P I N DA R I C M E T R E : T H E OT H E R H A L F

This page intentionally left blank

PINDARIC METRE:
THE OTHER HALF
K I I C H I RO I T S U M I

1
 3
Great Clarendon Street, Oxford ox2 6dp
Oxford University Press is a department of the University of Oxford.
It furthers the Universitys objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
# Kiichiro Itsumi 2009
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2009
All rights reserved. No part of this publication may be reproduced, stored
in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose the same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by SPI Publisher Services, Pondicherry, India
Printed in Great Britain
on acid-free paper by
CPI Antony Rowe, Chippenham, Wiltshire

ISBN 9780199229611

1 3 5 7 9 10 8 6 4 2
 Preface

Some understanding of metre is necessary for the full appreciation of

poetry. This conviction leads the study of Greek metre in two direc-
tions. One is towards research into the smaller units, i.e. verses and
phrases (cola or metra), while the other involves the interpretation of
the stanza-form (i.e. strophe/antistrophe and epode) as a whole. The
former raises the following questions. How should a verse be organ-
ically and logically analysed into phrases? How are the phrases related
to each other? Are the phrases common in the corpus, or not? How
do phrases tend to be combined to make up a verse? Finally, can
classes, major or minor, be established under which all the historic-
ally related phrases can be subsumed, together with material from
other poets also composed in the same tradition? The other direction
of metrical study raises similar questions, but in a broader perspec-
tive. How are the verses constructed in a stanza-form? To what extent
is each stanza-form unique in the corpus? Are there other similar
stanza-forms, and, if so, what are they?
This book aims to describe the metrical features of the twenty-two
Pindaric epinikia which are not composed in dactylo-epitrite (the
other half). Scholars currently assume, without examination in detail,
that all these epinikia are composed in a single type of metre which is
often called aeolic. Instead, I hope to prove that, if all the verses are
analysed in a consistent manner following objectively established rules,
there are in fact two types of metre, namely aeolic and freer dactylo-
epitrite. This is the main theme of Part I. The discussion is statistically
based on observation of all 235 verses of the eighteen major epinikia,
that is those in which the stanza-form is repeated several times (from
three to twelve times), so that the colometry can be established more
securely than in the four minor odes. Various metrical phrases are
classified according to the two types of metre. The various combin-
ations of phrases create such a multiplicity of verses as to enable each
ode to have its own peculiarities.
Nevertheless, Pindaric epinikia are not as polymetric as the odes of
tragedy. The metrical styles of the stanza-forms of the other half are
vi Preface

classified into three, according to the way in which the two metres,
aeolic and freer dactylo-epitrite, are knitted together. Strict defin-
itions of the three classes, based on the observations in the preceding
sections, are given in the final section of Part I. The differences in the
styles of these three classes are significant in the Pindaric corpus; it is
indispensable for the appreciation of each epinikion to notice to
which of these three classes its strophe and epode belong.
The second branch of metrical studies is pursued in Part II on the
basis of this classification. This Part consists of metrical commentar-
ies on all the non-dactylo-epitrite epinikia, both the major eighteen
and the minor four. The structure of the stanza-forms are each
analysed and compared. Metrical parallels are given in abundance,
both for the individual verses and for the stanza-form as a whole. In
general, I refrain from considering non-metrical questions, but in a
few passages textual problems are also discussed, for metrical study is
in part an auxiliary discipline of textual criticism. In particular,
metrical understanding is essential when one has to judge whether
or not exact responsion may be broken in a particular metrical
position. Thus in some places responsion between long syllable and
short may be admitted (as anceps), or responsion between long
syllable and two shorts (as resolution or biceps). In an Appendix to
this Part, the metrical features of the major fragments (most of which
are paeans) and their characteristics are discussed.
Part III consists of miscellaneous topics common to all the epini-
kia. It reveals some of the tendencies of Pindaric metre, with regard,
for example, to correspondence between repetitions, or the localiza-
tion of particular word-forms. Also discussed here are some of the
characteristics of dactylo-epitrite metre.
The germ of this work originated in the visiting fellowship granted
me by St Hughs College, Oxford, in 1989/90. Eighteen years have
passed since then. During these years St Hughs has given me hospi-
tality generously on every occasion I have visited Oxford. Any out-
sider would agree that one could not fully appreciate Oxford, the
paradise for classical scholars, without attachment to a college; but
what St. Hughs gave me is more than that.
At the initial stage, I believed in a single nature of non-D/e metre
(aeolic), and what I was planning to make was a kind of data-book.
However, the metre seemed so nebulous as to defy any classification.
 Preface vii
The major breakthrough occurred towards the end of the twentieth
century, when I realized that there were two different metres and
three different classes in non-D/e epinikia.
In different stages of evolution of this book very many friends not
only in Japan and Britain but also over the world gave me encour-
agement and assistance. They are too many to mention and I beg
their pardon if I skip most of their names. Dr B. K. Braswell gave me
many suggestions in his enviable study at Fribourg, and also sent me
a photocopy of Hohls dissertation. Professor Douglas Gerber was
the first person who accepted my idea of freer dactylo-epitrite and
propagated it in his paper, giving me confidence. Mrs Margaret
Howatson and Dr Patrick Finglass looked through the final draft,
and saved me from all kinds of errors. It is fortunate that Dr Leofranc
Holford-Strevens was my copy-editor. He scrutinized the draft, and
gave me acute comments. I now feel grateful especially for my two
old teachers: Professor Masaaki Kubo initiated me into the world
of philology when I started learning Greek poetry, and Professor
Elizabeth Craik introduced me to the worldwide scholarship by
assisting me to write papers in English on metrical study.
I cannot fully describe my debt to Dr Laetitia Parker. My gratitude
to her is immense. She gave me more than anyone would expect in
tutorial sessions. She read all my drafts word by word, corrected my
English, made my idea clearer by rewriting sentences, and discussed
every topic. It is absolutely true that, without her, I could not have
completed this book.
Last but not least I thank my wife Mutsuko. Eighteen years are so
long but she has incessantly supported my work and me.
Kiichiro Itsumi
Tokyo,
April 2008
 Contents

Symbols xii
Key to Terminology xv
Text xvii
Line-Numeration xviii

PART I . I N T RO D U C T I O N
1. The Other Half 3
2. Scope of the Investigation 8
3. Rules for Consistent Analysis 10
4. Two Metres in Non-Dactylo-Epitrite Stanza-Forms 21
5. Aeolic Phrases 24
A. The basic structure 24
1. The second position of reversed dodrans 24
2. Full base 25
3. Half-base 26
4. Catalexis 26
5. Prolongation 27
B. Classification by ending and frequency 30
C. The full-base group 31
1. The base qx 32
2. The base www 34
3. The base wq 34
D. The half-base group 35
1. Position of half-base in verse 36
2. Half-base ww 37
E. The first two positions of reversed dodrans 38
1. Reversed dodrans starting with qx 38
2. Reversed dodrans starting with www 39
F. Resolution 40
G. Acephaly 42
 Contents ix
6. Freer Dactylo-Epitrite 44
A. Basic structure and phrases 44
B. Acephaly 46
C. Resolution 48
D. Link anceps 52
7. Ambiguities between the Two Metres 60
1. Reizianum /x d x 60
2. Aeolic half-base ww /^ D 60
3. Wilamowitzianum /e w d 61
4. Aeolic full base wq /^ e 62
5. Pherecratean/rdod 1 63
6. Expanded aeolic 64
8. Composition of Verses 70
A. General 70
1. Demarcation of verses 70
2. Identical verses 72
3. Verse-length 73
4. Pendent and blunt endings 76
5. Ratio of short syllables (RSS) 77
6. Palindromic movement within verse 78
7. Repetition within verse 80
B. Aeolic verses: pure and composite 81
1. Contact between two aeolic phrases 81
2. Dovetailing 84
3. Verse-by-verse and phrase-by-phrase
construction 85
4. Aeolic phrases and d 86
5. Aeolic phrases and e 88
2
6. Aeolic phrases and e 90
3
7. Aeolic phrases and e 91
8. Aeolic phrases and ^ e 92
C. Freer D/e verses 92
1. Peculiarities of D and D 92
2. Verses containing two d 94
3. Short anceps and d (and a special
group of aeolic cola) 94
4. Verses with no double shorts 96
5. Double, triple, multiple e 97
6. e2 and other prolonged e 98
x Contents
7. d short anceps e 99
8. d e without intervening anceps 100
9. e d without intervening anceps 101
9. The Three Classes of Stanza-Forms 103
Appendix: Modern Emendations:
A Test Case (Nem. 6 str. 67) 111

PA RT II. THE EIGHTE EN MAJORS

Introduction 139
Olympian One 141
Olympian Two 154
Excursus A: Bacchylides 17 168
Olympian Nine 170
Olympian Ten 182
Olympian Thirteen (Strophe/Antistrophe vv. 15) 199
Excursus B: Bacchylides 3 204
Pythian Two 208
Pythian Five 220
Pythian Six 238
Pythian Eight 244
Pythian Ten 256
Pythian Eleven 267
Nemean Two 279
Nemean Three 281
Nemean Four 293
Nemean Six 300
Nemean Seven 315
Isthmian Seven 326
Isthmian Eight 333
Appendices 347
A. The Four Minors 347
Introduction 347
Olympian Four 349
Olympian Five 355
Olympian Fourteen 359
Pythian Seven 367
 Contents xi
B. Longer Fragments and Prolonged Ibycean 371
Introduction 371
Paean Two 373
Paean Four 378
Paean Six 384
Paean Nine (strophe only) 393
Partheneion One 396
Partheneion Two 398
Dactylic expansion revisited 399
Hephaestions Pindaric hendecasyllable (fr. 94c Sn.) 403

PA RT III. MISCELLANEA
A. Normal D/e epinikia 409
B. Tendency Concerning Bridge and Cut 426
C. The All-but-One Rule 431
D. Continuous Short Syllables 434
Addendum: Continuous Long Syllables 436
E. Verbal Assonance 438
F. Concentration of brevis in longo 441
List 1. Aeolic Phrases 443
List 2. Freer D/e Phrases 444
List 3. Identical Verses 446
List 4. RSS in Ascending Order 447
List 5. Bridge and Cut 449

Bibliography 455
Index 459
 Symbols

Aeolic phrases (for their relations with each other, see Part I, 5, A
and B)

adon adonean qwwqq

ar aristophanean qwwqwqq
dod dodrans qwwqwq
gl glyconic qwwqwq
gl 3 qwwqwqwqq
hag hagesichorean xqwwqwqq
hepta heptasyllable xqwqwwq
hepta 1 xqwqwwqq
hepta 2 xqwqwwqwq
hepta 2 3 xqwqwwqwqwqq
hepta 3 xqwqwwqwqq
hipp hipponactean qwwqwqq
ph pherecratean qwwqq
rdod reversed dodrans qwqwwq
reiz reizianum xqwwqq
tel telesillean xqwwqwq
tel 3 xqwwqwqwqq
wil wilamowitzianum qwqwwq
wil 1 qwqwwqq
wil 2 qwqwwqwq
wil 3 qwqwwqwqq
Freer dactylo-epitrite phrases
d qwwq
D qwwqwwq
 Symbols xiii
D qwwqwwqwwq
e qwq
e2 qwqwq
e 3
qwqwqwq
e 4
qwqwqwqwq
e 5
qwqwqwqwqwq
e 6
qwqwqwqwqwqwq
sp qq
^d wwq
^D wwqwwq

^D wwqwwqwwq

^D wwqwwqwwqwwq
^e wq
Phrases created by dactylic expansion (see Part I, 7, (6) and Part II,
Appendix B)
ibyc ibycean qwwqwwqwq

ibyc qwwqwwqwwqwq

ibyc qwwqwwqwwqwwqwq
ibyc qwwqwwqwwqwwqwwqwq
ribyc reversed ibycean qwqwwqwwq

ribyc qwqwwqwwqwwq

Phrase Names: n ( 2, 3) following a choriambic phrase wq

( 2), wqq ( 3); e.g. hepta 2 is xqwqwwqwq and wil 3 is
qwqwwqwqq. It must be remembered that hepta 2 or
wil 3 is a single phrase. The notation of n is introduced in order
to avoid further confusion by inventing new names or by using
ambiguous names like enneasyllable. The actual examples of n in
the eighteen majors are almost limited to 2 and 3, except for
one phrase of 1. However 1 (q), 4 (wqwq) etc. are occa-
sionally employed when they are useful to illustrate the structures
of hypothetical phrases.
xiv Symbols
Verse-numbering: P5s7 the seventh verse of the strophe/anti-
strophe of Pythian 5. O (Olympian), P (Pythian), N (Nemean),
I (Isthmian); s (strophe/antistrophe), e (epode). In the Excursus
and Appendices, I also use the following: Pae (Paean), Parth (Parthe-
neion), B (Bacchylides).
r the number of repetitions of a verse within the poem. So P5s7 8r
means that P5 contains a total of 8 strophes and antistrophes (4 strophic
pairs), so that the verse in question occurs eight times.
H (8, 16, . . . ) hiatus is attested at line(s) 8, 16, . . . Similarly, B
brevis in longo is attested, and HB both hiatus and brevis in longo
are simultaneously attested. Brevis in longo is abbreviated simply as
brevis when its meaning is manifest, especially in the combination
with hiatus, like hiatus/brevis.
 Key to Terminology

The eighteen majors is the name for all the non-D/e epinikia of
Pindar minus O4, O5, O14, and P7 (the four minors; I explain their
exclusion in Part II, Appendix A); that is: O1, O2, O9, O10, O13s15,
P2, P5, P6, P8, P10, P11, N2, N3, N4, N6, N7, I7, I8. All the data,
statistical and descriptive alike, for Part I are collected only from the
eighteen majors.
Stanza is used in its ordinary sense: each strophe or antistrophe
or epode is a stanza.
Stanza-form is used for the metrical scheme common to all the
stanzas of a particular form. Thus, all the strophes and antistrophes
of Olympian 1 have a single stanza-form, and all the epodes another.
O1s the strophes and antistrophes of O1, and O1e the epodes of
that poem. There is one exception: in O13, which changes metre in
mid-strophe from non-D/e to D/e, I treat the strophe/antistrophe as
two stanza-forms (O13s15 and O13s68).
Phrase is used as the nearest equivalent to colon which suits
Pindaric metres. See further, Part I, 3. It covers not only an aeolic
colon like glyconic but also a much shorter unit like e (qwq).
Single-short movement means a sequence of alternating short
syllable and long syllable ( . . . wqwqwqw . . . ). It may start with
either short or long. Every short syllable must be short in all the
repetitions, even though it may theoretically fill an anceps position.
Double-short movement means a sequence of alternating double
short syllables and long syllable ( . . . wwqwwqww . . . ). It may start
with either double short or long, but not with a single short. Double
short cannot be contracted into one long.
Leftwards and rightwards are used mainly with extension by
single-short movement. They mean that single-short movement
precedes (leftwards) or follows (rightwards) the choriambic nucleus.
Aeolic base: (Part I, 5, A). From the structural point of view, aeolic
phrases can be analysed as one of two basic asymmmetrical hexasyllabic
phrases, i.e. dodrans or reversed dodrans, preceded by aeolic base: (full
base, ), or by single anceps (half-base, x), or none (no base).
xvi Key to Terminology
n and zero-ending mean the number of positions following
choriambic nucleus. Thus, e.g. glyconic has plus-2 ending; Part I, 5, A
(5) and B.
Explicit anceps means an anceps position where a long syllable
and a short are actually found in responsion.
Implicit anceps means that the position is proved to be anceps by
the structure of the verse, but is always filled by a long syllable.
Bridge means absence of word-end at a particular place in a verse;
on the other hand, cut means occurrence of word-end. Bridge is
used in a looser, unorthodox manner in this book. People have used
bridge when word-end is avoided at a particular place in a metrical
pattern where all, or most, examples show the same tendency; e.g.
Maas, GM (1962), 47. My usage is simply descriptive as to whether
word-end occurs or not at a particular place in any given verse.
Example, when it is used in contrast with repetitions, means an
occurrence of verses which have a particular form. Thus, e.g. O9s3 is
an example of glyconic with the base qx and P2s2 is another, and,
counted with others, it is said that there are in total 11 examples of
this type of glyconics. An example is repeated several times within an
ode; so, glyconic in O9s3 is repeated in four strophes and in four
antistrophes. Repetition is used when we count every occurrence
individually; thus there are five repetitions of qw and three repeti-
tions of qq among the eight repetitions in O9s3.
The All-but-One rule means that exact responsion is broken only
once. Breach occurs by (a) anceps (short/long), (b) resolution (re-
solved/unresolved), (c) bridge/cut. All-but-One is not particularly
rare. See Part III, C.
 Text

My method is eclectic. I start with SnellMaehler (5th edn.) and

generally follow them. At the same time, I rely on Turyn for manu-
script readings and the ascription of emendations to mediaeval
scholars. My line-numeration is not Snells but Boeckhs; see below.
As for the emendations proposed by modern scholars since Erasmus
Schmid, I have been helped by Gerbers useful collection (Emend-
ations), although I have myself read and checked important editions
and papers, such as those of Hermann, Boeckh, Schneidewin, Bergk,
Mommsen, Christ, and Schroeder. The section of Textual Problems
in Part II deals with only the passages in which textual and metrical
questions matters are mutually entangled. I have no ambition to
solve cruces. This part of my book (in Slaters words) should not
be considered as an attempt on my part to improve the text of Snell,
but rather as a practical help to the reader to trace the history of
scholarship and to reach his/her own solution on the vital question
whether exact responsion should be pursued at all the repetitions or,
instead, the traditional text should be accepted as it is, by introducing
a metrical explanation, or licence.
The first occurrence of each stanza, that is, the text of the first
strophe and epode in each ode, is printed in Part II, to make the
colometry easier to grasp. I have not given an apparatus, since textual
problems are separately discussed.
 Line-Numeration

Line-numeration is, in a sense, a manifestation of an authors colo-

metry. Every editor seeks a rational system, but must at the same time
try to avoid confusion. I follow the numeration of Boeckhs major
edition. This is not only out of piety towards the scholar who first
gave the correct definition of a verse, but also for a practical reason.
Snells edition does not always reflect Boeckhs, in spite of its
marginal notation.1 In fact he follows Schroeder (BT), whose met-
rical theory has not found favour. When I disagree with Snell, not to
say with Schroeder, about colometry and divide his verse into two, or
combine two verses into one, amendment of Snells numbers would
add further confusion, since Snell himself occasionally uses numbers
such as 28b (in O1) or 7/8 (in P6). Turyn, in contrast, faithfully
adheres to Boeckh; when he adopts a different colometry, this is
reflected in his numeration, as 2b, which stands between 2 and 3
when Boeckhs 2 is divided, or 2/3 when two verses are combined into
one. I follow this system, but, of course, this does not mean that
I follow Turyns colometry. There are also important editions before
Snell, which any serious readers of Pindar will have to consult. The
line-numbers of these editions are various, each being different from
Boeckh (and, of course, from Snell).
At the same time Snells text has established itself as the standard;
I presume most readers of my book will consult SnellMaehler. His
numeration is in current use, for example, in Slaters Lexicon. Thus

1
Snells edition has to be used with care. It seems at first sight that we can use it to
restore the numeration of Boeckhs editio maior; Snell appears to have followed
Boeckhs editio maior in most odes because the notation Boeckhii is explicitly
printed in its left-hand margin in O1. When we have two numerations in O2, one
of which is in parentheses, it would be a natural guess that the one in parentheses was
the numeration of Boeckhs editio maior because the one not in parentheses is
indicated as Boeckh.2. In fact the notation in the left-hand margin is in a mess.
O2 and N6 are exceptions. Snell (actually Schroeder) without any notice diverges
from Boeckhs numeration in six non-D/e odes (O4, O9, O10, O13, P8, N3, I8). There
is even a case in which Snell becomes independent of Schroeder (I8; perhaps by a
simple mistake).
 Line-Numeration xix
for conveniences sake I occasionally add Snells numbering when it
differs from Boeckhs, like v. 25c ( v. 26 Sn.)
 , in I8. This
device is restricted to indication of a particular passage. When the
scheme of each verse is cited in Part I, Snells numeration is omitted.
Readers are kindly requested, if they are not sure, to consult the table
attached to the text in the initial section of each epinikion in Part II,
where Snells numerations are given in parentheses.
This page intentionally left blank
 Part I

Introduction
This page intentionally left blank
 1 . TH E OT H E R H A L F

The metre of about one-half of Pindars poems is easy to grasp. The

main features of dactylo-epitrite can be set out in half a page.1 More
space may be needed to investigate irregularities, but the basic sim-
plicity of dactylo-epitrite is undeniable.
The other half is still in obscurity. Recent commentaries keep a
cautious silence. The following passage, written forty years ago in an
obituary for A. M. Dale, a scholar of the greatest insight into Greek
metre, still remains valid: The metres of this group [ the other half]
used to be called logaoedic (which was almost a confession of failure
to understand them) and are now more commonly called aeolic,
because they seem to employ phrases familiar from the verse of Sappho
and Alcaeus. Yet how diVerently these phrases are used in the complex
organizations of Pindar and in the simple stanzas of Aeolic lyric!2
Since then, metrical studies have, of course, advanced. M. L. West,
developing the observations of Bruno Snell, introduces what I should
like to call dynamic analysis. After analysing an ode (P11) in aeolic
terms, he writes: This kind of analysis is necessarily of an ad hoc
nature (since the metrical scheme of each poem is unique), but it is
far from arbitrary. What we are trying to do in analysing these
musical paragraphs is to follow a train of thought. The basic ideas
are drawn from the common stock of metrical Wgures, but the poets
develop and embroider them in the course of composing a strophe,
producing sequences which appear bewildering when we look at
them in isolation and try to Wt labels to them, but which are easily
derived from what has gone before. Their etymology is more im-
portant than their deWnition.3
He invites us to pay more attention to the process of how a new
verse is generated from preceding verses and how it generates fol-
lowing verses than to the identiWcation of each verse. Thus we
should not hesitate to call e.g.


ww
ww
a mutant form
1
H. Lloyd-Jones, Pindar, Proceedings of the British Academy, 68 (1982), 138
( GELT 57), citing P. Maas, Greek Metre, trans. H. Lloyd-Jones (Oxford, 1962), 401.
2
R. P. Winnington-Ingram, Amy Marjorie Dale 19021967, Proceedings of the
British Academy, 53 (1967), 428.
3
GM 63.
4 Introduction
of glyconic when the phrase is better explained as derived from that
colon. To notate it as (gl) does not deWne the colon, [ . . . ] but it
interprets it.4 It follows that identical phrases may be explained in
diVerent ways, depending on the metrical context.
Certainly this method has merits. We need not torture a verse so as
to chop it up into accepted cola nor introduce multifarious, esoteric
names almost for their own sake. But at the same time it does not
make clear what forms are usual and what are not. It may embrace
curious forms, and we are left ignorant of the extent to which they are
curious. In view of the fact that metrical study is an auxiliary
discipline of textual criticism, this type of analysis may be counter-
productive.5 It is generally acknowledged that the text of Pindar is
less corrupt than those of dramatic poetry. However, there still
remain some passages in which a textual problem entangles itself
with a metrical problem: the most complicated case is N6s67.6
Snell, whom West follows, was not the only scholar to seek a uniWed
point of view for all the other half of Pindar. Dale herself proposed a new
system of metrical notation to avoid the danger of dividing a verse into
cola inorganically.7 If we set down a period in longs and shorts
and succeed in dividing these up into anapaestdochmiac

4
GM 64; (gl) is Snells symbol for a phrase which is not exactly glyconic but a
related form.
5
I do not wish to suggest that an interpretation of this type is useless. But each verse
should be explained from its relations not only with other verses within the ode in
question but also with verses outside the ode in the Pindaric corpus (and the Bacchy-
lidean, and any other Greek poetry). The two analyses are not mutually exclusive. In fact
I shall occasionally follow a train of thought (West, GM 63) in analysing each stanza-
form in Part II, although my analysis is not always the same as that of SnellWest.
6
Many eVorts have been made to restore the text and the metre. The metrical
forms assumed by scholars are so diverse that proposed emendations are abundant.
See Appendix.
7
Dales paper, The Metrical Units of Greek Lyric Verse was published in 19501
(CQ 44 and ns 1) before the Wrst edition of Snells Pindar (1953), which had in fact
been completed in 1943 (see the Praefatio). Since Dale included him in her acknow-
ledgements, I suppose that there was some interdependence between them. Willy
Theiler, too, tried to introduce a uniWed scheme for various metrical phrases and
invented for them new names which were free from traditional connotation (Zeit-
stufen). Theiler was a colleague of Paul Maas in Konigsberg, and it was by Snell that
Maas was literally rescued from the Third Reich. The metrical judgement of these
three (Maas, Theiler, Snell) have something in common. They, as well as Dale, stood
together against the idiosyncrasy of Schroeder, who was, in a sense, a successor of
Wilamowitz. Cf. T. Gelzer, MH 53 (1996), 281305.
 The Other Half 5
reizianum, this is worth no more as an elucidation than to call

w
ww
w
trochaiciambic metron.8 She warns us repeat-
edly of the pitfalls caused by inorganic dissection. Instead, she re-
commends describing a verse as a whole. Her newly proposed system
of description free from accustomed labels, is made up of symbols
like d(
ww
) or s(
w
), which are modelled on the Maasian
symbols for the D/e metre, D (
ww
ww
) or e (
w
). According
to her system, for example,
w
ww
w

w
is described as
sdsjs.9 Recently the system has been revived by Sicking.10
The system exempts us from controversies, sometimes futile, be-
tween diVerent colometries. However, metrical studies do not end at
describing a particular phrase like sds. A phrase should at least be
compared with other phrases. And even comparison inevitably en-
tails interpretation, to say nothing of further analysis from classiWca-
tion to poetic appreciation.
Dale seems to have developed the idea of the ds system from the
premise that some parts of the other half of Pindaric metre can and
should be analysed in the same manner as D/e. That means that the
concept of link anceps is indispensable for the analysis of some
verses. For example,
N6s6a
ww
ww


ww
rw

k
is easily analysable in D/e terms: D
d e
. This analysis is much
more natural than that of Snell: ^ chodim ^ glr sp. In a sense, intro-
duction of the concept of link anceps was a revolution in the history
of metrical study. Without it, the various phrases of the D/e had been
given multifarious names depending on where the intervening anci-
pitia were attached so that the structural simplicity of the metre was
obscured.11 For example,

8
Collected Papers, 467.
9
The mutant glyconic of West (


ww
ww
, see above) would be
designated either
sdd or

dd according to how it is interpreted. The ds-system is
not as neutral as it sounds.
10
Sickings method is duly criticized by West, GGA (1994), 18397.
11
I am not sure about Wests comment: It should be appreciated that this [ the
Maasian system] is merely a convenient method of notation, not an etymological
analysis. The link-syllable is a false concept as far as the process of creation is
concerned. Greek poets compose with cola and need no mortar to join them (GM
70). Is this not an a priori argument?
6 Introduction
O3s1
ww
ww


w


ww
ww
j D
e
D
was an encomiologicum catalectic (D
e)prosodiac dimeter
catalectic (
D) according to the metrical scholia, while
O3e1
w
x
w


ww
ww
j e xe
D
was a trochaic dimeter acatalectic (e x e
)dactylic penthemimer
(D), and
O3s4

w

j
e

an iambic penthemimer. However, it is a pity that Dale went on
applying the idea much more widely to other metres too. She even
tried to notate unequivocal aeolic verses in the same manner. If she
had not ventured to analyse all Pindaric verses on a single basis, she
would have anticipated what I shall call freer dactylo-epitrite.
At worst, Dales system discards any classiWcation. Thus her idea
and method can be said to have paved the way for the position which
West reaches by a diVerent route thirty years later, even though Dales
sds looks very objective, while Wests (gl) looks very subjective, and
West continues to use the appellation aeolic while Dale abandons it.
Glyconic and sds are not equivalent, because sds does not allow
for the phenomenon of aeolic base. It allows only for base in the form

w. The name glyconic does not simply trace the sequence of longs
and shorts. It implicitly suggests which position may be anceps, at
which position a word-end tends to fall, or to which metre the colon
may be related. It represents a history from Lesbian poetry, via
Anacreon, to dramatic poetry (and further), and also the accumu-
lated knowledge of ancient and modern scholars over centuries.12
If we try to observe every Pindaric verse in detail, and try to set
Pindaric metre in historical perspective, classiWcation is still indis-
pensable. A more common colon has to be distinguished from a less
common. Where there is a variant, the nature of the variation must
be deWned. In other words, a traditional, static analysis does not lose

12
As to the glyconic, I am sceptical about the views of Hephaestion and other
ancient scholars. It was Hermann in 1796 (De metris poetarum Graecorum et Roma-
norum) who rightly discarded the antispast and gave the correct deWnition of
glyconic, starting with aeolic base.
 The Other Half 7
its value. We should not prematurely give up hope that a consistent,
objective colometry may be available for Pindaric metres as well as
for other Greek poetry. Pindaric metre is certainly complicated, but it
is not so complicated that one can conclude that any and every
permutation of longs and shorts makes a metrically meaningful
pattern.
8 Introduction

2 . SC OP E O F TH E I N V E ST I G AT I O N

There are 44 Pindaric epinikia, counting Isthmian 3 and 4 as a single

unit (14 Olympians, 12 Pythians, 11 Nemeans, 7 Isthmians).13 Of
these 23 are made up of D/e. D/e odes can be separated from non-D/e
ones unambiguously. There is only one exception: Olympian 13,
which changes metre midway in the strophe from non-D/e to D/e.
No shift of metre is found elsewhere.14 In all the other 22 odes, O13
being counted twice, the strophes and antistrophes are completely
made up of D/e from beginning to end without any trace of other
metres.15 And so are their epodes. Moreover, the metrical identity is
kept throughout each poem. Every triadic ode whose strophe and
antistrophe are in D/e has its epode in D/e. In the D/e poems,
therefore, there are a total of 44 stanza-forms (on the term, see
p. xv).16
So much for the D/e odes. In the course of this study, I propose
to demonstrate that the non-D/e odes fall into three classes: Class I,
aeolic; Class II, freer D/e; Class III, amalgamated. For the moment,
however, I treat all the non-D/e odes together. There are 22 non-
D/e odes (16 triadic, 5 monostrophic, including the Wrst half of the
strophe/antistrophe of O13), which produces a total of 38 stanza-
forms. Of these, four minor odes, Olympians 4, 5, 14, and Pythian
7, are excluded from consideration at the Wrst stage. Since my
argumentation is cumulative, relying on parallels and statistics,
without frequent repetitions it is diYcult to decide whether
or not a verse-end is to be recognized in a particular position.
Accordingly, minor odes are unsuitable for this study, because

13
I avoid the question whether I3 and I4 are a single poem or not. Metrically, the
stanza-forms of I3 and I4 are identical.
14
There is one ode in Bacchylides which is made up both of non-D/e (the strophe)
and D/e (the epode): Ode 3. But there are some similarities between its two stanza-
forms. See further Part II, Excursus B.
15
Some irregularities are included in O6s5, O7e3, N8s1, N10s1; see 7. 6 below.
16
Of 23 odes made up of D/e (including O13), 21 are triadic and 2 are mono-
strophic (P12, N9); 2
212 44.
 Scope of the Investigation 9
they are too short to establish a certain division into verses.17 For
the same reason, the fragments too are excluded.18 Hereafter the
name the eighteen majors will be used, meaning O1, O2, O9,
O10, O13s15, P2, P5, P6, P8, P10, P11, N2, N3, N4, N6, N7, I7,
I8. Of these P6, N2, N4, I8 are monostrophic. Incidentally, the minor
odes being excluded, each stanza-form of the eighteen majors is
repeated not less than three times (the minimal cases are the epode
of N6 and that of I7).

17
They consist of one triad or one strophic pair only. O5 has three triads
(nevertheless the strophe and the epode are very short indeed), but its authorship
is questionable. The main reason for suspicion is a note inserted in the scholia
(K b K P q, although K
 is admittedly a strange word) but its
peculiar metre increases suspicion. The metre of each of these four odes will be
discussed later in Part II, Appendix A.
18
The longer fragments (Paeans 2, 4, 6, 9, and Partheneia 1, 2) will be analysed in
Part II, Appendix B.
10 Introduction

3. RU LES FOR CONSISTENT ANALYS IS

The strict deWnition of a verse (or period) is one of the basic

principles of modern metrical studies. SigniWcantly, it was from the
observation of Pindaric verses by August Boeckh that the correct
deWnition of verse-end was Wrst recognized. In the sung parts of
dramatic poetry, tragedy and comedy alike, verse-end is not always
clear. Its manifest proof, i.e. invariable coincidence of word-end
occasional hiatus/brevis, is not necessarily present, because the same
metrical scheme occurs only twice. Pindaric metre is diVerent.
Thanks to many repetitions, the beginning and the end of a verse is
Wrmly established in most cases. As a result, texts of Pindar have
come to be laid out in verses,19 while the texts of Attic drama tend to
be laid out in cola, that is to say, more or less well-deWned metrical
phrases, such as glyconic, iambic dimeter, etc. I shall generally avoid
the term colon and colometry, since cola in Pindar are often very
hard to demarcate, as we have already seen in D/e. Instead, I use the
term phrase, even for such familiar cola as glyconic and pherecra-
tean. It is important to note that phrases, however short or unfamil-
iar, cannot be combined in deWance of verse-end in order to make a
familiar colon.
The total number of verses in the eighteen majors is, by my count,
235. Of course, every scholar will count slightly diVerently. Judge-
ment will diVer in cases where hiatus/brevis is absent although word-
end is coincident at all the repetitions. There are also cases which
involve the emendation of the text. My calculation diVers from
Boeckhs as follows: I divide thirteen of Boeckhs verses into two,
and one verse into three; and there are 10 verses newly created by
combining two of Boeckhs into one. The shortest verses I found are
P5s7a, N6s1a, and N6s4b, which are composed of only three posi-
tions (for detailed discussion of the analysis of these verses, see Part
II, ad locc.):

19
I follow L. P. E. Parker, who always uses verse instead of period. The advan-
tages are that one avoids problems created by Dales major periods and minor
periods and that the term verse is internationally comprehensible.
 Rules for Consistent Analysis 11



k (P5s7a, 8r, on the notation, see above, p. xiv)
w

j (N6s1a, 6r)

w
k (N6s4b, 6r)
One position longer is:

wt
k (N6s6b, 6r)
which is preceded by unambiguous verse-end. Incidentally, this is the
same length of the shortest D/e verse:


w
k (O7s3, 10r)20
The longest verse in the eighteen majors is N2s4 (5r, 23 positions).
Its structure is simple, and naturally divided into three aeolic phrases,
namely, two glyconics and one pherecratean:
rw
ww
w



ww
w

x
ww

k
But not all the verses are segmented in such a clear, self-evident
manner. There have been controversies over diVerent analyses, and
controversy needs to be eliminated as far as possible so as to advance
static analysis. The Wrst step is to recognize in a consistent manner
the identical phrases that are dispersed in various stanza-forms.
Consistency is required not only for the analysis of the identical
form, but also for similar, structurally related forms. For instance, if

x
ww
w

ww
k (I8s6, 7r)
is divided into

x
ww
w
j
ww
k gl d
then
www
ww
w

w
k (P2e5, 4r; P11e4, 4r)
should be seen as
www
ww
w
j
w
k gl e

20
Turyn, dividing N1e1 into two verses, accepts
w
([N1e1b] 4r) as an
independent verse. I am inclined to unite [e1a] and [e1b]. Metrical lengthening of
 at the end of [e1a] is probable. See further Part III, A.
12 Introduction
The two verses are fundamentally in the same structure: glyconic
choriamb (d) or cretic (e).21 Every scholar would agree with this
division. Remarkably, a boundary is recognized between two (true)
longs:
. . . w
j
ww

. . . w
j
w

Consistency is, however, broken by both Turyn and Snell when the
division creates unfamiliar cola. Turyn expels all aeolic cola longer
than the glyconic (blunt ending) or hipponactean (pendent). He
does not set boundaries in the same manner as the cases above
between two longs in these two longer verses despite the relationship
of their structures:
N4s1, 12r x
x
ww
w
j
ww
k
N3s7, 8r
w
w
ww
w
j
w
k
He divides thus instead:
x
x
ww
j w

ww
k {dim chori Maecen}22

w
w
ww
j w

w
k {glyc do}
Snell diVers from Turyn, but he is inconsistent nevertheless. Snell
recognizes the similarity between N3s7 and those above:

w
j w
ww
w
j
w
k {cr ^ gl cr}
but on N4s1, he agrees with Turyn:
x
x
ww
j w

ww
k {^ chodim (^ chodim)}
Perhaps it may be said that Snell is even more inconsistent than
Turyn: Turyn has a principle at least, but Snell has none. When he
does not divide N3s7 in Turyns fashion, it does not mean that Snell is
conscious of the structure: he simply dislikes dochmiac (w

w
)
in Pindar. There is an obvious inclination common to Turyn and

21
In this chapter, I shall append abbreviated names of phrases to metrical
schemes. These names are given simply for the sake of convenience and will be
discussed in later chapters. It is more important at the moment to understand how
a verse should be divided than to classify phrases.
22
Metrical notations in { } mean that the analysis is oVered by Turyn or Snell (and
accepted by West), but rejected by me.
 Rules for Consistent Analysis 13
Snell: both are eager to mark out a familiar colon inside a verse, and
fail to recognize the fundamental structure shared by many others.
We need rules based on principles. A series of rules is indispensable
for reference when a verse seems to include ambiguities and more
than one analysis is conceivable. Established rules will demonstrate
why Turyns and Snells analyses are not always adequate.
First of all, we can and should agree not to set phrase-boundary at

j w
as Turyn and Snell occasionally do (see above). Instead,
boundary should always be set between true longs even if cola of
unusual lengths, or of unfamiliar forms, result before and/or after the
division. This principle is the Wrst step to guaranteeing objective
consistency.
The same principle can also be applied to double short. We should
not set phrase-boundary at
j ww
. Any length of dactylic run (or,
in other words, double-short movement),
ww
ww
[ww
ww . . . ],
is an indivisible unit. On the other hand,
. . .
ww

ww
. . .
must have phrase-boundary
. . .
ww
j
ww
. . . 23
When double short comes back after the intervention of short or
anceps, or after single-short movement, it is natural to suppose that
the sequence must have division somewhere between two
ww
.
For example, everyone would agree, following the principle above,
that there is a phrase-boundary between two true longs in the
following:
. . .
ww
w
j
w
ww
. . .
When an anceps intervenes, there are, in theory, two possibilities:
(i) . . .
ww
j x
ww
. . .
(ii) . . .
ww
x j
ww
. . .
I suppose that the majority of scholars would prefer (i) to (ii). In fact,
as will be demonstrated below, (ii) makes it possible that a catalectic
phrase like a pherecratean could stand in the middle of verse, which

23
There is no pure sequence . . .
ww

ww

ww
. . . in Pindar.
14 Introduction
is contrary to the modern idea of catalexis (see below). So (i) should
be accepted as a matter of principle. Thus I adopt two principles for
determining phrase-boundary: a boundary should be automatically
and exclusively placed:
(a) between true longs;
(b) before anceps Xanked by two longs (
j x
).
Most boundaries, perhaps as many as 90%, can be determined in this
way. Based on these two principles as well as others, the following
rules are established.
Rule 1. By (a), two juxtaposed
ww
with no position between
them belong to separate phrases.
P10e3, 4r x
w
ww
j
ww

w
k hepta d e
This means that so-called choriambic expansion of the type found
in asclepiadic cola is excluded. Snell gives the notation ^ glc w
to
P8e5, 5r

ww

ww
w
w
j
assuming it to be an expanded colon of

ww
w
w
(^ gl w
).
But I divide it after

ww
:


ww
j
ww
w
w
j
d d w e
Rule 2. Any length of dactylic run
ww
ww
[ww
ww . . . ] is an
indivisible unit:
ww
ww
and
ww
ww
ww
are taken to be
of the same nature as
ww
.24
So, where Snell and West divide P11s1, 8r (admittedly an unusual
verse) thus:


w
ww
j ww
ww
w


k {^ chodim ^ gl sp}
I should divide thus:


w
ww
ww
ww
w
j

k
ribyc 2 sp
It is an old device, going back to the Alexandrian editions, to cut a
dactylic sequence so as to produce a colon of a length deemed

24
No examples longer than three consecutive double shorts are found in the
eighteen majors.
 Rules for Consistent Analysis 15
appropriate, starting with ww.25 However, this division is unaccept-
ably artiWcial, especially when word-end is not found at the alleged
phrase-end.26
Rule 3. A position Xanked by two long positions, or by verse-opening
and a long position, is, if not short, assumed to be anceps.
This means that long is never to be understood as contracted
double short. For example, it is highly improbable that in the latter
half of
O9s6/7, 8r www
a
ww


ww

k
the sequence
ww


ww
is dactylic, with the middle long
contracted. As the context shows, the verse must be analysed as
wilreiz. A similar analysis applies to
N6s6a, 6r
ww
ww


ww
rw

k D x d e x27
Rule 4a. By (a), when two long positions are juxtaposed, both being
within the verse, a phrase boundary is always assumed between them.
N3e2, 4r
w
ww
j
w
ww

k rdod ph
I8s6, 7r
a
ww
w
j
ww
k gl d
P6s5, 6r
ww
w
j
w
w
ww
j dod wil

25
I believe that the division was not based on metrical principles; it was simply a
device for lay-out. There is a possible analogy here for the divisions. Cf. L. P. E. Parker,
CQ, ns 51 (2001), 48, and Itsumi, Whats in a Line?.
26
The case for the colon ww
ww
w
in mid-verse rests on P2s4 and N3e4,
which are exceptional; see below, 8. A. 7. The descriptions and statistics in Part I are
based on the following colometries for these two verses:
N3e4 ww
ww
w
j ww
ww
w

w
k tel tel e
P2s4 ww
ww
ww
w
j ww
ww
w
k (teld) tel
However, examination of expanded aeolics (Part I, 7. 6) will prove diVerent colome-
tries to be valid:
N3e4 ww
ww
j w
ww
ww
w

w
k ^D w ibyc e
P2s4 ww
ww
ww
j w
ww
ww
w
k ^D w ibyc
These divisions strictly follow Rule 5.
27
Theoretically, contraction of two shorts features in anapaestic as well as dactylic,
but anapaestic is a well-deWned metrical type admitting both t
t
and

t
t, and is not found in Pindar. Turyn Wnds sporadic anapaests in Pindar
(this analysis goes back to Wilamowitz), but following Maas and Dale we now have
clearer conceptions of metrical typology.
16 Introduction
This means that divisions which produce pherecratean (and other
pendent cola) within a verse, such as
w
ww

j w
ww

k
(Snells analysis of N3e2), are avoided. Whether or not catalectic
aeolic cola are permitted in synartesis (except under special condi-
tions such as repetition of the same colon) is controversial,28 but the
clausular function of catalectic cola is not in doubt.
This rule sometimes creates unfamiliar phrases, as has been illus-
trated above. One example is:
N4s1, 12r x
x
ww
w
j
ww
k hepta2 d
Its Wrst phrase is one position longer than the glyconic. However, it is
paralleled:
P11s5, 8r u
w
ww
w
j
w
w
k hepta2 e2
According to the colometry of Turyn and Snell, an aeolic colon,
w

ww
, turns out after the supposed division. Besides the neg-
lect of structural similarity (see above), the colometry has another
irregularity: w
as aeolic base is always preceded by word-end in
tragedy.29 As for P11s5, w

w
w
, which both Turyn and Snell
analyse as baia, is not to be favoured because the baccheus is
situated in mid-verse. For parallels for
w
w
, see below.
Rule 4b. The above Rule 4a also applies to two longs one of
which is the initial, or the Wnal, element of a phrase in single-short
rhythm:

28
e.g. Eur. Ba. 10510  1205 is naturally arranged thus:

ww
w



ww
w


wwwwwww



ww
w



ww
w
w


ww
w


The Wrst colon and the second are in synartesis in str. and in ant. (Bj 
j). For other examples, see Parker, Catalexis, CQ, ns 26 (1976), 1428
(esp. 205; ead., Songs, 9; and cf. Stinton, Pause and Period, CQ, ns 27 (1977), 3941
Collected Papers, 3258.
29
See Itsumi, Glyconic, 80. In Lesbian metre, iagl is occasionally realized as
w
w
j w

ww
w
; for example, Alcaeus 70. 10 LP.
 Rules for Consistent Analysis 17
P2e5, 4r www
ww
w
j
w
k gl e
O13s2, 10r a
w
j
w
ww

k x e ph
O1e6a, 4r

ww
w
j
w
w
k tel e2
Turyn occasionally sets a colon boundary before the Wnal jw


w
k instead of j
w
k, like the other case discussed above:
N3s7, 8r
w
w
ww
j w

w
k {glyc do}
As I shall argue later, w

w
cannot be entirely expelled from
Pindar, though it is inappropriate to call it dochmiac.30 N3s7 should
be analysed as

w
w
ww
w
j
w
k wil2 e
This is more consistent with other, structurally related forms, such as:
P8s5, 10r
ww
j
a
w
ww
w
k d wil2
Similarly, a longer phrase,
w
w
(e2), occasionally appears at
verse-end (see above, P11s5 and O1e6a). Snell expels anything that
looked like a dochmiac (current convention calls this phrase hypo-
dochmiac; Turyn simply calls it dochmiac) so that he analysed, for
example,
P2e3, 4r w

ww
w
j www
ww

j w
w
j {gl pher ia}
The phrase
w
w
must, however, be accepted. Pherecratean in mid-
verse entails abandoning the established notion of catalexis (see above).
Rule 5. By (b), when an anceps appears between two
ww
, phrase-
boundary is set immediately before the anceps. Even if anceps is not
explicitly realized as anceps (in other words, even if long and short
syllables are not actually found in responsion at the position in
question), the position is taken as anceps; cf. Rule 3.
N4s5, 12r
a
w
ww
j a
ww
w
k wil tel
O9s6/7, 8r www
a
ww
j

ww

k wil reiz
N4s6, 12r
a
a
ww
j w
ww

k wil reiz
Rule 6. By (b), when three positions, of which the middle one is long,
appear between two
ww
, if (i) either of the outer two is evidently
anceps (x
w or w
x), phrase-boundary is set immediately before
30
The phrase occurs only at the beginning of verse, and is analysed as ^ ee.
18 Introduction
the anceps; if (ii) both the outer two positions are short and it is not
possible to decide which is anceps, phrase-boundary is set immedi-
ately before the second short.
(i) O9s3, 8r
a
ww
w
j u
ww

k gl reiz
I7s5a, 6r


ww
w
j

ww
w
j u
w
ww
j gl tel
hepta
(ii) N4s4, 12r
w
ww
w
j w
ww

k gl reiz
In the eighteen majors there is no example of . . .
ww
u

u
ww
. . . with two explicit ancipitia. The sequence of positions . . .
x
x . . . is theoretically possible only in the wilamowitzianum
(
x
x
ww
) and heptasyllable (x
x
ww
). But in the eighteen
majors actual conWrmed examples of u
u even in these cola are scarce.
To assume phrase-boundary at
. . .
ww
w
j u
ww
. . .
and
. . .
ww
j u
w
ww
. . .
is natural, and no problem. In the case of
. . .
ww
w
w
ww
. . .
the choice between two possible positions is, in a sense, a matter of
taste. I prefer
. . .
ww
w
j w
ww
. . .
to
. . .
ww
j w
w
ww
. . .
because I assume that glyconic with its correlatives is the standard
aeolic phrase.31 Thus, N4s4 (above) is analysed as glreiz, not as

w
ww
j w
w
ww

k rdod hepta1
Rule 7. In spite of Rules 5 and 6, when the phrase after the anceps is
very short (
ww
or
w
), anceps is treated as a link, in the
manner of D/e.

31
Actually the sequence . . . ww
w
w
ww . . . is not common. There are 4
examples, collected at 8. A. 6. (b).
 Rules for Consistent Analysis 19
N3e1a, 4r
ww
jwj
ww
jaj
w
j
j dwdxex
P10s4, 8r w
j
ww
jaj
ww
j
w
ww
j
w
w
k ^ e d x d rdod e2
P8s7, 10r aj
w
jaj
w
w
k x e x e2
The concept of link anceps is desirable in particular when dealing
with the sequence . . . x
w
x . . . in order to avoid the suggestion of
iambic or trochaic movement.
The phrase x
ww

k remains ambiguous. It may reasonably be
taken as an aeolic colon, reizianum, i.e. the catalectic form of the
telesillean. At the same time, it is also possible to take it as
ww
(d)
preceded and followed by link anceps. Judgement depends on the met-
rical context. In practice, there are limitations on the use of link anceps.
It is, for example, highly signiWcant that phrases with long anceps, like


ww
,

w
,

w
w

are hardly ever preceded by an aeolic phrase. See below 8. B. 4 (x d); 5
(x e); 6 (x e2).
Rule 8. A prolonged sequence in single-short movement is not
divided into shorter phrases.
P2s1, 8r wwwwww
w
w
wwww
k e6
O1s6, 8r
w
wwww
w
w
jwj
ww
j
w
w
k e5 w d e2
N3s5, 8r

w
w
wwwjwj
ww
k x e3 w d
O1e3, 4r
w
w
jwj
ww
j
w
j e2 w d e
It is vital to recognize that these are sequences in which no long
syllable appears between longs in any repetition of the verse. Thus it
makes no sense to try to decide which of the shorts is theoretically
anceps. So I notate these sequences by en, so that e3, for example,
represents
w
w
w
. Cf. the treatment of sequences of double-
short movement in Rule 2.32
These are Rules.33 Before we go on, one curious fact deserves
notice. From the point of view of metrical study, it is unfortunate
that the traditional order of the poems sets the Olympians Wrst, since

32
Where a sequence begins with a short syllable, the notation ^ en is theoretically
possible, but it seems to me more natural to notate wen1. See below, 6. B.
33
I Wnd that my phrasing is basically identical with that of most 19th-c. scholars
before RossbachWestphal launched a new theory. The metrical charts attached to
20 Introduction
they include clusters of very unfamiliar verses. If we were to start, say,
with the Nemeans, our impression of non-D/e might be rather
diVerent. For reasons of general convenience, however, I present
the poems in the traditional order.

the text in the editions of Boeckh, Dissen, Schneidewin, Bergk, and Mommsen, all
analyse in the same manner. For example, O1s runs:
wwwlwllx wwwlu
wr wlwwlwwlwwlu
wlwlwl
lx wwwlwu
wlwlwu
wlwrwlwlwlwwwlwlwu
wlwlwlwwwlwl
wrwr
lwlwlwu
wt wlwwl
wrw lwlwu
Note the ictus put on the long (or resolved long). It may be said that the long with
ictus indicates the beginning of a new phrase. When the phrase starts with short, ictus
falls on the following long. When the phrase starts with aeolic base, the Wrst position
x , and the ictus falls on the Wrst true long in this case.
of anceps is indicated as (e.g.) l
Then the verses in O1s are analysed in the following manner:
wl lwwlwl lwlwwlu
wr lwl lwwlwwlwwlu
lwlwlwl
lwlwwlwu
lwlwlwu
lwlwrwlwlwl wlwwl lwlwu
lwlwlwl wlwwl lwl
wrwrlwlwlwu
wr lwl rwl
wl rwlwlwu
The diVerence between this and my analysis based on the Rules above, which will be given
in Part II, is not essential; I take wl lwwlwl as one phrase (glyconic), and wlwwl as
link anceps d. Rather, I should like to stress that the fundamental idea of Pindaric
colometry had been free from arbitrariness in the 19th c. until RossbachWestphal
introduced a false theory. Wilamowitz and Schroeder rightly denied RossbachWest-
phals theory but their analyses introduced another confusion as Snell and Turyn show.
 Two Metres in Non-D/e Stanza-Forms 21

4 . T WO ME T R E S I N N O N - DACT Y LO - E P I T R I T E
STANZA-FORMS

Snell and Dale, following Maas, introduced into the study of Greek
metre a much clearer and stronger conception of metrical typology.
Earlier, it was usual to take these poems as polymetric. Turyn, for
example, chooses labels for each constituent within a single verse
from such various classes of metres that, to the generation after Snell
and Dale, the analysis of a stanza looks like an assorted box of
chocolates.34
Dales ds notation (treated above, I. 1) seeks to exclude consider-
ations of metrical typology, and is, for that reason, comparatively
uninformative. Snell seeks to analyse all the non-D/e odes as aeolic.35
But here he goes too far. His uniform aeolic analyses produce phrases
for which there is no analogy in aeolic as we know it from the Lesbian
poets, Anacreon, and Attic drama. Even a phrase which has no
double short is included in aeolic by Snell,36 which would, inciden-
tally, be given by West the charming designation Wlleted glyconic.37
Moreover, because the D/e poems are metrically uniform, or very
nearly so, it tends to be assumed that the non-D/e poems are also
uniform. This leads to the conclusion that, since some poems are
clearly in aeolic metre, all must be.
According to my analyses, two types of metre are used in the non-
D/e poems: aeolic and what I shall call freer D/e. Out of all the 235
verses of non-D/e odes 101 verses (43%) are freer D/e (see below, 8.
A. 1). Also, these two metres can be combined in the same poem.
Aeolic is a general term for the metre whose characteristics are (1)
the asymmetrical units
ww
w
and its reversed form
w
ww

34
In general, Turyn follows Wilamowitz, but improves his analyses, and takes
great care to be consistent in labelling each constituent of a verse. If his analyses are
looked at closely ode by ode and colon by colon, he will be found more careful than
on Wrst impression.
35
To be strictly correct, both Snell and West treat the metre of O2 separately from
the rest.
36
e.g. x
u
w
at P8s7 and P10s6 are given the notation (^ gl).
37
West invents the phrase for the description of P11s5: at the same time the
iambic part w

w
w
appears as a Wlleted glyconic, with
w
taking
the place of
ww
(GM 63).
22 Introduction
and (2) absence of link anceps between phrases. When an aeolic
phrase is preceded or followed by e or d, it is included in aeolic in
the wider sense. If necessary, diVerent appellations, i.e. pure aeolic
and composite aeolic, are used. Freer D/e is a general term for the
metres which are composed essentially of (1) the symmetrical unit(s)
e (
w
), d (
ww
), and/or their enlarged forms e2 (
w
w
),
D (
ww
ww
), and their acephalous counterpart (^ e, ^ d, ^ D)
with/without (2) link anceps either between phrases or at the be-
ginning/end of the verse (its diVerences from the normal D/e will be
illustrated below). There are, however, some ambiguities between
aeolic phrases and freer D/e (see below, I. 7).
O9s19 is a good example of pure aeolic:
O9s1 ww
ww
w
k tel
O9s2 u
ww
w
j
u
ww
w
w

k tel gl3
O9s3
a
ww
w
ju
ww

k gl reiz
O9s4


ww
w
j

ww

k gl reiz
O9s5
a
ww
w
j

ww

k gl reiz
O9s6/7 www
a
ww
j

ww

k wil reiz
O9s8
u
ww
w
k gl
O9s9

ww

k reiz
The phrase-by-phrase construction is clear enough. Every phrase is
familiar, and the combinations are natural. The similarity with the
aeolics of tragedy or comedy is obvious.38 But the aeolics of Pindar
deviate from the mainstream of Greek metre in that such phrase-by-
phrase structure is rare. Rather, common aeolic phrases are very
frequently connected by shorter phrases, especially, by e (
w
),
and, to a lesser extent, by d (
ww
). This characteristic is well
illustrated by P6s (I cite the whole stanza):
P6s1/2 wj
w
jwww
ww
w
j
w
w
ww
k w e gl wil
P6s3 www
ww
wyj
w
k gl e
P6s4 ww
j
w
w
ww
k ^d wil
P6s5
ww
w
j
w
w
ww
j dod wil

38
But it must be observed that the following two verses, s10 and s11, are not given
here. O9s10 is a composite type of aeolic, illustrated in the next paragraph, and O9s11
is freer D/e. Moreover, most verses of the epode of O9 are not aeolic.
 Two Metres in Non-D/e Stanza-Forms 23
P6s6 www
ww
w
j
wrj
k gl e

P6s7/8 wj
w
j


w
ww

k w e wil1
P6s9 w
j
w
w
juj
w
k ^ e e2 x e

Note that the last verse (s9) is, unlike the preceding verses, composed
exclusively of shorter, symmetrical phrases.39 Moreover, a link
anceps appears in the middle of the verse. This verse is not aeolic
but freer D/e.
Freer D/e is exempliWed by the following two passages:
N6s4a ww
ww
ww
jrw
j
j ^D ex
N6s4b
w
k e
N6s5 ww
j
ww
ww
j ^d D
N6s6a
ww
ww
j
j
ww
jrw
j
k Dxdex
N6s6b
wtj
k ex
N6s7
wtj
ww
jwj
w
k edxe
P5s9 wj
w
j
w
j
w
k xeee
P5s10
w
j
ww
jrw
j
w
k edee
P5s11 w
j
jtw
jwj
w
k ^e x e x e

The similarity of freer D/e to the normal D/e is obvious. However,

freer D/e is diVerent from the normal D/e in some respects. They will
be discussed later in I. 6, but I summarize:
(i) the basic phrase of double-short movement is not D
(
ww
ww
) but d (
ww
);
(ii) other less common phrases are extensively used;
(iii) link anceps is not used so frequently as in D/e, especially
within the verse, rather phrases tend to be juxtaposed without a
link between them;
(iv) verses in freer D/e tend to be much shorter than those in D/e.
I now examine the two types of metric in greater detail.

39
The phrase w
, which I analyse as headless e, is, of course, asymmetrical itself.
It is a derivative of e.
24 Introduction

5. AEOLIC PHRASES

A. The Basic Structure

Aeolic cola are characterized by the presence of one or other of two
asymmetrical phrases:

ww
w
dodrans (dod )

w
ww
reversed dodrans (rdod)
Both phrases have six positions, and both have a double short and a
single short Xanked by longs. The diVerence lies in the order. They
are mirror images of each other.
From the structural point of view, these two basic phrases, the
dodrans and the reversed dodrans, can be preceded by aeolic base
(hereafter called full base, ), or by single anceps (half-base, x).40
no base
ww
w

w
ww

dodrans (dod ) reversed dodrans (rdod)
half-base x
ww
w
x
w
ww

telesillean (tel ) heptasyllable (hepta)
full base
ww
w

w
ww

glyconic (gl ) wilamowitzianum (wil)

1. The second position of reversed dodrans

The second position of the reversed dodrans is shown above as short. In
the mainstream of Greek metre it is anceps (
x
ww
). But in
Pindaric usage the position is hardly ever realized as long. Of 63
examples which contain the reversed dodrans, counting together the
reversed dodrans proper and those prolonged by half or full base, none
has the second position realized as long (i.e.


ww
) consistently
throughout all the repetitions. In most examples (56 phrases), a long
syllable never appears and a short syllable always Wlls the anceps
position. Even in the other seven, short anceps is dominant, and long
anceps is often (but not always) caused by proper nouns. Apparently in

40
The historical process is not taken into account here. Rather, the glyconic is the
most common of all the aeolic cola in Pindaric metre as elsewhere.
 Aeolic Phrases 25
Pindar, the long anceps is an occasional licence. But it must be noted
that not all the long syllables can be eliminated as exceptional. There are
a few examples which undoubtedly have anceps at the position, like
aeolic cola in tragedy. These defy easy emendation, and attempts to
emend them metri causa should be rejected. As I shall argue later, these
examples are judged to belong to stanza-forms in rigidly aeolic style
(Class I). For more details, see E below.

2. Full base
Aeolic base consists of two ancipitia in Lesbian poetry. Thus the
notation xx is appropriate there, and a base in this form is a dis-
tinctive characteristic of aeolic metre. But in Pindaric usage, as in
Attic poetry, the base is actually occupied by the following combin-
ations of syllables:

,
w, w
, and www.41 For a base in this
form, I use the symbol .
There are considerable diVerences in the frequency of each form of
the base. Pindar predominantly employs
w. Responsion between
diVerent forms is not entirely free, being admitted only between


and
w (therefore the notation
x is applicable for the
scheme for particular passages),42 but the other two forms (w

and www) correspond only with themselves, except in one or two
cases. So, while the notation x
can be used for tragedy and
elsewhere, it is not appropriate for Pindar.43 This peculiarity of
Pindar is shared without exception by all the phrases starting with
the base. For the statistical detail, see C below.

41
Some editors accept the responsion between ww
and

at N6e8. If the
text there is correct, we must admit not only ww
as a variation of aeolic base quite
possibly unique in Greek poetry, but also the responsion between it and

. This
is outrageous. There is a notorious glyconic starting with ww
: Aristoph. Ra. 1322
(
 ; t ; T ). Bacchylides 18 uses ww

ww
w
in the
environment of glyconics. I leave provisionally the transmitted text and the respon-
sion t
, but this does not mean that I accept it. See further, Part II, ad loc.
42
But even this responsion is, in fact, restricted to Class I stanza-forms. See below,
5. C.
43
For the responsion in tragedy (with statistics), see Itsumi, Glyconic, 678.
26 Introduction

3. Half-base
The aeolic base can be reduced to a single anceps: half-base. It can be

, w, or ww. The last form, ww, stands in the same relationship to

or w as www at full base does to

etc. Responsion is possible
between
and w, but not between ww and
or w. I use the symbol x
to cover all these forms (including ww) for conveniences sake. For the
statistics, see D below.

4. Catalexis
The Wnal w
of the dodrans can be changed into
; this process is
called catalexis. This is an a priori deWnition of catalexis, and it
implies that the Wnal position of, for example, pherecratean is not
anceps but triseme (equal to three morae, T).44
The following are catalectic cola:

ww

adonean (adon)
x
ww

reizianum (reiz)

ww

pherecratean (ph)
Catalexis is naturally not applicable to the reversed dodrans.45
Some scholars posit phrases truncated yet further; for example:

ww

Incidentally, this form subsumes as one of its realizations the re-
versed dodrans (
w
ww
). The doubly truncated phrase can be
dispensed with, so far as Pindaric examples are concerned:

w
ww
and www
ww
are the reversed dodrans, while
w

ww
is acephalous e (^ e)d.46 It is not necessary to assume

44
Cf. Parker, Catalexis, 15. Ancient metricians use the term catalexis for
mechanical amputation of the last syllable of a colon. That was a product of their
conception of Wnal anceps, which is now seen to be invalid. Cf. E. Rossi, RFIC 91
(1963), 5271.
45
Thus the latter phrase of the eupolidean dicolon

x
ww

x
w

is not the catalectic form of the former; contr. West, GM 95.
46
Whether an actual phrase
w
ww
or www
ww
should be taken (i)
as the base ()
ww
, or (ii) as reversed dodrans itself (of which the Wrst
position is resolved in the latter case) is, in a sense, a purely academic problem.
 Aeolic Phrases 27
the existence of still shorter phrases, since x
ww
and
ww
are
freer D/e phrases (x d and d ).

5. Prolongation
Just as phrases may be abbreviated (catalexis), so they may be
prolonged. Prolongation is possible in both dodrans and reversed
dodrans; for example,

ww
w

hipponactean (hipp)
is a glyconic prolonged by
. Presumably the value of this
is not
triseme (nor true long) but anceps (or perhaps the long similar to the
Wnal long in the dactylic hexameter), comparable with that of link
anceps in D/e, in Pindaric metre. Admittedly, the hipponactean
could be seen as a catalectic form of
ww
w
w
, but in
practice this phrase is extremely rare, and is not found with a
hipponactean as clausula. The same applies to the hagesichorean.
Some phrases prolonged by
have no name. This type of phrase is
designated by the notation 1, e.g. wil1 (see the chart below).
I analyse the aristophanean

ww
w


as dodrans prolonged by
(dod1). The aristophanean is certainly
the catalectic form of
ww
w
w
in Attic poetry, but it is not
found with that function in Pindar.47 Rather, one of the two ex-
amples functions as prolonged dodrans (dod1):
O1e7 w

w

ww
w

ww
w

k ^e e dod ar
A more important diVerence lies in the implication: if (i) is accepted, it means that
w

ww
is a correlative form of
w
ww
or www
ww
and even that
responsion between them is theoretically possible. There are 5 cases of w

ww
,
which do not correspond with other forms, in the eighteen majors (O1e2, P2e6, P5e5,
P10s2b, P10s4). Judging from the metrical context and consistency with other related
ones, it is better to take the phrase as acephalous e (^ e) d (i.e. a freer D/e phrase);
For acephale in general, see further 6, B and for the ambiguity between aeolic full base
w
and ^ e, see 7, (4). Thus I prefer (ii) to (i).
47
Even in tragedy, aristophaneans are found repeated in synaphea; see Eur. Ba.
105 V. cited in n. 28 above. Stinton supposes that link anceps occasionally
occurs in aeolo-choriambics (
ww
w
x in our case); at least, he admits the
diYculties about this type of colon. See Stinton, Pause and Period, esp. Postscript,
646 ( Collected Papers, 35861).
28 Introduction
Beside prolongation by
(1) there are longer ones like
x
w
ww
w
hepta2 48

ww
w
w

gl3 (phalaecian)
Outside Pindaric metre, the ending w

can be taken as bacchiac.
The name bacchiac is not neutral since it inevitably implies cata-
lectic iambic. Phalaecian may therefore be less tendentious than
glba. But whether the last postion is created by catalexis or not is
uncertain. It may be anceps like
(1). I use the notation 3
without any implication. For the same reason, the ending w
(2)
must be diVerentiated from the spondee following an aeolic colon.
In fact, we cannot always distinguish between catalexis and pro-
longation, although such a distinction may have been made in per-
formance, i.e. audibly by music and visibly by dance. For example, some
examples of
w
ww

could in theory be rdod1 (prolongation),
but in practice we have no way to distinguish it from pherecratean with
base in the form
w (catalexis). The most delicate case is
N3e2
w
ww

w
ww

k
where it is very tempting to suppose that reversed dodrans is
repeated, with the second dodrans followed by anceps. Also, the
pherecratean in the apparent priapean dicolon at O1s1 may not be
pherecratean, but in fact rdod1 (for the analysis of O1s1, see further
7. 5). On the other hand, so delicate a classiWcation risks producing
too many borderline cases. So hereafter I shall treat every example as
pherecratean. The same is the case with rdod2 (gl) and rdod3
(hipp).49
Below I set out all the theoretically possible forms. The forms in
square brackets are the ones which can or should be analysed in
diVerent ways, and * is attached to forms which do not occur in the
eighteen majors.

48
This aeolic enneasyllable has been regarded as anceps plus glyconic by some
scholars. This interpretation entails the unlikely supposition that there is a colon in
Greek metre which starts with three ancipitia. And the so-called Barretts scheme
(Hippolytos, Appendix I, p. 422) is refuted by Dale, LM 153 V.
49
Although the argument is circular, I am inclined to suppose that pherecratean in
Class I stanza-forms (aeolic) is really pherecratean (i.e. catalectic glyconic) but only
seems to be so in Class III (amalgamated), being in fact rdod1. So are other related
phrases. I shall come back to this problem in the Wnal chapter.
 Aeolic Phrases 29
dodrans reversed dodrans
prolonged by



ww
w

[
w
ww

]
aristophanean (ar) [rdod1 ph]
x
ww
w

x
w
ww


hagesichorean (hag) *hepta1

ww
w


w
ww


hipponactean (hipp) wil1
w

[
ww
w
w
] [
w
ww
w
]
[dod2 d w e] 50 [rdod2 gl]
[x
ww
w
w
] x
w
ww
w

[tel2 x d w e] hepta2
[
w
ww
w
w
]
w
ww
w

*[gl2 rdod w e] 51 wil2
w


[
ww
w
w

] [
w
ww
w

]
*[dod3 d w e
] [rdod3 hipp]
x
ww
w
w

x
w
ww
w


tel3 hepta3

ww
w
w


w
ww
w


gl3 (phalaecian) wil3
Although examples are very scarceindeed only one, P8e7, and that
a forced analysisthe following phrase (a type of dodecasyllable) is
taken as a single aeolic phrase:
hepta23 x
w
ww
w
w


The enneasyllabic phrase, hepta2, is prolonged in the same manner
as the glyconic in gl3 by w

. Its theoretical correlative
wil23
w
ww
w
w


does not in fact occur in the eighteen majors.

50
The metrical contexts of actual examples refute the aeolic analysis of d w e and
w d w e. They are collected at 8. C. 7.
51
The analysis, rdod w e, is concordant with that of

w
ww
w
w
w

which is easily accepted as glyconic link anceps (w) e. Note that the anceps is
always short in these cases. Examples are collected and discussed at 8. B. 5.
30 Introduction

B. ClassiWcation by Ending and Frequency

All the aeolic phrases are arranged below in a diVerent synoptic chart
according to the number of positions preceded by the choriamb. In
the following chart the number of occurrences is given in the right-
most column:
zero ending
reversed dodrans
w
ww
23
heptasyllable x
w
ww
12
wilamowitzianum
w
ww
15
1 ending
adonean
ww

2
reizianum x
ww

9
pherecratean
ww

11
*hepta1 x
w
ww

052
wil1
w
ww

1
2 ending
dodrans
ww
w
10
telesillean x
ww
w
29
glyconic
ww
w
49
hepta2 x
w
ww
w
3
wil2
w
ww
w
4
3 ending
aristophanean
ww
w

2
hagesichorean x
ww
w

1
hipponactean
ww
w

2
hepta3 x
w
ww
w

2
wil3
w
ww
w

2
[4 ending]53
5 ending
tel3 x
ww
w
w

4
gl3
ww
w
w

3
hepta23 x
w
ww
w
w

1
*wil23
w
ww
w
w

0

52
There are two examples outside the eighteen majors: Pae6e10, e11.
53
There is no phrase with 4 ending because I do not analyse phrases as aeolic
that could theoretically be included in such a category. Thus
ww
w
w
in
the list above I analyse as d w e, although it is equal to dod2.
 Aeolic Phrases 31
There are 145 phrases with zero or 2 ending against 40 with 1,
3, or 5. At the moment we should not jump to the conclusion
that blunt aeolic endings are preferred to pendent, because phrases
with zero ending and phrases with 2 can stand at the middle of the
verse as well as at verse-end and, consequently, have more chance of
being employed. The frequency of blunt ending will be discussed at
greater length later (8. A. 4).

C. The Full-Base Group

The following table consists of the same material as the table above,
but in a diVerent arrangement.
pherecratean
ww

11
glyconic
ww
w
49
hipponactean
ww
w

2
gl3
ww
w
w

3
wilamowitzianum
w
ww
15
wil1
w
ww

1
wil2
w
ww
w
4
wil3
w
ww
w

2
The full-base group can be further divided according to the nature of
the base.

w www w



x x
tx (t
)54 total
gl 12 12 8 3 11 1 1 (1) 49
ph 6 1 1 3 11
hipp 2 2
gl3 1 1 1 3
wil 9 1 2 3 15
wil1 1 1
wil2 1 2 1 4
wil3 2 2
32 15 12 6 19 1 1 (1) 87
The four left-hand columns after the stub indicate the number of the
phrases in which exact responsion is rigidly kept throughout all the
54
The last form in parentheses, t
, is based on a certainly corrupt text (N6e8);
cf. n.41 above, but is included in the calculation.
32 Introduction
repetitions. There are in total 65 instances (3215126).55 The
proportion of these is remarkably high (65/87 74.7%). In other
words, the anceps does not work as its name suggests. Another
characteristic is the rarity of examples of

(6 in total), compared
with 31 of
w and 15 of www. It should be remembered that

is
the most usual form in tragedy.56 In contrast, Pindar seems usually to
have avoided the heavy full base (

).

1. The base
x
Pindars preference for repeating the same form of full base is still
more evident when one examines each repetition of the 19 examples
of
x (for the terms example and repetition, see Key).57 Two
peculiar groups emerge. First, in six cases

occurs only in a
limited number of repetitions, in most cases one repetition only,
while the others are of the shape
w, and, moreover, all the


without exception involve proper nouns. For example, at six out of
seven repetitions of glyconic in I8s5c and I8s6, aeolic base is consist-
ently occupied by
w, the exceptions being in s5c at v. 55c ( v. 56
Sn.) and in s6 v. 16 ( v. 16a Sn.). At both v. 55c and v. 16, `Y, a
proper noun and a key word of I8, occupies the base

, and
consequently the scheme is given as
x
ww
w
. The same is true
in the other four examples (P2e5, N4s5, N4s6, I7e5). Secondly, there
are also some examples where a strong preference is observable, but
where, nevertheless, there is a single deviation from the norm which

55

w (gl) P5s3, P8s2, P8e6, P11s2b, N3s4, N4s4, N6s1b, N6s2, N6e2, N7e3,
I7e1, I8s4; (ph) O1s1, O1s4, O13s2, O13s5, P5e4 N3e2; (hipp) P2e8, N7s8 (wil) P5s2,
P5e1, P6s1/2, P6s4, P6s5, N6s2, I8s1/2, I8s3, I8s4; (wil2) N3s7; (wil3) P10s5,
N7s4.
www (gl) P2e1a, P2e4, P6s1/2, P6s3, P6s6, P8s1, P11e4, N2s4, N3s3, N4s7, N7e4,
N7e5; (ph) P11e2; (gl3) N7e5; (wil) O9s6/7.
w
(gl) O1s1, P2e1b, P2e2, P2e3, P5e2, P8e6, N2s1, N7s1; (wil) I8s1/2, I8s5a;
(wil2) P2e8, P5s8


(gl) O9s4, N2s4, I7s5a; (ph) P8e3/4; (gl3) I7s3/4; (wil1) P6s7/8
56
Itsumi, Glyconic, 678.
57

x (gl) O9s3, O9s5, O9s8, P2s2, P2e5, P8e3/4, P10e1, N2s3, I7e5, I8s5c, I8s6;
(ph) P8e2, P10s1, N2s4; (gl3) O9s2; (wil) N4s3, N4s5, N4s6; (wil2) P8s5. The
examples of the other two forms (x
, tx) are too rare to be taken into account.
For the detail of x
, see Part II, O10s6. tx (P5e9) will be illustrated later.
 Aeolic Phrases 33
does not involve a proper noun. Thus, at O9s2, P10s1, and P10e1,

w is found in every repetition of the verse except one, but no
proper noun is involved. This tendency is a manifestation of the
All-but-One rule (see Part III, C). At O9s5 and P8e3/4,
w being
the norm,

is used more than once, in all instances but one with
a proper noun. Conversely, at O9s8,

is the norm, but
w
occurs once, without proper noun.
So in 65 examples the aeolic base is of identical form throughout
and in another 12 examples it deviates from the chosen norm only
very rarely. Yet Pindar does not always show the same attachment to a
single form. In seven verses for which the notation
x is appropri-
ate, no such strong preference is observable: O9s3 (3 longs/8 repeti-
tions), P2s2 (4/8), P8s5 (4/10), P8e2 (2/5), N2s3 (3/5), N2s4 (2/5),
N4s3 (8/12). In these cases the anceps is really employed as anceps.
The situation is intriguing. A simplistic metrical rule is not appro-
priate. An explanation is to be sought not in a general metrical theory
but in the examination of each metrical context; in other words, of
the style of each stanza-form as a whole. Interestingly, the occurrence
of
x is concentrated in a limited number of stanza-forms or odes:
O9s (4 examples), P2e (1), P2s (1), P8s (1), P8e (2), P10s (1), P10e
(1), N2s (2), N4s (3), I7e (1), I8s (2). Adding to them the examples of


(6 in total), it is reasonable to suppose that Pindar employed


, whether in responsion with
w or not, only in some limited
stanza-forms. As will be argued later, the metrical context of these
stanza-forms is aeolic, not freer D/e. Thus the presence of

is to
be accepted as one of the criteria of Class I (aeolic) stanza-forms.
Here emerges an important admonitory remark. It is dangerous to
emend the text by introducing a word which has a long syllable at the
second position of a glyconic (or equivalents), without consideration
of the nature of the stanza-form as a whole. Thus, for example, I am
dubious about the transposition of
 at N7s8 (v. 37)
proposed by Boeckh and accepted by most editors.58

58
The paradosis of v. 37 does not make good metre and must be emended; but
Boeckhs transposition is not as easy a solution as has been supposed. See further Part
II ad loc.
34 Introduction

2. The base www

Pindar uses www frequently as full base. He is perhaps the Wrst Greek
poet to introduce resolution in the aeolic base; the Lesbian poets and
Anacreon did not use it, nor did Aeschylus except once.59 It becomes
common in later tragedy, especially in Euripides. Theoretically, www
can be taken as a resolved form of
w or w
; but as a rule www does
not correspond with the other forms, even in tragedy. The sole
exception in Pindar is found at the beginning of P5e9:
P5e9 tw
ww
w

ww
w rw

w
k gl d w e e
There, correspondence occurs between www o (31) and
w Z
(62) and h (124) and


 (93). Perhaps the pecu-
liarity of the style of P5 as a whole may be related to this irregularity
(see Part II, ad loc.).

3. The base w

The base w
is common neither in Pindar nor in tragedy. Of 12
examples in total, four are found in P2e, and two in I8s. This scarcity
is partly explained by the fact that Pindar does not use glyconic or other
phrases with full base a  (P2e1a4 are a rare exception).
Anacreon and the dramatic poets are fond of the regularity produced
by repetition, but with slight variation of the base. That is not Pindars
manner. Thus O1s1 (but see below, 7. 6) and P8e6 are exceptional:
O1s1 w

ww
w

w
ww

k gl ph
P8e6 w

ww
w

w
ww
w
k gl gl
In other words, since the phrase w

ww
w
and the like are not
invariably employed next to an unambiguous aeolic phrase starting
with full base, there are cases in which it is not easy to tell aeolic base
w
from acephalous e (w
). For example, the Wrst phrase of
P5e2 w

ww
w
w
w

w
ww
j
may be not a glyconic, but ^ edodrans followed by short
ancepserdod, since a number of verses in P5 start with ^ e (Part
II, ad loc.). Similarly,
59
Itsumi, Glyconic, 68.
 Aeolic Phrases 35
N7s1 w

ww
w

w
w
k gl e2
may be ^ edode2, taking into consideration that the resemblance
between the phrases
ww
w
and
w
w
is fully exploited in
this stanza-form (see Part II, ad loc.). As for the wilamowitzianum, I
analyse
I8s5a w

w
ww

w
ww

w
ww
j
as wilrdodrdod. But it may be w
(^ e)threefold
w
ww
.
But at the same time it must not be forgotten that there are some
examples of w

ww
w
which are, judging from the context,
most certainly glyconic: for example, N2s1, and three consecutive
verses of P2e (e1b3). It is impracticable to take all the examples of
w
as ^ e. For the ambiguity, see further 7. 4.

D. The Half-Base Group

reizianum x
ww

9
telesillean x
ww
w
29
hagesichorean x
ww
w

1
tel3 x
ww
w
w

4
heptasyllable x
w
ww
12
hepta2 x
w
ww
w
4
hepta3 x
w
ww
w

3
hepta23 x
w
ww
w
w

1

w ww x total
tel 14 3 6 6 29
reiz 4 3 2 9
hag 1 1
tel3 2 2 4
hepta 3 1 8 12
hepta2 1 1 1 3
hepta3 2 2
hepta23 1 1
24 9 8 20 61
36 Introduction
Generally speaking, long is preferred to short in half-base: 24 ex-
amples against 9.60 Especially when an aeolic phrase makes up a verse
on its own, long is dominant. P10s2 starts with short anceps, and is
the sole exception, while long anceps occurs at every repetition in
nine phrases ( verses). Besides, at explicit anceps (x, 20 examples)
where both long and short are freely employed, long is dominant.
Every repetition being counted, long occupies the anceps position 74
times, short 60 times. In some verses the anceps strongly tends to be
either long or short: e.g. N4s3 (long is used at 11 out of 12 repeti-
tions; All-but-One is a manifestation of extreme cases of this ten-
dency) or N4s1 (2 out of 10). These verses, and others of N4, indicate
that it is neither a metrical rule nor the metrical context that regu-
lates the realization of the anceps. N4 is a very simple aeolic stanza
(monostrophic), and its Wrst three verses start with hepta or hepta2.
But the ratio of long to short in each verse varies considerably: 2/10
(s1), 9/3 (s2), 11/1 (s3).61 This seems to be a matter of aesthetic
preference in the individual context.

1. Position of half-base in verse

Half-base is used either at the beginning of a verse (in cases where the
verse starts with the phrase in question) or in the middle (in cases
where the phrase is preceded by others). The dominance of long is
clearer at the beginning of the verse.

60 *
means that half-base is used in the middle of a verse.

(tel) O1e6a, *O9e4, O10e6, P11e3, *P2s2, P2s8, P5s7b, N3s1, N7s6, N7s7, N7s8,
*
I7s5a, I7e2, I7e3; (reiz) *O9s4, *O9s5, *O9s6/7, O9s9; (tel3) N2s2, N4s8; (hepta)
P2e7, P8s3, P11s2a; (hepta2) P11s5.
w (tel) O9s10, *I8s3, *I8s7; (reiz) *O1e4, *N4s4, *N4s6; (hag) *P2s8; (hepta) P10s2a;
(hepta2) P8e1.
ww (tel) O9s1, *P2s4, P10s6, N3s8, N3e4, *N3e4; (tel3) I7s1, I7e4.
x (tel) O9s2 (long at 7 repetitions out of 8), *P10e6 (1/4), *N4s5 (2/12), I7s2 (2/6),
*
I7s3/4 (1/6), *I7e5 (2/3); (reiz) *O9s3 (7/8), *N2s3 (2/5); (hepta) *O9e8 (2/4), P8s6
(4/10), P8e2 (3/5), P10e3 (2/4), P10e4 (3/4), N4s2 (9/12), N4s3 (11/12), *I7s5a (5/6);
(hepta2) N4s1 (2/12); (hepta3) *O9e8 (3/4), P10e5 (3/4); (hepta23) *P8e7 (3/
5).
61
Or 10/2 (s2) and 12 /0 (s3), if the initial syllable of  (v. 82) is scanned
long, and if P
 (v. 75) is read instead of 
 .
 Aeolic Phrases 37

w ww x total
beginning 18 3 6 10 37
middle 6 6 2 10 24
Long anceps is exclusively used in half the verses (18 out of 37).
However, there are 3 verses which invariably start with short anceps
in all the repetitions: O9s10 (tel), P8e1 (hepta2), P10s2a (hepta). At
the 10 explicit ancipitia at the beginning of the verse, no strong
preference is evident: long occurs in a total of 46 repetitions, short
in 31. Just one verse has a high occurrence of short syllables: N4s1
(hepta2, 10 short out of 12).
There are 6 examples of half-base in mid-verse where anceps
is invariably short. All these phrases are preceded by
. . .
w
w
, . . .
ww
w
or . . .
ww
, as the Wnal part of the
preceding phrase, so that the whole verse makes a sequence in which
single short or double short alternates with single long without inter-
vening long or anceps:
. . .
w
w
w
ww
. . .
. . .
ww
w
w
ww
. . .
. . .
ww
w
ww
. . .
These verses make a group of particularly Pindaric colour. The
examples will be collected and discussed in I. 8.

2. Half-base ww
In 8 cola, ww occupies the base. All the cola are telesillean or its
prolonged form (tel3). I do not classify
O9e3, O13s1 ww
ww

k
as reizianum, nor the Wrst phrase of
O1e5 ww
w
ww

w
w
k
as heptasyllable starting with ww: O9e3, O13s1 is acephalous D and
anceps, and O1e5 is ^ d w d.62

62
Some of the telesilleans, ww
ww
w
, may be interpreted in a diVerent
manner. See 7. 2.
38 Introduction
ww is never in responsion with
, let alone w. Even outside Pindar,
I do not know any certain example of t
ww
w
. Whether the
two double shorts in the phrase
ww
ww
w

are equal, and if so, whether they are distinguishable from those of
^D

ww
ww

is an interesting question.

E. The First Two Positions of Reversed Dodrans

As already demonstrated above (A. 1), there is no example which has



ww
at all the repetitions. Thus reversed dodrans itself and
phrases which include it are classiWed into three groups:
w
ww
,

x
ww
, www
ww
.

w
x www total
dod 13 1 9 23
hepta 10 1 11
wil 13 2 15
wil1 1 1
hepta2 2 2 4
wil2 4 4
hepta3 2 2
wil3 1 1 2
hepta23 1 1
46 7 10 63

1. Reversed dodrans starting with
x

The form
x
ww
, of course, permits initial

. Of the seven
examples of
x, four (O9s6/7, N4s1, N4s2, N4s6) occur in those
stanza-forms where aeolic characteristics are most evident, which
I shall call Class I. This tendency is obviously related to the fact
that

tends to be used at the full base in the Class I stanza-forms
(see C. 1 above). But even here, the repetitions with long anceps
 Aeolic Phrases 39
remain very much in the minority.63 Besides these four examples,
there remain a further three. In
P10s2a w
x
ww
j hepta
long anceps is found only at v. 8 (I). If this word can be
treated as a proper noun, then the irregularity would be mitigated.
The sixth example,
P11s4
xtww


wr
ww
k rdod x e d
is most irregular in that not only is long anceps used in the second
position, but there is resolution to the left of aeolic nucleus in half
the repetitions, and long medial anceps occurs after rdod.64 However,
even here, there are some limitations: in repetitions where resolution
occurs, the preceding anceps is always short. Thus there is no ex-
ample of

rww
, its avoidance agrees with the general rule that
long anceps does not precede resolved long; cf. 6. C (iii). The last
(P8e7) is a curious verse as a whole, and, whatever analysis is chosen,
its colometry will inevitably be a forced one. I oVer the following
provisionally:
P8e7

x
x
ww
w
w

k sp hepta23
The anceps position of the (assumed) reversed dodrans (hepta x
rdod) is Wlled with long at 3 repetitions out of 5. See Part II, ad loc.

2. Reversed dodrans starting with www

Ten verses have resolution of the initial position of the reversed
dodrans (www
ww
).65 There is no phrase in which resolved pos-
ition is in responsion with unresolved. Of the nine examples of
resolved reversed dodrans, Wve stand at the beginning of the verse.
One is tempted to ask whether reversed dodrans in this form
sounded identical in performance with the Wrst part of a glyconic
with base in the form www.

63
O9s6/7 (3 long out of 8), N4s1 (2/12), N4s2 (3/12), N4s6 (1/12). Moreover, four
of these long ancipitia are in proper nouns.
64
Resolution at aeolic nucleus is discussed below, 5. F. For the mid-long anceps
see 6. D.
65
(rdod) O1e6b, P2s2, P2s7, P2e2, P2e3, P5e3, P5e6, N6s3, N7e2; (wil3) N7s4.
40 Introduction
www
ww

www
ww
w

We cannot answer this question since we do not know the time-value
of anceps. It is, however, a reasonable hypothesis, and one which I
propose to adopt, that the two phrases did sound identical.
Compare the following verses, which make a spectrum:
P2e4 www
ww
w

w
k gl e
P6s3 www
ww
wy
w
k gl e
N7e4 www
ww
wr
w
k gl e
P2s7 www
ww
rw

w
k rdod e e
O1e6b www
ww

w
k rdod e
The Wrst two (P2e4, P6s3) are unambiguously gle. Following these
two verses, I analyse N7e4 as gle, not as rdode2. However, the
next, P2s7, is not glyconic followed by long ancepse, because there
is no certain case at all of gl
e.66 On the other hand, P2s7 is similar
to O1e6b, where the Wrst three shorts unambiguously belong to rdod.
The metrical context must be taken into consideration: reversed
dodrans of the form in www
ww
occurs in a limited number of
stanza-forms: O1e (1 example), P2s (2 examples), P2e (2 examples),
P5e (2 examples), N6s (1 example), N7s (1 example), N7e (1 example).
These are classiWed as Class II (freer D/e) or Class III (amalgamated),
not as Class I (aeolic), except for P2e (an ambiguous case). On the other
hand, the aeolic base in the forms

or
x is a mark of Class I.
It may not have been harmonized with resolved dodrans.

F. Resolution
Apart from www at full base and ww at half-base, a long position in
aeolic phrase is sometimes Wlled by two short syllables. This can
reasonably be regarded as resolution.67 Every long position of
dodrans and reversed dodrans can be resolved. For example, the
following forms of glyconic are found:

66
An aeolic phrase is, in general, not followed by a long anceps. There are only
two exceptions (P10s6, P11s4). See below, 6. 3.
67
Resolution of longs in aeolic verses is rare in Attic drama; cf. Itsumi, Glyconic.
It is almost restricted to later Euripides. The unique example of Aeschylus is Cho. 317:
 i  P (w
rww
w
). It worries many (e.g. Dale, BICS
 Aeolic Phrases 41
Left of the choriambic nucleus N6e2
wrww
w
j
resolved
Right of the choriambic nucleus P8s2
w
wwrw
k
resolved
Final position resolved I8s5c
x
ww
wt
w

k
P6s3 www
ww
wy
w
k
Resolution must be attested by responsion with unresolved pos-
ition. Examples of resolved long position corresponding with unre-
solved are found in four verses:
P11s4
atww


wr
ww
k rdod x e d
N7s7
tww
w
rw
w
k tel e2
P6s3 www
ww
wy
w
k gl e
I8s5c
a
ww
wt
w

k gl e

In the Wrst two (P11s4, N7s7), resolution is found at the left-hand
long position of the choriambic nucleus (tww
). Of these, the case
of N7s7 is not surprising, since resolution involves a proper noun (v.
70 P), and may be taken as a special licence. But P11s4 is
extraordinary. Of its eight repetitions, four have resolution, and none
of the resolutions involves a proper noun (v. 9 , v. 41 e b, v. 52
Ia, v. 57 ).68 In the other two (P6s3, I8s5c), resolution
occurs at the Wnal position of the glyconic, that is, of the dodrans
(
ww
wt). At P6s3, the position is unresolved in only one repe-
tition (v. 48 l). And out of seven repetitions at I8s5c, three are
resolved (v. 25c v. 26 Sn.
 , v. 35c v. 36 Sn. ,
v. 45c v. 46 Sn. K
). Here the situation is similar to P11s4 in
that resolved positions and unresolved are freely used in responsion.
There are 16 examples in which resolution is present at every
repetition (i.e. responsion between resolved and unresolved is

Suppl. 21.2 (1981), 15: resolution in 317 very ugly and hardly possible in Aeschylus),
but it may be accepted by referring to N6e2 and, especially, P11s4. It is noteworthy
that P11 treats the myth of Clytaemestra and Orestes. Perhaps Aeschylus is inXuenced
by Pindar. Further, Cho. 315 332 (the Wrst verse of the same strophe) may betray
another inXuence of Pindar.
ww
ww
w
is, most certainly, not glyconic in
Aeschylus (see Itsumi, Glyconic). Rather it may be related to P11s1 (for this verse,
see 7. 6 below).
68
For the avoidance of long anceps preceding the resolution, see 6. C above.
42 Introduction
absent). Of these, 10 are the initial of reversed dodrans, mentioned in
section E above. The others are: the Wnal of the dodrans, N7s2, N7e3,
N7e4; the left of the nucleus, N6e2 (gl); the right of the nucleus, P8s2
(gl), P11s2b (gl). In some verses it is diYcult to decide which
position is resolved. The sequence . . . www
w
is typically confus-
ing. For example, see the phrases at the end of
N7e3
w
ww
www
w
k
N7e4 www
ww
www
w
k
These are resolved forms of either (i) . . . w

w
or
(ii) . . .
w
w
. If (i) is chosen, the verses are glyconiccretic
(e). On the other hand, if (ii) is chosen, the verses are reversed
dodranse2. I choose (i) on the analogy of P6s3 above, but there is
no certainty here; see further Part II, ad loc.

G. Acephaly
Besides the phrases above, there is one more phrase that is probably
aeolic, judging from its context:
O13s5 ww
w

w
ww

k ^ dod ph
The Wrst half of this verse is explicable as acephalous dodrans. Acephaly
is rather common, as will be demonstrated, in freer D/e, but to what
extent it occurs in aeolic cola in general is not certain.69 At least there
are no other examples in the eighteen majors. Acephalous dodrans
(ww
w
) is familiar as a constituent of the third verse of the Attic
skolion, but it is doubtful whether such an extra-generic comparison
is meaningful. Rather, the proper comparandum is the Simonidean
poem 542 P. Here acephalous dodrans is most certainly attested. The
poem is monostrophic, with a seven-verse stanza:70

69
Note that O13s is the only stanza-form where the metre shifts from non-D/e to
D/e in the middle. The alternative interpretation is: ww e (ww being substituted for
anceps). See Part II, ad loc. Outside the eighteen majors, ^ dod occurs at Parth1s1; see
Part II, Appendix B.
70
I follow Page in the reconstruction of the text, but the colometry is my own.
West, GM 66 gives an analysis diVerent from both Pages and mine.
 Aeolic Phrases 43
1
ww
ww
w
w

j
2/3 ww


w
ww
w

w
ww
w
x
w
j
4/5 ww
w

w
ww
w

w
ww
w
k
6 ww
w

w
ww
j
7

w

w
ww
k
8 w

w

j
910
w
ww
x
ww
w

j
Although there remain some uncertainties in the reconstruction of
the strophe, it is certain that the phrase is situated at the beginning of
the two verses 4/5 and 6. The basic structure of the former (4/5) is the
repetition of
w
ww
w
. It is natural to extrapolate from it that
ww
w
is equal to
w
ww
w
minus the initial
w
. The
latter (6) is almost identical with O13s5. These two Simonidean
verses and O13s5 incorporate the palindromic movement
( . . . ww
w

w
ww . . . ) found in O1s1 and elsewhere (see 8.
A. 6 below).
Readers accustomed to Snells analysis may wonder why I excl-
ude from aeolic consideration of dactylic expansion (e.g.

ww
ww
w
). There are, however, problems of ambiguity
here, which I prefer to discuss following my deWnition of freer D/e;
see below, 7. 6.
44 Introduction

6. FREER DACTYLO-EPITRITE

A. Basic Structure and Phrases

The essential features of freer dactylo-epitrite have been sketched in
I. 4 above. Listed below are the phrases that constitute freer D/e.
Some are unfamiliar in normal D/e, but many are found in normal
D/e, even if only very rarely. The Wgures in the right-hand column
indicate the number of occurrences. Examples are collected in List 2.
d phrases
d
ww
55
D
ww
ww
9
D
ww
ww
ww
5
e phrases
e
w
158
e2
w
w
38
e3
w
w
w
1471
e5
w
w
w
w
w
3
e6
w
w
w
w
w
w
1
sp

12
Acephalous d and e phrases (verse-initial only, followed by true long)
^d ww
3
^ D ww
ww
4
^ D ww
ww
ww
1
^e w
23
It will be clear at a glance how these phrases are related to each
other. The basic ones are d (
ww
) and e (
w
). d is expanded by
repetition of
ww, the maximum length being three double shorts.
In normal D/e, where D is one of the two standard phrases, d is an
occasional variant, apparently the abridged form of D.72 In contrast,

71
Including a curious colon
x
w
w
in P8s6. For this, see 6. A.
72
It was once maintained by Schroeder and others that d was equivalent to e and
long anceps, and that even strophic responsion between d (
ww
) and e anceps
(
w

) or anceps e (

w
) was possible. Sandys (Loeb) and Puech
(Bude) followed them. That idea was refuted Wrst by Maas (Responsionsfreiheiten), and
then by Bowra, An Alleged Anomaly, who scrutinized every possible example.
 Freer Dactylo-Epitrite 45
D in freer D/e should be taken as secondary, an enlarged version of d.
There are considerably fewer examples of it than of d. The longest,
D , to which Maas gave the notation Dd2, is not uncommon either in
normal D/e (10 examples).73 In normal D/e, there is one example of a
much longer sequences (D Dd2d2 :
ww
ww
ww
ww
) at
P3s4, but there is none in freer D/e.74
e is expanded like d. Expansion of e is far more extensive than that
of d. The longest includes six single shorts (e6). There is no example
of e4, but this is probably accidental.75 The central short position of e3
(
w
w
w
) may conceivably have been anceps in origin
(
w
x
w
), and certainly some of the brevia of the longer
phrases (e5 and e6) must surely have been so. It is possible that
sequences of alternating short and long were derived from the famil-
iar iambo-trochaic sequence . . . x
w
x
w
x . . . , but even if that
is so, we are in no position to try to decide in Pindar which w in such
a sequence is truly anceps. It should be noted that, by deWnition,
when the symbol en is used, all the intervening brevia are short
throughout all the repetitions. In fact, it is the absence of long anceps
which most strikingly diVerentiates the freer D/e from normal D/e.
The longer sequences cannot be analysed a . Both Turyn
and Snell impose forced analyses on
P2s1 rwrw
w
w
wrw
k e6
as dochmiaciaia (Turyn) or criaiaw
(Snell).
O1s8 wrwrw
w
w
w
k w e5
seems to be analysed successfully as iambic trimeter. But it should
not be so described: not only does the term conceal the close relation
to P2s1 but it is also inappropriate for another reason. At six (vv. 19,
37, 48, 77, 95, 106) out of eight repetitions, word-end occurs at
wrwrw
jw
w
w
k

73
West uses the symbol D2 for
ww
ww
ww
. I do not use it because the
Wgure 2 is confusing.
74
Outside the eighteen majors, there is an example of ^ D in Parth1s3.
75
In P7, which I exclude from the eighteen majors, there is an example of e4
(P7s1).
46 Introduction
Such a frequent occurrence of medial caesura is incompatible with
the iambic trimeter.
The process of expansion and its reversed, contraction, is most
clearly visible in the following two pairs of successive verses:
P2s5

wrw
w
ww
k
e2 w d
P2s6

wrw
w
ww
ww
k
e2 w D
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
O1s7
w
w
w
w
ww

w
k e3 w d e
e2 (
w
w
) is uncommon in normal D/e, but occurs once in a
verse in that metre:
N8e4

w


w
w
k
It is a pity that Snell gives the verse the notation
E w
(here E is
used in the Maasian sense, of course) so that the phrase often escapes
notice. Incidentally, the identical verse, x e x e2, is found in the other
half of the Pindaric corpus:
P8s7 a
w
a
w
w
k x e x e2
There is a curious phrase
x
w
w
. It is similar to e3, but its
second position is occupied by a long syllable at seven repetitions out
of 10:
P8s6 x
w
ww

x
w
w
k hepta e3(aeol)
I take this as a variation of e3. Its initial two positions are treated like
aeolic base. The same phrase is used by Bacchylides in poem 18. See
further Part II ad loc.
A spondee is occasionally used both in normal D/e and in freer
D/e. In both metres its use is restricted to the beginning or end of the
verse (for a possible exception at O10s3b, see Part II, ad loc.). The
spondee is most naturally understood as e of which the short pos-
ition has been suppressed.

B. Acephaly
Acephalous phrases are occasionally used in freer D/e as well as in
normal D/e. The concept of acephaly has not been unanimously
 Freer Dactylo-Epitrite 47
accepted,76 but it is the most natural explanation for a certain group
of verses. Since acephaly occurs at the beginning of a verse, it may be
assumed that the missing position is absorbed in the vacuum before
the verse starts. I do not, therefore, call the telesillean an acephalous
glyconic, because a telesillean can stand in the middle of a verse. The
following verses can only be analysed when it is assumed that an
initial long is lacking:
N6s5 ww

ww
ww
j ^dD
P6s4 ww

w
w
ww
k ^dwil
O10s1 ww
ww

w
tw
k ^D e e
N6s4a ww
ww
ww
rw

k ^D e

By deWnition, every phrase of single- and double-short movement

may be acephalous at the beginning of the verse. But ^ e2 (w
w
), if
it occurs, is not distinguishable from short ancepse, and so is
disregarded. So are ^ e3 and longer phrases of the same type. Ex-
amples of the d-group are not so numerous as those of ^ e : 3 of ^ d, 4
of ^ D, and 1 of ^ D . Though there are exceptions, examples tend to
occur in Class II stanza-forms.77 Acephalous phrases are also used in
normal D/e.78
As for ^ e, the most certain cases are the ones where full e follows,
so that the verse starts with w

w
, or www
w
, as in
O2s2 wy
w

wr
wr
w
k ^e eeee
There are verses which are made up only of the phrase. For example:
P5s6 w

w
k ^e e
Of 23 examples of ^ e, 11 are followed by e, 2 by e2. Examples are
found concentrated in particular odes or stanza-forms: 6 in O1 (3
each in the strophe and in the epode), 6 in P5 (also 3 each), 3 in P10s,

76
For example, compare the analysis of O2 by West, GM 689, with that by
Parker, Trochee to Iamb. Also see the forced analysis of N6s5 and P6s4 (cited below)
by Turyn, Snell, or others.
77
1 each in O9, in O10s, in O10e, in O13s, P6s, and 2 in N6s. O1e5 is an
exception.
78
^ D is represented by Maas as d2d2, and ^ d as d2. At the same time d2 is used in
Dd2 ( D ). ^ D is used in P3e9, N8s5, ^ d in O7s1, O7s6, O7e6, O8s6, O13e6, P1e9,
P9s1, P9s3, N8e3.
48 Introduction
2 in O2s. To call the phrase w

w
dochmiac is misleading,
since it implies that the phrase can be in responsion with, e.g.

ww
w
.79 ^ ee2 (in O1s10, followed by another e, and in P6s9,
followed by x e) is traditionally analysed as bacchiaciambic (w


w
w
). However, the bacchiac metron is certainly alien to Pindaric
metre. Once e2 is accepted, the analysis as ^ ee2 is reasonably
concordant with the other verses.
I recognize ^ e in the following cases too, for which diVerent
analyses may be conceivable (the alternative would be aeolic cola
starting with the base w
).
w

ww

w
j ^e d e (P2e6, P5e5, P10s2b)
w

ww
ww
k ^e D (O10e2, I8s9)
In the following two verses, ^ e is followed by link anceps:
N6s1 w

j ^ e

P5s11 w


w
w
w
k ^e
e3
Perhaps the most perplexing case is
P10s5 w

w
w
ww
w

k ^e wil3
This is unique in that ^ e is followed by an aeolic phrase.80

C. Resolution
Every long position of phrases made up of single-short movement (e,
e2, e3, etc.) can be resolved, and such resolution is very common. As
for e, there are 33 examples of
wr (and another 4 of wr, i.e.
resolved ^ e) and 27 of rw
. The total number of e is 181 (including
23 ^ e). Thus 35.4% of e (and ^ e) are resolved at one long position or

79
Support for the term dochmiac might be sought in O1s2, where the paradosis
oVers
ww
w
(v. 89   ) corresponding with www
w
in the
other repetitions. But this reading is surely a false correction by someone who
thought the subject of  must be the mother. With  (Boehmer, according
to Gerber) the dochmiac disappears. The seeming dochmiac of the form u


w
at the end of Pae4s, l. 4 (cf. the analysis of West, GM 67), is, in fact, caused by
neglect of the verse-end after its initial anceps (brevis in longo).

w
is an
independent verse. See further, Part II, Appendix B, ad loc.
80
P10s5 may be ^ e e hagesichorean; see 7. 3.
 Freer Dactylo-Epitrite 49
the other (
wr and rw
), at least in one repetition. This high
percentage is worth noting.
O2 is noteworthy in that both of these two types of resolved form
coexist repeatedly in the same verse. But similar verses are also found
outside O2, for example,
I8s8 rw

wr
wr
k e e e

It is occasionally diYcult to decide which long position is resolved.
For example, I analyse
O10s3a x
w

wr

wr

j x e e x e sp
Another analysis looks possible:
x
w

rw

rw

j x e x e x e

I reject it, however, for two reasons: (a) should it be accepted, word-
end would occur very frequently after the two long mid-ancipitia;
this is against Pindars general tendency (see section D), and (b)
Pindar in general avoids long anceps preceding resolved long; see
below, (iii).
Responsion between resolved and unresolved is not rare.81 There
are 11 cases of
www 
w
(including 2 ^ e www  w
), and 6 of
www

w
. However, there does not seem to be complete
freedom in the matter of responsion between resolved and unre-
solved positions. A strong tendency can be observed. Either reso-
lution is dominant (as at O2s2, where resolution occurs at 9
repetitions out of 10) or it is a special licence (as at O10s1, where
resolution occurs in 1 repetition out of 10). The ratio of resolved to
unresolved is extreme in one direction or the other.82 This tendency
is a manifestation of the All-but-One rule (see Part III, C).
Longer phrases, too, are fairly frequently resolved. Since all the
intervening brevia of these phrases are short, resolution creates a long

81
^ e (wy): O1s9 (2 resolved out of 8 repetitions), O2s2 (9/10).
e (
wy): O2s3 (3rd e) (9/10), O2s6/7 (2/10), O2e1 (3/5), O2e4 (1/5), P5s2
(1/8), P5s5 (7/8), N6s6b (3/6), N6s7 (5/6), I8s1/2 (2/7).
e (yw
): O2s3 (5th e) (8/10), O2s5 (1/10), O10s1 (1/10), P5e7a (1/4), P5e7b
(1/4), P5e9 (3/4).
cf. e2 (
wyw
): O13s4 (1/10), N3e3 (2/4).
82
12 examples in total out of 19 listed above show this tendency.
50 Introduction
series of continuous short syllables. Unlike resolved e exact respon-
sion is observed in these sequences throughout all the repetitions and
produce a peculiar eVect: 7 shorts (O1s8, w e5), 6 shorts (P2s1, e 6;
P5s4, ^ ee; N7s6, e3, for emendation of the text, see ad loc.).83
About resolution in general, some tendencies are observable.
(i) Resolution does not occur simultaneously at both longs of e: i.e.
there is no example of rwr.84 But in expanded phrases, prolonged
sequences of short syllables do occur:
O1s8 wrwrw
w
w
w
k w e5
P2s1 rwrw
w
w
wrw
k e6
(ii) Between
wr and rw
, and between rw
and
wr, an
unresolved cretic, or an anceps at least, is always present.
There is just one exception for each:
P5s4 wr rw
k ^e e
I8s8 rw

wr
wr
k e e e

The latter (I8s8) would be eliminated if the direction of resolution
were diVerently assumed:
rw

rw
rw
k e
e e
This colometry, however, introduces long anceps followed by re-
solved long, which is not paralleled at all; see (iii) below.
(iii) Resolved long is never preceded by long anceps; i.e. l x rw
is
l
avoided (here x means anceps position Wlled by long; i.e. long
implicit anceps). In other words, when resolution occurs, the pre-
ceding position must be a real long (or occasionally, a short anceps;
see (v) below). There is no exception at all. This is an important
feature, but it is diYcult to explain it in a broader perspective. In
tragic trimeter, resolution occurs after the penthemimeral caesura
more frequently than elsewhere:

83
Continuous short syllables are found in the other type of metre too: 5 shorts (all
are in glyconic; N6e2, P8s2, P11s2b). See further Part III, D.
84
Outside the eighteen majors there are two other cases; one in O4e9/10 (the four
minors, see Part II, Appendix A, ad loc.) and another in Pae6s5a (see Part II,
Appendix B, ad loc.).
 Freer Dactylo-Epitrite 51
x
w
xjrw
x
w
k
(iv) In contrast, resolved long is occasionally followed by long
anceps:
wr l x. There are four examples of resolved longlong
implicit anceps (marked by l x) in mid-verse, one in O2, three in
O10:85
O2s3

w

wr
wy l x
w
tw
k
e e e
e e
O10s3a u
w

wr l
x
wr

j x e e
e sp
O10s4 w

wr l
x
w
k ^ e e
e
O10e3 t
wr
wr l x
ww
y
w
k xee
dxe

wr
occurs at verse-end in Wve verses:86
O10e1b x
wr
k xe

P6s6 www
ww
w

wr
k gl e

P11e5 w
w
w

wr
j w e2 e

N6s6b
wt
k e

I8s8 rw

wr
wr
k eee

(v) Short anceps precedes resolved long in three verses:
P5e7b w tw

j we

P5e9 ta
ww
w

ww
w yw

w
k gl d w e e
O1s8 w rwrw
w
w
w
k w e5
(vi) Short anceps follows resolved long in three verses (which all have
expanded e):
P11s3 rw
wr w
ww
k e2 w d
N3s5

w
w
wr w
ww
k
e3 w d
I8s3
w
w
ww

w
wr w
ww
w
k wil e2 tel
(vii) In one verse, an expanded e is both preceded and followed by
short anceps:
N3e1b w rw
w
wr w
ww
w
w

w
k w e3 w d w e e

85
Outside the eighteen majors there are two other cases. Both occur in Pae4s3, a
verse strikingly similar to O10s3a. See Part II, Appendix B, ad loc.
86
Some scholars, including Kayser and Maas, deny the resolution at the penulti-
mate position of a verse. The argument will be discussed below: Appendix, Adden-
dum II.
52 Introduction
(viii) When e is followed by anceps at the verse-end, the e is resolved
in many instances. Five examples of
wr
k are listed above in the
Wnal paragraph in (iv), and the examples of rw

k amount to
four:
O10e4/5

ww

ww
rw

j
dde

P5e7b w tw

k we

N6s4a ww
ww
ww
rw

k ^D e

N6s6a
ww
ww


ww
rw

k D
d e

Unresolved examples (
w

k) amount to six in total; interest-
ingly four of them occur in O2 (s1, s6/7, e2, e5; and N3e1, I8s5c).
These are characteristics of e. In contrast to e, resolution of d is an
oddity, like that of choriambic nucleus of aeolic metre (5, F). There
are only two examples. In both a resolved long position corresponds
with unresolved at only one repetition:
O10e10 w
wwt

wr
w

ww
k w d
e e d
N3s6

w
a
wwt
k
exd

In O10e10 resolution involves a proper noun (v. 110 105 Sn.
) but not in N3s6 (v. 14 I
). Furthermore, it is very
rare in any kind of metre for double short to correspond with long at
the penultimate position of a verse. For example, there is no pher-
ecratean in tragedy of the shape of
wwt
. I
 is highly
suspect.87 See further Part II, ad loc.

D. Link Anceps
Anceps can be located before, between, and after freer D/e phrases. In
addition, there are some cases in which aeolic phrases are followed by
short anceps and e (in most of them, d being occasionally used).
Here w
w
functions as a prolongation of the ending of the
preceding aeolic. This will be illustrated below. There are in total 61
ancipitia at the beginning of the verse, 45 in the middle, 28 at the

87
But outside the eighteen majors there are two examples of this type of reso-
lution in Pae4. See Part II, Appendix B.
 Freer Dactylo-Epitrite 53
end.88 So anceps is comparatively uncommon at the end of a verse.
This must be related to Pindars general preference in the eighteen
majors. He prefers blunt ending to pendent ending everywhere. For
the statistics of all the verses, taking also aeolic phrases into consid-
eration, see 8. 4.
Anceps may be followed by either (i) d (and D etc.) or (ii) e (and e2
etc.) when it is positioned at the beginning of the verse or in the
middle. It may also be preceded by a phrase of either group, or by an
aeolic phrase, in the middle or at the end of the verse. The position of
anceps makes a diVerence to its realization, as does the nature of the
adjacent group:
(i) at the beginning followed by
d group e group total

9 15 24
w 3 21 24
u 1 12 13
total 13 48 61
(ii) (a) in the middle preceded by
d group e group aeolic total

2 6 1 9
w 7 10 8 25
u 4 6 1 11
total 13 22 10 45
(ii) (b) in the middle followed by
d group e group total

2 7 9
w 14 11 25
u 6 5 11
total 22 23 45

88
These Wgures do not include 3 examples of biceps in which long anceps
corresponds with two shorts. In the following verses, t seems to be substituted
for anceps:
O10e3 t
wr
wr

ww
t
w
k
N6e6/7 t
w
w
w

j
Both verses are embroiled in textual problems, and biceps is not unambigously
attested; see Part II, ad locc.
54 Introduction
(iii) at the end preceded by
d group e group total

8 20 28
Note that the same ancipitia (45 in total) are repeatly counted in (ii)
(a) and (ii)(b). In (iii), I treat all the ancipitia at the end of verse as
realized as long even if the syllable is short.89
The number of explicit ancipitia ( anceps positions in which a
long syllable and a short are actually found in responsion) is in total
24 (13 at the beginning (i), 11 in the middle (ii)) and not as large as
that of implicit ones ( anceps positions in which either a long
syllable or a short is exclusively used throughout all the repetitions,
that is to say, in positions where the form of the verse requires anceps;
see Rule 6). This tendency is similar to the realization of aeolic full
and half-base which tends to keep the same form throughout all the
repetitions of the verse (see above, 5. CD). There are 10 examples of
anceps which is long in one repetition, short in all the others (6 at the
beginning of the verse, 4 in the middle).90 Conversely, there are four
examples of anceps which is short in one repetition, long in all the
others (2 at the beginning, 2 in the middle).91 This is one of the cases
to which the All-but-One rule applies. See Part III, C.
The diVerence between the d group and the e group is evident in
these points:
1. At the beginning of a verse, the anceps tends to be long when it
precedes d, but there is no such tendency in the e group.
2. On the other hand, the anceps tends to be short in the middle of
the verse either when it precedes or follows d. But again, there is no
such tendency in the e group.
Point 2 may be clariWed as follows. Apart from those following an
aeolic phrase, the mid-ancipitia can be classiWed into four classes
according to their surroundings:

89
Some will object to calling this feature brevis in longo because, even when verse-
end is recognized either by hiatus or by the structure of the following verse, it is
possible to take this Wnal position not as long but as anceps. There have been, and will
for ever be, controversies about interpretation of the Wnal anceps.
90
At the beginning: O1s11, O10s5, O13s2, P2e5, P10e1, P10e6; in the middle:
O13s4, P11e6, N3s6, N3e1a.
91
At the beginning: P11e6, I8s10; in the middle: P10s3, N2s5.
 Freer Dactylo-Epitrite 55
(i) d x d both preceded and followed by d group
(ii) d x e preceded by d group but followed by e group
(iii) e x d preceded by e group but followed by d group
(iv) e x e both preceded and followed by e group
In each of the classes (i)(iii), there are examples of (a) invariably
long, (b) invariably short, (c) responsion between long and short; in
class (iv), only (a) and (c) can exist since, by deWnition,

w
w
w
is not eancepse, but e3. In the following table,
(a) and (b) indicate the number of examples, and (c) indicates the
number of examples and, in parentheses, the number of repetitions
of long/short in the verses.
(a) (b) (c)
(i) d x d 1 2 3 (7/1, 3/5, 4/1)
(ii) d x e 1 5 1 (1/3)
(iii) e x d 1 10 3 (3/5, 1/3, 1/7)
(iv) e x e 5 [14]92 3 (3/3, 1/9, 4/6)
Using this table, point 2 above is reformulated, going into details:
2(a). When medial anceps is preceded by d but followed by e (ii), it
is mostly short. Exceptions are one verse (O10e10) and one repeti-
tion in another verse (N3e1a).
2(b). When medial anceps is preceded by e but followed by d (iii),
there is one example in which it is invariably long (O10e3), while
there are as many as 10 which are invariably short. Although this
tendency for anceps to be short is generally kept, there are a small
number of exceptions where the anceps is an explicit one.
If we look closely at actual metrical patterns of long and short
syllables in the six examples of elong ancepse, the following
should be noted:
3. The actual occurrence of
w


w
is remarkably rare,
although there are 5 verses which can be described structurally as e

e, including ^ e
e3 (P5s11) and w e2
e (O13s3).93 In fact, the

92
The number of e3 is given for reference.
93
Besides, the sequence
w


w
is incorporated at some of the repe-
titions in two verses: P6s9 (e2 x e), P8s7 (e x e2).
56 Introduction
long before the long anceps is often resolved.94 Needless to say, e
e
is the most popular combination in normal D/e.
There is a reason for 2(a) above. Short anceps is used in such verses
in a characteristically Pindaric manner. Typical of this group is:
O1s7
w
w
w
w
ww

w
k e3 w d e
The phrase w d is preceded by a long single-short movement (in this
verse, e3), giving the impression that the short anceps forms part of a
series. These verses amount to nine. They are all collected in 8. C. 3.
As in this group of verses, short anceps, in w d and w e alike, works
as if prolonging the preceding w
rightwards:
N3s1

ww
w
w
ww

w
k tel w d e
I7e1
w
ww
w
w
w
k gl w e
Here the short anceps follows an aeolic phrase ending with w
.
Including N3s1, there are 2 examples of this type of w d and,
including I7e1, 4 examples of w e. They are also collected below; in
8, B. 4. 3 and in 8. C. 7 respectively.
I stated above that short mid-anceps follows an aeolic phrase.
There are just two exceptions:
P11s4
atww


wr
ww
k rdod
e d
P10s6 ww
ww
w
a
w
w
k tel x e2 (long at 2 repetitions
out of 8)
However, P11s4 may be analysed in a diVerent manner,95 and P10s6
may also be diVerently interpreted.96 Then they are not altogether
exceptions; see Part II, ad locc.
There emerges an important feature when the positions of word-
ends are examined in relation to link anceps. Pindar has a strong

94
O2s3 (ey
e), O10s3a (er
er), O10s4 (er
e). The long is never
resolved after the long anceps; see above, 6. C. 3.
95
If this verse is analysed as rdod sp rdod, the irregular long anceps would be
eliminated. Verse-end might be set after the spondee in mid-verse. There are other
irregularities in this verse; see Part II, ad loc.
96
The Wrst phrase may not be telesillean but diomedean, a variation of D; see
further 7. 2.
 Freer Dactylo-Epitrite 57
tendency to avoid cut after long anceps at mid-verse.97 Where short
anceps occurs in responsion with long in some repetitions (i.e.
anceps is explicit), long anceps is always followed by bridge. The
extended Porsons Law seems to be applicable here.98 The examples
are found in 11 verses:
anceps A B C D
P10s3 wD xd
7 0 1 0
N2s5 d xd
4 0 1 0
P10s4 ^e d x d rdod e2 3 0 3 2
N3e1a d w d xe
1 0 2 1
N3s6
e xd
1 0 5 2
P11e6 xe xd
1 0 2 1
O9s11
d e xd
3 0 2 3
P8s7 xe x e2 4 0 3 3
O13s4
e2 xe 1 0 8 1
P6s9 ^ e e2 xe 3 0 2 1
P10s6 tel x e2 2 0 2 4
total 30 0 31 18
A Bridge after long anceps
B Cut after long anceps
C Bridge after short anceps
D Cut after short anceps

The metrical contexts are various. So are the phrases both before and
after the anceps. The anceps is realized as long at 30 repetitions
among 79 in total. Cut never occurs after these 30 long mid-ancipitia
(while it does after 18 short mid-ancipitia). But at the same time one
curious fact must be noted. The situation of the anceps which is
realized as long in every repetition is diVerent. This anceps is found in
the context of particular stanza-forms, and is usually, but not invari-
ably, followed by bridge. There are, in total, nine irregular cuts
among 74 repetitions.99 All these verses belong to what I shall call

97
By cut I mean word-end; in contrast, when word-end is absent, I use bridge.
I use these terms loosely: bridge has usually been employed for avoidance of word-
end at a particular place in a metrical pattern where all, or most, of the examples show
the same tendency. My usage is simply descriptive as to whether word-end occurs or
not at a particular place in an example ( a verse).
98
Cf. Parker, CQ, ns 16 (1966), 126.
99
O2s3 (10 repetitions), O10s3a (10r), O10s4 (10r), O10e3 (5r), O10e10 (5r),
O13s3 (10r), P5s11 (10r), P11s4 (8r), N6s6a (6r).
58 Introduction
Class II, except for P11s. In the following illustrations, l
x means
implicit long anceps.
O2s3

w

rw
wt l x j
w
tw
k
e e e
e e
1 case out of 10:
77 ( 70 Sn.)
,  e e
a  j   

O10s4 w

wr l x j
w
k ^ e e
e
1 case out of 10:
92 ( 88 Sn.) K
d
F g j

O10e3 t
wr
wr l x j
ww
y
w
k t e e
d y e
2 cases out of 5:
81 ( 78 Sn.) IE b
 
 j   K


103 ( 99 Sn.) P

  j
E Ke  

Notably in all the three verses above, the position preceding the
long anceps is resolved. Perhaps this may give a diVerent rhythm
from the ordinary case. Two occurrences of cut in O10e3 would
be eliminated if an alternative colometry were adopted; see Part II,
ad loc.
O10e10 w
wwt l x j
wr
w

ww
k w d
e e d
1 case out of 5:
88 ( 84 Sn.) H b 
a j
e
 I
 
This verse includes many irregularities as well, and I tentatively
propose a diVerent colometry; see Part II, ad loc. With that colome-
try the cut would be eliminated.
O13s3 w
wrw
l x j
w
k w e2
e
2 cases out of 10:
3  b 

, j 
25 ( 26 Sn.)  
, j F

,
It is uncertain whether or not it is more than a coincidence that cut
occurs between a short open vowel (note verbal assonance
-

 
) and double consonants in both lines.
P5s11 w
l x j
w
w
w
k ^ e
e3
2 cases out of 10:
22  j  H I,
53 I  j d

.
 Freer Dactylo-Epitrite 59
The avoidance of cut after long mid-anceps is one of the two
reasons why the alternative colometry is rejected at O10s3a (see
above, C). Two analyses of s3a are possible according to where
resolutions are supposed to occur:
(A) u
w
l x rw
l x rw

j x e
e
e

(B) u
w

wr l x
wr

j x e e
e sp
At Wrst (A) seems better in that the inner structure of the verse is well
articulated: ancepse ( iambic metron) is regularly repeated three
times. But it introduces cut between long anceps and the following
resolved long (5 repetitions after the Wrst long anceps and 9 after the
second). This is extraordinary. In contrast, bridge is perfectly ob-
served throughout all the repetition according to (B).
In contrast to long anceps, no preference in the matter of word-
end is observed after short anceps, either explicit or implicit. In 11
verses including explicit anceps, short anceps occurs at a total of 49
repetitions. Both bridge and cut occur indiVerently. As for implicit
anceps, there are in total 151 repetitions in 25 verses, excluding e3 (
e w e). Here, too, there is no observable preference.
60 Introduction

7 . A M B I G U I T I E S B E TW E E N T H E T WO M E T R E S

There are some ambiguities between aeolic phrases and freer D/e.
Five of these concern classiWcation of particular forms which can be
analysed in either way. The last is more serious: despite being asym-
metrical,100 some apparent aeolic phrases may, in fact, be freer D/e.

1. Reizianum /x d x
The ambiguity of the sequence x
ww

has already been men-
tioned above (3, Rule 7). It can be analysed in two ways: (i) as
reizianum, the catalectic version of telesillean and therefore a collat-
eral form with pherecratean, and (ii) link ancepschoriamb
(d)link anceps. The sequence which lacks the initial anceps is
similar:
ww

. This may be (i) an aeolic colon, adonean, or
(ii) choriamb (d)link anceps. The diVerence must have been clearly
audible in ancient times if the Wnal position of aeolic cola in catalexis
was given the value of triseme (T three brevia; for further, see
above on catalexis, 5. A. 4). For these phrases, the metrical context
should help us towards a decision.101

2. Aeolic half-base ww/^ D

A phrase starting with the movement ww
ww
. . . can be analysed
in two ways: (i) aeolic half-base (ww)choriambic nucleus and (ii)
^ D. Thus it is possible for a reizianum of the form ww
ww

to be
^ D
; cf. the reizianum of the form x
ww

above. Unlike
x
ww

, all the examples (two, in fact) of ww
ww

are
classiWed as ^ D. A question arises from its longer phrase
ww
ww
w
. It is, at Wrst sight, blameless telesillean. But some of
the telesilleans of this form may rather be related to D. Pindar once

100
The lack of symmetry is one of the two fundamental criteria which divide
aeolic phrases from freer D/e (above, I. 4). The other is absence of link anceps.
101
I classify 5 examples of x
ww

as x d
, and 9 as reizianum. The
examples of
ww

are rare (2 examples). Both of them I classify as adonean.
See Lists 12.
 Ambiguities between the Two Metres 61
uses ww
ww
w
at the beginning of a normal D/e ode which is
otherwise made up totally of D, e, and link anceps:
N10s1 ww
ww
w
x
w


ww
ww
k
The phrase ww
ww
w
x occupies the location where we should
expect x
ww
ww
x (x D x).102 The two phrases are related to each
other in this fashion:
x
ww
ww
x
ww
ww
w
x
The two double shorts of D go ahead, and the initial anceps, changed
into a real short, comes behind. According to this interpretation, the
apparent rtel is a derivative of D and the two shorts at the beginning
really are two shorts, like those in D. Now consider, for example, the
structural similarity between two verses in P10:
P10s6 ww
ww
w
a
w
w
k (tel) x e2
P10s3 w
ww
ww
u
ww

k w D x d

Beside P10s6 there four other delicate cases: P2s4, N3s8, N3e4 (bis).
All of them are treated as telesillean, but the question is unsolvable;
see Part II, ad locc. This question will be discussed again in a wider
perspective below, 7. 6. N10s1 will be cited there too.

3. Wilamowitzianum/e w d
A long single-short movement is deWned as a single unit (I. 3, Rule 8)
of freer D/e, like
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
O1s7
w
w
w
w
ww

w
k e3 w d e
Following this rule, it is possible to analyse the standard type of
wilamowitzianum,
w
w
ww
, in the same manner (e w d).103

102
I argued for this verse, ww
ww
w
x, in tragedy, collecting the examples:
BICS 38 (19913), 24361, giving it the name diomedean (248). It is extensively
employed later by Euripides.
103
There are in total 9 verses which include this type of wilamowitzianum.
Interestingly, their occurrences are concentrated in particular stanza-forms (e.g. 3
in P6s and 3 in I8s).
62 Introduction
Then the sequence is not aeolic but freer D/e. One of the most
delicate cases is
P10s5 w

w
w
ww
w

k ^e wil3
which may be better analysed as ^ eehagesichorean ( w dod
);
for ^ ee, see further below, 8. B. 6. I categorically reject this analysis
in Part I. Ambiguous cases will be examined in Part II.104

4. Aeolic full base w
/^ e

w
at the beginning of a verse can be taken in two ways: the aeolic
base or acephalous e (^ e). In the case of w

ww
, I take all the
examples as ^ ed (see above, 5. A. 3). As for w

ww
w
, the
situation is not clear-cut. In some contexts it is certain that (i) this
form is glyconic (pure aeolic); but in others it may be (ii)
^ edodrans (composite). I have tentatively discussed this problem
above (5. C) citing N2s1 and P2e1be3 for the certain examples of
(i), and P5e2 and N7s1 for (ii). Similarly, the wilamowitzianum in
I8s5a, w

w
ww
, may be ^ erdod. Here is another illustration
of this type of ambiguity:
O1s1 w

ww
w

w
ww

k
At Wrst sight, this is an innocent priapean dicolon. But (i) the
examples of priapean dicola are extremely rare, the other two
examples being P8e3/4 (
xgl

ph), N2s4 (

gl
xph); (ii) aeolic
phrases are scarce in O1; (iii) palindromic movement (for which see
8. A. 6) is conspicuous. And, although this is in general a Pindaric
tendency (5. C), (iv) the exact strophic responsion at the base (w
at
the glyconic,
w at the pherecratean at all the 8 repetitions) must be

104
Outside the eighteen majors I analyse the phrase in these verses as e w d:
O4e9/10 rwr w
ww
w
w
w
k e w d w e2
Pae6s5 rwr w
ww
w rw
j
ww
j e w d w e d
They are similar in that (1) they start with 6 continuous shorts and (2) single-short
movement extends rightward too. There is no example of totally resolved e wwwww
in the eighteen majors, but resolution occurs in expanded phrases, e.g.
P2s1 rwrw
w
w
wrw
k e6
See 6. C. 1.
 Ambiguities between the Two Metres 63
taken into account.105 Moreover, (v) w
at the beginning of this
verse harmonizes with other verses of this ode (s2, s9, s10, e2, e4, e7)
which start with w

w
(and www
w
) : ^ ee. (vi) Between
the full 2 ending of the glyconic and the full base of the pherecra-
tean, bridge is not observed at as many as four repetitions. This is
against the general tendency (cf. 8. B. 1). Last but not least (vii) the
similarity with the preceding verse (i.e. the Wnal verse of the epode) is
illuminating:
O1e7 w

w

ww
w

ww
w

k ^e e dod ar
O1s1 w

ww
w

w
ww

k
(the last colon of O1e7, aristophanean, is dod1 by deWnition in this
book; see above 5. A. 5). O1s1 is, then, ^ edodrdod1, i.e. w
is
not the base of a glyconic, nor
w of a pherecratean.
However, so delicate a classiWcation risks producing too many
borderline cases. Thus, for this type of phrase, I reject (ii) and accept
(i) and leave ambiguous cases as ambiguous. In the tables and
parallels given in Part I above, I treat all these examples as aeolic
phrases.

5. Pherecratean/rdod1
The apparent pherecratean in O1s1 is a related, but diVerent matter.
I have already raised the question (5. A. 5) whether
w
ww

is
really pherecratean in all contexts. The example I cited there was
N3e2
w
ww

w
ww

k
This verse may be analysed as two reversed dodrantes, the second
followed by anceps (rdod1). In O1s another pherecratean is sand-
wiched by two lecythia (e3):

105
The bases of the two cola (w
and
w) are in reversed order, and conse-
quently make a strong contrast. This usage is diVerent from the standard in tragedy
in two ways: Wrst,
w is often in responsion with

in tragedy (and so is w

with

though to a lesser extent). See Itsumi, Glyconic. But in O1 there is no
example of

in either colon. Secondly, in tragedy, if the glyconic starts with
w
, the following pherecratean usually starts with

. In short,
w does not
coexist with w
.
64 Introduction
O1s3
w
w
w
k e3
O1s4
w
ww

k ph
O1s5
w
w
w
k e3
It is a tempting suggestion that e3 and this pherecratean are the same
with exchange of the Wfth and sixth positions. Whether the Wnal
position of this pherecratean is triseme created by catalexis or anceps
is a recurrent question (see 1 above). Again borderline cases could be
produced, and I have rejected analyses like rdod1 in the tables and
parallels given in Part I.

6. Expanded aeolic
This last ambiguity does not concern the classiWcation of particular
forms in particular contexts but the deWnition of diVerent metres,
and, consequently, the historic perspective of the evolution of Greek
metre in general: namely, the question whether there is aeolic expan-
sion in Pindar. One might expect, after reading Snells Metrorum
Conspectus or Wests description, that there would be extensive use
of the dactylic expansion of aeolic cola in the Pindaric corpus. But
the examples are disappointingly scarce. What we might imagine as a
typical phrase would be, for example, gld, with which Snells table
begins:

ww
ww
w

which is absent, in fact, from the eighteen majors.106 Not only are
examples of expanded cola conspicuously few in number, but also
their range of variation is surprisingly narrow.107 I count only six
examples.

106
This phrase, gld, concludes a short strophe in Sappho (94 LP): gl k gl k gld k.
Here, the unexpanded and expanded formsare arranged side by side. This is the most
certain attestation of expansion.
107

ww
ww

is not an expanded adonean, of course, but D and anceps.
And w

ww
ww
k at O10e2, I8s9 is not an expanded reversed dodrans
(w

ww
) but acephalous e D. Certainly, there are a number of Ds which
can be explained best as an expanded form of d. But the relation between d and D,
and the longer form D , is a diVerent topic from the dactylic expansion of aeolic cola.
 Ambiguities between the Two Metres 65
Of these the most plausible examples amount to only three; the
notations in parentheses are given for conveniences sake, based on
the analysis taken in the following paragraphs:
[1] N6e4
ww
ww
ww
w
j (dod2d)
[2] O10e9

ww
ww
w
k (teld)
[3] P2s4 ww
ww
ww
w
ww
ww
w
k (teld) tel
[1] N6e4 could be dodrans expanded by two dactyls, [2] O10e9 tele-
sillean expanded by one dactyl, and so too the Wrst phrase of [3] P2s4,
with ww as base. This could be one of the two cases (the other is N6s3
below) in which an aeolic colon and its expanded cognate are located
side by side. There is a fourth example, though a diVerent analysis is
possible:
[4] N6s3 www
ww
www
ww
ww
ww
j rdod (rdod2d)
The latter phrase could be reversed dodrans, like the former, but with
expansion by two dactyls.108 Two further verses might be included in
this category. They are of a very unusual shape (but interestingly,
similar to each other even apart from expansion):
[5] O9e7

w
ww
ww


k (heptad) sp
[6] P11s1

w
ww
ww
ww
w


k ((hepta2)2d) sp
Thus the number of aeolic verses which have dactylic expansion
could be six at most.109
108
N6s3 has been analysed as gl D by many, supposing the eighth position of
the glyconic to be resolved. See Part II, ad loc.
109
If an emendation metri causa might be accepted, a seventh example would be
added:
O10e3 t
wr
wr

ww
y
w
k
In four repetitions out of Wve the latter part is the same as
[2] O10e9

ww
ww
w
(teld)
The exact responsion is broken only at the Wfth repetition: v. 103 ( v. 99 Sn.)
-
E Ke  
 ( is not found in the older MSS but was inserted
by Moschopoulos). Here the second double short is replaced by a long. This cannot be
contraction, for the double short of the choriambic nucleus of an aeolic colon is not
contracted into a long in the whole of Greek poetry, unlike in dactylic movement.
Admittedly, 
 is a proper noun, but the licence is extraordinary nevertheless.
I am tempted to think that 
 was originally a marginal gloss (referring to v. 2


E) which took the place of the true reading; e.g.
e PF , and
perhaps asyndeton is preferable. If 
 is kept, another analysis is required.
Then the possibility of expanded aeolics disappears. See further Part II, ad loc.
66 Introduction
There is an interesting feature common to these six verses: the
metrical context. None of them occurs among ordinary aeolic
phrases. Instead, there is a stronger association with freer D/e. N6e,
O10e, N6s, and O9e belong to the group of poems composed mainly
of freer D/e verses (the group of stanza-forms which will constitute
Class II in later chapters). P2s and P11s belong to the group in which
the two metres are closely amalgamated (Class III). It is especially
suggestive that Wve of the six verses have true dactylic phrases,

ww
ww
(D) or
ww
ww
ww
(D ), in their preceding or
following verses (N6e3, O10e8, P2s3, N6s4a, O9e6).110 So it seems
that expanded aeolic cola are associated with freer D/e more closely
than with aeolic proper.
Perhaps, indeed, these examples are not, in fact, aeolic nor related
to Lesbian metre at all. What we suppose to be expanded aeolic may
be a modiWcation of D or D . The chief reason for classifying them as
aeolic is the asymmetrical cadence [1, 2, 3, 6] and introduction [4, 5,
6]. But a few asymmetrical phrases are found in the normal D/e odes
of Pindar in three, or possibly four, passages:
[7] O6s5

w


ww
ww
w
k
[8] O7e2
w


w


ww
ww

ww
ww
w
k111
[9] N10s1 ww
ww
w
x
w


ww
ww
k
[10] N8s1
x
ww


w

j
[11] O13s6 x
w
ww


w


ww
ww

w
k
[11] O13s6 could be dismissed as a special irregularity because the
asymmetrical phrase is located at the turning point from non-D/e to
D/e. [10] N8s1 is usually emended and does not appear in the texts of
Snell or Turyn, but I believe the paradosis should be accepted.112 The

110
P11s1 is the exception. It is followed by hepta. Expansion, or more correctly
speaking, contraction, is evident here.
111
Colometry is uncertain; two other colometries are possible, but an asymmet-
rical phrase never disappears by any means; see Part III, A, ad loc.
112
The paradosis oVers at v. 40 h  I
; E K 
 , in which sense and syntax are blameless, but the Wrst part scans

w
ww
, an unconventional phrase in D/e. The other Wve repetitions give



ww
in responsion there. Snell emends to I  I
; E
K  , ending the verse with a lacuna of

. He then takes
 Ambiguities between the Two Metres 67
other three are certain, and interesting. The Wnal parts of [7] O6s5
and [8] O7e2 are equal to dodd or teld: (
)
ww
ww
w
. In [9]
N10s1, the phrase ww
ww
w
x appears where we should expect
x
ww
ww
x (D); cf. (2) above.
Pindar may have followed the Dorian tradition. Stesichorus and
Ibycus use a kind of asymmetrical phrase in the vicinity of dactylic
verses, or, it may be said, in prototype D/e. Note the epode of Ibycus
282 P:


ww
ww

k


ww
ww

k
y
ww
ww

k

w
ww
ww
ww

k

ww

ww
w
k
The fourth verse starts with
w
ww
ww
. Its similarity to [5]
O9e7 and [6] P11s1 is obvious.113 And the Wfth ends with

ww
w
.
Ibycus uses the so-called ibycean,
ww
ww
w
, alongside
dactylic phrases in another ode (286 P). The ibycean is often
taken as an aeolic colon, but it must be diVerentiated from a later
development of glyconic.114
Compared with Ibycus, the inXuence of Stesichorus is less con-
spicuous. In his S148 (Eriphyle, P. Oxy. 2618), the following verse is
found (str/ant. 1):

w


w
ww
ww
ww
ww
ww
ww
k
Note the similarity of the long phrase
w
ww
. . . to [5] and [6]
above. And also in his Iliou Persis,

ww
as a contraction of D, which is, however, totally unparalleled. If

ww
is accepted by emendation for all the repetitions, the initial


would be interpreted as a spondee followed by d, and then Christs claim, veterem
lectionem o glossa h summotam esse, would be preferable to Snells
emendation. See Part III, A, ad loc.
113
And the Wnal spondee of [5] and [6], which adds a further curiosity, may be
related to some verses of Alcman and others, collected by West, GM 523.
114
In his later plays Euripides occasionally uses
ww
ww
w
and

ww
x
ww
in aeolic contexts as variations of glyconic and wilamowit-
zianum.
ww
ww
w
is sometimes found in responsion with
www
ww
w
. See Itsumi, Glyconic, 72.
68 Introduction

y
ww
y
w

k
seems to be mingled (str/ant. 3) in the D/e context. But here double
short is in responsion with long. In these cases the possibility of link
biceps cannot be denied.115 If we had more of Stesichorus or Ibycus,
we might trace the origin of freer D/e to them, and Wnd some
asymmetrical phrases in freer D/e. Then so-called expanded aeolic
in Pindar might be related, not to Sappho and Alcaeus but to
Stesichorus and Ibycus. Or Pindar may be connected, in these
cases, with the Aeolic tradition via Ibycus. Ibycus moved to the
eastern Aegean and worked in the same ambience as Anacreon.
That means that he will have known aeolic verse.
Pindar seems to have used expanded aeolics elsewhere more ex-
tensively. There are in total 12 verses in his Paeans that appear to be
aeolics expanded by dactyls; see Part II, Appendix B. They are, in fact,
ibyceans and its prolonged forms. The metrical notation inevitably
associated with aeolic may not be suitable; Ibycean (ibyc) sounds
less tendentious. I therefore introduce two new terms:
expanded ibycean (ibyc ; cf. D ) for
ww
ww
ww
w
( dod2d)
reversed ibycean (ribyc; cf. rdod) for
w
ww
ww

Hereafter I shall notate the six examples in the eighteen majors in the
following manner, using and modifying these terms:

115
It is maintained by many that single short, long, and double short are well
attested as freely in responsion at link positions (cf. West, GM 49, 52); however, I
must stress that responsion between single short and double is precariously based on
some papyrus readings. The responsion at the Wrst position of the iambic trimeter of
tragic dialogue is, of course, another matter. And so is the beginning of the arche-
bulean of Callimachus (228 PfeiVer).
1. Stesichorus 222(b) (the Thebaid), 215 . Its -- corresponds with a
long in most repetitions and a short in one (285
).
2. Bacchylides 17 ( ), (i) str. 12. v. 101 : r responds with w in 2
repetitions (vv. 12, 78) and with
(v. 35). (ii) str. 17. v. 40 : r responds
with w in 2 repetitions (vv. 17, 106) and with
(v. 83). However, the double short in
both cases,  and , is easily removed; see Housman, CR 12 (1898),
138 Collected Papers, ii. 460 and Maas, Freiheiten II, 19. The transposition of
 can also remove metrical lengthening of . Snell keeps the papyrus readings
and the responsion.
 Ambiguities between the Two Metres 69
[1] N6e4 ibyc
[2] O10e9
ibyc
[3] P2s4 ww ibyc tel
[4] N6s3 rdod ribyc
[5] O9e7
ribyc sp
[6] P11s1
ribyc 2 sp
It would be better to follow Rule 3 above and to divide [3] P2s4 thus:
P2s4 ww
ww
ww
j w
ww
ww
w
k ^ D wibyc

Then it might be better to analyse N3e4 in the same manner:

[7] N3e4 ww
ww
j w
ww
ww
w

w
k ^ D wibyc e
The discussion will be resumed in Part II, Appendix B.
70 Introduction

8. C OM P O SI T I O N O F V E R S E S

The characteristics of each phrase have been discussed in the previ-

ous chapters. This chapter treats verses as a whole. It is divided into
three sections. In the Wrst, A, I discuss topics common to aeolic verses
and freer D/e verses, like verse-lengths or palindromic movements.
Especially important is RSS (ratio of short syllables to long), which
will be one of the strongest criteria to classify stanza-forms in the
next chapter. In the following two sections combinations of phrases
are discussed; aeolic verses in section B, freer D/e verses in section C.
Pindar shows his distinctive ways of making up verses of more than
one phrase. This distinctiveness is evident in particular patterns of
long and short syllables that are shared, in spite of various names
given to the phrases, by a number of verses either completely in the
same order or with slight modiWcations. All the examples are col-
lected for each distinctive pattern, and these collections function as
the parallels for each verse and will be frequently referred to in Part II.

A. General
1. Demarcation of verses
Hiatus and/or brevis conWrms verse-end when word-end occurs
consistently at all the repetitions. When both hiatus and brevis are
absent but word-end coincides, there are two options: (1) to recog-
nize verse-end there and to establish two independent verses, or (2)
to establish a long, united verse. Recognition of verse-end depends
on whether two separated verses are commoner, or in other words,
appear to be more natural, than one united long verse, and moreover,
whether the overall structure of a stanza-form becomes clearer by
dividing the verse. There are 235 verses in the eighteen majors. Of
these, 195 are marked by hiatus and/or brevis at the end and 40 verses
are not. Besides, there are 17 other places in 16 verses where I do not
recognize verse-end in spite of the coincidence of word-end without
exception at all the repetitions.116 Nobody would divide
116
Of these 16 verses, Boeckh separated 6 verses into two while I unite into one.
(And Boeckh wrongly separated another 4 verses, in spite of enclitics and/or elision.)
At the other 11 places in 10 verses, Boeckh and I agree on not recognizing verse-end.
 Composition of Verses 71
O2e3 rw

w
w
rw

jw
k
into two simply because of the coincident word-end before the
penultimate position, but there remain borderline cases. For
example, I divide N6s1a and N6s4a into two verses, although many
treat them as one:
N6s1a w

j ^e x
N6s1b
w
ww
w
rw
k gl e
N6s4a ww
ww
ww
rw

j ^ D e x
N6s4b
w
k e
And, on the other hand, I unite O9e1 with e2:
O9e1/2 w
w
w
w
jww
w
w
w

k w e2 w d w e2

The division of N6s into two verses makes its overall structure clearer.
The united verse O9e1/2 relates better to similar verses in other
stanza-forms. For further argument, see Part II ad locc. Colometry
is relatively unambiguous in non-D/e, aeolic or freer D/e alike. The
situation is diVerent in the normal D/e, since phrases, and conse-
quently verses too, are homogeneous. The diYculties will be de-
scribed at length in Part III, A.
Brevis in longo is not always a self-evident criterion. Brevis in longo
and recognition of verse-end are somehow interdependent when the
supposed verse-ends are pendent.117 For example, the Wnal position
of N6s4a above is occupied by both a long syllable and a short. The
short syllable in that position is brevis in longo if N6s4a is separated
from N6s4b. On the other hand, it would be link anceps if N6s4a
were united with N6s4b. Here too we should observe other verses. As
is demonstrated above (6, D), long anceps in mid-verse is hardly ever
followed by word-end in the eighteen majors, and, moreover, cut
never occurs after long anceps when it is in responsion with short
anceps. This gives support to setting verse-end after the anceps.118

117
A good illustration is given by Stinton, Pause and Period CQ, ns 27 (1977),
38 Collected Papers, 324.
118
Including N6s4 there are in total 7 verses which end with link anceps but
without hiatus between it and the following verse. There is a possibility that the Wnal
link anceps of these verses should be either mid-link anceps, when the initial position
72 Introduction
Even coincidence of word-end is not always deWnitely attested,
because deWnition of a word is not absolutely clear. There are cases
where a weak word like a preposition is located at the end of the
verse, while the verse-end is manifestly guaranteed at that point by
one or more other repetitions. Therefore a theoretical possibility
remains, which we must be prepared to accept, that there may be
verse-end even within a word-group which is normally inseparable.
But, of course, this type of verse-end should be avoided as far as
possible.
A verse may be made up of just one phrase or of several phrases.
When a verse is made up of more than one phrase, it does not
necessarily mean that those phrases are of the same metrical type.
Two metres, aeolic and freer D/e, can make up three types of com-
bination: verses composed (i) of an aeolic phrase or phrases only; (ii)
of a freer D/e phrase or phrases only; (iii) of both metres. I have
already called (i) and (iii) pure aeolic verses and composite aeolic
verses respectively (p. 22). (ii) is called freer D/e verse. The 235 verses
of the eighteen majors are classiWed thus:
(i) pure aeolic 62 verses
(ii) freer D/e 101 verses119
(iii) composite aeolic 72 verses
total 235 verses

2. Identical verses
Of the 235 verses in the eighteen majors, 142 are unique. The variety
of non-D/e verses in Pindar is certainly greater than that of D/e
verses. Nevertheless, this variety should not be over-emphasized,
since many unique verses are more or less closely related to others.
DiVerences sometimes amount to no more than the absence of one

of the following verse is a true long, or even a true long, when the initial position of
the following verse is a short or an anceps. Of the 7 verses, it is structurally
implausible that O10e4/5, P5e7b, N3e1, and N6e6/7 (corrupted text) should be
united with the following verses. For N6s1 and O2e5, there is a chance of combin-
ation; see Part II, ad locc.
119
This includes 6 verses which could in theory be expanded aeolic but are
classiWed here as freer D/e. See 7. 6.
 Composition of Verses 73
anceps or the addition of a Wnal phrase. In fact, family resemblances
between verses are easily recognizable.
The remaining 93 verses (39.6%) are repeated once or more (here
I ignore diVerences such as realization of link anceps, the form of
aeolic base, resolution). See List 3. It will be noted that the same verse
may be repeated consecutively (e.g. P2e1a, e1b, e2) or appear in a
number of diVerent stanzas; gle, which is found in seven verses and
is also incorporated in another three, can legitimately be regarded as
one of Pindars favourites.

3. Verse-length
The average length of the 235 verses is 12.2 positions. This is equiva-
lent to e.g. two dodrantes, or x eglyconic, or x D x e. Whichever
metre they may be composed in, non-D/e verses are generally short.
This is a signiWcant diVerence from the verses of D/e.120 No pure
aeolic verse consists of more than three phrases. Composite aeolic
verses are shorter: there is no verse which contains three or more
aeolic phrases plus e or other freer D/e phrase(s). For the sake of
convenience, I hereafter mean by the longer verses those that have
20 or more positions, although to separate 20 positions from 19 is, of
course, arbitrary.
The longer verses number 15. They are: (i) four pure aeolic; (ii)
nine composite aeolic; (iii) two freer D/e. In the following lists, the
Wgure in parentheses represents the number of positions.
(i) Pure aeolic
N2s4 www
ww
w



ww
w

x
ww

k gl gl ph (23)
I7s5a


ww
w


ww
w
x
w
ww
j gl tel (22)
hepta

120
It is less easy to set verse-end in D/e verses, for, before/after link anceps, word-
end occasionally coincides at all the repetitions without hiatus/brevis, as at O3s3 x D
x j e x D k or O3s4 x e x j e x e x D x e k. Moreover, even if a short syllable is located
at a link anceps which can be the last position of the verse, it does not necessarily
mean that brevis in longo occurs. If all these verses are divided into two or more, the
number of verses is 322, and the average length is 13.3 positions. On the other hand if
all these are connected, the number of verses becomes 226 and the average length
rises to 18.5 positions. The problem is fully discussed in Part III, A.
74 Introduction
P2s2 www
ww

x
ww
w


ww
w
k rdod (21)
gl tel
I8s5a w

w
ww

w
ww

w
ww
j wil rdod (20)
rdod
Only these four verses are made up of three aeolic phrases; there is no
verse with four phrases and fewer than 20 positions.
(ii) Composite aeolic: divided further into four sub-groups.
(a)
I7s3/4

w
x
ww
w



ww
w
w

k x e tel (22)
gl3
O9e8

w
x
w
ww
x
w
ww
w

k x e hepta (21)
hepta3
P6s1/2 w
w
www
ww
w

w
w
ww
k x e gl (20)
wil
In all three, anceps and e precede two aeolic phrases.
(b)
I8s3
w
w
ww

w
wr w
ww
w
k wil e2 tel (20)
P2s8

ww
w

w
w
w
ww
w

k tel e2 hag (20)
e2 (
w
w
) is sandwiched between two aeolic phrases, function-
ing as an extensive leftward prolongation of the following aeolic
phrase, which has a short half-base.
(c)
N7s2
ww
www w
w

ww
w

w
w
k dod w e (21)
dod e2
I8s1/2 w

w
ww
w
wt
w
w
ww
k wil w e wil (20)
w e, sandwiched between two aeolic phrases, in eVect prolongs the
preceding aeolic phrase rightwards.
(d )
N3e3
ww
w
w
ww

w
ww

wtw
k dod w d (22)
rdod e2
P10s4 w

ww
x
ww

w
ww

w
w
k ^ e d x d (22)
rdod e2
These two verses are structually identical in the last 16 positions. It is
highly unusual to Wnd this degree of resemblance in long Pindaric
verses outside D/e.
 Composition of Verses 75
(iii) freer D/e
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2 (21)
N3e1b w rw
w
wr w
ww
w
w

w
k w e3 w d w e e (20)
A prolonged sequence of single-short movement precedes the short
ancepsd.
The above 15 verses may also be divided into two groups another
way. Verses of groups (i) and (ii) (a) feature phrases with full aeolic
base in the form

,
x, or w
, and also half-base in the form
.
They thus tend to include more long syllables than those of the other
groups, giving a heavier impression.
The other eight verses (ii) (b), (ii) (c), (ii) (d), and (iii) are similar
to each other in structure. Short anceps, whether it be link before e or
d, or half-base, is not easy to distinguish from true short. Thus
phrase-boundary is not clear, and phrases tend to run together
neatly. In three of the eight verses, e2 is employed in a particular
manner, being juxtaposed with aeolic hexasyllables (
w
ww
/

ww
w
), which diVer from it only in having one double short
where e2 has single short; the similarity is fully exploited.
As will be discussed later, there is a diVerence between the metrical
contexts of the two groups. Most of the former seven verses are used
in the Class I stanza-forms (aeolic) while all the latter eight verses are
in Class III (amalgamated); and there is no example of the longer
verses in Class II (freer D/e), even when these verses are freer D/e in
character.
So much for the longer verses. Next are the shorter verses, mean-
ing those that have six positions or fewer. There are in total 17 of
these.
Two are aeolic:
P10e2
ww

k adon (5)
O9s9

ww

k reiz (6)
These two may be analysed diVerently, i.e. as (
) d
(
ww
)anceps (the unsolved ambiguity, 7. 1). Interestingly,
there are no examples where a dodrans or a reversed dodrans stands
as a verse by itself. This may be purely accidental, but may be
meaningful, since both are common cola and used frequently in
combination.
76 Introduction
Two others include d or D:
O10e8 ww
ww
k ^D (6)
I7e6
ww


k d sp (6)
The other 13 verses are made up of e, or its cognates, and anceps:
N6s4b
w
k e (3)
N6s1a w

j ^e x(3)
P5s7a


k x sp (3)
N6s6b
wt
k ex (4)
P5s4 wr rw
k ^e e (5)
P5s6 w

w
k ^e e (5)
P5e7a w
tw
j ^e e (5)
O10e1b x
wr
k x e x (5)
P5e7b w tw

j x e x (5)
P5e8
w

w
k ee (6)
O2s4
w
rw
k ee (6)
I7s5b,O10e7

w
w
k x e2 (6)
Nine out of the 17 examples are found in four stanza-forms (3 each).
Of these O10e, P5s, and N6s belong to Class II (freer D/e), and P5e,
to Class III (amalgamated).

4. Pendent and blunt endings

Pindar prefers blunt ending to pendent. Of 235 non-D/e verses, 165
verses (70.2%) end blunt and 70 verses (29.8%) pendent. When
verses end with an aeolic phrase, the diVerence is not signiWcant.
Out of 86 verses, 46 end blunt and 40 pendent. But there is a great
diVerence when the Wnal phrase of a verse is not aeolic but freer D/e,
like e or d: 119 verses end blunt, without Wnal anceps, while 30 end
pendent, with anceps.121 The preference for e at the verse-end is
especially remarkable: 89 verses (37.9%) end with e or its prolonged
cognates (e2, e5, e6).122 In contrast, only 20 verses end with e (e2, e3)
and anceps.

121
In this Wgure (119 verses), those 6 ambiguous verses (7. 6) which could be
expanded aeolic but is taken as freer D/e are included. They are all blunt, anyway.
122
The 6 verses which end with a spondee are excluded.
 Composition of Verses 77
Thus verses ending blunt predominate in many stanza-forms.
Only six stanza-forms out of the 31 in the eighteen majors have a
majority of pendent verses,123 and three of these belong to Class I.
P8s (also Class I) has no blunt verse-ends at all.

5. Ratio of short syllables (RSS)

When a verse includes many short syllables, it has light movement;
when light verses are repeated, the stanza-form feels light. A short
syllable is produced:
(i) in a short position;
(ii) when anceps is realized as short;
(iii) when a long is resolved into two shorts.
Thus, if a stanza-form includes (i) many short positions, (ii) short
ancipitia, (iii) resolutions, the rate of occurrence of short syllables,
which is hereafter abbreviated as RSS, is high.
RSS is calculated in the following manner: the number of short
syllables (not positions) divided by the total number of syllables (not
positions). Anceps position is counted either as long or as short,
according to its actual realization. Resolution is counted as two short
syllables. When long and short are in responsion at a particular
anceps position, and when resolved and unresolved long are in
responsion at a long position, the majority prevails.
RSS is calculated for all the stanza-forms. See List 4, in which not
only non-D/e but D/e values are calculated. In the latter, RSS is
remarkably low, mainly because of long link anceps and absence of
resolution. The average of the 40 stanza-forms of D/e is 40.1% (P4,
the longest epinikion, is typical: P4s 40.8%, P4e 40.0%). The highest
is 46.9% (O7e) and the lowest 34.3% (N5s). In contrast RSS of the
eighteen majors is higher. Even the lowest (O9e) is 42.1%. The
highest is as high as 61.8% (N7e). The average of the eighteen majors
is 50.9%. It is about 10% higher than that of the 40 D/e stanza-forms.

123
N2s (4 pendent verses/5 in total), O9s (7/10), O2e (4/6), O13s (3/5), P8e (3/6),
P11e (3/6).
78 Introduction
RSS helps our vague impression to become visible statistically.
This will serve as one of the powerful criteria for classifying stanza-
forms. In general, RSS is high in most cases of Class III (amalgam-
ated) and in some of Class II (freer D/e), but low in Class I (aeolic).
The minimum and maximum values in each class are: Class I 42.4%
(P8e)53.6% (P6s); Class II 42.1% (O9e)58.0% (N6s); Class III
50.0% (O1e)61.8% (N7e).

6. Palindromic movement within verse

By palindromic movement is meant a sequence composed of two
phrases which repeat each other in reversed, as for example,
O1s1 w

ww
w

w
ww

k

gl ph

Many more verses are partially palindromic. The following verses

include:
(a) . . .
ww
w

w
ww
. . .
P2e1a www
ww
w

w
ww
j
gl rdod

P2e1b w

ww
w

w
ww
k

gl rdod

P2e8 w

w
ww
w

w
ww
w

k

wil2 hipp

P8e6 w

ww
w

w
ww
w
k

gl gl

N7s8

ww
w

w
ww
w

k

tel hipp

cf. O13s5 ww
w

w
ww

k

^ dod ph

There are some other types of palindromic movements, such as:

(b) . . .
ww
w
w
ww
. . .
N3e3
ww
w
w
ww

w
ww

wtw
k

dod w d rdod e2
N3s1

ww
w
w
ww

w
k tel w d e

N4s4
w
ww
w
w
ww

k gl reiz

O1e4 w

w

ww
w
w
ww

k

^ e e dod reiz

(c) . . .
ww
x
ww
. . . and . . .
ww
w
ww
. . .

N2s5
ww
x
ww


k dxdx

N4s5
x
w
ww
x
ww
w
k

wil tel

P10s3 w
ww
ww
x
ww

k wDxdx

P10s4 w

ww
x
ww

w
ww

w
w
k ^ e d x d

rdod e2
 Composition of Verses 79

ww
w
ww
x
w

j dwdxex

N3e1a

ww
ww
w
ww
ww
w

w
k

N3e4 tel tel e (or

^ D w ibyc e)

x
x
ww
w
ww

k

N4s6 wil reiz

P2s4 ww
ww
ww
w
ww
ww
w
k

teld tel (or
^ D w ibyc)

(d) . . .
w
w
ww
w
w
. . .
O9e1/2 w
w
w
w
ww
w
w
w

k w e2 w d w e2 x

(e) . . . (
) w

w
ww
w

w (
) . . .

N3s4 w
w

w
w w
w

w
k
w e gl e

P2e8 w

w
w w
w

w
ww
w

k

wil2 hipp

(f) . . .
ww
w
x
w
ww
. . .
I7s5a


ww
w


ww
w
x
w
ww
j gl tel hepta

The following types include resolution, yielding the sequence

www
:
(g) . . .
ww
www
ww
. . .
N6s3 www
ww
www
ww
ww
ww
j rdod ribyc

P5e3
w
ww
www
ww
j

rdod rdod

P5e6
ww
www
ww

w
j d rdod e

(h) . . .
w
www
w
. . .
N7s7
tww
w
www
w
k

tel e2

O2e2
w
w
www
w

w

k e3 e e x

O2s8

w
www
w
w
ww
k x e e2 w d

(i) . . .
www
www
. . .
O2s2 wy
w

wr
wr
w
k

^e e e e e

t
wr
wr

ww
y
w
k x e e

O10e3

dxe
I8s8 rw

wr
wr
k eee

Perhaps the palindrome was not only audible but visible, if the
dance steps went, as it were, in reversed.
80 Introduction

7. Repetition within verse

It has already been mentioned that the combination of glph is
remarkably rare within a verse (3 examples). Not only glph but
repetition of identical or similar cola is rare. For example, glgl is
found in only 3 verses (including glgl3). Even when the identical
colon is repeated, the aeolic base is often changed. For example,
N2s4 www
ww
w



ww
w

x
ww

k gl gl ph
the Wrst glyconic of N2s4 starts with www while the second with

,
and for the Wnal pherecratean,
w is used at 3 repetitions out of 5.
There are even verses in which two aeolic cola have w
and
w
respectively, like O1s1 (w
gl
wph) or P8e6 (w
gl
wgl). This
combination is very rare in tragedy.
The following Wve verses therefore merit special mention in that
hexasyllables of identical form, sometimes even with identical base,
are repeated within the verse. Most examples belong to Class III
stanza-forms (amalgamated):
N7e5 www
ww
w
www
ww
w
w

k gl gl3
N3e4 ww
ww
w
ww
ww
w

w
k tel tel e
O9e8

w
x
w
ww
x
w
ww
w

k xe hepta hepta3
I8s5a w

w
ww

w
ww

w
ww
j wil rdod rdod
I8s5b
w
ww

w
ww
j rdod rdod
Whether they are freer D/e or expanded aeolic (7. 6), repetitive
structure is evident in the following two:
P2s4 ww
ww
ww
w
ww
ww
w
k (teld) tel
N6s3 www
ww
www
ww
ww
ww
j rdod (rdod2d)
In the following three verses, not whole phrases but parts of phrases
are repeated:
O1e7 w

w

ww
w

ww
w

k ^ e e dod ar
N3e2
w
ww

w
ww

k rdod ph
P2e8 w

w
ww
w

w
ww
w

k wil2 hipp
These are all there are, and the rarity is striking. Contiguous repeti-
tion is fundamental to Greek poetic rhythm, but Pindar deviates
from it to a great extent.
 Composition of Verses 81

B. Aeolic Verses: Pure and Composite

1. Contact between two aeolic phrases
There is no verse which has more than three aeolic phrases (above,
A. 3). Within that limit, the number of aeolic phrases within a verse
varies. Pure and composite aeolic verses are classiWed thus accord-
ing to the number of phrases:
pure aeolic composite aeolic
3 aeolic phrases 4 0
2 aeolic phrases 29 13
1 aeolic phrase 29 59
An interesting tendency can be observed in the matter of word-end
when two aeolic phrases are directly juxtaposed. In some of the
contact points, by which is meant the point of junction between
two phrases, either word-end coincides with the contact point (cut)
or a word extends into the following phrase (bridge). Bridge is strictly
observed throughout all the repetitions in some kinds of verse, while
not in others.
First we count the number of contact points. Each of the four
verses made up of three aeolic phrases (collected above A. 3) has two
contact points. There is one contact point each in the 29 pure aeolic
verses made up of two phrases. The 13 composite aeolic verses made
up of two aeolic phrases are classiWed into two groups: those in which
the two aeolic phrases are juxtaposed (7 verses) and those in which
one or more freer D/e phrases are sandwiched between them
(6 verses). The former have one contact point while the latter have
none. And I add one contact point of heptaaeolicized e3 at P8s6,
which, however, is one of the freer D/e phrases, but its Wrst two
positions are identical with aeolic base.
In total there are 45 contact points of two aeolic phrases. They can
be classiWed according to the ending of the Wrst phrase and to the
beginning of the second phrase. Theoretically six diVerent types of
contact are possible:124

124
By deWnition,1 ending and other pendent cola do not appear in mid-verse.
Also, by deWnition, the full aeolic base in mid-verse is not of the shape w
.
82 Introduction
example total instances
(i) from zero ending to full base rdodgl 5
(ii) from zero ending to half-base wiltel 4
(iii) from zero ending to no base rdodrdod 4
(iv) from 2 ending to full base glwil 15
(v) from 2 ending to half-base gltel 11
(vi) from 2 ending to no base glrdod 6
Combinations of phrases are various.125 A particularly strange
feature of the aeolics of Pindar is, as has been pointed out, the
shortage of examples of the combinations of glgl or glph within
a verse. There is an organic relationship between these cola, as is clear
from the usage of Anacreon and the Attic dramatists.
There is a strong tendency to avoid coincidence of word-end and
phrase-end between aeolic phrases, especially where the second
phrase starts with full base:
(i) . . .
ww

x
ww
. . . or . . .
ww

x
w
ww
. . .
(iv) . . .
ww
w

x
ww
. . . or . . .
ww
w

x
w
ww
. . .
In these cases, bridge is strictly observed in most examples. In contact
points between 2 ending and full base (iv), cut is totally absent at
nine out of the 15 cases,126 and so it is at four out of the Wve cases
where zero ending meets full base (i).127 This tendency cannot be
accidental. For example, glyconic is followed by wilamowitzianum
(iv) in 3 verses (I8s4, 7r; N6s2, 6r; P6s1/2, 6r). The repetitions
amount in total to 19; bridge is observed at them all without excep-
tion. To give another example: at N4s3, word-end never coincides
with the contact points between heptasyllable and wilamowitzianum

125
(i) hepta ph (P8e2), hepta wil (N4s3), hepta e3 (aeolicized) (P8s6), rdod
gl (P2s2), rdod ph (N3e2); (ii) wil reiz (O9s6/7, N4s6), wil tel (N4s5), hepta
hepta3 (O9e8); (iii) rdod rdod (P5e3, I8s5a, I8s5b), wil rdod (I8s5a); (iv) gl
gl (P8e6, N2s4), gl ph (O1s1, P8e3/4, N2s4), gl gl3 (N7e5), gl wil (P6s1/2,
N6s2, I8s4), dod wil (P6s5),^ dod ph (O13s5), tel hipp (N7s8), tel gl3
(O9s2, I7s3/4), wil2 hipp (P2e8); (v) gl reiz (O9s3, O9s4, O9s5, N2s3, N4s4),
gl tel (P2s2, I7s5a, I7e5), dod reiz (O1e4), tel tel (N3e4), tel hepta (I7s5a);
(vi) gl rdod (P2e1a, P2e1b, P2e2, P2e3), dod ar (O1e7), tel adon (I7e2).
126 127
See List 5. 1. See List 5. 2.
 Composition of Verses 83
(i) in any of the 12 repetitions. Such a Wrm preference must be
deliberate.
Here, however, as so often in metre, there is no single rule. The
situation is more complicated. Besides the majority, there are a few
verses where cut/bridge is chosen freely. For example, there are three
instances of glyconicpherecratean contact (N2s4, 5r; P8e3/4, 5r;
O1s1, 8r). At the Wrst two (N2s4, P8e3/4), cut is totally avoided at all
the 10 repetitions. So far the situation is the same as the glyco-
nicwilamowitzianum contact above. But O1s1 is quite diVerent:
there is cut at half the eight repetitions. Four repetitions are too many
to be considered as mere exceptions to an assumed rule. We are
impelled to judge that in this matter there is no such rigid rule as, for
example, Porsons Law in the iambic trimeter. Rather, we should admit
that there is no rule which applies to all the verses of a given form, even
if a strong preference is observable in more than half the cases. How the
situation should be grasped, or described, will be discussed in a wider
scope taking bridge in general into consideration in Part III, B.
The situation is diVerent in the cases where half-base is concerned:
(ii) . . .
ww
x
ww
. . . or . . .
ww
x
w
ww
. . .
(v) . . .
ww
w
x
ww
. . . or . . .
ww
w
x
w
ww
. . .
In these cases bridge may be expected to occur, if it occurs at all, at
two positions:
(a) between two aeolic phrases, i.e. before the half-base;
(b) after the half-base if it is realized as long:
(a) . . .
ww
 x
ww
. . . or . . .
ww
w
 x
ww
. . .
(b) . . .
ww


ww
. . . or . . .
ww
w


ww
. . .
(b) is based on the supposition that the half-base should be collateral to
the link anceps in freer D/e phrases; see 6, D above. However, examin-
ation proves that there is no strong tendency towards bridge before the
half-base (a) in either case (ii) or (v) although there are some conspicu-
ous exceptions.128 For example, bridge occurs at 11 repetitions out of
12 in N4s6 (ii) and all eight repetitions in P2s2 (v). But at the same
time there are verses in which cut is dominant like O9s6/7 (bridge at

128
See List 5. 3, 4.
84 Introduction
1 repetition out of 8). No strong tendency is not observed in case (v)
after the long half-base (b) either. On the other hand bridge is
observed, or in other words, cut is totally avoided, in case (ii), though
there are only three examples.129 See further Part III, B.
Examples are scarce too, in cases (iii) and (vi).
(iii) . . .
ww

ww
. . . or . . .
ww

w
ww
. . .
(vi) . . .
ww
w

ww
. . . or . . .
ww
w

w
ww
. . .
No tendency is observable in these cases.130

2. Dovetailing131
Dovetailing is a particular way of avoiding word-end at the contact
point of two phrases, such that, when two cola are joined by bridge,
word-end is regularly located after the Wrst position of the second R
phrase.
R Outside Pindar,
R it is most noticeable in the combinations gl
gl, gl ph, chia chia.
The following are examples of dovetailing between two aeolic cola
(each found in all 6 repetitions of its respective ode):
R
P6s1/2 w
w
www
ww
w
R
/ w
w
ww
k x e gl wil
I7s3/4

w
w
ww
w

/

ww
w
w

k x e tel
gl3
As the line-numbering of these two shows, Boeckh and his followers
divided them into two, following the cuts. Withthis division, dovetailing,
of course, disappears. But compare the following verses (the number of
instances and the total of repetitions are added in parentheses):
R
N2s4 www
ww
w

/

ww
w

w
ww

k gl gl ph
R (4/5)
N4s3

w
ww

/ u
w
ww
k hepta wil
(11/12)
Obviously, Pindar is conscious of dovetailing in particular cases.
The metrical contexts of all the four verses are similar. Standard
aeolic phrases come before and after the contact point. These stanza-
forms will later be classiWed as Class I. In these verses, it is preferable
to recognize dovetailing, even when there is coincidence of word-end
129 130
See List 5. 3 (b). See List 5. 5, 6.
131
For this term see Maas, GM (1962), 59. Also cf. West, GM, Glossary-Index.
 Composition of Verses 85
elsewhere. There are also verses in which the second phrase starts
with half-base. For example:
R
P2s2 www
ww

u
ww
w

/
ww
w
k rdodgltel(7/8)
However, Pindar diVers from the tragic poets. The tragic poets
tend to employ dovetailing in almost every case where word-end does
not coincide with phrase-end. In other words, a word which does not
end with the phrase overlaps the contact point by just one position.
In contrast, Pindar is not so rigid. Even in cases where bridge is
observed in all the repetitions without exception, word-end occurs at
various positions. For example, in I8s4, one of the three verses which
include glwil, dovetailing occurs at three out of seven repetitions.
In the other four, a word overlaps by two positions:

w
ww
w

wj
w
ww
k
In N6s2 (glwil), dovetailing occurs at two repetitions out of six,
while at three repetitions word-end occurs after two positions as in
the scheme above, and at one repetition it occurs after three posi-
tions. On the other hand dovetailing is rigidly observed in all the
repetitions in P6s1/2 (glwil, above).

3. Verse-by-verse and phrase-by-phrase construction

Some songs in tragedy and comedy are constructed almost entirely
by repetition of the same phrase (a H). The phrase is
repeated at length, and it is not clear how these phrases should be
grouped into verses. For example, in his later plays Euripides often
introduces a song in aeolic metre in which the glyconic or the
wilamowitzianum is repeated again and again with little modiWca-
tion no less than ten times in a manner that is almost mesmeric. In
Pindar, verses can, of course, be deWned and it can be observed that
he hardly ever repeats the same verse. Within a verse, repetition of the
identical phrase is rare as is demonstrated above (A. 7). Even in O2,
the verses, however homogeneous, vary considerably in length.132
132
But this does not mean that there is no structural principle. The structure is
more delicate and Xexible than simple repetition. The seven verses of O2s are
articulated into two halves, s8 being taken as a coda. And there are other devices.
At the introduction of these verses, extra positions are located and the extra decrease
in number one by one, and increase in reversed. See further Part II, ad loc.
86 Introduction
Verse-by-verse construction is rare, and when it occurs, its context is
distinctively aeolic (Class I). An example is O9s35 (glreiz; similar
too are s6/79) or P2e1a3 (glrdod, with a Wnal e2). Even repetition
with modiWcation is rare. The following case is noteworthy:
N4s4
w
ww
w
w
ww

k gl reiz
N4s5
a
w
ww
a
ww
w
k wil tel
N4s6
a
a
ww
w
ww

k wil reiz
N4s (monostrophic) is distinctively aeolic. It is conceivable that
Pindar regarded the repetition as one of the characteristics of purely
aeolic metre.

4. Aeolic phrases and d

In composite aeolic verses, one or two aeolic phrases are preceded or
followed by freer D/e phrases like e or d. Rarely, the freer D/e phrase
is sandwiched between two aeolic phrases. These freer D/e phrases
are short. e is used extremely often; e2 and d are markedly rarer. In
this section and the following, the characteristics of these freer D/e
phrases are illustrated; Wrst d and ^ d, then e, e2, e3, and ^ e (in this
order). Interestingly, D is never used in the same verse as an aeolic
phrase; that is, D does not occur in composite aeolics.
There are in total 55 examples of d and three of ^ d in the eighteen
majors. The majority of them (42 d; 2 ^ d) are used in freer D/e verses,
and will be discussed later (section C in this chapter). The examples
in composite aeolic verses (13 d; 1 ^ d) tend to be used in the
following ways.
(a) In seven out of 13 examples of (i), d immediately precedes an
aeolic phrase:
P5e6
ww
rw
ww

w
j d rdod e
P8s5
ww

x
w
ww
w
k d wil2
P10e1 x
ww

x
ww
w
k x d gl
I7e7

ww

w
ww
k x d rdod
P6s4 ww

w
w
ww
k ^ d wil
N3e3 . . . w
ww

w
ww
. . . see (c) below
P10s4 . . . w
ww

w
ww
. . . see (d) below
 Composition of Verses 87
An anceps occurs before d in four verses, and in one acephalous d is
used instead of d, but nonetheless these seven resemble each other
closely in structure. The aeolic phrases after d start either with full
base or no base (is there any diVerence between full base and the
initial two positions of rdod? cf. 5, E). There is no example that starts
with half-base. Bridge is in general observed between d and the
aeolic: at 32 repetitions out of 39 (82.1%).133 See further Part III, B.
(b) In another four verses, d follows immediately upon an aeolic phrase:
P10e3 x
w
ww

ww

w
k hepta d e
I8s6
x
ww
w

ww
k gl d
P5e9 ta
ww
w

ww
w rw

w
k gl d x e e
N4s1 x
x
ww
w

ww
k hepta2 d
Here too the structural similarity is evident. In contrast with e and e2,
the examples of aeolic phrase followed by d at the verse-end are
surprisingly few (22 e, 12 e2, see the next sections). Bridge is more
frequent than cut between the aeolic phrase and d, but it does not
seem to be absolutely required.134 Note that cut occurs at four
repetitions out of 12 in N4s1.
(c) In two verses, d is preceded by short anceps and further, by an
aeolic verse with 2 ending (w
). One of these (N3e3) is listed in
(a) above:
N3e3
ww
w
w
ww

w
ww

wtw
k dod w d rdod e2
N3s1

ww
w
w
ww

w
k tel w d e
As a whole, the palindromic movement (
ww
w
w
ww
) is
striking. Both examples occur in N3.135 An aeolic phrase is never
followed by long ancepsd.

133 134
See List 5. 7. See List 5. 8.
135
The segment . . .
ww
w
w
ww
is also to be found in two other
verses, but I prefer a diVerent analysis:
N4s4
w
ww
w
w
ww

k gl reiz
O1e4 w

w

ww
w
w
ww

k ^ e e dod reiz
See on individual verses, Part II below. Among freer D/e verses, a similar sequence
(d x d, often d w d:
ww
w
ww
) are found in 4 verses (see above on
palindromes).
88 Introduction
(d) The other two examples are irregular:
P11s4
atww


wr
ww
k rdod
e d
Aeolic phrase (rdod) and d are separated by intervening long anceps
(which is highly irregular too) and e. For this extraordinary verse, see
Part II, ad loc.
P10s4 w

ww
a
ww

w
ww

w
w
k ^e dx
d rdod e2
This verse includes two ds. It may be better to suppose that a
reversed dodrans intrudes into a fundamentally freer D/e structure.

5. Aeolic phrases and e

As d may precede and follow an aeolic phrase, so may e, which in
fact is employed far more extensively. The following table includes
only e, x e and e x; longer e sequences (e2, e3, etc.) and ^ e are
excluded. Also excluded are the examples in composite aeolic
verse in which an aeolic phrase and e are separated by other inter-
vening phrases. Thus only examples in which e precedes or follows an
aeolic phrase without any intervention, except for anceps, are
counted.
preWx full base half-base no base total
e 1 0 4 5
xe 10 3 4 17
suYx 2 ending zero ending total
e 15 5 20
ex 2 0 2
The above table shows that the common combinations are:
x
w
aeolic phrase, and aeolic phrase
w
. If the cases are
limited to those at the beginning and end of verses, the tendency is
stronger:
initial full base half-base no base total
e precedes 0 0 0 0
x e precedes 9 3 2 14
 Composition of Verses 89
Wnal 2 ending zero ending total
e follows 15 5 20
e x follows 2 0 2
For further description, I give glyconics as an example. There are
49 glyconics in the eighteen majors. Of these 13 are immediately
preceded at the beginning and/or followed at the end of the verse by
e: x egle (3 examples), x egl (1), gle (7), gle x (2).
There are in total 10 verses of gle at the end of verse. This Wgure
is remarkably high. Glyconic followed by cretic may not be totally
unparalleled outside Pindar, but the examples are scarce. As for egl,
Sappho (86 L.P.) uses creticgl, but Pindar does not. Oddly
enough, the anceps of x e is indispensable at the beginning of the
verse in the Pindaric cases. Perhaps this may be by sheer chance, and I
should not be surprised if counterexamples are one day found in
papyri. However, it must not be denied that there are some tenden-
cies or preferences in the non-D/e odes. These verses resemble each
other and their relationship is evident.
Another tendency is clearly observable. In the matter of word-end,
bridge is strongly preferred to cut between 2 ending aeolic phrases
and following e (15 verses) and e x (2 verses).136 Even extreme cases
are found. At 3 verses cut is completely avoided at all the 8 repeti-
tions. The proportion of bridge is, in total for all the repetitions, 77/
89 ( 86.5%; 15 verses of e) or 86/102 ( 84.3%; including e x). After
zero ending aeolic phrases, the tendency is less evident, 19/25
( 76.0%), perhaps because of the shortage of examples (5 verses).137
As for the beginning of verse, bridge between x e and following aeolic
phrases with full base occurs 52 repetitions out of 64 ( 81.3%).138
However, there are a few verses which run counter to the general
tendency; for example, at N7e3 (gle), cut is found in 2 repetitions
out of 5. In these bridge/cut is freely chosen. The situation is similar
to that between two aeolic phrases (B. 1): although a strong prefer-
ence is dominant, no metrical rule is oVered which regulates all the
verses of the identical form. A similar tendency is observable between
d (D, D ) and e (e2, e3) (C. 6), and perhaps also between 2 es (C. 9).
Bridge/cut will be discussed in a wider perspective in Part III, B.

136 137 138

See List 5. 9. See List 5. 10. See List 5. 11.
90 Introduction
To turn to link anceps, there is a group of verses in which x e
follows an aeolic phrase with 2 ending (4 verses) or with zero
ending (2 verses). In these verses, with only one exception, the anceps
is short. It is as if the ending of the aeolic phrase were prolonged
further by w
w
:
N7s2
ww
wr w
w

ww
w

w
w
k dod w
e dod e2
I7s2 x
ww
w
w
w

w
k tel w e e
I7e1
w
ww
w
w
w
k gl w e
P5e2 w

ww
w
w
w

w
ww
j gl w e rdod
I8s1/2 w

w
ww
w
wy
w
w
ww
k wil w e wil
P11s4, whose irregularities are repeatedly mentioned above, is the
exception:

xtww


wr
ww
k rdod
e d
At all the repetitions in P5e2 (4r) and I7s2 (6r) word-end occurs
either before or after the short link anceps.

6. Aeolic phrases and e2

e2 is used basically in the same manner as e in composite aeolic
verses. It follows an aeolic phrase with 2 ending (9 verses) or zero-
ending (3 verses) at the end of the verse. In the former case, bridge is
commonly observed (54/61 88.5%).139 The exceptions are con-
centrated in N7s2 (see Part II, ad loc.); if this verse is excluded, the
percentage of the observance rises to 92.5% ( 49/53). In the latter
case, cut is rather common.140 It is interesting that of these 12 verses
only two, P8e1 and P2e3, appear in stanza-forms composed mainly
of aeolic verses (Class I), the rest in forms that make extensive use of
both aeolic metre and freer D/e (Class III). The seven examples of e2
following gl, tel, and hepta2 are found in verses made up only of the
aeolic phrase and e2. The other Wve, which follow dod or rdod, occur
in a longer verse (for three of them, N7s2, N3e3, P10s4, see A. 3). In

139 140
See List 5. 12. See List 5. 13.
 Composition of Verses 91
these, the similarity between aeolic hexasyllable (dodrans or reversed
dodrans) and e2 is fully exploited.
There is no example of e2 x following an aeolic colon at the end of
the verse. As for x e2, there are two examples:
P10s6 ww
ww
w
x
w
w
k tel x e2
N3s8 ww
ww
w
w
w
w

k tel w e2

Interestingly, the anceps in P10s6 is really anceps (long in 2
repetitions, short in 6; the transmitted text requires emendation;
see Part II, ad loc.).
e2 and x e2 never precede an aeolic phrase at the beginning of
the verse, except for one verse which starts with x e2 and ends with
rdod:
I8s10

w
w
rw

w
ww
k x e2 e rdod
In the following two, which are also counted amongst the longer
verses above, e2 in eVect prolongs the following aeolic phrase left-
wards:
I8s3
w
w
ww

w
wr w
ww
w
k wil e2 tel
P2s8

ww
w

w
w
w
ww
w

k tel e2 hag

7. Aeolic phrases and e3

Four examples of e3 out of 14 are combined with an aeolic phrase. In
N7s, there are two verses in which an aeolic colon is followed by e3:
N7s6

ww
w
rwrw
w
k tel e3
N7s3
ww
w

w
w
w
dod e3
There is also one example of e3 preceding an aeolic colon:
I8s7
w
wrw
w
ww
w

w
k e3 tel e
In fact, e3 is treated like e2. The curious phrase
x
w
w
in P8s6
(on which see 6. A) also follows an aeolic phrase:
P8s6 x
w
ww

x
w
w
k hepta e3(aeol)
Longer phrases than e3 are not used with aeolics.
92 Introduction

8. Aeolic phrases and ^ e

By deWnition, an acephalous phrase can only stand at the beginning
of a verse. The following phrase is always, with one possible excep-
tion, freer D/e.141 The possible exception is:
P10s5 w

w
w
ww
w

k ^e wil3
This exception can be removed by analysing: ^ eehagesichorean
( w dod
); see further Part II, ad loc. ^ ee can precede an aeolic
colon. There are three examples:
O1e7 w

w

ww
w

ww
w

k ^e e dod ar
O1e4 w

w

ww
w
w
ww

k ^e e dod reiz
P5e1 w

wr
w
w
ww
k ^ e e wil

C. Freer D/e Verses

In freer D/e verses, d (and D etc.) and e (and e2 etc.) are combined
with or without link anceps. There are various kinds of combination,
but some tendencies are observable. A verse shares some common
characteristics with other verses; in other words, it has parallels. In
this section and the following ones, examples of peculiar combin-
ations are collected.

1. Peculiarities of D and D
D (
ww
ww
) and D (
ww
ww
ww
) are prolonged ver-
sions of d, but diVer from d in two points. First, two D phrases
(including D, D , ^ D, and ^ D ) are never found within the same
verse. Here, there is a contrast with normal D/e, in which the
repetition of D within the same verse is quite common. Secondly,
D or D is never associated with any type of aeolic phrases. In other
words, composite aeolic verses do not have any D in them. Aeolic
and D are incompatible, indeed.

141
If w
in the aeolic base of some cola, for example, O1s1 (w
gl ph), were
analysed as ^ e, the number of the occurrence of ^ e aeolic colon would be
considerably increased.
 Composition of Verses 93
There are in total 19 examples of D, D , and their acephalous
forms. That means the number of verses which include D etc. is 19
(no verse includes two or more); they are concentrated in eight odes,
in particular seven cases in N6 and four in O10 (their stanza-forms
will be classiWed as Class II: freer D/e). On the other hand, there are
stanza-forms (O1s, P2s, P10s, I8s; all Class III) in which D or D
occurs only once. The occurrence is distinct from the surroundings.
Especially in O1s2 (D ) and P2s3 (D ), the contrast with the other
verses is striking; see Part II, ad locc.
Verses are in general short. Three examples out of four of ^ D and
two out of four of x D are independent verses by themselves:
O10e8 ww
ww
k ^D
O9e3, O13s1 ww
ww

k ^D

O9e6, N6e9

ww
ww
ww
k x D
D and D phrases may be preceded or followed by d (6 verses). Three
examples out of nine of D are preceded by d (or ^ d) and one out of
Wve of D is followed by d without link anceps:
N6s5 ww

ww
ww
j ^d D
N6e1
ww

ww
ww
rw
j dDe
N6e5

ww

ww
ww
j xdD
N6e3

ww
ww
ww

ww
j x D d
It is interesting that all these are found in N6. In two other verses, D is
followed by x d:
P10s3 w
ww
ww
x
ww

k xDxd
N6s6a
ww
ww


ww
rw

k D x d e

From this point of view, N6 is the most similar to D/e of all the non
D/e odes.
e precedes or follows D etc. in 10 verses, two of which also include
d. Two es are possible, but three are not. Spondee is used in one verse
(O10s3b), and so is e2. There is no verse in which e both precedes and
follows. This is characteristically diVerent from the normal D/e
verses; for example, e x D x e is very common in them.
D preceded by e
I8s9 w

ww
ww
k ^e D
O10e2 w

ww
ww
k ^e D
94 Introduction
O1s2 wr
w

ww
ww
ww

k ^ e e D

O10s3b rw


ww
ww
k e sp D
P2s6

wrw
w
ww
ww
k x e2 x D
D followed by e
P2s3

ww
ww
ww

w
k x D e
N6s4a ww
ww
ww
rw

j ^D e

O10s1 ww
ww

w
tw
j ^ D e e
N6e1
ww

ww
ww
rw
j dDe
N6s6a
ww
ww


ww
rw

k Dxde

Between D etc. and e at the end of verse, bridge is the general
tendency. This will be discussed with the examples of de. See
further below, 8. C. 8.

2. Verses containing two d

As is stated above, D is not repeated within a verse, but D and d may be
included in the same verse. As for double d, there are two verses which
contain two ds without intervening anceps and four with anceps:
O10e4/5

ww

ww
rw

j
dde

P8e5

ww

ww
w
w
j
ddwe
O1e5 ww
w
ww

w
w
k ^ d w d e2
N3e1a
ww
w
ww
x
w

j dwdxe

N2s5
ww
x
ww

k dxd

P10s4 w

ww
x
ww

w
ww

w
w
k ^e d x d
rdod e2
The last (P10s4) includes rdod and is not a freer D/e verse but
composite aeolic.
Another two verses have one or more e phrases between two ds:
O9s11

ww

w
x
ww

k x d e x d

O10e10 w
wwt

wr
w

ww
k xdxeed

3. Short anceps and d (and a special group of aeolic cola)

There is a group of verses containing w d, which is peculiar to
Pindaric metre. 9 verses are made up in the same manner. In the
following chart they are arranged according to the length of the
phrase preceding w d:
 Composition of Verses 95
O1e3
w
w
w

ww

w
k e2 w d e
P11s3 rw
wr w

ww
k e2 w d
O2s8

w
rw
w
w
ww
k x e e2 w d
O9e1/2 w
w
w
w

ww
w
w
w

k x e2 w d
x e2

P2s5

wrw
w
ww
k x e2 w d
O1s7
w
w
w
w
ww

w
k e3 w d e
N3s5

w
w
wr w
ww
k x e3 w d
N3e1b w rw
w
wr w
ww
w
w

w
k x e3 w d
xee
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
Every verse starts with a single-short sequence. The short anceps of w
d is arranged in such a way that it carries on the preceding single-
short movement. In other words, the short anceps itself appears to be
a part of the movement. The double short contained in d seems as if
it were a temporal substitute for a single short. In two verses (O9e1/2,
N3e1b) the single short movement continues beyond the double
short.
At Wrst sight, an analysis which separates w d from the preceding
sequence may seem arbitrary. But it is necessary to divide the se-
quence; otherwise we are obliged to recognise an aeolic colon of
gigantic size (cf. I. 3, Rule 8). On the other hand, similarity with
the following four examples of w d and with another (P2s6), in which
w D is used instead of w d, suggests that this colometry is the least
inappropriate if we wish to be consistent:142
O1e5 ww
w
ww

w
w
k ^ d w d e2
N3e1a
ww
w
ww
x
w

j dwd
xe

N3e3
ww
w
w
ww

w
ww

wtw
k dod w d
rdod e2
N3s1

ww
w
w
ww

w
k tel w d e
P2s6

wrw
w
ww
ww
k x e2 w d
Moreover, there are another three composite aeolic verses which are
constructed in the same fashion as single-short movementw d. In

142
N3e3 and N3s1 have already cited as examples of aeolic cola followed by w d,
and O1e5 and N3e1a, as examples of 2 ds within a verse.
96 Introduction
these, instead of w d, an aeolic colon starting with short anceps, such
as a telesillean of the form w
ww
w
, follows single-short move-
ment:
I8s7
w
wrw
w
ww
w

w
k e3 tel e
I8s3
w
w
ww

w
wr w
ww
w
k wil e2 tel
P2s8

ww
w

w
w
w
ww
w

k tel e2 hag
More important than analysis is the general structure. Long anceps is
never used in verses of these types, except at the beginning of the
verse. In all the repetitions, the anceps preceding d is short, and so are
the ancipitia assumed to be present in e3 or e5. So short anceps
cannot be distinguished from true short. As a result, the verses
have a distinctively light movement, unlike normal D/e. This impres-
sion is enhanced by the length of the verse. Some of the verses are the
longer verses, collected above (A. 3). In general, they are very long,
longer than other non-D/e verses. The metrical contexts of all the
verses are interesting too. They are localized in particular stanza-
forms: O1, P2, N3, I8, which will be classiWed in Class III.

4. Verses with no double shorts

There are in total 45 verses which lack any double short, whether it be
the choriambic nucleus of an aeolic colon, a choriamb (d), or a
hemiepes (D). Twelve examples out of 45 occur in O2, but also a
further 33 verses in the other odes.143 It would evidently be forced to
classify all the non-D/e odes other than O2 into one group and to
interpret them as aeolic.
In these verses e is the commonest phrase. It is mixed with ^ e, or
e2, or the longer phrases. Link anceps is occasionally present: at the
beginning of a verse 23 times; at the middle eight times; at the end 11
times.144 Nevertheless, its rarity is striking. In contrast to the normal

143
O1s3, O1s5, O1s8, O1s9, O1s10, O1s11, O10s3a, O10s4, O10s5, O10e1b,
O10e7, O13s3, O13s4, P2s1, P5s1, P5s4, P5s6, P5s7a, P5s9, P5s11, P5e7a, P5e7b,
P5e8, P6s9, P8s7, P11e5, N3s2, N6e6/7, N6s1a, N6s4b, N6s6b, I7s5b, I8s8.
144
The number of mid-verse examples (11 examples) would increase if we
included as anceps the central position of e3 (6 examples) and some positions of e5
(2 examples) and e6 (1 example).
 Composition of Verses 97
D/e, e (and e2, e3, . . . ) and e are in general juxtaposed without link. It
is especially remarkable that long anceps between two es is avoided
to a considerable extent. There are only six mid-ancipitia in total
which are exclusively long.145 At another three ancipitia, long and
short are in responsion.146 As for word-end, cut is fairly strictly
avoided after the anceps (6 D).

5. Double, triple, multiple e

Double e and triple e are very common:
P5e8
w

w
k ee
P5s9 w
w

w

w
k x e e e
There are 20 examples of double e in O2, but also a further 18
examples in other odes.147 Except for the verses of O2, eee is
found in P5s9 and I8s8, and ee in O1s9, O1e1, O10s1, O10s3a,
O10s5, O10e3, O10e10, P2s7, P5s1, P5s10, P5e8, P5e9, N3e1b, I7s2.
The examples tend to be concentrated in certain odes; six in P5, Wve
examples in O10. Except for I7s, all the stanza-forms belong to Class
II (freer D/e) or Class III (amalgamated).
Resolution is frequent. The examples in which either or both of
two es placed side by side contains resolution amount to 28 out of
38. Either long of e may be resolved. For Pindars tendencies in
respect of resolution, see 6. C.
As has been demonstrated above (B. 5), at the end of verse e is
commoner than e x while at the beginning of verse x e is commoner
than e, when e is preceded or followed by an aeolic phrase. This
tendency is observed in the cases of double e too, and is to be
considered as a manifestation of link anceps being more frequent at
the beginning of verse than at the end (6. D). At the beginning of a
verse, x ee occurs in 12 verses, and x ee2 in two verses, while ee

145
(e
e) O2s3, O10s3a, O10s4, (e2
e) O13s3, P11e5, (^ e
e) P5s11. As has
already been remarked in 6. D (link anceps), the long is resolved either before or after
the long anceps so that three successive long syllables (
w


w
) tend to be
avoided.
146
(e2
e) O13s4, P6s9, (e
e2) P8s7.
147
In a verse made up three or more successive e, each e e is counted separately;
thus P5s9 above is counted twice.
98 Introduction
occurs only in three verses, and ee2 in one verse. At the end of a
verse, ee occurs in 17 verses, ^ ee in three verses, and e5 e in one
verse while ee x occurs only in four verses and e2 e x in one verse.
As for word-end between two es, bridge is fairly common. Cut is
generally avoided, both at the end of a verse (bridge is observed at
127 repetitions out of 156 in 21 verses; 81.4%)148 and at the begin-
ning (bridge is observed at 90 repetitions out of 110 in 14 verses;
81.8%).149 In other words, a single word of the shape
w
is rare at
verse-end, and so is a single word, or word group, of the shape
x
w
at the beginning.

6. e2 and other prolonged e

A prolonged single-short movement followed by w d in the group of
verses above is the most distinctive use of the prolonged e (e3, e5), cf.
section 3 above. But there are other examples. e5 and e6 make up a
verse either by themselves or with e:
O1s8 w rwrw
w
w
w
k x e5
N3s2
w
w
wrw
w
rw
k e5 e
P2s1 rwrw
w
w
wrw
k e6
In all three verses, frequent resolution is prominent: two out of six longs
are resolved at O1s8 and three out of seven at P2s1. Moreover, there are
seven short syllables in succession in O1s8 and six in P2s1 (for
continuous shorts, see Part III, D). Resolution always occurs through-
out all the repetitions. Certainly the short syllables have a striking eVect.
These three verses too are used in the Class III stanza-forms, like those
ending with w d in the section 3. There are no verses of the form e5 x
or e6 x. A preference for blunt ending over pendent is evident here
too.
There are 14 examples of e3. Some of these are listed in the section 3:
three followed by w d (O1s7, N3s5, N3e1b) and one followed tele-
sillean with short half-base (I8s7). Another four examples (including

148 149
See List 5. 14. See List 5. 15.
 Composition of Verses 99
I8s7) either precede or follow an aeolic phrase. They are listed at B. 8
above. The others are:
O2e2
w
w
wr
w

w

k e3 e e

O1s3, s5
w
w
w
k e3
O1e2 w

ww

wrw
w
j ^ e d e3
P5s11 w


w
w
w
k ^ e x e3
N6e6/7 t
w
w
w

j x e3

There are 38 examples of e2 in 37 verses. Of these examples, 17 are in
composite aeolic verses. The other 21 examples are in freer D/e verse
with e or other phrases; e2 is occasionally substituted for e, for
example:
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
O2s6/7 w

wr
wt
w
w

w

k ^ e e e e2 e

P2e3 w

ww
w
www
ww

w
w
j gl rdod e2
In O10e7 and I7s5b, x e2 (

w
w
) is a single verse by itself. As
in these, anceps usually stands before e2 at the beginning of the verse
(10 examples, as against 2 without anceps; cf. frequent x e/rare e, B.
5). If e2 were not recognized as a phrase, it would be diYcult to
analyse these simple verses:
O1s11 x
wr
w
w
k x e e2
P8s7 x
w
x
w
w
k x e x e2
O13s3 w
wrw


w
k x e2 x e

7. dshort ancepse
d w e is usually analysed as chia and is treated as a single colon, e.g.
in tragedy. But the metrical context in Pindar is against this treat-
ment. d w e is much less frequent in Pindar, and is included only in 4
verses.
P5e9 ta
ww
w

ww
w rw

w
k gl d w e e
N3e1b wrw
w
wr w
ww
w
w

w
k we3w d w e e
N6s7
wy
ww
w
w
k edwe
P8e5

ww

ww
w
w
j xddwe
No verses have long anceps before e, and verses in which d and e
are connected by explicit anceps are very rare. There is only one
example:
100 Introduction
N3e1a
ww
w
ww
x
w

j dwdxe

This verse (N3e1a) is irregular in that an anceps (link) is situated
after e at the end of the verse.
There is one verse which has d w e2
:
O9e1/2 w
w
w
w
ww
w
w
w

k x e2 w d w e2

This verse (O9e1/2) is included in the group discussed above (3),
short anceps and d, but is unique in that it has long single-short
movement both before and after the central
ww
. There is no
verse that ends with d w e
.
In the chart of aeolic phrases (5. A. 5) I mentioned a theoretical
possibility that d w e was the dodrans prolonged by w
(dod2) and
w d w e was telesillean prolonged by w
(tel2). But the metrical
contexts of the examples above refute this interpretation.

8. de without intervening anceps

In contrast with d w e, there are as many as 16 verses in which d is
directly followed by e (and e2, e3) without intervening anceps
(
ww

w
). The related phrase without link anceps, De, is
fairly common in the normal D/e odes. In the eighteen majors, the
phrase de often ends the verse. Pindar frequently uses e after an
aeolic colon for that purpose (see B. 5). Apparently the same eVect is
sought here too. There are nine verses which have de at the end of
verse, and 1, dee :
P2e6 w

ww

w
j ^e de
P5e5 w

ww
www
j ^e de
P10s2b w

ww

w
j ^e d e
N7e1 w
w

ww
rw
k wede
P10e3 x
w
ww

ww

w
k hepta d e
O1e3
w
w
w
ww

w
k e2 w d e
O1s7
w
w
w
w
ww

w
k e3 w d e
N3s1

ww
w
w
ww

w
k tel w d e
P5s10
w

ww
rw

w
k edee
Note that the Wrst long of e is resolved so as to produce a character-
istic sequence (
ww
www
) in three examples (P5e5, N7e1,
 Composition of Verses 101
P5s10). This sequence is included in De (N6e1) and in ^ D e x
(N6s4a); cf. 1 above.
The examples of de x are fewer. This is the same as the rarity of
e x after aeolic phrases:
O10e4/5

ww

ww
rw

j x d d e

N6s6a
ww
ww


ww
rw

k D x d e

Again the Wrst long of e is resolved in these two.
de2, and de3 are used in:
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
(cf. 3
above)
O1e5 ww
w
ww

w
w
k ^ d w d e2
N7s5 w
ww

w
wrw
k w d e3
O1e2 w

ww

wrw
w
j ^ e d e3
Bridge is the general tendency between d (D, D ; cf. 1 above)
and e at the end of verse, though there are conspicuous exceptions.
The examples of d (D, D ) and e amount to 10. The majority of them
strictly observe the bridge.150 Notably, cut is totally absent from all
the repetitions in as many as 6 examples. However, the situation is
complicated in the same manner as that between two aeolics (see B,1
above). In three examples (O1s7, N3s1, P10s2b) cut occurs in half or
more of the repetitions. Obviously there is no single rule that should
regulate all the examples. See further Part III, B.

9. ed without intervening anceps

In contrast with de, the phrase in reversed order, ed (
w


ww
), is less common (7 examples):
P5s5
wy
ww
k ed
N6s7
wy
ww
w
w
k edxe
P5s10
w

ww
rw

w
k edee
O10e10 w
wwt

wr
w

ww
j xdxe
ed

150
See List 5. 16.
102 Introduction
O10e1a w
wr
ww
j xed
N7e1 w
w

ww
rw
k xede
P11s4
atww


wr
ww
k rdod x
ed
Mirroring the sequence
ww
rw
, the last long of e is resolved
rather frequently (4 examples) so as to produce
www
ww
. Note
that x ed, which is a common phrase in drama (iach), is rare.
 The Three Classes of Stanza-Forms 103

9. THE THREE CLASSES OF STANZA-FORMS

We have discussed each phrase and combination of phrases up to

this chapter. Now it is time to discuss the metre of the stanza-
form as a whole. This is a diVerent topic from the metre of each
verse contained in a stanza. To comprehend a single stanza-form,
it would be suYcient to analyse its verses individually and de-
scribe, or label, the cola or phrases. However, the combination of
various phrases gives each stanza-form a kind of unity as well as
its particular character. Strictly speaking, every single stanza-form
is individual. Nevertheless, similarities can be distinguished be-
tween some stanza-forms. ClassiWcation is both required and pos-
sible. Moreover, parallels between diVerent stanza-forms are
required, in order to establish an appropriate colometry for am-
biguous verses.
Although there are some borderline cases, the 31 stanza-forms of
the eighteen majors can be classiWed stylistically into three groups,
according to the way in which two metres, i.e. aeolic and freer D/e,
are mixed. Stanza-forms which are made up mainly of aeolic phrases
belong to the Wrst group (Class I). The second group consists chieXy
of freer D/e (Class II). In the third group, the two metres are tightly
amalgamated (Class III).
ClassiWcation does not depend exclusively on the distribution of
the two metres. That is only one of the criteria. As has occasionally
been pointed out in the preceding chapters, there are other charac-
teristics peculiar to each group. For example, the aeolic cola of Class
I are diVerent from those of Class III in that they feature two longs at
the aeolic base. The criteria listed below are useful, and classiWcation
is acceptable when all the criteria are met.

Proportion of freer D/e to aeolic

By deWnition, the stanza-forms of aeolic style (Class I) contain none
or few verses made up only of freer D/e phrases. Indeed, all the
stanza-forms that I classify in Class I contain no freer D/e verse or,
at most, one. But the opposite is not true; a stanza-form without a
freer D/e verse or with just one does not automatically belong to
104 Introduction
Class I. Taking other criteria into consideration, I classify three
stanza-forms (N7s, N7e, P11s) as Class III, although they contain
only one freer D/e verse.
In contrast to Class I, the stanza-forms of freer D/e style (Class II)
cannot be judged by the number of occurrences of freer D/e verses,
for aeolic verses are ubiquitous. The proportion of freer D/e verses is
more signiWcant. More than half the verses are freer D/e in all the
stanza-forms that I classify as Class II. The opposite is true in this
case but one; O1s belongs to Class III though nine verses out of 11 are
freer D/e.

Ratio of short syllables (RSS)

Cf. 8. A. 5. The method by which RSS is calculated has been
explained above. The frequency of short syllables is a useful index.
In general, RSS is high in Class III, and quite high in some of Class II,
but low in Class I.

Length of verses
Cf. 8. A. 4. The length of each verse is, of course, not consistent
within a stanza. Admittedly, there is a great diVerence between three
positions and 22, but the average length of all the verses of a stanza is
nonetheless useful. It is high in most Class III stanzas, and in some
Class I stanzas. In contrast, it is low in Class II.
Fifteen verses have been deWned as longer verses. These are clas-
siWed into two groups: distinctively aeolic verses (7 verses) and
amalgamated verses (8 verses). The former tend to be used in Class
I, and the latter, in Class III. There are also 15 shorter verses. Of
these, two are aeolic, but the remaining 13 are made up of d, D, or
eanceps. It is, therefore, not surprising that verses in Class II are, on
average, relatively short.

Verse-by-verse construction
Cf. 8. A. 2. In Class I, there are some cases in which the identical
colon is repeated for 3 or more verses in cluster.
 The Three Classes of Stanza-Forms 105

Repetition of identical phrase within verse

Cf. 8. B. 3. In 10 verses, aeolic phrases, dodrans or reversed dodrans
are repeated within a verse with/without base. Most of them belong
to Class III.

Dovetailing
Cf. 8. B. 2. Dovetailing is a special method of linking two (or more)
aeolic cola by single-syllable overlap. It is peculiar to Class I.

Aeolic base of two longs

Cf. 5. B. 1. Aeolic base is a useful criterion. Full aeolic base is rarely
occupied by

(6 examples in total). Even when

corres-
ponds with
w (i.e.
x, 20 examples),

is exceptional in about
half the verses concerned; the exception often results from a proper
noun. In the majority of stanzas,

is avoided. This is the general
picture; but there are some exceptions in which

is freely used,
most of them belonging to Class I.

Reversed dodrans with long second position

Cf. 5. E. 1. In Pindar reversed dodrans is
w
ww
, but occasion-
ally
x
ww
occurs. The latter form is found, with few exceptions,
only in Class I.

Reversed dodrans with tribrach opening

Cf. 5. E. 2. There are 10 reversed dodrantes of the form (www
ww
).
Most of them occur in Class II or Class III.

Aeolic phrases and e2

Cf. 8. B. 6. In contrast to e, e2 is hardly ever used after an aeolic colon
in Class I. The similarity of e2 with aeolic hexasyllable (dodrans
and reversed dodrans) is fully exploited in some Class III stanza-
forms.
106 Introduction

Dactylic runs
Cf. 7. 6, 8. C. 1. D and D are rather rare phrases. Within the verse,
they are incompatible with aeolics. Nevertheless, the six verses which
look like expanded aeolics, but which are better analysed as special
D/e, are located next to verses containing D or D . The aYliation is
certain. Attention should be paid to the fact that an asymmetrical
phrase is occasionally mixed in with the genuine D/e. To sum up,
dactylic runs, whether contained in D or D or in expanded aeolics,
are absent from Class I.

Long link anceps

Cf. 6. D. In freer D/e, link anceps is not so frequent as in normal D/e.
Two longs tend to be juxtaposed without anceps in between. How-
ever, there are some verses which contain anceps in mid-verse. When
the link anceps is long in a verse, the stanza-form belongs to Class II.

Short link anceps and choriamb after long sequences

of single-short movement
Cf. 8. C. 3. 6. e5 and e6 are signals of Class III. e3 is used more
generally, but is absent from Class I. There is a group of verses
which are made up of single-short movement followed by w d.
These are peculiar to Class III.

Double, triple, multiple e

Cf. 8. B. 5. Successive e without anceps in between is a common
phrase, but is hardly ever used in conjunction with aeolic phrases, i.e.
it is absent in Class I.

Acephalous phrases
Cf. 6. B. Though there are a few exception, ^ d, ^ D, ^ D , and ^ e are
concentrated in some stanza-forms in Class II.
These criteria are arranged in the following synopsis:
 The Three Classes of Stanza-Forms 107
Class I Class II Class III
p
#1 no freer D/e verse or one at most
p
#2 fewer aeolic verses
p
#3 higher RSS
p
#4 lower RSS
p p
#5 longer verses
p
#6 shorter verses
p
#7 verse-by-verse construction
p
#8 construction in repetitive aeolic phrases
p
#9 dovetailing
p
#10 aeolic base of two longs
p
#11 rdod with long in second position
p p
#12 rdod with tribrach opening
p
#13 aeolic phrasee2
p
#14 parataxis of hexasyllabic aeolics and e2
p p
#15 dactylic run
p p
#16 dactylic movement with asymmetry
p
#17 long link anceps in mid-verse
p
#18 short link anceps and d
p p
#19 e3, e5, e6
p p
#20 multiple e
p
#21 acephalous phrases
In the following sections, a sketch of the ideal type of each Class is
oVered Wrst; a short historical survey then follows.
Class I: Aeolic
Examples: O9s, P2e, P6s, P8s, P8e, P10e, N2s, N4s, I7s, I7e.
Most certain cases: O9s, N2s, N4s, I7s, I7e.
Ambiguous cases: P2e (? Class III), P10e (? Class III)
Most of the verses, if not all, are made up of aeolic phrases with/
without preWx or suYx. They are made up of one or two, or,
exceptionally three phrases which make the verse very long (#5).
Freer D/e phrases are used as preWx and suYx, but verses made up
wholly of freer D/e are rare (#1). Aeolic phrases of the identical or
very similar type are repeated within a verse (#8), and two at a time
are sporadically connected by dovetailing (#9). The identical aeolic
phrase is also repeated for several verses in the same structure, and
verse-by-verse construction is clear (#7). Full aeolic base () may
take any of the forms:

,
w, w
, www; two longs (

) and
108 Introduction
long and explicit anceps (
x, i.e. long and short are in correspond-
ence in the second position) are peculiar to this class (#10). And the
second position of reversed dodrans is long occasionally at some
repetitions (#11). Partly because of these long ancipitia, RSS tends to
be low (#4).
It is reasonable to suppose that Pindars Class I belongs basically to
the aeolic tradition. The aeolic phrases of Pindar in Class I are the
same in detail as those of Sappho and Alcaeus on the one hand and
tragedy on the other, although the diVerences cannot be ignored. It is
not clear how far the diVerences are to be attributed to the genre and
how far to Pindars own artistic preferences. The judgement must
follow a fresh examination of other poets (Bacchylides, the three
tragic poets).
Class II: Freer D/e or choriambo-cretic
Examples: O2s, O2e, O9e, O10s, O10e, O13s15, P5s, N6s, N6e.
Most certain cases: O9e, O10s, O10e, O13s15, P5s, N6s, N6e.
More than half the verses are composed of the phrases e, d, and their
prolonged forms. Aeolic phrases are mixed in but are uncommon
(#1). Verses tend in general to be short (#6). Occasionally, there are
very short verses, made up of fewer than 6 positions. Freer D/e
phrases are combined with each other with or without link anceps.
Link anceps, especially long, in mid-verse is rare outside normal D/e,
but is occasionally found here in Class II (#17). The use of dactylic
run (D, D , and also ones with asymmetrical opening or cadence) is
a distinct feature, though some stanzas of Class III have them too
(#15, #16). Phrase e is sometimes repeated two or three times with-
out an intervening link (#19). It may be heavily resolved. At the
beginning of the verse acephalous phrases, mainly ^ e, but also others,
are found (#21). RSS varies according to stanza-forms, and is not a
valid criterion for the identiWcation of this class. O2 is peculiar in
that it contains no double short except s8, nor any aeolic phrase.
However, verses made up exclusively of e and e2 are not restricted to
O2 but are found in other stanza-forms.
Stesichorus is reasonably assumed to be the forerunner of the
normal D/e of Pindar. Freer D/e is another development from this
prototype, and the inXuence of Ibycus is noticeable in dactylic runs
with asymmetrical cadence/opening. In view of uncertainties of
 The Three Classes of Stanza-Forms 109
dating, it cannot be decided which of the two, the normal D/e or freer
D/e, is the earlier development in Pindars career. In normal D/e,
Pindar seems to have assigned himself rigid rules. In the other area he
chose to resort to rarer phrases, for example, d or the sequence of two
(or more) es without link anceps. Moreover, he introduced aeolic
phrases in this looser D/e environment (whether or not these aeolic
phrases are to be identiWed with those in Class I; see below on
Class III).
Class III: amalgamated style
Examples: O1s, O1e, P2s, P5e, P10s, P11s, P11e, N3s, N3e, N7s,
N7e, I8s.
Most certain cases: O1s, O1e, P2s, N3s, N3e, N7s, N7e.
Aeolic metre and freer D/e are neatly amalgamated; in other words,
verses can be divided into phrases as in freer D/e, but at the same
time a number of asymmetrical phrases of aeolic type are incorpor-
ated (#1, #2). As is demonstrated in the synopsis above, there are
many criteria common to Class II and Class III. The diVerence
between these two classes is, however, fundamental: one is the length
of verse (#5, #6), another is whether aeolic verses are frequent or not
(#2), and the third is absence of long mid-anceps (#17). RSS is very
high (#3), especially in the stanza-forms which contain long se-
quences of single-short movement with many resolutions (#19).
The single-short movement is often followed by w d (#18). Verses
of this type are in general very long and so are verses of another
group, in which paratactic use of aeolic hexasyllable with e2 is a
distinct feature (#14). Some stanza-forms are magniWcent both in
scale and in their complicated style. Class III is the most character-
istically Pindaric and the consummation of his art. It cannot be by
accident that not only the elaborate, grand-scale odes like O1 or P2
but also those most problematic in their content belong to this class:
I am thinking of N7 or I8.151
At the Wnal stage of this Part, I must ask the reader to be tolerant if
I raise again a residual question (cf. I. 7). I cannot help suspecting

151
Oustside the eighteen majors two beautiful poems belong to Class III: O14 and
Paean 4. The metres of these two are very sophisticated, comparable with O1 or P2,
or perhaps more. See each section of the Appendices to Part II.
110 Introduction
that the dodrans and the reversed dodrans, as used in Class III (and
in Class II, but not in Class I), are not really aeolic. Rather, Pindar
may have introduced these phrases (
ww
w
and
w
ww
) in
a much freer context and employed them like the symmetrical
components d or e2. They are even followed by anceps (for example,

w
ww

rdodanceps) or preceded by ^ e (for example,
w

ww
w
^ edod). He may have gone further, and
conXated these two asymmetrical phrases so as to invent a new
phrase
w
ww
w
, which happens to be equal to one possible
form of the glyconic, but is not in fact glyconic. The Simonidean
phrase which is repeated in 542 P (cited above in 5. G) is most
probably this non-glyconic type. If this hypothesis is right and Pindar
(and/or Simonides?) invented a diVerent metre, what seems to be
glyconic or pherecratean in Classes II and III is to be distinguished
from the aeolic glyconic or pherecratean in Class I, which starts with
the aeolic base. However, it is dangerous to push this argument too
hard.
Starting from the rules in Ch. 3, we have now established two
metres and three styles in the other half of Pindar. It remains to
assign individual odes to them.
 APPENDIX

Modern Emendations: A Test Case

(Nem. 6 Str. 67)

When there is a break in responsion between corresponding strophes, the

text is possibly corrupt. The exceptions are anceps and resolution, but
recognition of anceps or resolution itself occasionally raises a question.
Numerous emendations have been proposed to recover exact responsion
since the Byzantine period (notably by Moschopoulos and Triclinius), but
the nineteenth century was the heyday of textual criticism, especially in
Germany. In this chapter I cursorily trace the history of emendationes metri
causa. N6s67 is chosen as a test case to illustrate the inclinations of
prominent critics, from Hermann and Boeckh to Turyn and Snell. Other
passages will be each discussed in Part II.
The end of the strophe (s67) of Nemean 6 is heavily corrupt.1 Of its six
repetitions, str. 3 is the most contentious: proposed conjectures cover almost
a whole pages worth of Gerbers Emendations (11213). But the other Wve
repetitions are not exempt from suspicion either. They do not metrically
correspond with each other although the meaning is, in general, unambigu-
ous. Like all the other Pindaric passages where textual matters are entangled
with metrical ones, an attempt at emendation often involves a full-scale
rewriting of more than one corresponding passage. The emendations hith-
erto proposed are so huge in number that, before we examine them each by
each, some preliminary guidance would be helpful for grasping the scope of
the question.
First, it is convenient to arrange the six corresponding lines horizontally.
Instead of the manuscript readings I start from Snells text for practical reasons.
Not only is Snells text easily available, but it also prints N6s67 with the fewest
emendations among modern editions. In fact, Snell is certainly wrong in
assuming free responsion at the beginning of s7 (his 8th verse), but, thanks
1
One must remember that the Nemeans (and the Isthmians too) are not well
preserved in the MSS. The tradition is thin: Nemean 6 depends on only two, apart
from twelve verses in V; of these B is superior and has a scholarly text while D is
careless and unreliable. In contrast, the Olympians are preserved in a number of MSS,
including. A, which is the sole manuscript of the recensio Ambrosiana. The situation
of the Pythians is not so good, but far better than for the Nemeans and the Isthmians.
112 Introduction
to this alleged free responsion, his text generally keeps the manuscript readings
as they are (except str. 3, which must be rewritten in any case).
s6a
ww
ww


ww
rw

k
str. 1 (6) 
 K P N Pb a 
ant. 1 (13)
E K , n  
  r 
str. 2 (28) u I
e   h K
d F, , E,
ant. 2 (35) E 
 d H
 I
e 
str. 3 (50)  P fg I
 
 f  
ant. 3 (57) :   Kg  
  
s6b
wt
k
str. 1 (6b)
w




ant. 1 (13b)
w


F fg

str. 2 (28b)
wr
(K
 scanned trisyllabically)
s K

ant. 2 (35b)
w


x 


str. 3 (50b)
wr

E 
ant. 3 (57b)
wr

 ,
s7
at
ww
w
w
k
str. 1 (7)   E
d 
.
ant. 1 (14) P  Id

 ,
str. 2 (29) P
 a I,
ant. 2 (36) 

b  
str. 3 (51) 
 d d I 
,
ant. 3 (58)

 K
d Y F 

At the beginning of s7 (his 8th verse) Snell admits responsion between these
forms:

w

ww
. . . (str. 1; the Wrst syllable of  is scanned as long)

www
ww
. . . (str. 2, ant. 3; also str. 3, emendatio metri causa)


ww
ww
. . . (ant. 1, ant. 2).
The Wrst and the second are unproblematic. However, the responsion between
these two and the third is highly improbable. Snell postulates acephalous
choriambic dimeter, whose three positions are ancipitia (
ww
),2 but
the existence of three or more successive ancipitia must be denied in Greek
metre.3 Either
www or

ww (or, theoretically, neither) should be
retained, and the other must be emended. As I describe below, the majority

2
Snell, Griechische Metrik4 37. As is illustrated below, Snells text is identical with
Schroeders (BT2).
3
Itsumi, Choriambic Dimeter.
 Appendix: Modern Emendations 113
of critics before 1900 took

ww as the norm. This means as many as four
repetitions must be emended, namely, str. 1, str. 2, str. 3, and ant. 3.
Conversely, the idea that
www was the norm was Wrst proposed by
Hermann in his Notae of 1817 and, after a long interval, was revived by
Maas,4 Schroeder (ed. maior2, BT3), and Turyn. If the responsion between

www and
w
is allowed, the number of responsions that must be
emended is fewer: ant. 1, ant. 2, and str. 3. I discuss this Wrst.
Turyn, who considers
www as the norm, adopts two conjectures by
Hermann: P  at ant. 1 and 

 at ant. 2 (Hermann
himself emends the text more extensively; moreover, his ideas greatly
changed at least three times during his life; see below). I leave the emend-
ation of str. 3 aside for a moment. P  is a minor change but


 for 
 is not, and seems less plausible. Pindar
seems to have understood 
 , a Homeric epithet for Artemis
(see M. W. Edwards on Il. 20. 701), as with golden distaV (for the sense,
see R. Janko on Il. 16. 183). It is a favourite of Pindar (O6. 104 epithet
for Amphitrite; N5. 35 epithet for Nereids; fr. 29 epithet for Melia; and
Threnos 3. 1( fr. 128c) epithet for Leto, as in our passage). 

is not used anywhere by Pindar, though it is the epithet for Leto in h.Ap.
205.5
As noted above, other critics considered

ww as the norm. This idea
has its origin with Boeckh (1811), but actually the emendations which
Boeckh introduced were Wrst proposed by none other than Hermann in
1809 (De dialecto). Boeckh emends metri causa the initial part of three
repetitions thus (I leave str. 3 aside again):
str. 1     E
d 
.
(The Wrst syllable of  is scanned short, contra Snell and
Turyn Hermann, Notae)
str. 2 PE:
 a I
ant. 3

 K
d Y F 

Textually these emendations are not cogent.    (str. 1) and



(ant. 3) are hardly improvements. As for PE (str. 2), Pindar does not use
this Homeric form. He uses other forms for sg. acc.: P 5 times (O6. 76,
P8.65 ( 62 Sn), P9. 58 ( 56), N5.15, N6. 48 ( 46), and PA once (P12.
24 em. E. Schmid; P codd.). Nevertheless these three emendations
became indispensable for the text of N6s67 and established themselves as,
so to speak, a three-piece set. They had been adopted for a century by many,

4
According to Schroeders report, ed. maior2, 521.
5
Scholia register 
 (Drachamann iii. 109; 
 D).
114 Introduction
including Mommsen and Christ. It may be said that this three-piece set was
less oVensive than 

. An Entweder-Oder arose: either (un-
necessarily) to expel 
 or (unnecessarily) to introduce the
three-piece set. To avoid this dilemma, Bergk ingeniously invented a quite
diVerent metrical scheme with further sophisticated emendations (see
below), and Schroeder, instead of resorting to emendation, accepted some
of the manuscript readings as metrical licences, by introducing a theory
which supports these anomalous responsions.
So far I have left aside the emendations of str. 3. The paradosis Wts
metrically neither
www nor

ww at the beginning of s7:
[end of s6a] . . . f  
[s6b] E 

[s7] ` d a I 
,

Moreover it raises a question about the construction of E . Is it in appos-

ition to  (thus Hermann, who is followed, among others, by Turyn
and Snell)? Or should a transitive verb be introduced, e.g. E E
(Boeckh)? Note the scholia, which paraphrase thus: E b d K
B
c a  PE K
 f E `N, d
 e  Pe K H 
 . This reading includes
the verb K
 but not 
. Thus the whole sentence of str. 3 is
another big issue.
The third problem is the colometry, namely, the separation of s6b from
s6a. Some have felt that the length of s6b, a sequence of four positions, was
too short to be an independent verse. Thus they have tried to combine it with
either the preceding verse (s6a) or the following one (s7). But hiatus hinders
the combination in both cases: between s6b and s6a, E; j s (str. 2);
and between s6b and s7,
 j P (ant. 1; the paradosis is elided

 , which is unmetrical). Moreover, the metrical form of s6b itself is
questionable even if s6b is accepted as one verse. The responsion between
r
and

at verse-end is highly unusual. The penultimate position is
Wlled by a resolved long at str. 2 (K
 scanned trisyllabically ww
) and at
ant. 3 (and also at str. 3 in Snells text above; str. 3 is an emendation). Thus, to
restore the exact responsion at all the repetitions, Boeckh, for example,
introduces the unresolved form to ant. 3 by changing   into
 A and by changing the word-order in str. 2:
(MSS) K
d F, , E, j s K

(Boeckh) K
d F K
, t E , , s

In str. 3 the verse is occupied by E E ; a large-scale emendation (see

above).
 Appendix: Modern Emendations 115
Finally, some words or word-forms caused suspicion: for example,
a  (str. 1) or
 (str. 2). And, though there is no
metrical diVerence, there is a choice between K
A (plupf.) and 

(aorist). These are essentials for a rough sketch of the scope of the varieties of
emendations.6 Now we shall trace the history in chronological order.
It is generally acknowledged that the Wrst modern text of Pindar is
Boeckhs editio maior (1811). But before its publication, Hermann had
already published a dissertation, De dialecto Pindari observationes, in
1809, in which he proposed emendations for N6s67:
Nec praetereundum

 Ol. VIII 50 [ 38] correpta prima, pro


. [ . . . ].
ne quem moveat, ut a scribendum putet Nem. VI. 87 [53 51 Sn.]. in quo
carmine pene omnium stropharum Wnis corruptus est, moneo ibi sic legendum
videri: f   E 
, a f d z I 
, [ . . . ].

, ab sensu commendatum, ob metrum necessarium est, quum in vulgato

 , etiamsi vocalis sequatur, non possit brevem syllabam apostrophus ante
anacrusin excipere. . . . Sed ne cui in reliquis strophis metrum diYcultates obiiciat,
de his quoque dicendum puto. Ac v. 12. scribendum videtur    : v. 25.
F
 P  , [ . . . ] PB, forma ab Homero accepta, apud quem
est IE Iliad. XII. 318. denique v. 99   (seu  A)


 K
d Y. (Hermann, De dialecto, Opuscula, i. 2601.)

This is the Wrst occasion on which what above I call the three-piece set was
published. The corresponding lines are each adjusted to

ww. Also note
that Hermann implicitly accepts the manuscript word-order in str. 2. It must
be remembered that he follows the old division into lines preserved in
Heynes numeration.
Boeckh, in the Notae criticae incorporated in his editio maior of 1811,
argues at length against the metrical ideas in Hermanns Commentatio de
metris Pindari (which was published in 1799, ten years before De dialecto;
see Addendum I below). For our concerns, this part of Boeckhs argument is
not important. More important is his declaration that s6 should not be
divided but is a long verse, and that its Wnal part (i.e. what is later separated
as s6b) must not include resolution, which Hermann accepts (
wta) at
the penultimate position (see above). He gives a new colometry and emends
the text considerably (underlined) but, in fact, he accepts the three-piece set,
already proposed by Hermann.
s6
ww
ww


ww
rw
a
w
a
s7 t
ww
ww
w
w

6

 d (ant. 2 v. 36 35 Sn.) is an emendation of Triclinius for unmetrical
d (MSS). For this, see the Wnal paragraph of this Appendix.
116 Introduction
str. 1 
 K P N Pb a  

    E
d 
.
ant. 1
E K , n  
  r  F

P  Id

 ,
str. 2 u I
e   h K
d F K
, t E , , s
PE:
 a I
ant. 2 E 
 d H
 I
e  x 




b  
str. 3  P I
 K
A f   E 
 7
 
 a I 
,
ant. 3 :   Kg  
   ,  A,


 K
d Y F .

Hermanns De dialecto was reprinted in 1817 in Heynes three-volume

edition, and with it, he published another piece, Notae ad Pindarum, in which
he criticized Boeckh in turn and oVered diVerent solutions for several passages.
But at N6s67 he drastically changed his mind from De dialecto. His descrip-
tion in Notae is a kind of running commentary, but I reconstruct his whole
text, supplementing unmentioned words and metrical scheme:
s6b
wru
s7
wr
ww
w
w

str. 1 
 , P
  E
d 
.
ant. 1 F 

P  Id

 .
str. 2 s K

P
 a I,
ant. 2 





b  
str. 3 E 

 d a I 

ant. 3  ,


 K
d Y F ,

Underlined are the words or word-order which Hermann was the Wrst to
introduce. His approach is certainly drastic, but some of his emendations
have survived and are adopted in the modern editions:  (ant. 1),


 (ant. 2) are, as described above, printed by Turyn. Besides
these, both Turyn and Snell follow Hermann in changing word-order:
` 
 (str. 3). It must be remembered that 
 can be in this

7
Boeckh occasionally admits an elided vowel at the end of verse, contrary to our
notion of verse-end.
 Appendix: Modern Emendations 117
position because the second position of s7 is short. Boeckhs is proved to
be unnecessary (ant. 3). But instead, is introduced (str. 1;  is
scanned w
w). Hermann restores exact responsion in s6b, perhaps in
answer to Boeckhs critical comment. Interestingly, his exact responsion
includes resolution at all the repetitions (
wru), contrary to Boeckhs
(
w
w). The case of x  is changed into  (ant. 2).
 (<
) (ant. 1) is as bold as 

. From the point of view of
modern metrical scholarship, a diYculty is found in P (str. 1) at the
verse-end, which contradicts modern principles. One might combine s6b
with s7 in these texts. Then P would not be a problem, but correption of

 P in a non-dactylic context would produce a further problem.
Challenged by Hermann, Boeckh published a corrected edition (his editio
minor) in 1825. In that he changed str. 3 considerably:
s6
ww
ww


ww
rw
a
w
a
s7 u
ww
ww
w
w

str. 3  P I
 K
A f   E E
d a `f I 
,

The other emendations were left unchanged and the same as in his editio
maior. The emendation of str. 3 above was, in fact, proposed by Dissen
(1830) on vv. 52 V.: Quod nunc in textu vides, Boeckhius ex mea coniectura
reposuit, quum Scholiastes E videatur habuisse). Textually Dissen
(1830) is the same as Boeckh, ed. minor.
In 1840 Kayser published his Lectiones Pindaricae, which included a good
many conjectures. He explicitly maintained that E is in apposition
to  , against Boeckh: fortasse nullo mutato verborum ordo tantum
alius inferendus est: f  
  E d a f I

. [ . . . ] Dissenius in his (i.e. scholia) ipsius poetae manum agnos-
cebat, recipiens inde E E quod nobis secus videtur (Lectiones Pin-
daricae, 778). His metrical analysis of this text would be explained,
unsuccessfully, in his article of 1844 discussed below.
In 1843 two important editions were published. One was Schneide-
wins; although its title was Pindari carmina . . . ex recensione Boeckhii
commentario perpetuo illustravit Ludolphus Dissenius, editio altera auc-
tior et emendatior, the text is essentially Schneidewins, and he further-
more adds his own ample exegeses in the apparatus. Schneidewin divides
Boeckhs s6 into two verses (our s6a and s6b), which he gives the
following schemes:
s6a
ww
ww


ww
rw
a
s6b
wta
s7 a
ww
ww
w
wu
118 Introduction
str. 1 
 K P N Pb a 


    E
d 
.
str. 2 u I
e   h K
d F, , E,
s K

K I
 a I
str. 3  P I

 
 f  
 

 d a I 
,

The other three repetitions are identical with Boeckh (edd. mai. et min.).
I
 found followers. For example, half a century later, the manu-
script reading
 was denounced as unmoglich by Wilamowitz
(denn
 kann man von den Toten nicht sagen, GV 487), and LSJ
still has a vestige of this when it claims that
 in this passage as
lectio dubia.  instead of E was supported by Hermann in his third
publication (Nem. sextum, 1844). In reconstructing the text of N6s67,
Schneidewin owes a lot to Ahrens, who suggested the readings in str. 2 and
str. 3 (though Schneidewin did not follow him in str. 1). However, it was
Schneidewins own merit to divide Boeckhs verse into two, keeping the
modern notion of the verse-end. He further introduced in the apparatus two
conjectures by Ahrens as well as his own. One of them was adopted later by
Christ.8
Schneidewin app. crit.
str. 3 
 A a I 
,
Ahrens ibid.
str. 1 r    E
d 
.
ant. 3

  K
Y F 

The other edition to be published in 1843 was Bergks Poetae Lyrici Graeci
(Bergk1) which included Pindars epinikia. Bergk would introduce bold
conjectures in his later editions, but he was modest in this Wrst edition.
The text of s67 is basically the same as Boeckhs editio minor, except for
 instead of  (str. 1), proposed by Schneidewin.
In the next year (1844) Kayser reviewed both Bergk and Schneidewin
for Jahrbucher der Literatur. He makes several good comments and provides
his emendations as well. His metrical observation is keen. According to

8
Schneidewin contributes to the Pindaric text more than we see in the apparatus
of Snell or Turyn. He elucidates problems and passes sound judgements on the
proposals of his predecessors, especially Hermann. I presume his edition was enor-
mously helpful to later critics, e.g. Mommsen or Christ.
 Appendix: Modern Emendations 119
Schneidewins edition, the penultimate position is resolved at three repeti-
tions in N6s6b. Kayser remarks that resolution is unacceptable at this
position: In der AuXosung der trochaischen Dipodie scheint sich Pindar
zur Regel gemacht zu haben, den Tribrachys oder Anapast am Schlusse eines
Verses nicht zuzulassen (JL 106). He proposes the following emendations so
as to eliminate the resolution:
str. 2 s o (Wrst proposed by Kayser) instead of s K

ant. 3  A (already proposed by Boeckh) instead of  

For str. 3, he introduces his former suggestion slightly modiWed

(A a instead of d a ):
f  
  E A a  I 
.

Unlike the two emendations above, this text does not Wt Schneidewins
colometry, unless Kayser accepted a verse-end within a word: s6a ends
with f  
- and s6b is occupied by -  E (
w
w). Thus
Kayser must have assumed that s6a and s6b were one united verse (like
Boeckh). However, when s6a and s6b are united, hiatus occurs at str. 2
between E and s (Boeckh changed the word-order). It is not clear
how Kayser viewed the matter: he might have accepted the hiatus by some
unknown reasoning. At any rate, he was apparently the Wrst to suspect the
resolution at the penultimate position of a verse. This suspicion would be
revived after seventy years by Paul Maas (see Addendum II); and, the change
of word-order, f  
  E , would be accepted, along with
Boeckhs emendations and colometry, by Christ.
In the same year (1844) Hermann again published a newly revised text,
which is proved to be his last (Nem. sextum). As in the other two publications
on Pindar of his later years, namely Emend. P. (1834)which, despite its title,
includes emendations on the Pythians onlyand Quinque Ol. (1847), Her-
mann in this article makes many acute observations as well as bold conjectures.
He is especially keen to restore exact responsion: Accedit quod quaedam a
grammaticis metri maxime caussa videntur mutata esse: quo proclivior etiam
recentioribus criticis via ad errorem est, plus sibi licentiae sumpsisse Pindarum
putantibus, quam per severitatem lyricae poesis concessum erat. s6b is totally
regularized and united with s7:
s6a
ww
ww


ww
rw
a
s6bs7
wr

wr
ww
w
w

str. 1 
 K P N Pb a 

 i   E
d 
.
ant. 1
E K , n  
  r 
F 
 P  Id

 ,
120 Introduction
str. 2 u I
e   h K
d F; , E,
s K
 P
 a I
ant. 2 E 
 d H
 I
e 


 


b  
str. 3  P I
 K
A f  
  
; A d I 
,
ant. 3 :   Kg  
  
 

 K
d Y F 

The metrical scheme is not diVerent from that in Notae except that the last
position of s6b is Wxed as long without hiatus. In other words, Hermann
expelled
w

from, and introduced
wr
, to the corresponding
part of s6b at every repetition, and also adjusted the beginning of s7 to the
scheme
wr. Some of his conjectures are undeniably bold, however. Of
them, str. 1 L   (instead of  ) is a completely
new proposal. Str. 3 A d (instead of d a =a ) fol-
lows the A a of Schneidewin (app.crit.) or Kayser (JL). Note that the
monosyllabic word  now moves one position forwards (str. 1). It occupies
the position which had been occupied by P in Notae (
 is scanned as
double short as before). Also note that he revokes his rejection of the unusual
form a in De dialecto cited above. At ant. 1, F 
 , has some
manuscript support (F 
 ;  has been excluded).9  is from
SchneidewinAhrens; P, ; 

 are from Notae. It
should be noted that Hermann, here too, totally rejected Boeckhs three-
piece set (   =PE=

 ).10
Also in 1844, Rauchenstein published a pamphlet (Commentationum
Pindaricarum particula prima) containing some good proposals. In N6s6
7, he proposed an emendation of str. 3, which he would later withdraw:
s6
ww
ww


ww
rw
a
w
a
s7 u
ww
ww
w
w

str. 3  P I
 K
A f   B E
a d p f I 
,
or
a 
 f I 
,

Bergk, in his second edition (Bergk2) of 1853, establishes exact responsion

in a quite diVerent manner from both Hermann and Boeckh. He ingeniously
succeeds in expelling Hermanns  and 

 as well as the

9
For this emendation, see Addendum II below.
10
Hermanns contribution to the text of Pindar is enormous. But we should
admit that he was too keen to seek exact responsion, especially in his later years.
 Appendix: Modern Emendations 121
three-piece set of Boeckh. Whether his text is closer to the truth is, however,
another matter.11
s6a
ww
ww


ww
rw
a
s6bs7
ww
a
ww
ww
w
w

str. 1 
 K P N Pb a 


e   E
d 
.
ant. 1
E K , n  
  r 
F 
 P ::::::::::::
 Id

 ,
str. 2 u I
e   h K
d F, , E,
s K
 PE
 a I
ant. 2 E 
 d H
 I
e 
x 

 :::::::::::::::::::::
:::::::: 

b  
str. 3  P I
 K
A f  
 `  , s d a I 
,
ant. 3 :   Kg  
   ,
 K A (sic),

 K
Y F 

His text starts with accepting x

 
 (ant. 2) as the
standard. But, unlike the previous editions, he scanned
 
 . Following
this,
 (str. 1) and K
 (str. 2) too were scanned as w
. Thus what
had been scanned in all the previous editions as
w
a and
wr

became
ww
.  (ant. 1) occupies the position of , which is written in
the manuscripts but is meaningless, and had been deleted by many. Elided

 which had been recovered by Hermann (Nem. sextum) was intro-
duced. K- (ant. 3) is newly inserted. Thus all the s6b-parts became
ww
.
The most ingenious parts of Bergks new text are KA and  (not
 ). These words enable Bergk to keep the traditional 
 as
Boeckh does, but at the same time they dispense with

 K
d.
Compare:
Boeckh, ed. min.  A,

 K
d Y
Bergk2  KjA,

 K
Y
In this Bergks new scheme, -A occupies the position which
- occupies
in Boeckhs. Also compare:
Schneidewin  f 
 d a
Bergk2  j , s d a

11
Bergks sense, both metrical and stylistical, is very keen, perhaps next to Hermanns.
Like him, he does not hesitate to think radically and to propose an original colometry
which involves rewriting the majority of the repetitions. Bergk is also similar to Hermann
in discarding his previous emendations for the sake of a new idea in later editions. It is
important for us to note in which edition an emendation is proposed.
122 Introduction
` covers not only the latter half of s6b but the beginning of s7.
Instead, the main verb 
 (or a replacement such as Schneidewins

) is expelled.
From this time onwards, philology gradually declines into an intellectual
game. The grammar may become more correct, the meaning more precise,
and the metre more exact; but we may wonder whether Pindar actually
wrote like that. As an illustration, I give Hartungs text (1856):12
s6a
ww
ww


ww
rw
a
s6b
ww

s7
wr
ww
w
w

str. 1 
 K P N Pb  


 K E
d 
.
ant. 1
E K , n  
  r 
F K

P  Id

 ,
str. 2 u I
e   h K
d F  E,
s K

P
 a I,
ant. 2 E 
 d H
 I
e 
x 



b  
str. 3  P I

 
 f  
E d
E  a I 
,
ant. 3 :   Kg  
 
I


 K
d Y F 

Apparently Hartungs main concern was to recover
ww
at s6 in all the

repetitions, so scanning s K
 as well as 
 and
x

 . This idea may have come from Bergks second edition, but
his conjecture is too bold to be considered as likely.
In an article of 1858, Zu Pindars Nemeen, Rauchenstein approved of
Hartungs idea that, at str. 1, two interrogative pronouns, instead of one,
were necessary. He tried to be less radical. His metrical scheme is the same as
Boeckhs.

12
Hartung proposes many bold conjectures in all the odes. He occasionally has a
good idea; for example, the division of P10s2 into two verses with some emendations,
P10e34 (v. 1516), or P5 v. 118 E. But his text also includes many implausible
rationalizations. Most of his emendations seem to me to be Ximsy. In Part II, I often
ignore his proposals, unlike those of other big names.
 Appendix: Modern Emendations 123
s6
ww
ww


ww
rw
a
w
a
s7 a
ww
ww
w
w

str. 1 
 K P N Pb   

   E
d 
.
str. 3  P I

 
 f  
E  ,
d a f I 
,

Tycho Mommsen had worked for long years on Pindars manuscripts. It

was he who Wrst used two important vetusti, A and B. His edition was Wnally
published in 1864. Mommsen was more prudent than his contemporaries,
and his judgement was sober. Perhaps we do not Wnd his name as frequently
as Bergks, but it should not be forgotten that he scrutinized the emend-
ations which had already been proposed by his predecessors.13 In N6s67,
his text is basically the same as Boeckhs editio minor (note the adoption of
the three-piece set), except for str. 3, which is his own conjecture:
s6
ww
ww


ww
rw
a
w
a
s7

ww
ww
w
w

str. 1 
 K P N Pb a  

    E
d 
.
:::::::
ant. 1
E K , n  
  r  F

P  Id

 ,
str. 2 u I
e   h K
d F, K
 t ::::::::::::::::::::::::::
E , , s
PE
 a I,
:::::::::::
ant. 2 E 
 d H
 I
e  x 




b  
str. 3  P I
 K
A f   E 

a  K
 I 
,
ant. 3 :   Kg  
   ,  A,


 K
d Y F 

Two years later (1866) Mommsen published his editio minor. In the same
year Bergk published the third edition of PLG. As in the second edition, s6b
is not combined with s6a but with s7. But its form is radically changed, and
is diVerent from all the other editions:
s6a
ww
ww


ww
rw
a
s6bs7
wy

ww
ww
w
w

str. 1 
 K P N Pb a 

    E
d 
.
ant. 1
E K , n  
  r 
F
 P  Id

,  ,
13
Together with his major edition Mommsen published Annotationes criticae
supplementum ad Pindari Olympias. It is a pity his adnotationes did not cover the
other epinikia.
124 Introduction
str. 2 u I
e   h K
d F, , E,
s K
 PA N a I,
(nisi malis P I
)
ant. 2 E 
 d H
 I
e PF
 

b  
::::::::::::
str. 3  P I
 K
A f  
E f E , d a I 
,
::::::::
ant. 3 :   Kg  
   ,
 ,

 K
Y  F 
:::::::::

The most important change is the beginning of s7: the total length of the
second verse (s6bs7) becomes one position shorter. Compare Bergk3 with
some of previous editions and Snell:
Boeckh (ed. minor)
w
a k u
ww
ww
w
w

Schneidewin
wta k u
ww
ww
w
w

Hermann (Nem. sextum)
www

www
ww
w
w

Bergk2
ww
u
ww
ww
w
w

Mommsen
w
a k

ww
ww
w
w

Snell
wt
k
at
ww
w
w

Bergk3
wt

ww
ww
w
w

   (str. 1), one of the three-piece set ( BD) in Boeckh and others,
is changed into   . The idea of two interrogative pronouns comes from
Hartung. Later the paradosis  would be recovered in Bergk4, but in this
edition, Bergk believes the fourth position must not be anceps but long.

 is elided (ant. 1); the elision is actually found in the paradosis.


 is deleted while x is changed into  (ant. 2) with 
into P. One position is expelled from
N (str. 2) and
K
d Y (ant. 3). Presumably Bergk was eager to seek a diVerent way
from both Hermann and Boeckh. He rejected both Boeckhs three-piece set
and Hermanns 

. This aim may be admirable. However, it is
implausible that I
e PF  (ant. 2) is right.
In his fourth edition (1878), Bergk slightly modiWes the metrical scheme
and introduces the manuscript reading  at str. 1:
s6bs7
wy u
ww
ww
ww
w
w

The others are the same as the third edition, except for 
 (in the
third edition, 
 ). Bergk34 would later be adopted in principle
by Wilamowitz.
In 1869, three years after Mommsens editio minor and Bergk3, the
Bibliotheca Teubneriana published an edition of Pindar edited by Wilhelm
Christ. Christ later published his editio maior in 1896 with ample commentary.
 Appendix: Modern Emendations 125
Successively his editio minor (BT) was revised in the following year
(1897). Although it lacks the commentary, in text and apparatus it is
basically the same as the editio maior. Both his editio maior and BT are
excellent. His text would not have achieved much without thorough
examination of his predecessors; the report of previous conjectures is
very accurate.14 As for N6s67, his text is basically the same as Boeckh
Dissen, except str. 3, where he follows Kayser (Lectiones Pindaricae):
s6
ww
ww


ww
rw
a
w
u
s7 u
ww
ww
w
w

str. 1 
 K P N Pb a  

    E
d 
.
:::::::::::
ant. 1
E K , n  
  r  F

P  Id

 ,
str. 2 u I
e   h K
d F :::::::::::::::::::::::::::::::::::::::::
K
, t E , s
PE
 a I
ant. 2 E 
 d H
 I
e  x 




b  
str. 3  P I
 
 f  :::::::::::::::::::::::::::::

  E
d a f I 
,
:::::::::::::::::::::::::::::::::::
ant. 3 :   Kg  
    ::::: A::,


 K
d Y F 

Christ was a metrician too. He wrote a book on metre (Metrik der Griechen
und Romer, Leipzig, 1874). Like RossbachWestphal he introduces loga-
oedic with cyclic dactyl etc. His presentation of metrical scheme is dras-
tically changed from the ones from Boeckh to Mommsen.15 As for our
passage, one thing is certain. He does not like free responsion in Pindar.
He even tries to minimize actual occurrences of ancipitia. Note his conjec-
ture in the apparatus for str. 3: fort. a f d r , where the
initial syllable is long like the other repetitions.
Abhorrence of free responsion is noticeable also in RossbachWestphal,
the most celebrated book on metre for decades. I just cite the scheme from
the second edition of Griechische Metrik (1868), without any comment,
because their theories, based on analogy to modern Western music, have
already proved unncessary and futile. It is uncertain on which text West-

14
Christ usually abstains from printing his own conjectures in the main text in an
exemplary manner. At the same time, in the apparatus we Wnd many good proposals
which are based on deep consideration with keen sense. They are, and will for ever be,
a challenge for future generations.
15
Until Christ, the segmentation of verses had been basically the same as my
Rules. cf. the illustration above at pp. 1920 n. 33. Here in the Appendix, I omit ictus
() from their metrical schemes and reproduce their analyses in the current fashion.
126 Introduction
phals analysis is based. The line numbering is his; thus 7, 8, 9 correspond to
our s6a, s6b, s7.
2 Bd. II, 4. die logaodischen Metra.
61. Logaodische Strophen des Pindarischen Stils.
S 825 Nem. 6 str. Tetrapodisch-dipodisch
7
ww
ww


ww
www
a
8
w
a ^ ^
9 w
ww
ww
w
w
^

It is fair to note that, although it is less artiWcial, Christs idea of metre is not
far from Westphals.
The trend started changing at the end of the nineteenth century. 1897 was a
dramatic year in classical scholarship. The papyrus of Bacchylides was pub-
lished, and some apparent anomalies of responsion helped new theories of
metre to be launched. Sixteen years later, Maas described the situation as follows:
Bis vor etwa 16 Jahren wiesen die modernen Pindartexte keine anderen Freiheiten der
antistrophischen Responsion auf als Vertauschung von Lange und Doppelkulze in
einigen Hebungen . . . und Vertauschung von Lange und Kurze in einigen Senkungen.
Die ubrigen Inkongruenzen, die die Uberlieferung enthalt oder zu enthalten scheint,
waren seit dem 13. Jahrhundert allmahlich samtlich beseitigt worden, sei es durch
Deutung der Uberlieferung, sei es durch Annahme einer Korruptel. Nach den selben
Prinzipien wurden 1897 die neuen Lieder des Bakchylides in der Erstausgabe behan-
delt, wenn auch fur manche Einzelheit keine befriedigende Losung gegeben werden
konnte. (Freiheiten, 1.)

Schroeders editio maior was published in the last year of the nineteenth
century (1900); the publisher printed it as the Wfth edition of Bergks PLG.
Schroeder introduced into Pindars poetry a new metrical theory, which
resembles Wilamowitzs idea, but later develops it to the extreme.16 In
general, he tends to seek an example of irregular responsion in manuscript
readings, and to use it as proof of some more wide-ranging metrical theory,
even if it is not fully paralleled. Here in N6s67, he introduces two free
responsions:
s6a
ww
ww


ww
rw
a
s6b
w
a

ww
(s K
)
s7

ww
ww
w
w

16
To be correct, such metrical theory as Dreiheber/Vierheber or strophic analysis
based on the number of Hebungen is still absent from his editio maior. What he
introduced here is free responsion between e.g.
ww
and
w

in D/e.
 Appendix: Modern Emendations 127
w
ww (K I-)

w

( -)
ww
w ( ` -)
str. 1 
 K P N Pb a 17


  
d 
.
ant. 1
E K , n  
  r 
F

P  Id

 ,
str. 2 u I
e   h K
d F, , E,
s K

K I
 a I,
ant. 2 E 
 d H
 I
e 
x 




b  
str. 3  P I
 
 f  
E 

 d d I 
,
ant. 3 :   Kg  
  
 A,


  K
d Y F 

We need to go back to the chronological order. In 1890, ten years before

the publication of Schroeder, Bury had written a commentary on the
Nemeans. Bury honestly admitted his ignorance of metre and wrote: in
regard to Pindaric metres, I have adopted with hesitation the conclusions of
M. Schmidt. As I have not made a thorough study of Greek metric, I do not
feel competent to pronounce on a subject which demands the concentrated
powers of specialists (Preface, vi). In N6s67 he reads:
s6
ww
ww


ww
rw
w
s7
wty
ww
ww
w
w
18
str. 1 
 K P N Pb a 

    E
d 
.
ant. 1
E K , n  
  r 
F
  P  Id

 ,
::::::::::
str. 2 u I
e   h K
d F, , E,
s K
 PA N a I
:::::::::::::::::::::::::::

17
In apparatus: . . . a  exspectandum esse sensit Hartung, quod reposui
cum Wilamowitzio (Eur. Her2 II 232, v, 1106), Hartung ipse . . . ; for Hartungs text,
see above. If bei Tag oder Nacht (Wilamowitz) should be rendered from
K . . . a , not by  , hiatus would occur (as in str. 2) so that
s6b must be divided from s6a.
18
The metrical chart on p. 102 is misprinted; cf. on v. 50, p. 111.
128 Introduction
ant. 2 E 
 d H
 I
e 
x 

 

b  
str. 3  P I
 
 f  
E  A A a I 
,
:::::::::::::::::::::
ant. 3 :   Kg  
   ,
 ,

 K
Y F 

In s7, these free responsions are admitted:

str. 1 
 
w
ww

ant. 1 F
 P -
w
ww

str. 2 s K
 P-
wr


ant. 2 x 

 -
wr


str. 3 E  A-
wr


ant. 3  ,
-

wr


If Hermann had been alive, what would he have said about such extraor-
dinary metrical licence? However, it became a trend in the twentieth century
to defend abnormal responsion in manuscripts by introducing new metrical
theories. In 1908, Schroeder introduces new metrical schemes in his editio
minor (BT), based on the same text as his editio maior in N6s67. Here the
choriambic dimeter is explicitly launched for the Wrst time in a edition of
Pindar:
N6s6
ww
r


ww
rw
a
N6s6b
wy

N6s7
ww
w
w

represents any combinations of four syllables of either long or short:

w

( - ),

ww (P  ), w
ww (K
-),


ww (
-), ww
w ( ` -),
www (

 K
).
In 1914, Schroeder amended his text in the second edition of his editio
minor (BT2), after accepting the advice of Maas (cf. Praefatio: In plagulis
corrigendis amice et sollerter me adiuvit Paulus Maas non sine magno (velut
N vi str 9, . . . ) poetae ipsius commodo. A new metrical scheme is also
introduced for s7 (his str 9), which would be reproduced without any change
later by Snell:
s7
at
ww
w
w

str. 1 
 K P N Pb a 


  E
d 
.
ant. 1
E K , n  
  r 
F

P  Id

 ,
 Appendix: Modern Emendations 129
str. 2 u I
e   h K
d F, , E,
s K

K
::::::::
 a I,
ant. 2 E 
 d H
 I
e 
x 




b  
str. 3  P I
 
 f  
E 

 d d I 
,
ant. 3 :   Kg  
  
 ,


 K
d Y F 

Choriambic dimeter survives. It is unfortunate that, in our passages, Snells

text as well as his metre follow this edition of Schroeder.
Maas criticized the free responsion: ich habe . . . , wahrend ich fur Schroe-
der die Korrektur des Pindarstextes von 1914 las, die Uberzeugung von der
Unzulassigkeit der ganzen Gattung von Anomalien gewonnen (Freiheiten, 2).
But it took a long time for his ideas to be vindicated. The anomalies in the D/e
poems were wiped out comparatively early (however, as late as 1930, Bowra
still criticized the free responsions in D/e: An Alleged Anomaly), but those in
the non-D/e poems survived.19 It is uncertain to what extent the metrical
theories of Schroeder or Wilamowitz were correctly understood by non-
specialists. Apparently Sandys and Puech were under their inXuence.
I cite below in detail only the last part of s6 (the s6b-part) and the
beginning of s7 of Sandys (Loeb, published in 1915).
s6
ww
ww


ww
rw
a
wta
s7
ar
ww
w
w

The metrical scheme of Sandys is basically the same as that of Schroeder,
BT2 (1914). Textually Sandys diVers from Schroeder in two words (under-
lined).
str. 1 
    -
w



ww
ant. 1 F
 P 
w



ww
str. 2 s K
 P
-
wr

www
ant. 2 x 

 
-
w



ww

19
These are some of examples of licentiae antistrophicae which Schroeder
introduces in his text by keeping the paradosis: O9s10

ww
aw
(reading
 at v. 76), O10s
aw
y
w
(reading  at v. 15), N7s7a

taw
w
(reading  at v. 80).
130 Introduction
str. 3 E `
  -
wr

www
ant. 3  ,

 K
d
wr

www
The scheme of Puech (Bude, published in 1923) is much freer.
s6a
ww
ww


ww
rw
a
s6b
w
a
s7 uaau
ww
w
w

This is the metrical scheme Puech himself gives.20 Like the case of Sandys,
I cite s6b and the beginning of s7.
str. 1 
    -
w
w

w

ant. 1 F
 P 
w
w

ww
str. 2 s K
 PA
-
wr

w
w
ant. 2 x 

 
-
w



ww
str. 3 E 
 `
 -
w
w ww
w
ant. 3  A,

 K
d
w


www
Puechs own comment on s7 well explains his position:
La premiere dipodie de ce dernier colon ne peut etre ramenee, quoi quon fasse, au
meme type. Je nai admis aucune correction pour les vers de la 1re antistrophe, de la 2e
strophe, de la 2e antistrophe et de la 3e, qui ne paraissent suspects ni par le sens ni par
lexpression; jen ai admis une, qui est facile, au vers de la strophe 1; une autre qui est
plus douteuse, au vers de la strophe 3. On obtient ainsi partout quatre syllabes, mais
qui prennent les formes


w
,
w
w,

ww, ww
w,
www.
Cest un exemple unique chez Pindare. Vu lextreme diYculte dune solution
satisfaisante, je crois utile de renvoyer le lecteur au dernier ouvrage de Wilamowitz,
Griechische Verskunst, Berlin, 1921, ou il en trouvera, p.487, une autre assez diVerente.

Wilamowitzs text and the metrical form given in GV are based on the
fourth edition of Bergk (Bergk hat Recht gegen Boeckh und Schroeder).
The characteristics of Bergk4 are: (1) the total length of the second verse
(s6bs7) becomes one position shorter; (2) nevertheless, the paradosis is for
the most part preserved. Wilamowitz follows Bergk in elided
 (ant.
1); N (emendation, str. 2); K
Y (ant. 3) :Fur die Bestimmung
des Versmaes scheidet eine so verdorbene Stelle aus. Dann stehen vier
gegen eine, und es ist unerlaubt, die vier zu andern. The text in str. 3 is

20
The indentation seems to imply that s6b and s7 form one verse, but two
juxtaposed ancipitia defy this interpretation. Hiatus occurs at ant. 1, but Puech
may suppose epic correption occurs. Apparently he is not good at metre.
 Appendix: Modern Emendations 131
slightly diVerent, but basically the same as Bergk. The weak point of Bergks
text lies in ant. 2 as I have already pointed out above.
Aber auch die widerstrebende Stelle lat sich nicht zwingen; es ware doch bare
Willkur, anzunehmen, da 
 ein anderes Epitheton wie 
verdrangt hatte. Wir konnen nur konstatieren, da vor dem Hemiepes viermal

wtw, einmal
ww

steht, fur unsere Kenntnis unerhort. Also gehort
wieder die Alogos vor diesen Vers.

The Alogos in s6b is, however, not paralleled at all.

Wilamowitzs text reads:
s6a
ww
ww


ww
rw
a
s6b
wta 
ww


s7
ww
ww
w
w

str. 1 
 K P N Pb a 


  E
d 
.
ant. 1
E K , n  
  r 
F
 P
 Id

 ,
str. 2 u I
e   h K
d F, , E,
s K

PA N a I,
:::::::::::::::::::::::::::
ant. 2 E 
 d H
 I
e 
x 

 -


b  
str. 3  P I
 
 f  
E 
t d a I 
,
ant. 3 :   Kg  
  
 ,


 K
Y F 

Schroeder published the second edition of his editio maior in 1923. Although
the publisher called it pt. 1 vol. 1 of the six edition of PLG, revisions are not
printed in the main text butare collected inan appendix, where they easily escape
notice. It is evident that Schroeder owes much to Wilamowitz and Maas. In N6s7
he regularizes all the repetitions as
wy
ww
w
w
and notates ^ iat
ch ia:  - (str. 1), P  (ant. 1), P
-
(str. 2), 
 -d (str. 3),

 K
d (ant. 3). Most noticeable is


- (ant. 2), without mentioning Hermanns name.21

21
Instead, Schroeder mentions Maas (in idem inciderat P Maas, 521). Later in
his third edition of BT (1930), he prints 

 in the text, and notates
simply Maas in the apparatus.
132 Introduction
Bowra (OCT published in 1935) is in confusion. His text is fundamentally
the same as Schroeders (ed. maior2 BT3) on N6s67 and adopts


 (rightly citing Hermann) with others (see above), but at
ant. 1 he sticks to P  . Perhaps this is a simple mistake.
Now we are back at the point where we started: Turyn. Were he compared
with his predecessors in the main stream for the last hundred years, Turyn
might seem too conservative. But at N6s67, he virtually follows Schroeder
(ed. maior2 BT3). Remember that Snell follows Schroeders previous edi-
tion, BT2. I follow Turyn.
A word on 
 d (ant. 2 v. 36 35 Sn.). This is an emendation by
Triclinius, one of his best. All the critics mentioned above accept it in place
of the paradosis, d , which Hohl, however, defends:
Gerade die Seltenheit des Wortes spricht fur seinen Gebrauch durch Pindar. [ . . . ], es
einerseits wohl moglich ware, einer Erklarung fur die Anderung von d in

 d mit ziemlicher Sicherheit anzugeben, wahrend es anderseits unmoglich
ist, einen Grund fur die Anderung von 
 d in d einzusehen. [ . . . ]
Es erscheint kaum glaublich, da die angeblich authentische Leseart 
 d ,
welche metrisch einwandfrei ist, spurlos untergegangen sei und einer fur den damali-
gen Standpunkt fehlerhaften Glosse oder Interpolation Platz gemacht habe. (Respon-
sionsfreiheiten bei Pindar, 64.)

However, we know the text tends to be easily corrupted for various reasons,
carelessly or even consciously. On the other hand metrical irregularities (in our
case, contraction of double short in D which occurs at this repetition only) were
never introduced by the poet. Hohls case is extreme, but such an argument will
never cease. We should trust more the regularity of metre as well as grammar
than the manuscripts (including papyri) which are not always trustworthy.

Addendum I: Hermann, De metris pars ii (1799), 319

Hermann had already analysed the metre of N6, as well as all the other
epinikia, before writing the Notae (see above) incorporated in the third
volume of Heynes edition (Leipzig, 1817). The earlier analysis was also a
part of the third volume of an earlier edition by Heyne, that published in
Gottingen in 17989 (its third volume was published in 1799).22 This
included a two-part Commentatio de metris Pindari (Gerbers Herm. 1).
Parts I and II were later revised and published respectively as De metris

22
I have had diYculty in tracing the bibliographical history of Heynes editions.
The Wrst edition was published in Gottingen in 1773, and was reprinted, with revision
and/or additional material, several times by diVerent publishers. I have consulted the
copy in the Bodleian Library: 88 S 103108 Linc.
 Appendix: Modern Emendations 133
Pindari (Gerbers Herm. 5a) and Notae ad Pindarum (Herm. 6), in
Heynes 1817 edition.23
In part II Hermann oVers an analysis of the above verses in N6s, useless in
itself, but interesting as a historical document for tracing his own develop-
ment, which is, in a sense, to trace the origin of metrical study in modern
times. Hermann does not yet introduce radical emendations, but accepts
free responsions (as Schroeder would, though these two scholars theories
are completely diVerent). Acccording to Hermann in 1799, the metre of
N67 is ionic a maiore, which 9
he would later expel from Pindaric metre.
8
s6bs7a
w
a >
= >
<

ww
r

w
a
>
; >
: ww

s7b
ww
w
w

str. 1 
  -
w
w
w


 
d 
.
ant. 1 F
 P 
w
w
w
w24
Id

 ,
str. 2 s K
 PA
-
wr

w
w
 a I,
ant. 2 x 

 
-
w



ww

b  
str. 3 E 
-   -
w
w r

w
d a I 
,
ant. 3  ,

 h i K
d
wr


ww
Y F 

Hermann writes: Decimus strophae versus [our s6bs7a] Ionicus a maiore

est, . . . in Ionico ubique trochaei sunt. In the Wrst metron, anapaestus in
Ionico pro trochaeo, ut saepe (str. 2; ant. 3). Trochee is used also in the second
metron in str. 1, ant. 1 (note  which is the reading of the vulgata
( Triclinius) and D; B, which preserves  , was as yet unknown), str. 2
(the emendation of P into PA). In str. 3 anapaestus in principio
alterius Ionici pro trochaeo. Purus Ionicus occurs only in ant. 2 (note that

 is kept) and in ant. 3 (note the inserted ). This notion about
ionic a maiore is not a peculiarity of Hermann, but has its origin in Hephaes-
tion, from whose authority the young Hermann was not free;25 ionic a maiore
is treated at length in De metris poetarum Gr. et Rom. (1796) too.

23
The copy I have consulted is in the Bodleian Library: 88 M 3941 BS.
24
He writes: dipthongi emollit eius correptio.
25
Hermann does not totally discard ionic a maiore in Elementa doctrinae metricae
(1817). The examples he gives are what we now call Sotadean.
134 Introduction

Addendum II: Maas, Nachlese (1916)

After publishing the Wrst part of Die neuen Responsionsfreiheiten in 1913,
Maas wrote a short paper, Nachlese zu Pindar (reprinted as no. 6 of his
Kleine Schriften), in which he discussed several passages including N6s67.
Here he unites s6b with the following verse (s7). What prevents uniting these
verses is the hiatus between F fg
 and P  (ant. 1).
Maas emends F fg
 into F 
 , getting rid of the
hiatus. F 
 had already been proposed by Hermann in Nem.
sextum (see above), although Maas does not seem to have noticed it. The
meaningless  in the manuscripts, which all scholars delete, now becomes
explicable. For F  Maas cites two Aeschylean passages as parallels (see
below): Aeschylus hat F  im Sinn von verstarktem F. His main reason
for the uniWcation of s6b with s7 is metrical: the resolution at the penulti-
mate position in s6b is, according to him, totally unparalleled.26 Moreover, if
we adopt the following metrical scheme for s6bs7, none of the manuscript
readings, except for the transposition of 
  , needs to be
emended:

wtu
ay
ww
w
w
k
I am not sure how he interprets it but am rather sceptical about its possi-
bility. Especially, the two ancipitia divided only by one long are improbable.
However, even if we reject this scheme and follow Turyn, who adopts the
emendations proposed by Hermann (Notae 1817), such as 


in place of 
, Maass rejection of the resolution at the penulti-
mate position must be seriouly taken into consideration.
Two questions, (1) metrical and (2) grammatical, arise; and I judge Maass
proposal to be untenable.
1(a). There are four verses whose penultimate positions are resolved,
according to my analyses, in freer D/e verses in the eighteen majors. In
Part I, I prefer (i) the analyses which result in resolution at the penultimate
position, to (ii) the alternatives which eliminate this resolution, in all the
four verses. Presumably Maas will have preferred (ii) by assuming a diVerent
direction of resolution. The reason why I prefer (i) is that long anceps is
never followed by resolved long in the eighteen majors: see Part I, 6, C (ii).
Besides, there are other reason to prefer (i) in some individual cases.

26
This observation was, as far as I know, Wrst published by Kayser in 1844 (see
above).
 Appendix: Modern Emendations 135
O10e1b
(i) x
wr
k x e

(ii)

rw
k sp e
According to (i) the Wrst position can be Wlled by a short syllable. Thus
emendation of
 at 57b ( 55 Sn.) into
 proves to be unneces-
sary. If resolution were supposed to occur in the direction illustrated in (ii),
the initial position must be not anceps but true long; so the emendation
above becomes necessary.
P6s6
(i) www
ww
w

wr
k gl e

(ii) www
ww
w

rw
k gl
e
Long link anceps is hardly ever found after an aeolic phrase: see Part I, 8. B. 5.
There is no example of gl
e in the eighteen majors. In contrast gle
is
found in I8s5c (and besides, there are many examples of gle).
P11e5
(i) w
w
w

wr
j w e2 e

(ii) w
w
w

rw
j w e2
e
(ii) introduces cut after long implicit link anceps at two repetitions; but in
this position we usually encounter bridge (see Part I, 6, D); cut irregu-
larly occurs at nine out of 74 repetitions.
I8s8
(i) rw

wr
wr
k e e e

(ii) rw

rw
rw
k e
e e
According to (i), this verse would be unique in that two resolved e, whose
directions of resolution are diVerent, are juxtaposed. Usually, an unresolved
cretic, or an anceps, is found between rw
and
wr; cf. Part I, 6, C (ii).
(ii) would eliminate this irregularity, but, as is described above, it would
introduce long anceps followed by resolved long, which is totally unparal-
leled. (i) is preferable.
1(b). There is one case in the eighteen majors, and there are three cases in
Pae6, of correspondence between resolved and unresolved at the penulti-
mate position.
N3s6

w
a
ww t
k
e x d

Resolution occurs only at v. 14: I
. The reading is suspect; see Part II,
ad loc.
136 Introduction
Pae6s4 w
w

ww
wt
j w e ar
Resolution occurs at v. 149 : .
Pae6s6
ww

w
wwt
j d ph
Resolution occurs at v. 131 I
.
Pae6e6 w
[ ]
w
wwt
k [ ] ph

Interestingly, the same word causes the resolution as in Pae6s6: v. 176

I
.
Whether it is an aeolic phrase or freer D/e, the sequence
wwt
k is
diVerent from
wt
k. My intuition accepts the latter, but rejects the
former (there is no example of pherecratean with resolved penultimate
position in tragedy). In any case, N6b does not belong to the former group.
In sum, I judge that resolution at the penultimate position is certainly
rare, but may occasionally occur. Thus it is not wise to introduce emend-
ation just for this reason.
2. What Maas means by im Sinn von verstarktem F is, I think, equal to
the description of KuhnerGerth ii. 446 (566): Die Dichtersprache ge-
braucht zuweilen st. des einfachen F mit grosserem Nachdrucke F ,
d. i. F K, . Kuhner cites the same Aeschylean passages as Maas:
A. Sept. 705 and Suppl. 630. In these two passages F  stands at the
beginning of a sentence (or, you may say, a defective sentence), while in N6
v. 13b it does not. It comes in the middle of a relative clause.
Although Maas does not cite it, Hermann (Nem. sextum) refers to
S. Aj. 802 for introducing F 
 in place of F fg
. In
his edition of Sophocles Ajax, Hermann cites those two Aeschylean passages
as parallels, among others which are not necessarily so. However, the
sentence constuction of S. Aj. 8012 is diVerent from the others. F must
go with . Moreover, all these three passages in tragedy are emended, or at
least, suspected, by recent editors. My judgement is that F  may be
accepted at the two Aeschylean passages, although they are both suspect with
good reasons. S. Aj.802 is diVerent from these two, and is not a parallel.
Hermanns and Maass emendation, which depends on them, ought to be
rejected.
 Part II

The Eighteen Majors

This page intentionally left blank
 I N T RO DU C T IO N

Each of the eighteen majors is metrically analysed in Part II; for the
four minors see, the Appendix. Each analysis is organized as follows:
(1) metrical scheme; (2) lines with hiatus/brevis in longo at the end;
(3) sample text; (4) general remarks about the ode; (5) analyses of
strophe and epode; (6) textual problems; (7) individual verses.
In the metrical scheme, if verse-end is guaranteed by hiatus/brevis,
the mark k is placed at the end of the verse. If not, the mark j is used.
This mark is also found in the middle of a verse, where it means that
word-end coincides at all the repetitions. Structure of some stanza-
forms is easy to grasp when they are divided into sections. In these
cases, (1, 2, etc.) are attached to metrical scheme. Occasionally,
especially in stanza-forms of Class III, an alternative analysis is
given in parentheses with (?). This relates to the interpretation in
the Wnal chapter of Part I.
Lines where hiatus or brevis occurs are listed with H and B
respectively; lines in which both occur are listed in both places.
The Wrst repetitions of the two stanza-forms, namely str. 1 and ep.
1, are given with verse- and line-numbers (and Snells line-numbers
in parenthesis if they are diVerent from Boeckhs). The mark in the
text denotes a phrase-boundary. Every boundary is marked, except
for some next to link anceps. The mark is omitted after link anceps
(e.g. between w and e) and also before the anceps at the verse-end
(e.g. between e and xk), so as to avoid clutter.
Textual problems are discussed only when they have a bearing on
metrical analysis. Unless it involves a change of metrical scheme, an
emendation, however important it may be, is omitted. So too are
passages where at least one of the old codices reports the right
reading.1 In contrast, all emendations metri causa since Byzantine
1
I use the term paradosis loosely to mean the reading of all or some of the MSS
that Turyn gives roman capital letters as symbols. Those of the Moschopoulean and
Triclinian traditions are not included. In theory, there are chances that some of this
paradosis might not be the reading derived from the ancient world, let alone from
Pindar himself, but conjectures of Byzantine (or Italian?) scholars, especially when the
MSS were copied at a later date. To what extent Byzantine scholars were able to make
good conjectures without MSS now lost will remain a diYcult question. It is related to
the assessment of the MSS belonging to so-called contaminated traditions.
140 The Eighteen Majors
times are listed.2 I have done my best to check the ascriptions to
modern scholars since Hermann and Boeckh by consulting the
original publications, but occasionally I rely on Gerbers Emendations
(reliance is indicated). As for Byzantine scholars, my ascriptions
follow Turyns apparatus, because I have not examined any manu-
scripts (I have also consulted Mommsen and Christ). I occasionally
describe a diVerent approach, which sometimes involves rewriting
texts to a considerable extent. This may seem dispensable, but it is
still worthwhile to follow the process of how good conjectures, now
printed in Snell or Turyn, were made with the help of many others
which are dismissed (and I must say that Snells apparatus is not
always correct). Reconsideration of Hermanns and (to a lesser ex-
tent) Bergks proposals is, and will be, stimulating to serious critics.
If breaches of exact responsion occur, I discuss whether it is
normal or not verse by verse in the last part. I also collect here the
metrical parallels for each verse. Readers are asked to be tolerant
towards repetitions; description of each epinikion is self-contained
so that a reader may refer to it as a metrical commentary to an ode.

2
The list will help us also to survey quickly how scholars understood the metre in
historical perspective, although this survey needs more examinations. The history of
scholarship on metre is an interesting topic but is outside the scope of this book.
 Olympian One 141

OLYMPIAN ONE

Four triads, Class III

(1)
O1s1 w

ww
w

w
ww

k gl ph (? ^ e
dod rdod
)
O1s2 wr
w

ww
ww
ww

k ^e e D

(2)
O1s3
w
w
w
k e3
O1s4
w
ww

k ph (? rdod
)
O1s5
w
w
w
k e3
(3)
O1s6
w
wrw
w
w
w
ww

w
w
k e5wd e2
O1s7
w
w
w
w
ww

w
k e3wd e
O1s8 w rwrw
w
w
w
k we5
(4)
O1s9 wt
w
rw
k ^e e e
O1s10 w

w
w
rw
k ^ e e2e
O1s11 a
wr
w
w
k xe e2

O1e1 w
wr
w

ww
w

w
w
k we e dod e2

O1e2 w

ww

jwrjw
w
j ^ e d e3
O1e3
w
w
w
ww

w
k e2wd e
O1e4 w

w

ww
w
w
ww

k ^e e dod reiz
(? ^ e e dodwd
)
O1e5 ww
w
ww

w
w
k ^ dwd e2
O1e6a

ww
w

w
w
k tel e2 (?
dod e2)
O1e6b rw
ww

w
k rdod e (? ^ e d e)
O1e7 w

w

ww
w

ww
w

k ^ e e dod ar (? ^ e
e dod dod
)
s1 B 88, 91; s2 H 2, 42, 71; B 31; s3 H 43, 101; B 32, 72; s4 H 15, 91; B 62; s5 H 5, 34;
B 45; s6 H 35; B 46; s7 H 36; B 7, 76; s8 H 95, 106; B 19, 37; s9 H 96; B 20, 38, 78; s10
H 108; B 68, 97; s11 B 51; e1 H 52, B 23; e3 H 54, 112; e4 H 26, B 84; e5 H 85; B 27, 56,
114; e6a HB 86; e6b B 86b; e7 H 58

If s1 is analysed as suggested in parentheses (cf. Part I, 7. 4), the

alternative analyses should be adopted for s4 e4 e6a e7. See below.
142 The Eighteen Majors
s1 1 (1) @ b o, b e N
F
s2 2 (2)  

 d 
 

s3 3 (3) N   
s4 4 (4) ,  q,
s5 5 (5)  p 

s6 6 (6) 
 K  e  K 
N ,
s7 7 (7)  
 IH  P
:
s8 8 (8) 
 o I

s9 9 (9) H , E
s10 10 (10) 
E K Ia 
s11 11 (11) 
 ! "  ,
e1 23 (23) #


 B

   
e2 24 (24) K P
 F 
 I

e3 25 (25) F c K


e4 26 (26) 
, K
  F   ,
e5 27 (27) K  t 
e6a 28a (28a) q 

, 
  d H
e6b 28b (28b)
 $
b e IB 
e7 29 (29)  
 K
H F.

From the Alexandrian age onwards Olympian 1 has been the most
celebrated of all the epinikia, and deservedly stands at the head of the
collection. But one will Wnd oneself baZed at once if one tackles
the analysis of Pindaric metre by starting with O1, for the metre of
the poem is far from typical. There are a good many unusual verses
that refuse traditional analysis (or analyses). Nevertheless, since its
very Wrst verse appears at Wrst sight to be an unambiguously aeolic
phrase, the priapean dicolon (glyconic pherecratean), people are
often enticed into the search for an aeolic scheme for all the rest. But
even that is uncertain (see Part I, 7. 45). Aeolic features are not so
clearly present as might at Wrst be thought. The enigmatic phrase
`N 
A (v. 102) does not necessarily refer to the metre. It is
certainly technical but may possibly indicate the harmonia.3
We need to observe the style of O1 from a diVerent angle. To
understand the overall structure of the metre of O1, it is essential

3
For `N 
A, P2 v. 69 e   K `N E and N3.
76 ( 79 Sn.) I `N K
E PH should be taken into account
together. Gerber (Olympian One, ad loc.) considers that a reference to metre seems to
 Olympian One 143
to compare it with other poems, especially P2, which belong to
Class III. There are some signiWcant similarities between O1 and P2
(details will be extensively discussed in Part II, P2). The metre
apart, both are ranked as two of the most magniWcent epinikia.
Both celebrate victories of Hieron around the same time (O1 in 476
bc; P2 c.470 bc). In fact, these two odes are distinct even from
the other Class III stanza-forms. The maturity of Pindars metrical
technique can be observed in them (P1, another magniWcent poem,
is composed in normal D/e, but includes some unusual phrases).
Single-short movement is strikingly predominant. Such long
phrases as e2, e3, and e5 recur throughout. In contrast to the domin-
ance of single-short sequence, a long dactylic run (D
) features in
the earlier part (s2). Its length is equivalent to what is traditionally
called a dactylic tetrameter. This too is distinctive. There is no other
dactylic phrase in the ode. Even outside O1, the dactylic tetrameter is
unique in Pindar; it is unparalleled in non-D/e odes although there
are a few examples of D , notably P2s3 (see List 2).
RSS is fairly high: the Wgures for the strophe and the epode are
53.7% and 50.0% respectively.

be the likeliest explanation (i.e. likelier than dialect or musical mode) and suspects
that Pindar could have so designated virtually all the odes not composed of D/e
because it is doubtful whether Pindar would have used the general term 

as an
equivalent of the technical term . Parker, Songs, 161 suggests some special
appropriateness of the word aeolic for equestrian songs; cf. O1 v. 101

 
P2. 69 e . On the other hand Lloyd-Jones writes (Modern Interpretation
of Pindar, JHS 93 (1973), 123 GELT 132) the term Aeolic . . . will refer not to the
metre, but the music. West thinks O1, P2, and N3 are apparently in the Aeolian mode
(Music, 347) with a note (n. 87): Aeolian could mean merely Boeotian, i.e. from
Pindar, but close phraseological parallels with other passages favour the modal
interpretation. I agree with him. A particular name of harmony (mode) is mentioned
at N4. 45  f  . O14. 17 H (?) . . . K 
is certainly the
same one. The metre of N4 is not D/e, nor is that of O14. I do not know from when
glyconic or pherecratean started being called aeolic under a strict classiWcation.
Horaces usage of Aeolium carmen seems to imply it, but it may have been a rough
generalization. At least in Hephaestions classiWcation it is to the special type of
dactylic metre that the adjective `N refers, while glyconic and pherecratean
are never called aeolic. They belong to a diVerent group (antispastic). According to
West (Music, 147 n. 51), the use of the term aeolic with this denotation ( the
coexistence of single and paired short notes) is modern.
144 The Eighteen Majors

Strophe
The strophe/antistrophe can be divided into four sections: 1: s1s2;
2: s3s5; 3: s6s8; and 4: s9s11. 1 and 3 are composed of
longer verses (15.8 positions on average) and 2 and 4 of shorter
verses (8.0 positions on average, of half of the other). For the two
verses of 1, see above.
2 looks simple at Wrst sight; a pherecratean between two lekythia
(e3). But this apparent simplicity is deceptive. While outside Pindar
lekythion is very common, in the eighteen majors lekythion making
up a verse without additional phrases is found only here.4 A trickier
aspect is its juxtaposition with pherecratean starting with
w. Pher-
ecratean and e3 are equal in length. Both have seven positions. Even
their structures are similar. The Wfth and the sixth syllables being
exchanged, e3 turns into a pherecratean with base in the form of
w:

w
w
w
k

w
ww

k
One is tempted to guess that the Wrst two syllables of this pherecra-
tean are not two ancipitia but real long real short, equal to those of
e3. Cf. Part I, 7. 5.
3 consists mostly of long single-short sequences, with, in the Wrst
two verses, wd as the core. The way in which the verses relate to each
other becomes clear if they are arranged thus:
s6
w
wrw
w
w
w
ww

w
w
k
s7
w
w
w
w
ww

w
j
s8 w rwrw
w
w
w
k
The Wrst two longs of s8 are resolved to produce a sequence of seven
short syllables, the longest run of shorts in all the eighteen majors
(see Part III, D).
From s8 on, double short is totally missing. 4 starts with ^ e,
which resolved as w r has appeared already in s2 but is hereafter
repeated again and again, both in the strophe and the epode. s9 is

4
There are 9 verses made up of e3 other phrases. See Part I, 8. C. 6.
 Olympian One 145
enlarged in s10, the central e being changed into e2. The two verses
are identical in structure: even the resolution occurs at the same
position. The basic structure is maintained in the next (s11), too.
Here initial ^ e is replaced by anceps e and the Wnal e is dropped:
s9 wt
w
rw
k
s10 w

w
w
rw
k
s11 a
wr
w
w
k

Epode
The structure of the epode is clearer if the verses are arranged thus
(resolution being eliminated):
e1 w
w

w

ww
w

w
w
k
e2 w

ww

w
w
w
j
e3
w
w
w
ww

w
k
e4 w

w

ww
w
w
ww

k
(or w

w

ww
w
w
ww

k)
e5 ww
w
ww

w
w
k
e6a

ww
w

w
w
k
e6b
w
ww

w
k
e7 w

w

ww
w

ww
w

k
Every verse has a double-short phrase at its core: dodrans (
ww
w
)
in 4 verses, d (
ww
) in 3, and reversed dodrans (
w
ww
) in 1.
Before and after the core come single-short phrases. Typical ones are,
before the core, w
(e2, e4, e7) and
w
(e1, e4, e7), and, after the
core,
w
(e3, e6b)
w
w
(e1, e5, e6a) and
w
w
w
(e2).
There are some variations. In e1, the initial w
is prolonged (or
doubled). In e3, which is reminiscent of s6 and s7, w is placed between

w
w
and d. A long anceps is placed before the core in e6a too.
e5 has an unusual start. In e4 and e7 the core is repeated, once (e4)
tracing backwards (palindromic), once (e7) in the normal order.
Alternatively, e4 could be arranged as in the parenthesis. This
arrangement would be better in that it explains the short anceps of
the reizianum as equivalent to those of e3 and e5, and also the
relation between e3 and e4 would be clearer. The objection is that
its similarity with e1 and e7 would be concealed.
146 The Eighteen Majors
A diVerent analysis is applicable to e6b too. The initial long of the
reversed dodrans is resolved at all the repetitions. If www
ww
is not
analysed as rdod but as ^ e d, then the core of e6b is not rdod but d as
in e2 etc., and the similarity between e6b and e2 is more signiWcant:
e6b w

ww

w
k
e2 w

ww

w
w
w
j
Note that ^ e at the verse opening is a key phrase of this ode.
The connection between the two stanza-forms is organic. From the
strophe to the epode (s11 ! e1),
s11 a
wr
w
w
k xe e2
e1 w
wr
w

ww
w

w
w
k we e dod e2
e1 is developed from s11 by adding another e and dodrans between e
and e2. Even the position of resolution at the Wrst e is the same. For
the development from the epode to the strophe (e7 ! s1) see Part I,
7. 5:
e7 w

w

ww
w

ww
w

k ^ e e dod ar
s1 w

ww
w

w
ww

k ? ^ e dod rdod

As stated in Part I, this relation is one of the reasons why s1 may not
be gl ph. Even if s1 is analysed as gl ph, the similarity of the two
verses cannot be denied.

Textual problems
23 (e1). The paradosis # =# is unmetrical;
# Byz.
26 (e4). The paradosis KE is unmetrical;  Moschopoulos.
48 (s8). The paradosis K
is unmetrical. Boeckh adopts I
(Byz.), but Mommsens N is better. He is followed by later editors.
52 (e1). The paradosis
 is unmetrical;
 (Byz.).
57 (e6a). The paradosis is 
 , which is unmetrical. Hermanns 
 (Notae) is adopted by many editors (Boeckh, Schneidewin,
Mommsen, Christ, et al.) so that the penultimate syllable of the tele-
sillean may be short. Both Turyn and Snell follows a suggestion of
Fennell and print  , to avoid two dative pronouns (cf. 58 PH).
Corruption may lie deeper. Bowra adopts M. Schmidts x .
 Olympian One 147
64 (s6). The paradosis  Pe must be emended for metrical
reasons. The required scheme is
w. Hermann (Notae) and Boeckh
accept  (Byz.) with deletion of Pe. If this curious form
 is not a conjecture of the Byzantine scholars, it is conceivable
that, as Christ says, forma verbi ex falsa analogia verbi  Wcta
est; note that he adopts  (Byz.) at O9. 57 ( 53 Sn.), instead of
the unmetrical . Schneidewin, Bergk24 (quamvis dubitanter),
and Gildersleeve adopt  the reading of PQ (MSS eliminated by
Turyn) and the Aldine editor, accompanied by Pe. This grammat-
ically correct form is rejected by many because it introduces reso-
lution at a position which is not resolved at all the other repetitions.
It is certainly reasonable to hesitate before breaking exact responsion
by an emendation; but All-but-One may be applied. When Pe is
gone, the object of the verb is missing. Christ, following Bergk and
M. Schmidt, restores the accusative by emending x  (v. 63) into
x . But  seems to be too ungrammatical (as is , which
must be q). Thus Bowra and Turyn adopt B (Rauchenstein),
while Snell adopts   (Mommsen), keeping x  at v. 63.
Schneidewin seeks a diVerent verb from . His suggestion is
 (the parallels are cited from Aeschylus).
There is another fault in this verse. It lacks a short syllable either
before or after . hi (Byz.) has been adopted since Boeckh,
except by Mommsen and Christ () and Turyn (hi).
65 (s7). The paradosis has  in the expected second position from
the beginning of the sentence. This grammatical word-order is
unmetrical. Triclinius transposes.
71 (s11). The seclusion of the unmetrical  was Wrst proposed by
Bergk2 in the apparatus. Before that,  (Byz) was current.
80 (s11). The initial position of s11 is anceps. This analysis is based on
the recognition of the phrase e2 (xe e2). Some critics thought the
position had to be short, and it is in fact short except for
B here. Kayser (Lectiones), referring to Philostratus, Im-
agines, 1. 30, suggested KH . Bergk2 proposed an ingenious
emendation: B < , cf. Hom. Il. 9. 394 
),
which was adopted by Christ.
89 (s2). The paradosis gives L , which makes
ww
w

responding with www
w
. This is, of course, possible in dochmiacs,
but dochmiacs are alien to Pindaric metre (see Part I, 6. B); either
148 The Eighteen Majors
 (with asyndeton) or   (or ) must be accepted. The
subject of  is not here the mother, but the father (Gerber,
Olympian One, ad loc.).
104 (s6). Snell obelizes the unmetrical . Wilamowitz (GV
237 n. 1, 415) changes  into A on the assumption that in the
choriambic dimeter (on his deWnition)
ww

ww
(v. 104) can
be in responsion with
w
w
ww
(the other repetitions). This
emendation is accepted by Schroeder, but their deWnition of the
choriambic dimeter is invalid (cf. Itsumi, Choriambic Dimeter)
and so A must be rejected. It is certain that s6 is a long sequence of
single-short movement followed by wd, and, consequently, the posi-
tions which  falsely occupies must be
w. Triclinius  d is
not acceptable either; before him Moschopoulos had given  j,
but this would be semantically inappropriate (and I . . . 
. . . j is a strange combination). I am not sure whether Hermanns
Ia d (which is accepted by Bergk4, Christ, and Turyn) is grama-
tically plausible (no parallel for I . . .  . . . Ia  in
Denniston).5 Further possibilities include  d (Mommsen; fol-
lowed by Gildersleeve), K (or s) d (Maas, followed by
Bowra), z d (Von der Muhll). None of them is especially per-
suasive.
113 (e4). At the beginning, a syllable is missing. Since it must be
short (^ e), I (Blumenthal) or hI i  (Schroeder)
should be immediately rejected; either hKi or hK
i seems the inev-
itable choice. Mommsen (Annotatio, 1314) proposes a diVerent
solution by changing the metre: if v. 113 starts with  without
any preposition, the metre would be


w

ww
w
w
ww

k
e dod reiz
This is possible, and two correponding lines would be easily adjusted
to it:

 (84) and  (55) in place of

 and K:
A (26) might be the intrusion of a marginal gloss for
c
 which expelled an epithet of  in ori-
ginal, for example,
.

5
Those who adopt Ia d cite P4 v. 79 (I . . .  . . . ).
 Olympian One 149

Individual verses
s1. See Part I, 7. 4 (ambiguities), 8. A. 6 (palindrome), 8. A. 7 (contact
between gl ph). In v. 1, the rhythm w

w . . . is intensiWed by
articulation or word localizations:
` b o, w

w j w

b e N
F w

w j
ww

k
Compare and contrast v. 99:
I : w

j ww

e  Nd

 K w

j w
ww

k
But such localization is not signiWcant in the other six corresponding
verses. Against the general tendency (Part I, 8. B. 1), word-end occurs
four times between glyconic and pherecratean at four repetitions,
though not at the other four including the two above.
s2. See Part I, 6. B (^ e), 8. C. 1 (e and D within a verse). There are
Wve examples of D , but among them, O1s2 is unique in that D is
followed by anceps, and moreover, in that it is immediately preceded
by e. In the eighteen majors, even the ordinary D (
ww
ww
) does
not follow e with or without anceps (although it can precede it). In
the normal D/e, there are at least two examples of D following e,
but with link anceps between:
N1e3
w


ww
ww
ww


w


w
k
P4s6
w


w


ww
ww
ww
k
There is word-end either between e and D (3 times) or, more often,
after the Wrst long of D , a kind of dovetailing (6 times; both at v.
89, jj).
s3, s5. See Part I, 8. C. 6 (e3). For the verse-end after adverbial 
(s3 v. 14), cf. I8 v. 23b ( ).
s6. See Part I, 8. A. 3 (iii) (longer verse); 8. C. 3 (enwd), 8. C. 6 (en as
a substitution for e). Pindar favours long sequences of single-short
movement followed by wd, and this verse is the longest (6 times
w).
Another peculiarity is that the verse does not end at wd, but is
followed by e2. At the junction, two true longs are juxtaposed without
link, like the other two similar verses in O1, s7 (the next), and e3,
which have e after the sequence. But this is rare outside O1. There is
only one example:
150 The Eighteen Majors
N3s2
w
w
wrw
w
rw
k e5e
The third long of O1s6 is resolved. The resolution is never split
between words in this verse although Pindar does not in general care
about split resolutions. Word-end occurs frequently at these two
positions:

w
wr j w
w j
w
w
ww

w
w
k
There is a clear preference for absence of word-end between d and e2
(6 out of 8 verses). See Part I, 8. C. 8. and Part III, B.
s7. The preceding verse, s6, is organically shortened (see the chart
above). Reduction of length is a common development; but there is
no other example in single-short movement. (x) e3wd itself is a fairly
common length (N3s5, N3e1b). Observe too the similarity to I8s7:
O1s7
w
w
w
w
ww

w
k e3wd e
I8s7
w
wrw
w
ww
w

w
k e3tel e
Word-end is avoided at these positions throughout all eight
repetitions:

w
w
w
w
w w

w
k
The reason for the avoidance does not seem to be metrical; it is rather
the result of a similar phraseology, or verbal assonance as I should
like to call it. We can identify verbal assonance in the following two
verses, for example:
36 I
   47 P H 
Note the resemblance of sound of I=P, and
=
H at the same positions. Frequent word-end (4 out of 8) between
d and Wnal e is against the general tendency, but there are parallels;
P10s2b (5/8), N3s1 (5/8); see Part I, 8. C. 8.
s8. See Part I, 6. A (inappropiateness of the term iambic trim-
eter), 6. C. 3 (short anceps preceding resolved long), 8. C. 6 (e6 and
e5). This verse is a variation of the long single-short movement used
in s6 and s7. In these and other such verses the single-short move-
ment usually turns into wd, but there are two other exceptions (P2s1,
N3s2: Part I, 8. C. 6). The similarity between O1s8 and P2s1 is
especially noticeable. Both start with a cluster of short syllables.
 Olympian One 151
O1s8 has an additional short (O1s8 has 7 consecutive shorts, while
P2s1 has 6; in other words O1s8 starts in ascending movement while
P2s1, in descending). For continuous shorts, see Part III, D. The
seven short syllables are often chieXy occupied by a long word or
word-group: 8 
 , 19  $
e 
 , 37

K
, 66 a e 
, 77 Kb  K
d 
. Perhaps
these short syllables imitate boiling, bubbling water: 48 o  

d. Attic correption is admitted three times: 19 $
 
 , 66

, 106 K

 . And compare the diVerent treatment of
short open syllables in $
 - (19) with 109 
.
Furthermore, although the metrical location is slightly diVerent,
correspondence of 
 with 77 
 and 95 a
is interesting.
s9. See Part I, 6. B (^ e e), 8. C. 5 (double e). The phrase w


w
(^ e e) is very common at the beginning of a verse (11
examples), but there is no exactly identical verse, ^ e e e. Reso-
lution at the second position (wr
w
) occurs twice (20 
,
38 K ). The responsion between resolved and unresolved is
fairly free: it is not restricted to one irregular case nor to proper
nouns. At v. 96, I is scanned as w
.
The last e is resolved at all the repetitions. There is one parallel for
the phrase
w
rw
k at the end of a verse, namely O2s4.
s10. See Part I, 6. B (^ e e2). The preceding verse is expanded.
Instead of
w
(e) of s9,
w
w
(e2) is used at s10. Both e (s9)
and e2 (s10) are preceded by ^ e and followed by an e whose initial
long is resolved. Like I of s9,  in
 (v. 68) does not make
position. The sequence ^ e e2 is found also at
P6s9 w

w
w
x
w
k ^ e e2xe

and the sequence e2 e at

O2e3 rw

w
w
rw

w
k e e2e e
Note the resolution in the e after the e2, which occurs at the same
position as in our verse. And e2 re is structurally the same as
e re in the preceding verse.
s11. See Part I, 8. B. 6 (aeolic e2). As is shown by the chart in the
opening section, ^ e e2 in s10 develops into xe e2. Although it has
no exact parallels, the structure of this verse is well explained as a
152 The Eighteen Majors
variation of xe e, which is a very common sequence. The initial
anceps is basically short. Long anceps is used only at v. 80
(B ), under the All-but-One rule (Part III, C). Bridge is
observed between xe and e2 at all the repetitions; cf. Part I, 8. C. 6
e1. See Part I, 8. B. 6 (aeolic e2), 8. C. 5 (double e). The
juxtaposition of dodrans (or reverse dodrans) and e2 in a long
verse is a mark of Class III. Note the general similarity between
O1e1 and N7s2:
N7s2
ww
wr w
w

ww
w

w
w
k
O1e1 w
wr
w

ww
w

w
w
k
The other phrase-combinations phrases of O1e1 are also common.
xe e is used at the beginning of ten verses, of which O2s5 and O2e5
are resolved at the same position; e dod is found in six verses. Of
these, three are in O1e; the others are O1s2, P8s4, N7s2.
e2. See Part I, 5. A. 4 (analysis not as
ww
but as ^ e d); 8.
B. 4 (d e3); 8. C. 6 (e3). There are no exact parallels, but cf. P2e6,
P5e5, P10s2b (^ e d e), and N7s5 (wd e3). All four repetitions
follow the same pattern of word-end:
w

ww

jwrjw
w
j.
Turyn divides the verse into two, but his two verses (pherecratean
and xe2 respectively) do not harmonize with the rest of this stanza-
form. More probably O1e2 is an indivisible unit, as demonstrated in
the chart above.
e3. See Part I, 8. C. 3 (enwd). The structure of e3 is identical with s7,
but the length of the single-short movement is shorter by one
w.
They are diVerent in the preference for bridge. Word-end does not
occur before the Wnal e in e3, while in s7 it is found at Wve repetitions
out of eight.
e4. See Part I, 8. A. 6. (b) (palindrome), 8. B. 8 (^ e e aeolic).
e5. See Part I, 6. B (^ d), 8. B. 4 (d e2), 8. C. 2 (double d), 8. C. 3
(wd). The Wrst half seems strange, but is paralleled by a verse in the
four minors:
O4s1 ww
w
ww

ww
ww
j
In the normal D/e odes,
 Olympian One 153
P9s3 ww
w
ww
ww

k ^d w D

(accepting 
 instead of 
 at v. 37 as Snell does) is the
closest parallel. And, although the link anceps is always realized as
long,
P1e8 ww


ww
jrw

w
j ^d
d ee
is also similar.
e6a. Boeckh combines this verse with the next and makes one long
verse. He eliminates hiatus and brevis at the end of v. 86 by following
recentiores (K
 t j 
.). This reading is actually a Byzantine
emendation. The two halves, e6a and e6b, are metrically easily
explained. There is another verse (P11e3) which is made up of
telesillean e2, and a similar combination, dodrans e2, is used at
e1 above. Mommsen was the Wrst to divide e6, and is followed by
later editors.
e6b. See Part I, 5. E. 2 (resolution of reversed dodrans). N7e2 is the
identical verse. Including this, there are in total nine examples of
rwrdod. As is discussed above, the resolved position can be diVer-

ently analysed: wr
ww
(^ e d). ^ e d is found at e2 above.

e7. See Part I, 8. A. 7 (repetition within a verse), 8. B. 8 (^ e e
aeolic). As illustrated above, this verse is similar to e4 and s1. The
aristophanean is a surprisingly rare phrase in the eighteen majors:
besides O1e7, there is only one example, P11e1 which consists
entirely of an aristophanean. By deWnition, aristophanean is dod 1
in this book; see Part I, 5. B. For the repetition of dodrans with
prolongation by a long (or a long anceps), cf.
N3e2
w
ww

w
ww

k
in which reverse dodrans seems to be repeated with an added Wnal
position.
154 The Eighteen Majors

OLYMPIAN TWO

Five triads. Class II

(1)
O2s1 w
w

w

k we e

O2s2 w y
w

wr
wr
w
k ^e e e e e
O2s3

w

wr
wy

w
tw
k
e e e
e e
(2)
O2s4
w
rw
k ee
O2s5

wr
w
rw
tw
k
e e e e
O2s6/7 w

wr
wt
w
w

w

k ^e e e e 2e

(coda)
O2s8

w
rw
w
w
ww
k
e e 2wd
O2e1

w

wy
w
k
e e e
O2e2
w
w
wr
w

w

k e 3e e

O2e3 rw

w
w
rw

jw
k e e 2e e
O2e4

w

wt
wr

j
e e e sp
O2e5 u
wr
w

w

j xe e e

O2e6 w
w
w

k we 2

s1 H 9, 23, 67; B 75, 97; s2 B 2, 10, 32, 54, 68, 90, 98 (!! see below), s3 H 69, 91, 99, s4 H
34, B 78, s5 H 35, 57,6 s6/7 H 37, 51, 73, 81; B 7, 103, s8 B 38, 52, 74, e1 B 61, 83, 75, e2
H 40; B 62, e3 H 41, e4 H 64, 86; B 20, e6 H 66, B 22, 44.

s1 1 (1)  o,

s2 2 (2)  ,  l,    ;
s3 3 (3) X  b  

   ! 
s4 4 (4) I

s5 5 (5)  b A;   
s6/7 6/7 (6) , Z

 ,  
 ,
s8 8 (7) P 
  O

e1 17 (15) 
H : H b


e2 18 (16) K   d
a  I
 P 
e3 19 (17) %


c    
e4 20 (18)
b
 f P  .

6
If the emendation
 A is accepted (see below), hiatus does not
occur at the end of v. 57.
 Olympian Two 155
e5 21 (19) KH a $
e 

B 

e6 22 (20)
 ,
str. 2 2330 (21)(27) ant. 2 318 (28)(34) ep. 2 3944 (35)(40)
str. 3 4552 (41)(47) ant. 3 5360 (48)(54) ep. 3 616 (55)(60)
str. 4 6774 (61)(67) ant. 4 7582 (68)(74) ep. 4 838 (75)(80)
str. 5 8996 (81)(87) ant. 5 97104 (88)(94) ep. 5 10510 (95)(100)

The rhythm of Olympian 2 is essentially cretic, with sporadic

ancipitia, usually at the beginning or end of verses, but occasionally
within the verse (Parker, Trochee to Iamb, Iamb to Trochee).
Besides these, acephalous cretic (^ e) and prolonged cretic (e 2)
are admitted. The metre of O2 is indeed exceptionally homogeneous.
Among its 13 verses double short is found only once, in s8. Never-
theless this ode can be considered in perspective to be an extreme
case of freer D/e composition in which e is intensively and almost
exclusively employed. O2 has aYnities with others of the eighteen
majors: notably, there are in total 33 verses in the other stanza-forms
(Part I, 8. C. 4) that, like like those in O2, have no double short at all.
Some similar verses, and sometimes even identical ones, are found
here and there. Even s8, which has a double short at the end and
seems curious within the context of this ode, can be paralleled. It is
associated with a group of verses which are usually found in Class III
stanza-forms and which are made up of single-short movement
turning to d (Part I, 8. C. 3).
There are a number of phrases made up of three short syllables in
sequence. Two of the three must be a resolved long, but it is hard, and
occasionally impossible, to decide which two are resolved: either

wr
or
rw
. Besides, one or two of the three longs of e 2
(
w
w
) can be resolved. Consistent realization of anceps posi-
tions, either long or short, throughout all the repetitions makes
the decision more confusing. See, for example, the notes below on
e2, e3, and e4.
However, a strong tendency in the matter of resolution is observ-
able (cf. Part I, 6. C). First, when a long is resolved, the resolution is
usually repeated at the same position in every repetition. Thus, out of
39 es (including 2 examples of ^ e, and counting e 3 as 2 es) and four
e 2s, 12 es and one e 2 are always resolved at the same position
throughout all the repetitions (7
wr, 5 rw
, and 1
rw
w
). On the other hand, 20 es and three e 2s are totally
156 The Eighteen Majors
unresolved. In total, 36 phrases (32 es and 4 e 2s) out of 43, whether
resolved or unresolved, are consistently repeated 10 times in the
strophe/antistrophe and Wve times in the epode in the identical
form. Secondly, even in the remaining seven cases where resolved
and unresolved are found in responsion, preference for one or other
is manifest. The poet does not seem to have a completely free choice
of resolution at each repetition. The following chart shows the
relative frequencies of regular and irregular forms in each passage.
Thus s2 begins nine times with wr, only once with w
:
regular/irregular totals line Sn. irregular form
s2 wr/w
9/1 68 62 Y
s3 (3rd)
wr/
w
9/1 11 10 
s3 (5th) rw
/
w
8/2 3 ! j
25 23
j 
s5 (4th)
w
/rw
9/1 57 52

s6/7
w
/
wr 8/2 28 26 j #j
102 93 j  Kj
e1
wr/
w
3/2 17 15 H b

j
39 35 j 
j
e4
w
/
wr 4/1 108 98 j
 Ije
Emendation changes the statistics. If, for example, we read A
instead of IA at v. 57 (see below on textual problems),
w

emerges against the majority (rw
):
s5 (3rd) rw
/
w
9/1 57 52 A
Then the number of totally unresolved cretics will be 19 instead of 20,
while those in responsion with resolved will be eight instead of seven.
But, if we accept A, it is better to change the word-order as
well, giving
 A, so that the position of the resolved
long at third and fourth e is perfectly consistent with all the other
repetitions (in that case totally unresolved cretics will number 21).
There is also another possibility of an emendation: s3 (3rd), v. 11
(10 Sn.),  , hi. This increases the number of totally re-
solved cretics. So does s2, v. 68 ( 62 Sn.), Y  K  . For
these, see further below, on textual problems.
Resolution does not occur simultaneously at both longs of a cretic:
i.e. there is no example of rwr (for the general summary, see Part I,
 Olympian Two 157
6. C). Nor is there any example of in the eighteen majors
wr
rw
, nor more than one of rw

wr (I8s8). Between
wr
and rw
, an unresolved cretic, or an anceps at least, is always
inserted. Frequent resolutions raise RSS: 53.0% (O2s) and 47.3%
(O2e). Supposing that the whole ode were composed only of unre-
solved cretics and no ancipitia were included, RSS would be 33.3%,
because two-thirds of the syllables would be long.
Eight verses start with anceps. Realization of the anceps is rigidly
regularized as either long or short; Wve of these ancipitia (s3, s5, s8, e1,
e4) are consistently long, two are short (s1, e6). For a in v. 71 ( 65
Sn.), s5, see below. Only at e5 are long and short syllables in responsion
(2 longs vs. 3 shorts). Some editors unnecessarily regularize it at all the
repetitions; see the textual discussion on v. 107. With Wnal anceps six
verses end pendent. Pendent ending is rather uncommon in the strophe
(2 out of 6). Intermediate anceps between 2 es is very rare. I count only
one in the whole O2 (or 2, if the middle short of e 3 at e2 is included):
O2s3

w

wr
wt

w
t w
k
e e e
e e
For the possibility of another mid-anceps in e3, see below.
Bridge is observed at some contact points of 2 es. It is generally
recognized that Pindar observes it in xe e at the beginning of the
verse and in e e at the end; see Part I, 8. C. 5. But there is no rule
that can be applied to all contact points. Let us examine the strophe,
in which there are 15 contact points. There is only one contact point
where bridge is observed throughout all 10 repetitions. At another
four it is observed in nine. By contrast, there are points where it is
ignored in more than half the repetitions: between the third and the
fourth e in s3, in as many as seven repetitions, and between the
second and the third e, in six repetitions. There is no relation between
the presence of bridge and the resolution of e.

Strophe
Hiatus/brevis unambiguously indicates every verse-end.7 The length
of verses varies widely. But the structure of the stanza is articulated
7
Merkelbach, ZPE 12 (1973), 4555 grasps the essence of the metre of O2. Instead
of the traditional concepts of the (iambic, trochaic, etc.) metron, he introduces the
idea of e, e2, e3 (though his notations are diVerent from mine) and, above all, link
158 The Eighteen Majors
into two halves (1, 2), s8 being taken as a coda. Both halves are of
the same construction articulated by the contrast between a shorter
verse and two longer ones. Each starts with a shorter verse (s1 and
s4), which is made up only of two es, and two longer verses follow, of
which in each case the Wrst is shorter than the second. The numbers
of positions are 81417 in 1; 61317 in 2. Each of the latter
longer verses (s3 and s6/7) includes an irregularity: there is a mid-
verse anceps in s3, the only one in this ode, and a prolonged e (e 2) in
s6/7. The almost complete absence of mid-verse anceps and e 2 makes
the rhythm easy to catchperhaps even a triXe monotonous, in
contrast with the epode.
There are also devices that make the structure of s2s6/7 symmet-
rical, or circular. At the introduction of these verses, extra positions
are located that decrease in number from two (w
) through one
(
) to none, and then increase in reverse:
^e (s2) :
(s3) : none (s4) :
(s5) : ^ e (s6/7).
As for the cadence, the outer two verses (s1, s6/7) end pendent, the
inner four blunt. The last eight positions of s6/7 are exactly the same
as s1:
O2s6/7 w

wr
w t
w
w

w

k
O2s1 w
w

w

k
Thus it is possible, and may be sensible, to describe the Wrst phrase
of s1 as ^ e 2, instead of we as indeed Dale (Collected Papers, 78)
analyses it.
The last verse, s8, is the coda of the stanza-form. Though it starts
in the usual manner with anceps e (

w
) like s3, the move-
ment changes and surprisingly ends with a double short.

anceps. But it is a pity that he unites two verses without regard to the explicit verse-
end; for example, s1 and s2 are, as a whole, described as w 7 e (according to his own
notation, w 7 p). Certainly the Wnal long of s1 and the initial w
of s2 could make
another e, and this fact might illuminate the structure of these verses. But it should
not be ignored that there is verse-end between them. To ignore verse-end means
suicide for modern metrical scholarship.
 Olympian Two 159

Epode
In contrast with the strophe, the colometry of the epode is less
deWnitely settled. In three verses (e2, e3, e4) diVerent analyses are
possible, depending on which positions are taken to be resolved, and
on the presence and location of mid-verse anceps and e 2 (see below).
Pindar departs from the monotony of the strophe and deliberately
makes the rhythm hard to grasp. As demonstrated above, he tends to
regularize both the resolution of e and the realization of anceps, but
this tendency becomes less rigid in some verses of the epode; see e1
(resolution) and e5 (anceps). The Wnal two verses (e5 and e6) should
perhaps be combined into one:
e5 e6 u
wr
w

w

jw
w
w

k xe e e e 3

This is certainly long, but not exceptionally so. The number of
positions of this united verse amounts to eighteenone more than
the longest verses in the strophe, s3 and s6. The last phrase, e 3 (
ewe), is employed in e2 too. On the other hand, e 3 never follows e in
the eighteen majors.8 Moreover, if we separate e6, the transition from
the epode to the strophe is smoother:
O2e6 w
w
w

k we 2

O2s1 w
w

w

k we e

Hence, although I am strongly tempted by the united verse, I follow
the editions and divide e5 and e6.

Textual problems
6 (s6/7). By changing  of the paradosis into , Hermann
(Notae) restores perfect sense. The long
in Z

 should be
explained by a Homeric analogy, for example, 
(  ):
Chantraine, Gram. hom. i. 217.
11 (10 Sn.; s3). Hermann (Notae) proposes  , hi,
which produces the same form as the other repetitions. But he may
not have emended the text simply to tidy up the responsion. Neither
in the strophe nor in the epode is there any sequence of three long

8
B17s6 (xe e e3) is very similar; see excursus below.
160 The Eighteen Majors
syllables within a verse.  ,
F  is the sole exception. Cf.
Maas, Freiheiten II, 7. Similarly, Bergk4 writes K
Z.
34 ( 31 Sn.; s4) The paradosis
 is unmetrical. Most critics
accept
E (Moschopoulos), which is a term unattested elsewhere
(cf. LSJ
E). Mommsen changes word order: H 
 Pj
  
. It is preferable not to set verse-end after P, but it
may be accepted.
42 ( 38 Sn.; e4). The paradosis
 is unmetrical; A Her-
mann.  is unmetrical, too.  is recorded in some of
the scholia, and Byz.
46 ( 42 Sn.; s2). The paradosis
  is unmetrical; 
 
Triclinius. This emendation is plausible, but
 might still be right.
The corresponding part is www in all the other repetitions except 68 (
62 Sn.): Y (w
). But this line is corrupt and some critics try to expel
Y (see below). If www were introduced instead of Y in 68 and

 were retained in 46, the metre of s2 might be
yw
w

wr
wr
w
k e2e e e
Schroeders
 (=) (w
) cannot be accepted since digamma
does not in Pindar lengthen a preceding short syllable ending in
a consonant: Braswell, Pythian Four 184d.
57 ( 52 Sn.; s5). The paradosis  is unmetrical. I follow
Bowra (CQ 1936), Slater, et al. in reading I

(IA Mommsen, Turyn). The choice between I
(IA) and A is hard. I is attested in a papyrus
(P.Oxy. 2092). A (W. Dindorf; anxiety cf. H
Moschopoulos according to Willcock, Triclinius according to Christ),
is certainly germane in this sentence, but IA (foolishness
to remain as a fool equal to his fellow citizens) seems me to be
more likely as a Pindaric gnome.9 Metrically both are possible. As

9
Bowra, The Metre of Pindar cites P8. 734 for the illustration of Pindars
notion that to fail in the games . . . implies stupidity: N
 Ka

 c
f H
,
E e E
 I . . . (. . . for if someone has
gained success without long labour, he seems to many to be a wise man among
fools; tr. Race). But I am not sure that this sentence at once supports
I
 . . . (winning) releases from stupidity. Or, though this meaning
is not attested, does I perhaps mean deprived of P i.e. the status of
being deprived of celebration after the victory? At N7 v. 3, P (night) is
wrongly written as P in D.
 Olympian Two 161
demonstrated above, and also taking into account of the All-but-One
rule (Part III, C), one occurrence of an irregular form against the
others is not unusual. Thus it should not be regularized simply
because of its rarity. It is, in a sense, reasonable that Schroeder
(A
), despite accepting Wilhelm Dindorf s emend-
ation, keeps the word-order of the paradosis ( of the
paradosis is neither gramatical nor metrically possible). Schroeder is
followed by Snell. Nevertheless, if we prefer A to IA,
then
 A is better. This slight change removes two
successive irregular responsions. Pindar seems to have accepted
irregularities but, at the same time, to have been willing to remove
them as easily as possible.
68 ( 62 Sn.; s2). The paradosis Y  K  is unmetrical.
Both Turyn and Snell follow Mommsen, and delete K without the
change of the preceding Y b . The responsion between
Y w
here and wr in the other repetitions is not strange; see above,
and in a wider perspective, Part III, C (All-but-One). Other proposals
are: Y b : . . . Y  K . (Triclinius, Heyne); Y K b
: . . . Y  K . (Schwickert, Christ); Y b : . . . Y K
. (Boeckh), Y b : . . . Y K . (Bergk4), alii alia. Note
that these all seek exact responsion. The emendations depend on
our interpretation of the situation of the sun in Hades. Boeckh thinks
the sun always shines during both night and day. Bergk supposes
perpetual equinox (i.e. spring), and Mommsen interpretes Y as
d &E E H (when the sun rises in the upper world, it sinks in
the underworld).
71 ( 65 Sn.; s5). Most scholars, including Snell, adopt a
(Moschopoulos).  is attested at O3. 48, where a long syllable
() is necessary. Here in O2 v. 71, the position is anceps. The
paradosis is a. Mommsen keeps it, by scanning  as one long
syllable. He is followed by Turyn. a is preferable even metrically,
since anceps tends to be regularized in this poem.
845 ( 767 Sn.; e23). The word-order in the paradosis is in
complete confusion if the subject is taken to be Kronos, as the
majority of scholars since Didymus (ap. Drachmann i. 95) have
supposed. The paradosis is:
84 n
c  A E PH


85


 ! ' o
 K
E .
162 The Eighteen Majors
Remarkably, the metre is not so bad as some think, though a diVerent
articulation becomes necessary (for this analysis of these two verses,
see below):
O2e2
w
w
yw
w

w

k e 2e 2e

O2e3 rw

w
wtwy

w
k e e 3
e
The confusion arises conspicuously from two words: A and
E ;
hence Byz. introduced (1)  in place of A , (2) deleted
E ,
and supplemented the missing syllables by prolonging o
 into
$
. Boeckh and Schneidewin follow this path. Pauws emend-
ation  for A was restored by e.g. Christ and Schroeder, which
is now conWrmed by a papyrus (Snells 2 P. Oxy. 2092) and
adopted by both Turyn and Snell. It is a pity, however, that the next
verse (
 . . . ), which in the papyrus would have occupied
two lines, has dropped out. Whether the omission was noticed or not
cannot be determined, the margin on both sides being defective
(Hunt, The Oxyrhynchus Papyri, XVII 129). Reading (2) of Byz.,
deletion of
E and emendation to $
, is adopted by most
critics, including Turyn and Snell. Two questions arise: (i) why did

E wrongly intrude? (ii) were Byz. capable of emending the text so
brilliantly? Perhaps they had access to a good manuscript which is
now lost. If the subject is not Kronos but Zeus, as Aristarchus seems
to have interpreted, and moreover, the etymological meaning
(ruler) of
 , governing the genitive

, could have survived
till Pindars time, the paradosis in 85 might be acceptable.10 Then the
metre cannot be rejected at once but introduces two irregularities; see
below.
107 ( 97 Sn.; e3). All the manuscipts report   ,
which is one syllable too long. Editors emend  into ,
following Aristarchus, although the whole sentence has been vari-
ously construed with some emendations.
109 ( 99 Sn., e5). Snells report is careless. The paradosis is IE ;
d E is Mommsens emendation. The older editions before
Mommsen adopted KE . However, connective d is necessary for
10

E

 ! ' $
 K  (M. H. Rocha-Pereira accord-
ing to Gerbers Emendations) would be easiest if Zeus were the subject.
E
introduces
w where www occurs at the other repetitions. This is, of course, possible.
 Olympian Two 163
the sentence d . . .  to be included into the K
d clause. Metrically,
the long syllable does not pose a problem. The initial position of e5 is
unambiguously anceps. Snell marks K H (21 19 Sn.) to indicate that

(
its initial is short. But this is unnecessary.

Individual verses
s1. See Part I, 8. C, (5) (double e). The verse is short, its structure
simple; but the exact parallel (we e x) is not found elsewhere. The
closest verses are:
P5s1 w
w

w
k we e
P5s9 w
w

w

w
k we e e
Including these two and O2s1, there are in total 15 verses that start
with short anceps and e (w
w
).11 Six of them are found in P5. As
demonstrated above, we here may be ^ e 2. It is impossible to select
from the other 14 examples those that may be acephalous e 2.
Interestingly, when e is followed by anceps at the verse-end, the e is
resolved outside O2; see Part I, 6. C. (viii). There are only two
examples of unresolved (
w

k)
I8s5c
a
ww
wt
w

k gl e

N3e1
ww
w
ww
x
w

j dwdxe

s2. See 8. A. 6 (i) (palindrome). This verse starts with an aceph-
alous e which is followed by a full e. This combination (w

w
) is
fairly common (11 in total); see 6. B. The full e is repeated four times
without intervening link anceps, unparalleled except O2s5. The sec-
ond e and the third are resolved at the same position (
wr). They
create a palindromic sequence
www
www
. There are two other
parallels.
Brevis in longo occurs frequently, at seven verses out of 10. Inter-
estingly, all seven end with a short vowel  , ,
, I, I
,
, 
 (an eighth would be
added if I (v. 46 v. 42 Sn.) were not followed by a consonant).
Perhaps this fact is another example of verbal assonance.

11
O1e1, O2s1, P5s1, P5s2, P5s3, P5s9, P5e4, P5e7b, P6s1/2, P6s7/8, P8s4, N3s4,
N7s4, N7e1.
164 The Eighteen Majors
s3. See Part I, 6. C. 3 (er
e), 6. D (long mid-anceps). Five seems
to be the maximum for full e in one verse. There is no such verse in
the eighteen majors. Between the third and the fourth e, a link anceps
occurs which is realized as long at all the repetitions. Long mid-
anceps is rare not only in O2 (cf. s2 and s5, in which four es are
juxtaposed) but in Pindar. The position before the anceps is resolved
at nine repetitions out of 10 (except v. 11 v. 10 Sn.; see above). The
sequence er
e (
wr

w
) occurs also in
O10s4 w

wr

w
k ^e exe
Bridge is observed between long anceps and the following e in most
of the repetitions. There is only one exception: v. 77 ( 70 Sn.)
j  (see parallels in 6, D). Note the words of the shape
ww
at the verse-end: 33 , 77 
, 99 K
 . This is
naturally common in s8, which ends with d, but is supposed to be
rare in the verses made up of single-short movement.
Verbal assonance is observed: 33 "E 
 55 "E 
  69 Kd .
s4. See Part I, 8. A. 3 (shorter verse). Outside this ode there is a
single verse (P5e8) made up only of two es. Five repetitions end with
words of the shape ww
: 4
, 34 
, 48
 (again),
92  , 100  . Verbal assonance may occur between
these, especially 92  Kd   100 
  .
s5. This verse is one position shorter than s2. Without its initial
short, s2 would be structurally equal to s5. However, their diV-
erent resolved positions give them such a completely diVerent
impression:
O2s2 wt
w

wr
wr
w
k ^e eeee
O2s5

wr
w
rw
tw
k
e e e e
Verbal assonance: 5    49   57
.
s6/7. See 8. C. 6 (en as a substitution for e). This verse starts with
acephalous e (^ e) like s2, and ends with anceps like s1. In the middle
3 es and 1 prolonged e are connected without mid-anceps. s6/7 as a
whole structurally resembles the preceding s5 (and s2 too):
O2s5

wr
w
rw
tw
k
O2s6/7 w

wr
wt
w
w

w

k
 Olympian Two 165
At nine out of ten repetitions, word-end regularly occurs unam-
biguously in the following manner:
w

w r
w t
w
jw

w

k
The exceptional case is v. 29 ( v. 26 Sn.): E j  . Boeckh
occasionally neglects postpositives at the beginning of a verse. Our
verse is one of the examples. He wrongly divides this into two.
Perhaps he may have wished to expel an unfamiliar phrase,

w
w
(e 2). In fact, it is fairly common for this to be adjacent
to e. The combination e e 2 is incorporated, outside O2, in
O1s11 a
wr
w
w
k xe e 2
and e 2 e in
O1s10 w

w
w
rw
k ^e e 2e
For the ending . . .
w

k, see s1 above.
s8. See Part I, 8. A. 6 (h) (palindrome); 8. C. 3 (enwd). A double
short exceptionally occurs at the end. But this verse is not extraor-
dinary in the eighteen majors; it is associated with a group of verses,
usually found in Class III stanzas, which are made of up of single-
short movement ending with d. These sequences vary in length. That
found here (e 2wd) is attested in:
O1e3
w
w
w
ww

w
k e 2wd e
P11s3 rw
wr w
ww
k e 2 wd
O2s8

w
rw
w
w
ww
k
e e 2wd
O2s8 is diVerent from the others in that it has a short phrase (anceps
e) which precedes the sequence. At the same time this phrase and
the initial part of the following e 2 are partially equal to some of the
preceding verses:

w
(s3),
w
rw
(s4, s5),
wr
w

(s5). Resolution of the Wrst long of e 2 is also found in P11s3, and in
N7s7
yww
w
rw
w
k tel e 2
Other colometries are theoretically possible:


w
wr
w
w
ww
k
e 2e w d


w
rw
w
w
ww
k
e e hepta
166 The Eighteen Majors
However, these analyses fail to bring out the similarity with the
sequences mentioned above.
Verbal assonance: 8 P 
  52 K 
,  38 PA   104 P  16 h .
e1. The structure of long anceps 3 es is identical with the Wrst
half of s3. But words are localized in a completely diVerent manner.
In s3, the following cuts are frequent (612 and 7 out of 10 repetitions
respectively):


w

jwrj
wy . . .
while in e1 word-division is absent from all Wve repetitions at both
boundaries.
This verse includes a rare case where a resolved long and an
unresolved are in responsion rather freely (resolution at 3 repetitions
against 2). Interestingly, two repetitions which lack resolution have
words that could be scanned as if resolution might have been con-
sidered for these verses too till some relatively late stage of the
composition: 17 b

 39 
. These words
must be scanned in this context w

w
but could be ww
w
,
and if another w were supplemented, the total verse would be equal to
the other three. Conversely, if w were taken away from the other three
(61 - , K 83 - ! '
 105 - K
  ), the
verse would be


w

ww
w

at all the repetitions. This is exactly what the Byzantines did; their
text has: 61 K (instead of K), 83 OE (instead
of OE), 105  (instead of K
). Boeckh, Dissen, and Schnei-
dewin followed them. Bergk2 is the Wrst major edition to adopt the
currently favoured metre (but he actually follows Mommsens sug-
gestion: creticum numerum restituendum esse vidit Mommsen).
e2. See Part I, 8. A. 6 (h) (palindrome), 8. C. 6. (e 3). This verse
includes e 3 uniquely in O2 (in other words, it uniquely has a short
link anceps between 2 es throughout all the repetitions). The alter-
native analysis

12
7, if O ; Ng j  (11) is included.
 Olympian Two 167

w
w
rw
w

w

k e 2 e 2 e

is less likely since there is no verse elsewhere in the eighteen majors
that includes two juxtaposed e 2s. The reading of the paradosis
requires this analysis (see above), but the emendation is indispens-
able; A should be changed into  .
e3. This verse too could theoretically be analysed in a diVerent
way:
rw

w
w
wr

w
k e e 3
e
It may be better, however, to avoid e 3 in the middle of the verse. e 3
always occurs either at the beginning or at the end of a verse without
exception. The long mid-anceps is rare, but not a problem (cf. s3
above). Note that bridge is observed between it and the following
long at all the Wve repetitions. If the paradosis were accepted (see
above), this verse would be
rw

w
wtwy

w
k e e 3
e
The responsion between resolved and unresolved at two diVerent
positions of e 3 is highly irregular, since there is no example of
responsion in the eighteen majors between
wr and rw
at e.
Moreover, the paradosis introduces cut (K j) after the long
anceps.
Verbal assonance: 19 %

  63    85



.
e4. See Part I, 6. D (long mid-anceps). This verse is very similar to
O10s3a
u
w

wr

wr

j xe e
e sp
and the colometry of O2e4 should be taken into consideration.
O10s3a should not be analysed as
u
w

rw

rw

j xe
e
e

because (i) this analysis entails too many cuts after the second long
anceps (see Part I, 6. D, and Part II ad loc.) (ii) it also entails long
ancipitia preceding resolution (see Part 1, 6. C. 3). In fact, a long
mid-anceps is, in general, very rare. As for O2e4, although it does not
entail the cut after long mid-anceps, the colometry
168 The Eighteen Majors


w

wt
wr

k
e e e sp
is preferable to


w

wt
rw

k
e e
e

The latter analysis would introduce a long anceps followed by reso-
lution. This is totally unparalleled. The analysis apart, it is important
to notice that the ending wr

k is similar to
w

k, which is
repeated in O2 (s1, s6/7, e2, e5).
e5. There is no manifest indication of verse-end. When e5 is united
with e6, a long verse will result; see Epode above.
e6. There is no other verse made up of xe 2
, but two verses
consist of xe 2 (O10e7, I7s5b).
Verbal assonance: 22
   44

.

Excursus A: Bacchylides 17
The metre of Bacchylides fr. 17 ( ) is similar to that of O2. The
similarity is more evident, and easier to grasp, when both are analysed
according to the same rules. The text of  still remains uncertain
even after the challenge of R. Fuhrer. There are a great many disagree-
ments between scholars about the extension of Responsionsfreiheiten. If
one could say that Maass position is to the extreme left of the spectrum,
then that of Fuhrer would be to the extreme right (there is no political
implication; Fuhrers arguments cannot be said to be conservative in
the usual sense: he is as radical as Maas) and Snell would be located
somewhere in the middle. Fuhrers text is very sensitive to what I would
like to call verbal assonance (see Part III, E) between the correspond-
ing repetitions as they stand on the papyrus. Consequently, he is very
faithful to the word-order on the papyrus and does not hesitate to
widen the limitation of acceptance of irregular responsions. This is not
the place to argue the text in detail. I simply propose the analysis of the
Wrst strophe using my method of notation, to show how the metre
of B17 resembles that of O2. The verses which include similar phrases
are given in the right column. In the Wrst strophe of B17 there is
no signiWcant disagreement on the text. The metrical scheme is
 Excursus A: Bacchylides 17 169
unambiguously Wxed. What intrigues us is the responsion between
this strophe (str. 1) and the other three repetitions, namely, ant. 1, str.
2, ant. 3.
B17s1 ^e e e 2e e 2k O2s6/7 (^ e e e e 2e
)
B17s2 we e 2 j O2s1 (we e
), O2e6 (we 2
)
B17s3 e e
k O2s1, O2e5 (xe e e
)
B17s4
e
e e e ej O2s3 (
e e e
e e)
B17s5 we e spk O2e4 (
e e e sp)
B17s6 xe e e 3j O2e5 O2e6 (xe e e e 3
)
B17s7 dwewDwe 2j O2s8 (
e e 2wd )
B17s8 we e 3e 3e ek
B17s9 e e e 3k O2e3 (e e 2e e)
B17s10
e 4e e
k
B17s11 xe e 2k O2e5 (xe e e
)
B17s12 e e 3ej
B17s13 e e 2ek O2e3 (e e 2e e)
The following characteristics are observable:
1. The basic metre is the same as O2: e; also its prolonged phrases
(e 2, e 3, e 4) are used together here and there.
2. Link anceps is rare. e and other phrases are directly connected.
3. d and D are used in only one verse (B17s7). Thus the metre of
this ode too can be described a special case of freer D/e.
4. There are a considerable number of similar verses shared be-
tween B17s and O2. But verses are, in general, longer, and short
anceps (incorporated in e 3 or e 4 ) is more frequent in B17s than O2.
170 The Eighteen Majors

O LYM P I A N N I NE

Four triads. Class I (strophe/antistrophe); Class II (epode)

(1)
O9s1 ww
ww
w
k tel
O9s2 u
ww
w

u
ww
w
w

k tel gl 3
(2)
O9s3
a
ww
w
u
ww

k gl reiz
O9s4


ww
w


ww

k gl reiz
O9s5
a
ww
w


ww

k gl reiz
O9s6/7 www
a
ww


ww

k wil reiz
O9s8
u
ww
w
k gl
O9s9

ww

k reiz
(3)
O9s10 w
ww
w

w
k tel e
O9s11

ww

w
a
ww

k
d exd
(?
d e
reiz)
O9e1/2 w
w
w
w
jww
w
w
w

k we 2wdwe 2

O9e3 ww
ww

k ^ D

O9e4



ww
w


k sp tel sp
O9e5
ww




j d sp sp
O9e6

ww
ww
jww
k
D
O9e7

w
ww
ww


j
ribyc
( heptad) sp
O9e8

w
u
w
ww
u
w
ww
w

k
e hepta hepta 3
s1 H 72; s2 H 13, 73; B 32; s3 B 93; s4 B 15; s5 B 46, 65; s6/7 H 7; s8 H 19, 98; B 19, 68,
79, 109; s9 H 110; B 20, 39, 50, 69; s10 H 70, 100; B 51, 81, 111; s11 H 112; B 22; e1/2 B
22; e3 B 115; e4 H 26; B 116; e6 H 28, 58; B 88, 118; e8 B 30

For the analysis and the notation of e7 (
ribyc ), see Part I, 7. 6 and
Appendix B.
s1 1 (1) (e b  
s2 2 (2) A 
 ,  
 
s3 3 (3)  
Z F
s4 4 (4) 
   f 
s5 5 (5) Ia F  A I
e 
s6/7 6/7 (6)   
   K

 Olympian Nine 171
s8 8 (7) I @
s9 9 (8) E ,
s10 10 (9) e 
 e l 
s11 11 (10) K

  ! "


e1/2 23/4 (21) Kg   
 E
(22) K
 IE ,
e3 25 (23)  I



e4 26 (24) A d e $


 A
e5 27 (25) I
 ,
e6 28 (26) N   


e7 29 (27) K %  A
:
e8 30 (28) E a
 a 
 Id b d d
a  
str. 2 3141 (29)(38) ant. 2 4252 (39)(48) ep. 2 53/460 (49)(56)
str. 3 6171 (57)(66) ant. 3 7282 (67)(76) ep. 3 83/490 (77)(84)
str. 4 91101 (85)(94) ant. 4 10212 (95)(104) ep. 4 113/4120 (105)(112)

The two stanza-forms (i.e. the strophe/antistrophe and the

epode) of Olympian 9 belong unambiguously to diVerent classes.
This is highly unusual. Even in a few exceptional poems (P2, P5,
P10) where the classiWcation changes from strophe to epode, the
diVerence between the stanza-forms is subtle. Some similarities are
undeniably visible. O9 is, however, a one-oV. The strophe is very
simple aeolic. Its structure is one of the simplest in Pindaric
aeolic. It thus belongs to Class I without any doubt. In contrast,
the epode includes few asymmetrical phrases. There are verses
which have features in common with those of Class II odes, but
many unfamiliar, sometimes perplexing, phrases are included. It is
possible to speculate that the metrical contrast may be a device to
reXect the contrast between the simple structure of Archilochean
songs and the far more elaborate songs of Pindar himself; str. 1
starts with e b   while ep. 1 starts with Kg 
 
 E K
  IE . For the assumption
that the central part of the epode may be related to the metre of
Stesichorus or Ibycus, see Part I, 7. 6. RSS is low both in the
strophe (43.8%) and in the epode (42.1%, the lowest of all the
non-D/e stanza-forms).
172 The Eighteen Majors

Strophe
In the central part (2), the identical or nearly identical verses are
repeated Wve times from s3 to s9. The Wrst three (s3, s4, s5) are
completely identical: glyconic is followed by reizianum. Instead of the
glyconic, the fourth (s6/7) introduces its variation, wilamowitzianum.
At s8 and s9 the sequence gl reiz returns, but this time as separate
verses. Such a straightforward structure is unusual in the aeolics of
Pindar. He occasionally repeats the same or similar verses as at P2e13
or N4s46, but never at such length. Our case has aYnities with the
short stanzas of Anacreon (though gl reiz is not his choice), or some
simple odes of dramatic poetry (Euripides or Aristophanes).
Another unusual feature is the variety of forms of the aeolic base of
glyconics. In his usual practice, Pindar maintains the same form
throughout all the repetitions, and, even when a diVerent form is
used in responsion, it is restricted to one, or two at most, repetitions
(Part I, 5. C. 1). Generally he prefers
w (and even www) to

,
contrary to the tendency of Anacreon or the dramatic poets. In this
ode, however, the ratios of
w to

in the four glyconics from s3
to s8 are 5/3, 0/8, 5/3, 1/7 respectively. 0/8 means perfect regulariza-
tion and 1/7 accords with the All-But One rule, but the predomin-
ance overall of

is signiWcant.
Returning to the beginning of the strophe (1), the Wrst two verses
are also simple aeolic. The ancipitia in s2 are generally long. In 3 the
penultimate verse (s10) starts with a simple aeolic colon, but notably
the anceps is always short. And it uses e as suYx. Here, the Pindaric
Xavour is prominent for the Wrst time. The last verse (s11) is situated,
as it were, at the midpoint between aeolic and freer D/e.

Epode
In contrast with the strophe, the epode is very unusual. Except for the
Wrst and last verses, which are adjacent to the strophe, single-short
movement is totally absent. These two verses (e1/2, e8) resemble each
other, although the analysis above conceals the resemblance:
e1/2 w
w
w
w
ww
w
w
w

k
e8

w
u
w
ww
u
w
ww
w

k
 Olympian Nine 173
The diVerence lies in two explicit ancipitia (and the initial anceps)
as well as the inserted ww
in e8.
The central Wve verses are made up mostly, though not completely,
of spondees and dactyls. Spondee occurs at the beginning and end of
e4, duplicated in e5 after d, and also placed at the end of e7. Snell, in
the metrical scheme printed at the head of his text, moved the second
spondee in e5 at the beginning of the next line, though the line is
indented (the lay-out of the text is diVerent). Whatever he may have
intended, verse-end should not be supposed between the two spon-
dees; see v. 57 ( 53 Sn.)   j q ( Snell) and further,
textual problems ad loc.13 Spondee is not a common phrase in the
eighteen majors (12 in total), but, signiWcantly, about half of the
examples are concentrated in this epode. (The others are found at
O2e4, O10s3a, O10s3b, P5s7, P8e7, P11s1, I7e6; some of them will be
cited below as parallels.) Spondee is occasionally found at verse-
beginning and verse-end in normal D/e. For example,
(before e) P1s3


w


w
k sp e
e
(after e) P9s2
w


w


k e
e sp
(after D) P1s2
w


ww

w


ww
ww


k e
d e

D sp
Most certainly the Wnal two longs in e4 and e7 are comparable with
those in P9s2 or P1s2, and the initial ones in e4 with those in P1s3.
Another peculiarity is dactylic movement, which is not very
common either, even in freer D/e, except for N6. It occurs in three
verses: e3 (^ D), e6 (
D ), e7 (

wD
ribyc heptad).
The arrangement of e4e7 is symmetrical and, so to speak,
chiastic:
0
O9e4



ww
w


k sp tel sp
B

B
B O9e5
ww




j d sp sp
B
@ O9e6

ww
ww
ww
k
D
O9e7

w
ww
ww


j
ribyc ( heptad) sp

13
The discrepancy between the layout of the text and the metrical scheme in
Snells edition may be explained thus: Snell carelessly follows Schroeders metrical
scheme while, in the main body of the text, changing the layout. According to the
idiosyncratic theory of Schroeder, the initial short syllable of  does not matter. He
deletes  before it.
174 The Eighteen Majors
The inner two verses, e5 and e6, are similar. The similarity would be
strengthened if the spondee at e5 were supposed to be equivalent to a
dactyl. The structure of the outer two, e4 and e7, are also similar.
The central part of both is a sequence which combines double short
(s) and a single short, the order being reversed. A spondee follows it.
As is demonstrated in Part I, 7. 6, asymmetrical cadence and
introduction in the vicinity of dactylic verses occur in the prototype
D/e of Stesichorus and Ibycus. The Wnal spondee may be related to
Alcman and others. For the initial spondee, Stesichorus S222(b) (the
so-called Lille Thebaid) ep. 6 can be cited:



w

k
In Wve dactylic verses (e37), word-end between two shorts
tends to be avoided. Localization of words like B  

(
wjw
wjw
) is absent. There are in total nine paired shorts,

(
)ww
, in each of these verses and they are repeated four times.
Word-end occurs between the two shorts only in one out of nine
pairs in each repetition; thus there are in total four irregular cases.
Moreover, each of these four irregular cases occurs in a diVerent
verse (e4: ep. 1 v. 26 e $

 e5: ep. 2 v. 57  E; e6: ep.
3 v. 88  b d; e7: ep. 4 v. 119 , H . As a result the
rhythm of these verses feels very square, for instance
e6 v. 28 N   



wwj
ww
jww
j.

Textual problems
1820 (s6/79). All modern editors follow Bergk2 and adopt his
ingenious emendation (felicissimo ingenii acumine, Schroeder, ed.
maior):

  IE
 , ,

F  :
Bergk, in fact, owes much to Boeckh. His colometry is fundamentally
the same as Boeckhs division. It was Boeckh who Wrst rejected the
vulgate text and established verse-ends: IEk and

k.
Synartesis is impossible because verse-end is attested at each (s7, s8,
 Olympian Nine 175
s9) by hiatus/brevis in one or more other repetitions. It must be
remembered that A, which preserves    ,

had not
been discovered yet; thus Boeckh conjectured As reading without
knowledge of it (it was discovered by Mommsen). He also expels r 
(sic in all MSS except A; see below) and scans F disyllabically
against the usual trisyllabic scansion; cf. not. crit., p. 397. Bergks
merit lies in the introduction of the invocation to Castalia in s8 by
deducing  from r. Hermann (Notae) insists on retaining the
(unmetrical) vulgate:

  IE , Y-
  
-
F  
He writes: si hi versus . . . asynarteti sunt, nihil opus est mutationi-
bus, quas Boeckhius durissimas et minime elegantes fecit. It is one of
the cases in which Hermann makes use of asynarteta as a weapon to
reject the deWnition of verse-end established by Boeckh. As late as in
1847 (Quinque Ol.), he did not abandon the idea of asynarteta but
united s8 with s9:
 , 
-jF  
His deWnition of asynarteta is: [versus] qui sicut placeret poetae
nunc cohaerentibus numeris decurrerent, nunc ex duobus constarent
non continuatis numeris (Opusc. viii, 115). I record Hermanns text
here, not because it should be adopted nor because his metrical
theory on asynarteta is right, but because it shows how diYcult it
had been to recognize verse-end as the primary necessity of modern
metrical studies.14 For

at the verse-end, o
 (O1 v. 57; pre-
verbal) is a remote parallel. As for disyllabic F, there is an
example outside Pindar: Ba. 6.v. 3.
Kayser introduces

 

because Quum spondeus in omnibus ceteris strophis octavum ver-

sum inchoet, non credibile est, in hac sola trochaeum a poeta fuisse
admissum. But this is not strange (All-but-One).

14
For asynarteta, see Itsumi, Whats in a Line?.
176 The Eighteen Majors
57 ( 53 Sn.; e5). Snell retains the paradosis . The Wrst
position is long at two other repetitions in e5. The fourth, 87 ( 81
Sn.), is indeterminate: . Turyn introduces q. For q, see
Braswell, Nemean Nine, on v. 17 and Gerber, Olympian Nine, ad loc.
Boeckh, Mommsen, Christ, et al. adopt  (Byz.). Schroeder
retains  and sets verse-end after the Wrst spondee. Supposing
asyndeton, he deletes  before . Thus the two positions for 
stand at the beginning of the next verse. The Wrst syllable of  is
accepted as anceps. He introduces anceps also at the beginning of the
following verse, e6 (for 
 89 ( 83 Sn.), see below). Presumably
he Wnds some kind of parallelism between e5 and e6. Bowra exactly
follows Schroeder; his colometry is the same as Schroeders, and he
chooses  and 
. But how he analyses the metre is uncertain.
Wilamowitz (Wil.3 in Gerbers Emendations) supposes w
to be
required at all the repetitions and, retaining  and  as they
are, emends two repetitions: 27 ( 25 Sn.) 
 (for ), and
117 ( 109 Sn.) Z IF  (for Z  ).
Snells metrical analysis (chotrim) must be rejected, together with
Schroeders theory. Dale retains  in a diVerent metrical analysis
(the proper Pindaric form  need not be emended to the dubious
 or q, Collected Papers, 73), maintaining that the seventh
position is, as well as the Wfth and others in the preceding verse (e5),
is a drag:

ww


u
d
s(
s)
Her conception of drag is unorthodox, and she resorts to it too
frequently.
60 ( 56 Sn.; e8). The paradosis F is unmetrical; A
Triclinius.
82 ( 76 Sn.; s11).  in the paradosis is unmetrical. Snell obelizes
it. Various emendations have been proposed. The easiest is either
F (Mingarelli, followed by Mommsen and Christ) or changing
the word-order with insertion of :  P  (Hermann,
Notae, followed by Boeckh and Schneidewin). More sophisticated are
r  (Bothe), Z (Bergk, LPG 2),  (Hermann,Quinque
Ol.), keeping the word-order of the manuscripts. Turyn, on the same
lines, writes  . Schroeder retains the paradosis by defending the
 Olympian Nine 177
awkward responsion
uw
. This idea of responsion itself is rejected
by Maas, but he classiWes the passage amongst nicht evident geheilte,
weil auf verschiedene Weise heilbare Stellen (Freiheiten, 11, Beleg 75).
He cites F (Mommsen, in apparatus) as a sample of the emend-
ations. See further Gerber, Olympian Nine, ad loc.
89 ( 83 Sn.; e6). The majority of manuscripts have 
.
Moreover, A has unmetrical N after it. Most editors adopt

, except Bergk1 (but 
 Bergk24), Schroeder, and
Bowra. These adopt 
 (HN). Two questions arise: (i) whether
Pindar used the aorist-stem starting with 
-, and (ii) whether the
present tense is preferable to the aorist. There is no metrical problem,
because the position is anceps. 
 produces exact responsion,
but Pindar may not have been worried about breaking it (All-but-
One). Braswell, Glotta, 58 (1980), 205 V. refutes Schroeder, who
banned all non-indicative forms of 
-, and supports 
 in
our passage: in the non-indicative moods early Greek uses either the
present or the aorist. And the unmetrical gloss N: It was obviously
added by someone who felt that the notion of duration was not
adequately expressed by the aorist optative that he found in his text.
Nd (A, see above, Csl) is expelled by most, but there may be a
chance for it to be retained. Mommsen suspects
 and estab-
lishes the text: 
 N:  . . . Instead of
,  may be
used in similar contexts (P10. 64, and N7. 65, where Mommsen
introduces it in place of
 to heal a metrical irregularity). In
the same line, the manuscripts unmetrically read X, which must
be q (Byz.).
106 ( 99 Sn.; s5). The paradosis K is unmetrical; N
Byz.
120 ( 112 Sn.; e8). `N
  K  "
 N
 H
(MSS) seems grammatically impossible even if we disregard the
metre. was introduced by Byzantine scholars who thought the
preceding  (acc.) and H (nom.) were incongruous. had
been accepted by older scholars (E. Schmidt, Boeckh, Hermann
(Notae), Schneidewin), with various emendations. But since  is
present, sententia eadem manet, sive pronomen addideris sive
non addideris (Christ). Perhaps it may be helpful to put a colon after
the preceding verse, I
 (Hermann, Mommsen), rather than a
178 The Eighteen Majors
comma (Christ, Snell). It was Hermann (Quinque Ol.) who Wrst
deleted ; his conjecture `r ,   is now generally accepted.

Individual verses
s1. There is no strictly identical verse. However, the telesillean start-
ing with double short is quite common (6 in total, and 2 others
suYxed by w

). Also common is the telesillean that makes a verse
by itself, not combined with any other components (4 in total):
O10e6, P5s7b, I7e3 (all of these start with long anceps).
s2. The half base of the telesillean is always long except for v. 62 ( v.
58 Sn.) (  ), and the full base of the gl 3 is
w, except for v. 73
( v. 68 Sn.) (

, A,  ). Both follow Pindars general ten-
dency. Bridge between the two cola is observed in general, but breached
at v. 2 and v. 103 ( v. 96 Sn.). There are two other verses in which a
telesillean precedes another aeolic colon starting with full base: I7s3/4
(xe tel gl 3) and N7s8 (tel hipp). The examples of gl 3
amount to three (others are: I7s3/4 (above) and N7e5 (gl gl 3).
s39. See Part I, 5. E. 1 (rdod starting with
x) and for s9, 8. A. 3
(shorter verse). For the repetition of the same structure, and for the
general tendency of realization of anceps, see above. The combination
of gl/wil reiz is found in N2s3 (gl reiz), N4s4 (gl reiz), N4s6
(wil reiz). All are found in Class I odes. No tendencies are observable
about bridge either before or after the anceps ( half-base) in s3, s4,
and s5. The most interesting case is s6/7. Except for one repetition,
v. 47/8 ( v. 44 Sn.), where elision of  is involved, word-end always
falls at the end of the wilamowitzianum (thus Boeckh separates s6 from
s7 to make two verses). In other words, s6/7 is virtually two verses, like
the following s8 s9, in most repetitions.
Verbal assonance: 75
E    94
E  K 
 ;
18
  IE  108
d b E .
s10. The anceps of the telesillean is always short. This is against
Pindars general tendency: usually long is preferred to short at the
half-base, especially in the cases at verse-beginning; see Part I, 5. D. 1.
The suYx of e is typically Pindaric: there are in total 15 examples of e
following an aeolic with 2 ending; see Part I, 8. B. 5. Bridge between
the telesillean and e is always observed in our verse,
 Olympian Nine 179
s11. See Part I, 6. D (bridge after long anceps), 8. C. 2 (double d).
The Wrst eight positions (anceps d e) do not constitute an aeolic
colon; accordingly, it seems natural to describe the latter half of this
verse as anceps d anceps. It is, in fact, equal to the reizianum of
the preceding verses. The arrangement of the last two-thirds (9
positions) of s11 is completely identical with that of s36/7; more-
over, the verse is exactly as long (14 positions). The diVerence lies in
the Wrst Wve positions:
s11

ww

w
u
ww

k
s3
u
ww
w
u
ww

k
The unsettled question emerges whether every reizianum is the cata-
lectic version of the telesillean, so that the Wnal long of the reizianum
was given the value of three shorts, i.e. longer than an anceps. If so, the
diVerence between reizianum and anceps d anceps must have been
audible. The sequence of anceps d e is paralleled:
N7s5 w
ww

w
wrw
k wd e 3
Alternatively, the last nine positions in s11 might be analysed as an
aeolic colon (
x
ww

). But that is less plausible. There are
few examples of this colon (P6s7/8), and even if it is accepted, the
anceps at the fourth position is against the general tendency of
Pindaric reversed dodrans. The position is predominantly realized
as short, whereas in our case it is long at as many as three repetitions
(41 A, 71  , 101

).
e1/2. See Part I, 8. A. 6 (d) (palindrome); 8 C. 3 (enwd). Word-end
coincides after the eighth position throughout all the repetitions.
Many editors (but not Turyn) divide this verse into two at this point.
But in that case, the second verse (e2) is diYcult to analyse. Instead,
the whole of e1/2 should be taken as a variation of a long sequence of
single-short movement turning to d. Rightward extension by single-
short movement is not paralleled by so many verses as leftward
extension, but there is one example:
N3e1b w rw
w
wr w
ww
w
w

w
k we 3wdwe e
This verse is longer on the left side and shorter on the right than our
verse (but e is attached without link), but its basic structure is similar.
Outside the eighteen majors two verses are very similar to ours:
180 The Eighteen Majors
O14s2 w
w
w
ww
w
w

j we wdwe

O9e1/2 w
w
w
w
ww
w
w
w

k w e 2 wd we 2

O4s5/6 w
w
w
ww
w
w
w

k w e wd w e 2

The uniWcation of e1 e2 is now supported by Barrett, who avoids a
short vowel at the end of v. 83 (77 Snell).
e3. There are in total four verses that start with acephalous D.
Among them,
O13s1 ww
ww

k ^D

is identical with ours. Others are: O10e8 (^ D) and O10s1 (^ D e e).
e4. The Wrst three long syllables of our verse are here analysed as
spondee initial anceps of telesillean. For the spondee, see above,
epode. Instead one may posit the anceps true long anceps of the
aeolic enneasyllable:
x
x
ww
w


k
I reject this analysis, which requires the supposition that the second
position of the reverse dodrans (
x
ww
), in our case the third
position or the second anceps, is realized as long at all the repetitions.
This is most unlikely (see Part I, 5. A. 1); in Pindar the position is
mostly short.
e5. There is a possible parallel for this unusual verse, though
spondee is not duplicated there:
I7e6
ww


k d sp15
e6. There are four examples of
D ; among them,
N6e9

ww
ww
ww
k
is completely identical. Others are N6e3 (
D d) and P2s3 (
D e).
e7. This strange verse must be compared with another strange
verse, P11s1:
O9e7

w
ww
ww


j
P11s1

w
ww
ww
ww
w


k

15
The verse may be united with e7 by adopting Barretts emendation; see ad loc.
Then the Wnal spondee will disappear.
 Olympian Nine 181
The identical structure is astonishing. Both include what I call
reverse ibycean or its cognate (ribyc 2) at the core. Link anceps
stands at the beginning of the verse, and a spondee at the end.
e8. See Part I, 8. A. 3 (ii) (a) (longer verse), 8. A. 7 (repetition
within a verse). The colometry is clearly indicated by two ancipitia.
Long and short appear 2/2 and 3/1 respectively. Duplicated heptasyl-
lable is absent elsewhere. Heptasyllable preceded by e is paralleled:
P2s8

ww
w

w
x
w
ww
w

k tel e hepta 3
182 The Eighteen Majors

O LY M P IA N T E N

Five triads. Class II

(1)
O10s1 ww
ww

w
t w
k ^D ee
O10s2 u
w

ww
w
k xe dod
O10s3a u
w

wr

wr

j xe e
e sp
O10s3b rw



ww
ww
k e sp D
(2)
O10s4 w

wr

w
k ^ e e
e
O10s5 a
w

w
k xe e
O10s6 a

ww
w
k gl
(3)
O10e1a w
wr
ww
j we d
O10e1b u
wr
k xe

O10e2 w

ww
ww
k ^e D
O10e3 t
wr
wr

ww
y
w
k xe e
dxe (here x
means a biceps)
O10e4/5

ww

ww
rwj

j
d d e

(4)
O10e6

ww
w
k tel
O10e7

w
w
k
e 2
O10e8 ww
ww
k ^D
O10e9

ww
ww
w
k
ibyc ( teld)
(coda)
O10e10 w
wwjt

wr
w

ww
k wd
e e d
s1 H 1, 73; B 7, 23, 29, 89, 95; s2 H 68; B 2, 8, 24, 46, 90, 97; s3b H 31b; B 91b; s4 H 4,
26, 92; B 4, 10, 26, 32, 48, 54, 76, 92; s5 B 55, 71; s6 B 6, 28, 72, 78, 100; e1b B 57b; e2 H
102; B 36, 58, 80, 102; e3 H 59, 103; B 15, 37, 59; e6 B 18, 62, 84; e7 B 63, 85, 107; e8 H
86, B 108; e9 H 65, 109; B all (!); e10 B 44

The analysis above, on which Part I is based, presents quite a few

diYculties in e3 and e10. An alternative, experimental analysis,
which tries to eliminate these diYculties, will be given in a separate
section at the end. For the analysis and the notation of e9 (
ibyc),
see Part I, 7. 6 expanded aeolic, and also Appendix B, Dactylic
expansion revisited.
 Olympian Ten 183
s1 1 (1) (e 
 I
 
s2 2 (2) 

E,
 
s3a 3a (3) KA 
 f a PH  O
s3b 3b (3) K
 t E , Ia f d

s4 4 (4)
  , O A 
`
s5 5 (5) K 
s6 6 (6) K
a I.
e1a 13a (13)  a `

e1b 13b H ,
e2 14 (14)    

e3 15 (15) d
 @ : 

 b  
d $

e4/5 16/7 (16) ! 
  K 

 H
e6 18 (17) " 

e7 19 (18) < ,
e8 20 (19) E
 .
e9 21 (20)
    I A

e10 22 (21)
 
  Ic F f

 .
str. 2 238 (22)(27) ant. 2 2934 (28)(33) ep. 2 3544 (34)(42)
str. 3 4550 (43)(48) ant. 3 526 (49)(54) ep. 3 5766 (55)(63)
str. 4 6772 (64)(69) ant. 4 738 (70)(75) ep. 4 7988 (76)(84)
str. 5 8994 (85)(90) ant. 5 95100 (91)(96) ep. 5 10110 (97)(105)

The rhythm of Olympian 10 is basically freer D/e, like P5 or N6

(Class II). It includes four asymmetric phrases (s2, s6, e6, e9), but the
essential phrases are e and, to a lesser extent, d. They are occasionally
linked by anceps; mid-verse long anceps is one of the main charac-
teristics of Class II stanza-forms, and is comparatively frequent in
O10. There are in total four examples (s3a, s4, e3, e10). Dactylic
movement is also conspicuous, being present in six verses (s1, s3b,
e2, e3, e8, e9). One of them, e9, can be analysed as an expanded aeolic
as well, teld, and likewise the end of e3. Expanded aeolics (see Part I,
7. 6) are, in spite of the appellation, peculiarities not of Class I stanzas
( aeolic), but of Class II ( freer D/e), like O9e7, N6s4, or N6e3.
Thus I analyse e9 as link anceps ibycean, ibycean being associated
with D. The long of e is frequently resolved. Especially the form

wr returning at every repetition is remarkable (3 examples in
strophe, 6 in epode).16 Resolution of either long of e and the presence
16
In the eighteen majors
wr occurs 33 times and rw
27 times: the
numerical diVerence is statistically insigniWcant; see Part I, 6, C. But in O10, the latter
is scarce.
184 The Eighteen Majors
of adjacent link anceps make colometry ambiguous here and there,
especially in s3a, s3b, s4, e1b. Two irregularities are perplexing. One is
the biceps, used twice in e3. The other is the spondee in the middle of
verse at s3b. Moreover, the glyconic in s6 is abnormal in the strophic
responsion of aeolic base, and the long of d is resolved at e10. The
transmitted text is certainly corrupt in some passages which also
contain metrical problems. The metre of this ode is perhaps the
most diYcult to analyse of all the Pindaric epinikia.
It is interesting that brevis in longo occurs with an extraordinary
frequency at no less than half of the verse-ends: in the strophe (10
repetitions in total), Wve at s1, six at s2, eight at s4, Wve at s6; and in
the epode (5 repetitions in total), four at e2, three each at e3, e6, and
e7, Wve at e9. For other examples of concentrated brevis in longo, see
Part III, F. This means that, if the ends of the other repetitions are
modiWed, the verse in question can be united with the following verse
and the united verse will take on quite a diVerent form. For example,
if we conWne ourselves to the Wve repetitions of the verses (ant. 1, ant.
2, str. 3, ant. 3, ant. 4), s4 and s5 could be run together to make
w

wr

wwjw
w

w
k ^e e
e e e
Could it be that Pindar was wavering till some relatively late stage of
the composition of the ode between two ways of dividing the verses?
SigniWcantly, all the verses except one (s3b) start with an anceps or
a short position, and most end blunt (there are 3 exceptions: s3a, e1b,
e4/5). As is usual in Class II odes, many verses are short. The average
number of positions is 10.3 in the strophe and 9.5 in the epode. The
strophe is considerably shorter than the epode: 72 positions in the
strophe to 98 in the epode. The ratio (s/e) is 0.73 ( 72/98), the
lowest of all the eighteen majors, and even taking the normal D/e
odes into consideration, it is still the lowest.

Strophe
The main part of the strophe is composed of e. It resembles O2,
though e is not repeated within the verse as frequently as in O2. There
are two dactylic phrases (s1, s3b), which frame the Wrst half of the
strophe (1) (e.g. 1 e 
 and 3b Ia f d
).
1 both starts and Wnishes with ww
ww
. Also, two dodrantes at
 Olympian Ten 185
verse-end have an important function in articulating the structure
(e.g. 8  f  and 12  K
). One of
them is a dodrans in itself (s2) and the other forms part of a glyconic
(s6), but the similarity is obvious:
s2 u
w

ww
w
k xe dod
s6 a

ww
w
k gl
The initial part of s2 (x
w
) is repeated in s3a and s5. And there is
one more structural correspondence:
s3a u
w

wr

wr

j xe e
e sp
s4 w

wr

w
k ^ e e
e

Note the resolution and mid-long-anceps at the same position.

The responsion a
at s6 is highly irregular. The correspondence
between w
and

in aeolic base is common, of course, in
Lesbian poems and in drama. But this verse is the only example in
the eighteen majors that testiWes to the correspondence, not only in
glyconic but also in all the other aeolic cola starting with full base; cf.
Part I, 5. C. Long syllable occurs at two repetitions (v. 50 v. 48 Sn.

 , v. 94 v. 90 Sn. 
). At all the other eight repeti-
tions, the position is regularly realized as short. It is very strange that
such an unusual practice occurs in a Class II stanza-form where
glyconic is not otherwise present. Even in Class I (aeolic) where
typical aeolic characteristics such as responsion between
w and


or dovetailing are retained, this responsion is totally avoided.
As far as Gerbers Emendations reports, only Hermann (Quinque
Ol.) seems to have doubted the paradosis. His proposal is I

for 
 , and  or
 for 
. But these are too
ingenious to be accepted. Glyconics can start with x
anyway.
Perhaps Pindar introduced a one-oV licence in this ode, like that in
e3 below, and tried to substitute

for w
at the beginning of a
verse, as if to give a Wrst foot of x
w
. Then our verse is literally a
shorter version of s2:
s2 u
w

ww
w
k xe dod
s6 a

ww
w
k x
dod
And if this supposition were accepted, then the metrically suspect
word H at v. 26 ( v. 25 Sn.; s4), which must be rejected
186 The Eighteen Majors
according to current metrical theory, could be left unemended for
the same reason:
s4 a

wr

w
k x
e
e
Here the regular w
(^ e) is replaced by

. For the emendation of
H, see further on textual problems. Perhaps this type of licence
may be paralleled outside the eighteen majors. In Paean 9, x
at the
beginning of two verses may not be aeolic base; see Appendix B.

Epode
I tentatively divide the epode into two halves. The structure of the
Wrst half (3) is easily grasped when the phrases are arranged hori-
zontally (irregular responsion between long and double short at three
positions is ignored):
e1a w
wr
ww
j
e1b u
wr
k
e2 w

ww
ww
k
e3

wr
wr


ww
ww
w
k
e4/5

ww


ww
rw

j
cf.
e10 w
ww



wr
w

ww
k
An anceps stands in the left-hand column.
wr is located in the
centre. The phrase is occasionally modiWed: it is cut short into
in
e2, and doubled in the Wrst third of e3. Finally in the right-hand
column
ww
is situated. In e2, this is expanded to D. These
phrases are inverted in e4/5: the d phrase precedes the e (
ww

rw

). Moving on to e10, we Wnd the same structure repeated.
The initial anceps of e1b is short at one repetition (57b 55 Sn.)
although it is regularized by all the critics; see textual problems
below.
The main part of 4 consists of three short verses. The Wrst two are
similar. The double short in e6 is replaced by a single short in e7.
There is no single short in e8. E
 (v. 20 19 Sn.) is an
 Olympian Ten 187
impressive phrasing. The whole of e8 is the same as the beginning of
s1, which frames the Wrst half of the strophe (1, see above). e8 is
incorporated in the next verse too:
e9

ww
ww
w
k
At the same time e9 is an expanded form of e6, and has already
occurred in e3 (except at one repetition; see the Wnal section). The
Wnal verse, e10, has characteristics common to 3. See the chart
above.

Textual problems
8 (s2). The paradosis  is metrically wrong; 
Boeckh.
9b (s3b). H should be emended both for semantic and met-
rical reasons. Without H, s3b starts with www
(e) at all the
repetitions. Many are suspicious of H, but not Schroeder, Turyn,
or Verdenius. It is not the K
a of people in general (H) but
of Hagesidamos, or his father Archestratos, which  , interest,
can absolve (see Erbse, and Barrett, Collected Papers, 5472). Very
many emendations are proposed for H. They are words of the
shape w

starting with a vowel, which makes exact responsion
with all the other repetitions. Hermanns O
 (Notae) (cf.
`(*=`(*') was adopted by Boeckh and Schneidewin. It
is still kept in LSJ (v. O) and in the apparatus of many editions.
Like O
, most critics look for a word which qualiWes  : e.g.
O
 (Mommsen) or  (Christ). Erbses I
 qualiWes
F, but it is, in a broader sense, on the same line. On the other hand
some propose a verb which is attached to the following sentence, like
Schneidewins (Gerbers Schn. 8, monente Wiesler) A t. This
was adopted by Bowra. Fennell reWned it to
 (the subject is
Hagesidamos), which he later withdrew in his new edition, but
Barrett revives (Collected Papers, 61).
However, the metrical problem does not seem to have been thor-
oughly examined. I myself believe that the text is corrupt, but there is
some chance of H being right. If it were accepted, s3b would be
analysed as
188 The Eighteen Majors
wy



ww
ww
k ^ e
sp D
and  Hw


respond with wr

in the other nine
repetitions. The lack of exact responsion at only one of the repeti-
tions is a common feature in Pindar (All-but-One; Part III, C). ^ e
with resolution at the beginning of the verse is paralleled:
O2s2 wy
w

wr
wr
w
k ^e eeee
(here resolution occurs at 9 repetitions out of 10; the same frequency
as our case.)
Unlike O2s2, ^ e in O103b is followed by three longs. Most possibly
the Wrst of them is anceps (resolved long followed by anceps is
paralleled by O2s3 and O10s4). The next two must be a spondee
(sp D cannot be D ; double short in D and D are not be
contracted). This analysis is not a very welcome one, but, even if
H is expelled, the spondee remains. At two repetitions of
O10s3b word-end occurs after the anceps; i.e. bridge is ignored
between the long anceps and the following long. This is a counter-
argument against the analysis above, but not an absolute one (for
long anceps and bridge, see below). Alternatively, s3b may be ^ e
sp
D. This analysis hardly diVers from the one above.
Or should s3a and s3b be united, as most editors have believed
since Boeckh? Turyn is an exception (and perhaps Snell too, because
his metrical analysis seems to require two independent verses; if so,
he should not have indented the text there). Then another
wr

is created instead of ^ e:
u
w

wr

wr

wy



ww
ww
k
xe e
e
e
sp D
This is very long (25 positions), but not excessively so. There is a
verse of 23 positions in the eighteen majors (N2s1), and a verse of 27
positions is found (N5s1) in the D/e odes. I prefer the separation
because the other verses of this stanza are short and because the
shortness of verses is one of the characteristics of Class II. But this, of
course, is not compelling.
21 ( v. 20 Sn.; e8). The paradosis d is unmetrical;  Byz., Npc.
 Olympian Ten 189
26 ( v. 25 Sn.; s4). Except here, every repetition starts with w
,
acephalous e, which is common before full e. Snell obelizes H and
records two conjectures in the apparatus, g (Theiler) and g
(Erbse). Christ in his apparatus proposes an ingenious emendation,

 (cum H metrico vitio laboret et arae nihil ad certamen
attineant, H ex
 corruptum esse duco, ut quae infra sex
genera certaminum, cursu lucta pugilatu quadrigis iaculo disco,
referentur, iam hoc loco signiWcentur.). Bowra and Turyn adopt it.
If Christs
 is rejected, H, which in fact gives good sense,
cannot be accepted metrically, unless we accept an over-ingenious
interpretation (see on the strophe). H 
 K is
the reading of A; the other manuscripts are more unmetrical, reading
e H !  K (or K:! .). The older edi-
tions (Heyne, Boeckh, Schneidewin) accept Triclinius reading 
!  K. Hermann too (Quinque Ol.) takes it with
 in place of K at v. 25 ( 23 Sn.). Mommsen,
Bergk4, Christ (in the main text), and Schroeder (BT) keep H
H, but without persuasive metrical explanation.
36 (35 Sn.; e2). The paradosis Z
 is unmetrical. Z
 is the
reading of Byz. and Npc.
57b (55 Sn.; e1b). The older manuscripts read
:

(Byz.) regularizes the quantity of the Wrst position, which is long in
the other four repetitions. It is adopted by almost all editors. Snell
analyses the Wve positions of e1b-part as sp cr. If this analysis is
adopted,
 would be indispensable. However, the regularization
is unnecessary because, according to my analysis, the position in
question is anceps. Cf. Mommsen (annotatio p. 139): Possit dividi
[ep. 1] ut etiam
 vs. 55 teneas. Omnino haec numerorum
deWnitio suavior (si recte sentio) quam vel Bockhiana . . .. This verse
is one of the cases where recognition of link anceps is vitally important,
otherwise spondee or dochmiac would be introduced. For the reso-
lution of the penultimate position, see Part I, Appendix, addendum II.
59 (57 Sn.; e3). For I, see the Wnal section.
67 ( 64 Sn.; s1). The paradosis is unmetrical (P). Pf
 (Thiersch) is generally accepted. Critics since Schneidewin and
Bergk2 construe it with . Dissen and Mommsen are the ex-
ceptions who adopt the Hermanns emendation  (Hermann
(Notae): Pf
).
190 The Eighteen Majors
73 ( 70 Sn.; s1). The paradosis gives the meaningless A
I #A  is the emendation of Boeckh
based on the scholia; the last two words are scanned
rw
with
crasis of - (some accordingly write ).
74 ( 71 Sn.; s2). Asyndeton is introduced by Moschopoulos, who
deletes the unmetrical b.
90 ( 86 Sn.; s1). The pardosis u is unmetrical; z Boeckh.
95 ( 91 Sn.; s2). The unmetrical b is deleted by Byz.
103 ( 99 Sn.; e4).  is inserted by Moschopoulos. For
`
, see the Wnal section.
10910 ( 1045 Sn.; e910). Boeckh introduces
 instead of

 and establishes the verse-end at e9. Then brevis in longo occurs
at the end of e9 in all the repetitions.
110 (105 Sn.; e10). The paradosis
  is one
syllable too long and metrically wrong.
  (Hermann,
Notae) is adopted by Boeckh, Schneidewin, Schroeder. Momm-
sen proposes  , and is followed by Christ and Turyn.
Maas (Nachlese) retains
 by changing  into p,
which Bowra and Snell accept. The presence of
 in exactly
the same metrical position in v. 44 ( 42 Sn.) is an argument for
reading it here as well, but, in view of Pindars usual practice, not
so strong an argument as to warrant introducing the unusual
verbal form p or p (see Part III, E). Verdenius accepts the
correspondence between cretic and iambic, citing Wilamowitzs
authority. But the responsion based on the assumption of triseme
is not paralleled by unambiguously certain examples, at least in
Pindar.

Individual verses
s1. See Part I, 6. B (acephaly), 8. C. 1 (e and D within a verse), 8. C. 5
(double e), 8. C. 8 (D e). It is rare for dactylic sequences to be
followed by e. Of four examples of acephalous D (^ D), O10s1 is
unique. N6s4 is most similar though its dactyls are extended:

N6s4 ww
ww
ww
rw

j ^D e

In the normal D/e odes, there are two examples of ^ D followed by e,
though a link anceps is inserted between them:
 Olympian Ten 191
P3e9 ww
ww


w


w
k
N8s4 ww
ww


w

ww
ww

k
The Wrst long of the second e is resolved at one repetition (v. 73
v. 70 Sn.) out of 10. It is caused by a proper noun which is the
emendation of Boeckh based on the scholia (see above, Textual
Problems).
s2. The initial anceps is not regularized; long at seven repetitions
and short at three. There is a verse which is of the same structure:
P8s4 w
w

ww
w
k we dod
Verbal assonance: 2
   8 f  , 30 I  74
.
s3a. See Part I, 6. C. 3 (er
e), 8. C. 1 (e and D within a verse),
8. C. 5 (double e). Although hiatus/brevis is absent, the coincidence
of word-end throughout all the repetitions suggests the division into
two verses, s3a and s3b. For the possibility of a united verse, see
above, Textual Problems. The extraordinary internal spondee remains
whichever verse-division is adopted.
Another analysis of s3a is conceivable, according to where resolu-
tions are supposed to occur:
u
w
l x rw
l x rw

j xe
e
e

But this colometry introduces two irregularities and must be
rejected. In the eighteen majors (i) resolution is never preceded
by a long anceps (see Part I, 6. C. 3) (ii) cut tends to be avoided
after long anceps (Part I, 6. D); it would occur at Wve repetitions
after the Wrst long anceps and at nine repetitions after the second,
according this colometry. In contrast, bridge is observed through-
out all the repetition according to the other colometry. Moreover
s3a harmonizes with another verse in this stanza: s4 (see above).
O2e4 is structurally similar, and another colometry seems possible
(see ad loc.):


w

wt
wr

k
e e e sp
Outside the eighteen majors, there is a verse which is strikingly
similar to ours:
Pae4s3

w

wr

wr

jwr

k
e e
e

e sp
192 The Eighteen Majors
The verbal unit of wr

j at verse-end is distinct. In nine out of
10 repetitions, word-end occurs at
u
w

wr

jwr

j.
The exception is v. 9 OEj- K
a. Moreover, at four repeti-
tions, the same verbal unit is repeated:
u
w

jwr

jwr

j.
(v. 3 KA 
jf a PHj  O). Although the
analysis is diVerent, rw

j is used at the end of e4/5 too; for its
diVerence from wr

, see below on e4/5.
Verbal assonance: 3 KA 
  9  b F  97 Ic

s3b. The metre picks up the same pattern as the Wnal word-group of
s3a: jrw

j (see above). The two longa between rw
and D are
diYcult to analyse. This is the only case of a spondee in mid-verse.
Analyses apart, the similarity between O10s3b and O1s2 is undeniable:
O1s2 www
w

ww
ww
ww

k ^ e e D

O10s3b www



ww
ww
k e sp D
s4. See Part I, 6. C. 3 (er
e), 6. D (long mid-anceps). The second
long of the full e is resolved. The sequence w

wr is also found in
O2s6/7 and P5e1. The latter half is analysed as long anceps another
e. The combination er
e is rare, but not unparalleled:
O2s3

w

wr
wy

w
tw
k x e e e x e e
Bridge is observed between long anceps and the following e in most
of the repetitions. There is only one exception: v. 92 ( 88 Sn.) g
j
. This cut could be eliminated if another colometry were
adopted:
wr
rw

w
k ^ e
e e
But this must be rejected. First, there is no example of a long anceps
followed by resolution. And acephalous e is rarely followed by anceps
(the examples are P5s11, N6s1a).
s5. See Part I, 8. C. 5 (double e). Anceps e e is a very common
phrase, and there is another example where the phrase makes a verse
by itself:
 Olympian Ten 193
P5s1 w
w

w
k x e e
The anceps at O10s5 is short, except for one: v. 71 ( v. 68 Sn.)
(().
s6. For the correspondence between w
and

at aeolic base,
see above, strophe.
Verbal assonance: 34 l @ 50
 F.
e1a. See Part I, 8. C. 9 (e d).
e1b. See Part I, 6. C. 3 (er anceps at verse-end). As stated above
(Textual problems, 57b), the initial position is not long but anceps.
The alternative analysis (

rw
), based on a diVerent direction
of resolution and on an emendation (
), is unnecessary and
should be rejected. Hiatus/brevis is absent at the end of e1a; metrical
tendency supports the division neverthless, because (1) if the two
verses were combined, d would be followed by long anceps e, which
is extremely rare (
ww
w
w
is common,
ww


w

irregular17), and (2) there are some examples of short verses com-
posed only of e and link anceps, like
N6s6b
w t
k e

Verbal assonance: 79
a   101 

e2. See Part I, 6. B (^ e d, ^ e D); 8. C. 1 (e and D within a verse).
This strange form must be acephalous e and D. The same verse is
used at I8s9. The nearest parallel is
N6s5 ww

ww
ww
j ^d D
e3. See Part I, 6. C. 3 (er
e); 8. A. 6 (i) (palindrome); 8. C. 5
(double e). Irregular responsion between ww and
, which occurs
twice, makes the analysis of the verse diYcult. The Wrst case comes at
the beginning; N6e6/7 may be cited as a parallel (but see ad loc.).
Double short appears only at v. 59 ( v. 57 Sn.) I while long

17
One of the three cases of d
e is found at only one repetition (v. 57 ( 59 Sn.)

E) out of four in
N3e1a
ww
w
ww
a
w

j d w d x e

The others are e3 of this ode (v. 102 99 Sn.: `
) and e10; see the Wnal
section.
194 The Eighteen Majors
syllables Wll the position at all the other repetitions. The position is
most probably anceps from the structural point of view, since the
following phrase is e. Contrary to the general tendency, double short
is replaced by long only in ep. 5 in the latter half of the verse
( `
). This responsion may be explicable as a licence for a
proper noun, and perhaps the irregularity may be mitigated by the
fact that the regular form has been repeated four times before this
irregularity occurs. However, it would be an extravagant licence.
There is no evidence in all Greek poetry that the choriambic nucleus
of an aeolic colon is ever contracted, nor the double short of D. This
cannot, then, be contraction. Inevitably, the double short in question
should be taken as biceps like that at the beginning of the verse.18
Then the phrase is anceps d biceps e. However, there are still
further oddities; see the Wnal section.
e4/5. See Part I, 8. B. 4 (d ex), 8. C. 2 (double d). As the line-
numbering suggests, Boeckh divided this verse into two:
e4

ww


e5 ww
r


But this yields verse-end after
  at v. 16, and must be rejected.
In this verse, however, there is another place where word-end co-
incides perfectly:
e4/5

ww

ww
rwj

j
If the verse ends there, supposing brevis in longo at all the repetitions
(this is not implausible since brevis occurs with high frequency in this
ode; see examples cited above), and the two Wnal longs are shifted to
the beginning of the next verse, the following colometry, which
Turyn adopts, emerges:
[e4/5]

ww

ww
ww
k
d D
[e6]

j

ww
w
k sp tel
The new e4/5 sits well in the chart above (Epode), without introducing
the inverted e4/5. On the other hand, the new e6 is not good because

18
Dale (Collected Papers 75) accepts contraction in general, although ascertain-
able contraction is extremely rare. The only possible case is in a proper noun [ . . . ].
The contraction in `
 is clearly genuine.
 Olympian Ten 195
of the uncommon combination of sp tel. Certainly O9e4 (sp tel
sp) can be cited as a parallel, but its environment is too peculiar to
be comparable. Turyns colometry will be better rejected.19
Under the accepted colometry, parallels for e4/5 can be cited: for
xd d, P8e5; for d re (
ww
rw
), P5s10, P5e5, N6s6a, N7e1.
As stated above, the sequence of . . .
www

j is found at the end
of verse at s3a too, but the direction of resolution is diVerent. The
analysis there is
wr

j (er sp) and is applied to O2s4 as
well, while rw

j (re
) must be applied here and to N6s4a,
N6s6a, because the preceding
is not anceps but unambiguously
belongs to the preceding phrase (d or ^ D ). Word end does not
function for the decision because, unlike tragic poets, Pindar does
not avoid split resolution but puts frequently word end between two
shorts.
Verbal assonance: 17 K 

 H  61 f 



 .
e6. Telesillean with long anceps at half-base makes a verse by itself.
There are examples in P5s7b and I7e3 too.
e7. See Part I, 8. A. 3 (shorter verse). Anceps e 2 is a common
phrase, and makes a verse by itself in I7s5b too.
e8. See Part I, 8. A. 3 (shorter verse). Acephalous D is used already
in s1. There are two verses made up of only ^ D
(O9e3, O13s1), but
in consisting of ^ D without Wnal anceps our verse is unique.
e9. Hiatus attested at v. 65 ( v. 62 Sn.) and v. 109 ( v. 104 Sn.)
guarantees the verse-end, so that brevis must be accepted even
though all Wve verses end with a short syllable. This verse is related
to the telesillean of e6 as well as to the dactylic movement of the
previous verse (e8).
Verbal assonance: 21 (I) A
  65   109
; 87
e I  109 ,
.
e10. See Part I, 8. C. 2 (double d), 8. C. 5 (double e), 8. C. 9 (e d).
At v. 110 ( 105 Sn.) double short () corresponds with
long. Resolution of the latter long of the d is highly irregular. There
are other oddities too; see the next section.

19
Barrett (Collected Papers, 190) gives another ground for rejecting Turyns
colometry: it introduces frequent occurrences of a short vowel at verse-end (SVE).
196 The Eighteen Majors

Alternative Colometry for e3 and e10

This is the colometry on which the discussion in Part I is based:
e3 t
wr
wr l x
ww
y
w
k te e
dye
e10 w
wwjt l x
wr
w

ww
k wd
e e d
(here l
x means implicit long anceps)
This analysis of these two verses includes several irregularities. Of
these the Wrst three are unmistakable:

(i) the initial biceps in e3: at v. 59 (57 Sn.) double short

(I) corresponds with long in the other four repetitions;
(ii) the second biceps in e3: at v. 103 ( 99 Sn.) double short is
replaced by long ( `
).
(iii) the resolution of the second long of d in e10: at v. 110 ( v.
105 Sn.) its place is occupied by double short ().
The others are not self-evident; they concern long anceps:
(iv) d
e (d followed by long anceps e) in e10.
(v) d
e at the Wfth repetition in e3, with ` 
 retained.
(vi) cut between long anceps and the following true long at two
repetitions in e3; v. 81 ( 78 Sn.) 
 j   and v. 103 ( 99
Sn.)  j
E .
(vii) cut between long anceps and the following true long at one
repetition in e10: v. 88 ( 85 Sn.) 
a j
e .
For (iv) and (v): examples of d
e are extremely rare; cf. on e1b above.
For (vi) and (vii): Pindar tends to keep bridge after long anceps.
Where short anceps occurs in responsion with long in some repeti-
tions, there long anceps is always followed by bridge. Where anceps is
always long (implicit), there are in total nine exceptions among 74
repetitions of which one-third are constituted by the three cases in
O10e3 and e10.
The alternative colometry removes (i), (iii), (iv), (vi), (vii). In-
stead, uu
should be accepted, see below. For (ii) and (v), emend-
ation of `
 is desirable.
e3 t
rw
rw

ww
ww
w
k t
e e ibyc
e10a w
w
k we
e10b t

wr
w

ww
k t
e e d
 Olympian Ten 197
At all the repetitions in e10 there is coincidence of word-division
after the fourth syllable. In other words, if we suppose brevis in longo
there, we can divide e10 into two verses. Then (iii) the irregular
resolution of d and (iv) the mid-anceps are eliminated together with
(vii), the irregular cut, and instead t
emerges. When, interpret-
ation apart, this t
is acknowledged as a phrase and the direction
of the resolution of e is changed in e3, the long mid-anceps is
eliminated (vi). Note the similarity between e10b and e3: these verses
both start with spondee, which is replaced by r
in one repetition
each, and have two es after it. The Wnal phrase is d or its prolonged
form (ibycean). The curious phrase t
may be paralleled; the
outrageous phrase at N6e8 (t

ww
w
), which cannot be
glyconic, becomes explicable once t
is accepted. Perhaps it
should be, on the following grounds: (1) the long of e may be
resolved, and responsion between resolved and unresolved e is no
problem; (2) spondee may be interpreted as a variation of e; (3) there-
fore t
may be an e whose middle short has dropped out. This is
contrary to current metrical theory, indeed, but Pindar may have
experimented as with u
in the strophe (above).
Without `
, the latter half of e3 would be ibycean and is
repeated in e9 with the preceding anceps. 
 may be an
intrusion of a marginal explanatory gloss, referring to 


E at v. 2. The original text may have been, for example,
e
PF . Note that  is not the reading of the paradosis; it was
inserted by Moschopoulos (perhaps for metrical reasons). The asyn-
deton can be classiWed into the group of Hummel, Syntax 461
(asyndete, implication illocutoire et performativite); cf. Y
(full-stop is rather desirable after ).
The irregularities of these verses have been explained in various
ways and various improvements have been sought. In the nineteenth
century (Boeckh, Schneidewin, Mommsen, Christ), I at v. 59
(57 Sn.) seems to have been mistakenly scanned as
rw. These
critics give a diVerent metrical scheme to e3:

twr
wr

ww
y
w
k
Bergk2 knows the syllable is long and scans I- as
w
. At
the same time he maintains that
w
may be in responsion with


(quod si cui insolens videatur, meminerit solutionem, quam
198 The Eighteen Majors
statuunt,
t vel t
magis etiam oVensioni esse). Hermann
(Quinque Ol.) tries to expel not only this irregular resolution caused
by I but also the synaloepha of  in
 (37 36
Sn.). He unites e3 with e4 (note that brevis in longo occurs in quite
a few verses at the end of e3) and separates e4 from e5:
[e3/4]

wr
wr
wr w
ww
y

ww

j
This colometry is not necessarily good. Moreover it requires too
many emendations to be accepted (I forbear to cite them).
To expel oddities, Hermann proposes a diVerent colometry for e10
as well. He takes the last position of e9 as short (note brevis in longo at
all Wve repetitions) and unites e9 with the Wrst half of e10. A new
verse starts with the eighth position of e10:
[e9]

ww
ww
wr
ww


j
[e10] wr
w

ww
k
But not only brevis but hiatus occurs at the end of e9 at ep. 3 (65 62
Sn.) and at ep. 5 (109 104 Sn.). To avoid hiatus, Hermann (1)
introduces Bs  instead of , and (2) changes word-
order with large-scale emendations: u  ,  
 I. Besides, verse-end falls between
e and

 at ep. 4 (88 84 Sn.) according to this colometry.20

20
A diVerent colometry is also proposed by Bergk2. He sets a division after the
fourth position of e10:
[e9]

ww
ww
wr
w
k
[e10] u

wr
w

ww
k
Brevis in longo is posited at the end of the new [e9] in all the repetitions. His text of
ep. 5 is: . . .
 I k   . . .
 Olympian Thirteen 199

OLYMPIAN TH IRTEEN (STROPHE/ANTISTROPHE

V V. 1 5)

Five triads. Class II

O13s1 ww
ww

k ^ D

O13s2 a
w

w
ww

k xe ph (? xe
rdod
)
O13s3 w
wrw


w
k we2
e
O13s4

wtw
a
w
k
e2xe
O13s5 ww
w

w
ww

k ph (? re
^ dod
rdod
)
s1 B 75; s2 B 10, 46, 54; s3 H 11, B 69; s4 B 56, 70; s5 B 5, 13, 49, 57, 93
O13s6 u
w
ww


w


ww
ww

w
k
xrdod
e
D e
O13s7 u
ww
ww


ww
ww

k xD
D

O13s8
w

w


ww
k e e
d
O13e1

ww
ww

ww
ww


w

k
D D
e

O13e2
w


ww
ww
j e
D
O13e3
w
u
ww

w

k exd e

O13e4 tw


w


w
k e
e
e
O13e5
w


w


ww
ww
k e
e
D
O13e5b ww

w

tw
k ^ d e
e
O13e6a
w


w
j e
e
O13e6b
w

tw

k e
e

s1 1 (1) (

s2 2 (2) K
 r  l IE ,
s3 3 (3)  b 
, 
s4 4 (4) a O , "
s5 5 (5)
 A , I
s6 6 (6) K  A a P   , 


 I ,
s7 7 (7)  d  N, 
 I

,
s8 8 (8) 
E P 
e1 17 (17) ! *

 IE  :
 
$ .
e2 18 (18) d 
 K
e3 19 (19) f 

 
;
200 The Eighteen Majors
e4 20 (20)  a

 K  ,
e5 21a (21) j H E NH  
e5b 21b (22) K
 ; K b E 
 ,
e6a 22a (23) K  @ IE 
e6b 22b (23) P NE IH.
str. 2 2330 (24)(31) ant. 2 318 (32)(39) ep. 2 3944 (40)(46)
str. 3 4552 (47)(54) ant. 3 5360 (55)(62) ep. 3 616 (63)(69)
str. 4 6774 (70)(77) ant. 4 7582 (78)(85) ep. 4 838 (86)(92)
str. 5 8996 (93)(100) ant. 5 97104 (101)(108) ep. 5 105110 (109)(115)

In Olympian 13, the metre changes midway through the strophe/

antistrophe from so-called aeolic to the normal D/e. This is unique
among all the Pindaric epinikia. Comparable is Bacchylides 3, of which
the strophe is non-D/e but the epode is D/e. But there are signiWcant
diVerences; see further the excursus below. The changing-point of O13
is between s5 and s6. This analysis is accepted unanimously by every
scholar, whatever deWnition he may give to aeolic. The D/e section
does, however, begin with an asymmetrical phrase (heptasyllable, or
rather, anceps rdod x
w
ww
). The reversed dodrans is closely
associated with the two pherecrateans in s2and s5; on which see below.
The epode conforms metrically to the latter half of the strophe and is
made up purely of D/e. The metrical change after s6 is evident not
simply because the verses can be described using the Maasian symbols.
Other characteristics of D/e are also present. Verses become much
longer than before. Link ancipitia are abundantly employed and mostly
realized as long. Semantic break too occurs between s5 and s6 in eight
out of 10 repetitions; the exceptions are: str. 3 (49 51 Sn., but there
ends a participial clause), ant. 3 (57 59 Sn.), a case of enjambment.
Nevertheless, it must be recognized that the aeolic Xavour is
already not very strong even in the non-D/e part (s1s5). Asymmet-
rical phrases, the most distinctive feature of aeolic, are rather scarce.
Only three such phrases are interspersed among Wve verses: one in s2
and two in s5. The other three verses are to be analysed as freer D/e,
not a very diVerent metre from the normal D/e.
One of the aeolic phrases (s5) is unique in Pindar. It is an acephalous
dodrans (see Part I, 5. C); for another interpretation, see below on s5.
The other two are pherecrateans. Pherecratean is not a very common
phrase in Pindar. It can even be imagined that here in this context the
 Olympian Thirteen 201
pherecratean is used as a kind of variation of D/e phrases, like the initial
phrase of the D/e part at s6. It is possible to analyse the pherecratean
with base in the form
w as the reversed dodrans link anceps (cf.
Part I, 7. 5), while the initial of s6 as link anceps rdod:

w
ww


u
w
ww

The juxtaposition of pherecratean and heptasyllable is found in another
stanza-form. P10s1 is a pherecratean, and s2 is a heptasyllable.
The association of the D/e part with the non-D/e part is clear. Some
of the verses in the normal D/e part have similarities with freer D/e,
and even with aeolic. The most irregular is the asymmetrical phrase at
s6 mentioned above. In addition, some other phrases are introduced
which are fairly rare in the other typical D/e odes: choriamb, d, is used
twice (s8 and e3), and acephalous d, once (e5b). Moreover all these
three verses share another peculiarity: link anceps is often omitted.
Thus these verses would be easily accommodated even if they were
transferred into freer D/e surroundings:
O13s8
w

w


ww
k ee
d
O13e3
w
u
ww

w

k e x d e

O13e5b ww

w

tw
k ^d e
e

The transition is smooth from strophe to antistrophe and from

epode to strophe. The last verse of the strophe (s8), mentioned
above, is akin to freer D/e, and so is the last of the epode (e6b):
O13e6b
w

tw

k e
e

In short, it is certain that the metre of O13 changes between s5 and
s6, but the diVerence between the two metres is not fundamental.
RSS is 49.0% in s1s5, and 42.0% in s6s8 in the strophe, and 37.2%
in the epode.

Strophe (s1s5)
The non-D/e part has a simple, transparent structure. Of its Wve
verses, s3 and s4 are identical except for the realization of anceps
and the admission of resolution in e 2. They are sandwiched between
202 The Eighteen Majors
s2 and s5 which are also very similar to each other. The diVerence lies
in the initial position:
O13s2 a
w

w
ww

k xe ph
O13s5 ww
w

w
ww

k ^ dod ph
For ^ dod, see on the individual verse s5 below. The very Wrst line is
occupied by one, essential word ((
). Its metre, ^ D
,
is common to the normal D/e, but the brevity of this verse is typically
pertinent to non-D/e.

Textual problems (s1s5 only)

78 ( v. 81 Sn.; s4). Both Turyn and Snell follow Mommsen and
accept the paradosis I. Not only Mommsens predecessors
(Hermann, De dialecto, Boeckh, Schneidewin, Bergk2) but some
amongst his successors (Christ and Bowra) reject this form. Their
choice is I (Callierges; for this Homeric form, see LSJ I).
Both I and I introduce resolution at the position, which is
long at all the other repetitions. Exact responsion is unnecessary but
I (Triclinius) is supported by Hermann (Quinque Ol.).
79 ( v. 82 Sn.; s5). The paradosis

 is unmetrical;


Moschopoulos.
101 ( v. 105 Sn.; s5). The deletion of the unmetrical is possibly
based on the judgement of Byz. (deleted in N too).

Individual verses
s1. The same verse appears in O9e3 and, without the link anceps at
the Wnal position, twice in O10.
Verbal assonance: 1 (
  97 a  
 PH;
23 Pf I
  89 Pf I.
s2. Anceps e (iamb) often precedes an aeolic colon with full
base (in total 9 examples). The anceps of O13s2 is generally short.
Long is found only at v. 68 ( v. 71 Sn.) ( ). Two
positions are ambiguous (10 + ,; 46 ), but probably short
(All-but-One). Bridge between e and the pherecratean is irregularly
neglected. Word-end occurs after e as many as at Wve repetitions (vv.
2, 10, 24, 54, 68; and one more at 76 (K
  ). See Part I, 8. B. 6.
 Olympian Thirteen 203
s3. See Part I, 6. D (long mid-anceps); 8. C. 6 (e 2), Anceps e 2,
x
w
w
, is a well-paralleled phrase. Here it is followed by anceps
e. There is one verse made up of the same combination:
P11e5 w
w
w

rw
k
And one more with the order of the phrases inverted:
P8s7 x
w
x
w
w
k
Two ancipitia are realized regularly without any exception throughout
all the 10 repetitions: short (the Wrst, before e 2) and long (the second,
before e). Cut after the long anceps, which is highly irregular in the
eighteen majors, occurs at two repetitions: 3 

, j ,
25
, j F

 (Part I, 6. D). Is it a simple coincidence that
the anceps position is Wlled in both repetitions by a short vowel
followed by double consonant (note the verbal assonance under-
lined)? In addition, the resolution of the second long of e 2 is regu-
larized. The form
wrw
is quite common; in total there are
six examples; see Part I, 6. C. Of these P2s5 is very similar even in
the structure of the following verse in that anceps stands before and
after it:
P2s5

wrw
w
ww
k xe 2xd
Verbal assonance: 55 K
I A  99 
 K   i
s4. This verse is identical in structure with s3, but diVers in two
features. First, the regularization of anceps is reverse: the Wrst (before
e2) is always long and the second (before e) is basically short. The sole
exception is v. 78 ( v. 81 Sn.) . However,  too may
be scanned as short; cf. M. L. West on Hes. Th. 15. Secondly, the
resolution of e2 is found at only one repetition, v. 78 ( v. 81 Sn.)
I, see textual problems ad loc.
s5. See Part I, 5. G (acephaly and the Simonidean parallels), 8. A. 6
(a) (palindrome). The verse starts with ^ dod. This is the sole example
of acephalous aeolic in the eighteen majors.21 Acephalous dodrans is

21
Outside the eighteen majors, there is one:
Parth1s1 ww
w


ww
w
ww
j ^ dod
dwd
204 The Eighteen Majors
most certainly attested in the Simonidean poem (542 P) cited in Part
I. That poem includes two verses which start with acephalous
dodrans, and one of them (v. 6) is very similar to O13s5. Interest-
ingly, the next verse (v. 7) is similar to O13s2. The relation between
v. 6 and v. 7 of Simonides 542 P is the inverse of that between O13s2
and s5: ^ dod in v. 6 and s5 is replaced by xe in v. 7 and s2. Acephaly is
a common feature of O13: we Wnd ^ D in s1 and ^ d in e5b. Thus the
head of dodrans may be dropped. Another, but less probable, inter-
pretation is that this phrase starts with biceps, instead of anceps, at
the initial position, which is followed by e. Structural similarity of s5
with s2 may support this interpretation. The biceps may be compar-
able with that of Stesichorus. Pindar seems to have resorted to this
licence in O10e3. But the usage of biceps in lyric poetry is more
limited than is usually supposed. The initial of dramatic iambic
trimeter is another matter. In short, acephaly is preferable to biceps.
Verbal assonance: 5 I  93 - K
  27
 s  71  y.

Excursus B: Bacchylides 3
From a metrical point of view, O13 is comparable with B3. Like O13
the metre of B3 changes from non-D/e to normal D/e. It is true that
the strophe of B3 is non-D/e while the epode is D/e. The relation
between the two stanza-forms is, however, not so clear-cut, nor are
their structures so simple.
B3s1 a
w
a yw
w

k
B3s2 a
ww
ww
w

k
B3s3 a
ww
ww
w
w
w
a
ww
w

k
B3e1 u
ww
ww


w
u
w
k
B3e2
w
u
w


wt

w
k
B3e3 tw
u
w
u
w

w
u
w
k
Before analysing each stanza-form, it must be recognized that both
strophe and epode are very short. This brevity is alien to the eighteen
majors of Pindar. Comparable is O5, whose brevity is one of the
reasons why its authenticity is questioned. Each stanza-form consists
of only three verses though they are repeated as many as seven times.
 Excursus B: Bacchylides 3 205
The third verse is longer than the other two, both in the strophe and
in the epode, especially in the former. It is like those of the Sapphic
stanza or the Alcaic stanza (Wests aaA pattern; their traditional
four-line layout does not represent their structures).22
It is inappropriate to call the metre of the strophe as a whole aeolic;
at least s1 is not aeolic, having no double short in it. For this reason
s1 is similar to e2 and e3. Another similarity between the strophe and
the epode is dactylic sequence. The Wrst half of e1 is identical with the
Wrst eight positions of s2 and s3. This phrase is anceps D, and this
D is the only D in the epode. The epode is usually analysed as normal
D/e and the analysis is not wrong, but after this D all the rest is
exclusively made up of e (and link anceps). I shall come back to this
topic later.
s1 is usually analysed as iambic trimeter catalectic (ia ia ba),
as if it were the same verse as in drama. Such an analysis in Pindaric
metre should be rejected (see Part I, 5. A. 5). The bacchiac, whose
last position is possibly triseme created by catalexis, is alien to Pindar.
Admittedly, Bacchylides is not Pindar. It is, nevertheless, reasonable
to think that the last position of the apparent bacchiac in s1 is not
triseme but anceps. The reason is as follows.
The ending of s1 is identical with that of s2: . . .
w

k. It is
natural to take both Wnal positions to be of the same value. On the
other hand, s2 is repeated in the Wrst half of s3. Here in s3 the
position corresponding with the Wnal of s2 is realized as short, but
its position is deWnitely anceps, for its left short is really short and so
is its right. Then, if the Wnal position of s2 is anceps, so is that of s1.
The situation is similar to the Sapphic stanza:

w
x
ww
w

k

w
x
ww
w

k

w
x
ww
w
x
ww

k
Here too the Wnal position of the Wrst verse and that of the second are
most probably anceps, like the corresponding position in the third
verse.
Pindar once uses a seeming iambic trimeter catalectic. It is not
found in the eighteen majors but in the four minors:

22
Itsumi, Whats in a Line?
206 The Eighteen Majors
P7s1

w
wrwrw

k
e 4

But B3s1 is diVerent from P7s1, in that its Wfth position is not
invariably short; it is therefore preferable to describe it as x e x e2
.
It resembles rather
P8s7 a
w
a
w
w
k x e x e2
With the additional Wnal anceps attached, B3s1 is one position longer
than P8s7.
In the next two verses, B3s2 and s3, the
w
ayw
of s1 is
replaced by
ww
ww
. In other words, the two verses include so-
called expanded aeolic. It has been argued that these expanded
aeolic phrases are, in fact, freer D/e (Part I, 7. 6). This is true in B3
too. Conventionally, I use the notation ibyc(ean) ( dodd) for

ww
ww
w
(dodd makes it easier to recognize the structure at
a glance). Then the verses are analysed thus:
B3s2 xibyc
k
B3s3 xibyc w e x dod
k
The phrase
ww
ww
w
(ibyc dodd) and its shorter form

ww
w
(dod) are used as if they were variations of D. The
relation of ibyc to dod is the same as that of D to D. Like D or
D they are preceded and followed by link anceps. And e intervenes
between them in s3 too. The strophe of B3 is, as a whole, freer D/e.
As described above, the epode is totally made up of e except for
the D at the beginning. In this sense, it is similar to O2. At the same
time, there is a diVerence from freer D/e, which Maasian symbols
clarify:
e1 x D
e x ek
e2 e x e
e
ek
e3 e x e x e e x ek
Link anceps is regularly used except between the third e and the
fourth in e3. This is certainly the usual manner of normal D/e.
Conversely, link anceps is rare in freer D/e. But there is still a
diVerence. The number of e in e2 (four) and e3 (Wve) is not typical
even in normal D/e. Of all the verses in Pindars normal D/e odes (in
total 321), only seven verses are made up of four es without D,
and only one verse of Wve es without D (O12e7, 1r); see Part III,
 Excursus B: Bacchylides 3 207
A. Incidentally, his maximum number of es without D is 6 (N10s6).
Although link is often absent and this absence is a crucial diVerence,
some of the verses of O2 includes as many es as B3:
O2s3

w

wr
wy

w
tw
k
e e e
e e
O2s5

wr
w
rw
t w
k
e e e e
208 The Eighteen Majors

P Y T H I A N T WO

Four triads. Strophe/antistrophe Class III: Epode Class I (?Class III)

(1)
P2s1 rwrw
w
w
wrw
k e6
P2s2 www
ww

u
ww
w


ww
w
k rdod gl tel
(2)
P2s3

ww
ww
ww

w
k
D e
P2s4 ww
ww
ww
w
ww
ww
w
k wwibyc tel ( teldtel)

[or ww
ww
ww
w
ww
ww
w
k ^ D wibyc (see
below)]
(3)
P2s5

wrw
w
ww
k
e 2wd
P2s6

wrw
w
ww
ww
k
e 2wD
(4)
P2s7 www
ww
rw

w
k rdod e e
P2s8

ww
w

w
w
w
ww
w

k tel e 2hag
P2e1a www
ww
w

w
ww
j gl rdod
P2e1b w

ww
w

w
ww
k gl rdod
P2e2 w

ww
w
www
ww
k gl rdod
P2e3 w

ww
w
www
ww

w
w
j gl rdod e 2
P2e4 www
ww
w

w
k gl e
P2e5 a
w

a
ww
w

w
k xe gl e
P2e6 w

ww

w
j ^e d e
P2e7

w
ww
k hepta
P2e8 w

w
ww
w

w
ww
w

k wil 2 hipp
s1H 9; s2 H 2, 82; B 10, 26, 34, 58, 74; s3 H 59, 75; s4 H 84, B 4, 12, 28, 76; s5 BH 61; s6
H 78; s7 BH 31; s8 B 40; e1b H 65b, B 89b; e2 B 66; e4 B 44; e5 B 93; e7 H 95; e8 H 24

For the analysis and the notation of s4 (wibyc ), see Part I, 7. 6

(expanded aeolic), and also Appendix B.
s1 1 (1) 
 t #
, 

s2 2 (2)  @ , IH

  A
 ,
s3 3 (3) h  A
A I
e A 
s4 4 (4)   I  K ,
s5 5 (5) P
 ! "  K v 
s6 6 (6)  I  
 ,
 Pythian Two 209
s7 7 (7)
   , v P 
s8 8 (8)  IE K d
 K

 .
e1a 17a (17)     b

e1b 17b 
  Id  O

e2 18 (18) b  , t 
E, 
e 
e3 19 (19) d
 I
,
 
 K I

e4 20 (20) a a  E I
e5 21 (21) H  KE "  d F E
e6 22 (22)  K
 H
e7 23 (23)
 A 
e8 24 (24) e P IE IE K
 .
Like other odes dedicated to Sicilian tyrants and placed early in the
collections, Pythian 2 is very long. Not only is the triad repeated as
many as four times, but also both the strophe and the epode are long
in themselves. The number of positions amounts to 124 in the
strophe and 122 in the epode, making them respectively the third
and the fourth longest among the eighteen majors.
The style of the strophe is majestic, and unambiguously classiWed as
Class III. There are many peculiarities in each verse (for the details, see
below). The epode does not lack majesty either, but its structure is
based on regular repetition of identical aeolic phrases, and feels
diVerent from the strophe. The epode is located somewhere between
Class III and Class I (aeolics). I classify it tentatively as Class I. But its
similarity to the strophe is undeniable; see the section below. Frequent
resolution makes RSS high, especially in the strophe: 58.3% (the
second highest). In the epode it is 52.4%. This Wgure is not exceed-
ingly high among all the stanza-forms, but deWnitely high for Class I.
A number of verses of P2s and P2e8 (the Wnal of the epode)
conspicuously resemble those in Paean 6; see Appendix B.

Strophe
The strophe has the most magniWcent structure of all the non D/e
stanza-forms. The only comparable one is the strophe of Olympian 1.
It is remarkable that P2s and O1s share two impressive but rare types
of verse. Both stanza-forms include (i) a verse composed solely of
a long sequence in single-short movement (P2s1/O1s8), and (ii) a
verse/verses including a long sequence in double-short movement
210 The Eighteen Majors
(P2s3, P2s4/O1s2). Moreover, (iii) a verse made up of multiple aeolic
phrases (P2s2/O1s1) deserves notice. There is a strong tonal contrast
by nature between (i) and (ii), and the contrast is more signiWcant in
P2s than in O1s because (i) and (ii) occur nearer together. P2s starts
with an extraordinarily long single-short movement (the longest of
the eighteen majors), which is diversiWed by resolution. The six
continuous short syllables at the opening are especially striking.
The eVect is enhanced by a single word which occupies the very
beginning of the odes (
 ). Resolutions, which give
bright, dazzling eVects, are also to be found as well here and there
in other verses: s2 and s7 start with tribrach; s5 and s6 (the latter is an
enlarged version of the former) have resolution at the same position.
As for (ii), two long dactylic sequences are incorporated in P2s: s3
(h  A
A I
e A) and s4 ( 
I). As in the Wve dactylic verses of O9e (see ad loc.), word-
end between two shorts within these sequences is, in general,
avoided. As a result the rhythm feels square. We may imagine that
a kind of solemnity was given by this movement. The former is
analysed as anceps D , which is used in O1s8 too, but fairly
uncommon. The latter is analysed as ^ D (or a part of teld), a very
unusual phrase (see on individual verses below, s7).
A verse made up of three aeolic phrases (iii) is rare in the eighteen
majors, and P2s2 is one out of only four examples (the Wrst verse of
O1 is also aeolic, and this combination of two aeolic phrases is rare in
Pindar: the priapean dicolon). The three phrases in P2s2 are designed
to be diVerent from each other: no base (rdod)/full base (gl)/half base
(tel); reversed dodrans (rdod)/straightforward dodrans (gl, tel). In
O1s1 there remains a possibility that the pherecratean and the
glyconic should be diVerently interpreted (see Part I, 7. 4), but in
P2s2 the glyconic is unmistakably identiWed by the free responsion at
the aeolic base (see below).
A long single-short sequence followed by d is a characteristic of
Class III stanzas. Here in P2 an additional technique is signiWcant:
expansion of d (s5) into D (s6). Interestingly, a similar compression
is applied to the transition from O1s6 to s7, but in reverse:
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
O1s7
w
w
w
w
ww

w
k e3 w d e
 Pythian Two 211
In general, the structure of the strophe of P2 is more complicated
than that of the strophe of O1, which is divided into four clearly
discernible sections. P2s is divided into four as well, but the relation
between each section is more subtle. Each section is made up of two
verses. In 1 (s12) there are two motifs which will be developed later.
The two verses are completely diVerent. s1 is e 6. It is as a whole
symmetrical, although frequent resolutions occur in it. By contrast,
s2 is made of three aeolic phrases whose essence is asymmetry. When
we move to 2 (s34), we notice at once that the main feature is an
important type of metre which is diVerent from both s1 and s2 and
which has been absent up to now. This is double-short movement,
and most of the second part is based on it. The two verses in 3 (s5
6) resemble each other, to a greater degree than those in 2. s5 itself is
a typical verse of the Class III stanza-forms in which single-short
movement ends with a double short. s6 is almost the repetition of s5.
But the diVerence between the two is signiWcant. By the expansion of
the Wnal phrase (d ! D), s6 has a resemblance to the dactylics in 2
(s34). At the same time, the initial part of s5 and s6 has a similarity
with s1:
s1 rwrw
w
w
wrw
k
s5

wrw
w
ww
k
Thus various metres merge and develop. Then comes the Wnal
section (s78). The two verses of 4 are basically a recapitulation of
s2, as well as an enlargement. The initial phrase of s7 is identical with
that of s2. This phrase appears at Wrst sight to be repeated once more
in the same form, but, by changing a short into a long ( . . . ww

! . . .
w
), it turns into two es. The Wnal e is not new; it is a
reminiscence of s3. Again, s8 starts with an aeolic phrase, which is
identical with the third ( Wnal) phrase of s2, a telesillean. Then
follows a single-short movement, which resembles s5 and s6, but
without resolution. If it ended without prolongation beyond

ww
, the latter half of s8 would be identical to s5:


ww
w

w
w
w
ww
w

k
But the sequence has an appendage w

, so that the whole
sequence shares a common feature with aeolic phrases. For example,
a glyconic, the second phrase of s2, can be detected in it:
212 The Eighteen Majors


ww
w

w
w
w
ww
w

k
The appendage w

makes s8 end pendent. s8 is the only verse in
the strophe that has pendent ending. Compare the last verse of the
epode (e8). Interestingly, O1e7, the last verse of O1, ends with
w

. Pindar may have intended a special eVect and reserved it to
round oV a stanza. These aeolic phrases, hagesichorean (O2s8),
hipponactean (O2e8), aristophanean (O1e7), all of which have
w

at the end, are highly unusual.

Epode
The epode is composed of aeolic phrases, and its composition is one
of the most straightforward in all the Pindaric non D/e stanza-forms.
The generative procedure from the preceding verse to the next is
evident in each case. In short, the basic principle is repetition ac-
companied by a slight variation. The Wrst three verses (e1a, e1b, e2)
are the same in structure (gl rdod). e3 starts in the same way, but is
extended by suYxed e 2. The glyconic is repeated again at e4, but in
this verse the reversed dodrans is dropped, and e 2 is shortened to e.
Then the new combination, glyconic e, is repeated in the next
verse, e5, but with preceding anceps e. The next verse, e6, is an
abbreviated version of e5: the initial xe is shortened into ^ e, and the
glyconic, into d. At the same time, the combination ^ e d is equal to
three-quarters of the glyconic starting with w
, which is used
repeatedly for three successive verses from e1b to e3. The Wnal e is
the same as in e4 and e5. The next verse, e7, is composed only of a
heptasyllable. This can be analysed as anceps reversed dodrans,
which is repeated four times successively from e1a to e3 (for verse-
by-verse construction, see Part I, 8. A. 2). When the heptasyllable is
prolonged both forwards and backwards, a decasyllabic aeolic phrase
(wil 2) results. This is the Wrst half of the Wnal verse, e8. Its aeolic
base w
has been much used already. The decasyllable incorporates a
glyconic, and the glyconic is repeated once more in the latter half of
the verse with one additional position (hipponactean), which makes
the last verse the only pendent verse in the epode.
One of the peculiarities of the aeolic base of this epode is frequent
w
. It comes at the opening of Wve verses out of nine (including e6,
 Pythian Two 213
where it is analysed as ^ e). The opposite form,
w, which is the
commonest in the Pindaric odes, is never used. When w
is absent, a
neutral base, www, is used (e1a, e4). wwwgl echoes wwwrdod in two verses
of the strophe (s2, s7). At the same time, wwwrdod is used after a
glyconic (e2, e3). The initials of glyconics and reversed dodrantes
make a contrast: www cannot stand at opening of both of glyconic and
reversed dodrans within the same verse. In other words, it is
restricted to one or the other. The aeolic base of the glyconic in the
middle of e5 is designated as
x. It is in fact
w;

is a licence
for a proper noun (21 "). There are two verses (e5, e7) which do
not start with w
, but a true long is avoided at the beginning in these
two, as in all the other verses. They start with anceps.

Textual problems
7 (s7). The paradosis A is unmetrical; v Hermann, Notae.
12 (s4). The paradosis

 is unmetrical;

 Heyne.
16 (s8).The paradosis gives  , which makes the verse


ww
w

w
x
w
ww
w

k tel e hepta 3
This analysis is not impossible (cf. O9e8, in which e precedes a
heptasyllable whose half-base is x), and Liberman accepts the anceps.
But a better and well-paralleled scheme is obtained by a slight
change; so E (Moschopulos). For the similarity of s8 to
s5, see above, strophe. See also below on s8.
28 (s4). The paradosis K (and K
) is unmetrical; N Triclinius.
30 (s6). , like E above, is as one of the best emend-
ations of Moschopoulos (for an evaluation of whom cf. Gunther, Ein
neuer metrischer Traktat).  is metrically impossible.
31 (s7).  at the verse-end is deleted by Hermann (Notae).
41b (e1b). The paradosis I produces resolution at the Wnal
position of the glyconic. It is not impossible, but I (Moscho-
poulos) is deWnitely preferable.
65 (e1a). The paradosis I
 is unmetrical;  Hermann,
Notae. Gentilis I
 is impossible. Even if
w
ww
were
analysed as
ww
(cf. Part I, 5. A. 4), it would not respond
with w

ww
. Full base is not free in Pindar. There is no example
of responsion between
w and w
.
214 The Eighteen Majors
66 (e2): The paradosis
d

  is one short syllable too
short. Following Boeckhs
d 

 , Bergk proposed 

d

 , based on scholia.
72 (e8). The paradosis   (most MSS;   B) is unmet-
rical;  Triclinius.
76 (s4): of A is treated as long, as in the dactylic hexam-
eter. Christ cites Theognis 324 
.
79 (s7). The paradosis K
 is unmetrical; N
 Byz.,  

Schroeder.
82 (s2): The paradosis gives unmetrical . Both Turyn and Snell
adopt Heynes proposal . Boeckhs Ia (accusative of I) is,
according to Schneidewin, approved by Hermann, and adopted by
Mommsen, Christ, et al. LSJ records it (3. curve, bending . . . hence
. . . in the sense of crooked arts, deceit). Bergk2 proposes  (i.e.
rete), alii alia. Gentili, by deviating from the consensus on metre
among modern scholarship, keeps . According to his colometry,
a pherecratean which ends with a short syllable can be followed by a
glyconic which starts with an iamb (
u
ww
w u

ww
w
)!
89 (e1). b  s is one syllable too long.  was deleted by
Triclinius.

Individual verses
s1. See Part I, 8. C. 6 (e 6 and e 5). The verse consisting only of e 6 is
unique; there are no other examples. Comparable is the single-short
movement of O1s6, which is of the same length but ends in d. As
stated above, the consecutive six short syllables in the initial positions
of an ode are striking. However, if we search for parallels, not just at
the beginning of an ode but at that of any verse, two other verses are
found to start with several shorts: O1s8 (7 shorts), P5s4 (6 shorts); cf.
Part III, D. In P2s1, there is another resolution near the verse-end.
The ending of
wrw
k has parallels: N3s3 (analysable as e 2), N7s5
(analysable as a part of e 3).
s2. See Part I, 5. E. 2 (resolution of reversed dodrans); 8, A. 3 (i)
(longer verse). This verse has characteristics of typical aeolic phrases:
(i) multiple phrases; (ii) dovetailing; (iii) Xexibility of aeolic base. It
is composed of three aeolic phrases, rdod gl tel, comprising in
total 21 positions. Comparable is
 Pythian Two 215
N2s4 www
ww
w



ww
w

u
ww

k gl gl ph
which is of 22 positions in length and has some structural similarities
(note the tribrach opening). Word-end tends to be avoided at junc-
tions both between the reversed dodrans and the glyconic (except for
v. 26 , j e) and between the glyconic and the telesillean
(avoided without exception). Instead, overlap by one position
(dovetailing) is a general tendency from the glyconic to the tele-
sillean. Dovetailing happens at seven repetitions out of eight (the
exception is v. 50 jE). Reversed dodrans starting with www
at verse-beginning is common (5 examples in total). Thus in the
other examples the resolved long of s2 is not in responsion with
unresolved, and the initial anceps of the telesillean is regularly long.
In contrast, the aeolic base of the glyconic retains Xexibility:


and
w are used equally (4 times each).23 This glyconic provides
strong counterevidence against the hypothesis that the phrase

w
ww
w
in Class III can be diVerentiated from the glyconic
as an aeolic phrase.
s3. See Part I, 8. C. 1 (e and D within a verse), 8. C. 8 (D e). The
nearest parallel is:
N6e3

ww
ww
ww

ww
j
D d
The diVerence lies in the Wnal phrase. In our verse e is used instead of
d. There are in total Wve examples of D , in four of which a long
anceps precedes D as here. Phrase D, which is more frequently used
than D , is combined with e without link anceps in freer D/e (N6e1)
and more rarely in the normal D/e. Our verse is its prolonged variant
and is not strange. Avoidance of word-end is strictly observed at the
joint of D and e; see Part I, 8. C. 1.
s4. See Part I, 8. A. 6 (b) (palindrome), 8. A. 7 (repetition within a
verse). This verse is one of the six expanded aeolics discussed at
length in Part I, 7. 6. The following colometry is adopted in the
main body of Part I, the latter half of P2s4 being counted as one of
examples of the telesillean whose half-base is ww:

23
Or 5 long /3 short, . . . -A: c being scanned as


.
216 The Eighteen Majors
(1) ww
ww
ww
w
jww
ww
w
k teldtel ( wwibyc tel )
However, if the argument that expanded aeolics are not aeolics but,
in fact, a kind of freer D/e should be convincing, a diVerent colo-
metry, which strictly follows Rule 5 (Part I, 3), would be admissible
and possibly ought to be accepted:

(2) ww
ww
ww
jw
ww
ww
w
k ^D wibyc
No strong tendency can be observed concerning word-boundary in
the eight repetitions.
Leaving analysis and appellations aside, the similarity to N3e4 is
striking:
N3e4 ww
ww
w
ww
ww
w

w
k tel tel e or ^ D w ibyc e
The Wrst half is expanded by another dactyl in our verse. Also
similar is
N6s3 www
ww
www
ww
ww
ww
j rdod ribyc
Outside the eighteen majors, teld (or wwibyc) is used as a verse by
itself:
Pae4s1 ww
ww
ww
w
j
Short anceps between two d is paralleled:
O1e5 ww
w
ww

w
w
k ^ dwd e2
N3e1a
ww
w
ww
a
w

j dwdxe

s5. See Part I, 8. C. 3 (enwd). This verse is a shorter form of those
sequences of single-short movement that turn into d. The sequence
in O1s7 is one position longer, and, although it does not turn into d,
O13s4 is the nearest parallel at the same time:
O1s7
w
w
w
w
ww

w
k e 3 wd e
P2s5

wrw
w
ww
k
e 2 wd
O13s4

wrw
a
w
k
e 2xe
s6. See Part I, 8. C. 1 (e and D within a verse). 8. C. 3 (wd). The d in
the preceding verse is expanded into D; one of the clearest cases of
internal expansion according to Wests classiWcation of the genera-
tive procedures (645). Our verse is unique in that D appears as the
 Pythian Two 217
second unit after link anceps; there is no other example of this
feature.
s7. See Part I, 5. E. 2 (resolution of reversed dodrans); 8. C. 5
(double e). The alternative analysis of this verse, glyconic long
anceps e, should certainly be rejected. A glyconic whose last pos-
ition is resolved is attested four times elsewhere, but in all these cases
the glyconic is followed not by an anceps but by a true long. In fact
there is no example of glyconic (or dodrans) followed by long anceps.
Double e after an aeolic phrase is not paralleled either, but these two
verses are related:
P5s10
w

ww
rw

w
k e d e e
O10s1 ww
ww

w
tw
k ^D e e

Note especially P5s10, in which the same position (the Wrst long of e
after d) is resolved as our verse.
s8. See Part I, 8. A. 3 (ii) (b) (longer verse); 8. C. 3 (e 2w half-base).
A telesillean precedes the single-short movement. The nearest parallel
is I8s3:
I8s3
w
w
ww

w
wr w
ww
w
k wil e 2tel
P2s8

ww
w

w
w
w
ww
w

k tel e 2hag
Both P2s8 and I8s3 start with an aeolic phrase which is followed by
e 2. They are almost identical in structure and the number of posi-
tions is the same. As we have seen above, without the appendage
w

the latter part of P2s8 would be e 2wd. There are in total three
verses in which an aeolic phrase is used instead of d after a long
single-short movement (the third is I8s7). Hagesichorean is very rare
in any context. There is no other example in the eighteen majors.24
e1a3. See Part I, 5. E. 2 (resolution of reversed dodrans); 8. A. 6
(a) (palindrome); 8. C. 6 (en as a substitute for e). There is no hiatus
or brevis in longo after e1a and after e3, but the repeated structure
guarantees the division. The sequence incorporated in these verses,
dodrans reversed dodrans,

ww
w

w
ww

24
There will be another if P10s5 is analysed not as ^ e wil 3 but as ^ e e
hag; see ad loc.
218 The Eighteen Majors
makes a type of palindrome. Bridge between glyconic and reversed
dodrans is observed when reversed dodrans starts with
w (e1, e1b),
but is not observed when it begins with www (e2, e3). Interestingly,
there are no other examples in the eighteen majors of the combin-
ation of aeolic with 2 ending (like glyconic) and reversed dodrans.
The e 2 at verse-end in e3 following an aeolic phrase is common (12 in
total), including three examples of rdod e 2 (the others are N3e3,
P10s4). Bridge between rdod and e 2 is not the rule, being observed at
only two repetitions (43, 67).
e4. Glyconic e is a very common phrase (10 in total). Though
bridge between glyconic and e is a general tendency (see Part I, 8. B.
6), cut is admitted once at v. 68 (e j

).
e5. See Part I, 5. C. 1 (proper noun involving aeolic full base of the
shape
x). Anceps e (iambic metron) often precedes glyconic
and other aeolic cola starting with full base (9 examples in total; see
Part I, 8. B. 5). There are two other verses with the identical structure:
N3s4 w
w

w
ww
w

w
k xe gl e
P5s3 w
wr
w
ww
w
rw
k xe gl e
The aeolic full base of the glyconic is mostly
w, as stated above.
e6. See Part I, 5. A. 4 (analysis not as
ww
but as ^ e d); 6. B
(^ e d, ^ e D); 8. C. 8 (d e). There are two other verses (P5e5,
P10s2b) of the identical structure: ^ e d e (and one of ^ e d e 3:
O1e2).
e7. A verse made up only of hepta is not rare; there are in total Wve
examples.
e8. See Part I, 8. A. 6 (a, e) (palindrome); 8. A. 7 (repetition within
a verse). There are four examples of wil 2, of which two, P5s8 and
P2e8, have w
as aeolic base, and it is possible to analyse them as
^ e gl (see below; in P5s8, wil 2 makes a verse by itself). In our
verse, bridge is strictly observed between wil 2 and the following
hipponactean without exception. Hipponactean is a rare phrase in
Pindar: there are only two examples, including our verse (the other is
N7s8). Both are located at the end of their stanzas.
Alternatively our verse is analysable as
w

w
ww
wil
w

w
ww
w

k wil 3
 Pythian Two 219
This analysis has its merit in its recognition of the repetitive struc-
ture. However, I reject it because (1) to cut at
ww
jw
is against
the rules established in Part I, and, more importantly, (2) the analysis
w

w
ww
w

w
ww
w

k ^e gl gl x
becomes more meaningful once
w
ww
w
is recognized as a
non-aeolic unit in some stanza-forms (see the Wnal section of Part I).
220 The Eighteen Majors

P Y T H IA N F I V E

Four triads. Class II (strophe/antistrophe); Class III (epode)

(1)
P5s1 w
w

w
k we e
P5s2 w
wt
w
w
ww
k we wil
P5s3 w
wr
w
ww
w
rw
k we gl e
(2)
P5s4 wr rw
k ^e e
P5s5
wy
ww
k ed
P5s6 w

w
k ^e e
P5s7a


k
sp
(3)
P5s7b

ww
w
k tel
P5s8 w

w
ww
w
k wil 2 (? ^ e gl)
(4)
P5s9 w
w

w

w
k we e e
P5s10
w

ww
rw

w
j edee
P5s11 w


w
w
w
k ^e
e 3

P5e1 w

wr
w
w
ww
k e wil ^e

P5e2 w

ww
w
w
w

w
ww
j
gl (? ^ e dod)
we rdod
P5e3
w
ww
rw
ww
j rdod rdod
P5e4 w
w

w
ww

j we ph (? dod
)
P5e5 w

ww
rw
j ^e d e
P5e6
ww
rw
ww

w
j d rdod e
P5e7a w
tw
j ^e e
P5e7b w tw

j we

P5e8
w

w
k ee
P5e9 ta
ww
w

ww
w yw

w
k gl dwe e
s1 H 74, 94; B 94; s2 H 44, 75; s3 H 34, 65; s4 H 97; B 77; s5 H 78; s6 B 37; s7 H 49, 80;
B 7, 18, 38, 49, 80, 111; s7b H 69b; B 7b, 69b, 100b, 111b; s8 H 8, 70, 112; B 19, 81,
112; s9 H 20, 51; B9, 20, 51, 71, 102, 113; s10 ? H 72 see below; s11 B 42, 73, 84, 115; e1
H 54; e8 H 61, B 30; e9 H 62

s1 1 (1) !
F P ,

s2 2 (2)   I A  A
s3 3 (3)  Ic

 Pe I

s4 4 (4)
 
.

 Pythian Five 221
s5 5 (5) t  ,
s6 6 (6)    A
s7a 7a (7) NH
s7b 7b IA 

s8 8 (8) f P 
s9 9 (9)  

 :
s10 10 (10) P n a  Z 

s11 11 (11)  
 .
e1 23 (23) 
  H  c ,
e2 24 (24)
 f Id A
  I,
e3 25 (25)
d b e Y $
,
e4 26 (26) E b
  
e5 27 (27) n P a
 
e6 28 (28) O   A
e7a 29a (29a) I 
e7b 29b (29b) 
e8 30 (30) I I

e9 31 (31) o  d  I E  ,

Pythian 5 celebrates the same victory as Pythian 4. In contrast with

P4, however, which is a long D/e ode and by far the longest of the
Pindaric epinikia, the metre of P5 lacks grandeur and, indeed, creates
a light tone. The eVect is caused by (i) frequent cretics (e) and
avoidance of dactyls, (ii) frequent resolutions, (iii) short verses.
Freer D/e is the basic metre of this ode. More correctly, e is dominant
while d appears sporadically (in total only Wve: at s5, s10; e5, e6, e9).
There are a considerable number of verses which are made up only of
e and anceps (s1, s4, s6, s9, s11; e7a, e7b, e8). Indeed, there are several
parallels for these verses in O2, which is made up almost completely
of e. There is no dactylic sequence at all.
Aeolic phrases are intermixed both in the strophe and in the
epode: four phrases in four verses in the strophe (s2, s3, s7b, s8)
and eight phrases in six verses in the epode (e1, e2, e3, e4, e6, e9). The
occasional intrusion of aeolic phrases is not rare in the odes of Class
II (freer D/e), and the strophe can be classiWed with certainty as
belonging to Class II. But as for the epode, I am tempted to suggest
Class III (amalgamated), because of the noticeable presence of the
reversed dodrans (4 times) and its variant, pherecratean (see below).
Nevertheless, the similarity of the two stanza-forms is undeniable.
222 The Eighteen Majors
There are some common features, for example, a heavy use of
acephalous e (w
). RSS is 51.8% (the strophe) and 52.6% (the
epode). ClassiWcation is not always self-evident. This epode is a
borderline case.
The aeolic base (full base) of three aeolic phrases (s8, e2, e5) located
at the beginning of the verse is Wlled with w
. This is equal to ^ e, which
frequently appears in this ode (5 times in total). The two are not always
easy to distinguish. The ambiguity of w
sometimes makes the colo-
metry uncertain. Compare these two verses, which both have 10 posi-
tions (the diVerence lies in the central 3 positions, marked by [ ]):
P5s8 w

[w
w]w
w
k wil 2
P5s11 w

[
w
]w
w
k ^ e
e w e
Thus it is not impossible to analyse the Wrst, s8, as ^ e and glyconic
(Part I, 7. 4; for wil 2, see below).
Another ambiguity is found in wilamowitziana starting with
w
(Part I, 7. 3). That forming the latter half of
P5e1 w

wr
w
w
ww
k ^e e wil
can also be analysed as e short anceps d. This ambiguity is
exploited in s2 in a sophisticated manner (see below). Moreover,
resolution adds further ambiguity. For example, the sequence

ww
www
ww

which occurs twice in e3 and e6, can be analysed in two ways,
depending on how www is treated.
There is yet another diYculty. No clear indications of verse-end
are found in eight successive verses from e2 to e8. Hiatus/brevis are
totally absent. For the arguments for placing verse-end in these
verses, see below on each verse.
Verses are in general short, especially in the strophe (here, verse-
end is clearly guaranteed in most cases). Half of the 12 verses are
made up of fewer than eight positions: s1 (7), s4 (5), s5 (7), s6 (5),
s7a (3), s7b (7); cf. three verses in the epode: e7a (5), e7b (5), e8 (6).
The average length is 8.8, the shortest of all the Pindaric odes. In the
epode the verses are longer (11.6).
Pindar generally prefers blunt ending to pendent (see Part I, 8.
A. 4), but the preference is extreme in this ode. Except for e4 and e7b,
 Pythian Five 223
all the verses end blunt; 14 with
w
(or rw
), Wve with
ww

(s7a is exceptional, see below).

Strophe
The strophe is divided metrically into four sections. In 1 (s13)
verses are gradually augmented. In spite of diVerent terminology, the
Wrst seven positions of s2 are identical with s1, if we ignore the
resolution at the fourth position that appears at only one of the
repetitions (see below). The same sequence is repeated again in the
next (s3), but wilamowitzianum in s2 is replaced by glyconic, the
reversed phrase of the same length, and the verse is prolonged
furtherly by suYx e. Both the wilamowitzianum and the glyconic
start with
w and end with w
. The diVerence between them lies in
the order of single and double short:

w
w
ww
k

w
ww
w
k
2 (s47a) consists of four very short verses. Despite the brevity, the
rhythm of the Wrst (s4), is striking because of the six successive short
syllables (cf. Part III, F). After a slightly longer verse (s5), the same phrase
returns, but this time without any resolution (s6). The Wnal verse of this
section, s7a, is again striking. It is made up of just three longs (for the
separation of s7a from s7b, see below on the textual problem in v. 100).
Aeolic phrases do not appear in 2 but return in 3 (s7b8). First
comes a telesillean (s7b). The next verse, s8, ends with the same six
positions (
ww
w
), but they are preceded by w

w, instead of
just
. So s8 starts in the same manner as s6, the key phrase of 2
(w

w
):
s6 w

w
k
s7b

ww
w
k
s8 w

w
ww
w
k
s8 is, in a sense, a combination of s6 and s7b.
The key phrase of 2 is augmented into we e in the Wrst verse of
the last section (s9). This is the return of the initial verse of the
224 The Eighteen Majors
strophe (s1) with an additional e. e is dominant in the next verse
(s10), but a double short, d, is introduced. Thus, the Wrst half of s10
(e d) is identical with s5. The last verse (s11) starts with ^ e again,
but unlike the other examples a long anceps intervenes before a
sequence in single-short movement (e 3 e w e).

Epode
The epode contains here and there several phrases reminiscent of the
strophe. At the same time, aeolic phrases, which are rare in the
strophe, appear more frequently. In particular, reversed dodrans is
repeatedly used and distinctive. Another diVerence between the
epode and the strophe is verse length. Verses tend to be longer in
the epode. Transition from one verse to another is gradual and
smooth. The contrast between one verse and the next is not so strong
as in the strophe, which are divided into 4 sections. It is not possible
to divide the epode into sections in this way.
e1 (^ e e wil) as a whole is very similar to s2, which, however,
starts with we instead of ^ e e. ^ e e already appears in s4 and s6
(the key phrase of 2). The latter half of e1, wilamowitzianum, is the
same as s2. The next verse, e2 (gl we rdod), again starts with w
.
In e2 this phrase is the aeolic base, but is nevertheless identical with
^ e in the preceding verse. This glyconic is one of those that may be
better analysed as ^ e dod. Note the twofold structure of e2:
e2 w

ww
w

w
w

w
ww
j
The Wrst half ends with dodrans, the latter with reversed dodrans,
which is incorporated in the wilamowitzianum of e1. For the neat
relationship between the localization of words in e1 and that in e2,
see individual verses, e2 below.
Reversed dodrans has not been used in the strophe, but occurs at the
end of e2 for the Wrst time and is repeated as many as twice in e3. The
resolution of the initial position of the second dodrans makes a char-
acteristically impressive palindromic phrase,
ww
www
ww
. This
phrase is, in fact, reminiscent of
ww
www
which has already been
incorporated in s10 (as d re), and also of
www
ww
, in s5 (as
er d).
 Pythian Five 225
In the next verse, e4, reversed dodrans is enlarged into the pher-
ecratean (we ph). Preceding it, we is used in the same way in s2
(we wil) and s3 (we gl e). Pherecratean is, theoretically, aeolic
full base catalectic form of dodrans, but this pherecratean in e4
should rather be analysed as rdod anceps (Part I, 7. 5). Note the
similarity between the latter half of e2 and e4:
e2 w

ww
w
w
w

w
ww
j
e4 w
w

w
ww

j
^e comes back again in e5. The sequence ^ e d e is contained in
s10 (e d e e), and even the resolved position is the same in these
two verses. At the same time, names apart, the phrase e5 is similar to
e3 in structure:
s10
w

ww
www

w
j
e3
w
ww
www
ww
j
e5 w

ww
www
j
This similarity is carried over into the next verse:
e6
ww
rw
ww

w
j
The characteristic movement of e3,
ww
www
ww
, returns
again in e6. The next verse, e7a, is ^ e e, identical with s4 and s6
(and with the beginning of s8 and e1). e7b also is made up of just Wve
positions (we
). Several verses start with this movement (s1, s2, s3,
s8, e4).
In e8, two es are juxtaposed. The combination e e is very
frequent in the strophe (s1, s9, s10). The Wnal verse, e9, contains
glyconic, the aeolic phrase used occasionally both in the strophe and
in the epode. For the unusual responsion in base, see below.

Textual problems
2 (s2). Resolution of the fourth position is found, among eight
repetitions, only at v. 2 (  I A). Hermann (Emend. P.,
145), with his usual keen sense, doubts the resolution and proposes
O A: a rational but radical solution. But it is in the Pindaric manner
to use a diVerent form only once in corresponding repetitions (All-
but-One; see Part III, C). Thus we need not doubt the innocent word
226 The Eighteen Majors

I A, though I cannot be so conWdent that I A A is a key-

phrase as some commentaries maintain.
5 (s5). E. Schmid corrects the unmetrical paradosis  into
 . The position is resolved at all the repetitions except v. 36 (see
ad loc. below). Even if consonantal treatment of  of  , or
omission of  ( ), were posible (cf. Snell in apparatus; Liberman
p. 258 n. 6),  would be more probable.
12 (s1). The paradosis  =  is unmetrical;   Moscho-
poulos.
17 (s6). I must confess that I am unable to understand vv. 1719
( c Oe . . . ). The diYculties arise, above all,
from the Wrst three words both in the meaning and in the logical
construction of the sentence(s). Christ seems to me to be the nearest
to a solution: immo Oe est ipsius Arcesilai, qui vultu gravita-
tem et maiestatem regiae dignitatis prae se ferebat, quam innatam
( ) et a maioribus acceptam habebat. However, his text is
metrically untenable:    = Oe N  :
according to modern scholarship,  should not be introduced at the
verse-end by emendation. Verse-end is certainly attested by brevis at v.
37. Recently Liberman has taken Oe as Arcesilaus. His solution
is:   , = Z, = N  . . . Since there are
good grounds for dividing s7 into two verses (see v. 100 below),
Z as one independent verse can be accepted. Another direction
of emendation and interpretation has been proposed by Hermann
(Emend. P. 1456), who changes  into K
, keeping, of course,
 : te ius servantem magna sequitur felicitas, partim quod rex
es magnarum urbium: nam gentile lumen est haec maxime vener-
abilis dignitas tuae sociata sapientiae; partim beatus es, quod vicisti
nunc in ludis Pythiis. Lloyd-Jones, in Philanthropia kai Eusebeia:
Festschrift fur Albrecht Dihle (Gottingen, 1993), 304 n. 5 The Fur-
ther Academic Papers of Sir Hugh Lloyd-Jones (Oxford, 2005), 113 n. 5,
suggests that Oe means a source of light.
18b (s7b). The paradosis N is unmetrical; N
E. Schmid. Gentili keeps the paradosis and explains the apparent
anceps as the last position of a syncopated trochaic ( palimbac-
cheus/molossus), followed by an iambic. However, his metrical ex-
planation is hardly acceptable. The juxtaposition of two ancipitia is
improbable; if anything, the positions of both would be regularized,
 Pythian Five 227
either long or short. Above all, palimbaccheus and molossus, and
their responsion, are admissible only in a purely trochaic context,
which is totally alien to Pindar.
23 (s1). E. Schmid corrects the unmetrical word-order of the
paradosis   into  c.
36 (s5). The paradosis 
 is unmetrical;  (Pauw) is
accepted by Hermann (Notae) and Boeckh. Since resolution occurs
regularly at all the other repetitions except for the preceding
 (All-but-One), Hermann introduces exact responsion
(Emend. P.): certo aestimari posse arbitror, Pindarum 
  scripsisse, ut A pro substantivo esset. Gentili
keeps 
 . According to his theory, or nella dottrina metrica
antica e nella prassi poetica, responsion between cretic and trochaic
(here in this case, epitrite of trochaic form) is possible. But respon-
sion of
w

with
wr is outrageous.
42 (s11). Snell, following Schroeder (Pythien, 58), supposes that
the initial  of 
 is prolonged ( making a preceding short
vowel to make position), so that the Wnal  of the preceding
 is metrically long. The paradosis is also accepted by
Fuhrer, who thoroughly examines the possible cases of irregular
responsion afresh. The prosodical licence of lengthening the Wnal
short syllable followed by a single consonant is applied to a few
isolated cases in D/e; but in non-D/e, it is extremely rare. P11s1
(38) is perhaps the only case in the eighteen majors (see ad loc.).25
Our case is diVerent from that and the examples in D/e in three
points:
(i) The consonant which can make a position is usually  (espe-
cially the combination of ), and ocasionally . There is no other
case involving .
(ii) There is a word-division between the vowel  and the nasal  in
our case while word-end comes after  in most cases. This is a
diVerence, though all the words in a verse are metrically continuous,
so that 
 . may be compared with a 
  (P1

25
Snell does not recognize brevis in longo but posits prosodic lengthening at the
end of P11. 8. He thinks that the Wnal  of  followed by a vowel (Z) is given
a syllable-closing pronunciation. But colometry is good enough even if verse-end is
established there; see P11s3.
228 The Eighteen Majors
v. 45). Snell correctly classiWes these in the same sub-category in his
metrical conspectus. But P5 v. 42 is still unique. All the other similar
cases involve . There are no other examples in which the nasal (, )
or  stands at the beginning of a word, and  is not a light
word like a preposition.
(iii) The metrical context is diVerent. Our case occurs neither (i) at
the end of double-short movement (Fuhrers Class I (a), in which he
posits prosodical lengthening) nor (ii) in the sequence
ww
taking
the place of
w

(Fuhrers Class II, in which he posits metrical
freedom).
On the other hand there are some similarities. P5s11 is the Wnal
verse of the strophe: Fuhrer observes that this type of freedom of
responsion tends to occur there, and that the normal rhythm has
already been repeated twice. If the metrical licence were accepted, it
would mean that more Xexibility should be introduced in our case
than others. Fuhrer himself admits: Diese Annahme wurde eine
besondere, die Sprache gewaltsam in seinen Bann ziehende Festigkeit
der Rhythmus gegen das Ende der Strophe hin inplizieren.
The older critics did not accept the licence but tried either to
emend the text (e.g. Boeckh) or to explain the irregurality metrically
(Hermann). The manuscripts are divided between two readings,
 and  e. Boeckh proposes
, e

, . He paraphrases  thus: natum, quum lig-
num natura Wguram humani corporis haberet, nec nisi expolitum
esset arte (Dissen, 1830, Comm. 231). Hermann, who, in general,
does not hesitate to introduce emendationes metri causa, nevertheless
accepts  . in Notae and tries to give metrical justiWca-
tion to the responsion between short and long at this position: nihil
in  de metro laborandum. Neque opus ut aut

H quod in diss. de dial. P.p.6 conieci, aut
, e 
,
 scribatur, quod edidit Boeckhius, commate per typothetae
errorem, ut puto, post 
 posito. Versus es asynartetus ex
antispasto et iambico ischiorrhogico, quo genere saepe usi tragici.
His metrical explanation is invalid. Later (Emend. P.) he changes his
mind and accepts e 
  without the comma
(Librorum scripturam,  : ., ego quidem non propter
metri legem, sed propter diligentiam, qua in hoc metro syllabas
 Pythian Five 229
aequavit Pindarus, repudiatam velim.) Several emendations are pro-
posed. Boeckhs, adopted by Schneidewin and Christ, requires an
unattested adjective  , meaning either P or of the tree:
cf. Christs paraphrase: arboream statuam ex uno trunco factam,
non ex pluribus partibus compositam. Thus Kayser writes: et prop-
ter constructionem 
, e :. et propter insolitam ea
notione adjectivi verbalis genus non accipio. But his proposal

, F 
 F is not acceptable. Nor is Momm-

sens in his ed. maior (withdrawn in his ed. minor):
, F

 , i.e. huius (ipsius saltus Parnassii) plantam singu-
larem. Turyn adopts the emendation of Pauw (, e

 ). But this is even less plausible.
Contrary to his predecessors, Wilamowitz (GV 307) retains the para-
dosis just with an orthographical change ( 

) but introduces a metrical licence instead:
w

w

w
w
 w


w
w
w
.
Responsion ist streng; aber nach den Erfahrungen mit Bakchylides
werden wir uns nicht entsetzen, wenn in den letzten [iambischen]
Trimeter einmal das zweite Metron statt kretisch bakcheisch ist.
What Wilamowitz means by Bacchylides is either 18. 7 (our aeoli-
cized e 3; cf P8s6) or 17 in which free responsion seems to occur
between iambic and cretic and between iambic and bacchiac (the
responsions are still controversial). But the responsion between cretic
and bacchiac is, even in Bacchylides, unparalleled.
Unless an overall emendation is adopted (like e  
,
van Herwerden post Hartung, according to Gerber, Emendations;
Hartungs own proposal is   ), it seems
inevitable to adopt Boeckhs (above).
49 (s7). See below, v. 100.
52 (e10). The paradosis IH (most MSS)/IH (V) is un-
metrical; Moschopoulos IH is a good emendation and is gen-
erally accepted. Mommsen proposes $.
53 (e11). The paradosis
 is unmetrical;

Moschopoulos.
57 (e4). The paradosis  is possible because the initial pos-
ition of e4 is anceps. But the start of most verses is predominantly short.
Exact responsion is easily recovered by  (Moschopoulos).
230 The Eighteen Majors
58 (e5). The paradosis F is unmetrical;  Moschopoulos.
69 (s7). See below, v. 100.
72 (s10). The MSS  cannot be accepted. Both Snell and
Turyn take  (Wilamowitz; Apollo proclaims), but 
(Hermann (Opuscula VIII, 938), followed by Bergk24 ()
and Christ, and recently by Race; it is mine to proclaim) seems to
me to be better. Metrically both are good ( makes hiatus).
Mommsen, following scholia, introduces  by uniting s10
with s11:

w

ww
rw

wtw


w
w
w

This is not impossible, but is inferior even metrically.
80 (s7). See below, v. 100.
100 (s7). Snells text is metrically untenable. He gives the following
metrical scheme and analyses it as sp gl:


a

ww
w
k
This interpretation is hard to accept. First, responsion between w
and


at the aeolic base is extremely rare; there is only one example
(O10s6; see Part I, 5. C. 1). And an aeolic phrase with the base w

usually, perhaps always, comes at the beginning of the verse. Secondly,
there is no other example of spondee preceding an aeolic phrase.
Two metrical schemes are conceivable:
(i) s7

w

ww
w
k
e dod
(ii) s7a


k
sp
s7b

ww
w
k tel
The metre of (i),
e dod, is common; there are two parallels: O10s2
and P8s4. It requires that the third position be short. This is short in
four lines (vv. 7, 18, 38, 59), and in another three (vv. 49, 69, 80),
some minor, orthographical emendations make it short; see below.
Substantial change, however, is inevitable in v. 100. Even if A
(gen. pl.) is rejected and E (acc. sg.) is adopted, the following
 is impossible. The word must start with a vowel. Then,
E (acc. sg.) o would be a most plausible conjecture
(see below). Alternatively, if (ii) is chosen, the third position
does not need to be short. A (gen. pl.) is possible. So is
.
 Pythian Five 231
I adopt (ii), following Hermann (Emend. P.), and divide between
s7a and s7b. Hermann is followed by Schneidewin, Bergk, Schroeder
(ed. maior), Turyn, and Liberman; but the last two read E
(acc. sg.). They all accept .
Boeckh combines s7a and s7b into one verse, keeping . But his
text is hard to accept. He introduces 
 Ia . . . E (nom.
sg.)  with full-stop after $
e . Mommsen revives the old
emendation o (and adopts E acc. sg.). Christ and Bowra
follow him. o (Beck according to Mommsen and Christ; but ed.
Oxon. 1697 had already adopted it, according to Bowra and Gerbers
Emendations) would be a good conjecture: songs can be poured, H
cannnot. (Barrett, Collected Papers, 191). However, the metaphor is not
necessarily bold, and the genitive plural (A) follows Pindars
tendency more than the accusative singular. For   I . . .
, variis accentibus codd.; scholia utroque modo tum genitivos
plur. tum accusativos sing. interpretantur (Turyn).
Metrically, (ii) is certainly uncommon, but acceptable. s7a is
anceps spondee. It ends with a short vowel (Barretts SVE) at two
repetitions, but avoidance of SVE is not an absolute rule. The brevity
of the verse composed of only three positions is not a diYculty: there
are two other comparable verses (see Part I, 8. A. 3). Rather, the
analysis has its own merit. As is explained above in the general
introduction to this strophe, three longs provide a strong contrast
to six shorts within the same block (s4). As Turyn points out, there is
strong semantic pause at three of the repetitions (Hermann has
already laid stress on this point). This argument is not decisive
because of frequent discord between semantic and metrical pause.
At vv. 49, 69, 80, some emendations are unavoidable whichever
colometry may be chosen. DiVerences between Hermann (ii) and
BoeckhMommsen (i) are more or less orthographical and are not
decisive on their own:
49 E = (codd. - or -)
69 E H d = H d (codd.  H d)26
80 E, = , (codd. E )

26
Bergk24 establishes the trisyllabic verse, s7a, like Hermann (ii), but introduces
B and B.
232 The Eighteen Majors
At 100, should be deleted.
118 (e3). The paradosis 
e t  is two positions too
short. Both Turyn and Snell, following Christ or Bowra, adopt 
e
E,  (Hartung; who proposes E at v. 120 too). Schroe-
der (Pythien, 51) writes: nach Hartungs glucklicher Deutung des
uberlieferten t. I am not sure of this emendation. E has the
following competitors: 
 (Hermann, Notae), Z
 (Boeckh),

E (Mommsen), 
e H , 
  (Bergk2), 
e
H , 
  (Bergk4). Liberman proposes  in the
apparatus.

Individual verses
s1. See Part I, 8. C. 5 (double e). It is one of Pindars favourite
techniques to Wll the whole of the Wrst verse with a gnomic manifesto.
Here
F P makes up one verse. The combination of
anceps e e is a common one, but it usually occupies the initial
part of a longer sequence. Verses of two es length are rare. There is
only one exact parallel, O10s5.
s2. As observed above on the strophe, the Wrst seven positions are
completely identical with the preceding verse (s1); after the seventh,
ours is ampliWed by anceps d. Thus s2 may be analysed as we ewd.
At the same time, however, the similarity of s2 with the following
verse, s3, is not negligible. In s3, the wilamowitzianum of s2 is
replaced by a glyconic (and expanded by e). The sequence

w
w
ww
is highly ambiguous between wilamowitzianum
and e wd in this verse, where metrical context is not necessarily
decisive (Part I, 7. 3). Compare our verse with these:
P5e1 w

wr
w
w
ww
k ^e e wil
P10s5 w

w
w
ww
w

k ^e wil 3
P6s4 ww

w
w
ww
k ^ d wil

Bridge from e to wilamowitzianum is generally observed, except v. 33

(
 j  ); cf. Part I, 8. B. 5..
s3. The verse starts with the same phrase (we) as s2, but here the
fourth position is resolved throughout all the instances. Both xe as
preWx and e as suYx of glyconic are common (Part I, 8. B. 5). Bridge
 Pythian Five 233
from gl to e is strictly observed in all eight repetitions while cut
occurs between xe and gl at no fewer than four.
Excluding the initial position, the verse is palindromic on a gigan-
tic scale. Even resolutions are located symmetrically at either side.
There is no exact parallel but one very similar verse:
N3s4 w
w

w
ww
w

w
k we gl e
s4. See Part I, 6, C. 2 (er re); 8. A. 3 (shorter verse). Two
analyses are possible: (i) e2 (rwrw
) or (ii) ^ e e (wr rw
).
Neither has parallels; see Part I, 6. C. I prefer (ii), because ^ e is
frequently employed in the ode and because ^ e e is used again in
s6. Six successive short syllables are impressive. See Part III, F. Word-
end is almost regularly located at wwwwjww
j, except one: v. 35
(
 A).
Verbal assonance: 4
  77
, and 46 
 
97  .
s5. See Part I, 8. C. 9 (e d). There are Wve examples of e d at the
verse-end, but it is usually part of a longer phrase. The nearest is
O10e1 w
www
ww
j we d
Contrary to the resolution at s2 (once in 8 repetitions), here reso-
lution occurs almost regularly at all the repetitions but one at v. 36
(); see textual problems.
s6. See Part I, 8. A. 3 (shorter verse). ^ e e is a very common
phrase (10 examples in total), but there is no verse which is solely
made up of it excepting the two verses of this ode (s4 is the other).
s7a. See Part I, 8. A. 3 (shorter verse). There are two possibilities
for the analysis of the three longs. I prefer anceps spondee to its
reversed (spondee anceps), because anceps at the beginning of a
verse is more frequent than that at the verse-end.
s7b. There are in total four verses made up of only one telesillean
(the others are I7e3, O9s1, O10e6).
s8. There are in total four examples of wil 2 (of which N3s7 is the
most certain case; the other two are P2e8 and P8s5). P2e8 starts with
the identical form of our verse:
P2e8 w

w
ww
w

w
ww
w

k wil 2 hipp
234 The Eighteen Majors
s9. See Part I, 8. C. 5 (triple e). The triple e combined continuously
without link has many parallels. Besides seven verses in O2, there are
two others, O1s9, I8s8. The exact parallel, xe e e at the beginning
of a verse, is found three times in O2, at s3, s5, e1. Resolution is
frequent in these parallels while in our verse none of the three e is
ever resolved. Word-end between two es is generally avoided, except
at v. 9 (
 j
 : ).
s10. See Part I, 8. C. 5 (double e), 8. C. 8 (d e), 8. C. 9 (e d).
Although d and e are combined with each other rather freely, there is
no exact parallel to e d e e. The resolved e, rw
, located after
d (or D) is common (7 examples in total). As in s9, word-end
between two es at verse-end is generally avoided, except at 72 (e 
Ke j ).
s11. See Part I, 8. C. 6 (e 3). In the lyric iambic in tragedy, this verse
would be easily identiWed as bacchiac lecythion. But in Pindar the
initial w

is rare; N6s1, which is made up only of w

, is the
sole parallel. Certainly w

is not bacchiac ( syncopated iambic).
I take the third syllable to be long anceps. Word-end after w


occurs only at 22 (), 53 (I ), and in a looser sense, 115
(
 ). e 3 (lecythion) is a common phrase. But it is not
preceded by anceps, let alone by bacchiac. The sequence w


w
is not diYcult to use at the middle of verse in D/e, the third position
being long link anceps; but for whatever reason it is extremely rare in
the eighteen majors. Besides our verse, there is only one example:
O13s3 (included in the phrase we 2
e).
e1. See Part I, 8. B. 8 (^ e e aeolic). The sequence of ^ e er
(w

wr) is employed in O2s6/7 and also O10s4. The following
wilamowitzianum includes reversed dodrans, which will be repeated
in the following verses. But it is analysable as ewd too. As in the latter
half of s2, ambiguity remains.
e2. See Part I, 8. B. 5 (aeolic we). This verse is one of the four
examples in which w
w
(we) functions as a prolongation of
the ending of the preceding glyconic and the like ( . . . w
). As is
described above under the epode, e2 is in a sense an ampliWed version
of e1. The start and the ending of the two verses are composed of an
identical sequence of longs and shorts. The similarity is strengthened
by localization of the words or word-groups of the same shape. The
clearest case is found in the fourth epode:
 Pythian Five 235
116   Nd K
 H ,
117 .    e F 
 E ,
116 w

wj r
w
j w
jww
j
117 w

wj w
w
jw
wj

j w
jww
j
Hiatus/brevis does not occur at the end of the verse in any repe-
tition. But it is reasonable to assume verse-end because sentences are
round oV there in all the four repetitions, and because otherwise the
length of the verse would be extreme.
Bridge is strictly observed at the two borders between e and dod/
rdod (i.e. between ^ e and dod, and between we and rdod).
e3. See Part I, 5. E. 2 (resolution of reversed dodrans); 8. A. 6 (g)
(palindrome). The latter half of the preceding verse, reversed
dodrans, is repeated twice. A longer series, of Wve rdod, is found in
I8s5ab; but whereas the Wrst position of all Wve is of the same form,
the second rdod in our verse is resolved. Resolution of the initial
position of rdod is common itself: there are in total nine examples
(Part I, 5. E. 2), of which four, including our verse, have the resolved
rdod in question in the middle of verse (the others stand at the
beginning of the verse). Interestingly, they are found in P2e and
P5e only, and their structural similarities seem suggestive:
P2e2 w

ww
w
www
ww
k gl rdod
P2e3 w

ww
w
www
ww

w
w
j gl rdod e2
P5e3
w
ww
www
ww
j rdod rdod
P5e6
ww
www
ww

w
j d rdod e
In the two examples of P2, the preceding phrase is not (
w)
ww

but obverse dodrans:
ww
w
. But they are based on the same
construction.
Because (
w)
ww
stands before www
ww
, a long palin-
dromic movement is created. This movement is found only in
N6s3 and the two P5e verses.
e4. Manifest indications of verse-end are lacking at the end of the
pherecratean. Then, if e4 is combined with the following verse, e5, the
pherecratean changes into rdod which is repeatedly used in the preced-
ing verses and e5 becomes identical with s10, but for the Wnal e:
w
w

w
ww

w

ww
rw
j we rdod e d e
236 The Eighteen Majors
This colometry is not bad in itself, but a strong sense-break at two
repetitions, v. 26 and v. 88, suggest that the separation of e5 from e4
would be better. Pherecratean is not a common phrase in Pindar, nor
there is any other example in this ode, although a verse of identical
composition with ours is used in O13s2 (in this verse too, the
pherecratean is peculiar). The illustration above in Epode show how
the pherecratean in our verse may be analysed as reversed dodrans
long anceps.
e5. See Part I, 5. A. 4 (analysis not as
ww
but as ^ e d); 6. B
( e d, ^ e D); 8, C, (8) (d e). Acephalous e standing before d is
^
not rare. There are two verses identical with ours, P2e6 and P10s2b.
Moreover, O1e2 (^ e d e 3) and O10e2, I8s9 (^ e D) can be
paralleled.
e6. See Part I, 5. E. 2 (resolution of reversed dodrans), 8. A. 6 (g)
(palindrome), d aeolic: 8. B. 4.
e7ab. See Part I, 6. C. 3 (short anceps preceding resoloved long);
8. A. 3 (shorter verse). These two verses are combined into one in
most editions. The result is analysed as
w
tw
wtw

j ^e e 3

This analysis is not impossible, but there is no parallel in the eighteen
majors for e 3 preceded by ^ e. Instead, the divided verses are paral-
leled, and moreover are susceptible to an organic explanation (see
above). Note that e7a, ^ e e, is identical with s6 (and s4). e7b is
we
, which is an independent verse in O10e1b (x
wr
k) too.
Outside the eighteen majors there is a good parallel for the division:
P7s7 wr
w
k ^e e
P7s8 u rw

k xe

P7s7 s8 is the same as P5e7a e7b, and at the same time the verse-
end of P7s7 is guaranteed by brevis (v. 7).
There are two resolutions, each in one repetition: v. 29 I
(Hermanns regularization (Emend. P.), IE, is not only un-
necessary but untenable) and v. 91

.
e8. See Part I, 8, A, (3) (shorter verse), 8. C. 5 (double e). There is a
parallel (O2s4) for the verse made up of only two es.
e9. See Part I, 6, C, 3 (short anceps preceding resolved long), 8. B. 4
(aeolic d); 8. C. 5 (double e), 8. C. 7 (dwe). This is the sole example
 Pythian Five 237
of irregular responsion at the full base between wwwo (31) 

wZ (62) and h (124) 

a (93). Not only
the responsion between

 www, which is manifestly irregular,
but also that between
w  www, which seems less irregular, are in
fact, unparalleled. Mommsen proposed an emendation: $ A for
o. Pindar does not admit irregular responsion in typically aeolic
stanza-forms (Class I), where only responsion between the identical
forms and


w are found. Perhaps because O5e (and O5s
too) is not a Class I stanza-form, Pindar may have been able to
introduce this irregularity. It is suggestive that another irregular
responsion is found not in a Class I stanza-form but in a Class II:
O10s6 a

ww
w
k gl
238 The Eighteen Majors

PYTHIAN SIX

Six strophes (monostrophic). Class I

P6s1/2 w
w
www
ww
w

jw
w
ww
k we gl wil
P6s3 www
ww
wt
w
k gl e
P6s4 ww

w
w
ww
k ^ d wil
P6s5
ww
w

w
w
ww
j dod wil
P6s6 www
ww
w

wr
k gl e

P6s7/8 w
w

j

w
ww

k we wil 1
P6s9 w

w
w
u
w
k ^ e e2xe

s1/2 H 46/7; s3 B 30, 39; s4 H 40; s6 H 6; s7/8 H 7/8; s9 B 27

s1/2 1/2 (1/2)  q a 
  

j %
s3 3 (3) I
, Oe K
s4 4 (4) e K 


s5 5 (5)   O 
s6 6 (6)
  
 d a 

s7/8 7/8 (7/8) E o e K

s9 9 (9) 
  


As is usual with monostrophic odes, the structure of Pythian 6 is
simple and clear. The basic metre is aeolic. Glyconics and wilamo-
witziana return, and some additional minor phrases produce slight
modiWcations. The simplicity of the ode is enhanced by repetition of
identical forms. Aeolic base in all three glyconics takes tribrach form.
And three wilamowitziana are of the standard type
w
w
ww
.
As the line-numbers indicate, Boeckh divides two verses, s1/2 and
s7/8, into two. Turyn follows him. But Snell does not divide them. He
is probably right in the former case, and arguably right in the latter
(see below). Two phrases are fused into one verse while word-end is
regularly placed one position after the colon-end (dovetailing; see
Part I, 8. B. 2).
Though the ode is simple as a whole, it includes some rare phrases:
^ d (s4), e
at the end of a verse (s6), and wil 1 (s7/8). And the last
verse (s9) is irregular in this context, in that it is not aeolic; double
short movement is totally absent. RSS is rather high (53.6%), the
highest of the Class I stanzas.
 Pythian Six 239

Textual problems
4 (s4). The paradosis e is unmetrical. Many, including Snell and
Turyn, adopt 
 (see Hermann, Merkwurdige Art von Logik, in
Opuscula, vi. 2869), but the phrase Z . . . K 
 arouses
suspicion in a series of critics from Dissen to Liberman. If K e is
an intruding gloss as Mommsen suspects, there are many possibilities
as to the word which was expelled; cf. Liberman, 2734.
18 (s8). The paradosis I
 is unmetrical; I
E Byz.
21 (s3). The paradosis 
 is unmetrical; 
E. Schmid.
28 (s1). The paradosis K is not impossible, but  (Tri-
clinius) is easy to accept.
31 (s4). The paradosis I is unmetrical; I (Boeckh;
I Triclinius).
32 (s5). The paradosis  is unmetrical;  Byz.
46 (s1). The paradosis I 
 is too long; either
 or
 must be deleted. Bergk2 deletes ; he is
followed by Christ and later editors. Before Bergk, critics used to
follow Byz. and delete
. This is grammatically easier, but is less
probable for the same reason.
48 (s3). l destroys exact responsion. This is not a problem in
general (All-but-One); responsion between resolved and unresolved
long at the end of glyconic occurs at I8s5c (resolved at three repeti-
tions out of seven; cited below). However, introducing a tribrach like
all the other repetitions has a special merit in our case. If l were
replaced, the metre might not be gl e but rdod e 2;27 cf.
P5e3
w
ww
rw
ww
j rdod rdod
P2s7 www
ww
rw

w
k rdod e e
Kayser was the Wrst explicitly to reject l on the basis of metrical
classiWcation (JL 106): Aber einen so tribrachisch ausgehenden Glyko-
neus hat erst Euripides sich erlaubt, Pindars fester Bau kann sich mit
solchen, gleichsam in der Luft schwebenden Formen nicht vertragen..

27
The rhythm is ambiguous in N7e3 and e4:
N7e3
w
ww
wr
w
k gl e (? rdod e2)
N7e4 www
ww
wr
w
k gl e (? rdod e2)
see ad. loc.
240 The Eighteen Majors
Hermann, Emend. P. supposes
 (46, above) to be redundant
and writes: videtur id ex adscripta vera lectione P
 ortum esse.
Citing P2. 28 P
 $

, he proposes P
 in the place of
l. Kayser. Lectiones, 56 makes a comment on it: nimirum P

quidem $

 dici potuit, h  h $

 non potuit,
quia nec injustitia nec superbia caret unquam. His proposal is 
.
But what poses a problem for P
 is not the two adjectives but, as
Schneidewin remarks, the verb 
. Kayser answers Schneidewin
(JL 107): Aber das Particip [ 
] gehort zunachst zu
F
und das vorgeschlagene Substantiv [ 
] ist nur Apposition
dazu. Mommsen modiWes Hermanns proposal in a diVerent manner:
$

 Ia (in apparatus). For the single h, see Hummel,
Syntax 389. Hartung construes diVerently:  h $

 goes
with
F (without a comma after ), and the object of 

becomes . Then he recovers the exact responsion by introducing
IH, which go with .
50 (s5). The paradosis OE

 n

 , is metric-
ally wrong, and needs a verb in the relative clause. Locus antiquitus
corruptus . . . sed etiamnunc credo OE

 esse parepigraphen
adscriptam ad v. 51 
I  (Bergk4). Following Christs
O A n

A K, Bowra introduces  n

A K,
the most natural emendation, which is accepted by Snell. Turyn
adopts a palaeographically less radical  A n

  (Rau-
chenstein). Wilamowitzs
 n requires responsion between
www and
w at the aeolic base. That is improbable in Pindar.
51 (s6). The paradosis
 does not make sense, nor is it
metrical;
 E. Schmid.

Individual verses
s1/2. See Part I, 8, A. 3 (ii) (a) (longer verse), 8. B. 1 (two aeolics in
contact). There are many examples (9 in total) of an iambic metron
(anceps e) preceding an aeolic phrase starting with full aeolic base.
In these the base is mostly
w or
x. On the other hand, when the
aeolic base is occupied by a tribrach as in our verse, it usually stands
at the beginning of verse (13 examples) and is not preceded by any
phrase. Besides our verse there is none in which a tribrach follows e
(there is one example of tribrach in mid-verse: N7e5, where two
 Pythian Six 241
aeolic verses, both of which start with tribrach, are repeated). Bridge
is observed in our verse at the junction of xe and glyconic at all six
repetitions; cf. Part I, 8. B. 5.
There are two other examples of a glyconic followed by a wilamo-
witzianum within a verse (I8s4, N6s2). And later in this ode (s5), a
dodrans ( glyconic without base) is combined with a wilamowit-
zianum. Between the glyconics and the wilamowitziana, word-end is
strictly avoided; cf. Part I, 8. B. 1.
s3. See Part I, 5. F (resolution of aeolic phrase). Glyconic suYxed
by e is Pindars favourite (10 examples), but in this verse there is a
small diVerence from the majority: resolution at the eighth position
of the glyconic. The resolution occurs regularly at almost every
repetition. The only exception is v. 48 (l); see above on textual
problems. Bridge between glyconic and e is observed at all the
repetitions but one (v. 12 jK  ). Perhaps it is most interesting
that word-division is found to be coincident before the sixth (4 out of
6 repetitions), after the seventh position (4 out of 6) of the glyconic,
or in both places, while it is avoided between the sixth and seventh:
www
ww :
w : t
w
k
Verbal assonance: 3 -e K  30 K.
s4. See Part I, 6. B (acephaly). There are four examples of aceph-
alous d, but this case is unique in that an aeolic phrase follows it.
Even examples of full (not acephalous) d preceding an aeolic phrase
are scarce: only
P8s5
ww

x
w
ww
w
k d wil 2
can be cited for a parallel in a broader sense. There is one verse where
acephalous e, instead of d, precedes wil 3:
P10s5 w

w
w
ww
w

k ^e wil 3
s5. Verse-end is not manifested by hiatus/brevis, but the overall
structure may probably guarantee it. No exact parallel for dod wil is
found elsewhere, but there are three examples of gl wil (one of
them occurs in this ode, see s1/2 above). As in these cases, bridge is
observed here in s5 too.
Verbal assonance: 32

  K
  50

A K.
242 The Eighteen Majors
s6. See Part I, 6. C. 3 (eww anceps at verse-end), 8. B. 4 (d
aeolic), 8. B. 5 (aeolic e x). The glyconic is followed by
www

This phrase must be either e anceps (
wr
) or anceps e (

rw
). The former analysis is preferable, because long anceps is
never followed by resolved long in the eighteen majors: see Part I, 6.
C. ii. Besides, there is no example of gl
e (
w
ww
w



w
). By contrast, gl e
is found, and the example supports the
former interpretation:
I8s5c
x
ww
wt
w

k gl e

Nevertheless, e anceps is signiWcantly scarce after any type of aeolic
phrase. The examples are limited to I8s5c and our verse, contrary to
plenty of examples of
w
following glyconic and its kindred.
Word-end never falls between gl and e. At four repetitions (vv. 6,
33, 42, 51) word-end occurs at
www
ww
w

jwr

and at the other two (vv. 15, 24) at
www
ww
w

wjr

Interestingly the short syllable before the word-end in both these
lines is . Moreover, an identical tendency for word-localization is
found in I8s5c too; see ad loc. Parallels for resolution at the penul-
timate position of a verse end are collected and discussed in: Part I,
Appendix, Addendum II.
s7/8. It is not easy to decide whether two verses in Boeckhs edition
should be fused into one, because the phrases created by either
colometry are not supported by parallels. But one combined verse
(i) with dovetailing would be preferable to two separate verses (ii):
(i) With fusion, a rare verse, wil 1, appears. This enneasyllabic
phrase is itself unique. Moreover, there is no other example of
wilamowitzianum and the like (wil 2, wil 3) starting with the
aeolic base of two long syllables. On the other hand, the structure of
the verse is well explained. A verse starting we and followed by an
aeolic phrase has been used in this ode, s1/2.
(ii) w
w

j


w
ww

k
 Pythian Six 243
With separation, a short phrase, xe x, must be recognized as an
independent verse. This is possible (cf. O10e1b u
wr
k x e x),
but this type of short verse is rare in the aeolic context (Class I). And
the ensuing octosyllabic aeolic phrase, whatever name is given to it, is
never used in the eighteen majors and very rare outside the Pindaric
corpus too.
s9. See 6, B (^ e e 2), 6, D (bridge after long anceps). Uniquely in
this ode, this verse is not aeolic. In dramatic poetry it would be
analysed as iambic: bacchiac two iambics. But consistency with
other Pindaric examples suggests the analysis as acephalous e e 2
link anceps e. The nearest parallel is
O1s10 w

w
w
rw
k ^e e 2e
Similar too are:
P11e5 w
w
w

rw
k we 2
e
O13s3 w
wrw


w
k we 2
e
O13s4

wrw
a
w
k
e 2 x e
The fact that double short is, exceptionally, absent from the Wnal
verse in a stanza form which otherwise would be made up totally of
aeolic phrases can be paralleled by
P8s7 u
w
u
w
w
k x e x e2
though P8 is triadic.
The anceps between e 2 and e is realized as long at three repetitions
(or 4, if the Wrst syllable of 27

 is counted long) and short
at three.
244 The Eighteen Majors

P YT H I A N E I G H T

Five triads. Class I

P8s1 www
ww
w
k gl
P8s2
w
wwrw
k gl
P8s3

w
ww
k hepta
P8s4 w
w

ww
w
k we dod
P8s5
ww

a
w
ww
w
k d wil 2
P8s6 a
w
ww

u
w
w
k hepta e3(aeol)
P8s7 a
w
a
w
w
k xexe2
P8e1 w
w
ww
w

w
w
k hepta 2 e2
P8e2 u
w
ww

a
ww

k hepta ph
P8e3/4
a
ww
w

j

ww

k gl ph
P8e5

ww

ww
w
jw
j
d dwe
P8e6 w

ww
w

w
ww
w
k gl gl
P8e7

u
u
ww
w
w

k sp hepta 2 3
s1 H 29; s2 H 44, 51; s3 H 52, 94; s4 H 46, 88; B 53, 74; s5 B 33; s6 H 97; B 27, 69;
s7 H 56, 91; B 98; e1 B 57, 99; e2 H 79; B 16, 58; e3/4 B 38/9; e5 HB none; e6 B 20;
e7 H 21, 42

s1 1 (1) / ! , 

s2 2 (2) t 
 ,
s3 3 (3) A  d

s4 4 (4)  ~  $


s5 5 (5)  a  .
s6 6 (6) f a e e   d
E H
s7 7 (7) K
 H f IE
e1 15 (15)  b d 
  K .
e2 16 (16) (g   h  ,
e3/4 17/18 (17) Pb a f
 A b H
e5 19 (18)   
 n PE 
e6 20 (19) 
   K
e7 21 (20) e
  E  .
str. 2 228 (21)(27) ant. 2 2935 (28)(34) ep. 2 3642 (35)(40)
str. 3 439 (41)(47) ant. 3 506 (48)(54) ep. 3 5763 (55)(60)
str. 4 6470 (61)(67) ant. 4 717 (68)(74) ep. 4 7884 (75)(80)
str. 5 8591 (81)(87) ant. 5 928 (88)(94) ep. 5 99105 (95)(100)
 Pythian Eight 245
The basic metre of Pythian 8 is aeolic. Overall, it is homogeneous.
It starts with three short verses which all consist of a single, unam-
biguously aeolic phrase. They are very simple, like some others.
Moreover, even the intermingled combinations are simple, and fa-
miliar in Anacreon or dramatic poetry, like gl ph (e3/4) and gl gl
(e6). The ode is classiWed as Class I without doubt. However, on
closer inspection, the apparent simplicity proves deceptive. The ode
as a whole is not so simple.
The lengths of the aeolic phrase in this ode are diverse. Each of
the Wrst three phrases ( verses) is of normal length, seven or
eight positions. But after them comes a phrase of 10 positions (s5)
and another of 9 positions (e1). Furthermore, there is one of 12
positions (e7), though the colometry of the verse is not certain. As
for shorter ones, there is a six-position phrase (s4). Some would
like to include among the aeolics

ww
in e5, which counts
Wve positions. Moreover, it is not impossible to take
ww
at s5
as the minimum length for an aeolic phrase. But I reject these
classiWcations. Rather,

ww
and
ww
should be classiWed
as d-phrases. Like d-phrases, e-phrases are also used; w
w
as a
preWx (s4), and
w
w
as a suYx (e1). There is one phrase
which I classify as freer D/e:
ww
w
w
(e5), but which
could also be included in aeolic (Part I, 5. A. 5). In contrast,
double short is absent from s7, and the verse cannot possibly be
aeolic.
At s7 and e1, the paradosis includes some ancipitia that cannot
easily be analysed either as link or as aeolic base:
s7 u
w
u
a
w
k
e1 w
w
ww
a

a
w
k
The irregular long occurs at one repetition out of ten or Wve each.
Dale (Collected Papers, 71) accepts long syllables at all these positions,
and calls the result drag. But this is a confusing extension of the term
drag, which should be limited to the exceptional prolongation of the
penultimate position of glyconic and other blunt aeolic cola (which,
mostly, occur in tragedy). In fact, in the above verses the licences are
prosodic, not metrical. At e1, F (v. 57 55 Sn.) should
undeniably be scanned as w
w. Similarly, the initial syllable of

 (v. 49 47 Sn., w
) and the middle syllable of N

246 The Eighteen Majors
(v. 36 35 Sn.,
w
) scan short; cf. West, GM 12, 17.28 If these
scansions are adopted, the positions shown as (a) prove to be short
in all repetitions.
There still remain two cases which seem to be purely metrical
irregularities:
s6 u
w
ww

u
w
w
k hepta e 3(aeolicized)
e7

u
u
ww
w
w

k sp hepta 2 3
The analyses given are tentative; see ad locc. below.
RSS is very low in the epode (42.4%, the second lowest of all the
non-D/e stanza-forms). In the strophe, RSS is neither high nor low
(52.1%), but the average number of positions is distinctively small
(10.1 pos.).

Strophe
There are seven verses in the strophe. All are short: even the longest
has no more than 14 positions. Although the metre is not straight-
forwardly aeolic, each verse is closely related to those preceding and
following it. s2 is identical with s1: both are glyconics. At the same
time s2 can be taken as a wilamowitzianum, supposing a diVerent
combination of two shorts to be a resolved long (
w
wrww
k).
When the initial syllable of the wilamowitzianum is removed, it turns
into a heptasyllable: s3. In s4 the reversed dodrans included in the
heptasyllable is changed into the obverse one, and the half base of s3
(
) is augmented into w
w
. The dodrans appears to be repeated
again at the beginning of s5, but its last two positions are cut oV
(
ww
w
!
ww
). Then a new, longer phrase starts within the
same verse. This phrase is a decasyllable and starts like a wilamowit-
zianum but is carried on further and includes a dodrans at the end.
The next verse, s6, starts with a heptasyllable (i.e. half base reversed
dodrans). Interestingly, s6 contains a sequence of positions incorp-

28
West comments that shortening is not found with ; , [e]xcept in late
poetry where the prosody is otherwise faulty. [ . . . ]
 is a doubtful example
(GM 17). If this prosodical licence is too great, we should admit an irregular phrase
which is similar to e 2 but its second position may be anceps: aeolicized e 2
(
x
w
) like e 3. See futher below (s7).
 Pythian Eight 247
orated in s5 in the same manner, and both s5 and s6 have the same
number of positions (14). Their relation becomes clear when they are
arranged thus:
s5
ww

a
w
ww
w
k
s6 a
w
ww

u
w
w
k
The latter half of s6 (
x
w
w
) is a curious phrase which I
tentatively call aeolicized e3 (see below). This phrase anticipates the
following verse, s7. If one omits the choriamb, s7 can be seen as, in a
sense, a curtailed version of s6:
s6 a
w
ww

u
w
w
k
s7 a
w
a
w
w
k
There is no pendent ending in the strophe. All seven verses have
blunt ending.

Epode
There are six verses in the epode. Half have pendent ending, which is
totally absent from the strophe. Another diVerence from the strophe
lies in the length of the verses: all those in the epode are long, none
less than 13 positions.
As in the strophe, a close structural relationship is observable
between successive pairs of adjacent verses. e1 starts with an unfamil-
iar phrase (hepta 2), but the verse as a whole has clear aYnities
with s6:
s6 w
w
ww

u
w
w
k
e1 w
w
ww
w

w
w
k
The diVerence between e1 and s6 is so slight that one would be tempted
to take the latter half of e1 (w

w
w
) as an independent phrase
equal to the aeolicized e 3. But it is not wise to set phrase boundary at

jw
against the General Rules (especially, Rule 4a, Part I, 3).
Besides the similarity with s6, e1 is also similar to s7. The phrase e 2
(
w
w
) repeats the close of the preceding verse: s7. The next
verse, e2, starts with a real heptasyllable, like s6, but it is followed by a
pherecratean. In other words, inversion of long and short occurs
twice from e1 to e2:
248 The Eighteen Majors
e1 w
w
ww
w

w
w
k

e2 u
w
ww

a
ww

k
The next verse, e3/4, consists of two phrases and has almost the
identical structure with e2. A pherecratean is used again as the latter
half, while a glyconic is substituted for the heptasyllable. From e3/4
to e5, we can observe a greater evolution. Snell introduces so-called
choriambic expansion as an explanation:
e3/4
a
ww
w



ww

k
e5

ww

ww
w
w
j
He gives e5 the notation ^ glcw
; I give it
d dwe.
A diVerent colometry is just worth considering. Consistent word-
end makes it possible to move the Wnal w
of e5 to the beginning
of e6:
[e5]

ww

ww
w
j
d dod
[e6] w
w

ww
w

w
ww
w
k we dod gl
The Wrst half of [e6] is then identicalwith s4. However, the division I have
chosen (that of Snell and others) is more compatible with the sense,
keeping together PE  (19)  (40)K 
(103).
The Wnal verse, e7, is hard to analyse, in that it includes the
sequence x
x, which is very rare in Pindar. A glyconic (a key
phrase in this ode) can be distinguished, and the whole verse is
certainly a prolonged aeolic phrase of some kind. Barretts deWnition
of aeolic (Hippolytos, 422) would suit it, although his scheme is not
always appropriate for describing aeolic cola.

Textual problems
4 (s4). ~  is rightly scanned by Moschopoulos.
16 (e2).  is restored by E. Schmid.
21 (20 Sn.; e7). The paradosis
 is unmetrical;
 Boeckh.
33 (32 Sn.; s5). Schroeder regularizes the aorist of  and
emends  into . Metrically, this is unnecessary. The position
is the kind of anceps in which a short and a long freely correspond.
 Pythian Eight 249
45 ( 43 Sn.; s3). The paradosis z  r
 is unmetrical; z r

Triclinius.
62 ( 59 Sn.; e6). The paradosis $
 introduces resolution
of the choriambic nucleus. It is extremely rare; $

 E. Schmid.
Also unmetrical is  ; A is Triclinius correction.
80 ( 77 Sn.; e3). The paradosis    is unmetrical;
  Byz.
82 ( 78 Sn.; e5). The paradosis  K cannot be right. Byz.
deletes K, and scholia reads  in the sense of  

E. Some critics (including Mommsen, Schroeder, and Snell) follow
it, but this reading is hardly convincing. There is no parallel of
 
. Bergk2 was the Wrst to punctuate at the
end of the preceding verse (H). A new sentence starts with an
imperative addressed to Aristomenes, 
 . The subject is
changed from  by asyndeton: ne nimis concupiscas in certa-
men descendere, multas iam partas habes victorias. Thus he retains
K but expels a superXuous syllable from the paradosis .
Christ and Turyn follow him. Liberman follows Bergk4: $
H:
  fKg 
 . . . ;  as imperatival.
84 ( 80 Sn., e7). The paradosis  is unmetrical;  Byz.
91 ( 87 Sn., s7). The paradosis  is unmetrical;
 Boeckh.
100 ( 96 Sn.; e2). The paradosis 
 is certainly inferior
to 
 not only stylistically but also metrically; correption in
non-dactylic movement is unusual.

Individual verses
s1. A glyconic with tribrach opening is common (12 examples). Also
common is a glyconic which stands for a verse (8 examples). These
two features are shared by N4s7 too.
Verbal assonance: 8 f 
  64 f  , + ; 50 b
g  92 b .
s2. See Part I, 5. F (resolution of aeolic phrase). There is one
parallel for the glyconic with a resolved sixth position:
P11s2b
w
wwrw

w
k gl e
250 The Eighteen Majors
Word-end regularly falls either before or after the third short of
the successive Wve shorts; thus
w
wwjrw
k (5 repetitions) or

w
wwwjww
k (the other 5 repetitions).
Verbal assonance: 2 t 
 23 
 ; 9 
 K
  86 
 a ; 65   86
a .
s3. A verse made up of a single heptasyllable is not rare; there are
three other examples: P2e7, P11s2, N4s2.
s4. The initial anceps is always short. There is a parallel for anceps
e preceding a dodrans:
O10s2 u
w

ww
w
k x e dod
though its anceps is long at seven repetitions out of 10. Bridge is quite
strictly observed between the e and the dodrans: word-end does not
occur at that point except in v. 95 (v. 91 Sn.) $

 j.
Verbal assonance: 4   25 E.
s5. See Part I, 8, B. 4 (d aeolic). This verse is one of the two cases
where the unfamiliar decasyllable, wil 2, is most certainly used (the
other is N3s7; and there are two other examples). The anceps is
realized both as long (4 repetitions) and short (6 repetitions). Such
a free responsion is typical of the full base used in aeolic phrases of
the Class I stanza-forms.
Seven verses start with d, but an aeolic phrase following the d is not
common. There is another example:
P5e6
ww
www
ww

w
j d dod e
but a diVerent colometry is possible: dod d e. As in s4 bridge is
observed: between d and wil 2 word-end is avoided except in v.
47 (v. 45 Sn.) K
 j
d. Dovetailing is frequent (8
repetitions), even followed by strong sentence pause at four repeti-
tions: 12 o K ,  26 ( v. 25 Sn.)  I
IA .  54
(52 Sn.) I

.  75 ( 72 Sn.)  N.29

29
The exceptions are:
47 K

jd B:  
68
 , Y jb

 
This is one of the two reasons why Hermann (Emend. P.) suspects 47 (turpissimum
vero est vitium in tertia stropha; but 68 does not seem to be ugly all the same, since
 Pythian Eight 251
s6. Juxtaposition of two true longa Wrmly indicates the phrase
boundary illustrated in the chart above. There is always bridge between
the two phrases, which indeed are even fused by dovetailing: word-end
occurs after the initial position of the second phrase at seven repeti-
tions, at six of which the second position is realized as long. Thus the
phrasing

w
w
at verse-end is very frequent: 13  b
, 27 ( 26 Sn.)  d E , 48 ( 46 Sn.) A
K
I
 , 55 ( 53 Sn.) F,  H, 69 ( 66 Sn.) $E
K

 , 76 ( 73 Sn.) c f H
. At 34 ( 33 Sn.), the
syllable after the cut is short:  H. Interestingly, the same
phraseology is observable in the other three repetitions, where dove-
tailing does not occur, being cancelled by enclitics: 6  d
E
H , 90 ( 86 Sn.)  KH I

, 97 ( 93 Sn.) b d

. This is suggestive. The enigmatic phrase
x
w
w
should
be explained not only from the metrical but also from the phraseo-
logical, or even grammatical, point of view. It may be said that Pindar
sought the following colometry until some stage of the creation:
a
w
ww

k octosyllable (hepta 1)
u
w
w
k x e2
The Wrst phrase is a prolonged heptasyllable; its half-base is real-
ized freely, as long in four repetitions and short in six. The second
phrase is puzzling. Its second position is beyond doubt anceps: long
in seven repetitions and short in three.

he, like Boeckh, always admits word-division before ). The other reason is the
inconsistency of the anceps. It is realized as long (four repetitions) and short (Wve
repetitions), and undecided (68). But in three of the Wve repetitions at which the
anceps is short, the value can be changed orthographically;  (33),  (54),
and  (75) become ; F (Byz.), and  (Byz.) respectively.
At 47 the anceps might be lengthened too (
), but that would worsen the word-
division. Hermann supposes that the verse should be regularized at all the repeti-
tions:

ww

j


w
ww
w
k
This is his emendation of 47: K
;
E; j d B. But  is not emphatic; it
should not stand at the beginning of the verse. At 96, where the short is not easily
changed to long, Hermann proposes  for . His colometry is
certainly invalid.
252 The Eighteen Majors
There are a few examples of the phrase
x
w
w
outside the
Pindaric corpus. Bacchylides uses it once as if it were a variation of
glyconic (18. 7), and the second colon of the eupolidean dicolon (Aris-
toph. Nub. 51862) is of the same form, whatever its origin may be:

x
ww

w
w

Perhaps this is a kind of experimental phrase made either from a
lecythion with modiWcation of the Wrst two positions to make an
aeolic base or from a wilamowitzianum with the change of
ww

into
w
;30 but in Pindars case a third explanation is possible. e2
of this ode is similar to s6, but instead of the aeolicized e 3, a
heptasyllable is followed by a pherecratean with the base
x. The
pherecratean is made up of seven positions; if the penultimate and
the antepenultimate are exchanged, the aeolicized e 3 appears:

x
ww



x
w
w

There is another example in which Pindar may exploit the similarity
of lecythion and pherecratean: O1s35, but there the initial two
positions are always
w.
s7. See Part I, 6. D (bridge after long anceps), 8. C. 6 (e 2). There is
no double short in this verse. This type of verse may be used at the
end of stanza-forms made up of otherwise totally aeolic phrases. The
last verse of P6 (^ e e 2 x e) is comparable. There is no other example
of x e x e2, but the reversed combination x e2 x e is used three times:
O13s3, O13s4, P11e5.
Both ancipitia are freely realized, each being long in four repeti-
tions and short in six. When the second anceps is long, word-end is
totally avoided after it; this avoidance is characteristic of Pindar (cf.
Part III, B). At the end, the phrasing formation

w
w
is
repeated as in the preceding verse, though less frequently: 7 H
f IE, 28 ( 27 Sn.) l K 
, 56 ( 54 Sn.) H f
IE, 98 ( 94 Sn.)  .
I introduce irregular scansion (
 w
) into the latter half of
v. 49 ( 47 Sn.) to regularize the metre:

w
w
(
e2). The
position marked by w is short at all the other nine repetitions. This

30
See Itsumi, Choriambic Demeter.
 Pythian Eight 253
position would be anceps, if metrical licence were preferable to
prosodical one. There is a parallel to the phrase
x
w
outside
the eighteen majors:
Pae2s2 u
a
w
j x e 2(aeol )
Here long is attested at one repetition out of the three which are
established: A 
E (2). Note that irregularity is caused
by a proper noun, like
. For the interpretation of aeolicized
e 2, see ad loc.31
e1. The enneasyllabic aeolic, hepta 2, is well established here. As
in s5, which includes the decasyllable wil 2, phrase boundary is
indicated unambiguously by two adjacent longa. Including our verse,
there are three examples of hepta 2. Of the other two (N4s1,
P11s5), P11s5 is exactly identical:
P11s5 x
w
ww
w

w
w
k hepta 2 e 2
As for e 2, there are in total nine examples following an aeolic phrase
with 2 ending. Bridge is observed between the two phrases in
four repetitions; 36 ( 35 Sn.) N j  is the sole
exception. I scan N irregularly as
w
(see above). The
normal scansion would create drag in the penultimate position of
a blunt aeolic phrase. Then the licences would be metrical, not
prosodic.
Verbal assonance: 15   78 (75 Sn.) .
e2. Pherecratean is not a very popular phrase in the Pindaric
corpus, and it is not surprising that there is no other example of
the combination with a heptasyllable. Bridge over phrase boundary is
observed at all the repetitions.
The initial anceps of the heptasyllable is long in three repetitions
and short in two (including 
 at v. 37 v. 36 Sn.; if it is
changed into P
 , short appears only at v. 79 v. 76 Sn. a 
P). The aeolic base of pherecratean is

(2 repetitions) and
w
(3 repetitions).

31
Some editors accept
x
w
at P10s6 (long at vv.30 and 60); see ad loc.
254 The Eighteen Majors
e3/4. Word-end is totally avoided at the end of the glyconic;
instead there is dovetailing in all repetitions. In the strictly aeolic
stanza-forms (Class I) fusion of two phrases is preferred to separ-
ation, and the colometry is better explained in the wider context, as is
described above.
e5. See Part I, 8 C. 2 (double d); 8. C. 7 (dwe). Verse-end is not
manifestly indicated by hiatus/brevis. For a diVerent colometry, see
above on the epode.
e6. See Part I, 8. A. 6 (a) (palindrome). The combination of two
glyconics, which is very frequent in dramatic poetry, is surprisingly
rare in the eighteen majors. Furthermore, contrary to usual practice
in dramatic aeolics, where aeolic bases in two successive phrases are
not reversible, the Wrst is w
and the second
w. This reversing is
comparable with the combination of w
gl
wph at O1s1. This
anomaly would be avoided if the alternative colometry were adopted.
Word-end occurs at the phrase boundary at three repetitions.
Interestingly, a monosyllabic word construed semantically with its
neighbour is used before or after the boundary at all three: 41 ( 40
Sn.)  j
E , 62 ( 59 Sn.) A j Oe, 104 ( 99 Sn.)
 j .
e7. There are a number of verses in which a prolonged sequence of
alternating long and short is prolonged leftwards of the choriamb;
see Part I, 8. C. 3. But P8e7 is unique in that anceps, not short, takes
turns with long. SpeciWcally, two verses, 63 and 105 ( 60, 100 Sn.),
are of the shape


w
w
ww
w
w

k
and the other three, 21, 42, 84 ( 20, 40, 80 Sn.) are


j

j

ww
w
w

k
Note the regular coincidence of word-ends: 21 e j
 j,42 f j
 j,84  j E j.32

32
PfeijVers notation, which follows Sicking, is misleading. Not all the Wrst Wve
positions are ancipitia.
 Pythian Eight 255
There is no other example of the sequence . . .

x
x
. . . .
Two ancipitia cannot be a part of e or its cognates. Aeolic base,
whether it be full or half, preceding a reversed dodrans is the only
possible explanation. Thus, although it is artiWcial, I analyse this as a
spondee a long aeolic phrase, hepta 2 3. An enneasyllabic
phrase, hepta 2, has been used in e1. In e7, this phrase is prolonged
further by w

. See Part I, 5. A. 5; For rdod starting with
x, see 5.
E. 1.33

33
Both Hermann (Emend. P.) and Bergk4 try to regularize the metre, but in
diVerent directions. Hermann changes two short ancipitia at vv. 63 and 105 into
longs and produces two spondees, while Bergk supposes the third position to be short
and changes long into short at three repetitions (vv. 21, 42, 84):
Hermann





ww
w
w


Bergk

w
u
ww
w
w


Bergks scheme is better. It is common one and easy to analyse:
e tel 3. However,
his emendations are improbable; v.21 with scansion of
 as w
, v.42
 f with scansion of f as w
, v.84  A ; d v , (v  ). Her-
mann regularizes not only the two positions but the word-division at the beginning
of all the repetitions:
63 
 , j H j  I    .
105 E j  KH, j d f (H   E.
These are implausible.
256 The Eighteen Majors

PYTHIAN TEN

Four triads. Strophe/antistrophe Class III; epode Class I (? Class III).

P10s1
a
ww

k ph
P10s2a w
a
ww
j hepta
P10s2b w

ww

w
k ^e d e
P10s3 w
ww
ww
u
ww

k wD x d

P10s4 w

jww
a
ww

w
ww

w
w
k ^ e d x d rdod e2
P10s5 w

w
w
ww
w

k ^ e wil 3
P10s6 ww
ww
w
a
w
w
k tel x e2
P10e1 a
ww

a
ww
w
k x d gl
P10e2
ww

k adon ( d
)
P10e3 u
w
ww

ww

jw
k hepta d e
P10e4 u
w
ww

w
k hepta e
P10e5 u
w
ww
w

k hepta 3
P10e6 a
w
a
ww
w

w
k x e tel e
s1 B 43; s2b B 2; s3 H 21, 27, 45; B 9; s4 H 4, 10; s5 H 41, B 5, 59; s6 B 6, 12, 42; e1 H
49; e2 B 68; e3 BH 69; e4 B 34; e5 B 17; e6 B 54

s1 1 (1)  ,
s2a 2a (2) 
.
s2b 2b
e  I K 
s3 3 (3) I
  !    .
s4 4 (4)  

a ; I
   d e
E I

s5 5 (5)  
E , ! "

 
s6 6 (6) IE K
 IH a Z
.
e1 13 (13) 
 d K

e2 14 (14) @

e3 15 (15)  d  $
e 
A
e4 16 (16) Ig 
 /.
e5 17 (17) 
 E d $
e6 18 (18) K  I

F IE 
Pythian 10 is the earliest epinikion in Pindars career. But this does not
mean that its metre is simple. Metre does not develop from simple to
complex along with Pindars age. Perhaps the young Pindar was keen to
employ his polymetrical technique colourfully, but, if so, his intention
was not carried out. Its realization is rather crude. The harmony which
maturity would bring to him had not yet been achieved.
 Pythian Ten 257
The colometry of this ode is not diYcult in itself. Phrase-boundary
is unambiguously settled in most verses with the help of juxtaposed
longs or of anceps sandwiched between two longs. But the interpret-
ation of the ode, especially in the strophe, is not easy. What makes the
interpretation particularly diYcult is the variety of diVerent metres.
A variety of phrases, some of which are hardly ever used outside P10 in
the eighteen majors, are mixed together in one stanza-form, as if in a
ragbag. Compare P10 with simple but sophisticated structure of, for
example, N3 or N7 (Class III). The structural relation of the phrases
in P10 to each other is not evident, especially if we keep the
paradosis at v. 38. Separating s2 into two as above does something to
clarify the way in which the verses develop throughout the stanza (see
below).
I place the strophe in Class III. Some of the characteristics of the
later Class III stanza-forms are already recognizable. There is a long
verse (s4) in which the similarity of an aeolic hexasyllable
(
w
ww
) to e 2 (
w
w
) is exploited. Acephalous e (w
) is
used rather as if it were the full base of aeolic phrases (s2b, s4, s5).
Link anceps is used between phrases (s3, s4, s6). But it is undeniable
that the two metres, i.e. aeolic and freer D/e, are not well amalgam-
ated. At one extreme, there are aeolic phrases in which the ancipitia
are not completely regularized so that they retain a strong aeolic
Xavour (s1 and s2). Moreover, after an interval a long aeolic phrase
returns (s5). It is preceded by ^ e, and the combination is totally
unparalleled. At the other, there is a dactylic verse, whose construc-
tion is more suitable to normal D/e odes (s3).
As for the epode, I tentatively propose Class I, because the hep-
tasyllable is repeated in verse after verse (e35) in a cluster and
because RSS is comparatively low: 45.5% (in the strophe, 50.0%).
Other aeolic phrases too are found here and there, and the base of the
glyconic is not regularized. But the epode is undeniably similar to
the strophe. This ode has one peculiarity common to the strophe and
the epode: resolutions are completely absent.

Strophe
Although at s2a 2b the paradosis gives one continuous long verse
with no break in the middle, I prefer to divide it into two verses, s2a
258 The Eighteen Majors
and s2b, by adopting a slight emendation at v. 38, Hartungs -

 A for 
 A (MSS) so as to clarify the whole
structure better. This produces word-end at the same point through-
out all the repetitions (for the unmetrical paradosis in v. 15, see
below, textual problems). Sentences break clearly between s2a and
s2b at vv. 2 and 44.
With the verse-division as I give it, the strophe begins with a pair
of similar, but contrasting verses. Both are aeolic phrases of seven
positions, but the Wnal position of the Wrst verse becomes, as it were,
the Wrst position of the second:
s1
a
ww

k ph
s2a w
a
ww
j hepta
The relation of the two verses would be closer if the Wnal position of
this pherecratean were not triseme derived from catalexis but anceps.
See Part I, 7. 6. This verse-division is particularly eVective in the Wrst
stanza, where metrical similarity and contrast coincide with semantic
similarity and contrast:   and 
;
cf. vv. 7, 19, 37, 61 too.
From s2a to s5 all the verses start with w
, though analysis diVers
according to the context of each verse. The initial short at s2b, s4, and
s5 is real short while it must be short anceps at s2a and s3. But the
similarity is fully exploited: transitional change is ingenious.
w
s2a x
a
ww

s2b w

ww
. . .
w
s3 x
ww
ww
. . .
s4 w

ww
. . .
s5 w

w
. . .
The second anceps in s2a is dropped from s2b and d emerges as not
merely part of an aeolic phrase. This d is expanded into D in the next
verse (s3). But there is no general similarity in these two verses as a
whole. The next verse (s4) shares the characteristics of the preceding
two verses. It starts in the same manner as s2b, and at the same time it
has a common sequence with s3 in the middle:
s3 w
ww
ww
u
ww

k
s4 w

ww
u
ww
(
w
ww

w
w
)k
 Pythian Ten 259
but its second half, here in parentheses, is not related to any other
verse. The next verse, s5, starts with ^ e in a similar manner to some of
the others, but goes on to be totally independent. In s6, the double-
short movement of s3 comes back again. We can discern a structural
similarity in a broader sense between these two verses:
s3 w
ww
ww
u
ww

k
s6 ww
ww
w
a
w
w
k
Here ww
ww
w
x functions like x
ww
ww
x (x D x). This
phrase is extensively used later in tragedy, especially by Euripides
(see Itsumi, Enoplian in Tragedy); Pindar uses it even in one of the
normal D/e odes (N10s1). See Part I, 7. 2, 6 and Part III, A.

Epode
In the centre of the epode, hepta is repeated three times (e35). In
each case, a short phrase follows and relations between the verses are
easy to grasp. The other three verses too are composite aeolic phrases.
However, they are not closely associated with the central three, nor
with each other. There are as many as seven ancipitia which are not
regularized as short or long throughout.

Textual problems
8 (s2). Most editors adopt Boeckhs H  I (MSS
I). This leaves the third position (anceps) to be irregular.
In the Pindaric examples, the position of reversed dodrans is basically
short; cf. the examples at Part I, 5. E. 1. Here it is short in general,
except v. 8. Pindar may have not succeeded in pruning away long
anceps totally, like the ancipitia in s1 (v. 25) or e1 (v. 49). Hermann
(Emend. P.) introduces short: H
 . This is
accepted by Schneidewin, and may be right.
11 (s5): The paradosis E is unmetrical; E Triclinius.
1516 (e34). Snells apparatus is too simple. The unmetrical
word-order in the manuscripts ( d  Ig $
e

) had already been corrected by Triclinius. He reduced
one syllable by elision and changed the word-order:  d
 $
e  Ig
. Boeckh accepted this, except
260 The Eighteen Majors
, and suggested besides in his notae criticae changing 
d into B b d for metrical regularization. Hartung Wrst real-
ized that the transmitted  should not be accusative but
nominative (but nominative is preserved in D, according to Christs
report). His text is: B b d  $
e 
 Ig.
Christ supported it (!  non  scribendum fuit,
quoniam ut planities cirrhaea, in qua certamina Webant, pratosa recte
dicitur, ita rupes supra illum campum pratosa dici non potuit;
construe  Ig $
e 
 B / -

: ). Christ, however, makes a further change. His main text is
B b d  $
e  Ig
, while in the
apparatus he writes, fortasse  $
e 
A Ig.
This is what Turyn means by Hartungio praeeunte Christ. However,
both Turyn and Snell prefer  d (MSS) to B b d. From
the metrical point of view, both are possible, so they are right. The
Wrst syllable is the initial position of the heptasyllable, and the
position is actually realized both as short (69) and long (33, 51).
The same is true in the next verse (e4); both the word-order of
Triclinius (Ig k
 ) and that of HartungChrist (apparatus)
(
 (or
A) k Ig) are possible, although in e4 the position
is always long at all the other three repetitions.
17 (e5). Moschopoulos (according to Christs report) changed

 (MSS) into 
. The reason for the change seems to
have been simply metrical because the corresponding syllable is
long at all the other repetitions. 
 is accepted by Boeckh,
Schneidewin, and Christ. Schroeder wrongly banned all non-indica-
tive forms of 
; but Braswell, Glotta, 58 (1980), 210, is right:
If we adopt 
 at O9 83 [ 89 Sn.], it would be consistent to
adopt it here as well, although few recent editors have done so.
29 (s5). The paradosis is one syllable too short and the optative
requires :  is supplemented by Hermann. i (Boeckh, etc.) goes
back to Moschopulos. Ng 
o Bergk4.
30 (s6). a (E. Schmid) is necessary instead of a; cf.
Braswell, Pythian Four, 241c. The nineteenth-century editors (Boeckh,
Schneidewin, Mommsen, Christ, etc.) all adopt it. a may have
been introduced from epic usage.34 See further, below on v. 60.
34
LSJ ( ) cites P10. 30 and O1. 28 as examples of this as a poetic equivalent
for  . However, the latter, 

, is the plural of the noun F.
 Pythian Ten 261
38 (s2ab). For 
 A (Hartung), see above, Strophe.
58 (s4). The paradosis   (most MSS)/  d (B) is unmet-
rical; K d Triclinius. According to Gentili, pherecratean ends with
anceps and it can be followed by another anceps (in our case, the
short anceps of iambic:
w
ww
u w
w
); he therefore retains
 . This is another manifestation of his ultra-conservatism on both
textual tradition and metrical theory (cf. P2, v. 82, s2).
60 (s6). The paradosis is corrupt and one position too short:
     . Triclinius supplements rather
mechanically:   ha i  . Gentili adopts this. Boeckhs
 $
 restores exact responsion. Schneidewin and Momm-
sen follow him. Hermanns  $
 introduces an unusual
sequence x
x
w
(see below). I do not understand why Christ
adopts this, while giving the metrical scheme as x
w
w
. Snell
and Turyn, following Bowra, accept   (Mair). This, like
Hermanns conjecture, leave two positions as ancipitia:
ww
ww
w
x
x
w
k
The Wrst anceps is short in six repetitions, long in two, 6 IH, 24
 ), which creates no problem. It is a link, although an aeolic
phrase long mid-anceps is very rare (Part I, 6. D); of the two
examples, the other is P11s4. The second anceps is more diYcult,
since it produces the phrase
x
w
(instead of the normal
e 2
w
w
).35 In fact, the position in question is short except in
vv. 30 (above) and 60. Short could be read throughout by reading
  (or  $
) in 60, and a for
a in 30. Then the phrase x
x
w
becomes x
w
w

(x e 2) in all the repetitions. This is what Boeckh and others have
already done, and Liberman restores after a long interval. The aorist
- is attested at in P11 v. 23b (s2b) and P8 v. 33 ( 32 Sn.) (s5).
Schroeder does not admit the stem -- but -- and changes all
the examples into --. He is wrong. I agree with Liberman (in
apparatus on P8 v. 32): Pindare . . . utilise -- ou -- (voyelle
breve, syllabe longue) selon la commodite metrique.

35
There is an example of the phrase outside the eighteen majors in Pae2s2. It may
be that P8s7 includes this; see ad locc.
262 The Eighteen Majors
69 (e3). The paradosis I  K
 K is one short
syllable too short. Editors emend the text variously; e.g. I b
K
. (Boeckh), I b K
. (Hermann, Emend. P., Schneide-
win), I K
  N (Bergk24), I 
-
 (Mommsen), I   N  (Christ), alii alia.
Turyn prints his own conjecture and is followed by Gentili: I
 K
. This seems metrically improbable because of the epic correp-
tion in non-dactylic movement. Recently Liberman has revived
Housmans I   K
 (ap. Sandys). Wilamowitz
(Pindaros), citing fr. 43 (H
  K
 ), changes accusa-
tive into dative (IE  . . . KE ), which Snell accepts.

Individual verses
s1. The verse consists only of a pherecratean, a relatively unfamiliar
phrase in the Pindaric odes, which nevertheless make up an entire
verse twice elsewhere: O1s4, P11e2. The aeolic base is of the shape

w at seven out of eight repetitions. Only at v. 25 does

(d
) appear (All-but-One). The long anceps is irregular except for
Class I (aeolic) stanza-forms, but  is certain.
s2a. See Part I, 6. B (^ e d, ^ e D). A verse made up of only a
heptasyllable is not rare (5 examples in total). The third position
(anceps) is irregular at s2a as well as s1. In the Pindaric examples, the
position of reversed dodrans is basically short. Here it is short in
general, but long syllable is used at v. 8 (I). Cf. the
examples at Part I, 5. E. 1. If I is not an ordinary noun
(neighbours) but a proper noun (the Amphictionic oYcials, pace
Braswell, Pythian Four, 66e), the irregularity may be mitigated as a
special licence. Otherwise, Pindar may not have succeeded in totally
pruning away long anceps, like the ancipitia in s1 or e1. For Her-
manns regularization, see above, textual problems.
When s2a and s2b are combined, an enneasyllabic aeolic phrase
(hepta 2) appears:
w
a
ww
w

ww

w
k hepta 2 d e
This alternative colometry is possible, but less likely. Hepta 2 is not
a problem in itself. It is certainly an established phrase in the eighteen
 Pythian Ten 263
majors (3 examples in total). The diYculty of the combined verse lies
in that it is less organically related to its surroundings. The emend-
ation is slight, and the separation of two verses is preferable.
s2b. See Part I, 6. B (^ e d, not
ww
); 8, C, (8) (d e). The
combination, ^ e d e, is fairly common. There are two other
verses which are the same: P2e6, P5e5. And O1e2 can be cited as
another parallel:
w

ww

wwww
w
j ^e d e3
P2e6 is especially interesting because it is followed without interrup-
tion by heptasyllable, in the reversed order to P10s2a and s2b:
P2e6 w

ww

w
j ^ e d e
P2e7

w
ww
k hepta
s3. See Part I, 6. D (bridge after long anceps), Part I, 8. A. 6 (c)
(palindrome); 8. C. 1, 2 (D and d). This verse is divided into two
similar halves:
w
ww
ww

u
ww

k
Obviously, D in the Wrst half functions as an expansion of d in the
second. This diptych structure is surprisingly rare in the eighteen
majors. Some similar verses may be cited, but the resemblance is not
very close:
N6e5

ww

ww
ww
j x d D
N2s5
ww
w
ww

k d w d

The nearest parallel is found outside the eighteen majors. In one of
the normal D/e odes, there is a very similar verse:
I3/4e1

ww
ww


ww
k
D
d
As is typically observed in
D
D, the diptych construction is
common in D/e. But Pindar avoids it in non-D/e. Or rather, what
Pindar avoids is D itself. Except in N6, D is very rare (see List 2).
Apart from our verse, the examples of D preceded or follwed by d are
conWned to N6. Anceps D is much rarer, two in total; the other is
found at P2s6, which is the expansion of anceps d in the preceding
verse. There is no other example of x D at verse-beginning, although
264 The Eighteen Majors
seven verses start with x d and four with x D . The anceps is regu-
larly short, which is diVerent from the normal practice in ordinary
D/e. The other anceps, which is placed between D and d, is long
except in one instance (v. 27 h
 Ie ). Bridge is regularly
observed between the anceps and the following d.
s4. See Part I, 6. B (^ e d, not
ww
), 6. D (bridge after long
anceps), 8. A. 3 (ii) (d) (longer verse), 8. A. 6 (c) (palindrome), 8. C. 2
(double d). Symmetrical and asymmetrical phrases of Wve or six
positions are connected together neatly. A long verse made up of
them is one of the most distinctive characteristics of Class III, and
certainly becomes Pindars favourite later. Even this being admitted,
the similarity of P10s4 with N3e3 is astonishing:
P10s4 w

ww
u
ww

w
ww

w
w
k ^ e dxd rdod e 2
N3e3
ww
w
w
ww

w
ww

wrw
k dodwd rdod e 2
Three-quarters of these two long verses are completely identical for
as long as 16 positions. There is a diVerence only in the initial
quarter: ^ e d in P10s4 contra dodrans in N3e3 (and both of these
have 6 positions; ^ e d is equal to dod when ^ e and d are inversely
placed). And there is another parallel:
P2e3 w

ww
w
rw
ww

w
w
j gl rdod e 2
Not only its latter two-thirds are identical (rdod e 2), but the initial
phrase starts with w
(should this phrase be analysed not as glyconic
but ^ e dod?).
The link anceps after the Wrst d is not regularized. It is short at Wve
repetitions and long at three, 10  I
, 22 P, 58
e. Word formation is often localized. Bridge is, in general,
observed between in two true longs: between ^ e and d at all eight
repetitions, between d and rdod except for 28 I j), between
rdod and e 2 with two exceptions, 22 Icj and 64 Kaj (and further,
46 dj, 58 Kj). Two coincidences of word-division are remarkable:
after the initial w

j in all repetitions, 58  included) and in
the rdod,
wj
ww
except 28
j (interestingly, the single
short position is Wlled by  in 3 instances).
s5. The phrase
w
w
ww
is often ambiguous between wila-
mowitzianum and ewd (Part I, 7. 3). To be consistent, I analyse it as
wilamowitzianum in every verse. Thus s5 is ^ e and wil 3. But
 Pythian Ten 265
ambiguity is more prominent in this verse than elsewhere. There are
in total 23 examples of ^ e, and none of them is followed by an aeolic
phrase (Part I, 6, B). All the others are followed by freer D/e phrases
like e, d, or D. The analysis
w

w
w
ww
w

k ^e e hag
is attractive. If ^ e e is accepted, the remaining part will be hage-
sichorean (for ^ e e aeolic, see Part I, 8. B. 5). Hagesichorean is
unusual, but appears once in a similar context:
P2s8

ww
w

w
w
w
ww
w

k tel e2hag
Both P10s5 (if this colometry is accepted) and P2s8 end with a hage-
sichorean, which is preceded by e (P10s5) or its expanded form, e2
(P2s8). In a sense, the colometry of P2s8 above is a compromise so as
to analyse a long single-short movement a double short, which is
quite normal so far, but here further prolonged by w

. Whichever
analysis is accepted for P10s5, it is certain that this verse is unusual.
The initial w
cannot be aeolic base: there is no example in Pindar
or elsewhere of such an aeolic decasyllable as
x
x
ww
.
s6. 6, D (aeolic phrase long mid-anceps). See above, Strophe, and
Textual problems. Anceps e2 is rare, but compare:
N3s8 ww
ww
w
w
w
w

k tel w e2 x
P8s7 x
w
x
w
w
k x e x e2
Notably, the Wrst phrase of N3s8 is telesillean starting with ww in the
same manner as our verse; but the anceps is always short. The pivotal
position of the link anceps in our verse is discernible in bridge and
cut after it. When the anceps is short (6 repetitions), word-end may
fall after it (4 repetitions), but, when the anceps is long (2 repeti-
tions), word-end never so falls; bridge between the anceps and the
following long is obligatory. See Part III, B.
Verbal assonance: 60    66   .
e1. See Part I, 8. B. 4 (d aeolic). Anceps d is common at the
verse-beginning, but it rarely precedes an aeolic phrase. There is only
one parallel:
I7e7

ww

w
ww
k x d rdod
266 The Eighteen Majors
Moreover, there is no other example of glyconic preceded by d at any
position of the verse. Both in P10e1 and in I7e7, bridge between xd
and the aeolic base is strictly observed. The initial anceps is realized as
short except for v. 67
H. In the base of the glyconic,


appears only at v. 49 
 P.
e2. See Part I, 8. A. 3 (shorter verse). I take this phrase as the
minimal length of aeolics, but whether it is really adonean (as a short
aeolic phrase) or, alternatively, d anceps (as a freer D/e phrase) is
not easy to decide. There is one identical verse, I7e2, where the
metrical context more plausibly suggests adonean. Here the context
does not help; d is used in the preceding and the following verses, and
aeolic phrases are used extensively in others.
e35. See Part I, 8. B. 4 (aeolic d); 8. C. 8. (d e). Three
successive verses start with the identical phrase, a heptasyllable.
Comparable is N4s13 (hepta 2 d k hepta k hepta wil k). The
transition is easy to understand: excision of d (from e3 to e4) and
change of e to bacchiac (from e4 to e5). Surprisingly, however,
neither of these simple transitional processes recurs in the eighteen
majors, though each verse on its own is well paralleled. The initial
anceps of the heptasyllable is not regularized in any verse: short is
used in respectively one, two, and two repetitions out of four. In e3
bridge is strictly observed, both before and after d.
e6. See Part I, 6. D (bridge after long anceps). There are two other
examples in which anceps e (iamb) precedes an aeolic phrase
starting with half base:
O9e8

w
x
w
ww
x
w
ww
w

k
e hepta
hepta 3
I7s3/4

w
x
ww
w



ww
w
w

k
e tel gl 3
The initial anceps is short except for v. 36  and so is the second
except for v. 72 d.
 Pythian Eleven 267

PYTHIAN ELEVEN

Four triads. Class III

P11s1

w
ww
ww
ww
w


k
ribyc 2
( (hepta 2)2d ) sp
P11s2a

w
ww
j hepta
P11s2b
w
wwrw

w
k gl e
P11s3 rw
wr w
ww
k e2wd
P11s4
atww


wr
ww
k rdod
e d
[?
atww


rw
ww
k rdod sp rdod]
P11s5

w
ww
w

w
w
k hepta 2 e2
P11e1
ww
w

j ar
P11e2 www
ww

k ph
P11e3

ww
wj

w
w
k tel e2
P11e4 www
ww
w

jw
k gl e
P11e5 w
w
w

wr
j we2e

P11e6 u
w
a
ww

k xexd

s1 H 54;36 B 1,37 6, 49; s2b H 39b; B 18b, 23b, 55b; s3 B 8;38 s4 H 52, 57; s5 H 10, B 26,
42;e2 H 27; B 59; e3 B 29; e4 H 46

For the alternative analysis of s4 in square brackets, see the last

section.
For the analysis and the notation of s1 (
ribyc ), see Part I, 7. 6
(expanded aeolic), and also Appendix B.
s1 1 (1)
 , # b 

 IA ,
s2a 2a (2) "g b 
s2b 2b
A 
 ,
s3 3 (3) Y f !  I
s4 4 (4) d
a  H K  

s5 5 (5) , n
 K  ,
e1 11 (11) 

 
e2 12 (12)
 IH   ,
e3 13 (13) K H  A  

36
The text is uncertain at the beginning of the following verse (v. 55).
37
It is unnecessary to change IA to IA (Christ in his apparatus, adoted
by Snell); cf. Barrett, Collected Papers, 187.
38
Snell does not recognize brevis in longo at the end of v. 8; see below, strophe.
268 The Eighteen Majors
e4 14 (14)  

 ,
e5 15 (15) K IE I

e6 16 (16) H 
 .
Pythian 11 creates a confusing impression. It is, in a sense, a metrical
potpourri. Aeolic verses, made up in a typical Pindaric manner, are
juxtaposed with typical freer D/e verses; for example, s2b with s3, or e3
4 with e56. A reversed dodrans has a strange resolution of the chori-
ambic nucleus, complicated by free responsion at the second position
(s4). Moreover this aeolic phrase is followed by long anceps and freer
D/e phrases (e d). Most extraordinary, and hard to analyse, is the Wrst
verse. I classify this ode as Class III, not so much because it has
characteristics common to Class III stanzas, but because it includes
everything. Besides P10, P11 is most crudely structured.
Resolution is frequent, but it is not always easy to identify the
resolved positions; see s4 below. RSS is 53.3% in the strophe, and
50.0% in the epode. Verses are in general short. The average number
of positions is 12.0 in the strophe, and 9.5 (the second shortest in the
eighteen majors) in the epode. The stanza-forms themselves are short.
The epode is the shortest (57 positions) of all the non-D/e stanzas.

Strophe
As stated above, there are a number of peculiarities. Nevertheless, a
rationale is discernible in the strophe through the scheme below, in
which resolution is eliminated. A diVerent analysis is given by Snell
and eloquently explained by West, GM 623; but it seems to me
strained. For Snells somewhat idiosyncratic line-division see below:

ww
ww
#
s1

w
ww
w


k
s2a

w
ww
j
s2b
w
ww
w

w
k
s3
w
w
w
ww
k
s4
a
ww


w

ww
k
s5

w
ww
w

w
w
k
[or s4
a
ww




w
ww
k ]
 Pythian Eleven 269
This chart is based on the assumption that, when the Wnal w
is cut
oV from glyconic, the phrase can be felt as equivalent to reversed
dodrans:

w
ww
[w
]
w
ww

This applies only to one form of glyconic, but that is the only form
used in this stanza. The basic phrase of the stanza is found in the
former half of s5. I call it hepta 2, but, in this chart, it may well be
analysed as anceps reversed dodrans w
. This phrase is
repeated, with modiWcations, from beginning to end. In s1, its central
part is expanded by two dactyls. Elsewhere, either the Wnal w
(s2a;
hepta) or the initial anceps (s2b; gl) or both (s4; rdod) is missing. In
s3 the phrase is augmented by single-short movement. After the basic
phrase comes e (s2b) or the like: spondee (s1) or e2 (s5). The latter
half of s4 is irregular.
Snell combines into one long verse not only s2a s2b but
s3 s4.39 The former is, in a sense, a matter of taste but the latter
involves a theoretical problem. He does not recognize brevis in longo
at the end of v. 8. He thinks that the Wnal  of  followed by a
vowel (Z) is given a syllable-closing pronunciation. Thus, accord-
ing to him, the syllable  is long by position, like all the other seven
repetitions, and, consequently, s3 s4 is one verse. However, it is
much more natural to admit brevis ( verse end) and to separate s3
from s4. Metrical lengthening is very rare and should not be lightly
introduced. It may possibly occur at v. 38 in this epinikion: 

K: . . . But this is the only case in the eighteen majors (or in all the non-
D/e epinikia); see below, Textual problems. For the further discussion
accompanied with the assessment of Fuhrers argument, see P5,
Textual problems, v. 42.

Epode
The structural principle of the epode is obscure. Even verse-end is
not guaranteed at the end of e1 or e5 (here Snells mark k is wrong)

39
Note that the verse number allotted in his metrical chart and the line number in
his main text are inconsistent. v. 5 in the Wrst strophe is the fourth verse in his
metrical analysis.
270 The Eighteen Majors
and, moreover, there are two other instances of coincidence of word-
end within a verse: e3 and e4. In the colometry, however, echoes of
some parts of the strophe are discernible. Two verses are almost
identical with verses in the strophe:
e3

ww
w

w
w
k tel e2
s5

w
ww
w

w
w
k hepta 2 e2
e4 www
ww
w

w
k gl e
s2b
w
wwwww

w
k gl e
Inside the epode, too, there are some echoes. e1 and e2 are enlarged
into e3 and e4 respectively:
e1
ww
w

j ar
e3

ww
w

w
w
k tel e2
e2 www
ww

k ph
e4 www
ww
w

w
k gl e
Here the pherecratean of e2 is analysed as rdod anceps, like some
examples in the Class III strophes. The similarity between it and e1 is
clear, for the aristophanean is dod anceps.
If e1 and e2 were combined into one verse, as by Snell, the
colometry would be diVerent:

ww
w

wr
ww

k dod e d

Then the sequence, eww d, would be identical with that of the latter
half of s4:
s4
atww


wr
ww
k rdod
e d
This might be a merit, but on the other hand this colometry destroys
the inner structure analysed above. The separation of e1 from e2 is
preferable.
None of the aeolic phrases is used in the last two verses (e5 and e6).
They are distantly related to s4.

Textual problems
2b (s2b). The paradosis  is unmetrical;  Byz.
4 (s4). The paradosis N is unmetrical; K Triclinius, but cf. Indi-
vidual verses, s4.
 Pythian Eleven 271

6 (s1). The paradosis =E is unmetrical; 

(Hermann, Notae) is semantically better as well.
10 (s5). The paradosis B introduces an undesirable hiatus;
so,  (Heyne).
18 (s2a). The paradosis H may be right. The initial anceps of the
heptasyllable is occasionally short; for example, P10e4 (short at one
repetition, v. 16 Ig, out of four. But H (E. Schmid) is a very easy
correction and the form Pindar usually uses. It should be adopted.
23b (s2b)  must be retained. (The paradosis is ;
 is a Byzantine conjecture, according to P. Finglass.) Schroeder
regularizes the aorist stem of this verb and introduces ; ed.
maior, Prolegomena II, 62 (p. 32). The second position of glyconic is
theoretically anceps, but in Class III strophes it is very rarely long.
Moreover,  is attested in the manuscripts at P8 v. 33. Regular-
ization is unnecessary.
25 (s4). e c is unmetrical; e b Pauw. Gentili keeps the paradosis
by analysing the latter half of s4, our
e d, as dimeter polyschema-
tistos. According to him, both E; e c  I (25,


w

ww
, with  scanned as one syllable) and - Ia
 f @ (36,


w
ww
, see below) can be in respon-
sion with the other repetitions:

wr
ww
. The Wrst half of
wilamowitzianum ( choriambic dimeter, polyschematist), however,
has not such a freedom; see Itsumi, Choriambic Dimeter.
34b (s2b). The word-order of the paradosis,  , is
unmetrical. Byz. transposed.
36 (s4).  is Triclinius emendation (but G has it too, accord-
ing to P. Finglass), possibly for a metrical reason. It is adopted by
many.  f @ is comparable with O9 v. 82 ( 76 Sn.) P
 K @. Most manuscripts have . It may be kept, see the Wnal
section.
37 (s5).
  (Moschopoulos) improves the metre, because
the consistently long anceps certainly accords with the Pindaric
manner. It is stylistically better too than the paradosis 
.
38 (s1). This is perhaps the only case in the eighteen majors (or in
all the non-D/e epinikia) in which prosodical lengthening of the Wnal
syllable by consonant (
 K . . .) may occur. P5s11 (v. 42

 
) is quite diVerent and under suspicion (see
ad loc.; Fuhrers argument is summarized there). Our case is more
272 The Eighteen Majors
plausible. Even Maas, who categorically denies a freer metrical
responsion (so-called anaclasis), rejects any textual emendations
for our case: (e.g. I
  (Hermann; accepted by
Schroeder and Turyn), Die zur Beseitigung der drei gelangten End-
silben [the others are Ba. 5. 187, 10. 113] vorgeschlagenen Emenda-
tionen verschlechtern alle den Text, Freiheiten, 18). When both
textual emendation and metrical explanation are rejected, prosodical
licence remains as the sole explanation. Snell follows him (see
Metrorum conspecutus D). Fuhrer, after thorough re-examination,
conWrms the prosodical lengthening for our case. This passage shares
these characteristics in common with some of the most certain
examples in D/e odes:
(i) The lengthening in question occurs in .
(ii) The syllable in question is placed at the end of double-short
movement. The inXuence of epic is certain.
(iii) Normal rhythm has already been repeated several times (Wve, in
our case) before this irregularity occurs.
At the same time the following peculiarities should be noticed:
(i) Prosodical lengthening is extremely rare in Pindar even in D/e.
(ii) Whatever analysis is given, the verse P11s1 itself is metrically
very unusual.
(iii) In its six repetitions the metrical irregularity is made more
prominent by four-syllable words (K
 in our case; see s1
below) located at the end of the verse.
41 (s4). The paradosis H 
 is one syllable too
short. Christs E is adopted by Snell and Turyn. For other
possibilities, see the Wnal section.
43 (e1).  is Triclinius emendation ( MSS).
46 (e4). K is Triclinius supplement.
53 (e6). f is deleted by Triclinius. Christ adopts Bergk23s

 instead of  (BV; 
 rell.) and keeps .
In his 4th edition Bergk retracts the emendation (praepositio novis-
simo versus loco incommoda).
548 (ant. 4) are a monstrous textual crux (David Young, Three
Odes, 2). There are considerable diVerences between editors even
about what these sentences mean. The biggest issue is whether to
 Pythian Eleven 273
delete  (55) or not. But, in a sense paradoxically, the metrical
scheme is not aVected, for every editor tries to give a text which
metrically corresponds to the established scheme. I resign myself to
leaving the question unsolved.

Individual verses
s1. This verse is peculiar in two respects: the dactylic expansion and
the Wnal spondee. But the two verses in O9e suggest it is not as
strange as it looks:
O9e4



ww
w


k sp tel sp
O9e7

w
ww
ww


j
ribyc (heptad) sp
P11s11 starts in the same manner as O9e7 and ends like O9e4. It is
almost symmetrical, or palindromic, with the expanded nucleus
situated at its centre:


w
ww
ww
ww
w


k
There is no dactylic movement elsewhere in the stanza. This is a
diVerence not only from the other Wve examples of expanded aeolics
in the eighteen majors (Part I, 7, 6) but also from the examples in the
Paeans (Part II, Appendix B).
The eVect of the Wnal spondee is enhanced by long words located
there: 1 IA , 17  , 22 K
P
, 33

(or
), 38 K
, 54 I. Perhaps ,
an indispensable word in this ode, may have prompted the poet to
introduce a spondee at the end of the verse: words of the shape
w


can indeed be located in other positions of other metres
indeed; for example, over two aeolic cola from 2 ending to full
aeolic base, and above all, in normal D/e, by locating its penultimate
syllable on link anceps, but the spondee is essential to locating the
word in question at verse-end.
s2a. A verse made up of a single heptasyllable is not rare (N4s2,
P2e7, P8s3, P10s2).
s2b. See Part I, 5. F (resolution of aeolic phrase). There is one
parallel for the glyconic with sixth position resolved:
P8s2
w
wwrw
k gl
274 The Eighteen Majors
No example is found in Pindar of the responsion between resolved
and unresolved at this position. The Wve successive shorts are occu-
pied by one word at v. 2b 
 and v. 23b 

 and by
a metrical word at 7b ()  K
; 55b   ; for the
other examples of successive shorts, see Part III, D. Including these,
six repetitions out of eight have word-end after the successive shorts.
gl e is a familiar combination; cf. Part I, 8, B, 5. Bridge is observed
here at all the repetitions.
s3. See Part I, 6. C. 3 (resolved long followed by short anceps), 8. C.
3 (enwd). Continuous shorts created by resolution occupy the central
part again in this verse. P11s3 is one of the verses made up of a long
single-short movement followed by wd; see the complete list at Part I,
8. C. 3. O1e3 has a sequence of the same length, and P2s5 is very
similar in construction.
P11s3 rw
wr w
ww
k e2 w d
O1e3
w
w
w
ww

w
k e2 w d e
P2s5

wrw
w
ww
k x e2wd
Resolved e2 is quite common, but the form rw
wr has no parallel.
But the sequence w
wrw
ww
is also incorporated in the fol-
lowing verses (marked by ):
N3e1b w rw
w
wr w
ww
w
w

w
k we3wdwe e
I8s3
w
w
ww

w
wr w
ww
w
k wil e2tel
In P11s3 at seven repetitions out of eight, word-end falls at

wrjw
ww
in this sequence (the exception is v. 40; 
j). The identical division occurs at all seven repetitions in
I8s3, but never in N3e1b.
Alternatively, P11s3 can be analysed as e wrdod (or e hepta). This
analysis harmonizes with other verses as illustrated under Strophe. It
is especially recommended if s4 is analysed in the alternative manner
set out below. However, the similarity with O1e3 etc. is lost.
s4. See Part I, 5. E. 1 (rdod starting with
x), 5. F (resolution of
aeolic phrase), 6. D (aeolic phrase long mid-anceps); 8. C. 9
(e d). As many as three idiosyncrasies are packed into this verse:
(1) responsion between long and short at the second position of the
reversed dodrans; (2) resolution at the left of its nucleus; (3) long
link anceps between rdod and e.
 Pythian Eleven 275
(1) and (2) are related. Eight repetitions are classiWed into three
groups:
wr (4 repetitions) 9 Z , 41 E, e b, 52 H
a Ia
, 57 ;
w
(2 repetitions) 4 d
a
, 25 ;


(2 repetitions) 20 
, 36
F. The last two cases are proper nouns. Thus, long syllable
at the second position is acceptable as a special licence for proper
nouns. In contrast the resolution of the nucleus is extraordinary,
being is almost unparalleled in the eighteen majors. There are three
other examples, N6e2, P8s2, P11s2b. In these, resolution occurs at all
the repetitions; in P11s4, on the other hand, it occurs only at half the
repetitions, and in the other half does not.40 Thus, strictly speaking,
what is peculiar in our verse concerning (2) above is free responsion
between resolved and unresolved at the left of the nucleus.
The resolution is followed by the two original shorts of the nu-
cleus. As a result, there are Wve successive shorts in these four
repetitions, the same number as in s2b. Presumably Pindar sought
a certain eVect by the use of runs of shorts in three consecutive verses:
s2b
w
rwr

w
k gl e
s3 rw
wrw
ww
k e2wd
s4
wrww
. . . . . . . . . k rdod . . . . . . .
but then why did he not resolve the position in more of the repeti-
tions of s4?41
As for the link anceps (3), short anceps is the rule after the nucleus
or d. The combination of long anceps e itself is not rare when it is
preceded by e or e2. However it is highly unusual that it should be
preceded by
ww
. The exceptions, besides our verse, are two:
O10e10 w
wwt

www
w

ww
k w d
e e d
(this verse has many irregularities, see ad loc.)
N3e1a
ww
w
ww
a
w

j d w d x e

(long anceps occurs once out of four repetitions)
40
The number of occurrences of resolution may be changed by slight emend-
ations: 52 Ia may be i (
) (Hermann, Notae), which reduces the number. 4

a may be
a, which increases it.
41
Hermann (Emend. P.) adjusts all the other repetitions to
wr and restores
exact responsion: besides 4
a, 20 N (for 
), 25 K


 n  I , 36 e $
e
 (for F
). Certainly,
nobody could believe in these rewritings. The old Hermann reveals vehement intoler-
ance of any irregular metrical responsion, especially in Emend. P. At the same time we
should not forget that there are still good emendations in this article.
276 The Eighteen Majors
Strictly speaking, these are not parallels. In our verse, the phrase
preceding long anceps is not d but an aeolic phrase (rdod). Bridge is
observed between the anceps and the following long throughout all
the repetitions both in our verse and in O10e10. To eliminate this
irregular long anceps, an alternative colometry is conceivable. See the
Wnal section.
s5. The enneasyllable, hepta 2, is employed in two other verses
(N4s1, P8e1). In P8e1 it is followed by e 2, as in our verse. Bridge is
strictly observed between hepta 2 and e 2 (there is one breach in
P8e1; Part I, 8, B, 6).
e1. Aristophanean is a rare verse in Pindar. There is only one
example elsewhere:
O1e7 w

w

ww
w

ww
w

k ^e e dod ar
e2. Pherecratean is less rare, and in three examples out of 11 stands
alone and makes up a verse (O1s4, P1s1, and P11e2). But the
pherecratean starting with www has no parallel.
e3. Bridge is always observed between telesillean and e 2; cf. Part I,
8. B. 6. There are in all three verses composed of tel e 2, and two of
gl e 2.
e4. As in the previous verse, bridge is observed before the suYx.
Glyconic starting with www is common, and so is gl e.
e5. See Part I, 6. C. 3 (er anceps at verse-end)
Two analyses are possible, depending on which position is re-
solved:
(i) w
w
w

rw
j w e 2
e
(ii) w
w
w

wr
j w e 2 e

(ii) is preferable to (i) because (i) introduces cut after long link
anceps (31 l j  , 47 I j

). Cf. the ana-
lyses of O10s3a:
(i) u
w

rw

rw

j x e
e
e

(ii) u
w

wr

wr

j x e e
e sp
Here too (i) is to be rejected for the same reason, see Part I. 6. C.
Parallels for the resolution at the penultimate position of a verse end
are collected and discussed in Part I, Appendix, Addendum II.
 Pythian Eleven 277
The combination of e2 e has parallels: O1s10, O2s6/7, O2e3,
I8s10. The anceps at the beginning of verse is always short; thus we2
may be ^ e3.
e6. There is one identical verse in the other odes, and one similar
one:
N3s6

w
x
wwy
k xexdx
O9s11

ww

w
x
ww

k x d e x d x

Alternative Colometry for s4

Leaving the irregularities in the Wrst half aside, the long anceps in the
other half can be eliminated by another colometry:
s4
wtww


rw
ww
k rdod sp rdod
The second rdod is the same as in the latter half of the preceding
verse:
s3 www
w rw
r
k
Reversed dodrans is a key phrase of this ode, and its recurrence
matches better than
e d (see the introductory chart above,
Strophe). Moreover, it saves  (36), without emending it un-
necessarily into , because
w
ww
can correspond with
rw
ww
(the Wnal of Ia making position before ), although
there are no examples of this responsion in the eighteen majors. This
analysis is what Hartung oVers. But instead, this analysis introduces a
spondee in the middle of the verse. This is highly unusual, and the
colometry feels artiWcial. There may be a parallel for the spondee in
mid-verse, but the analysis of this verse is uncertain; see ad loc.:
O10s3b ww w



ww
ww
k e sp D
Spondee is used in s1 indeed, but at the verse-end. Rather, if it were
possible, it would better to set verse-end after the spondee in s4 too.
Then s4 would be divided:
[s4a]
wtww


k rdod sp
[s4b] yw
ww
k rdod
278 The Eighteen Majors
There are some textual problems in the manuscripts here and there.
Word-end is certainly attested after the spondee at Wve repetitions
out of eight: 4  j, 20 H j, 25 E j (sentence-end), 52
$ j, 57 lacuna? (the following j 
 gurantees the
word-division there). According to the text of Snell and Turyn (and
others), there are three repetitions in which word-boundary does not
coincide with the end of [s4a]. In fact, it does coincide at two
repetitions in the paradosis.
At 36, the paradosis  is blameless.  is Triclinius
emendation, and most editors adopt it. If asyndeton were acceptable,
delete Ia instead. Then F
  k hKi  f
@ would be a simple solution, both brevis and hiatus being
supposed at the end of [s4a]. K introduces a long in the place of
double short at the initial of the rdod. Admittedly, this corresponsion
is unparalleled in the eighteen majors, but it may be related to the
irregular resolution in [s4a].
At 41, Ej is Christs emendation. The paradosis is H, and
it covers the spondee. However, the text is one position too short.
Before Christ, editors supplemented the latter half: H j
hi
 (Boeckh) or H j
  (Her-
mann).
The repetition where word-end does not occur after the spondee
is 9. I tentatively suggest the followings: Z   A E (or
a ) d hei O.42

42
Bergk2 too retains the paradosis at 36 and 41,  and H, but at the same
time emends the text in all the other repetitions: 4 N deleted (K Triclinius), 9 
(H
paradosis), 20 f deleted (omitted in. V), 25 b deleted. 52 {a} 

  ( paradosis, deleted by Triclinius). The metre he oVers is:

wtww


ww
ww
k rdod
D
This colometry is not good, however: D is never combined with an aeolic phrase in
the eighteen majors.
 Nemean Two 279

NEMEAN TWO

Five strophes (monostrophic). Class I

N2s1 w

ww
w
k gl
N2s2

ww
w
w

k tel 3
N2s3
u
ww
w
u
ww

j gl reiz
N2s4 www
ww
w



ww
w

u
ww

k gl gl ph
N2s5
ww
u
ww

k dxd

s1 B 16; s2 H 2, 22; B 7; s4 H 4, 24; B 19; s5 H 5, 15; B 10

s1 1 (1)
+ 
 d !
s2 2 (2)

H K
 a
 I
s3 3 (3)
, e K
, d  I
s4 4 (4)
a H I  
H,

s5 5 (5) K
 e .
Nemean 2 is one of the simplest of all the Pindaric epinikia. It is
monostrophic, and composed of the smallest number of verses
among the eighteen majors. Its metre is aeolic in a plain structure.
Particularly noteworthy is s4, in which three aeolic phrases, two
glyconics and a pherecratean, are connected with total avoidance of
word-end at junction. The construction resembles that of a tragic
verse, and is typical of Class I (aeolic). Another characteristic of Class
I is the long anceps in the second position of full aeolic base (s3, s4).
Except for s1, all the verses end pendent. This is fairly unusual in the
eighteen majors. RSS is comparatively low (45.5%).

Textual problems
19 (s4). The paradosis
a introduces the improbable responsion

w in the base;
a (Triclinius) is necessary.
24 (s4). P (Callierges) is rightly scanned thus. The paradosis
is PE.

Individual verses
s1. Some Pindaric glyconics which start with base w
could be taken
as acephalous e dodrans (e.g. O1s1, P5e2, N7s1 cf. ad locc.). But
280 The Eighteen Majors
N2s1 is most certainly glyconic because of the aeolic context and the
absence of e elsewhere in this ode.
s2. In the eighteen majors there are in total six examples of tel 3,
half of which are concentrated in I7. This phrase usually, as here,
makes a verse by itself.
s3. Neither hiatus nor brevis is present, but the pendent ending
and the aeolic base of the following verse conWrm verse-end. As is
occasionally found in typical aeolic phrases, ancipitia are realized as
long or short indiscriminately at each repetition without any pre-
dominant tendency. The aeolic base of the glyconic is

three
times,
w twice (cf. Part I, 5. C), and the ratio of the initial anceps of
the reizianum is three long to two short (cf. Part I, 5. D). Word-end is
strictly avoided between glyconic and reizianum in all Wve repeti-
tions, in three of which dovetailing is observed.
s4. See Part I, 5. C. 1 (the base), 8. A. 3 (i) (longer verse), 8. A. 7
(repetition within a verse). Combinations of gl gl or gl ph are
most common with Anacreon and in Attic drama, but surprisingly
scarce in the Pindaric corpus. The second glyconic starts with the full
base

at all the repetitions. The rarity of this type of base is
another peculiarity of Pindar. Bridge is observed at all ten junctions
between two phrases (Part I, 8. B. 1). Between the two glyconics
dovetailing is found at four repetitions out of Wve, and between the
glyconic and the pherecratean at three.
s5. See Part I, 6. D (bridge after long anceps), 8. A. 6 (c) (palindrome);
8. C. 2 (double d). The latter half of this verse may be either reizianum, as
in s3, or anceps d anceps. Whichever analysis is accepted, there is no
exact parallel for the verse as a whole. Reizianum is employed at s3 after
glyconic. If we accept the parallelism and the latter half of s5 is analysed as
reizianum, the former half of s5 may be regarded as the shortest phrase
among aeolic cola. I prefer to take it as d, consequently the latter is d and
anicipitia. Thus a residual ambiguity (Part I. 7. 1) creeps into this simple
stanza-form. There is a parallel for the double d:
N3e1a
ww
w
ww
a
w

j d w d x e

On the other hand there is no example of d (or xd) preceding the half base
of an aeolic phrase in the eighteen majors. The anceps between two ds is
long except in v. 20 (). There are in total seven verses which
start with d. Of these, only our verse and N3e1 have d followed by anceps.
 Nemean Three 281

NE MEA N TH RE E

Four triads. Class III

N3s1

ww
w
w
ww

w
k telwd e
N3s2
w
w
wrw
w
rw
k e 5e
N3s3 www
ww
w

wrw
k gl e2
N3s4 w
w

w
ww
w

w
k we gl e
N3s5

w
w
wr w
ww
k
e3wd
N3s6

w
a
wwt
k
e x d

N3s7
w
w
ww
w

w
k wil 2 e
N3s8 ww
ww
w
w
w
w

k telwe2

(? ^ Dwe3
)
N3e1a
ww
w
ww
a
w

j dwdxe

N3e1b w rw
w
wr w
ww
w
w

w
k we3wdwe e
N3e2
w
ww

w
ww

k rdod ph
(? rdod rdod 1)
N3e3
ww
w
w
ww

w
ww

wtw
k dodwd rdod e2
N3e4 ww
ww
w
ww
ww
w

w
k tel tel e
(? ^ Dwibyc e)
s1 H 9, B 9, 29, 69; s2 H 2, 22; B 2, 22, 62; s3 H 51; s4 H 64; s5 H 13; s6 H 6, 26; B 34,
46; s7 B 55; s8 H 68; e1b H 17b; e2 H 18, 38; e3 H 19; e4 B 20

s1 1 (1) *
 E, A , ,

s2 2 (2) a
 K  

s3 3 (3)   A `Y o

s4 4 (4)  K

  

s5 5 (5)  ,  Z
.
s6 6 (6) B b
A  b ,
s7 7 (7) I b 
 Ia E,
s8 8 (8) 
 IA  
 O

e1a 17a (17)
   b
A
e1b 17b (18)  $e K 
  e  .
e2 18 (19) N  Kg e   K  A
e3 19 (20) I $

 K

E 
  ,
P

e4 20 (21) I
  o
 ! 
A P ,
282 The Eighteen Majors
str. 2. 218 (22)(29) ant. 2. 2936 (30)(37) ep. 2. 3740 (38)(42)
str. 3. 418 (43)(50) ant. 3. 4956 (51)(58) ep. 3. 5760 (59)(63)
str. 4. 618 (64)(71) ant. 4. 6976 (72)(79) ep. 4. 7780 (80)(84)

The style of Nemean 3, of strophe/antistrophe and epode alike, is a

representative sample of Class III. Abundant short syllables are con-
spicuous. Frequent resolutions apart, anceps is very often realized as
short. RSS is high: 53.1% (strophe) and 54.4% (epode). Short anceps
and true short are diYcult to identify in some cases. Three verses (s2,
s5, e1b) start with a long sequence of single-short movement (e 5, e 3,
e 3 respectively), and one (s8) ends with it (equivalent to e 4, for its
analysis, see below). Also remarkable are two palindromic sequences,
which are quite uncommon, especially the former:

ww
w
w
ww
(s1, e3)

ww
w
ww
(e1a)
There are many lengths of aeolic phrase in this ode (the indicative
phrase, I `N K
E PH (76 79 Sn.), is
discussed with O1 v. 102; see ad loc.). Anceps in the aeolic phrases
is predominantly short, except at the beginning of the verse. When a
phrase starts with full aeolic base, the two positions are never Wlled
with

, so that long syllables (sporadically resolved into two
shorts) and short syllables take regular turns, contrasting with the
double shorts of the choriambic nucleus. And the phrase
ww
,
which can be analysed as d of freer D/e, appears here and there. It is
always preceded or followed by short anceps; that is, it is never
Xanked by two long syllables. Thus the movement runs very lightly
from beginning to end.
The eVect of lightness is enhanced by the frequent attachment of
suYx e. In Pindar it is common for an aeolic phrase to be followed by
e, and, to a lesser extent, by e 2; see Part I, 8. B. 56. But in this ode the
extensive use of these is greater than usual. Two-thirds of all the
verses (8/13) end with them. And three of them are resolved, regu-
larly or partly, at some repetitions.
Dactylic movement is totally absent. But there are three examples
of telesillean starting with two shorts, which produces two consecu-
tive double shorts. SigniWcantly, the last verse of both strophe/anti-
strophe and epode starts with this type of telesillean. The identical
Wnale gives unity to the two stanzas. These telesilleans may be a
 Nemean Three 283
derivative of D in the normal D/e; see Part I, 7. 6. Consequently, s8
and e4 can be interpreted as a variation of freer D/e:
s8 ww
ww
w
w
w
w

k
e4 ww
ww
w
ww
ww
w

w
k
(A diVerent colometry is conceivable for these verses, which is
indicated in parenthesis in the chart above; see on e4 in individual
verses.)
Verses are in general long. The average number of positions is 13.4
(strophe) and 17.2 (epode). The latter is the biggest in the eighteen
majors.

Strophe
Every verse in the strophe has as its main body a sequence in which
long syllables and short syllables regularly alternate. At some places a
double short, which is either choriambic nucleus of an aeolic phrase
or d of freer D/e, replaces a single short. The sequences are long.
Occasionally anceps or short phrases are attached before or after.
This structure is more evident in the strophe than in the epode, and
will be easily visible in the following chart in which resolution is
eliminated and the irregular long anceps at v. 46 ( 48 Sn.) in s6 is
ignored:
s1

ww
w
w
ww

w
k
s2
w
w
w
w
w

w
k
s3
w
ww
w

w
w
k
s4 w
w

w
ww
w

w
k
s5

w
w
w
w
ww
k
s6

w
w
ww

k
s7
w
w
ww
w

w
k
s8 ww
ww
w
w
w
w

k
Three verses (s1, s5, s6) start with long anceps. Two verses (s6, s8)
end with it. Some short phrases precede or follow the main body:
w
w
at the beginning of s4,
w
at the end of four verses (s1, s2,
s4, s7) and
w
w
at the end of s3. Except for these, there is no
place where two long syllables are juxtaposed, whether the position is
a real long or a long anceps. All the central sequences are quite long,
284 The Eighteen Majors
and regularly constructed. Even the shortest has four longs (s3, s4,
s6). The longest has six (s2, s5, s8). Two verses have two double shorts
(s1, s8), one has none (s2), but the others do not lack one. Double
short tends to be located towards the end: before the Wnal long (s5,
s6, and one of s1) or before the second Wnal (s3, s4, s7). Each of the
verses has its own characteristics but at the same time they resemble
each other.

Epode
Unlike the strophe, the Wve verses of the epode are not constructed in
a uniform manner. Only e1b is of exactly the same type as those of
the strophe, and has a long sequence of regularly alternating longs
and shorts as its core, which is followed by e. Others have similar
sequences only in part. Two juxtaposed longs are found to a greater
extent. For example, e3 starts in a very similar manner to s1, but its
latter half is diVerent. A reversed dodrans, which is itself followed by
e2, follows the sequence. Thus there are two collisions of two longs in
this verse, and the long verse is chopped up into shorter phrases.
Reversed dodrans is repeated in e2 as well (if it is assumed that the

w
ww
in the pherecratean is equivalent to it). Repetition is
certainly an important motif of the epode. e4 has the same telesillean
as the initial phrase of s8, but it is here repeated twice, instead of a
single-short sequence. In spite of these diVerences, the fundamental
similarity of the epode with the strophe is undeniable, as has been
described in the opening section. For example, the palindrome of e1a
is a modiWcation of that of s1, which is exactly repeated in e3, while
the Wnal part of e1a is similar to the preceding verse, s8, and so on.

Textual problems
2 (s2). The paradosis
 is unmetrical;
 Moscho-
poulos.
14 (s6). The penultimate position is resolved only at v. 14: I
.
The metrical context of our verse strongly suggests that a
wwt
k
should not be analysed as a reizianum but as xd
. The resolution of
the second long of d is an oddity (see Part I, 6. C). There are only two
examples. In the other (O10e10 v. 110 v. 105 Sn. ) a
 Nemean Three 285
proper noun is involved and the phrase is not located at verse-end.43
Even if our case were a reizianum, the resolution would be totally
unparalleled. In tragedy or comedy, when the right of the choriambic
nucleus is resolved, it must be followed by w
(Part I, 5. F).44
PfeijVer 619 does not grasp the essence of the metrical problem.
What is irregular is not a non-responding resolution of the longum
of a double short element nor four short syllables in a row, but that
the resolved long is (i) the second long of the choriamb, whether it is
d or choriambic nucleus, and (ii) the penultimate position of a verse.
The sequence
wwt
k is diVerent from
wt
k. I accept the
latter (see Part I, Appendix, Addendum II), but it does not guarantee
the former.
I
 is highly suspect. Kayser, JL 108 introduces . Rau-
chenstein, Zu Pindars Nemeen, 250 searched in a diVerent direction.
His emendation is 
 or I
 (cf. Hom. Il. 16. 157, on the
legendary (
) I
 of the Myrmidones). I
 may
have been a gloss which displaced the original word; hence Stadt-
muller and Mair conjectured Y, which is adopted by Bowra. For a
defence of I
, see PfeijVer 61924.
18 ( 19 Sn.; e2). The paradosis J is unmetrical; Kg Heyne.
19 ( 20 Sn.; e3). To obtain the exact responsion (
wrw
)
throughout all the repetitions, Hermann (Notae) emends the text:
Vulgaris error
. Scribe
. Scholia
 explicant.
His emendation is adopted by Christ. At 39, too, he scans diVerently
and writes h
 I, instead of IE; cf.

 (22) at the
end of s2. He may be right.

43
Even this example may be eliminated by establishing a diVerent colometry; see
ad loc. Then the irregularity of our case is enhanced.
44
Outside the eighteen majors, there are in total three examples of resolution of
the penultimate position of aeolic phrases in Pae. 6. Two of them are found in
pherecrateans ( . . .
wwt
k): s6 (ph): 131 I
 ( 91 `
),
e10 (ph): 176 I
( 54 E, 115 r ). It is noteworthy that the
same word, I
, is used at both. Perhaps Pindar may have been conscious of the
extraordinary nature of the resolution, but nevertheless had no other means to expel
it. Maas supposed diVerently (Freiheiten, 11, Belege 789): I
may have been a
substitute for another word of the form

. The third example is the resolution of
the penultimate position of aristophanean ( . . .
ww
w

k): s4: 149  (
88 I, 128 IA). See Appendix B.
286 The Eighteen Majors
23 ( 24 Sn.; s3): The paradosis $
 is unmetrical; $

Moschopoulos.
33 ( 34 Sn.; s5). The paradosis F is unmetrical;  Byz.
41 ( 43 Sn.; s1). The paradosis If is unmetrical; f
G. Morel.
43 ( 45 Sn.; e3), The paradosis r  is unmetrical; Y
E. Schmid, Y  Moschopoulos. Also I must be I
(Byz.).
44 ( 46 Sn.; s4). The paradosis K 
is unmetrical; 

Triclinius.
53 ( 55 Sn.; s5). The paradosis  is unmetrical;
 Byz.
56 ( 58 Sn.; s8). The paradosis is a syllable too short; <K>
E. Schmid.
69 ( 72 Sn.; s1). The paradosis is K  I
; f g Hermann.
Turyns comment, metro non cogente, is wrong; the position must
be short.
72 ( 75 Sn.; s4). <> Triclinius.

Individual verses
s1. See Part I, 8. A. 6 (b) (palindrome), 8. B. 4 (aeolic d); 8. C. 3. wd,
8. C. 8 (d e). The verse is a palindrome, except for the last two
positions. There are four verses which include this type of palin-
drome. For the decision of where to set phrase-end between
two double shorts, see on Part I, 3, Rule 6, which leads us to prefer
(a) to (b):
(a)

ww
w
w
ww

w
k telwd e
(b)

ww
w
w
ww

w
k
d hepta e
There are in total three verses which include the sequence x
ww


w
(xd e). Contrary to Pindars general tendency to maintain
bridge over the Wnal e (Part I, 8. C. 8), word-end falls before it at Wve
repetitions (or six, if 49  is included). The e is in general
occupied by one word (1 , 21 K
 , 29  , 41 K
 , 49 , 61 P).
s2. See Part I, 8. C. 6 (e 6 and e 5). The verse starts with a very long
sequence in single-short movement. The length of the sequence is
 Nemean Three 287
eleven positions (e 5). Of the other examples, O1s8 (xe 5) and P2s1
(e 6 ) are comparable. Like these two and unlike the majority, the
sequence at N3s2 does not run into double short at the end. And,
peculiarly, it is followed directly by e. At the junction of e 5 and e, two
longs are juxtaposed without brevis/anceps in between, but the
transition is smooth because the initial long of the e is resolved.
This sequence of e
re at the verse-end is found also at O1s9,
O1s10, O2s4. Word-end does not occur before the resolved e except
at v. 2 ( j 
). Rather, word break tends to be . . . j

rw
k (six repetitions out of eight).
It is conspicuous that two repetitions end with a tetrasyllabic word
of which the last syllable is not only short (brevis in longo) but open
(2 
, 22

). These two make hiatus. Short open end is
found at another repetition (42 
) too. Though the end is
closed, a short vowel is distinctively used at four other verse ends
(10 , 30  , 62 , 70 ). That is, seven out of
eight repetitions have a short syllable. Compare the similar ending of
N7s5.
There is another resolution at the fourth long of e 5. Here too
words of three or more consecutive short syllables are used (2 K
 , 10
, 22 
 b, 30
, 42 
, 50
). There is a strong sense break either before or after the
third short syllable:

w
w
: w : rw
w
rw
k
The pattern of bridge and cut is interesting; the Wrst three continuous
instances, i.e. 2, 10, 22, have cut regularly before that position while,
the rest, as if Pindar changed his mind after the third repetition, have
cut after the position throughout all the Wve repetitions. Moreover, in
three instances out of the Wve there is a similar grammatical con-
struction: a Wnite verb of the shape w
w Wlls the positions from the
second short to the third short: (30 
, 42 , 50
).
This may be regarded as an example of verbal assonance in a wider
sense (Part III, E).
s3. The Wnal e 2 is a common feature. In the eighteen majors there
are in total nine verses in which e 2, following an aeolic phrase with
2 ending ( . . .
ww
w
), comes at the verse-end (Part I, 8. B. 6).
Bridge between the aeolic phrase and e 2 is generally observed, and in
288 The Eighteen Majors
our verse too, but there is an exception: 63 K
 j
.
Resolution of e2 is not very common. The form
wrw
occurs
in six examples (Part I, 6. A). It is worth noting that a verse that ends
with this form is found only in N3 (here and e3). In the other verses it
stands at the beginning, preceded by link anceps.
s4. See Part I, 8. A. 6 (e) (palindrome). Except for the initial
position, this verse is again palindromic. This type of palindromic
sequence (e gl e) is without parallel. The metrical structure of
P2e5 x
w

x
ww
w

w
k
is identical with that of N3s4, except that the aeolic base is not
restricted to the form
w.
Anceps e ( iambic metron) often precedes an aeolic phrase.
Combined with glyconic, it is found in three other examples (P2e5,
P5s3, P6s1/2). Here the anceps is always short. Bridge is generally
observed between e and glyconic, and between glyconic and e. In our
verse there are exceptions, one each: 24
A j

, and 4
 j  .
s5. See Part I, 6. C. 3 (resolved long followed by short anceps); 8. C. 3
(enwd). The single-short rhythm of s2 reappears. But here, unlike
the previous case, the sequence starts with anceps and turns to
double short at the end in the usual manner. There is another
diVerence from s2 at the beginning; s5 starts with anceps. The
position where cut appears in most repetitions is also diVerent.
Here a strong sense break is found after the second short in six
instances. Compare:
s2
w
w
: w : rw
w
rw
k
s5

w
w :
w rw
ww
k
At s2, three or four consecutive short syllables are often allotted to
one word (see above). Here this is rather rare (13  , 45
 b, 53 ).
s6. See Part I, 6. D (bridge after long anceps). Like s5 this verse also
starts with xe. The initial anceps is always long. Then comes xdx, or
reizianum (for the ambiguity of this sequence see Part I, 7. 1). The
link anceps (or the half-base of reizianum) is irregularly long at v. 46
( v. 42 Sn.) I. For the resolved choriamb (d) at v. 14, see
under Textual problems above.
 Nemean Three 289
s7. This verse is one of the four examples in which an aeolic
decasyllable (wil 2) occurs. The phrase is used unambiguously at
P8s5
ww

x
w
ww
w
k d wil 2
(the others are P2e8 and P5s8). Long and short syllables appear
alternately, similarly with e3 or e5 of the preceding verses, at the Wrst
Wve positions. Bridge is regularly observed between wil 2 and e,
except at 55    j  (and, with elision, at 85  j
I
).
s8. Telesillean with two short initial syllables is fairly common
(eight examples in total, including tel 3). Although the length is
diVerent, these three are comparable:
I7s1 ww
ww
w
w

k
I7e4 ww
ww
w
w

k
P10s6 ww
ww
w
x
w
w
k
Here single-short rhythm is considerably prolonged. Counting the
penultimate short position of the telesillean together, there are four
shorts. Such a long sequence of single-short movement usually
precedes a double short, but rarely follows it. The only comparable
verse is
O9e1/2 w
w
w
w
ww
w
w
w

k we2wdwe2

The structural similarity with P10s6 (above) suggests that the tele-
sillean may be interpreted as the phrase ww
ww
w
x which func-
tions like x
ww
ww
x (xD x) and is extensively employed later by
Euripides (Itsumi, Enoplian in Tragedy). For another possible colo-
metry, see e4 below.
e1a. See Part I, 6. C. 3 (resolved long preceded and followed by
short anceps); 6. D (bridge after long anceps); 8. A. 6 (c) (palin-
drome); 8. C. 2 (double d); 8. C. 3 (wd); 8. C. 8 (dwe). Verse-end is not
marked by hiatus or brevis in longo, but the verse formations of e1a
and e1b, and the length of e1b, establish the division with certainty.
The verse e1a is analysable into d, e, and link anceps (dwdxe
). In
the normal D/e, D x D x e x is a typical combination, but in the
eighteen majors, such a verse with frequent use of link ancipitia is
highly exceptional, even if the diVerence between D and d is ignored.
Our verse is only partially paralleled. For the former, d xd (x),
290 The Eighteen Majors
parallels are collected at Part I, 8. C. 2. In these the combination xd x
is found in two cases (N2s5, P10s3), but the phrase comes at verse-
end. In our instance, xd x is followed by e x. Surprisingly, d xe x is
unparalleled. Even xe x is not common in freer D/e: xe x is usually
preceded by e. d xe (our phrase minus the last anceps) has a few
parallels, but the middle anceps is usually realized as short at all the
repetitions (
ww
w
w
), like that one in the next verse, N3e1b.
In our verse, the long anceps occurs only at v. 57 ( v. 59 Sn.)

E. If it were realized as short at all the repetitions, the phrase
would be
w
ww
w
w


and analysed as tel 3, which is a common phrase, used, for ex-
ample, in the preceding verse (s8).
e1b. See Part I, 8. A. 3 (iii) (longer verses); 8. C. 3 (enwd); 8. C. 8
(dwe); 8. C. 5 (double e). This verse is another example of a sequence
mostly in single-short movement. The verse starts in the same man-
ner as s5, and even the location of resolution is the same:
s5

w
w
wr w
ww
k
e3 w d
e1b w rw
w
wr w
ww
w
w

w
k we3 w d w e e
However, it is further prolonged after d. The movement frames the
central d. The prolongation of two w
has no precedent in the
strophe. Word-end is found regularly after the choriamb (17b
 , 37b I
 , 77b  ) except for 57b (but
even here word-end is found;  j).
Although the single-short movement is fundamental, it is not
necessarily felt as a galloping rhythm. The reason lies in that
words which would even suit dactyls or anapaests are used rather
frequently, thanks to split resolution; for example,
57b $
e ( 
 Ia 
jwwww
jw
wwjww
jww
j
77b  
j
wwjww
ww
j
e2. See 8, A. 7 (repetition within a verse). The pherecratean is
longer than the preceding reversed dodrans by the Wnal long. Both
phrases start with
w. It is as if reversed dodrans were repeated
 Nemean Three 291
twice, and prolonged by a long (or long anceps). For the interpret-
ation of this apparent repetition in general, see Part I, 5. A. 5, and 7.
5. Word-end is totally avoided at the junction of the two phrases.
There is no other example of this type of repetition, but the last two
phrases in the following verse are comparable:
O1e7 w

w

ww
w

ww
w

k ^e e dod ar
in which dodrans seems to be repeated with an added Wnal position
(this combination too is unique).
e3. See Part I, 8. A. 3 (ii) (d) (longer verse), 8. A. 6 (b) (palin-
drome), 8. B. 4 (aeolic d), 8. C. 3 (wd). Verses of this type are
characteristic of Class III. The verse should be compared especially
with the following:
N3e3
ww
w
w
ww

w
ww

wtw
k dodwd rdod e2
N7s2
ww
wt w
w

ww
w

w
w
k dodwe dod e2
P10s4 w

ww
x
ww

w
ww

w
w
k ^ e dxd rdod e2
Note the juxtaposition of dod/rdod and e2. The second long of the
Wnal e2 is resolved only in the last two repetitions (59  O
,
79  ). In the Wrst two repetitions the position is not
resolved, but it is interesting that both involve the combination of
light plosive and liquid making position: 19 P
 39 h

IE, as though, at Wrst reading, they might be expected to scan

www
, like
wrw
in the other two repetitions (Attic correp-
tion is less common in Pindar, but not rare). Perhaps Pindar may not
have intended to admit the responsion between
wrw
and

w
w
and have tried to regularize all the repetitions in either

www
or
wrw
, but changed his mind at some later stage.
Instead, it may be better to suspect the text, because the responsion
between resolved form and unresolved is unparalleled (the sole
exception is O13s4, v. 78 ( v. 81 Sn.) I; but this is not beyond
suspicion). Hermann may be right to restore exact responsion by
changing only
 into
; see Textual problems above.
Remarkably, bridge between these two phrases is completely absent
at all the repetitions (including 59 c j  O
). Bridge
before the Wnal e2 is occasionally absent, especially when the phrase
preceding e2 is an aeolic phrase with zero-ending, but this case is
extreme; see Part I, 8. B. 6.
292 The Eighteen Majors
e4. See Part I, 8. A. 6 (c) (palindrome); 8. A. 7 (repetition within a
verse). In Part I, the following colometry is adopted for the statistics:
ww
ww
w
jww
ww
w

w
k tel tel e
This colometry helpfully speciWes the repetition of the telesillean.
On the other hand, it is one of the two examples in which the rules of
colometry (Part I, 3) lose consistency and require exceptions; accord-
ing to them, phrase boundary should be set before the short (which is
regarded as anceps), not between a long and double short. In all the
other verses the consistent division is:
ww
ww
jw
ww
ww
w

w
k
The other exception is P2s4, which is structurally similar to N3e4:
P2s4 ww
ww
ww
jw
ww
ww
w
k
This can, or should, be analysed as ^ D w ibycean if the argument
that expanded aeolics are not aeolics but, in fact, a kind of freer D/e is
accepted. And, if P2s4 is ^ D wibyc, it is reasonable to analyse N3e4
as ^ Dwibyc, too (and e, of course). Then, should N3s8 not be telwe2

but ^ Dwe3
? This analysis would hide the similarities with the other
telesilleans cited above (on s8). I must admit that there is a dilemma,
and I have no clear-cut solution at the moment.45
Between the two telesilleans, bridge is observed at three repeti-
tions out of four. However, elision or postpositive is involved in all
cases: 20 =o
, 40 A  =IA, 80
  =
.

45
The fundamental problem is our ignorance of the time value of two shorts in
the place of half-base of telesillean: ww
ww
w
. Is the initial ww equivalent to
ww in the nucleus when this telesillean is used in unambiguously aeolic context? We
should remember too that, as far as I know, there is no example in Greek metre of
responsion between ww
ww
w
and

ww
w
.
 Nemean Four 293

NEMEAN FOUR

Twelve strophes (monostrophic). Class I

N4s1 a
a
ww
w

ww
k hepta 2 d
N4s2 u
a
ww
k hepta
N4s3

w
ww

u
w
ww
k hepta wil
N4s4
w
ww
w
w
ww

k gl reiz
N4s5
a
w
ww
a
ww
w
k wil tel
N4s6
a
a
ww
w
ww

k wil reiz
N4s7 www
ww
w
k gl
N4s8

ww
w
w

k tel 3
s1 B 33; s2 H 10, 34, 90; B 82; s3 H 3, 27, 43, 91; B 35, 59, 75; s4 H 28, 92; B 36; s5 H 69;
B 21, 45; s6 H 6, 46, 86; B 70; s7 B 23, 47

s1 1 (1) @ P
 

s2 2 (2) N b 
s3 3 (3) A  Id  
.
s4 4 (4) Pb e o   a 
s5 5 (5) E,  P  
 .
s6 6 (6) B  K
  ,
s7 7 (7)   f % 
s8 8 (8) H e K  .
Nemean 4, the longest of the monostrophic poems, is constructed
very simply. All the verses can be analysed without ambiguities. The
metre is completely aeolic. Repeated use of wilamowitzianum (3
times) and its acephalous form, heptasyllable (twice), is a peculiarity
of this ode.
The following features of this ode show that it is purely aeolic,
Class I.
1. At some anceps positions long and short syllables occur quite
freely. Long syllables sporadically appear even at the second position
of reversed dodrans, against Pindars usual practice.
2. Resolution is very rare (only in base at s7).
3. Acatalectic and catalectic verses appear in sequence to make an
evident contrast at s5 and s6.
4. Dovetailing occurs at 11 repetitions out of 12 in s3.
5. e is never used as a suYx.
294 The Eighteen Majors
6. Verse-by-verse structure is evident: identical or similar cola are
repeatedly employed.
To illustrate the last point: s13 all begin with hepta, or its related
colon hepta 2. In s3, heptasyllable is followed by wilamowitzianum
(
hepta). s4 begins with glyconic, the reverse form of wilamowitzia-
num. From s4 to s5, and from s5 to s6, these modiWcations are clear:
s4 ! s5: gl ! wil (inversion); reiz ! tel (acatalectic)
s5 ! s6: wil ! wil (repetition); tel ! reiz (catalectic)
The next (s7) is glyconic again (inversion), which is followed by
tel 3; s7 and s8 are divided by verse-end, manifested by brevis in
longo at two verses (but see below, s7). Compare the construction of
O9s39: gl/wil reiz is repeated four times; then broken into two
verses: gl k (s8) and reiz (s9).
It is important to notice that colometry and word-patterns are
related, of course, but at the same time are independent of each other.
In this poem word-end often coincides at particular positions in
most, but not all, of the twelve repetitions. This makes a dominant
word-pattern; eventually a particular, conspicuous rhythm emerges
superimposed on the colometry. For example, the Wrst seven posi-
tions of s1 are repeated in the following verse (s2). The repeated
rhythm is highlighted by the presence of word-end after the seventh
position in nine out of 12 repetitions of s1 (including 41 Kd  j):
x
x
ww
jw

ww
k
The very Wrst verse of the ode is typical (@ P j

) in that a syntactical break is found there. Similarly, in s3
(hepta wil), word-end almost always falls after the Wrst position of
the second phrase (wil), e.g. A  j Id  

:


w
ww

j

w
ww
k
There is only one exception (v. 51). Moreover, the word-pattern is
reversed in the following verse with a minor change (Pb e
o j   a ) in three repetitions in the following
verse (s4):

w
ww
jw
w
ww

k
 Nemean Four 295
In eight other repetitions another neat construction can be dis-
cerned: the verse is given a twofold structure too, but slightly diVer-
ent from the above. The structure is clearest at v. 44 (str. 6):
K, E,
w
ww
w
d  P, ,
w
ww


The pivotal nature of the seventh position can be felt at almost all the
other repetitions:

w
ww
: w :
w
ww

k
Only one (v. 68) has word-end neither after nor before it. The ninth
position in the following verse (s5) is again of a pivotal nature; six
verses have word-end after it, and three before:

w
w
ww
: x :
ww
w
k

Textual problems
3 (s3). The paradosis  is unmetrical;  E. Schmid.
9 (s1). The paradosis d is unmetrical; d Boeckh.
12 (s3). The paradosis h
 is unmetrical; M0
 E. Schmid.
20 (s4). The paradosis o is unmetrical; o Triclinius.
23 (s7). The paradosis  is unmetrical;  Triclinius.
25 (s1). The paradosis ( is unmetrical; ( G. Morel. So
is the paradosis e ; e E. Schmid. Hermanns -
e is implausible, since it involves resolved d.
37 (s5). The paradosis  is unmetrical; so  (Triclinius).
54 (s6). E. Schmid regularizes the initial anceps of reizianum by
changing  into . The anceps is short at all the other
repetitions.  gives the right sense (a handmaiden; 
scholia), Iolkos being personiWed. Here in this verse, regularization is
preferable to All-But-One.
55 (s7). The paradosis d is unmetrical; d Triclinius.
62 (s6). The paradosis A is unmetrical; most editors
adopts Hermanns emendation (Notae) 
.
634 (s78). The verse-division in the paradosis is
Z O
 I

 
 
 O
296 The Eighteen Majors
The older editors accept it, but  at the beginning of a verse is highly
improbable. Ahlwardts emendation of  into d is accepted by e.g.
Schroeder and Snell, although it would make the word-order less
grammaticalor it may be for that reason that the manuscripts
introduce . Carey (p.147) cites k  in our case as well as k  at I8.
10, as parallels to k  a at N7 v. 25, and acknowledges an occasional
licence that a postpositive can stand at verse-beginning. Certainly 
at I8. 10 is diYcult to emend (see ad loc.), but this is not strong
enough reason to support the others. And  at N7. 25 need not be
recognized as a postpositive. k  in our case is better emended.
Turyn adopts Ahlwardts d, and further transposes I
 and

 . Turyn registers Wilamowitzs name in his apparatus for this
change, but it had been already proposed by Kayser (JL). Wilamowitz,
Pindaros, 175 n. 3 suggests it as a possibility but rejects it.
68 (s4). The paradosis K 
is unmetrical; Kb Ritterhausen.
90 (s2). The paradosis e I,
E is metrically wrong.
Boeckhs emendation e 
,
E is a solution and is
adopted by most scholars including Snell (his apparatus is wrong).
Boeckh basically follows Hermann, De metris: e 
,

E. This is metrically inferior, and the usage of is not in Pindars
manner. The change of the tense of the verb from future to aorist was
Wrst proposed by Pauw. The assumed situation is: Euphanes, the
grandfather of Timasarchos and, most probably, the father of Kalli-
kles, whom he celebrated, is now dead. Some scholars, including
Bury and Farnell, maintain that the change of the tense is unlikely to
have occurred in transmission. Among these, Mommsen is the least
radical, and is followed by Turyn and Race (Loeb): I,
E,
 . This requires correption (
E, ) in non-dactylic movement.
Moreover, belongs to a monosyllabic word with circumXex accent.
The correption is highly suspect; I Wnd no examples in Pindar (for
O4 v 7 6 Sn. 
E, n , see ad loc. in Appendix A).
91 (s3). The paradosis    is wrong; Both Snell and Turyn
adopt I (Mingarelli). Hermanns   Y (Notae) is fol-
lowed by the older editions: " scholiastes iam legisse videtur, ita
scribens;  a, d, Y F $

 K,

Pe r .
95 (s7). The paradosis a is unmetrical; a E. Schmid.
 Nemean Four 297

Individual verses
s1. See Part I, 5. E. 1 (rdod starting with
x), 8. B. 4 (aeolic d). The
analysis of aeolic enneasyllable (hepta 2) followed by d is guaran-
teed by the similarity with P8e1 and P11s5, in which hepta 2 is
followed by e 2:
x
w
ww
w

w
w
k
The use of d with a longer type of aeolic colon resembles (in reverse
order)
P8s5
ww

u
w
ww
w
k d wil 2
Short syllables are dominant at both the two ancipitia of the
enneasyllable (hepta 2). Long syllables are used only twice at each
anceps. Proper nouns are involved (17 , 49 P) at
the second anceps (thus, proper nouns apart, rdod is not
x
ww

here either), but not (49 K  , 81 
) at the Wrst.
In the eighteen majors there are four verses in total that are
composed of an aeolic phrase ending with w
( 2 ending) and d
(Part I, 8. B. 4). In these examples word-end tends to be avoided at the
junction (bridge), but N4s1 is an exception: word-end occurs at four
repetitions out of 12. Interestingly,
ww
is occupied by one word in
all four (1 , 9 d  , 65 , 81 ).
s2. See Part I, 5. E. 1 (rdod starting with
x). At the second anceps
short syllable is dominant as usual in reversed dodrantes, but there
are three exceptional cases: 10 (
 (proper noun), 66
h, 90  (emendation; see above, Textual problems).
Long and short syllables seem to be used rather freely at the Wrst
anceps (long/short 9/3 or 8/4 (depending on 82 -).
s3. Heptasyllable is repeated once more, but this time followed by
wilamowitzianum. Except for v. 51 (Ij
), there is dovetailing in
11 out of 12 repetitions after the Wrst position of the wilamowitzia-
num. Bergk24 divides the verse into two by regularizing v. 51; nec
obstat v. 51 / 
 b k I
 
 , si  in unam
syllabam coalescere statuis.46 But the regularization is unnecessary;
this is a typical case of All-but-One.

46
There is another reason for Bergk to divide s3. He scans disyllabically the two
sequential vowels of  (35) and  (75) and suggests that r
w
ww

is possible for s3b, as when in Euripides resolution at the initial of the heptasyllable is
298 The Eighteen Majors
At the aeolic base of the wilamowitzianum, responsion between


(8 times) and
w (4 times) are freely interchanged. But the
other ancipitia tend to be of Wxed value: the half base of the heptasyl-
lable is always longreading 
 at 75 (E. Schmid); for
-= -, see Braswell, Pythian Four, 214(d)whereas the anceps
in the reversed dodrans incorporated in the heptasyllable and that in
the wilamowitzianum are always short.
s4. See Part I, 8. A. 6 (b) (palindrome). For the palindrome . . .

ww
w
w
ww
. . . and the rule of setting colon-end before its
second breve, see N3s1. Following the rule, this verse is analysed as
glyconic and reizianum. This analysis explains particularly well the
relationship between s4 and s56 (see above).
s5. See Part I, 5. C. 1 (proper noun involving aeolic full base of the
shape
x), 8. A. 6 (c) (palindrome). The aeolic base of the wilamo-
witzianum here is in general
w, but at three repetitions

is
used for proper nouns only (21 E, 53 , 69
; at v. 29 l
 should be read; see Braswell, Pythian
Four, 58(a). The anceps of the telesillean is short 10 times out of 12.
Long syllable appears at two repetitions (one of them with proper
noun): 5  13 ( , and perhaps at two further verses
(61

, 77

).
s6. See Part I, 5. C. 1 (proper noun involving aeolic full base of the
shape
x), 5. E. 1 (rdod starting with
x), 8. A. 6 (c) (palindrome).
As in s5, the aeolic base of the wilamowitzianum is generally
w;
only for proper nouns (22 `N , 46 N ) is

permitted.
And the second anceps is short, except in one proper noun (70
P
). The initial anceps of the reizianum is always short.

accepted. The accepted colometry requires  to be irregularly scanned as w
,

instead of (ww
), which is normal in all the other examples (the example at N6 v. 20
is diYcult to decide, see ad loc.). Certainly this would be a bonus. But, as Hermann
(Notae) points out,  is disyllabically scanned at I6. 55. As for , there is
no problem in supposing synizesis. Metrically Bergks colometry is not good. The
resolution of the half-base of hepta is not paralleled in Pindar. One may cite:
O1e5 ww
w
ww

w
w
k ^ dwd e2
But the metrical context suggests a diVerent analysis, Above all, the responsion
between double short and single short is extraordinary. Bergks proposed division
should be rejected.
 Nemean Four 299
s7. Glyconic with base www is common in the eighteen majors,
commoner than that with

. Verse-end between s7 and s8 is
weak; I
(47) and K
 (31) at the verse-end are unparalleled. If
s7 were united with s8, the united verse would be very similar to s4,
s5, and s6. Attestation of verse-end depends on brevis in longo at 23
and 47, but these cases could only be eliminated if it were possible to
assume prosodical licence there: syllable-lengthening before  at 23/4

 j !  and before at 47/8 I

j `Y . However,
examples of prosodical lengthening are very few, fewer than some
might assume, at least in Pindar. The only probable case in the
eighteen majors is P11. 38. P5, v. 42 is most certainly not an example,
and should be emended. See ad locc.47
s8. The aeolic decasyllable (tel 3) are used at N2s2, I7s1, I7e4 too.
All of them stand as independent verses by themselves.

47
Bergk2 eliminates verse-end by another means; by intoducing two conjectures:
E (with K in the preceding verse changed into q) and  .
300 The Eighteen Majors

NEMEAN SIX

Three triads. Class II

(1)
N6s1a w

j ^ e

N6s1b
w
ww
w
rw
k gl e
N6s2
w
ww
w

w
w
ww
k gl wil
(2)
N6s3 www
ww
www
ww
ww
ww
j rdod ribyc ( rdod2d)
N6s4a ww
ww
ww
rw

k ^ D e

N6s4b
w
k e
(3)
N6s5 ww

ww
ww
j ^d D
N6s6a
ww
ww


ww
rw

k D
d e

N6s6b
wt
k e

(coda)
N6s7
wy
ww
w
w
k e dwe
N6e1
ww

ww
ww
rw
j dDe
N6e2
wrww
w
j gl
N6e3

ww
ww
ww

ww
j
D d
N6e4
ww
ww
ww
w
j ibyc ( dod2d)
N6e5

ww

ww
ww
j
d D
N6e6/7 y
w
w
w

j ye3

N6e8 y

ww
w
k ygl
N6e9

ww
ww
ww
k
D

s1b H 54b; s2 H 25, 32, 48, 55; s4a B 11;5 s4b B 50; s6a H 29; B 6, 29, 59; s6b H 13b; B
6b; e8 B 22; e9 HB 23

For the analysis and the notation of s3 (ribyc ) and e4 (ibyc ), see
Part I, 7. 6 expanded aeolic, and also Appendix B.
s1a 1a (1) 1  IH,
s1b 1b  H  K A b

s2 2 (2) e I  b
A 
s3 3 (3)  , e b P, b
 Ib Nb 
s4a 4a (4)  P : I

 

s4b 4b j 
 Nemean Six 301
s5 5 (5)  X  I
 ,
s6a 6a (6) 
 K P N Pb a 
s6b 6b (6b) 

s7 7 (7)   E
d 
.
e1 15 (15) Y K 
 e
 
e2 16 (16)


  .
e3 17 (17) E a 
 Kg `N
e4 18 (18) 
H <  > I
` F,
e5 19 (19) d

"E 
 ,
e6/7 20/1 (20)  b E , 


e8 22 (21) # , n $

e9 23 (22) <
 $ .
str. 2. 2430 (23)(29) ant. 2. 317 (30)(36) ep. 2. 3846 (37)(44)
str. 3. 4753 (45)(51) ant. 3. 5460 (52)(58) ep. 3. 619 (59)(66)

Nemean 6 is a representative sample of a small-scale Pindaric

epinikion in language, use of motifs, structure, and its praise for
the victor and his family. But metrically, it includes most unusual
phrases both in strophe/antistrophe and epode. Even the verse-end is
not always Wxed; see s1a, s4a. Metrical problems are entangled with
textual problems. The transmitted text is undoubtedly corrupt here
and there. The Wnal part of the strophe (s67) is especially diYcult,
and inevitably, involves wider emendations. Using this part as a
model case, I try to trace emendations proposed by critics historically
in detail (Part I, Appendix). The Wnal part of the epode is also
diYcult. For this, see below Textual problems.
The dactylic phrase
ww
ww
(D) and its prolonged form

ww
ww
ww
(D ) are used extensively in N6. Both are often
preceded or followed by
ww
(d). The combination of D or D
and d is peculiar to this ode. Outside N6 only one example (P10s3) is
found in the eighteen majors (Part I, 8. C. 12), another in O4
outside them. Dactylic movement occurs in the expanded aeolic
too. Two out of the six expanded aeolics in the eighteen majors are
used in this ode (e4, s3). These dactylic movements may possibly
raise RSS: 58.0% (the strophe) and 55.9% (the epode). The usual
type of glyconic and some other aeolics are mixed, but most verses
are best analysed as freer D/e, and both stanza-forms are classiWed
unambiguously into Class II.
302 The Eighteen Majors

Strophe
The strophe is metrically divided into three blocks, all of which
consist of three verses each (s1a2, s34b, s56b), and a coda (s7).
After an impressive short verse (+  IH), simple aeolic phrases
are repeatedly employed in the Wrst block. In the Wrst verse of the next
block (s3), double-short movement appears in an expanded aeolic
phrase: ribyc ( rdod2d). It is a transitional point from aeolic to
freer D/e. The second and the third blocks resemble each other in the
overall structure. Their Wrst verses (s3 and s5) end with dactylic runs.
The next, s4a and s6a, start with dactylic movement again, and their
Wnal parts are exactly the same:
. . .
ww
rw

k
This peculiar sequence is paralleled only by O10e4/5 (cited below, ad
s4a). d e is common in the eighteen majors, but d e x is not (8, C,
8). Then comes a very short verse in both blocks: s4b and s6b. The
coda, s7, is certainly corrupt textually in the Wrst three positions, and
the metrical analysis remains uncertain. But its other part oVers no
problem. The Wnal sequence of w
w
connects with the beginning
of the following antistrophe. When an epode follows, there is a
stronger sense of pause ( . . . w
k
w . . . ). There is one long anceps
in mid-verse as a link (for its rarity, see Part I, 6. D). This shows
strong aYnity with the normal D/e. Four verses end with pendent
anceps (s1b, s4, s6a, s6b).

Epode
As in the strophe, dactylic phrases are much used in the epode. The
phrase d D is twice repeated (e1, e5), and its reversed form with
expansion, D d, appears too (e3). Phrases with three dactyls, D
(e3, e9) and ibyc ( dod2d) (e4), are also signiWcant. These are not
common in the eighteen majors. With them two glyconics (e2, e8), if
they are really so, and a single-short verse starting with biceps (an
oddity) are mixed (e6/7). Hiatus/brevis is absent till e6/7, but the
coincidence of word-ends accompanied by rather frequent sense
pauses, and lengths of verse support the divisions accepted since
Boeckh.
 Nemean Six 303

Textual problems
Besides s67 (Part I, Appendix), there are quite a few verses in which
textual and metrical problems are so closely intertwined that even
their metrical scheme is uncertain. The repetitions in e4 are discussed
together Wrst, and those in e68 too, before other individual prob-
lems are introduced.
e4 (vv. 18, 41 40 Sn., 64 62 Sn.):
ep. 1 
H I
F,
ep. 2   
ep. 3
E,   K d  (
E  D)
Apparently there is a lacuna in ep. 1. The other two repetitions are
(
metrically the same if synizesis of - is accepted at ep. 2. (i.e.

w
):

ww
ww
ww
w
j
Based on this scheme, Bergk2 supplies  after
H at ep. 1
(18). Both Snell and Turyn adopt it. Henry, following Barrett, sup-
poses the lacuna to come at the end of the verse after I
`F:
words more often drop out at the end of manuscript cola. Barrett
supplements . Correption (F -) is acceptable in quasi-dactylic
movement.
Before Bergk, scholars thought diVerently. A Wnite verb is not
necessarily required. Even metrically there are other possibilities.
Without assuming the synizesis of  the metre at ep. 2
becomes longer by one short position than Bergks proposal:

ww
ww
rw
w

This scansion had been accepted by critics before Bergk, and, after
him, by Mommsen. The supplement of Boeckh (ed. minor), K
(18) is based on this metre though the resolution does not occur at
the position next to D :

ww
ww

w
w
(   ).
Kaysers proposal K is also based on this scheme (Bergks
 is a development of Kaysers). Exact responsion including
resolution is introduced by Hermann (Nem. sextum), K, and
304 The Eighteen Majors
by Mommsen, K
d 

 (in the main text) and I

g (pro-
posed in apparatus). K requires correption (- I
) in a non-
dactylic context and is doubtful. To make ep. 3 correspond with ep. 2,
either a short syllable should be added or a short should be replaced
by a long. Boeckh writes:
E,   K d .
Hermann (Nem. sextum), after proposing Mb , prints
in the main text:
E,   K d b, . He is
basically followed by Mommsen:
E,   K d b, -
. The metrical form of Boeckh, Hermann, et al. must be
analysed as D e 2, which is totally unparalleled in the eighteen
majors. Even taking account of the scarcity of expanded aeolics
into consideration, Bergks scheme (and/or Barretts supplement) is
better; and it requires only one lacuna without any further emend-
ations.
e68 (vv. 2022, 435 424 Sn., 668 646 Sn.):
ep. 1  b  , 

 # , n $

ep. 2 
  
  
  
ep. 3 E  
   Y Y
 
The Wrst metrical question is whether e6 and e7 should be divided
or not. Unless Hermanns E for  (Nem. sextum) is
adopted at ep. 1, the syllable in question is brevis in longo, and
e6 must be separated from e7. On the other hand,
 at 44 is
enclitic, and supports combination into one verse. In Part I, e6
and e7 are combined and are treated as a single verse, with the
adoption of Hermanns E . Neither separation nor combination
creates metrically irregular verses, except for the double short at
the beginning.
The initial position of e6 is diYcult. 
 at ep. 2 starts with
double short, E at ep. 3 with long. Either scansion is possible
for the remaining repetition  (ep. 1): ww
or w
(for syniz-
esis of  cf. N4. 75). Unless 
 is emended and something like

 (Hermann) is introduced to make the position in question a
simple anceps, the responsion between double short (ep. 1, ep. 2) and
long (ep. 3) should be accepted as a rare example of link biceps. This
type of correspondence at link position occurs only at O10e3. Both
O10e3 and N6s6 have, or are supposed to have, the link anceps at the
beginning of the verse. Perhaps we had better accept the licence of the
 Nemean Six 305
responsion between long and double short only at the initial position
of the verse, as in the tragic trimeter.48
At e8, the paradosis gives worse responsion than at e6/7. Neither
Turyn nor Snell provides a good text from a metrical point of view.
E. Schmids  I (45 43 Sn.) for   (the
paradosis:  I is a correction of Triclinius, according to
the reports of Mommsen and Christ) was accepted by the old edi-
tions (e.g. Boeckh and Schneidewin). The supposed metre is
t

ww
w

Hermanns X  (Nem. sextum) is a further step to
preserve  . In consequence, he also introduces H in
place of 
 (the paradosis) to expel one syllable and to adjust
the length to other repetitions. Some, including Snell, suppose a
diVerent metre so as to keep 
 , but I leave this aside for the
moment. The corresponding lines in Hermanns text are:
ep. 1 # , n $



ww
w

ep. 2 H X 


ww
w

ep. 3 Y Y
  r

ww
w

Turyn follows Hermann. Mommsen rejects H , but his emend-
ation is Xat (
 ,  ). The verse could be analysed as
glyconic with base in the form r
at ep. 3. But it is highly
questionable whether there is a glyconic of the form r


ww
w
, and still more questionable, even if it is admitted,
whether r
can respond with

.49 Thus Bergk dubitanter
proposes N
 for Y Y
 to restore exact responsion

48
There seems to be another biceps in the middle of O10e3, but I strongly believe
that emendation is necessary there; see ad loc. There is no other certain example. For
I8s2, some editions accept the responsion but I reject it; see ad loc.
49
If, however, the phrase t
were accepted as a possible resolution not of
glyconic but of a freer D/e phrase, two other irregularities would be eliminated in
O10: namely link biceps in e3 and resolution of d in e10; see Alternative Colometry
for e3 and e10 at the end of the section for Olympian 10. It might be explained
that the phrase t
is a variation of e. The responsion between rw
and

w
is common, and

(spondee) is easily analysed as e with suppression
of the central w. Then t
might be tw
minus the central w. I am not sure of
Henrys analysis: (tia^ ) tia. The initial anceps of iambic may not be resolved
outside spoken parts of drama.
306 The Eighteen Majors
throughout all the repetition. Bergk is followed by Christ. Maas is
more radical. He changes the potential optative  . . . Y
 (ep. 3)
into indicative d . . . r   (Freiheiten 19). Bowra accepts
Postgates N
, which must be rejected even metrically unless
the Wrst syllable () is long: w
is hardly ever in responsion with


in Pindar.
Wilamowitz (Herakles, on v. 681) supposes a diVerent metrical
scheme instead of glyconic. Based on 
 X  (cer-
tainly good Greek, see above), the scheme is



ww
w
. It
is one position longer than Hermanns. To add another mora to the
other repetitions, # (ep. 1) and Y Y
 (ep. 3) are
changed into # and r  < > Y
 (with  in the
preceding verse changed into d) respectively. Snell follows Wila-
mowitz unreservedly while Schroeder does so with minor changes
(Y in place of r). The problem of this sort of solution lies in the
metrical form itself. The phrase x
x
ww
w
, an aeolic enneasyl-
lable (hepta 2), occurs three times (P8e1, P11s5, N4s1) but its third
position is in general realized as short. The three continuous long
syllables preceding dodrans are unparalleled.50
Errors and emendations in the other verses:
23 ( 22 Sn.; e9). The paradosis H K is unmetrical; $
W. Schultze,  Triclinius. The older editions write  and
scan

.
26 ( 25 Sn.; s3).
 is supplemented from the scholia by
E. Schmid.
28 ( 27 Sn.; s5). The unmetrical word-order in the paradosis,
 
F, is corrected by Mingarelli: 
F  (epic correption
in dactylic movement). This word-order is now supported in a
papyrus ( 41 SnellMaehler).
31 ( 30 Sn.; s1).  is, pace Gerber, unmetrical;  Hermann.
So is K; K Triclinius.
36 ( 35 Sn.; s6). 
  is one of Triclinius best emend-
ations. The paradosis is d (a rare verb). Hohl is certainly
wrong to defend it. See Part I, Appendix.
52 ( 50 Sn.; s6). The unmetrical and ungrammatical  is omit-
ted by Triclinius.

50
Schroeders scheme a


ww
w
is more improbable.
 Nemean Six 307
55 ( 53 Sn.; s4a). F (Pauw) certainly gives better sense than
the manuscript reading . Metrically too it is better, although it
is not impossible to accept, like Turyn, the responsion of

with

w and to keep .
57 ( 55 Sn.; s4a). The paradosis Id is unmetrical; so Nd.
60 ( 58 Sn.; e2). Snells apparatus is at its least informative.
Before Turyn proposed the text that Snell accepts, there was a long
and complicated history. The paradosis is   K


 A  A. The nominative   is metrically wrong.
The required metre is
wrw. First,  (genitive) was
introduced (E. Schmid, Hermann; this was further elaborated by
Boeckh, who made a 
the subject). Then Bergk suggested
 (Aeolic nominative), which had been already printed by
Ahlwaldt (so reported by Mommsen, and by Gerber, Emendations).
At the same time Bergk writes in the apparatus: quae ( nominativi
forma aeolica) si cui displicet, poterit is ,  (vel  )
K

 . The Aeolic nominative was accepted by Mommsen, but
was suspected by Christ (nescio tamen an ad metricam licentiam
nominis proprii confugere praestet). On the other hand the form
was championed as a vocative by Maas (Freiheiten, 15), who intro-
duced  ( ) instead of  , and K
 (inWnitive), which
was Wrst proposed by Schroeder (below). Bowra prints , f
 K

 (which has been already proposed by Maas: ist nicht
unmoglich, aber schwacher). Finally, Turyn establishes , 
K
. Meanwhile Hartung proposed: . . . f I E
j  K

 . . . . Metrically, this introduces an unusual
responsion:
ww
w
w
in place of
wwwww
w
. The required
metre is not dwe (or ch ia) but glyconic, because resolution is
rare in d (see below, e2). The same is true of Schroeders reading
(ed. maior):  K
.

Individual verses
s1. See Part I, 8. A. 3 (shorter verse). Word-end falls after the initial
three positions without fail at all the repetitions, but lacks hiatus or
brevis in longo there. Whether these three make an independent verse is
diYcult to decide. Most editors, including Turyn and Snell, do not
acknowledge independence; Schroeder and Bowra are exceptions. In
308 The Eighteen Majors
Pindar there is no exact parallel bacchiac at the initial position of a
verse (whether these three syllables should appropriately be called
bacchiac is a diVerent matter; I use the name for the sake of conveni-
ence at the moment). Both O10s4 and P5s11 start with w

fol-
lowed by a true long (in O10s4, the true long is resolved). Neither has
word-end consistently after the bacchiac. At the same time, the
coincidence of word-end at all the repetitions does not necessarily
mean the verse-end, of course. For example, at P10s4
w

jww
u
ww

w
ww

w
w
k ^e d x d rdod e 2
word-end occurs after the initial w

at all eight repetitions, but
phrase-end (let alone verse-end) here is improbable.
However, the possibility that the bacchiac in our case is a verse by
itself is strong. Its shortness does not matter; there is a certain example
of a verse composed of four positions in Pindar (O7s3 in D/e). And it
is worth noticing that one or two other short verses are found in our
strophe: s6b most certainly, and (very probably) s4b too. Generally,
every verse tends to be short in Class II odes. Perhaps the strongest case
for verse-end is the impressive, paratactic structure of the very Wrst
verse of the ode 1  IH, j  H  : K
  j (24) does not
weaken the case; cf. N4. 79 N   j and even N4. 31 K
 j.
Unless s1 and s1b are separated, the combination of bacchiac
glyconic is highly exceptional. Glyconic is undoubtedly preceded on
occasion by other phrases within a verse. But there is no example for
the preceding phrase ending pendent; all the examples end with a
true long. For the last position of our phrase there are two possible
interpretations; and whichever of the two is accepted, it is question-
able whether the three syllables are attached to those following.
(i) Suppose the three syllables are a true bacchiac. Then the last
syllable is most probably three brevia long, longer than true long as
a result of syncopation. There are cases in which bacchiac is followed
by cretic or iamb in lyric iambic, but hardly ever followed by non-
iambic phrases. In general, bacchiac comes at verse-end and its
catalectic eVect is usually accepted. (ii) Alternatively, suppose that
three syllables are acephalous e plus link anceps. If so, the link
preceding glyconic within a verse will be unique (see above). To
sum up: an independent verse of w

j seems more plausible,
and the phrase is ^ e
, not bacchiac.
 Nemean Six 309
s1b.e following glyconic is very common (12 examples). Reso-
lution at the same position has a parallel:
P5s3 w
wr
w
ww
w
rw
k we gl e
Our verse illustrates Pindars tendency to avoid word-end at the junc-
tion; it has no repetition with cut at the position. See Part I, 8. B. 5.
s2. This is one of the three examples of a glyconic followed by a
wilamowitzianum. Word-end never occurs at the junction in any
repetition of these three verses. Cf. Part I, 8. B. 1.
s3. See Part I, 8. A. 6 (g) (palindrome); 8. A. 7 (repetition within
a verse). Although there is no certain indication (hiatus/brevis),
verse-end is almost certain because, if this verse is combined with
the next, an unusually long dactylic run results (but see below), and
because there is a sense-break here three times (vv. 10, 49, 56 10,
47, 54 Sn.); note especially the anaphoric sentences at 1011  b
t . . .   s ). The diYculty lies in colometry. Two analyses are
possible, but both are unparalleled:
(i) www
ww
wr
ww
ww
ww
j gl D
(ii) www
ww
www
ww
ww
ww
j rdod ribyc ( rdod2d)
cf. P5e3, which is ambiguous like our verse:
(i)
w
ww
wr
ww
j gl d
(ii)
w
ww
www
ww
j rdod rdod
I reject (i) and adopt (ii), against Snell (and West, GM 65). If (i) were
correct, the eighth position of glyconic would be resolved; if (ii), the
initial position of what I call reversed (and of course expanded)
ibycean (
w
ww
ww
ww
) is resolved, like that of the second
reversed dodrans in P5e3 (see further ad loc.).
As to (i): although a glyconic with two resolutions at either end has
parallels in P6s3 (and its eighth positions, resolved and unresolved, are in
responsion) and N7e4 (Part I, 5. F), and the combination gl d also has
a parallel in I8s6 (8. B. 4), D following gl is unique. The phrase D
makes a verse by itself or combined with e or d. (It is the same with its
shorter, much more usual form, D.) If analysis (i) were accepted, our
verse would be the only example in which an aeolic phrase and a
dactylic phrase (D or D ) were united in the same verse; this is the
strongest argument against it. Repeated word-end is not helpful:
310 The Eighteen Majors
there are three coincident cuts in Wve out of six repetitions, but
neither of them shows phrase-division, positively or negatively:
www
wwj
wjww
ww
ww
jww
j
And bridge is perfectly observed both before and after the second
tribrach. That is, whichever of (i) and (ii) is adopted, two phrases are
closely knitted together.
In (ii), for the so-called dactylic expansion of the choriambic
nucleus of an aeolic phrase see Part I, 7. 6. Examples are surprisingly
rare. There is no certain example of rdodd (
w
ww
ww
) or of
rdod2d (
w
ww
ww
ww
) in the eighteen majors. However, in
N6 there is a plausible case of dod2d (e4), though it is not paralleled.
As I have argued in Part I, what we suppose to be expanded aeolic
may be a modiWcation of D or D . Like the other examples in the
eighteen majors, a true dactylic phrase (s4) follows our verses too.
This is a good reason for preferring (ii). For resolution of reversed
dodrans, see Part I, 5. E. 2.
In (ii) we see an expansive diptych structure, i.e. the repetition of
similar phrases of which either the former or the latter is longer than
the other. The Wrst reversed dodrans is expanded by two dactyls,
with the second keeping resolution at the initial. Pindar occasionally,
but less often than is expected, uses expansive diptych structure, e.g.
N7e5 www
ww
w
www
ww
w
w

k gl gl 3
P2s4 ww
ww
ww
w
ww
ww
w
k (teld) tel
s4a. See Part I, 6. B (acephaly); 8. C. 1 (e and D within a verse); 8.
C. 8 (D e). The double-short movement of the preceding verse is
ampliWed. There is no hiatus or brevis at the end of the preceding
verse (s3), and s4a is headless, starting with ww
ww
. . . Theoret-
ically it may be conceivable to make one composite verse, which
includes a very long dactylic run. But such a length is improbable.
There is one example of a similar combination of two verses among
the normal D/e odes. In N8, the Wnal phrase of s3 is D, and s4a starts
with ^ D, but hiatus occurs between them at vv. 20 and 37: they are
manifestly independent verses. For the acephalous dactylic, there is
no other example of the length of N6s4a: three double shorts. Shorter
lengths (^ D and ^ D
) are found at four passages: O10s1, O10e8
(^ D); O9e3, O13s1(^ D
). Of these,
 Nemean Six 311
O10s1 ww
ww

w
rw
k ^D ee
is similar to our verse in that acephalous dactylic is followed by two es.
s4b. See Part I, 8. A. 3 (shorter verse). Most editors do not separate
s4b from s4a. The sole dissident is Hermann (Nem. sextum), who,
however, combines s4b with s5. When s4a and s4b are combined into
one verse, the Wnal position of s4a is considered as link anceps, . . .
rw
xj
w
k, realized as long at all the repetitions except for v. 11
(: j q ). Word-end occurs at the end of s4a in every
repetition. Full sentence-end is found in two repetitions
(11 ., 27

 .), and a syntactical division occurs here at
v. 4 (
 
 j j ). Metrically, the combined verse
contains an irregularity. Contrary to the ordinary D/e, the phrase,
e x e whose anceps is exclusively long or long/short in responsion, is
unusual in freer D/e (whereas e 3,
w
w
w
, is fairly frequent).
The examples are limited to the verses which have no double short at
all but are totally composed of e, as in O2. Even in the verses where
medial long link anceps appears, word-end after it is generally
avoided (Part I, 6. D). In other words, bridge is observed as a rule
between long link anceps and the following long.
By giving independence to s4b, the overall similarity becomes
transparent between two blocks: s3s4as4b and s5s6as6b. See
the analysis in the opening section above. Except for the brevity of

w
k, there is no objection to its separation as an independent
verse. Such a short verse is not strange (Part I, 8. A. 3). In the normal
D/e odes, there is a verse of similar shape, just one position longer:
O7s3

w
k (3 ); (H9, B22, 66).
A similar but not identical phrase is used as an independent verse at
s6b of our ode too (but some scholars do not accept its very brevity
and have tried to expel it).51

51
Hermann adds after  (50b 48 Sn.) so as to expel the brevis in longo
and to connect s4b with s5. But apart from the meaningless , his colometry, though
eliminating  at the beginning of v. 35 ( v. 34 Sn.), is not good in itself:

w
ww

ww
ww
j
It is a combination of an aeolic phrase D within a verse; which is unparalleled (see
above on s4).
312 The Eighteen Majors
s5. See Part I, 6. B (acephaly). Acephalous d has a parallel in freer
D/e. There is no other case of ^ d D; but ^ e D (O10e2, I8s9) is
structurally comparable. Among the normal D/e odes, the same
phrase (^ d D) is found at O7e6:
ww

ww
ww


w
k
s6a. See Part I, 8. B. 4 (d ex); 8. C. 1 (e and D within a verse).
Unless we follow Boeckh and transpose F (see Part I, Appen-
dix), verse-end is Wrmly established by hiatus (breve at 29 and 59 is
not helpful, because the last position is anceps). This verse ends with
an anceps (pendent ending); there is also a long anceps between D
and d. Our verse is akin to the normal D/e Cf. N11e4:

ww
ww


ww

w
j D
d e
In contrast to D/e odes, the combination Dxd (let alone DxD, one of
the most popular phrase in D/e) is rare in non-D/e odes. Among the
eighteen majors, the nearest parallel is
O10e4/5

ww

ww
rw

j
d d e

But in this verse link anceps is not placed after the Wrst d.
There is cut in our verse between D and the intermediate link
anceps at all the repetitions, except for v. 52 ( 50 Sn.) (P
I
j ). Without this, our verse would be divided into two.
Verbal assonance: 13 
  r   59 
 
 .
s6b. See Part I, 6. C. 3 (er anceps at verse-end); 8. A. 3 (shorter
verse). Most scholars recognize verse-end at the end of s6b, whether
s6b is separated from s6a or not. The exceptions are: Kayser (JL),
Hermann (Nem. sextum), Bergk2, Bergk34 (diVerent metre from
2nd edn.), and Maas (Nachlese); see Part I, Appendix. The brevity
of this verse is not a problem; see s4b above. Although the diVerences
are many in the surrounding verses, the texts of Snell and Turyn are
the same in s6b. However, there remains an oddity, which is not fully
considered in the Appendix. The third position of e is resolved at
three repetitions: v. 29b ( 28b Sn.) s K
 (K
 being
scanned trisyllabically), v. 52b ( 50b Sn.) E  (emend-
ation metri causa), v. 59b ( 57b Sn.)  . There are
 Nemean Six 313
also three unresolved repetitions: v. 6b 
 , v. 13b F fg

 (emendation), v. 36b ( 35b Sn.) x 

 . It is
Pindars normal practice, when resolved and unresolved are in
responsion, for one form to be dominant. FiftyWfty is unusual.
Kayser (as early as 1844) and Maas reject resolution at the penulti-
mate position of the verse. Although resolution is certainly avoided,
there are some cases. For discussion, see Addendum II to Appendix.
s7. See Part I, 8. C. 7 (dwe), 8. C. 9 (e d). Snells text cannot be
accepted. It would be the best solution, by emending 

(37 36 Sn.), to introduce a short syllable into the second position.
Hermanns 

 is, so far, virtually the sole proposal along
this line (Part I, Appendix). If 
 were still preferred along
with emendations in other repetitions, the verse would be analysed as


ww
ww
w
w

Dwe
e1. See Part I, 8. C. 1 (e and D within a verse); 8. C. 2 (d d), 8. C.
8 (D e). There is no exact parallel for this verse, but its construction
is easily analysed from analogous verses. Supposing that D is an
expanded form of d, the verse can be taken as a variation of d d e
without any intervening link anceps. There are seven verses in which
d is situated in the initial position. For d d there are two (but both
are preceded by anceps).
e2. See Part I, 5. F (resolution of aeolic phrase). Three analyses are
theoretically conceivable for this verse, but (ii) and (iii) are without
parallels and must be rejected.
(i)
wrww
w
j glyconic with the third position resolved
(ii)
wwr w
w
j choriamb (d) with the second long
resolved short anceps e
(iii)
rwr
w
j long anceps totally resolved cretic
(e) another cretic
(i) Two other examples of glyconic contain Wve successive shorts.
From their metrical context they are certainly glyconic, but the
resolved position is not the third but the sixth.
P11s2b
w
wwrw

w
k gl e
P8s2
w
wwrw
k gl
314 The Eighteen Majors
(ii) There are two cases in which one or other of the two longs of a
choriamb is resolved: N3s6 (14 I
), O10e10 (110 ).
Responsion guarantees the resolution here. But the resolution is
irregular; the position is not resolved at most repetitions. For another
reason, I suspect N3s6 (14 I
), and there is a possiblity of a
diVerent colometry for O10e10; see ad locc.
(iii) Anceps two es is a common phrase in Pindar. As for the
resolution,
N7s6

ww
w
rwrw
w
k tel e 3
is a remote parallel, but a totally resolved cretic is never found in the
eighteen majors (including O2).
e3. D , the doubly expanded d, is used in
D (four verses) and

D
(one verse). The nearest one to our verse is
P2s3

ww
ww
ww

w
k
D e
There is no example of D d or D d in the eighteen majors. In
D/e odes, there is one case of D d ( . . .
ww
ww

ww
. . . ):
I1s6
w


ww
ww

ww

w


w
k e
D d
e
e
e5. The order of the longer phrase (D, D ) and the shorter phrase
(d) of e3 and s6 is reversed, in a similar manner to e1 or s5. The
combination d D is found once in the normal D/e odes:
N5e4

w


ww

ww
ww
u
w
j
e
d Dxe
e9. This verse, anceps D , has an exact parallel at O9e6.
 Nemean Seven 315

NEMEAN SEVEN

Five triads. Class III

N7s1 w

ww
w

w
w
k gl e 2 (? ^ e dod e 2)
N7s2
ww
wr w
w

ww
w

w
w
k dodwe dod e 2
N7s3
ww
w

w
w
w
k dod e 3
N7s4 w
w

wrw
ww
w

k we wil 3 (? we
ewdod
)
N7s5 w
ww

w
wrw
k wd e 3
N7s6

ww
w
rwrw
w
k tel e 3
N7s7
tww
w
rw
w
k tel e 2
N7s8

ww
w

w
ww
w

k tel hipp
N7e1 w
w

ww
rw
k we d e
N7e2 www
ww

jw
k rdod e
N7e3
w
ww
wr
w
k gl e (? rdod e 2)
N7e4 www
ww
wr
w
k gl e (? rdod e 2)
N7e5 www
ww
w
www
ww
w
w

k gl gl 3

s1 H 43, 85; s2 H 23, 31, 44, 86; B 2; s3 H 74, B 53; s4 B 54; s5 H 5, 13, 68, 89; B 5, 55,
76, 97; s6 H 14, 48, 56; s7 H 49; B 15, 57; s8 B37, 92; e1 B 17; e2 H 81; B 18, 102; e3 H
103; B 19, 61; e4 H 83; e5 H 84; B 21, 63

s1 1 (1) ,

 A ,
s2 2 (2)
E  , , + , 
  
s3 3 (3) P
 , P   P
s4 4 (4) a Ia K
 I + .
s5 5 (5) I
  P
 K
d Y
s6 6 (6) Y b
   : f b 
s7 7 (7) d
E  I A 
s8 8 (8) h I  a
 .
e1 17 (17) d b  E 
e2 18 (18) , P $
e  

e3 19 (19) Ie
  


e4 20 (20)  : Kg b
 

e5 21 (21)   j

 a e 
B  ! 
Nemean 7 is fairly typical of Class III. In its many verses, both in
the strophe and in the epode, aeolic phrases and freer D/e phrases are
316 The Eighteen Majors
closely combined. RSS is 53.3% in the strophe, and, in the epode,
61.8%, the highest of all the Pindaric odes. Sequences of three or
more short syllables are very frequent (s2, s4, s5, s6, s7, e1, e2, e3, e4,
e5). Note four continuous shorts in s2, s4, s5 and six shorts in s6 (cf.
Part III, D). Some of these make colometry ambiguous, for it is not
always easy to identify the resolved position. The sequence
www
w
is typically confusing. For example, see the phrase at the
end of
N7e3
w
ww
www
w
k
N7e4 www
ww
www
w
k
It is a resolved form of either (i) w

w
or (ii)
w
w
. If (i) is
accepted as in the analysis above, the verses are glyconic e. The
combination is without doubt a favourite with Pindar, and glyconic
with its eighth position resolved can be paralleled. Thus (i) is con-
sistent with other examples. On the other hand, with (ii), the alter-
native analysis in parenthesis, the verses are reversed dodrans e 2.
Reversed dodrans,
w
ww
, and e 2,
w
w
, are structually
similar to each other: the former is equivalent to the latter, with

ww
taking the place of
w
. Pindar occasionally exploits such
a similarity in this poem (e.g. s2) and elsewhere; see the analysis of
the overall structure of the strophe below. The phrase is used in a
similarly ambiguous manner at N7s7.
In the freer D/e parts of this ode, metre is almost restricted to
single-short movement. It is signiWcant that e 3 (
w
w
w
),
which is a rather unfamiliar phrase (14 examples; Part I, 8. B. 7, 8.
C. 6), less frequent than e e, is repeatedly used (but much longer
ones, like e 5 or e 6, characteristic of Class III, are absent). It is a
plausible guess that Pindar employed this especially in this ode
(and O1 too, although its use is diVerent there). d is occasionally
employed, but double-short movement is completely missing.

Strophe
As is noted above, resolution sometimes obscures colometry, especially
at s2, s4, and s7. But a consistent overall structure for the strophe can
be clearly grasped. Let us suppose that repeated dodrans (
ww
w
)
and single-short movement (
w
w
,
w
w
w
) are the basic
 Nemean Seven 317
units of N7, and adjust resolved positions to these schemes. Obviously,
the similarity between
ww
w
(dod) and
w
w
(e2) is fully
exploited. In the following chart, resolution is totally eliminated:
s1 w

ww
w

w
w

s2
ww
w
w
w

ww
w

w
w

s3
ww
w

w
w
w

s4 w
w

w
w
ww
w


s5 w
ww

w
w
w

s6

ww
w

w
w
w

s7

ww
w

w
w

s8

ww
w


w
ww
w


Dodrans and e 2 juxtaposed without intervening link is the base for
all the verses. It occurs by itself at s7, except for a preceding anceps.
Dodrans is followed by e 2 and its prolonged form
w
w
w
, e 3,
but never vice versa. Only single e can stand before dodrans. At s1 a
short phrase (w
) precedes dod e 2. This latter combination is
almost exactly repeated twice at s2, though, the initial long of e 2
being suppressed (w
w
), the Wrst half of this verse is one position
shorter than the latter. The combination of dodrans and e3 is also
used by itself (s3) and in a longer form (s6). Some additional changes
occur: at s5, wd (or headless reversed dodrans, w
ww
) is used
instead of dodrans; at s8, dodrans is repeated in the form of hippo-
nactean, which starts like e 2 and incorporates dodrans (or one could
say that hipponactean appears instead of e 3); and at s4, e 2 is missing.
Note that s4 and s8 end in the same manner ( 3 ending); except for
these two verses, the others end blunt. Also note that the initial
anceps of s4 and s5 is short anceps, that of s6, s7, and s8 long.

Epode
The structure of the epode is simple and its analysis is easy:
e1 w
w

ww
rw
k
e2 www
ww

w
k
e3
w
ww
www
w
k
e4 www
ww
www
w
k
e5 www
ww
w

www
ww
w
w

k
318 The Eighteen Majors
Four verses out of Wve end with cretic (e). The frequent resolutions
are remarkable, and create ambiguities. Seven longs are resolved out
of 33. If e3 and e4 are divided into two similar halves (rdod e 2), as
is alternatively proposed in the opening section above, e4, for ex-
ample, will be arranged in the same manner as e2:
e4 www
ww
rw
w
k
The tribrach in e2 is also ambiguous. It is most naturally taken as a
part of reversed dodrans (examples collected and discussed in Part I,
5. E. 2), but it may be ^ e, an abridged form of we of the preceding
verse (see further, below). The Wnal verse, e5, is the only verse that
ends blunt. Its structure is very close to that of the Wnal verse of the
strophe (s8).

Textual problems
12 (s4).  d (MSS) is one syllable too short. Both K d
and K are possible as further improvements of E. Schmids
K d.
16 (s8). Deletion of  (Hermann, Notae) is indispensable.
1920 (e34). The paradosis gives 


k A  :


at verse-end can be paralleled and is not problematic (examples
cited from the eighteen majors): O1e6, v. 57 o
 k  (but
strictly, here o
 is not a preposition but adverb) and O9s8, v. 19 
, ,

k `F   (but strictly, here

refers back to   , as well as forwards). The paradosis

has caused suspicion on other grounds. Its main diYculties lie in the
curious expression 
 A itself, and in the lack of a word
which means similarly or the like (paraphrased in the scholia by
 ). The paradosis also has another  after Ie which is
metrically unacceptable. Furthermore I suspect that
a A
 does not mean to die but to visit a tomb (of a dead person).
The meaning of the text itself apart, the paradosis is suspect metric-
ally. With A  the base of the glyconic in v. 20 is
w while
the other four corresponding instances have www. This correspond-
ence is found only at P5e9 (see ad loc. and Part I, 5. C on aeolic base
in general). Pindar does not generally admit responsion between
diVerent forms of base, except between
w and

, mainly in
 Nemean Seven 319
Class I poems. Schneidewin accepts Wieselers emendation, 


 k  , which is superior to other proposals and
accepted by many (e.g. Mommsen, Christ). That also removes syn-
izesis of . Snell nevertheless returns to the paradosis. He is
now fully refuted by Barrett (Collected Papers, 188).
22 (s1): Insertion of  (Hermann, Notae) is necessary.
334 (s45). Snells text adopts several emendations; H
participle ( MSS),
a (
MSS),  1st pers. sg.
( MSS). He puts a full stop before H. Boeckhs text is:
 H, d
a : O: P:   : d is rel. pron.,
and  is of course 3 p. Boeckh is followed by Mommsen and
Turyn.  H is suspect. Hermann (Notae) writes:
Locus hic antiquo vitio laborat, . Tum scholiastes pro a
habuit
a, idque recte et metro exigente. [ . . . ]  est is, qui
ad clamorem accurrit. Inde E auxilio venire clamantibus: unde
recentior forma  :  simpliciter de viris fortibus dictum
esse, valde dubito.  autem   , ut Olymp.I,
177. K
 $g e . Similiter Ol.XIII 137
   K
 "E 21. Dissen follows Her-
mann (  recte dici Graeci mortuos viros fortes,
heroes, dubitat Hermannus, cui plane assentior. At , h.e.
K
, sensu aptissimum, immo prope necessarium mihi videor
ob consilium loci) and so does Schneidewin. Christ expels .
His emendation is 
 . . . g: (mortem occumbere omnes
homines, sed honore ornari quorum famam mortuorum deus
augeat, exemplo Neoptolemi poeta illustrat, qui mortuus antistes
pomparum Delphicarum a diis constitutus sit). I am not sure who
was the Wrst to introduce the participle H and the Wrst person
 together. The participle is ascribed to Farnell by Snell (Gerbers
Emendations seems to support this), but the main verb there is .
35 (s6). I take Christs proposal 
 ( is scanned ww).
Most editors accept the paradosis 
 by scanning it as

www, with crasis of . This irregularly makes the initial position
of e 3 unresolved here but resolved at all the other repetitions. This
responsion may seem accepteble as an attestation of the All-but-One
Rule (see Part III, C), but when long sequences of short syllables
occur in expanded e phrases, like O1s8 (we 5) or P2s1 (e 6), exact
320 The Eighteen Majors
responsion is never breached throughout all the repetitions. At v. 103
(e3), 
 is scanned as w
www.
37 (s8). The paradosis oVers:
# b :
  N 
 .


ww
w

w
ww



j
The problem here is that, while drag is well attested in aeolic phrases
in dramatic poetry which end blunt, there is no parallel for a pendent
phrase (here hipponactean) ending


instead of w

. Mod-
ern editors (Turyn, Snell, Bowra) generally accept Boeckhs transpos-
ition:
# b ,
  N  .


ww
w



ww
w

j tel hipp
This, however, produces base in the form

, whereas in all the
other repetitions the base of the hipponactean takes the form
w.
Moreover, elsewhere in this poem correspondence in aeolic base is
exact, as is characteristic of Class III stanza-forms. Hermann
(Notae) sought to preserve the MS order by reading
 , an
unattested form. Bergk2 in the apparatus, proposed

 vel

B.

 had been proposed by Ahlwaldt. Mommsen
introduced
A in the main text. I regard
A as preferable
to transposition. Note that the subject of  is implicit in the
manuscripts. Barrett supposes
A , to avoid SVE; see Collected
Papers, 189.
41 (e4). ( (E. Schmid) is necessary for (.
43 (s1). Transposition of
a and 
 (E. Schmid) is
necessary.
46 (s4). & had been scanned
ww
. Schroeder introduced
& .
5960 (e12). The paradosis gives I : I
 (Hermann, Notae) recovers both sense and metre.
61 (e3). For the prosodic licence of N -, see West, GM 17.
Boeckh writes , and Bergk24, in the apparatus, haud dubie
vitiosum, . . .  postea K . . ..
65 (s2). The paradosis oVers
" $
b e N d


 ,   
 Nemean Seven 321
Scanning N disyllabically (Turyn and Snell), this produces

ww
wwww



ww
w

w
w
k
At v. 86, the MSS read

H b, where, metre apart, b is
completely inappropriate. Turyn (

H ) and Snell
(


 ) produce the same scansion as v. 65. Thus in
these two verses Turyn and Snell introduce long where there is
short in all the other eight repetitions. Both posit phrase-division
before the anceps thus produced:

ww
wrw
u

ww
w

w
w
k52
It is, however, doubtful whether a glyconic of the form w


ww
w
is found except at the beginning of a verse. And respon-
sion between w
and

at base is highly exceptional. An alter-
native would be to analyse

ww
rww
u

ww
w

w
w
k d dod dod e 2
but the Wrst dodrans, with resolved Wrst long and drag, is most
peculiar. The position in question is not anceps but a real short.
It is much simpler to eliminate both longs by following Hermann
(Notae) in deleting d at 65 (d is not connective but aYrmative,
thus there is asyndeton, anyway Carey ad loc.), scanning N
trisyllabically, and in reading the epic form  (Jurenka, Novae
lectiones Pindaricae, 28) at 86. So Bowra. Instead of deleting d at
65, Mommsen proposes d  (cf. his change of
 into
 at O9 v. 89 83 Sn.). He is followed by Christ (et apud
peregrinos et inter cives incedo . . .).
66 (s3). $
 (E. Schmid) is necessary for $

.
77 (s6). I
 (Triclinius) is necessary for I
.
78 (s7). For the irregular scansion e, see Maas, Freiheiten,
Beleg 57.
81 (e2). The word-order of the paradosis (o  ) does
not make correct responsion. Emendation is easy by transposition; of
two proposals,  o  (E. Schmid, Hermann (Notae),
Snell) is better than  o  (Turyn) because it does not
require correption in non-dactylic movement.
52
Strictly speaking, their metrical analyses are diVerent. Turyns analysis (nota-
tion) is glyc glyc do, while Snells is 2 cho hipp ia.
322 The Eighteen Majors
83 (e4). & (Benedictus) is necessary for A or A.
86 (s2). See above on v. 65.
98 (s6). The paradosis  is unmetrical.  (Triclinius) is an
easy correction. Snell accepts Maass f .
104 (e4). 
(E. Schmid) is necessary for 
.

Individual verses
s1. There are eight examples of the phrase w

ww
w
k. Some of
these, like N2s1, are beyond doubt glyconic with aeolic base w
. Our
verse is one of the most probable cases of ^ e dodrans. On this
ambiguity in general, see further, Part I. 7. 4. e2 is common after an
aeolic phrase with 2 ending: excluding the examples in N7, there
are six verses in which the combination occurs. Bridge between the
two phrases is generally observed (Part I, 8. B. 6); this is the case in
our verse too, word-end occurring only at v. 51 e  out of
ten repetitions.
s2. See Part I, 5. F (resolution of aeolic phrase), 8. A. 3 (ii) (c)
(longer verse), 8. B. 5 (aeolic we), Part III, B.
The analysis of s2 is complicated by textual problems at v. 65 and v.
86. See above. Unlike s4 or s5, the sequences of four short syllables are
not occupied by one long word (except v. 65 $
b e ) but divided
into two. At six repetitions out of ten the division is between the
second short and the third so that the introductory movement of this
verse may sound dactylic; the typical case is v. 44 Ia e  j
I
:
ww
wwjww
wj; the others are vv. 10, 23, 31, 52, 73. At
the other three repetitions, the rhythm is
ww
wwwjw
, e.g. v. 2

E  . Contrary to the general tendency (Part I, 8. B. 6),
bridge between the Wnal e2 and the preceding dodrans is frequently
ignored: 10, 73, 94. Besides there are two repetitions in which a
sentence ends there, but with an elided vowel: 65, 86.
s3. See Part I, 8 B. 7 (aeolic e 3). Of 14 examples of e 3 (including a
strange form at P8s6; see below) in the eighteen majors, the totally
unresolved form
w
w
w
occurs only seven times, including
N7s3. The others include resolution somewhere, like s5 and s6.
Unlike e or e 2 it is surprisingly rare for e 3 to follow an aeolic phrase
(and there is no example of exe doing so). There are only three
occurrences, including P8s6. The other two are found in N7.
 Nemean Seven 323
Word-end is avoided between dodrans and e 3 except at v. 66
 
; and elision is found at v. 53 a 
  .
s4. For the ambiguity of the phrase
w
w
ww
between wila-
mowitzianum and ewd, see Part I, 7. 3. ClassiWcation is rendered too
delicate by the many uncertain borderline cases. Thus, without con-
sidering each passage in its metrical context case by case, I have
classiWed all the examples as wilamowitzianum, giving preference
to consistency. In N7s4, the phrase in question is prolonged by
w

, yielding wil 3. But it is one of those for which the alterna-
tive is highly probable: we ewar (aristophanean is dod
by deWni-
tion). The chart given in under Strophe is based on this analysis.
When the phrase is accepted as wil 3, we have a sequence we wil
3 similar to those in P5s2 (we wil ) and P6s7/8 (we wil 1). As an
example of wilamowitzianum(and its cognates with n) our verse is the
only case that starts with
wr. In other words, the Wrst long of reversed
dodrans following the base is resolved. This is not rare in reversed
dodrantes which lack base (9 examples, Part 5, E. 1).
s4 is one of the four verses of N7 which includes a sequence of four
or more short syllables in mid-verse (Part III, D). Characteristically,
all four shorts often belong to the same word: 4 K
, 12 K
(emendation), 33
a , 54 , 88 I.
s5. See Part I, 8. B. 4 (d e 3). Note the relatively frequent occur-
rence of brevis in longo (4 times). Moreover it is striking that a word
or word-group of four short syllables is used there: 5 K
d Y, 55 
I, 76 , 97 a . The ending of N3s2 is very
similar. Compare the use of tetrasyllabic words at s4.
This verse begins with short anceps d. Anceps d is quite a
common phrase at the beginning of a verse (8 examples in total).
However, in all the other instances it is long, except O10e10 (wd
e e
d) and P10e1 (xd gl) of which three repetitions out of four have it
short. Between wd and the following e 3, bridge is observed in eight
repetitions out of 10. There is word-end with elision at v. 5
(I
  P) and full word-end at v. 97 (IA

).
e 3 is rare after d, as it is after aeolic phrases (see s3 above), in other
stanzas: there is only one parallel:
O1e2 w

ww

wrw
w
j ^e d e3
324 The Eighteen Majors
s6. See Part I, 8. B. 7 (aeolic e 3). Telesillean is followed by e 3.
Unlike s3 and s5, bridge before e 3 is occasionally neglected (full
word-end is found at three repetitions; vv. 35, 48, 69; elision at v.
6). Both the Wrst and the second longs of e 3 are resolved, so that there
are six consecutive shorts (cf. Part III, D). Note the following words
or word-groups (cf. s4, s5): 6  , 27  , 35

 (see under Textual problems), 48  
, 56
I, 69
 Z, 77 I
, 90 n K
 , 98
K
 .
s7. See Part I, 5. F (resolution of aeolic phrase), 8. A. 6 (h)
(palindrome), 8. B. 7 (aeolic e 3). The resolution at the second
position occurs only at v. 70 (n.pr. P). As indicated in the
opening section, the colometry of s7 can be taken in two ways,
according to how we connect two shorts in resolution, and I take
(ii) as more probable (see the chart arranged horizontally):
(i)
tww
w
wr
w
k
d
e e
(ii)
tww
w
rw
w
k tel e2
The combination of tel e 2 is used at O1e6 and P11e3.
s8. See Part I, 8. A. 6 (a) (palindrome). Hipponactean is rare;
Besides our verse, there is only one (P2e8). Both are located at the
end of a stanza-form. Emendation at v. 37 is necessary; see textual
problems.
e1. See 8. C. 8 (d e); 8. C. 9 (e d). we d, (in drama, ia ch), is
surprisingly rare; it is found in only one other verse, O10e1. On the
contrary, the phrase d e is incorporated in many verses. The re-
solved form
ww
rw
is quite frequent. In N7e1, word-end
at . . .
wjww
k in three repetitions is conspicuous, perhaps more
conspicuous than ordinary cases because of verbal assonance (17
E , 38 K O, 101  Z
).
e2. For the ambiguity of the initial www, see above. There is another
verse which is identical:
O1e6b rw
ww

w
k rdod e
And the combination rdod e is incorporated in these:
P5e6
ww
rw
ww

w
j d rdod e
P2s7 rw
ww
rw

w
k rdod e e
 Nemean Seven 325
Alternatively, there are three examples of ^ e d e (P10s2b, P2e6,
P5e5); the initial ^ e is not resolved in these. Word-end is strictly
avoided between reversed dodrans and e at N7e2.
e3, e4. See Part I, 5. F (resolution of aeolic phrase). As discussed in
the opening section, and as presented in the chart of Epode above,
two colometries are possible, depending on how the second www is
taken. I analyse gl e, giving preference to consistency with verses in
other odes. The eighth position of glyconic is surely resolved; respon-
sion between unresolved and resolved at P6s3 and I8s5c attest it.
e5. See Part I, 8. A. 7 (repetition within a verse). The aeolic
hendecasyllable (gl 3) is found elsewhere only at
O9s2 x
ww
w

x
ww
w
w

k tel gl 3
I7s3/4

w
w
ww
w



ww
w
w

k
e tel
gl 3
Repetition in sequence of the glyconic with base www has no parallel,
but the pair of telesilleans or reversed dodrantes in the following
verses is comparable:
N3e4 ww
ww
w
ww
ww
w

w
k tel tel e
N6s3 www
ww
www
ww
ww
ww
j rdod ribyc
Word-end clearly occurs after the Wrst glyconic at three repetitions
(vv. 21, 42, 63). At one of the other two, bridge is not strongly
marked: 84 `N .
326 The Eighteen Majors

ISTHMIAN SEVEN

Three triads. Class I

I7s1 ww
ww
w
w

k tel 3
I7s2 a
ww
w
w
w

w
k telwe e
I7s3/4

w
a
ww
w

j

ww
w
w

k
e tel gl 3
I7s5a


ww
w


ww
w
u
w
ww
j gl tel hepta
I7s5b

w
w
k
e 2
I7e1
w
ww
w
w
w
k glwe
I7e2

ww
w

ww

k tel adon
I7e3

ww
w
j tel
I7e4 ww
ww
w
w

k tel 3
I7e5
a
ww
w
u
ww
w
k gl tel
I7e6
ww


k d sp
I7e7

ww

w
ww
k
d rdod
s1 H 35; s2 B 2, 36, 41; s3/4 H 25; B 8, 37;s5b B 39b; e1 H 28, B 11; e2 H 29; e4 H 48; e5
HB 32; e6 H 33; B 16, 33, 50;53 e7 H 17

s1 1 (1) ( H

 , t 
,
s2 2 (2) H K
 
 e 
s3/4 3/4 (3) h ; q 


(4) 
  P
s5a 5a (5)  , j H  

s5b 5b e  H,
e1 11 (11)  
 K @

;
e2 12 (12) j  I
 o OH
e3 13 (13)  K
d H
e4 14 (14) ,   
e5 15 (15) `NE  ,   ;
e6 16 (16) Ia
a

e7 17 (17) o
, I
 b ,
Most of the verses of Isthmian 7 are unequivocally aeolic. Both the
strophe and the epode can be reasonably attributed to Class I (aeolic

53
See on Textual problems.
 Isthmian Seven 327
odes). The ancipitia, both those of half-base in the telesillean group
and those of full base in the glyconic group, are realized as long in
several examples. RSS is comparatively low, 44.0% (strophe) and
46.0% (epode).
Typical features of aeolic metre are identiWable at s3/4, s5a, and e5.
At s5a, three phrases, gl, tel, and hepta, are set out in that order
(22 positions). Verses composed of three aeolic phrases are found
in another typical Class I stanza-form, N2s4. Phrase-boundary
tends not to agree with word-end, though there are exceptions (1/6
and 3/6). All the ancipitia of full and half-bases in s5a are long except
one. The combination of gl tel is repeated at e5. In s3/4, dovetailing
is found between tel and gl 3.
Other verses are fairly short. In some of them an aeolic phrase is
expanded by an apparent bacchiac (s1, s4, e4). Such expansion is a
common feature of tragic aeolics, but less common in the eighteen
majors. There are in total seven examples of the phrase expanded
by 5 ending (tel 3, gl 3, hepta 2 3) in the eighteen majors,
and its concentration in I7 is noteworthy.

Strophe
Of the aeolic phrases, telesillean appears most frequently. In fact
every verse includes it except for the last (s5b). In the Wrst two verses
(s1, s2) telesillean is expanded to the right; in the next (s3) it is
expanded to the left and followed by a hendecasyllable (gl 3). The
ending of this verse ( 5, w
w

) is common to s1. The number
of positions in s3/4 amounts to 22, of which, the Wnal position being
taken as anceps, 10 are true long. Interestingly, both Wgures are
exactly equal to those of s5a. The two verses are related in the
following manner:
s3/4

w
a
ww
w



ww
w
w

k
s5a


ww
w


ww
w
u
w
ww
j
Note the locations of

gl and hepta. Unlike the others, the last
verse (s5b) has no double short. Though the scale is smaller, it is
comparable with the last verse of the strophe of P6 (^ e e 2 xe) or
that of P8 (xexe 2), for these stanza-forms are, except for the last,
made up exclusively of aeolic verses.
328 The Eighteen Majors

Epode
The key phrase of the strophe, telesillean, is heavily used in the epode
too. It stands as a verse by itself in e3. And e4, tel 3, is identical with
s1. In e5 two thirds of s5 is repeated (gl tel). The expansion of
glyconic by w
w
(we) in e1 is quite similar to s2 (telwe e).
However, there are some diVerences. In e2 telesillean is followed by
adonean, which, alternatively, may be d anceps. d is also used in the
last two verses (e6, e7).

Textual problems
8 (s3). The word-order in the paradosis, 
E , is un-
metrical, and is corrected by Pauw (
E ().
11 (e1), The paradosis N is unmetrical; K E. Schmid.
23 (s1). The paradosis  N

 does not respond. Bergk24
restores responsion with b N
 (he also introduces
E
N
 at O6. 30 for the unmetrical
E N
). The older
editions adopted  N (E. Schmid).
28 (e1). The paradosis I must be emended. Since I
seems an intrusion from the preceding verse (27 I), there are
many possibilities for emendation: I (Hermann, Notae),
Ia  (Thiersch, Christ, Willcock), I

g (Mommsen),
Ig (Mair, Turyn), alii alia. Snell obelizes e I.
33 (e6). Brevis occurs at every repetition. This is certainly unusual
(the other example is O10e9; cf. Part III, F). It is impossible to
combine this verse with the following. Verse-end is attested by hiatus
(it occurs only at v. 33). Also unusual is the verse-form which
I analyse as d spondee (see below). Some editors try to expel the
hiatus at 33. For example, Christ, in his apparatus, proposes
 instead of P at the beginning of 34. The newly com-
bined verse would be

ww

w

ww

w
ww
k d e d rdod
This is acceptable although there is no example of d e d in the
eighteen majors. Bergk4 introduces a quite diVerent metre. Starting
with the suspicion about 
  (Amphiaraum non decuit
componere cum Meleagro et Hectore, qui patriam fortiter defendentes
 Isthmian Seven 329
occubuerunt; the so-called Amphiaraus diYculty; see David Young,
Isth. Seven, 21 nn.72, 74), he intoduces I 
  (sc. temple of
Amphiaraus). The metre so produced is:
x
ww

k reiz
Reizianum harmonizes with the Class I context. But the two corre-
sponding lines must also be emended, though the emendations are
very slight: 16 I
a a, 50 E  fg. Bergk is
followed by Bury (with a further change: I 
). I am now
persuaded by Barrett, who brilliantly solves the Amphiaraus diY-
culty (Collected Papers, 1979). His supposition is I e E .
He unites e6 with e7 (see above on Christs conjecture). As stated
above, the combination of d e d rdod is unparalleled, but is
explicable enough to be accommodated in Class I context.54
36 (s2). The unmetrical K
is deleted by Callierges.
43 (e3). The paradosis  is improved by Benedictus:  .

Individual verses
s1. Telesillean and tel 3 with the initial anceps (half-base) replaced
by two shorts are quite frequent (8 in total; Part I, 5. D. 2). Verbal
assonance: 18   40 .
s2. See Part I, 8. B. 5 (aeolic we); 8. C. 5 (double e). Telesillean
comes again, but the initial anceps is literally anceps in this verse
(short 4 times, long twice). Contrary to the frequent occurrence
of cretic (e) after telesillean or glyconic, iambic (we) is rare (4 ex-
amples). The last two positions of telesillean iambic metron pro-
duce the sequence of w
three times. At all six repetitions, word-end
occurs either before or after the Wrst short link anceps. Between
we and e, bridge is observed at Wve repetitions among six, with the
exception of v. 36 K
 K
.
s3/4. See Part I, 8. A. 3 (longer verse). Since Boeckh, this verse has
been divided into two at the point where word-end occurs in all the
repetitions:
I7s3

w
w
ww
w

j
e hag
54
I retain the analysis and the statistics in Part I unchanged so as to avoid
confusion at the Wnal stage.
330 The Eighteen Majors
I7s4

ww
w
w

k tel 3
But apart from coincidence of word-end, there is no manifest indi-
cation of verse-end. It is more probable that s3 and s4 make one
verse, and that the coincidence of word-end is a manifestation
of dovetailing; see above. Here Pindar seems to share a common
tendency with the tragedians. The current colometry makes s3 a
hagesichorean, which is hardly ever used in the Pindaric corpus
(only one certain example, in P2s8). At Wrst sight, anaphora at
vv. 89 may be thought to support the old division into s3 and s4:
8 j Id
E ( E ;
9 j I "

;
But that is not necessarily so. Compare the following verse:
10 j #
H IA; j  A @ K IA
Here the second j occupies not the Wrst but the second position of
the second phrase by dovetailing. Lines 8/9 and 10 are in anaphora on
a larger scale. Note also the second X clause at v. 5, which starts in
mid-verse (s5). Note too the similarity of the meaning of the two
sentences at the same metrical position (verbal assonance in a wider
sense): 25 z
 b @   42 
 a H

 .
s5a. See Part I, 8. A. 3 (longer verse), 8. A. 6 (f) (palindrome).
Though manifest indications are absent, verse-end is almost certain
after the third phrase; otherwise the verse would be extraordinarily
long (28 positions). Besides this verse and e5, there is one example of
glyconic followed by telesillean, P2s2; there bridge is strictly observed
between glyconic and telesillean but almost totally neglected after the
long anceps (the half-base). 5. Here, in I7s5a, the tendency is not
quite so strong. Between glyconic and telesillean, bridge is observed
at nearly all the repetitions (one exception, v. 39 I ;
but v. 22   
 ). After the long anceps, word-end
occurs at two repetitions (vv. 10, 44).
The half-base of the heptasyllable is short only at v. 27
(
  ). Heptasyllable following telesillean or other
phrases ending with w
is totally unparalleled.
 Isthmian Seven 331
s5b. See Part I, 8. A. 3 (shorter verse). Anceps e 2 can stand for a
verse by itself. There is one certain case, O10e7; it is interesting that
this too is preceded by a telesillean. Anceps e 2 may be associated
with telesillean, the key phrase of this poem. If its Wrst single short is
changed into double short, the phrase
e 2 becomes telesillean.
e1. See Part I, 8. B. 5 (aeolic we). It is as if the ending of the
glyconic were prolonged further by w
w
( 4 ending, so to
speak). The examples are rare. The other is P5e2. The ending is
articulated in all three repetitions: v. 11 K @

, v. 28
K H, v. 45 K PF  ; note the verbal assonance
in a looser form. At v. 45, the Wrst syllable of  is scanned as
short, as at P2e2, v. 90, and P6s9, v. 45.
e2. This verse is unique in that
ww

follows an aeolic phrase.
I analyse the phrase as another aeolic phrase, adonean, in this
context. Adonean is the catalectic version of
ww
w
, dodrans,
which is incorporated in the preceding telesillean. As a catalectic
phrase,
ww

by deWnition does not stand before other phrases
within a verse. Even when we take this into account, the phrase is
surprisingly rare. There is only one parallel at most (P10e2), and that
may not be adonean but d anceps.
e3. Although hiatus/brevis is lacking, verse-end is certain. The
following verse e4 is a well-identiWed form, the same as s1. Telesillean
makes a verse by itself in O10e6 (half-base
), P5s7b (
), I8s3b (w),
and O9s1 (
).
e4. The half-base of the telesillean is changed from long to double
short, the inverse relation to s12.
e5. See Part I, 5. C. 1 (proper noun involving aeolic full base of the
shape
x). The full base of the glyconic is

only at v. 15
`NE, a proper noun; it responds with N and   , t; a
verbal assonance? In contrast to s5 and P2s2, bridge is neglected
between glyconic and telesillean in two repetitions out of three, and
perfectly observed after the long anceps.  (15) and  (49)
occupy the same metrical position.
e6. See Part I, 8. A. 3 (shorter verse). All repetitions end with a
short syllable, but hiatus at v. 33 proves verse-end, unless we adopt
Barretts emendation (above, under textual problems). This unusual
verse may be analysed theoretically
332 The Eighteen Majors
(i) dodrans (
ww
w
) dragged;
(ii) hemiepes or D, with the second biceps contracted;
(iii) d and spondee.
In Part I, I reject (i) and (ii) because there is no certain example of
either drag or contraction. Spondee takes the place of cretic (e) in
D/e and freer D/e. The closest parallel is:
O9e5
ww




k d sp sp
e7. See Part I, 8. B. 4 (d aeolic). There are in total seven verses
which start with anceps d, but no exactly identical verses are found.
The nearest parallel is
P10e1 x
ww

x
ww
w
k xd gl
and the two phrases anceps d and reversed dodrans are included in
the same order in the following verses:
P10s4 w

ww
x
ww

w
ww

w
w
k ^edxd rdod e2
N3e3
ww
w
w
ww

w
ww

w
w
k dodwd rdod e2
Here in I7e7, the relation between the two phrases is comparable with
that between telesillean and glyconic.

ww
is telesillean minus
w
and
w
ww
is glyconic minus w
. Bridge is observed after

d, as in P10e1 (at all 4 repetitions), P10s4 (with one exception out
of 8 repetitions). Note the verbal assonance 34 P 51 P
(and 17 o).
 Isthmian Eight 333

ISTHMIAN EIGHT

Seven strophes (monostrophic). Class III

(1)
I8s1/2 w

w
ww
w
wy
w
w
ww
k wilwe wil
I8s3
w
w
ww

w
wrj w
ww
w
k wil e 2tel
I8s4
w
ww
w

w
w
ww
k gl wil
(2)
I8s5a w

w
ww

w
ww

w
ww
j wil rdod rdod
I8s5b
w
ww

w
ww
j rdod rdod
I8s5c
a
ww
wt
w

k gl e

I8s6
a
ww
w

ww
k gl d
I8s7
w
wrw
w
ww
w

w
k e 3tel e
(3)
I8s8 rw

wr
wr
k e e e

I8s9 w

ww
ww
k ^e D
I8s10

w
w
rw

w
ww
k
e 2e rdod
s1/2 B 22, 32; s3 H 43, 53; s4 H 4, 14, 24; B 64;s5c H 15c, 35c; B 15c, 35c, 45c, 65c; s6
B 56; s7 H27, 67; s8 H 18; B 8, 18, 28, 38,55 48, 58; s9 H 39, 69; B 29, 49; s10 B 10, 20,
30, 40

s1/2 1/2 (1) 

     h, t , 

s3 3 (2)
e Ie (

a

(3) Ng I
s4 4 (4) H, "
  
, d 
s5a 5a (5) I 
 KF H d K, 
 I
s5b 5b (5a) , N  
s5c 5c (6) E: K 
 b
  
s6 6 (6a)  K O
 
,
s7 7 (7) 
 



  I

 H
s8 8 (8)    d a
:
s9 9 (9) K
c e $
b A
s10 10 (10) y (
 

    ,
Isthmian 8 is a monostrophic ode with seven repetitions. In spite
of the large number of repetitions, verse boundary cannot always be

55
Word-order is corrected; see below, Textual problems.
334 The Eighteen Majors
settled with certainty. See s1/2 and s3 below. Although I8 is mono-
strophic, its stanza is extremely long. The total number of positions
is by far the biggest (163 positions). Compare N2, the shortest
monostrophic ode (65 positions). It is even bigger than the total of
the strophe and the epode of P11 (129 positions) or of others (I7,
P10, P8, O2 (only the eighteen majors being taken into consider-
ation); but it is smaller than the total of any whole triad; a triad of
P11 amounts to 201 positions). Its structure is not so simple as those
of the other monostrophic odes (P6, N2, N4). They are all simple
aeolic (Class I) whereas I8 is not. O14, one of the four minors, is
very similar to I8: it too is monostrophic, belongs to Class III, and
is 157 positions long.
Certainly there are various aeolic phrases in I8. Among them,
reversed dodrans
w
ww
is the key phrase. Especially conspicu-
ous is its frequent repetition, no less than Wve times, from s5a to s5b
(the Wrst is incorporated in a wilamowitzianum). Its longer cognates,
glyconic (three times) and wilamowitzianum (four times), are used
frequently too. But I8 is diVerent from the stanza-forms of Class I. In
these aeolic phrases tend to be constructed colon by colon, while in I8
they are not. At s5c and s6, the base of glyconic is represented as
a,
which may sound characteristic of Class I but is in fact a special
licence to admit a proper noun, `Y, the key word of I8 (Part I, 5.
C). On the other hand, there are similarities to N3 or N7 (Class III):
preponderance of short anceps, long runs of single-short movement
(s3, s7), and freer D/e, which appears in the Wnal part of the ode (s8
10) and includes even D. RSS is 52.7%. This is not signiWcantly high,
but higher than that of O1e, one of the typical Class III stanzas.
The structure of I8 is indeed complicated, but less complicated
than the variety of names given to each phrase suggest. The process of
generation from the preceding verse to the next is fully comprehen-
sible. The stanza can be divided into three sections: 1 s1s4;
2 s5a7; 3 s810.
In 1, what I conventionally called e 2 tel (s3) may be analysed as
e wgl; the analysis is less important than the fact that s3 as a whole is
of equal length and structure to the preceding verse, s1/2:
s1/2 w

w
ww
w
wy
w
w
ww
k
s3
w
w
ww

w
w rw
ww
w
k
 Isthmian Eight 335
Both verses start with wilamowitzianum, and end with it or its
cognate, glyconic. A tetrasyllabic single-short phrase is sandwiched
between them. The direction of each single-short movement corres-
ponds to the beginning of the verse (the aeolic base): iambic (w
)
in s1/2, trochaic (
w) in s3. Two longs are juxtaposed without an
intervening short at either border: that after the iambic phrase in s1/
2, that before the trochaic in s3. In other words, it is as if the iambic
phrase were attached to that preceding to make up a dodecasyllabic
aeolic colon, while the trochaic phrase were attached to that follow-
ing to make up an aeolic of the same size. These two verses have a
further aYnity: the last phrase of s1/2 is the same as the Wrst of s3.
In the next verse (s4), the sandwiched phrase, the trochaic, drops out,
and the wilamowitzianum and the glyconic are interchanged.
2 starts at s5a in the same manner as s1/2, after which comes
its principal part, reversed dodrans, which is repeated four times.
Glyconic is found again in s5c, where it is followed by e anceps
(interestingly, the anceps is never Wlled with a syllable long by nature;
it is as if this phrase were equal to the Wrst four positions of glyconic).
The same phrase is repeated in the following verse (s6), but the suYx,
e anceps, is changed into d. The diVerence between the two suYxes
is in the penultimate position alone: long in the former; short in the
latter. The main part of s7 appears unusual at Wrst sight, but is in fact
a longer form of the latter half of s3:
s3 (end)
w
wr w
ww
w
k
s7
w
wr w
w
ww
w

w
k
e is attached to it, and this phrase is repeated in the next verse (3, s8)
with abundant resolution.
Aeolic phrases are absent from 3 except at the end of the stanza
(s10). The key phrase, reversed dodrans, returns and provides a
transition to the beginning of the stanza like a sort of metrical
ring-composition. The dactylic phrase D appears for the Wrst time
in the middle of this section (s9), but the underlying structure of the
verse is not so diVerent from s8:
s8 rw

wr
wr
k
s9 w

ww
ww
k
The Wnal verse starts again with single-short movement.
336 The Eighteen Majors

Textual problems
10 (s10). Emendation is necessary to avoid  (B, Scholia; D) at the
initial position. Carey tries to acknowledge an occasional licence,
whereby a postpositive can stand at verse beginning. But his parallels
are weak; see Part II, N4, Textual paroblems, 634. Bowra and Turyn
accept a mechanical transposition (Bergk1):   (
. This
is inelegant: stone does not need to be emphasized. It is not good
metrically either. At all the other repetitions the initial position is
Wlled with a long syllable. The position is anceps, and a short syllable
is, of course, acceptable, but in all the other eight examples of xe 2 at
verse beginning Pindar tends to keep the same length in responsion.
We should thus avoid introducing a breach of the tendency. Momm-
sen introduces d (
 ; but emphatic d is meaningless.
Bergk4 admits the responsion between double short and long here:
 (
 . It is highly exceptional in Pindar. Privitera
follows Bergk, and analyses  (
 as iambic. But the reso-
lution of the initial anceps is permitted only in the spoken part of
drama. Snell obelizes the text.
1112 (s12). I am not conWdent of understanding this notori-
ously diYcult sentence. I provisionally read I Kd E b

 a 
 , construing

with E (so Hermann and others, including Boeckh and Christ
et al., who print I
). The paradosis gives
. Some
take
 as neuter (the events passing by; Mommsen
(with
 instead of E!), Schroeder, Snell). Bowra and Turyn
adopt Boehmers Kb with A . . . A. For the discussion
of its metre, see below on individual verses, s1/2.
13 (s3). The unmetrical paradosis Nd is emended by E. Schmid to
give Id. Between it and B a lacuna is discernible. A papyrus
(P. Oxy. 2439) provides 
, which is adopted by Snell.
15 (s15a). The paradosis  is unmetrical;  Triclinius.
18 (s8) The paradosis  B is unmetrical;   
Triclinius.
212 ( 21 Sn.; s1/2). The paradosis Kg A is unmet-
rical according to my analysis (which follows Hermann and Maas).
See below on s1/2.
24 (s4). The paradosis K
 is unmetrical; K
 Triclinius.
 Isthmian Eight 337
25b ( 25a Sn.; s5b). The paradosis   is unmetrical;
 Triclinius.
29 (s9). The paradosis  is unmetrical;  Triclinius. So is
;  E. Schmid.
31 (s1). The paradosis X is metrically wrong. The chor-
iamb in the wilamowitzianum may not be replaced by three longs.
Carey, citing Radt, Paian, 93, tries to defend the paradosis, but the
examples given by Radt are not necessarily appropriate. K


(Triclinius) is adopted by both Turyn and Snell, KB (Bergk4) by
Christ.
33 ( 32 Sn.; s3). The paradosis   
e
E needs another short. Various words have been proposed for
insertion after : for example, (E. Schmidt),  (Boeckh, but
correption before digamma in this pronoun is improbable), 
(Bury; followed by Turyn). Bergk4 proposes    and
  i and prefers the former; Christ adopts the latter.
But is the potential optative suitable for a destined event? Schroeder
(BT), following Ahlwardt and followed by Bowra and Snell in turn, is
more drastic: 
   E. Mommsens
  
e  is on a similar line. Cf. Barrett,
Collected Papers, 185 n. 195, who supposes    K

 .
35 (s5a). The paradosis d is unmetrical; d Triclinius. So is
; Snell adopts Triclinius , Turyn Hermanns   (Notae).
Bergk4 proposes d .
37 (s7). The paradosis  E is unmetrical; Hermann
(Notae) transposes: E @, and Boeckh inserted  .
389 (s89). The paradosis
E 
 O


 N
 does not correspond with the metrical scheme. Hermann
(Notae) transposes  to follow  and introduces the
forms , O

, `N . He is followed by editors.
40 (s10). The paradosis d is unmetrical.
 (Bothe) is
generally accepted.
41 (s1). Pf should just be changed to Pf. See below on s1/2.
43 ( 42 Sn.; s3). The paradosis  is unmetrical; 
Boeckh.
45b ( 45a Sn.; s5b). The paradosis K
 is unmetrical;
K
 Triclinius.
338 The Eighteen Majors
46 ( 46a Sn.; s6). The paradosis  is unmetrical; 
I Hermann (Notae).
47 (s7). At Wrst sight  (Bergk24, adopted by Turyn; i.e.
Zeus and Poseidon), which introduces long corresponding with the
third short, is metrically blameless, notwithstanding the absence of
long anceps at the position elsewhere. However, though it is analysed
as the half-base of telesillean, the position is always realized as short
in such a long single-short sequence in leftward extension from d.
Thus similarity with a number of examples, including I8s3, is
broken. From this point of view, the paradosis  (implicitly
Zeus only; so Snell) or  (Triclinius, Bowra) is preferable.
a  is the correction of E. Schmid. The paradosis is 
I.
48 (s8). The paradosis I =I is unmetrical; 
E. Schmid.
52 ( 51 Sn.; s1/2). The paradosis  is unmetrical; (
E. Schmid.
55b ( 55a Sn.; s5b). The paradosis I is unmetrical;
 E. Schmid.
56 ( 56a Sn.; s6). The paradosis Id 
 produces an
improbable hiatus which preserves long ; so, rather mechanically,
Id 
, Hermann (De dialecto). Snell introduces
hK
i
, which is improbable because of the correption in non-
dactylic movement. In his apparatus Snell registers

 with the
name of Maas, but it is in fact Schroeders (ed. maior 2).
58 (s8). The paradoses  and  are unmetrical; 

Heyne,  E. Schmid.
59 (s9). The paradosis is unmetrical; p Boeckh.
60 (s10). The paradosis K   is nonsensical and unmetrical;
K  Callierges.
62 (s1/2). The paradosis  is unmetrical; 
Bothe.
63 (s3). The paradosis I

 is nonsensical and unmetrical; 


 Hermann (Notae).
65 (s5). The paradosis I d is nonsensical and unmetrical;
I d Triclinius.
65b ( 65a Sn.; s5b). The paradosis e . . . a is wrong in both
sense and metre; e . . . a Triclinius.
 Isthmian Eight 339
68 (s8). The word-order of the paradosis ( 
d )
does not correspond with the metrical scheme. Hermanns transpos-
ition with correction (  
d, de metris) is accepted
by Snell, Bergk34s 
d   by Turyn.
70 (s10). The paradosis $
e 
is nonsensical and unmetrical.
Editors accept Triclinius $
e  A, with some hesitation.

Individual verses
s1/2. See 8. A. 3 (ii) (c) (longer verse), 8. B. 5 (aeolic we). It is
diYcult to decide whether to divide this verse into two (Boeckh,
Turyn) or not (Hermann, Snell). The metrical problem involves a
textual problem. It is also complicated by the interpretation of the
perplexing passage at vv. 11/12. Emendations (and preservation of
the paradosis) can be classiWed into three from a metrical point of
view;56 my analysis above is based on (iii).
(i) Division into two verses:
w

w
ww
w

k
t
w
w
ww
k
(ii) Combination into one verse, leaving an unparalleled responsion
at mid-verse:
(ii) (a) w

w
ww
w
at
w
w
ww
k or
(ii) (b) w

w
ww
w
at
w
w
ww
k
(iii) Combination into one verse, with emendation metri causa at
mid-verse:
56
Or four, since Mommsen acknowledges two verses in the same manner as
(i), but at the same time establishes verse-end within a word: Kd. He calls it
asynarteton. The idea of asynarteton is, for some scholars even after Boeckh, a last
resort to make compatible both verse-end and carrying-over of a word. It survives as
late as Christs editio maior (1896), although it should be noted that Christ does not
admit verse-end inside a word: Versus primum et secundum, cum et numeris et
sententiis (cf. vv. 11.12; 21.22; 31.32) cohaereant, in unum versum Hermannus
Bergkius Mommsenius coniunxerunt, quorum auctoritati quominus obtemperarem,
syllaba media ancipiti versuum 21. 41 prohibitus sum. Divisi igitur versus, ita tamen,
ut ex duobus versibus unum asynartetum versum eYci putarem. According to our
diWnition of verse, anceps iuxta anceps is a mark of division of verses. For asynarteta,
see Itsumi, Whats in a Line?. Privitera adopts the same text (colometry) as Momm-
sens, and analyses the Wrst half as Pindaric hendacasyllable. For this, see Appendix B.
340 The Eighteen Majors
w

w
ww
w
wt
w
w
ww
k
(i) is accepted by Boeckh, Christ, Schroeder (ed. maior 1), Bowra,
Turyn. Boeckh divides into two verses, rather navely. His text in-
cludes irregularities in both verses: I
k  (v. 11; prepositive at
verse-end followed by postpositive at verse-beginning), and  k (v.
31; elision at verse-end). Ia at verse-end seems to be paralleled by
O9e3 v. 55 ( 51 Sn.), but we must still ask whether that passage is
really parallel (see below), and  can be expelled to produce asyn-
deton. The most serious problem is  in the initial position. Thus,
Bowra and Turyn, following Boehmer, change  into K (two
shorts must be accepted anyway for v. 51 ! , but see below).
Certainly, K could stand at the beginning, and sense seems to be no
problem when A (genitive) is introduced. But I Kd=Kb is
a transitional idiom, as Carey points out. The division I
k Kb is
inelegant, even if grammatically acceptable. And at O9e3, v. 55 ( 51
Sn.), Ia is not construed with the next word but with the whole
sentence (Ia k e  I

 K k  E.)
When the verse is divided, the Wrst verse is metrically perfect:
w

w
ww
w

k wil 3
The second is less easy to accept:
t
w
w
ww
k te w d
Double short corresponding with long in (supposedly) anceps pos-
ition can be compared with O10e3, but this verse is highly unusual
metrically (see ad loc.):
t
wr
wr

ww
t
w
k te e
dte
Another, N6e6/7, is a crux textually as well as metrically (see ad loc.).
Turyn gives six parallels, but they are not comparable. At I8s2 the
double short is found in the paradosis only at v. 52 ! , which
may be a licence for a proper noun. Introducing double short by
emendation (Kb) is another matter. Thus, the emendation entailed
by (i) is the least probable.
(ii) (a) is adopted by Wilamowitz (GV 237), Schroeder (BT1), and
Snell. Following the metrical theory of Wilamowitz and Schroeder,
Snell calls the latter half of the united verse choriambic trimeter:
 Isthmian Eight 341
(2a) w

w
ww
w
ut
w
w
ww
k chodim chotrim
According to their theory, free responsion at the Wrst metron of
choriambic trimeter between w
w
 w


 w
wr is
admitted. But the idea is groundless (see Itsumi, Choriambic Dim-
eter). So (ii) (a) must be rejected.
Carey, too, rejects the division into two; but he does not seem to
accept (ii)(a). Presumably he takes ut as aeolic base, judging from
his citation of P5e9:
tu
ww
w

ww
w tw

w
k gl dwe e
(ii) (b) w

w
ww
w
ut
w
w
ww
k is worth serious
consideration. The former half is no problem. There are four ex-
amples of this type of decasyllable (wil 2;
w
ww
w
). Of
them, P5s8 and P2e8 start with w
and are identical with our case.
However, the latter part of (ii) (b) is questionable. We must ask (1)
whether the colon, which can be described as another type of deca-
syllable,
x
x
ww
, has any parallels; (2) even if that is
admitted, whether Pindar really admitted the responsion between
w


 www at aeolic base; (3) even if the responsion is
admitted, whether the form may occur in the middle of a verse, not
at the beginning.
1. There is no certain example of this type of decasyllable. Two
sequences seem to be applicable to the scheme:
O1e3
w
w
w
ww

w
k
P10s5 w

w
w
ww
w

k
But they can be better analysed in a diVerent manner. There is no
reason to pick out O1e3 and adapt it to this particular scheme,
separating it from the group of verses which have sequences in
single-short movement of various lengths turning to d (Part I, 8,
C). As for P10s5, the initial two positions should be taken as aceph-
alous cretic (^ e), judging from the metrical context.
2. In Pindars usage, responsion at aeolic base is restricted to iden-
tical forms and


w (Part I, 5. A. 2). There are only two
exceptions to that dominant tendency, O10s6 and P5e9 (see ad locc.)
Both occur unambiguously at the beginning of the verse.
342 The Eighteen Majors
3. Aeolic base w
comes as a rule at the beginning of a verse.
Admittedly, that is a matter of deWnition, because I consistently set
a boundary between two true longs (Rule 4a). If, for example, I8s6
were divided as

x
ww
jw

ww
k
the result would be diVerent. However, I have not so far found any
case that decisively refutes this rule of division. In short, the colo-
metry (ii) (b) is hardly plausible.
Hermann, Bergk4, and Maas are the champions of (iii). Metrically,
their proposals are perfect: an iambic metron (or short anceps e)
is sandwiched between two wilamowitziana:
w

w
ww
w
wt
w
w
ww
k
This structure is related well to that of the following verse (s3), as is
described in the opening section. The sandwiched iambic metron is
paralleled by
P5e2 w

ww
w
w
w

w
ww
j glwe rdod
N7s2
ww
wr w
w

ww
w

w
w
k dodwe dod e2
(emended; see note
ad loc.)
In order to adjust to this colometry, which demands short, instead of
anceps, at the third position of the iambic metron, two verses of I8
must be emended. One is simple; at v. 41, Pf should be changed to
Pf. The other is not so easy. At vv. 212, Hermann changes Kg

A into   A . But the active voice of 
does not
have the meaning to have sexual intercourse with. Bergk proposes
 KA (and in his apparatus  KA).57 But those do
not explain the intrusion of the less common form K. Maass
proposal is , `Y (vocative). He may be right to suppose
that A was originally a paraphrase of another word and in-
truded into the text (although the scholiast actually read the word
A in his text and noted that it was used euphemistically). My
tentative proposal is  :  is found at Hom.

57
Bergk34 introduces a bold emendation in vv. 1112, though this does not aVect
the metre: I K P E b
 A 
 A:
 Isthmian Eight 343
Od. 292, 254. Barrett, Collected Papers, 1913 adjusted Maass
emendation, proposing  , `Y.58
s3. See Part I, 6. C. 3 (resolved long followed by short anceps), 8, A.
3. (ii) (b) (longer verse), 8. C. 3 (e 2 w as half-base). Turyn divides s3
into two at the point where word-end occurs at all seven repetitions:

w
w
ww

w
ww
kB
w
ww
w
k
He is followed by Snell. The Wnal syllable of the former half is short
throughout, so brevis in longo must be assumed at every repetition.
That is exceptional, though a high degree of coincidence of brevis in
longo is not particularly unusual; for example, at I8s8, six repetitions
among seven have it (cf. Part III, F). As for colometry, the divided
verses are plain. The former one (s3a) is wil rdod. The latter (s3b) is
telesillean. It is metrically much better to unite them, however. The
structure of the stanza is well suited by the united verse, as is
demonstrated in the opening section. The long phrase following
the initial wilamowitzianum may look strange at Wrst sight (hence
Turyns division):

w
wr
ww
w

But this is a variation of sequences occasionally found in Class III
stanzas. They start with long single-short movement and usually turn
to double short. Here the double short is followed by w
(called
e 2 telesillean, but in fact, equal to e 2wdw
). e 2 telesillean can be
compared with e 2 hagesichorean (e 2wdw

), in
P2s8

ww
w

w
w
w
ww
w

k tel e 2 hag
And, as has already been remarked, I8s7 has a longer sequence of the
identical structure (e 3 tel e 3wdw
). As for e 2wd, there are two
parallels: P11s3, O1e3.
s4. This verse is one of the three examples of gl wil. Bridge is
observed at all the repetitions, and dovetailing occurs at three repe-
titions out of seven.

58
Two metrical notations in Barretts book are wrong, one in the main text (192)
and one in a footnote (n. 212).
344 The Eighteen Majors
s5ac. See Part I, 5. F (resolution of aeolic phrase), 8. A. 3 (i)
(longer verse), 8. A. 7 (repetition within a verse). The recurrence of
the identical phrase
w
ww
is noteworthy. The phrase itself is
repeated four times, and also forms part of the preceding wilamo-
witzianum and the following glyconic. Its second position is always
short, except once (v. 55c `Y, in glyconic). The whole of s5 can
be described as
w
6

w
ww
wt
w


From the viewpoint of colometry, there is not much diVerence
whether or not the whole is divided into three? The main reason
for the division is its gigantic length (44 positions in total).
A stronger impression peculiar to these verses is produced by the
coincidence of word-end at the end of the metrical unit
w
ww
.
Besides the end of s5a and s5b, where word-end is found regularly at
all the repetitions, word-end also occurs at six repetitions out of
seven at the end of the Wrst reversed dodrans in s5a; only v. 65 is an
exception (I ). And in s5b, two reversed dodrantes are divided
by word-end at four repetitions; there is elision at a third.
The eighth position of the glyconic at s5c is resolved at three
repetitions: 25c
, 35c , 45c K
 (whereas at 5c, -
of
 is scanned as one long). Responsion between resolved and
unresolved at the same position also occurs at P6s3.
At s5c an apparent trochaic metron (e anceps) comes after the
glyconic. Various types of aeolic colon are frequently followed by e,
but rarely by e anceps (Part I, 8. B. 5); besides our verse there is just
one example:
P6s6 www
ww
w

wr
k gl e

I8s5c displays a special relationship with s8, in which the Wnal e is
followed by anceps.
Word-end falls after the Wrst position of e at four repetitions, and
in the other instances the boundary shifts one position later (i.e. after
the short of e) always with postpositive (25c  45c ; 65c  ).
Interestingly, the identical tendency of word-localization is found in
P6s6 too. Brevis in longo occurs at four repetitions (hiatus too occurs
at two), and in the other three repetitions there is a short vowel in a
syllable lengthened by position, as if the four syllables of e
had
 Isthmian Eight 345
been intended to work as the four initial syllables of a glyconic
(
w
w) till a certain stage of creation. Also noteworthy is the
coincidence of word-end after the penultimate position of glyconic
(6 repetitions):

u
ww
wjt
jw

k
s6. See Part I, 8. B. 4 (aeolic d). Aeolic base of glyconic is
w,
except for one

(v. 16), to admit `N , exactly as in the
preceding verse (s5c). Unlike gl e but like gl e
(s5a), gl d is rare.
s7. See Part I, 8. B. 7 (aeolic e 3), 8. C. 3 (e 3 w half-base). For
the similarity to s3, see above. Moreover, with change of the tele-
sillean ( wdw
) to wd, O1s7 can be oVered as a parallel:

w
w
w
w
ww

w
k e3 w d e
s8. See Part I, 6. C. 2 (re er), 6. C. 3 (er anceps at verse-
end), 8. A. 6 (h) (palindrome); 8. C. 5 (triple e). Another colometry is
conceivable, but should be rejected:
rw
l x rw
rw
k e
e e
Long anceps is never followed by resolved long in the eighteen
majors; see Part I, 6, C. Parallels of the resolution at the penultimate
position of a verse end are collected and discussed in Part I, Appen-
dix, Addendum II.
A verse made up of three es, each of which is resolved, is familiar
from O2, e.g.:
O2s5 x
wr
w
rw
rw
k xeeee
But in O2 and in other stanzas, there is no example of the sequence
rw

wr, juxtaposition of two es resolved in diVerent positions
(Part I, 6, C). At verse-end a word (or word-group) of four short
syllables is employed at almost every repetition: 8 a
, 18
, 28 
 , 38 , 48  , 68 
 (Her-
manns transposition). That causes frequent brevis in longo (six out of
seven). Even the sole exception (v. 68) could be eliminated by
changing e (v. 69) into n.
s9. See Part I, 6. B (^ e d, ^ e D); 8. C. 1 (e and D within a verse).
The same verse is used at O10e2.
346 The Eighteen Majors
s10. A verse starting with anceps e 2 is quite common. The closest
parallel is:
P11e5 w
w
w

rw
k w e 2
e
The combination of e 2 and
e is also paralleled:
O1s10 w

w
w
rw
k ^ e e 2e
O2e3 rw

w
w
rw

w
k e e 2e e
As described above, the Wnal phrase
w
ww
provides a transition
to the beginning of the stanza. At the same time, it reminds us of the
Wnal verse of O2s, which is unique in having double short in that ode.
 APPENDIX A

The Four Minors

I have excluded from the analysis in Part I these four non-D/e odes: O4, O5,
O14, and P7. Their shortness is the main reason for the exclusion. O4 and P7
are each composed of only one triad. O14 is even shorter; it consists of just
one strophic pair. Thus O4s, P7s, and O14s have just two repetitions, and
O4e and P7e are never repeated. O5 is diVerent. It is composed of three
triads, but the verses are unusually few. Its authenticity has been questioned
not only on the basis of the scholia but for a metrical reason (see below).
When a stanza-form is not repeated many times, coincidence of word-end
seems less striking. When word-end coincides here and there, the colometry
of these stanza-forms tends to be uncertain, for it is theoretically possible
that every coincidence of word-end should be verse-end. For example, four
coincidences are found in O4s1 s1b:
ww
w
wwj

ww
ww
j
w
wj

k
(The Wnal one is unambiguously verse-end. It is attested by hiatus in both
vv.1b and 11b). Boeckh established one long, undivided verse as above. I
prefer to divide his long verse into two after the second coincidence, for this
division makes the metrical articulation clearer (see below). But admittedly
this argument is not conclusive. It would be possible to divide the verse
further, as Bergk4 does, after the Wrst coincidence, by the assumption of
brevis in longo: ww
w
w
k. This division does not make for good metre,
and can be dismissed, but the possibility of brevis at all the repetitions
should not in itself be summarily rejected. Thus Turyn divides
P7s2 w
wwj

w
u
w
ww
j
after the fourth position and makes a short verse: w
w
k. I prefer to reject
this too, but the possibility remains. In astropha like O4e and P7e, colometry
is desperately uncertain. Every word-end is in theory a candidate for a
verse-end.
The observations of the eighteen majors are of course helpful in some
cases for the decision of verse-ends. For example, O4s2a and s2b, which
make up one long verse in Boeckhs edition,
348 Appendix A. The Four Minors
ww
ww

ww

jw
ww
w

k
should be divided into two because there is no verse in the eighteen majors
in which both a dactylic movement and an aeolic phrase appear together.
Parallels also help. O4s5 and the following s6 should not be treated as two
independent verses, but be united into a single verse:
w
w
w
wwj
w
w
w

k
For this O9e1/2 is a good parallel:
w
w
w
w
ww
w
w
w

k w e2 w d w e2

It is itself paralleled by other verses; see Part I, 8. C. 3.
The following analyses of the four minors are necessarily tentative. I try to
cite parallels from the eighteen majors so as to place them in the perspective
of the whole of Pindaric metre.
 Olympian Four 349

Olympian Four
One triad. Class II

(1)
O4s1a ww
w
wwj

ww
ww
j ^ d wd D
O4s1b
w
wj

k e2

O4s2a ww
ww

ww

j ^D d

O4s2b w
ww
w

k hag
O4s3

ww

ww

j
d d

O4s4





j sp sp sp
O4s5/6 w
wj
w
ww
j w
w
wj

k wewdwe2

(2)
O4s7

w

w

w
j
e e e
O4s8a
w
ww
rw
ww
j rdod rdod
O4s8b

w
w
k
e2
O4s9 u
ww

j xd

O4s10 a
jww
w
w
w

k xdwe2

O4e1/2

ww
w
w
ww
w

j tel hag
O4e3 w
w
w
w
j we3
O4e4

ww
ww

w
k
D e
O4e5 w
w
ww

w
wrw
j hepta e3
O4e6
ww
ww

k D

O4e7

ww
ww
j D
O4e8 w

w
w
w
w
ww
j ^ e e3wd
O4e9/10 rwr w
ww
w
w
w
k ewdwe2

s1b H 1b, 11b; s2b B 2b; s5/6 H 5/6; s8b H8, B18; s10 B20

s1a 1a (1) c $
 A I


s1b 1b F d a ! *
s2a 2a (2) $
e
 IA
s2b 2b   

s3 3 (3) $
 
 I:
s4 4 (4)   s

s5/6 5/6 (5)  P I
d E K :
s7 7 (6) Ia 

E , n `Y 
s8a 8a (7) r
 I 

s8b 8b (H O,

s9 9 (8) 

s10 10 (9)  %   H,
e1/2 21/2 (1920)
 
E 
 H
e3 23 (21)  K I .
350 Appendix A. The Four Minors
e4 24 (22)   K  H 
e5 25 (23) 
 ! ,
 a  N:
e6 26 (24) y Kg A:
e7 27 (25) E b d q Y.
e8 28 (256)  b d  K I


e9/10 29/30 (27) 

a e  K .

Although some aeolic phrases are mixed in here and there, the basic metre
of Olympian 4 is freer D/e, both in strophe and epode. It is remarkably
similar to N6. As there, double-short phrases are extensively employed.
There are 10 ds and 5 Ds; D is a somewhat infrequent phrase in the eighteen
majors; for its use, see Part I, 8. C. 1. In three verses (s1, s2, s3) two ds (or D)
are juxtaposed without intervening anceps ( . . .
ww

ww
. . . ). This
type of juxtaposition is rare in the eighteen majors: 4 examples of d D and
D d (all concentrated in N6; Part I, 8. C. 1), and 2 examples of d d
(8, C. 2). RSS is rather high: 45.4% in the strophe, 53.8% in the epode.

Strophe
The strophe is divided into two sections. Each section is articulated by long
single-short sequences following the core d:
s5/6 w
w
w
ww
w
w
w

k
s10 a
ww
w
w
w

k
The relation of the Wrst 6 verses of 1 is easily grasped from this chart:
s1a ww
w
ww

ww
ww
j s1b
w
w

k
s2a ww
ww

ww

j s2b w
ww
w

k
s3

ww

ww

j s4





j
Double-short movement occurs in alternate verses. These three verses (s1a,
s2a, s3), each of which has two d (or D) in the centre, are gradually
shortened. The other three verses (s1b, s2b, s4) resemble each other. s1b
and s4 each consist of six positions, but in s1b two of them are short, while
in s4 all are long. The last Wve positions of s1b are the same as those of s2b
(and s5/6). It must be remembered that the verse-ending . . .
w

k is
remarkably rare in the eighteen majors; see below, Individual verses.
So far, all the phrases employed are symmetrical except for the hagesi-
chorean in s2b. An asymmetrical phrase, the reversed dodrans, appears
in s8a of 2 and is repeated twice (for s7, which is analysed as heptasyllable
anceps rdod in Snell and Turyn, see Textual problems). All the other
verses, however, are symmetrical. There are similarities with the verses of 1:
s8b (
e2) is the inverted form of s1b (e2
), and s9 (
d
) is, in a way,
 Olympian Four 351
included in s2 and s3. The Wnal verse (s10) is, as described above, a shorter
version of the Wnal of 1 (s5/6). At the same time it includes s8b and s9:
s9 a
ww

j
s8b

w
w
k
s10 a
ww
w
w
w

k

Epode
Most verses are freer D/e. Four (e3, e4, e6, e7) can be described in the normal
D/e manner. D is more frequent than d, unlike in the eighteen majors. Also,
long single-short movements are conspicuous (e3, e5, e8, e9). The Wnal
verse, e9/10, is very similar to the central verse of the strophe, s5/6:
s5/6 w
w
w
ww
w
w
w

k wewdwe2

e9/10 rwr w
ww
w
w
w
k ewdwe2

The similarity between s2b and e1/2 is also noteworthy. The rare phrase
hagesichorean (w
ww
w

) is repeated.

Textual problems
9 ( 8 Sn.; s9). Most editors accept P
 (Byz.) to regularize
the initial position. The initial position is anceps, and the paradosis
( .) goes well. However, the statistics may favour the regularization; see
below, s9.
10 ( 9 Sn.; s10). Boeckh, Schneidewin, Mommsen, and Bergk24 all
adopt  (Byz.), instead of , to get exact responsion. They also move
 and the corresponding word, , to the preceding verse (s9 in the
chart above):
[s9]

ww

w
j
d e
[s10] ww
w
w
w

k

Christ Wrst restores  and establishes the colometry which I adopt. He is
followed by later editors, including Turyn and Snell.
17 ( 16 Sn.; s7). All modern editors since Boeckh accept the paradosis
 . Their colometry is

w
ww

w
(hepta e). That
supposes correption in the corresponding line in the strophe: v.7 ( 6 Sn.)

! , n . Correption in non-dactylic movement is, in general, disfavoured.
Our case is especially suspect because it involves
E. Rather we should scan

E, n as
w, and introduce  (scholia, recc.) as Heynes text and
Hermanns colometry in Notae.
352 Appendix A. The Four Minors

Individual verses
s1a. The Wrst half, ^ dwd is paralleled:
O1e5 ww
w
ww

w
w
k ^d wd e 2
As for the latter, D,
N6s5 ww

ww
ww
j ^d D
is similar.
s1b. The phrase e2
is, in appearance, identical with the ithyphallic in
drama. Interestingly, there is no example of e2
in the eighteen majors,
whereas the reverse form
e2 is attested; see below on s8b. Outside the
eighteen majors it is repeated in as many as Wve verses of O5. See ad loc.
s2a. There is no exact parallel except for N6s5, cited above ad s1, which is
structurally close (d and D are interchanged). ^ D is used in O10s1 and O10e8.
s2b. Hagesichorean is a rare phrase in the eighteen majors (and so are
other phrases with 3 endings; i.e. aristophanean and hipponactean). There
is only one secure example: P2s8. However, s2b is similar to s10, in spite of
their diVerent names:
s2b w
ww
w

k
s10 a
ww
w
w
w

k
s3. d d are incorporated in O10e4/5 and P8e5.
s4. Three consecutive spondees are unparalleled. The nearest case is
O9e5
ww




j d sp sp
s5/6. The closest parallel is O9e1/2, as seen in the general introduction to
this Appendix.
s7. See Textual problems. The triple e combined continuously without link
has many parallels. See note on P5s9.
s8a. The same verse is used with resolution at the same position at
P5e3
w
ww
rw
ww
j rdod rdod
s8b. The same verse is found at O10e7 and I7s5b.
s9. At v. 9, the choice between P
 (Byz.) and 

(the paradosis) is not easy (and there must have been no diVerence in the
alphabet of Pindars age). The position is anceps, but in the eighteen majors
exact responsion is a strong tendency when d is preceded by anceps at the
beginning of a verse. There is only one exception (P10e01). And the anceps is
long more often (5 examples in the eighteen majors) than short (2 ex-
amples). See Part I, 6. C. The same pattern appears in
O9s9

ww

k
 Olympian Four 353
although this should certainly be analysed as reizianum from the metrical
context.
s10. Although single-short movement extends leftwards from wd in nine
verses (Part I, 8. C. 3), extension rightwards is rare.
e1/2. The palindromic sequence of
ww
w
w
ww
is incorporated
in four verses of the eighteen majors; see Part I, 8. A. 6 (b). Of these,
N3s1

ww
w
w
ww

w
k tel wde
is the most similar. For the hagesichorean, see s2b above. The uniWcation of
e1 e2 is now supported by Barrett, who avoids SVE at the end of e1.
e3. This verse is equivalent to the iambic dimeter. Surprisingly, there is no
example in the eighteen majors.
e4. D e without intervening anceps is rare at the end of a verse. It is
paralleled in the eighteen majors by
N6e1
ww

ww
ww
www
j d D e
P2s3

ww
ww
ww

w
k
D e
It is rare even in the normal D/e odes (5 examples).
e5. Two verses may be cited as remote parallels:
P10e4 x
w
ww

w
k hepta e
P8s6 x
w
ww

x
w
w
k hepta e3(aeol)

e67. Brevis in longo must be assumed at the syntactical break. Both D

and
D by themselves are familiar verses in normal D/e, but in the eighteen
majors there is no exact parallel. e6 ends with a short vowel (SVE). To avoid
it, Barrett, Collected Papers, 185 proposes a diVerent colometry:
y Kg A E b d q Y:  b d 

ww
ww
w

ww
ww
w

w
w
j ibyc ( dodd) ibyc e2
Ibycean suits this Class II ode. But two ibyceans without diaeresis are not
paralleled in the eighteen majors nor in Ibycus 286 P. This is highly unusual,
I presume, in any other Greek poems and should not be adopted even for the
sake of avoidance of SVE.
e8. e3wd is common, but the initial ^ e is rather strange in this context. The
nearest case is:
P10s5 w

w
w
ww
w

k ^e wil 3

The parallels for e3wd are:

O1s7
w
w
w
w
ww

w
k e3wd e
N3s5

w
w
wr w
ww
k
e3wd
354 Appendix A. The Four Minors
e9/10. If the rule is strictly observed, the Wrst half should not be notated as
ewd but as wilamowitzianum. There is no example of e resolved totally in the
eighteen majors. Outside the eighteen majors, there is one verse whose Wrst
half is the same:
Pae6s5a rwr w
ww
w rw
j e w d w e
 Olympian Five 355

Olympian Five
Three triads. Class II (?)
O5s1


ww

ww
w

w
j sp d (? rdod) dod e
O5s2


ww
ww
ww

w
w

k sp D (? rdod2d)
e2

O5s3 ww
ww
w

w

w
w

k tel e e2

O5e1


ww
ww
ww
w
w
w

k sp D (? rdod2d)
we2

O5e2


jww
ww
ww

w

w

w
w

k sp D (? rdod2d)
e e e2

s2 H 2, B 13, 18; s3 HB 3; e1 H 23

For the diVerent colometry of e2 adopted e.g. by Turyn, see the Wnal
paragraph.
s1 1 (1) ! ,A IA d 
  
s2 2 (2) H P
 , *F ,  E
s3 3 (3) I
  I
  2  H
e1 7 (7)

 &  
 : d b F 
e2 8 (8) 
 I, d n
 @ K
 d
a  .

The Scholia record the suggestion that Olympian 5 may be a spurious

work (K b K P q), though its meaning is not necessarily clear
(K
 is a strange word).1 Besides, there are so many metrical features
uncommon in the other Pindaric epinikia that its authenticity is questioned
by many; cf. Boeckh i. 372: Metrum plane est eximium, quamquam a ceteris
Pindari carminibus mirum quantum distans.
First of all, both the strophe and the epode of O5 are extremely short.
Though each individual verse is long and the triad is repeated as many as
three times, the number of verses is only 3 (strophe) and 2 (epode),
extremely few and far from the average of the eighteen majors. Parallels
are rather found outside the Pindaric epinikia: Partheneion 1, 2 (see Ap-
pendix B) and Bacchylides 3 (see Excursus to O13).
Secondly, the structures of the two stanza-forms are unusually simple and
clear. All the verses both in strophe and in epode are almost identical. Three
of them (s2, e1, e2) start with sp D , and the other two also start in a
similar fashion. All end with e2
(apparently ithyphallic) except s1. Even
the ending of s1, e, can be explained as a variation. e is repeatedly inserted in

1
A stimulating explanation is now given by Barrett, Collected Papers, 47.
356 Appendix A. The Four Minors
e2. Such simplicity and clarity is totally absent from the non-D/e odes of
Pindar, except for partheneia.
Thirdly, both these key phrases, namely sp D and e2
, are unfamil-
iar in the eighteen majors. The former could be analysed in two other ways:
(1) expanded aeolic ( rdod2d); (2) dactylic with its Wrst dactyl replaced by
spondee. Neither is likely.
1. Interpretation apart, the expanded aeolic is rare (six in total) in the
eighteen majors (Part I, 7. 6). A seeming rdod2d is once used, but it starts
quite diVerently:
N6s3 rw
ww
rw
ww
ww
ww
j rdod ribyc ( rdod2d)
More peculiar than expansion is the initial pair of longs, which, together with the
following
ww
, could be seen as making up a reversed dodrans. In the
eighteen majors the second position of the reversed dodrans is short, at least in
some, if not in all repetitions:


ww
at all the repetitions is unparalleled
(see Part I, 5. A. 1). Another explanation might be that the Wrst two positions
are aeolic base; s1 does not include expansion but


ww
is followed by

ww
w
. The combination of these two is equal to the asclepiad of Lesbian
metre, and has been taken thus by many scholars. But this cannot be so either,
because (i) there is no certain example of
ww
, the doubly
truncated glyconic (Part I, 5. A. 4) (ii) the asclepiad is never used in the eighteen
majors. The interpretation of s2 as beginning with rdod2d should be rejected.
2. The whole of


ww
ww
ww
cannot be dactylic either. Con-
traction of two shorts of a dactyl is totally unparalleled in freer D/e as well as
in the normal D/e. It would be an Aeschylean-type verse if it were dactylic.
If this phrase is analysed in the same manner as those in other Pindaric
verses, its initial two longs must be an independent spondee. Spondee is
used three times in the normal D/e odes. It comes at the beginning of the
verse in P1s3 (but it is followed not by a D-type colon but by e
e). Since the
two longs of s2, e1, and e2 are spondees, it is rational to take rdod in s1 as
sp d. Its reversed form, d sp, is used in I7e6 (and d sp sp, in
O9e5).2 In the Paeans of Pindar there are some examples of spondee
preceding dactylic sequences; for example,
Pae4s5


ww
ww
ww
ww
w
j sp ibyc
These are discussed in Appendix B.
Like sp D , e2
(apparently ithyphallic) is not familiar either, in the
eighteen majors. The ithyphallic is a very common colon outside Pindar, but
contrary to expectation, there is no example at all in his poems, although
2
If we adopt Barretts correction, the colometry of I7e6 will be diVerent; see ad loc.
 Olympian Five 357
we2
(apparently equal to iambic bacchiac) at e1 is paralleled (O2e6,
N3s8, O9e1/2).
At the same time, however, it should be pointed out that there are
some similarities to the eighteen majors. The appendage e following a
dodrans (s1) is very common, and can, indeed, be regarded as one of the
characteristics of Pindaric metre. The combination of dod e is also
included in s3 in the form of tel e. The double es in e2 are paralleled;
see Part I, 8. C. 5.
The Wnal verse of the strophe, s3, starts with the telesillean with half-base
in the form ww, which is common (6 examples, including tel 3). The
following verse is especially similar:
N3e4 ww
ww
w
ww
ww
w

w
k tel tel e
This type of telesillean may also be compared with ww
ww
w
x, the
phrase which is used in one of the normal D/e odes: N10s1; see Part I, 7. 2. It
is related to xD.
It is interesting that the combination of e e2
(cretic ithyphallic) is
used, outside epinikia, in the three verses of Parth1 (s3, e1, e2). They are
similar to the verses in O5, and, especially, the similarity between Parth1s3
and O5s3 is striking:
Parth1s3 ww
ww
ww
ww

w

w
w

k ^ D e e2

O5e2


ww
ww
ww

w

w

w
w

k sp D e e e2

As far as we believe in the reconstruction of Partheneion 1 and its ascription

to Pindar, we should not deny the authorship of O5 outright, at least on
metrical grounds.
Turyn, following Schroeder, divides e2 into two verses:
e2a


ww
w
k gl
e2b
ww

w

w

w
w

k d e e e2

This colometry assumes verse-end at these word-boundaries: v.8  ,
v.16  s, v.24  P. According to these divisions, we
encounter hiatus (vv.8, 24) and brevis (v. 16).  at the verse-end is not
itself a problem. It is paralleled. But here in this metrical context, i.e. the
dactylic movement, epic correption is acceptable.3 If  is scanned as a
short syllable, the dactylic movement is carried on further, and metrical

3
Barrett, Collected Papers, 51 writes: Pindar invariably observes the digamma in
the pronoun , , and its adjective . [ . . . ] It looks to me as if we have got
correption of d before ; and if we have, the ode is not by Pindar. Incidentally,
the examples of the adjective are scarce.
358 Appendix A. The Four Minors
consistency is acquired. The combination of e2a and e2b as one verse is
preferable.

Textual problems
16b (e2). The paradosis gives an improbable hiatus: s b  . Both Snell
and Turyn adopt Boeckhs emendation: s b  . Hermann suggests M0
  , and is followed by Dissen, Mommsen, and Christ.
18 (s2). Christ keeps the paradosis,  "E and scans - " - as ww
( " E , non Cretense, sed Olympicum). Schroeder, Snell, and
Turyn follow his scansion. Hiatus should be acknowledged as a licence for
a proper noun in dactylic movement. Boeckh and others write 
"E by scanning the initial "- as a long vowel. However, this scansion
requires responsion between a long and double short, and should certainly
be rejected.
 Olympian Fourteen 359

Olympian Fourteen
Two strophes (monostrophic). Class III
O14s1

w
ww
j hepta (?
rdod)
O14s2 w
w
w
ww
w
w

j wewdwe

O14s3
ww
w
w
ww

k dod reiz (? dodwd
)
O14s4 www
ww
w
w
ww
w
w
j glwdwe
O14s5
ww
w

w

w
w
k dod e e2
O14s6
ww
w

ww
k dod d
O14s7
ww
ww
w
ww

k Dwd

O14s8
w
ww

ww
w
j rdod dod
O14s9
ww
ww

w
w
w

ww
k D e3d
O14s10

w
w

w

jww
w
k
e2e dod
O14s11
yjw


jw
j
e
e
O14s12
ww
w
w
ww
w
wj

k dod tel 3 (? dodwdwe
)
s3 HB 15; s5 H17; s6 B18; s7 B 19;4 s9 H9; s10 B 10

s1 1 (1)  $

s2 2 (2) E   
 ,
s3 3 (3) t
A I 
s4 4 (4) %
 F,
 A K

,
s5 5 (5) F , K
d h f a $E 

a 
s6 6 (6) a  

 E ,
s7 7 (7) N  , N  , Y  Ie I.
s8 8 (8) Pb a d A % 
s9 9 (9)  f h E Ia

 
s10 10 (10)  K PH,  

s11 11 (11)  
  ,

s12 12 (12) N 
e 
 
.

The stanza is repeated, like strophe and antistrophe in tragedy. This

structure escaped the notice of the metrical scholiast and Byzantine scholars.
They treated the whole as one long astrophon. There had been some
attempts to search for strophic responsion, but in Heynes edition published
in 17989, Hermann took the Wrst step worthy of consideration to recover it
(De metris Pindari, 259). His analysis was republished in almost the same
form in his Notae. There are a number of textual uncertainties which
conceal the strophic responsion (see below, Textual problems), but respon-
sion between the two strophes is undeniably established. The total number
of positions in each strophe amounts to 158. This Wgure is very high; higher
than in most of the stanza-forms of the eighteen majors, except for I8s (169),
360 Appendix A. The Four Minors
which is also monostrophic. RSS is 48.7%. Resolution is conspicuously rare;
it occurs only in s4 and s11.
Olympian 14 is a short poem, but its metre is sophisticated. The stanza-form
is unambiguously classiWed in Class III. Perhaps it represents the most devel-
oped stage of Class III metre, comparable to O1 or P2. Asymmetrical phrases,

ww
w
(dod) and
w
ww
(rdod), are used as if they were variations
of symmetrical phrases like
ww
ww
(D),
ww
(d) and
w
w

(e2). The similarity of dodrans to e2 is exploited in N7s, but its resemblance
to D on such a scale is unparalleled. Aeolic phrases appear to be multifarious
at Wrst sight, but are in fact rather simple, and the development from verse to
verse is clear. The whole structure is well articulated in the chart below:

(single-short/anceps) (dodrans/D/e2 ) (single-) (dodrans/D/e2)

s1

w
ww

s2 w
w
w
ww
w
w

j
s3
ww
w
w
ww

k
s4 www
ww
w
w
ww
w
w
j
s5
ww
w

w

w
w
k
s6
ww
w

ww
k
s7
ww
ww
w
ww

k
s8
w
ww

ww
w
j
s9
ww
ww

w
w
w

ww
j
s10

w
w

w

ww
w
k
s11

w


w

s12
ww
w
w
ww
w
w

j

Dodrans is the principal phrase, used as many as ten times. Its place is taken
twice by D (s7, s9), four times by d (s3, s6, s7, s9), and twice by e2 (s5, s10).
Reversed dodrans too is used twice (s1, s8). With the exception of s1, s2, and
s11, each verse comprises two parts, each of which contains one of the above
phrases; it is noteworthy that dodrans and D never occur in the same verse.
The pairings are two dodrantes (e.g. s4), dodrans and d (e.g. s6), dodrans
and e2 (e.g. s5), and D and d (e.g. s7). This accords with the general
observation that there is no verse in which both dactylic movement and
an aeolic phrase appear together. The principal phrase may be immediately
preceded by a single-short one: e before the dodrans in s5 and s10, e3 before
d in s9. In s11, single-short movement (
e
e) alone makes a verse without
the dodrans. Dodrans is prolonged by single-short movement leftwards (s2,
s4) and rightwards (s2, s4, s12). D and d are connected by short link anceps
(s7), which also stands before or after dodrans (s3, s4, s12). Long anceps
is rare, especially in mid-verse. The sole example is found in s11, where
one word (
 11, P 22) occupies the consecutive three long
 Olympian Fourteen 361
syllables. There are no other words of this shape, so that the whole poem
gives a peculiar impression of lightness.
Another outstanding feature of this ode is its variety of palindromes.
Seven out of twelve verses contain palindromic movement. These are the
examples, with their parallels, if any:
s2
w
w
ww
w
w
(1)
s3
ww
w
w
ww
(2)
s4 www
ww
w
w
ww
w
none
s7
ww
w
ww
(3)
s8
w
ww

ww
w
(4)
s9
ww

w
w
w

ww
j none
s12
ww
w
w
ww
s3 (2)
(1) O9e1/2, O4s5/6
(2) O1e4, N3s1, N3e3, N4s4
(3) P2s4, N3e1, N3e4, N4s6
(4) none, but its reversed form,
ww
w

w
ww
, is found in 7 verses.

Textual problems
2 (s2). The third position of s2 must be short. Boeckhs emendation,
E for E, is still worth considering (verisimiliter Christ); it
expels hiatus between E and . Snell, following Mommsen, accepts
the paradosis and supposes correption: E ; but correption in
single-short movement is improbable. Turyn too keeps the paradosis
E, but adopts  (Bergk2) to avoids hiatus. Then the position in
question becomes anceps, and s2 cannot be analysed as a separate verse from
s1. Metrically, separation is more attractive (the united verse of Turyn will be
discussed below). s2 is well paralleled by O9e1/2 and O4s5/6. All three
belong to the group of long sequences of single-short movement turning
to d (9 verses in total in the eighteen majors, see Part I. 8. C. 3).

O14s2 w
w
w
ww
w
w

j wewdwe

O9e1/2 w
w
w
w
ww
w
w
w

k we2wdwe2

O4s5/6 w
w
w
ww
w
w
w

k wewdwe2

When s2 is united with s1, as in Turyns colometry, the phrase boundary
must be established before the anceps:


w
ww
w
u
w
ww
w
w

j hepta 2 hepta 2 3
This analysis is acceptable, but not altogether a good one. The former
phrase (hepta 2) is well paralleled but the latter is not (cf. P8e7, which is
quite unique and its analysis is uncertain). Rather, the diYculty lies in the
362 Appendix A. The Four Minors
combination of the phrases. There is no example in the eighteen majors of an
aeolic phrase of 2 ending followed by another aeolic starting with x
w.
5, 17 (s5). Strophe and antistrophe do not respond:
F , K
d h f a $E 

a  


a (most MSS); 


 (C)
F H  a `
 K 

5
ww
w

w

w
w(w)

17
ww
w
w
w

ww
w

The text has been variously emended both in the strophe and the antistrophe,
according to various metrical schemes; my analysis above is based on (vi).

(i)
ww
w

w

ww
w
k dod e dod

(ii)
ww
w
w


ww
w
k dwe tel
(iii)
ww
w
w
w

ww
w
k dodwe dod
(iv)
ww
w
w

ww
w
k dwe dod
(v)
ww
w
w

k u
ww
w
k dwe
j tel
(vi)
ww
w

w

w
w
k dod e e2

(i) Hermann (Notae), Bergk2, Christ, Turyn, Snell

F , K
d h f a h 
hi 
a 
F HH a 
 K 

(ii) Boeckh
F , K
d h f h a 
hi 
a 
F H   
 K 

(iii) Hermann (Quinque Ol.)

F , K
d P f a h 
hi 
a 
F H  a `
 K 

(iv) Mommsen
F , K
d h f fag h 
hi 
a 
F H H fag `
 K 

(v) Bergk4
F , K
d h f fag h j 


   j
I . . .
F H  a j 
 K 
j
fKg   I-
4
If  is adopted. Turyn reports . Boeckh prints , which
is now given strong support by Barrett, Collected Papers, 187.
 Olympian Fourteen 363
(vi) Schroeder, Bowra
F , K
d h f a h 

a 
F H: H a 
 fKg 

Metrically, (ii) and (iv) are less likely. The long anceps of the telesillean does
not suit Class III (ii); and the correption of H before ` is implausible (iv). At
Wrst sight (v) is too drastic to accept. But, although this colometry is totally
diVerent from the others, the textual changes are not many, except
I in the following verse. The  after 
a is the reading of C.
Even I should not be dismissed at once, for exact responsion is
lost in the following verse too (see below; Bergk does not adopt Kaysers
 for  at v. 6). However, deletion of a (iv)(v) may not be
syntactically preferable. The choice would be either (i) or (vi). Although
there are no strict metrical parallels to (i) or (vi), both are possible com-
binations of phrases. I adopt (vi); see further individual verses.
6, 18 (s6). Turyn retains the paradosis with the following analysis:
a  

 E , wwr
w

ww
k
K   I ;
ww
w

ww
k
However, the resolution of the choriamb (or choriambic nucleus of
dodrans if phrase boundary is changed) is rare; emendation is necessary.
Hermann transposes . Boeckh, Mommsen, et al. follow him:
a  

 E ,
wr
w

ww
k e e d
K hi  f g I ,

I prefer Kaysers emendation a    K   I-), like

Christ, Schroeder, Snell. This gives

ww
w

ww
k dod d 5
8, 20 (s8). Again, exact responsion is lost:6
Pb a d A % 
w
w


ww
w

F :  F 
w
ww

ww
w

Neither the metrical form in the strophe nor that in the antistrophe is
exactly paralleled in the eighteen majors. From the metrical point of view,
the latter is slightly preferable (rdod dod). This combination of two aeolic

5
According to Bergk4 (see above, s5, (v)), s6 becomes
ww
w

ww
k ^ dod d
^ dod is found in O13s5.
6
Schroeder (BT) accepts the irregular resposion. According to his theory, both

w
w

and
w
ww
belong to the same metre.
364 Appendix A. The Four Minors
phrases harmonizes well with the other verses; see the horizontally arranged
chart above. The diYculty of the former (e2 tel) lies in the fact that e2
rarely stands at the beginning of a verse in the eighteen majors. Also rare is
the long half-base in combinations like xe tel.
Various emendations have been proposed to recover the exact responsion.
Some scholars emend v. 20: thus Boeckh introduces . Respon-
sion is seemingly recovered, and Mommsen adopts it. But resolution before
half-base,
w
wr a
ww
w
, is hardly acceptable; instead, Maas
(Nachlese) proposes , which is accepted by Turyn.7
Emendation should be rather looked for in v. 8, but a convincing proposal
has not yet been made. Snell, following Wilamowitz (GV), scans d
monosyllabically and supposes responsion
w
y
. Contraction of
the choriambic nucleus is the least probable option, and should be rejected.
Kayser proposes A instead of A. Christ follows him. Exact respon-
sion is restored (
w
ww
), but correption in an aeolic phrase (d
A ww

) is doubtful. Bergk4, having found a rare word in Hesychius
(  ,  , P ), reads Pb a A   .8
9 (9). See below, on v. 21.
11 (s11). See below, on v. 21.
13 (s1). t is Boeckhs supplement (om. MSS).
15 (s3). K

 F (MSS) is one syllable too short. Bergk2 changes the
adjective into a verb: K
E F. Both Snell and Turyn adopt this.
Hermann retains the adjective. After two proposals, K

  (De
metris; adopted by Boeckh) and K

 F (Notae), his Wnal one is
K

 e F (Quinque Ol.), which is adopted by Mommsen.
21 (s9). Snell accepts the paradosis. It requires responsion between two
shorts and a long in D, which is improbable.  (Ahrens, followed by
Christ) seems to be a simple correction, but this conjugated form of X
is not otherwise attested. Y (Byz., adopted by Turyn), too, is simple, and
may be right instead of hiatus. The dactylic rhythm might admit it.9 On the

7
Bergk2 introduces E in place of F and changes the word-order: Pb d a
in the corresponding v. 8. This makes
ww
y
, which is improbable.
8
Hermann (Quinque Ol.) is drastic: he not only emends v. 8 but changes the
colometry by moving F  to the preceding verse:
Pb A j  % d
w

kw

ww
w

F : j  F 
w

kw

ww
w

The second phrase, ^ e dod, does not suit the context well; there is no example of ^ e
in this song.
9
Turyn supposes the digamma before E. I am incapable of judging whether his
citation from epigraphic material is right.
 Olympian Fourteen 365
other hand Boeckh emends the strophe (v. 9), introducing , and
retains  here (the hiatus remains). The metre then becomes
ww


w
(d e). That is possible, and Mommsen accepts it.
23 (s11). The paradosis P is unmetrical, so P (Boeckh), or
P (Bergk). The responsion between  in v. 11 and 

(FEN) in v. 23 involves resolution, which is possible. To produce the
exact responsion, Ahrens accepts 
 (the other MSS) and introduces
H in the strophe (v. 11):


ww


w
j
d
e
This may be right, because resolution is in general avoided in this ode. He is
followed by Christ.

Individual verses
s4. The sequence wdwe is unusual; it should perhaps be reanalysed as tel 2
( wdod w
), as is demonstrated in the table above by which I analyse
the structure as a whole. There is only one parallel for wdwe, in which it is
preceded further by we3:
N3e1b w rw
w
wr w
ww
w
w

w
k w e3 w d w e e
s5.  at verse-end is paralleled. There is no example of dod e, but its
correlative, tel e, is one of Pindars favourites. e e2 is used in three
verses in O2. The three phrases, dod, e, e2 are all included in O1e1, but in a
diVerent order:
O1e1 w
wr
w

ww
w

w
w
k w e e dod e2
The text of Turyn and Snell (see above, Textual problems) yields a peculiar
colometry since e is never used between two aeolic cola in the eighteen
majors.
s6. d is not a common suYx of aeolic phrases; there are only two parallels
(I8s6, P5e9).
s7. There is no exact parallel, but the following verses are similar:
N6s6
ww
ww


ww
rw

k D
d e

N2s5
ww
x
ww

k d x d

N3e1
ww
w
ww
x
w

j d w d x e

s9. Three D/e phrases are connected without link anceps. One expects to
Wnd parallels in normal D/e, but connection without link is fairly rare. There
are six examples of D followed by e
e (not by e3): P3s6, P4e7, P9s7, N1s7,
N8e1, I3e6, and three of d preceded by e (not by e3): O7e7, N1e4, I5e4.
366 Appendix A. The Four Minors
s10. There is one verse of similar structure in the eighteen majors:
I8s1 x
w
w
rw

w
ww
k x e2 e rdod
s11. This verse is equal to the iambic dimeter, of which, however, there is
no example in the eighteen majors (there are many in the D/e odes, of
course).
 Pythian Seven 367

Pythian Seven
1 triad. Class II (strophe/antistrophe); Class II (or Class III; epode)

P7s1

w
wrwrjw

k
e4

P7s2 w
wwj

w
u
w
jww
j wd e hepta
P7s3

w

j
e

P7s4

wjw

k
d

P7s5 w
wr
ww

j we d

P7s6

jww
w
k tel
P7s7 wr
w
k ^e e
P7s8 u rw

k xe

P7e1
w
wwj

w
ww
j rdod rdod
P7e2 w


w
j

wjw
w
j ^ e
e tel
P7e3a rw
ww
j rdod
P7e3b rw
wj
ww
j ewd
P7e4

w
w
ww

ewd
P7e5

w

ww
j
e d
P7e6 w
ww

k wd

s1 H 9; s3/4 H 4; B 12; s5/6 H 6; s7 B 7

s1 1 (1)
 
 `A
s2 2 (23)
 `A PE  A
s3 3 (3/4)
E IA
s4 4

 .
s5 5 (5/6) K
d 

,  r 
s6 6  O

s7 7 (7) K

s8 8 (8) !

;
e1 17 (17) t 
 , $  d
.
e2 18 (?)   P
   e  ,
e3a 19a (?)  I
e3b 19b (?) a a :   

e4 20 (20) o  Id

e5 21 (21)
 P
e6 22 a d a .

Strophe
The colometry of the strophe is uncertain. Coincidence of word-end is
frequent. There are also some textual uncertainties. Notably, Snells colo-
metry (taken from Wilamowitz, GV) is unique, and diVers from all its
368 Appendix A. The Four Minors
predecessors. The colometry above is Boeckhs. It follows, more or less,
semantic divisions, and is adopted by most editors either without change
(Schneidewin, Mommsen, Christ, Schroeder) or with a minor change
(Bowra unites s3 with s4, and Turyn divides s2 into two after the fourth
position). The exceptions are Snell, as has been mentioned, and Bergk24,
who deletes the two initial longs of s6 and unites s6 with s5 (see below,
Textual problems, on v. 6). Snell is no less radical. He not only changes the
metre at the junction between s2 and s3 (see below, Textual problems, v. 2)
but unites s3 with s4, and s5 with s6.
If s3 is united with s4 (Bowra and Snell), and s5 with s6 (Snell), two aeolic
phrases starting with heavy full base (

) emerge (with dovetailing):
P7s3/4

w

j

ww

k
e ph
P7s5/6 w
wr
ww

j

ww
w
k we d gl
This aeolic base is, however, peculiar to Class I (aeolic), but there are no
other Class I features elsewhere in this stanza-form. When separated from
each other, the four verses, s36, become short, but their shortness can be
accepted. It is paralleled by s7 and s8, between which verse-end is demon-
strated by brevis in longo at v. 7.
This stanza-form is, as a whole, similar to P5s and belongs to Class II (freer
D/e). As in P5s, e is dominant while d appears sporadically. Asymmetrical
phrases are occasionally mixed in: the heptasyllable, which may be analysed as
link rdod, at s2, and telesillean, link dod, at s6. It is remarkable that all the
verses start with anceps or acephalous e. None starts with a true long.

Epode
The colometry of the epode is more uncertain. I start from the hypothesis
that it is not very diVerent from the strophe (and P5). My colometry is
basically the same as Christs and Schroeders. Before Christ (Boeckh,
Schneidewin, Mommsen) e3b was divided into two, its Wrst half (rw
w,
a a ) united with the preceding verse, and its second half (
ww
,
  
), with the following verse. The analysis above is considerably
diVerent from those of Turyn and Snell. Following Wilamowitz, they expel
 (MSS) at v. 20 (e4) and read . They then analyse:
P7e3b www
w
ww
j wil
P7e4


w
ww
j wil
P7e5

w

ww
j wil (Turyn), ia ch (Snell)
Snells nomenclature is right here. With my analysis, the two wilamowitziana
disappear. It is common in Class II or Class III (for example, P5e or I8s) that
reversed dodrans is repeatedly used amongst es.
 Pythian Seven 369

Textual problems
2 (beginning of 3 Sn.; s2). Snell scans  A as a disyllable and moves it to the
begining of the following verse. In his corresponding verse 1011 Snell takes
oV a syllable to make responsion, by establishing   (V) where
most editors adopt   , Moschopoulos emendation of the other
MSS   . According to Snells colometry (but not his notations),
the verses may be analysed thus:
[s2] w
wwj

w
u
w
j wd e x e
[s3] w
j

w

j

wjw

k ^ e
e ph
[s3] is similar to e3a. The disyllabic scansion of  A and the expulsion of 
are already suggested by Mommsen in his apparatus, though his colometry
is diVerent:
s2 w
wwj

w
u
w
jw
j wd e x e2
4 (s4).

 (MSS) is unmetrical;

 Triclinius.
5 (s6).  (=)E (Boeckh) is accepted by many. The paradosis is 
r (unmetrical) or   r .
6 (s6).  is dubious. Scholiasts try laboriously to explain it. Proposed
emendations are various: N (Kayser, followed by Christ, and recently,
Liberman), H (or H, Schneidewin), NA (i.e.
; Mommsen). If
a word starting with a vowel like N is selected, brevis in longo occurs at
the end of the preceding verse. Combining s6 with s5 is more unlikely in this
case because Pindar hardly ever uses u
at the opening of glyconic. Bergk2
deletes it, and the corresponding E (14) too, as a gloss.
6 (s6). The paradosis (O
=O
=O
) is one syllable too
short. Boeckh, Schneidewin, Mommsen, Christ et al. adopt O

(Triclinius). O
 is Boeckhs suggestion in not. crit. It is adopted by
Schroeder, Turyn, and Snell.
9 (s1). The paradosis (
  =
  =
 
 ) is unmetrical. Many adopt
  (Byz.), except Turyn
(
  ).

Individual verses
s1. In the iambic context, this verse would be a catelectic trimeter (ia ia
ba); but the bacchiac as the catalectic form of the iambic metron is alien
to Pindaric metre. Except in the initial and the Wnal position, all the
intervening brevia are short. Thus, like e5 and e6 in the eighteen majors,
370 Appendix A. The Four Minors
this verse should be taken as one of the long sequences of single-short
movement (en), preceded and followed by ancipitia, even though e4 is not
paralleled in the eighteen majors. Frequent resolution is a characteristic
common to the examples of long en (O1s8, N3s2, P2s1). Our verse includes
seven successive shorts, as many as O1s8. It is localized by 

A (and
  E), which reminds us of the beginning of
P2s1: 
 t #
.
s2. The combination of xd e is frequent:
N7s5 w
ww

w
wrw
k wd e3
O9s11

ww

w
x
ww

k
d e x d

For e hepta
O9e8

w
u
w
ww
u
w
ww
w

k
e hepta hepta 3
is a remote parallel.
s3. There are two verses made up only of xe
: O10e1b and P5e7b.
s4. No example is found of xd
as an independent verse (O9s9 may be
one, though I analyse it as reizianum).
s5. Very similar is
O10e1 w
wr
ww
j we d
s6. A single telesillean is found as an independent verse in Class II stanza-
forms (O10e6, P5s7b).
s78. Exactly the same combination of short verses at
P5e7 w
tw
j ^ e e
P5e7b w tw

j we

e1. There are two other verses made up of two rdod: I8s5b, P5e3.
e2. Acephalous e is hardly ever followed by long anceps. The unique
parallel in the eighteen majors is
P5s11 w


w
w
w
k ^e
e3 (see ad loc.)
It may be better to divide after P
 to make the telesillean an inde-
pendent verse like s6.
e3b4. x e w d
is paralleled at N3s6, P11e6.
 APPENDIX B

Longer Fragments and Prolonged Ibycean

It is much more diYcult to establish colometry in the fragments of Pindar

than in the small-scale epinikia, to say nothing of the eighteen majors. Of
the two main sources of fragments, citations by ancient authors are not
necessarily reliable in detail. They often change the word-order. Even the
omission of a particle could change a metrical scheme. Moreover, verses
are hardly ever repeated, for it is rare that two or more stanzas are cited.
On the other hand, the fragments recovered from papyri have diYculties
pertinent to papyri. First of all, one must be a papyrologist to read the
letters correctly. Furthermore, the restoration of strophic responsion from
scattered letters on more than one scattered column of a papyrus is itself
sometimes dependent on a supposed metrical scheme. The colometry
laid out in papyri is often mysterious. Whether it deserves to be treated as
a real colometry should be seriously examined. Even when strophic
responsion is conWrmed, the number of repetitions is generally small.
That means it is not easy to aYrm verse-end by hiatus/brevis. For the
identiWcation of hiatus/brevis, both the Wnal letter and the initial letter of
the following line must be preserved in the papyrus. We are not always so
lucky.
But, in spite of these diYculties, some fragmentary odes of Pindar can be
used for metrical studies. West chooses Paean 4 as a good specimen to show
generic development of verses in an aeolic ode (GM 67). Recently, Ruther-
ford has remarked that such sequences as gl@da [ gld] and gl@2da [ gl2d]
are rare in the Epinikia. They are, on the other hand, characteristic of the


 genre (p. 449). It is interesting and even desirable to investigate how
the major fragments are found to be related to the epinikia when the rules on
which my analysis is based (I. 3) are applied to them.
Some of the fragments are evidently D/e, and are consequently excluded
from further consideration here. Of the non-D/e poems, these six are long
enough to be analysed in detail: four paeans (Pae2 D2 Rutherford, Pae4
D4, Pae6 D6, Pae9 A1) and two partheneia (Parth 1, 2). It must be
stressed that my observation is limited to metrical analysis. For the reading
372 Appendix B. Longer Fragments
of papyri, I follow the report of the previous editions, especially Grenfell
Hunt, Lobel, SnellMaehler, Radt, and Rutherford.1
In general I do not specify how the colometry of the papyrus diVers from
modern analyses. How papyri lay out verses and whether they correctly
analyse metrical units are interesting questions about scholarship in an-
tiquity, but outside the scope of this book.2
In the Paeans, especially in Paean 4 and Paean 9, there are a number of
verses which seem to be created by so-called dactylic expansion of aeolic
phrases. The analysis of these is a big issue, and a special section is devoted to
them after the sections for each poem. At the end of this Appendix, I discuss
fr. 94c Snell. This is not a papyrus fragment but a citation by Hephaestion.
The name he gives to the metrical form in question is interesting: the
Pindaric hendecasyllable. I investigate two questions: one is what Hephaes-
tion understood by the Pindaric hendecasyllable; the other is what the
Pindaric hendecasyllable really is.

1
Snells layout is a compromise between the reproduction of the layout in the
papyrus and his own metrical analysis. He keeps the line-division of the papyrus to
print the text, but makes indentation where he judges the verse does not end.
Nevertheless, in his metrical chart he freely moves syllables to clarify his analysis.
Consequently, the line division in his chart does not accord with the line division of
the text. For example, the second verse is divided into two lines of 7 syll. (ph) 7
syll. (ia ba) in his metrical chart, while in the text it is divided into two of diVerent
lengths (6 8).
2
Cf. Itsumi, Whats in a Line?
 Paean Two 373

Paean Two
Three triads. Strophe: out of the six repetitions, some parts of str.1, str.2,
str.3 and ant.3 are preserved: s1a 2r;3 s1b 34r; s2 3r; s34 24r; s5 23r; s6
13r; s7 12r. Epode: survives better than the strophe: e1 13r; e2 3r; e3/4 2
3r; e5e9 3r.
Strophe: Class I. Epode: Class III
Pae2s1a
w
ww
j rdod
Pae2s1b u
w
w

j xe2

Pae2s2 u
a
w
j xe2(aeol)
Pae2s3 rw
ww
w

u
ww

k gl ph
Pae2s4

ww

a
ww

w
k
d rdod e
Pae2s5
w
ww
rw
ww

j rdod ph
Pae2s6

ww
w

w
ww

j tel ph
Pae2s7

ww

w
ww
w

j
d hipp
Pae2e1

ww

ww
w rw
w

k
d d we2

(or
d dodwe
)
Pae2e2
ww

w
ww

w
k d rdod e
Pae2e3/4




ww

ww
ww
ww
w
j sp sp d ibyc
Pae2e5 w rw
w

w
ww
rw

j we2 rdod e

Pae2e6 rw
ww

a
ww

w
k rdod rdod e
Pae2e7

w
ww
w

k hepta 3
Pae2e8
w
ww
w


ww

k gl reiz
Pae2e9 w
ww

w
ww

w
ww

k wd rdod ph

The colometry of the papyrus is obviously wrong in two verses (s4, e2). It
ignores the mark of verse-end (hiatus/brevis), and makes a continuous line
beyond the mark.
s4. Colon-end is placed after the double short of the reversed dodrans,
and its last long and the following e are united with s5, at all the correspond-
ing repetitions. The vertical stroke here indicates line-boundary, not
phrase-division:
41/42 ] [
51/52 . . .  . . .
77/78 . . . 
 . . .

3
r means the number of repetitions which suvive in the papyrus. Needless to say,
where the number is small, the analysis is uncertain. The number is uneven within a
verse because (i) surviving parts are diVerent according to repetitions, (ii) the
colometry of the papyrus is diVerent from mine.
374 Appendix B. Longer Fragments
The colometry of the papyrus ignores brevis in longo at 78 (
 P-),
rendering the division impossible.
e2. Colon-end is placed between the reversed dodrans and the following e,
and the e is united with e3, at all the corresponding repetitions:
25/26 . . . 
 
 . . .
61/62 . . .

 . . .
97/98 . . . 

 . . .

Hiatus at 26 (d P-) makes the division impossible.

The structure of the strophe is fairly simple. Except for s2, aeolic phrases
are repeated. Responsions between long and short both at aeolic full base
(s3) and at the second position of reversed dodrans (s4, e6) are character-
istics of Class I. So is dovetailing (s3). Occasionally
d (s4, s7) and e (s4)
are combined with aeolic phrases. Parallels are many in the eighteen majors.
s1b and s2 are irregular; see below.
The epode is more complicated. Reversed dodrans is frequent and freely
combined with d and e (e2, e5, e6, e9). Particularly unusual is e3/4, which is
the only verse that includes double-short movement.

Individual verses
s1ab Editors including SnellMaehler do not divide s1 into two halves.
However, word-end is regularly attested after the reversed dodrans (s1a).
The anceps at the beginning of s1b is long at two repetitions (1 @, 37
Id) and short at one (73 e). In the eighteen majors aeolic phrases
are hardly ever followed by long link anceps a freer D/e phrase. There are
only two exceptions:
P11s4
atww


wr
ww
k rdod
e d
in which the anceps is always long, but would be eliminated with an
alternative colometry, rdod sp rdod (see ad loc.);
P10s6 ww
ww
w
a
w
w
k tel x e2
in which the anceps is long at two repetitions out of six.
Thus it is better to separate s1a (rdod) from s1b. xe2
(ia ba in
drama) is not very common in Pindar, and when it occurs, the anceps is
always short in the eighteen majors (3 examples). we2
either functions as
a prolongation of single-short movement (N3s8, O9e1/2) or stands as an
independent verse (O2e6). There is no example that follows an aeolic phrase
with zero ending like rdod.
 Paean Two 375
Furthermore it is possible to unite s1b with the following verse, s2. The
Wnal part of s1b is preserved in four repetitions. There is no hiatus/brevis to
prevent two lines being united.
s2. Two ancipitia, which make this verse unique and diYcult to analyse,
are attested beyond doubt by the reading of the papyrus. The restoration of
2, A 
E w


w
, is certainly right while the whole of 74
E f ,

w
w
, survives in the papyrus. The third (38)
o ,

w
w
, is a restoration from the scholia, and is
accepted by many. Note that one of the two ancipitia, either the former or
the latter, is short, while the other is long, at all the three repetitions. Perhaps
the phrase x
x
w
may be a variation of an aeolic phrase,
x
x
ww
, in the same manner as the latter half of the eupolidean
dicolon (
x
w
) is equal to its Wrst half (
x
ww
) with
cretic taking the place of choriamb. In the eighteen majors, there is a
verse which seems to be related somehow:
P8s6 a
w
ww

u
w
w
k hepta e3(aeol)
Our verse and P8s7 may support each other:
P8s7 u
w
u
w
w
k
I introduce unusual scansion (
 w
) into the latter half of v. 49 47
Sn. to regularize the metre: e2. The position marked by w would be anceps
if metrical irregularity were preferable to prosodical irregularity, and aeoli-
cized e2 would occur. The other examples that Snells metrorum conspectus
collects under the head anceps in dactylo glyconei (pt. 2, 174) are not
really parallels.
s3. Glyconic pherecratean is not a usual combination in the eighteen
majors. There are two examples: O1s1 and P8e3/4. In Pae2s3 the combin-
ation shows typical characteristics of the pure aeolic, or Class I: (i) dove-
tailing occurs at three repetitions out of four (vv. 3, 39, 75), (ii) the full base of
the pherecratean is

(vv. 3/4, 75/76) corresponding with
w (50/51).
s4. The second position of the reversed dodrans is, like that in the
pherecratean in s3, explicitly realized as anceps: long at one repetition (52

), short at three (5, 41, 77).
d followed by full aeolic base is
paralleled in the eighteen majors:
I7e7

ww

w
ww
k
d rdod
P10e1 a
ww

a
ww
w
k xd gl
There are Wve examples in the eighteen majors of e following an aeolic
phrase with zero-ending, see Part I, 8. B. 5.
376 Appendix B. Longer Fragments
s5. There are examples, one in the eighteen majors (P5e3), and another in
the four minors (O4s8a), that are perfectly identical. Even the position of
resolution is the same:

w
ww
rw
ww
j rdod rdod
s6. Exact parallels for the combination, tel ph, are not found in the
eighteen majors, but there are two places in which telesillean is followed by
an aeolic with full base: N7s8 (tel hipp), O9s2 (tel gl 3). Against the
general tendency in the eighteen majors, bridge is not observed. Word-end
occurs at the joint of two phrases in the two repetitions that survive in the
papyrus (55, 80).
s7. For
d aeolic phrase with full base, see s4 above. Hipponactean is
a rare phrase in the eighteen majors. There are only two examples:
N7s8

ww
w

w
ww
w

k tel hipp
P2e8 w

w
ww
w

w
ww
w

k wil 2 hipp

It is interesting that all three, including our verse, are situated at the end of a
stanza.
e1. The nearest parallel is
P8e5

ww

ww
w
w
j
d d w e
For the other examples of d d, see Part I, 8. C. 2. It is least probable that
wwww
is d (choriamb), for the resolution of d is extremely rare. Rather,
the sequence
ww
wrw
w
is, as a whole, either d w e2 or dod w e.
I adopt the former analysis for two reasons: (i) P8e5 (above) is quite similar;
(ii) we2 resolved at the same position is used in e5. But the alternative is not
impossible; compare
N7s2
ww
wr w
w

ww
w

w
w
k dod w e dod e2
e2. Without the suYx e, this verse is equal to
I7e7

ww

w
ww
k
d rdod
e3/4. In the papyrus this verse is divided into two lines after the Wrst
positions of ibyc . This division is accepted by many. But the sequence
. . .
ww

ww
. . . is well attested, e.g.:
N6e1
ww

ww
ww
rw
j d D e
For ibyc , see Dactylic expansion revisited in this Appendix. The Wrst
half, two spondees and choriamb, would, as a whole, be analysed as a
 Paean Two 377
wilamowtizianum in drama. However, this form of wilamowitzianum is
totally alien to the eighteen majors. Rather,
O9e5
ww




j d sp sp
is a possible parallel.
e5. There are in total nine examples of xe2 in the eighteen majors, but
none of them is followed by reversed dodrans. Nor are there any examples of
the combination of rdod e
. However, the sequence . . .
ww
rw


at the end of the verse Wnds parallels:
O10e4/5

ww

ww
rw

j d d e

N6s6
ww
ww


ww
rw

k D
d e

e6. Five reversed dodrantes are repeated in I8s5 and the following, but a
closer parallel is found in the Wrst two phrases of
P2s2 www
ww

u
ww
w


ww
w
k rdod gl tel
The anceps of the second dodrans is long at 67 (P) and short at 31
(
E N). At 103 the text is missing.
e7. There are two examples of hepta 3 (O9e8, P10e5).
e8. In O9, three successive verses are made up of gl reiz.
e9.
d rdod is used in I7e7 (see above, on e2). For rdod ph, see
N3e2
w
ww

w
ww

k rdod ph
Like this (N3e2), the pherecratean in our verse may be better analysed as
rdod anceps (Part I, 7. 5). Moreover, it is possible that the initial wd
should be acephalous reversed dodrans (^ rdod); but the metre of ritual cries
(Nc N,
) is far from certain.
378 Appendix B. Longer Fragments

Paean Four
Two triads. Strophe/antistrophe: s13 23r; s3b8 24r. Ant. 2 is the only
stanza that preserves the whole text. Str. 2 too is in good condition but
some metrically crucial parts are absent. Str. 1 and ant.1 are only partially
preserved but they compensate metrically the missing part of Str. 2.
Epode: totally 1r. Only ep. 1 survives in the papyrus. The Wrst two
lines of ep. 2 are the rearrangement of the citation of Plutarch according
to the metrical scheme of ep. 1. Thus they are not helpful for metrical
study.
Class II.
(1)
Pae4s1 ww
ww
wwj
w
j wwibyc ( teld)
Pae4s2

w

ww
w
k
e dod
Pae4s3a

w

wr

wr

jwr

k
e e
e
e sp
(2)
Pae4s3b

w
k
e
Pae4s4 w
w



ww
ww
w
k we sp ibyc
Pae4s5


ww
ww
ww
ww
w
j sp ibyc
(3)
Pae4s6
ww
j d
Pae4s7 w

ww
ww


k ^ e D sp
Pae4s8 w

ww
ww
ww
ww
w
k ^ e ibyc

Pae4e1a

ww
ww
j
D
Pae4e1b

ww

w
rw
ww
j
d e rdod
Pae4e2

w


w
w


ww
w
k
e
e2tel
Pae4e3/4

ww

ww
ww
ww
ww
ww
w
k
d ibyc
Pae4e5 w
w
w

ww
ww
w
k we2ibyc
Pae4e6/7
wr
w

w
w

wr

j e e e2e sp
Pae4e8

wr

ww
w
k
e tel
Pae4e9 w
w

ww

j we d

Telesilleans at e2 and e8 should be analysed in this context as link anceps
dodrans. Ibycean is, by deWnition, expanded dodrans (dodd).

Strophe
The colometry above diVers considerably from those of the papyrus and of
modern editors.
 Paean Four 379
(i) s3ab. The third and the fourth lines on the papyrus are thus represented
(the wide spaces are given here for a practical reason, to highlight coinci-
dences of word-end within the line):

Third
str.1 3 ] P: 

ant.1 13 ] %

-
str.2 34   b [ ]
ant.2 44 K e f (
 Ka -

Fourth
str.1 4 H 
ant.1 14 b K  
str.2 35 a K        P
ant.2 45 
 d  r  P

The third line cannot be a verse in the modern sense:4 a word is divided at
the end of the line:
- (ant.1) and - (ant. 2). On the other hand,
there are two coincidences of word-end inside the fourth line:
3.

w

rw


4. rw

jrw
uj

w
k
Snell unites these two lines into one verse:


w

rw

rw

rw
x

w
k
Following him, West analyses it as 4 iambics dochmius. Instead of the
dochmius, however, Snell simply leaves the metrical symbols as they are,
x

w
, without giving them any name, apparently because of his total
avoidance of dochmiac metre in Pindar. These Wve positions remain per-
plexing. The diYculty lies in the initial anceps, long at two repetitions (4,
14) and short at the other two (35, 45), which is, nevertheless, followed by
(possibly) another anceps (always long). It should be separated from the
following four positions (

w
), contra Snell and West. Notably, word-
end is observed after it in all four repetitions: that is, there is verse-end. The
short syllable in this position is in fact brevis in longo.

4
Apparently Bowra makes a mistake. His reconstruction of the third line does not
make good metre.  P: 
: (str. 1, 3 according to his line-numbering) is one syllable
too long, and
(ant. 1, 12) is two syllables too long to correspond with Ka
(ant. 2, 40).
380 Appendix B. Longer Fragments
s3a

w

wr

wr

wr

k
e e
e
e sp
s3b

w
k
e

The similarity of s3a to

O10s3 u
w

wr

wr

j xe e
e sp
longer than O10s3 by another
wr
. Note too that
is striking.5 Pae4s2 is
the verses preceding these two are the same (
e dod). There are parallels for
the shortness of s3b. Apart from the examples in the eighteen majors, O7s3
in a normal D/e ode is remarkable; the same form (
e) is used as an
independent verse.
(ii) s5 (end)s8. These are the seventh, eighth, ninth, and tenth lines on the
papyrus. Again, wide spaces indicate coincidence of word-end:
Seventh

str.1 7 &
: 
ant.1 17  H
str.2 38   
 : <f> -
ant.2 48   A ; 
Eighth

str.1 8
ant.1 18
str.2 39 E
:   e r -
ant.2 49   
 Y-
Ninth

str.1 9
ant.1 19 
str.2 40
   

ant.2 50 . , ,

-
Tenth

str.1 10 

ant.1 20  . N
str.2 41 e   
.
ant.2 51 ,  b e 
.

5
O10s3 should be analysed as above, not as xe
e
e
, because the latter
analysis introduces cut between long anceps and the following resolved long against
Pindars tendency; see further, Part II ad loc.
 Paean Four 381
7. ww
w
j
ww
jw

8.
ww
ww

9.
wjw

ww

10. ww
ww
ww
w
k
Snell rightly separates the Wve initial positions of the 7th line on the papyrus,
  (str.2) and   (ant.2), from the following parts,
and unites them with the preceding line (s5).6 I follow him. After that,
however, my colometry diVers from Snells. Snell makes a very long, single
verse (his 6th):

ww
jw

ww
ww

wjw

ww
ww
ww
ww
w
k
However, it is better to divide this long verse into three verses at the two
points where coincidence of word-end occurs. The initial choriambic, which
is written in the middle of the seventh line on the papyrus, makes an
independent verse (s6). Note that syntactical division occurs after these
four positions at two repetitions (17, 48). Furthermore, the ninth line
should be divided after -
 : k (40) and - .k (50) (s7). Thus the
choriamb located in Snells chart with indentation at the beginning of the
second line of his 6th verse becomes

kw
, with brevis in longo
supposed after the second position (but  followed by  is not neces-
sarily brevis). Again, syntactical division occurs there.
The separation of s6 from s7 is necessary. If s6 were united with s7, the
united verse would include an aeolic phrase (dodrans) and D in one verse:

ww
w

ww
ww


k
There is no parallel at all for this in the eighteen majors.
The overall structure of the strophe is quite easy to grasp. It is divided into
three blocks composed of three verses each. Transition between the verses in
1 is smooth. The last six positions of s1 and s2 are identical (
ww
w
),
and so are the Wrst six positions of s2 and s3. The structures of the next two
blocks (2, 3) are similar to each other. Both start with a short verse made
up of four positions. All the other verses in both blocks include dactylic
movement of diVerent lengths. Of the two verses in each block, the number
of dactyls is smaller in the former. The shorter ones are similar to each other:
when the penultimate position of the ibycean in s4 is replaced by long, the D
sp of s7 appears. The dactylic movements of the longer ones (s5, s8), are

6
It must be remembered that Snell reproduces the layout of the papyrus. His own
analysis, which is diVerent from the papyrus, can be deduced from the metrical chart.
He gives the notation (gld) 5 da at the Wfth verse, which is meaningful when the
initial Wve syllables of the sixth verse are added.
382 Appendix B. Longer Fragments
completely identical. Spondee is found here and there (s4, s5, s7). For its
interpretation, see below.
(iii) The other verses of the strophe
s1. Exactly the same phrase is used in the Wrst half of
P2s4 ww
ww
ww
w
ww
ww
w
k wwibyc tel ( teldtel),
or rather, ^ D wibyc

And, although the beginning is diVerent,

O10e9

ww
ww
w
k
ibyc ( teld)
is the same in that the verse is made up only of this phrase.
s2. There are two verses of the same combination:
O10s2 u
w

ww
w
k xe dod
P8s4 w
w

ww
w
k we dod
s4. The Wrst four positions are we and cause no problem. What I describe
as spondee in the following part is taken by many as aeolic base. Thus, what I
analyse as sp ibyc is, according to them, gld. The restoration of the text
and the metrical theory are interdependent at the crucial point. Only one
line is completely preserved in the papyrus (the underlined letters make the
spondee):
ant.2 46 


H 
  K


The latter half of this part,
ww
ww
w
k, is preserved in the other three
repetitions too. The Wrst half, w
w


, is missing in two repetitions (5,
15). In the other (36), one of the two syllables which I suppose to be spondee is
missing. GrenfellHunt supplement the defect with a short syllable:
str.2 36 K
 :   n I

Housmans supplement, which Snell accepts in his edition, restores the same
metre as in 46:
f: g K
: H  n I

The reason why I prefer a spondee to aeolic base will be discussed in a

separate section (Dactylic expansion revisited).
s5. The initial spondee is attested in two repetitions. In the other two the
corresponding part is missing. The long dactylic sequence (ibyc ) is not
paralleled in the eighteen majors.
 Paean Four 383
s7. For ^ e preceding D, these verses must be compared:
I8s9 w

ww
ww
k ^e D
O10e2 w

ww
ww
k ^e D

Epode
The metre is far from certain. I give here a tentative analysis. It is diVerent
from that of Snell and West in the following cases: e3 and e4 are united; so
are e6 and most of e7;
 (29, e7) is scanned not as www but as
ww;
@ (29) is transposed from e7 to e8.
384 Appendix B. Longer Fragments

Paean Six
The metrical scheme is the most certainly attested of all the Paeans in the
non-D/e metre.
Three triads. Strophe/antistrophe: each verse is repeated Wve times at most
(ant. 1 is totally missing; str. 2 and ant. 3 are partially preserved), three times
at least. Epode: e15 1r; e67 12r; e810b 23r; e1114 3r.
Class III.
(1)
Pae6s1 ww
w
w
w
w
ww
w

j ^ dwe2hag
Pae6s2
w
ww

w
w
w

k rdod e3

Pae6s3
ww
w

w
k dod e
Pae6s4 w
w

ww
wt
j we ar
Pae6s5a rwr w
ww
w rw
j ewdwe
Pae6s5b
ww
j d
Pae6s6
ww

w
wwt
j d ph
Pae6s7 w
w
ww
w

w
ww
w

k hepta 2 hipp
(2)
Pae6s8a
wrw
j e2
Pae6s8b
ww
ww


ww
ww
k D
D
Pae6s8c ww

ww
wwj
ww


j ^ d D sp
Pae6s8d ww
ww
rw

w
k ^D e e
Pae6s9 ww
w
jw
w

wj
ww
k ^ d w e2rdod
Pae6s10 w rw
w
w
w
wrw
j w e6
Pae6s11
w
jw
ww
k wil (?e w d)
Pae6s12
ww
w

k ar
Pae6e1
ww
w
w
w
k dodwe
Pae6e2

w
ww
w

j hepta 3
Pae6e3 rw
ww
w
ww

j rdod reiz
Pae6e4


ww
j sp d
Pae6e5
ww

w
ww

j d ph
Pae6e6 w
[ ]
w
wwt
k [] ph
Pae6e7

w
ww
w

k hepta 3
Pae6e8
ww
ww
ww
ww
ww
w
j ibyc
Pae6e9/10a w
w
ww
jw

w
w
k hepta 2 e2
Pae6e10b
ww
w
w
w
ww

j dod hepta 1
Pae6e11
ww
w

y
ww
w
w
ww

j dod D (?)
hepta 1
Pae6e12
w
ww
w
j gl
Pae6e13a
wwrw

w
ww

k rdod ph
Pae6e13b w
w
w
ww

w
w
ww
k we w d e w d
 Paean Six 385
The colometry of the papyrus is wrong at two or, possibly three, verses;
unambiguously wrong at s8d, and, unless we admit a licence, at s8b. At these
two it ignores the mark of verse-end, hiatus, and makes a continuous line.
There is one more error at e10a, as long as we accept the reading of Vitelli
Radt.
s8b. In the 13th line in the papyrus the last half of s8b (D) is united with
the initial part of s8c (^ d ) in spite of hiatus at 95:

ww
ww
ww

str.1 13

ant.2 95 
::
str.3 135

. . .


The corresponding parts in str. 2 (74) and in ant. 3 (156) are not preserved.
The supplement 95 $ !  is beyond doubt. This means that the
papyrus ignores hiatus. Snell accepts this hiatus in mid-verse, apparently as a
special licence for a proper noun in dactylic movement. This is possible, but
unnecessary. The united verse, s8b s8c, would be of unparalleled length
in Pindar; cf. Radt 95. There are in total six examples of D k ^ d
( . . .
ww
ww
kww
. . . ) in the normal D/e; see Part III, A, under
Uncommon phrases.
s8d. The 16th line in the papyrus runs over the verse-end and includes the
Wrst half of s9 (ww
w
):
str.1 16 
ant.2 98

 
str.3 138  


The corresponding part in line 78 is not preserved. Hiatus is observed at 138

  K
-.
e10a. According to Radt, there is a position in the middle of a line on the
papyri (both 4 and 5) which is, in fact, the Wnal position of the verse.
Although the preceding position is short, long and short syllables neverthe-
less respond in this position. This is the Wnal position of e9/10a (Radts 9th
verse). Snell diVers from Radt. He reads short syllable throughout all the
three repetitions. This is Snells text:
53 . . . I
 $:
114 . . . X e K
-
175 . . .  : O
H  : -

The underlined syllables occupy the position in question. Both by keeping

this short and by supposing that word-end does not occur after it (note
O
H), he unites the latter half of e9/10a with e10b to make one verse. Radt
conWrms the reading of the editio princeps (Vitelli), which is:
386 Appendix B. Longer Fragments
175 . . . . . . : H  : -

Rutherford follows Radt, while Maehler (1989) follows Snell. From the
metrical point of view, Radts division is easier to explain; see below.
For the reading of the papyri, I follow Radt. My colometry too is essen-
tially the same as his (cf. his argument in detail, 93 V.), except in two places:
the separations of s5b from s5a, and of s8b from s8a. Snells colometry is
considerably diVerent. He does not accept s8c (Radts 9th verse) as an
independent verse, nor s8d (Radts 10th verse), but makes a gigantic verse
by uniting s8a/8b/8c/8d into one.
The metre of Paean Six is striking at Wrst sight (Das Metrum ist wieder
aolisch, aber die aolischen Grundformen sind hier viel freier variiert als im
zweiten Paian, Radt 93). But its peculiarity should not be exaggerated. As is
demonstrated below, the metre as a whole unambiguously belongs to Class
III. Each verse is well explained by the parallels collected from the epinikia
(the eighteen majors). The similarity with P2 is especially striking. There are
at least seven pairs of verses which are almost identical, common to both
Pae6s and P2s (with one being P2e). It is possible that Pindar may have
modelled one upon the other.
Pae6s1 ww
w
w
w
w
ww
w

j ^ dwe2 hag
P2s8

ww
w

w
w
w
ww
w

k tel e2 hag
Pae6s5a rwrw
w w
wrw
j ewdwe
P2s1 rwrw
w
w
wrw
k e6
Pae6s7 w
w
ww
w

w
ww
w

k hepta 2 hipp
P2e8 w

w
ww
w

w
ww
w

k wil 2 hipp
Pae6s8b
ww
ww


ww
ww
k D
D
P2s4 ww
ww
ww
w
ww
ww
w
k ^ D wibyc

Pae6s8c ww

ww
ww
ww


j

^ d D sp
P2s3

ww
ww
ww

w
k
D e
Pae6s8d ww
ww
www

w
k ^D e e
P2s7 www
ww
www

w
k rdod e e
Pae6s10 w rw
w
w
w
wrw
j we 6
P2s1 rwrw
w
w
wrw
k e6

Other parallels are frequently cited from O1, N3, and, outside the eighteen
majors, from O14. At the same time some characteristics of Class II are
mixed in: O10, N6, and, O4 outside the eighteen majors.
Resolution is frequent. Continuous shorts are remarkable: 6 shorts (s5),
and 4 shorts (s5, s8, and twice in s10). Amazingly, the penultimate position
of aeolic phrases with pendent ending is occasionally resolved: s4: 149 

:
( 88 I, 128 IA), s6: 131 I
 ( 91 
), e10b:
 Paean Six 387
176 I
( 54 E, 115 r ). These may support the unusual
resolution of w
wwt
k, caused by I
 at N3s6. Partly because of
frequent resolutions, RSS is very high.
Palindromic movement on a grand scale is prominent. Most examples are
unparalleled in the eighteen majors.
s1 ww
w
w
w
w
ww (1)
s5a wwww
ww
wwww none
s7
w
ww
w

w
ww
w
(2)
s8b
ww
ww


ww
ww
(3)
s10 wwww
w
w
w
wwww none
e2

w
ww
w

(4)
e3
ww
w
ww
(5)
e7

w
ww
w

e2
e10b
ww
w
w
w
ww
none
e11
ww
w
w
ww
(6)

(1) None, but a shorter form,
w
w
w
w
, is found at O9e1/2.

(2) Cf. P2e8 w

w
ww
w

w
ww
w

k; and
ww
w

w

ww
in a further six verses.
(3) None, but
ww
x
ww
in four verses, and
ww
w
ww
in
another four.
(4) None, but it includes glyconic.
(5) Cf. (2) above.
(6) O1e4, N3s1, N3e3, N4s4.

Strophe
Most verses of the Wrst half (1) include aeolic phrases. Particularly noteworthy
are those which have 3 ending ( . . .
ww
w

); hagesichorean (s1),
aristophanean (s4), hipponactean (s7). They articulate the structure of 1. The
3 ending is quite rare in the eighteen majors, and these phrases are unfamil-
iar. An aeolic phrase with pendent ending is used in s6 too (pherecratean).
This pherecratean is possibly one of the cases which may well be analysed as
rdod anceps (cf. Part I, 7. 5). Lengthy single-short movements are included in
s1, s2, and s4. This movement culminates in 2, s10.
Aeolic phrases are nearly absent from the latter half of the strophe (2).
Double-short movement becomes a key. D and its related phrases are
remarkably repeated (s8b, s8c, s8d). After s8d some resemblances to 1 are
pursued. s9 is a variation of s1, and, as mentioned above, single-short
movement, which appears here and there, is extended and attains its max-
imum length in s10. Finally, an aeolic phrase of 3 ending (aristophanean)
388 Appendix B. Longer Fragments
comes back. It echoes one of the themes of 1, and at the same time
introduces the beginning of the following stanza, whether strophe or epode.

Epode
Because restoration of the metrical scheme of the Wrst half (e17) relies on
essentially only one repetition (ep. 2), some uncertainties inevitably remain.
But the similarities with the strophe are evident; see below on each verse.
About half of the verses end pendent. Repetition of 1 ending, like that of
pherecratean, should be noted.

Individual verses
s1. For the similarity to P2s8, see above. Two-thirds of the latter part of
the verse is completely identical. The long single-short movement whose
initial short is replaced by double short has no parallel. Nearest to a parallel
are:
O1e5 ww
w
ww

w
w
k ^d w d e2
O4s1 ww
w
ww

ww
ww
j ^d wdD
They have only one single short between two doubles. Instead, our verse has
four singles.
s2. There is no exact parallel. Of the verses which start with reversed
dodrans, the following are partially similar:
I8s5b
w
ww

w
ww
j rdod rdod
N3e2
w
ww

w
ww

k rdod ph
O1e6b, N7e2 www
ww

w
k rdod e

There is one example of e3
(the preceding position, a double short, is

suspect but the sequence itself is certain):
N6e6/7 r
w
w
w

j ? xe3

A long single-short movement ends pendent in these verses too:
O2e6 w
w
w

k we2

N3s8 ww
ww
w
w
w
w

k telwe2

O9e1/2 w
w
w
w
ww
w
w
w

k we2wdwe2

s3. Two other verses include the same sequence as ours:
N7s3
ww
w

w
w
w
k dod e3
P6s5
ww
w

w
w
ww
j dod wil
 Paean Six 389
s4. These two end blunt, but are quite similar:
O10s2 u
w

ww
w
k xe dod
P8s4 w
w

ww
w
k we dod
Aristophanean ( dodrans anceps, by deWnition) is rare in the eighteen
majors. It is used at P11e1 (by itself) and O1e7 (end of verse).
s5a. See above for the similarity to P2s1, which would become identical
with Pae6s5 if the central
w
w
were changed into
ww
. Outside
the eighteen majors, there is one verse whose Wrst half is the same:
O4e9/10 rwr w
ww
w
w
w
k e w d w e2
ewd may be analysed as a wilamowitzianum which starts with six shorts. This
is not rare in tragedy, especially in Euripides. But it is unparalleled in the
eighteen majors. I prefer to analyse the whole verse (as well as O4e9/10) as a
long single-short movement with d at its core (the examples are collected in
Part I, 8. C. 3). There is no example of totally resolved e (rwr) in the
eighteen majors. But in expanded phrases, long sequences of short syllables
do occur, as in P2s1 above. Although resolution gives a diVerent impression,
the sequence ewdwe is incorporated in
N3e1b w rw
w
wr w
ww
w
w

w
k w e3 w d w e e
s5b. I separate d (
ww
) from the preceding verse because none of the
examples similar to s5 have another d after rightward extension. Word-end
coincides before and after d, and moreover, d is occupied by one word at all
three repetitions that can be read in the papyri (8  ; 90 ;
130 
). There is no verse in the eighteen majors that is made up
only of d. Thus it is possible to combine this d with the following s6; this is
preferable to doing so with s5.
s6. The parallel for initial d is
P5e6
ww
rw
ww

w
j d rdod e
Furthermore
N3e2
w
ww

w
ww

k rdod ph
is very similar, except for the initial
w.
s7. For the similarity to P2e8, see above. The diVerence lies only in the
initial position.
s8a. It is uncertain whether
wrw
should be separated from the
following D. There are no verses made up only of e2, but two examples
390 Appendix B. Longer Fragments
of
e2, O10e7 and I7s5b. Nor is there any example of e2 followed by
D. Resolution of the third position of e2 occurs at N3s3.
s8b. Note the essential feature of P2s4 and its similarity to Pae4s8b, see
above. Both Ds are prolonged to ^ D and ibyc, and the names given to them
conceal the similarity. There are no other verses that include two D in the
eighteen majors.
N6s6
ww
ww


ww
rw

k D
d e

is a remote parallel. Of course, in the normal D/e, D
D is familiar.
s8c. For the similarity to P2s3, see above. The Wnal spondee of Pae4s9
is replaced by e in P2s3. The Wnal spondee following D can be compared
with
I7e6
ww


k d sp
O9e7

w
ww
ww


j
ribyc ( heptad) sp
As for the initial ^ d followed by D ,
N6s5 ww

ww
ww
j ^d D

is similar though the length is diVerent.

s8d. For the similarity to P2s7, see above. Even the resolved position is
the same. Nomenclature apart, the diVerence lies only the absence of
the initial short. Identical too except for the diVerence in resolved
positions is
O10s1 ww
ww

w
tw
k ^D ee

s9. For ^ dwe2, see s1 above. Although there are three examples of rdod
e2,there are none of e2 rdod.
s10. See above for the similarity to P2s1, which lacks the initial w. On the
other hand,
O1s8 w rwrw
w
w
w
k we5
is one w
shorter but similar in that it starts with w.
s11. The wilamowitzianum of this shape is frequently used, but perhaps
this phrase is better analysed in context as e w d. Whichever it may be, there
is no other verse made up only of this phrase.
s12. P11e1 too is made up only of aristophanean.
e1. There are two examples of gl w e (I7e1, P5e2) and one of telwe (I7s2),
i.e. the sequence of threefold w
after the choriamb.
e2. This ten-syllable phrase, hepta 3, is used twice (P10e5, O9e8). Of
these P10e5 is made up only of this.
 Paean Six 391
e3. Reizianum and xd
are ambiguous (Part I, 7. 3). There are Wve
examples of wil reiz (O9s6/7, N4s6). Resolution of the Wrst position of
reversed dodrans is frequent (5 examples).
e4. Since the second position of reversed dodrans is regularly short, this
phrase is to be taken as sp d. There is no parallel, but in the reverse order see
I7e6
ww


k
e5. There is no exact parallel but
I7e7

ww

w
ww
k
d rdod
P10e1 w
ww

w
ww
w
k xd gl
are similar enough.
e6. The reading and restoration of v. 109 proposed by GrenfellHunt and
accepted by all later editors, e b (
, is rejected by Radt: so
scheint es mir fast sicher, dass hier   oder ]  stand.
Corresponding lines are absent.
e7. See above, e2.
e8. This line is a prolonged version of so-called ibycean (
ww
ww
w
);
see below, Dactylic expansion revisited. A shorter one is used in
N6e4
ww
ww
ww
w
j ibyc ( dod2d)
In our verse there are Wve dactyls. This is by far the longest. In the eighteen
majors D is the longest. D is used once in normal D/e.
e9/10a. These two verses are identical:
P8e1 w
w
ww
w

w
w
k hepta 2 e2
P11s5

w
ww
w

w
w
k hepta 2 e2
e10b. Although the analysis is diVerent, the sequence

ww
w
w
ww

(i.e. 2 single shorts sandwiched between doubles) is to be found in:
N3e3
ww
w
w
ww

w
ww

wtw
k dodwd rdod e2
N3s1

ww
w
w
ww

w
k telwd e
N4s4
w
ww
w
w
ww

k gl reiz
O1e4 w

w

ww
w
w
ww

k ^ e e dod reiz
(?^ e e dod w d
)

Our verse is longer by one more
w (three single shorts between doubles);
cf. s1 (four single shorts between doubles).
e11. Radt scans 117 - : I 
 as
ww
ww
. This scansion

seems blameless, and is certainly more plausible than the crasis (scanning 
392 Appendix B. Longer Fragments
I as one longum) proposed by Turyn; but it creates two metrical irregularities.
One is noticed by Radt himself: responsion between
ww
ww
and



ww
(56
d  ; the initial part of the other verse, 178, is
missing: :

). The double short of D is never replaced by a long,
in contrast to true dactyls.7 Snells analysis, ^ gld, should certainly be rejected.
His metrical note in the Metrorum conspectus is groundless.8 The other
metrical irregularity is that D is connected with reversed dodrans within a
verse. In the eighteen majors, unlike e or d, D is never used in the same verse
with an aeolic phrase. I leave the question open.
e12. Glyconic quite often makes a verse by itself.
e13. Resolution of the choriambic nucleus of an aeolic phrase is rare, but
there are examples (Part I, 5. F). The following verses are comparable:
P8s2
w
wwrw
k gl
P11s2b
w
wwrw

w
k gl e
N6e2
wrww
w
j gl

For the combination of rdod ph, see the verses cited above on s2.
e14. The sequence x e x d occurs in
P11e6 u
w
w
ww

k xe x d

N3s6

w
w
wwt
k
e x d

and a closer case is
O4s5/6 w
w
w
ww
w
w
w

k we wd we2

It is most likely that the latter half should not be analysed as wilamowitzia-
num but e wd in this case.

7
Radt gives four parallels: O10. 99, O14. 8, 21, N6. 64. Of these, only the Wrst is
textually certain (see ad locc.). But it should be interpreted as link biceps. Moreover it
involves a proper noun. I myself doubt the transmitted text.
8
Longum pro duobus brevibus semel a Pindaro in aeolicis ponitur: pae. 6, 56 . . .
(ii. 173).
 Paean Nine 393

Paean Nine (strophe only)

Two triads at least.
s12 2r; s3 3r; s49 4r; s10 23r. It is fortunate that Dionysius of Halicar-
nassus cites in total 19 lines ( the whole of 10 lines of str.1, 8 lines of ant. 1 and
1 line of ep. 1). Moreover, it is a happy coincidence that some letters of these
lines are conWrmed in the papyrus (fr. 126, col. ii). A practically complete
column follows this small fragment and gives most of another strophic pair (
6 lines of str.2 and 9 lines of ant. 2). Consequently, we have in total four
repetitions in the central Wve or six lines. The epode is missing except for one
line cited by Dionysius; I exclude it from consideration.
Class II
Pae9s1 u

ww
ww
ww
w
k x
ibyc [gl2d]
Pae9s2

w
w

ww
w
k
e2dod
Pae9s3/4 w
w

ww
ww
ww
ww


w
k we D
e
Pae9s4/5
ww
www
ww
ww
ww
k d ribyc
Pae9s6 u

ww
ww
w
k x
ibyc [ gld]
Pae9s7
ww
ww
ww
w
k ibyc
Pae9s8 ww
ww
w
k tel
Pae9s9


wr
w

k sp e e

Pae9s10

ww


w
k
d
e
At s3/4s4/5, Snell follows GrenfellHunt and keeps the colometry (or
layout) of the papyrus which establishes three verses. At Wrst sight this
colometry appears to be in harmony with other verses:
Pae9s3 w
w

ww
ww
ww
w
k we dod2d
Pae9s4


ww
ww
w
k sp dodd (or gld)
Pae9s5 w
ww
ww
ww
k wD
But, in fact, it assumes brevis in longo at all the repetitions of both s3 and s4
and sets verse-end after the last syllable, which is always short: s3 3r; (str.1) 3
I
 j N, (ant. 1) 13  j j, (ant. 2) 44   j H s4 4r; (str.1) 4
 j K
, (ant. 1) 14  j $
, (str.2) 34   j
, (ant. 2) 44
 j I. Consequently, it conceals the fact that
the preceding long double-short movement is carried on. It is theoretically
not impossible for brevis in longo to occur at every repetition, but it is highly
unusual for it to be repeated in two successive verses. Moreover, there is
another diYculty in this colometry. The Wfth syllable of s4 is, in fact, not
short at all the repetitions on the papyrus. It is long in ant. 2. 44:
Pae9s4


wa
ww
w
k
H
 e d 


394 Appendix B. Longer Fragments
Thus, instead of the blameless d, an emendation metri causa,  (Wilamo-
witz, GV 490), is introduced to expel the long syllable. Housman diagnoses the
error of the colometry and writes in such a punctilious manner that no
addition is needed (CR 22 (1908), 12 Collected Papers, 769):
It will be observed that this, the fourth line of the two strophes and anti-
strophes, always ends with a short syllable, which is never lengthened by
position nor followed by hiatus; and that the end of the third line, wherever
extant, obeys the same restriction. The fourth line therefore is apparently in
synaphea with what precedes and follows, and is merely a scribes line, not a
verse [my italics]. The division of the verse lies . . . where it is revealed by the
syllaba anceps.
I follow him and rearrange the three lines on the papyrus into two verses.
Rutherford does the same, although his analysis is diVerent.
Snell supplements  at the end of 33 (str. 2), the following word
being K
. This creates hiatus according to my colometry.

Individual verses
s1. The initial anceps is realized as long in str.1 and as short in ant. 2. The
other two corresponding lines are missing. For the analysis, see below,
Prolonged ibycean.
s2. There are in total nine verses in the eighteen majors which start with

e2, but none is followed by an aeolic phrase. A remote parallel is
I8s10

w
w
rw

w
ww
k
e2e rdod
s3/4. Although the verse as a whole can be described in D/e symbols, no
parallels are found in D/e odes or the eighteen majors, even if D were
equated with D (i.e. we D
e). The following are all those verses in which xe
precedes double short:
In D/e
N1e4

w

ww
j
e d
In the eighteen majors
O10e01 w
wr
ww
j we d
N07e01 w
w

ww
rw
k we d e
O10s2 u
w

ww
w
k xe dod
P8s4 w
w

ww
w
k we dod
There is no verse in the D/e odes, nor in the eighteen majors, that starts with
xe D. Even e D at verse-beginning is rare. It occurs only once in D/e:
 Paean Nine 395
O8e7
w

ww
ww
k
As for parallels for the latter half of s3/4,
e is never preceded by dactylic
movement and is hardly ever preceded by other phrases in the eighteen
majors, while there are in total six verses in the normal D/e odes that end
with D x e.
s4/5. The sequence
ww
www
ww
is included in the following
verses:
N6s3 www
ww
www
ww
ww
ww
j rdod ribyc ( rdod2d)
P5e3
w
ww
rw
ww
j rdod rdod
P5e6
ww
rw
ww

w
j d rdod e

N6s3 is similar to our verse in that expansion occurs in the latter half. But in
the paeans there are no other examples of reversed ibycean of any lengths.
s6. The initial anceps is once (ant. 1) realized as long (j
 ) while in

the other three it is short. For its interpretation, see below.
s7. If s7 could be united with the following s8, the united verse would be
very similar to
P2s4 ww
ww
ww
w
ww
ww
w
k wwibyc [ teld] tel)
But brevis in longo in ant. 1, 17  o prevents this.
s8. Verse-end is conWrmed by the hiatus at 8. Thus brevis in longo must be
supposed to occur at all of the four repetitions. Telesillean starting with
double short is common in the eighteen majors.
s9. The spondee preceding e is unparalleled in the eighteen majors.
However, in the normal D/e odes, there is a close parallel:
P1s3


w


w
k sp e
e
s10. There is no parallel for this simple verse in the eighteen majors, nor in
the D/e odes, except
O06e03
ww


ww


w


w
k d
d
e
e
which includes the same sequence as our verse.
396 Appendix B. Longer Fragments

Partheneion One
A pair of strophe/antistrophe and, if we rely on SnellMaehler, almost the
whole of an epode survive. This song may not be a partheneion because a
masculine participle is used for the singer ( 11).
Parth1s1 ww
w


ww
w
ww
j ^ dod

dwd
Parth1s2
ww
ww
w

k ibyc 1
Parth1s3 ww
ww
ww
ww

w

w
w

k ^D

e
e2

Parth1e1

ww
ww

w

w

w

w
w

k
D e e
e e2

Parth1e2

ww
ww
ww

w

w

w
w

k
D e
e e2

The stanza-forms of Partheneion 1 are short. The strophe/antistrophe is

made up of only three verses. The identical metre and paragraphoi at the
beginning of ll. 6 and 11 conWrm that these parts make up a strophic pair.
Therefore the lines starting at l. 16 with another paragraphos belong to the
epode, and so does the part preceding the strophe. SnellMaehler assume
that ll. 4 and 20 are in responsion so that the epode is as short as the strophe.
The scheme above is based on this assumption. The shortness is similar to
Partheneion 2 and, interestingly, to O5, which is made of three verses
(strophe) and two (epode) too.
All the verses, in strophe and epode alike, are unfamiliar in the eighteen
majors. On the other hand, s3, e1, and e2 are decidedly similar to some
verses in O5, whose authenticity has been questioned on metrical grounds
too. This feature is intriguing.
s1. There are two peculiarities: (i) ^ dod at the beginning; (ii) the following
long link anceps. There is the unique example of ^ dod in O13s5, but its
metrical context is diVerent:
O13s5 ww
w

w
ww

k ^ dod ph

Moreover, an aeolic phrase of 2 ending is never followed by long anceps in

the eighteen majors. Even e is rarely followed by
d; the exception is O10e3
(but the long preceding the anceps is resolved, unlike our verse).
d w d is uncommon too; an exact parallel is found only in N3e1, but there
is one example of ^ dwd and two of dxd; see Part I, 8. C. 2.
s2. This is equal to the alcaic decasyllable. It is used nowhere in the all the
non-D/e epinikia, nor in the paeans. The nearest parallel is
 Partheneion One 397
O10e9

ww
ww
w
k
ibyc
Our verse is the only case of expanded aeolic ending with w

; see Part
I, 7. 6 and dactylic expansion revisited below. In Simonides 542 P, which I
cite in Part I, 5. G to illustrate ^ dod, there is another similar verse:
1
ww
ww
w
w

j ibyc 3
This may be analysed as D w e
, however.
s3, e1, e2. These three verses are structurally similar: double-short move-
ment e or some es e2
with total absence of link anceps between
phrases. This structure is shared by some verses in O5. e2
, which would be
analysed as ithyphallic outside Pindar, is unfamiliar, except for O5; see above
ad loc.
There are in the eighteen majors four verses in which D or its cognates are
followed by e without an intervening link:
P2s3

ww
ww
ww

w
k xD e
N6s4a ww
ww
ww
rw

j ^ D e

O10s1 ww
ww

w
tw
j ^D e e
N6e1
ww

ww
ww
rw
j dDe

But the most similar verse is

O5e2


ww
ww
ww

w

w

w
w

k sp D e e e2

In O2, es are frequently juxtaposed without link anceps, but there are no
double-short movements there.
There are in total four examples of
D (O9e6, P2s3, N6e3, N6e9), but

^D is totally unparalleled.
398 Appendix B. Longer Fragments

Partheneion Two
Seven triads at least. Strophe: s1 59r; s2 710r; s3 6r. Epode: e1 23r; e2
23r. Class I
Parth2s1 ua
ww
w
u
w

u
ww
w
u
w
k gl x e gl x e
Parth2s2

ww
w

u
ww
w
k tel gl
Parth2s3

ww

k reiz
Parth2e1


ww
w

u
ww
w


ww

k gl gl reiz
Parth2e2
u
ww
w


ww

k gl reiz

Accustomed to the non-D/e metre of the epinikia and Paeans, one may be
surprised that Pindar wrote Partheneion 2 in such a simple metre, as if it
were a short song of Aristophanes. Whether this simplicity is related to the
genre of Partheneia, in other words to the Chorus made up of girls, is not
easy to answer.
Like the anceps in Class I stanza-forms, the anceps in this song is often
long, aeolic base or link alike (perhaps xe of this song should be iamb; then
link is not a suitable word). The base of glyconic is freely variable; in s1
(1st):

(4r),
w (4r), w
(1r, 61 KB); in s1 (2nd):

(7r),
w
(2r); in s2:

(6r),
w (4r); in e1 (1st):

(3r, all); in e1 (2nd)


(1r),
w (2r); in e2

(2r),
w (2r). s1 is made up of two identical
halves. Word-end coincides exactly after the Wrst half except in (36)
j, and perhaps also in (91) [ . . . ]j[ (  + ). If a special
licence is admitted in that the adverbial preWx of a verb (K) is treated like a
preposition, since a preposition occasionally stands at verse-end, s1 can be
divided into two equal verses.
 Dactylic expansion revisited 399

Dactylic Expansion Revisited

In Part I, 7. 6, I demonstrate that verses which seem to be created by the so-
called dactylic expansion of aeolic phrases are, in fact, related to freer D/e.
I argue here that this is conWrmed by the observation of the verses in the
fragments.
Expanded verses are used particularly in Paean 4 and Paean 9 but are rare in
Paean 2 and Paean 6 (only one each). Rutherfords comment that they are
characteristic of the

 genre is slightly exaggerated. These are the examples
collected from the four Paeans. First I set them out according to my analysis.
(I) Ibycean (for its deWnition, see below)
[1] Pae4s4 w
w



ww
ww
w
k we sp ibyc
[2] Pae4e5 w
w
w

ww
ww
w
k we2 ibyc
[3] Pae9s6 u

ww
ww
w
k x
ibyc ( ? gld)
[4] Pae4s1 ww
ww
ww
w
j ribyc ( teld)9
(II) Ibycean (and reversed ibycean) expanded by one dactyl
[5] Pae2e3/4




ww

ww
ww
ww
w
j sp sp d ibyc
[6] Pae9s7
ww
ww
ww
w
k ibyc
[7] Pae9s1 u

ww
ww
ww
w
k x
ibyc ( ? gl2d)
[8] Pae9s4/5
ww
www
ww
ww
ww
k d ribyc

(III) Ibycean expanded by two dactyls

[9] Pae4s5


ww
ww
ww
ww
w
j sp ibyc
[10] Pae4s8 w

ww
ww
ww
ww
w
k ^ e ibyc
(IV) Ibycean expanded by three dactyls
[11] Pae4e3/4

ww

ww
ww
ww
ww
ww
w
k
d ibyc
[12] Pae6e8
ww
ww
ww
ww
ww
w
j ibyc

Seeming aeolic base

I analyse the two positions preceding dactylic movement, for example,


of [1] Pae4s4, as spondee, and w
of [10] Pae4s8, as acephalous e. This
analysis is diVerent from the usual interpretation. Instead of spondee or
acephalous e, Snell, West, Rutherford and others take them as aeolic base. I

9
The notation ^ ibyc may be given instead. [4] is the acephalous form of [6]
below.
400 Appendix B. Longer Fragments
now pick out the examples that appear to start with full base and rearrange
them according to the shape of seeming aeolic base. The designation in
square brackets is that produced by taking the two positions as base.
(a)


[1] Pae4s4 w
w



ww
ww
w
k xe [gld]
[9] Pae4s5


ww
ww
ww
ww
w
j [gl3d]
(b) w

[2] Pae4e5 w
w
w

ww
ww
w
k xe [gld]
[10] Pae4s8 w

ww
ww
ww
ww
w
k [gl3d]
(c) x

[3] Pae9s6 x

ww
ww
w
k [gld]
The initial anceps is long at one repetition and short at three.
[7] Pae9s1 x

ww
ww
ww
w
k [gl2d]
The initial anceps is long and short at one repetition each.

It is noteworthy that there is no example whose full base is of the shape (d)

w or (e)
x. (d) is most frequently used in the aeolic phrases of the
eighteen majors, while (b) w
is rare (Part I, 5. C). If (a) and (b) were aeolic
full base, (d) or (e) would be expected to occur at least once.10
The phrases follow (a) spondee and (b) acephalous e are not only asym-
metrical (i.e. ibycean-type, ending with . . .
w
) but also symmetrical (i.e.
D). The latter is attested in Pae4s7, which is not expanded aeolic:
Pae4s7 w

ww
ww


k ^e D sp

This supports my analysis of the two examples of (b). It is interesting that

there are two verses similar to this in the eighteen majors:
O10e2, I8s9 w

ww
ww
k ^e D

And three verses out of Wve in O5 always start with spondee, for example:
O5s2


ww
ww
ww

w
w

k sp D (? rdod2d) e2

However, the two examples of (c) x
raise a diYcult question. If x
is
not aeolic base, how should [3] Pae9s6 and [7] Pae9s1 be analysed as a
whole? It may after all be better to give them the notation gld and gl2d.
But the comparison with the eighteen majors is still illuminating. The
aeolic base x
occurs only once in the eighteen majors, at
O10s6 a

ww
w
k gl

10
The metre of the second book of Sappho is gl2d: xx
ww
ww
ww
w
.
The aeolic base of the 27 verses of fr. 44 is Wlled with

(19 exx.),
w (5 exx.),
w
(3 exx.).
 Dactylic expansion revisited 401
The initial anceps is long in two repetitions and short in eight. O10s (and
O10e too) includes very few aeolic phrases and certainly belongs to Class II
(freer D/e). Moreover, in this stanza-form there is another verse, O10s4,
which may well start with x
. At v. 26 ( 25 Sn.), most editors (except
Mommsen and Schroeder) adopt Christs
 metri causa, which gives
initial w
in all repetitions. But H (the reading of some MSS) makes
good sense, and would produce one initial

:
O10s4 a

wr

w
k x
e
e
Perhaps Pindar may have introduced a licence that

at verse-beginning
may stand instead of the normal w
, especially in some stanza-forms. Note
that x
is Wlled with w
more frequently than

, in O10s and Pae9s
alike. This suggests that w
is the norm. The two Pae9s verses in group
(c) above may be variants of group (b), with ^ e being occasionally replaced
by spondee.

The metrical context

The metrical context indicates that the examples of expanded aeolic in the
eighteen majors are not really aeolic but freer D/e. Such verses do not occur
among ordinary aeolic cola in the Paeans either. Instead, D and its related
phrases tend to be used with them. For example,
Pae4s7 w

ww
ww


k ^e D sp

which precedes [10] Pae4s8, or

Pae9s3/4 w
w

ww
ww
ww
ww


w
k we D
e
which precedes [8] Pae9s4/5.
Most of the expanded aeolic verses are found in Paean 4 and Paean 9,
both of which belong unambiguously to Class II (freer D/e). On the other
hand, the strophe of Paean 2, which is the most aeolic of all the stanza-forms
of the Paeans, does not include any expanded aeolic. Paean 6, which
resembles Class III stanza-forms in its peculiarities, includes only one.11

Ibycean
All twelve examples listed above under (I)(IV), except [8], end with the
movement . . . ww
ww
w
. They are apparently similar to the verse

11
There are many D in Pae6: ^ D (s10), DxD (s8b), D (s9).
402 Appendix B. Longer Fragments
which has been traditionally called ibycean by some (
ww
ww
w
).
I adopt this name because it is the least tendentious. I also introduce the
expression expanded to ibyceans. This does not imply historical develop-
ment, but remains as a structural explanation. Thus the examples above can
be called, more or less, expanded ibycean. The designation, ibyc , is used to
represent (6)
ww
ww
ww
w
(Pae9s7, N6e4) for convenience sake.
The mark is used as in D (
ww
ww
ww
).12 As is demonstrated in
Part I, the ibycean used by Ibycus himself is not an aeolic colon. It is likely
that Pindar succeeded Ibycus in further developing such cola.
Ibycean itself could be described as dodd. Snell (and West) would use
dod2d for ibyc (in fact, Snells notation for N6e4 is gld). Certainly symbols
such as gld or gl2c are very helpful for the understanding of some Lesbian
metres. However, since ibycean-type phrases are not aeolic but freer D/e, it
would be misleading to use Snells symbols while at the same time rejecting
his interpretation. D w
would be a solution. However, such a notation is
misleading in another sense, suggesting a verse made up of D and suYx
w
. Another notation for ibyc would be 4da w
. That inevitably re-
minds us of the idea of lyric dactyls accepted by some scholars, including
Snell himself and Korzeniewski (and at one time Eduard Fraenkel). Old
ghosts like aeolic dactyls or logaoedic are not welcome.

w
ww
ww
is reversed ibycean (ribyc) in the same manner as

w
ww
is reversed dodrans. It is used in the Paeans only at [8] Pae9s4/
5, but is paralleled by
N6s3 www
ww
www
ww
ww
ww
j
which can be designated rdod ribyc . Outside Pindar, a verse of Stesi-
chorus (S148 Eriphyle, P. Oxy.2618, str/ant. 1) is e
ribyc :

w


w
ww
ww
ww
ww
ww
ww
k
and one of Ibycus (282 P, 4th verse) is ribyc
:

w
ww
ww
ww

k
The unusual phrase
w
ww
ww
ww
w
in P11s1 can be taken as a
conXation of ibycean and reversed ibycean (also expanded). The name, or
the notation, which I give to the phrase is ribyc 2, in the same manner as
hepta 2: x
w
ww
w
.

12
The alternative would be ibycd; but to designate [12] Pae6e8 (
ww
ww
ww

ww
ww
w
) ibyc is slightly less misleading than ibyc3d, just as D is
preferable to Wests D2 (see Part I, n. 73).
 Hephaestions Pindaric Hendecasyllable (fr. 94c Sn.) 403

Hephaestions Pindaric Hendecasyllable (fr. 94c Sn.)

In 14 (
d B  I

  ) of the Enchiridion (p. 44 Consbr.),
Hephaestion cites these two lines:
!   E F.
 t a 

 E
w

w
ww
w


The metre is called by an interesting name: Pindaric hendecasyllable. Some
of the letters in the Wrst citation have now been identiWed in a papyrus (P.
Oxy. 2438; see below), and Snell plausibly ascribes both lines to a Parthe-
neion (94c). We are accustomed to analyse the metrical form above as
wil 3 (or wil ba), based on the supposition that its Wrst two syllables
(w
) are the aeolic base and that the fourth (w) is anceps:

x
ww
w

.
If this analysis were rightly applied to the two Pindaric lines, the appellation
Pindaric hendecasyllable would become nothing other than a Hephaestionic
name given to wil 3. But there remains a possibility that what Hephaestion
means by the Pindaric hendecasyllable might be restricted only to the verse that
starts with w
(and the fourth position not anceps but regularly short; but I
leave aside this topic). Note that the two citations both start with w
.
Furthermore, another possibility must be taken into account; unlike wil 3,
the two positions may not be ancipitia. Instead, they may be Wxed as w
. It
should be remembered that, although the Sapphic hendecasyllable appears to
be equal to one of the possible permutations of wil 3, its Wrst two positions
are not ancipitia but must be
w unconditionally:

w
x
ww
w


Needless to say, the Sapphic hendecasyllable and wil 3 are diVerent metres,
and should be separately treated because their metrical contexts are diVer-
ent. At least it is impractical and confusing to treat the Sapphic hendecasyl-
lable as one of the possible permutations of wil 3. The same may be true
with the Pindaric hendecasyllable; it should be asked whether it is really
related to wil 3. Hephaestions manner of citation is interesting and may be
suggestive. The Pindaric hendecasyllable comes immediately after the Sap-
phic hendecasyllable as if the Pindaric were a variation of the Sapphic.
Compare these two hendecasyllable lines:
Sapphic hendecasyllable
w
x
ww
w


Pindaric hendecasyllable w

w
ww
w


404 Appendix B. Longer Fragments
Hephaestion classiWes both hendecasyllable lines under the name of epi-
choriambic trimeter catalectic. The Sapphic is trochaic syzygy choriamb
iamb I
. And, according to him, when the trochaic syzygy is
replaced by antispastic, the Sapphic hendecasyllable becomes the Pindaric. It
is not important whether his analysis is right or not. What concerns us is the
nature of the Pindaric hendecasyllable itself.
First it must be asked what Hephaestion means by antispastic. Strictly, by
the antispastic is meant only the metron w

w. But Hephaestion uses it
in a looser manner in the chapter on antispastic: e I
e c b

   
 a e

 N a  F

 ; . . . (10, p. 31 Consbr.). That is, the following are all
permitted to stand at the beginning of the antispastic metre:


w,

w
w, w

w, ww
w. If this interpretation were applied to the
Pindaric hendecasyllable too, what Hephaestion calls the Pindaric hendeca-
syllable would be equal to wil 3.
However it seems to me more likely that here in the description of the
Pindaric hendecasyllable the antispastic means only the exact antispast,
namely, w

w. There are several reasons. First, as is already pointed out
above, the actual citations both start with w
. Secondly, if the Wrst two
positions of his Pindaric hendecasyllable were capable of any combination of
, it would also cover the Sapphic hendecasyllable as one of its possibil-
ities. Hephaestion does not say so, though admittedly he is not always
logical. Thirdly, polyschematist (i.e. wilamowitzianum) is fully treated in
a diVerent chapter (16). It would be natural, and more logical, for the
Pindaric hendecasyllable to be treated there if it were virtually wil 3. Thus I
presume what Hephaestion regards as the Pindaric hendecasyllable is no
more wil 3 than the Sapphic hendecasyllable is. Short syllable is compul-
sory for the Wrst position of the Pindaric as well as the second position of the
Sapphic. And possibly also the fourth position of the Pindaric is short. But it
must be stressed that this deWnition of the Pindaric hendecasyllable is merely
what Hephaestion (or some of his predecessors in the Hellenistic period)
conceived. Hephaestion does not always correctly understand the metre of
archaic poetry.13 For Pindar himself, it probably was diVerent. We should
ask why, or in which historical process, the metre in question was named the
Pindaric hendecasyllable.
It is easily observable that most of the citations by Hephaestion come
from poems, or parts of poems, which are made up of the identical verses
a . Otherwise they are cited from short strophes like the Sapphic
stanza. At the same time it is unlikely that the Pindaric hendecasyllables were

13
Cf. Itsumi, Whats in a Line.
 Hephaestions Pindaric Hendecasyllable (fr. 94c Sn.) 405
arranged a  by Pindar himself. As far as we know none of his
poems was made up of verses a . Perhaps a Hellenistic poet may
have excavated this verse from the Pindaric corpus and used it a .
The title
H $ of this verse is, rightly or wrongly, ascribed to
Pindar by Hephaestion, but the propagator may have been another.
We go back to the two lines cited by Hephaestion and speculate as to their
context.
They are from the same poem without doubt, but syntactically the second
does not immediately follow the Wrst. There seems to have been an inter-
mediate line, or lines, of a diVerent metre, because, if this (or these) were
also the hendecasyllable, it would have been more natural for it (or them) to
be cited instead of the second citation. At the same time, if there is more
than one intermediate line, they must not have been numerous, or else the
hendecasyllable would not have been conspicuous or comprehensible
enough to cite the two lines above as a suitable example of the same metre.
It is fortunate that we have now P. Oxy. 2438. This papyrus is a part of a
version of the Vita Pindarica. In its 30th line we read eleven letters of the Wrst
of the two lines Hephaestion cites:
28    : 
: 
: d
: h-
:
29  z   K B TB w & I -
30   E F 
: :
:
:
It is in Hephaestions manner to cite the Wrst lines of a poem. The descrip-
tion in 29 (K B TB w# & I: ) supports the restoration of 30. In other
words, the two lines cited by Hephaestion come from the poem that the Vita
mentions.
Snell is certainly right to classify this poem among the Partheneia (fr. 94c).
Even if we were sceptical about the credibility of ancient Vitae in general, a
Partheneion would be the most appropriate genre for the two female proper
names, Protomache and Eumetis. They are cited as Pindars two daughters
not only in P. Oxy. 2438 but also in the Vita Ambrosiana. From the metrical
point of view, it is suggestive that the two verses belong to a partheneion.
The metre of Partheneion 2 is surprisingly simple (see above). It accords
with the manner of Hephaestion too, for he is interested only in simple
metres. Except for the hendecasyllable, his citations from Pindar are verses
made up of dactylo-epitrite.
It is time to speculate as to the exact nature of the Pindaric hendecasyl-
lable. In the eighteen majors there are a number of verses which start
regularly with w
. It is sometimes diYcult to identify these two positions:
either aeolic base or ^ e (Part I, 7. 4). Many of these cases are followed
by dodrans (w

ww
w
. . . ) and not immediately relevant to our
406 Appendix B. Longer Fragments
problem. But there are four which are followed by reversed dodrans,
i.e. parallels to the Pindaric hendecasyllable:

I8s1/2 w

w
ww
w
wr
w
w
ww
k wilwe wil

I8s5 w

w
ww

w
ww

w
ww
j wil (?^ e rdod) rdod rdod
P5s8 w

w
ww
w
k wil 2 (?^ e gl)
P2e8 w

w
ww
w

w
ww
w

k wil 2 hipp
cf.
P10s5 w

w
w
ww
w

k ^e wil 3

I am inclined to believe the following: the verse which Hephaestion calls the
Pindaric hendecasyllable was repeatedly used in the Partheneion 94c Snell.
The poems structure was very simple, and the verse was conspicuous. The
metrical nature of the verse was ambiguous like the examples above; it may
have been either wil 3 or ^ e hipp (or gl x), for its Wrst two positions were
w
in all repetitions. Their nature may have been discernible in the original
performance, but it was not so in late antiquity. Hephaestion noticed its
similarity to the Sapphic hendecasyllable.
 Part III

Miscellanea
This page intentionally left blank
 A. NORMAL D/E EPINIKIA

Determination of verse-ends
There are 23 epinikia which are composed of D/e (counting I3 and I4
as a single ode, and O13s68 O13e as an entity diVerent from
O13s15). The verses in these epinikia amount to a maximum of
321. Half (159 verses; 49.5%) are unambiguously identiWed as indi-
vidual verses; not united with the preceding or the following verse
because both the verse-beginning and the verse-end are marked by
hiatus/brevis in longo.1 The Wrm indication, hiatus/brevis, is absent
from the other verses either at the verse-beginning or at the verse-
end, or both, although word-end is coincident at both ends through-
out all the repetitions. Thus these verses may be united with the
preceding and/or the following verse. This situation may be said to be
basically the same as for the non-D/e verses. In some of the non-D/e
verses too verse-end is not deWnitely attested, but the number of
these verses is less, and colometry is easier to Wx; parallels helps us to
decide whether it is common or not for a particular combination of
phrases to make a single verse. In contrast, the D/e verses are much
more homogeneous, and more ambiguities remain.
For example, there is coincidence of word-end among all the
repetitions in O3e4, but hiatus/brevis is absent. Thus it is possible
to unite it with the following, O3e5, into one verse as Snell does:2
O3e4/5
ww
ww
j

w
x
ww
ww
k D
e x D
But at the same time it is also possible to establish two independent
verses:
O3e4
ww
ww
j D
O3e5

w
x
ww
ww
k
e x D

1
The statistics would be diVerent if verses were counted based on Snells edition.
Snell occasionally (32 cases in total) unites two verses of the 321 into one. Between
these two, of course, hiatus/brevis is not attested. According to his colometry, 179
verses out of 289 are certainly Wxed (61.9%).
2
This combined verse was given a single number, e4, by Snell, following Schroeder.
Snells line-numbering is diVerent from Boeckhs in this ode. Snells e5 is Boeckhs
(and Turyns) e6. The situation is the same as in the non-D/e odes. Cf. Line-Numer-
ation (p. xviii).
410 Miscellanea
The parallels are not helpful. Both colometries, the combined and the
separated alike, can be paralleled. For the former, there are seven
verses whose form is exactly identical; e.g.
O7s5
ww
ww


w


ww
ww
k D
e
D
Unlike O3e4/5, this cannot be divided into two because in six
repetitions out of ten word-end does not fall coincidentally after
the Wrst D. And for the separated colometry, two pairs of verses are
found to be parallels; e.g.:
N1s3
ww
ww
k D
N1s4

w


ww
ww
k
e
D
(the other is I1s4 s5). At the end of N1s3, hiatus occurs in two
repetitions, and brevis in one, out of eight; thus the D cannot be
united with the following
e
D. In O3e4, perhaps on the assump-
tion that a verse made of a single D is too short to be an independent
verse, Snell unites it with the following. But this is groundless. There
are in total four manifest examples of a verse made up of a single D
(and, in addition to O3e4, another seven that I take as independent
verses but whose independence is not unambiguously certain).
Metrical context is not a good criterion either. Returning to O3,
there is another verse of the identical form in the strophe with O3e4/5:
O3s1
ww
ww


w


ww
ww
j D
e
D
But this does not necessarily mean that O3e4 and e5 must be one
verse. Pindar may have used identical verses in the strophe and in the
epode, but at the same time it can also be maintained that he may
have slightly modiWed the verse in the strophe by dividing it into two
in the epode. Both explanations are possible.
We may wonder whether sense-breaks help us to decide whether or
not to divide verses. In fact they do not. Syntactically there is no
stronger or weaker pause after the initial D in e4 than at the end of e5.
The same is true of the preceding verse(s), O3e3a 3b, which is
similar in structure and has cut throughout all the repetitions at
the same position, though the sequence is reversed after the break.
Here too we Wnd no stronger or weaker pause after the initial D :
O3e3a 3b
ww
ww
j

ww
ww


w
k D
D
e
 Normal D/e epinikia 411
Thus the question remains open.
The example we have just been examining is O3e4 and e5, verses in
the epode. One may imagine that, if a verse were in the strophe and
were repeated more, hiatus/brevis would eventually occur. But the
number of repetitions does not always help. For example, hiatus/
brevis is missing at the end of O3s1 at all the 6 repetitions. Thus O3s1
and the following, O3s2, could be one gigantic verse:

ww
ww


w


ww
ww
j

ww
ww


w
k
D
e
D
D
e
None of the editors adopts this interpretation (nor do I), presumably
because the combined verse sounds too long. But this is not a
decisive argument. Its length, 31 positions, is not, in fact, extraor-
dinary. There is an example of a similar length (30 positions) and a
similar structure, and unambiguously a single verse:
P1s6
ww
ww

w


ww
ww


ww
ww

w

k
De
D
De

There is a more delicate case in the strophe: O3s4a 4b (24 posi-
tions):


w

j
w


w


ww
ww


w
k
e
e

e
D
e
Word-end is coincident after the Wfth position (
e
) at six
repetitions. Even a sentence ends (full-stop) in two of them. Thus
Turyn posits the verse-end there, rightly, I think, against Boeckh,
Bowra, and Snell. They must have thought that Wve positions are not
enough to form a verse. But this is a groundless assumption.
We should admit that there is no objective rule as to whether two
verses should be united or separated in these cases. Thus consistency
is required at least at the initial stage of metrical study. What I do is to
divide every verse wherever possible. This means that where word-
end coincides in all repetitions (i) before or after link anceps or
(ii) between two metrical phrases like D or e where link anceps is
absent, verse-end is posited even if manifest indication (hiatus/bre-
vis) is lacking. Thus, to cite the examples discussed above, I separate
not only O3s1 from O3s2 (where nobody would object), but also
O3e3 from O3e3b and O3e4 from O3e5 (where some, like Turyn and
412 Miscellanea
Snell, would object), and moreover I divide O3s4 into two at the
border after
e
. Consequently 34 of Boeckhs verses are divided
into two (Snell basically follows Boeckh except two cases). There are
some reservations. Even if word-end coincides before or after the
link, I cannot recognize unconditionally a short sequence like
j
w
j or j

w
j as an independent verse. Inconsistency
would be removed if the minimal length of verse were made smaller.
There is one verse made of four positions (see below). But I hesitate
to separate all cases like j
ej or je
j from a sequence and call them
a verse. I take j
e
j as a verse of the minimal length without
manifest indication. P1s6, cited above for its 30 positions, is the
longest verse; the shortest is
O7s3

w
k
e
Some editors, including Bowra and Turyn, accept a shorter verse
than this:
N1e1b
w
k e
Its length is not a problem, and we need not change word-order, as
Hermann and Bergk do, to expel the seeming brevis between [e1a]
and [e1b]. It occurs at two repetitions: 51 . . .  = I and
v. 69 . . .  = hKi . Rather we should accept prosodical
licence of lengthening the Wnal short syllable (-
) here. According
to Fuhrer, this prosodical lengthening tend to occur (i) in -, (ii) at
the end of double-short movement, (iii) after normal rhythm has
already been several times repeated before this irregularity occurs.3
Our case fulWls all three conditions.
Thus, as far as the D/e of the Pindaric epinikia are concerned, the
range of length of a verse which is unambiguously certain is between
four and 30. The actual variety of lengths is much wider than might
be expected.
One thing remains to be said about the determination of a verse.
One of the two manifest indications of verse-end, brevis in longo,
does not work when pendent verse-end is possible. For example,

3
See Part II, P5, Textual problems, on v. 42, and P11, on v. 38. Cf. Braswell,
Pythian Four 184(d), West GM 16. On the other hand, Dale prefers to accept the
division (Collected Papers, 75).
 Normal D/e epinikia 413
the last position of I1s1 is once Wlled by a short syllable (v. 35
E 
 ;  . . .). If I1s1 is independ-
ent of the following verse, the Wnal syllable of v. 35 is brevis. But if we
assume that the position is a link anceps, s1 and s2 could be one long
verse, though I reject this uniWcation for the reason discussed above.
Incidentally, when s1 is acknowledged to be independent, two ways
of notation are still conceivable. I do not notate
I1s1
ww
ww


w
x j
I1s2
ww
ww
x
w
k
but
I1s1
ww
ww


w

k
I1s2
ww
ww
x
w
k
Scholars have joked about the huge format of Boeckhs editio
maior and said that, without it, it would have been impossible to
print a verse on one line. But if a verse were divided wherever word-
ends are coincident as is my practice, more verses could be printed on
a line, even in the format of OCT or BT, with fewer indentations than
Bowra or SnellMaehler.

Verse-length
When a verse is established by dividing wherever possible, the average
number of positions in a verse is 13.3. But the average and the
divergence from it is not very useful for judging whether a particular
verse is common or not. There is too great a diVerence between the
longer and the shorter ones to make the average meaningful: as
mentioned above, the longest (P1s6) has 30 positions while the
shortest (O7s3) has only four. Moreover, because the standard
phrases of D/e are D (7 positions) and e (3 positions), we may easily
anticipate that verses of certain numbers of positions, like 12 (one D,
one e, and two links), will occur very frequently.
It is more promising to count double shorts and single shorts in
each verse. Then the peculiarity of P1s6 emerges. This is the only
verse that has three D (i.e. 6 doubles) in it. The other verses have two
D at most. Even the verses which have two D are not many (see the
chart below, where rows are arranged by pairs of double shorts).
414 Miscellanea
Since each D contains two such pairs, it is natural that verses with
odd numbers of them (1, 3, 5) should be rare. D , which contains
three pairs, and d, which contains one pair, are not such common
phrases as D (see below). There are in total 74 verses which have no D
(nor d, D ) and 156 verses which have one D, and so on. The
columns are arranged according to the number(s) of single shorts.
There are in total 52 verses which have no e, 97 verses with one e, and
so on. Thus, all the D/e verses are classiWed and counted according to
the number of D and e. For example, there are in total 62 verses
which have two pairs of double short (one D in most cases) and
one e. Ancipitia are totally ignored, and some rare phrases (see
below) are simpliWed.
! single 0 1 2 3 4 5 6 Total
# double
0 0 5 36 24 7 1 1 74
1 1 6 12 2 0 1 0 22
2 28 62 46 14 4 2 0 156
3 5 9 7 5 0 0 0 26
4 18 14 8 1 0 0 0 41
5 0 1 0 0 0 0 0 1
6 0 0 1 0 0 0 0 1
Total 52 97 110 46 11 4 1 321

Some interesting tendencies may be observed.

(i) Verses with four single shorts (4 es or equivalents; to some of
which Maas would give the symbol, e.g. E
E
) are decidedly rare
(11 in total), let alone those with Wve or six. They are much rarer than
those which have four double shorts (2 Ds or equivalents; 41 verses in
total). The extraordinary verse with six es (the highest Wgure) and no
D is N10s6.
(ii) More than one-third of the verses (126 52 74) completely
lack either double or single short. That means there is a fairly large
number of D/e verses in which D and e are not united.
(iii) We expect that, because of the rarity of D and d, verses with
odd numbers of double-short pairs (1, 3, 5) will be uncommon. But
we might not have expected there to be some equally uncommon
combinations with even numbers of pairs. Verses of two doubles and
three shorts (for example, P3s2, e
D
e
ek) are not numerous;
 Normal D/e epinikia 415
there are 14, ignoring permutations of phrases and ancipitia. Nor are
the verses of four doubles and one short (for example, O3s1, D
e

Dk), also 14 in total. We might easily have conceived these to be
typical, and expected them to be used much more frequently.
(iv) Pindar occasionally creates very lengthy verses like P1s6 (6
doubles and 2 singles). They are exceptional, despite being completely
made up of the two commonest phrases (D and e):
N1s7 e
D e
e
e
ek
N5s1
e x e e
e
D
ek
O12e7 e
e
e e
e
k
N10s6 e
e
e e x e
ek
Each Wgure in the chart above conceals the variety inside. For
example, there are 46 verses of two doubles and two singles. This is
undeniably a high frequency. These 46 verses are classiWed into as
many as 13 diVerent types, some of which are very common while
others are not:
e x D x e 12 e x D x e x 1 exD e 2 xexDxe 3
Dxexe 4 D exe 2 xDxexe 5 xD exe 2
exexD 6 exexDx 5 xexexD 1 dxdxexe 1
d e exd 1 x e x ibyc 1

In short, D/e verses are certainly homogeneous but at the same time
multifarious.

Link anceps
Like non-D/e verses, the D/e verses end blunt more often than pendent.
321 verses are classiWed, according to their ending, as follows:
. . .
ww
k (or j) 103
. . .
w
k (or j) 113
. . .
ww

k (or j) 42
. . .
w

k (or j) 60
others 3
In other words, two-thirds (68%) of the D/e verses end without link
anceps. Similarly, verses start without link anceps more frequently
than with it:
416 Miscellanea
k (orj)
ww
. . . 80
k (orj)
w
. . . 157
k (orj) x
ww
. . . 32
k (orj) x
w
. . . 36
acephalous 12
others 4
Consequently, it is rare that a verse both starts and ends with anceps.
Out of 42 verses above that end with . . .
w

, only seven start
with anceps, including Wve short verses made up just of x D x . And,
out of 60 verses that end with . . .
w

, only 12 start with anceps,
including four short verses made up of x e x . Similarly, out of 32
starting with x
ww
. . . nine end with anceps (of which 5 are x D
x ) and out of 36 starting with x
w
. . . 10 end with anceps (of
which 4 are x e x ).
On the other hand, link anceps is common inside a verse and is
normally posited between two phrases, e and D alike. There are 512
places where two phrases are located next to each other either with or
without intervening link anceps. Out of these 512 places, 459 have a
link (89.6%). When we count only the typical cases (i.e. the combin-
ations of D and e), the following Wgures emerge. Note the contrast
between the Wgures in the left and the right columns:
With intervening link Without intervening link
DxD 26 DD 1
Dxe 81 De 14
exD 111 eD 3
exe 185 ee 15
subtotal 403 33
It is evident that link anceps is a necessity in most cases. However,
there is one exceptional case that is not given in the table above.
When d (not D) is followed by e, anceps is often absent (d e 8
times in contrast to d x e twice). d e is the same length as e
e,
and their diVerence lies in the third position (
ww

w
:

w


w
). This fact gave rise to the illusion that these two
combinations could be in strophic responsion. Maas Wnally
showed that to be false.4 It is interesting that, contrary to d e, the

4
Bowra also argues against Wilamowitz and Schroeder in An Alleged Anomaly.
 Normal D/e epinikia 417
reversed combination, e d, is rare (e d 3 times in contrast to e x d 20
times).
There are in total 50 places where two phrases are connected
without intervening link anceps. Of these, 33 combinations are
cited in the chart above; and there are eight examples of d e and
three of e d. The other cases are: d D 1; ^ d D 1; D d 1; ^ d e 2; D e 1.
These places (missing links) are distributed unevenly in 43 stanzas.
There is no missing link in 12 stanza-forms while in two stanza-
forms there are four. The number of stanza-forms are arranged below
according to the number of missing links inside:
Missing links Stanza-forms
0 12
1 18
2 9
3 2 (O7e, O13e)
4 2 (N8e, P1s)
It is no accident that the stanza-forms which include three or four
missing links show unusual features in other respects too. Various
types of irregularities tend to be concentrated in particular stanza-
forms. O7e includes an extremely curious verse (O7e3, discussed
below); O13e belongs to the unique ode whose metre changes from
non-D/e to D/e (and O13s68 has two missing links in its D/e part
which has only three verses); N8e includes an unusual phrase, e2
(N8e4; and N8s is peculiar too); P1s is very complicated in its
construction (for example, two spondees in s2 and s3).
Some verses even include two missing links (these are counted
twice in the chart above). They are the following Wve verses:
P1s6 De
D
De

N10e6 r
eeD
e
e
N11s5 dee
d
I1s6 e
Dde
e
I6e8 e
Dee
e

Resolution of e
One of the two long positions of e may be resolved: rw
or
wr.
The other phrases may not be resolved. The exception is a very strange
418 Miscellanea
phrase included in O7e3 which I tentatively notate as ribyc (discussed
below). In 12 verses resolution of e recurs throughout the repetitions
(and also in O7e3).5 Besides there are 14 cases among 13 verses (there
are two resolutions in O3s5) in which resolution occurs at only one
repetition (11 cases)6 or some of the repetitions (3 cases);7 in other
words, resolved and unresolved positions are in correspondence.
When resolution of e occurs in all repetitions, it is almost exclu-
sively restricted (9 out of 12 cases) to the beginning of the verse,
which thus starts with three shorts, e.g.:
P1e7 rw


w

j
There are only three exceptions:
I2e6
wr

w

k
This is the sole example in which the latter long of e is resolved. And
in these two verses resolution occurs in mid-verse:
P1e3
w


w


w
rw
k
O7e5
ww
ww
w rw


ww
ww
k
O7e5 is highly unusual in that it includes four successive shorts (the
anceps preceding the resolution is always short).
When resolution occurs at only one or some of the repetitions, no
strong tendencies can be observed. Both tw
and
wt are found
(8 and 6 verses respectively). Resolution at the beginning of the verse
occurs in six verses out of eight examples of tw
.

Uncommon phrases
Apart from D and e, the following phrases are used in the D/e stanzas.

D (
ww
ww
ww
): 11 examples
This is an expanded form of D. Its usage is basically identical with
that of D. It stands at the beginning of a verse without preceding link
anceps (4 examples, P4s4, N1s6, I5e8, N1e2; the last one makes a

5
O7e5, O11e3, P1e3, P1e7, P1e8b, N1e1, N10e6, I2e6, I5s2, I5s6/7, I5e6, I6s6.
6
O3s5 (bis), O13e4, O13e5b, O13e6b, P1e5, P4e7, P9e10, N5s4, I3e6b, I6s3b.
7
P4s8 (23 resolved/3 unresolved), N5s6a (2/4), I6s6 (4/2).
 Normal D/e epinikia 419
verse by itself) or with anceps (1 example, P4e5). The other six
examples are preceded by e
(2 examples, N1e3, N5e6) or e

e
(4 examples, O6e2, P4s6, I3s5, I6s3). There are no verses in
which D and D are used together. When D is followed by one or
more phrases, anceps and e are commonest:
e or
e
or
e

e. The following are exceptional cases: e
without link in between
(I3s5) and
d (O6e2).
There are two examples that regularly have word-end after

ww
ww
in all repetitions:
I5e8
ww
ww
jww


w
j D
e
N1e2
ww
ww
jww
j D
It is possible to divide D there into two verses, e.g. Dj^ d
e (I5e8).
Indeed, there are six other pairs of verses in which D is followed by
^ d, but is manifestly separated from it by hiatus/brevis (see ^ d
below). N1e2a is followed by e2b
e
e without hiatus/brevis,
and is similar to I5e8 in structure.

D (
ww
ww
ww
ww
): 1 example (P3s4)

This appears to be the longest of the dactylic expansions in the
Pindaric epinikia (among non D/e odes, D is the maximum). P3s4a
is made up of D alone, but there is no manifest indication of
verse-end between it and the following verse:
P3s4a
ww
ww
ww
ww
j
P3s4b

w


w


ww
k
At v. 50, sense-break is very weak: (j H
d
 
j j H, . . .). Since X at verse-end is certainly unique, but not
impossible since there are some cases of j, the uniWcation of
s4a s4b is not compulsory.

d (
ww
): 35 examples
This is a rather common phrase and its usage is various. There are 22
examples at verse-end, in all but one of which there is no following
anceps. Only O7e7 ends with d
. Four verses start with d (there is
no verse starting with
d). Two verses are peculiar in that they have
two d:
420 Miscellanea
O6e3
ww


ww


w


w
k d x d x e x e
N11s5
ww

w

w


ww
k d e exd

^ D (ww
ww
): 2 examples (P3e9, N8s4)

Acephalous phrases are used in the other half too (2 examples of ^ D
in the eighteen majors). In the D/e stanzas, ^ D is rarer than ^ d,
contrary to the relative frequency of D and d. In both examples, ^ D is
followed by
e (and this in turn by D
in N8s4, by
e in P3e9).
The verse preceding N8s4 ( N8s3) ends with D. If it were
not for hiatus after v. 20 (
 j K ) and v.37 ( j
I
) a prolonged sequence of double shorts ( . . .
ww
ww

jww
ww
. . . ) would result.

^d (ww
): 9 examples
This may be followed by D, d, or e, with or without link, in a total of
nine verses, of which six (marked by #) it stands verse-initially
immediately after verse-Wnal D ( . . .
ww
ww
kww
. . . ):

D
P9s1, P9s3
D #O7e6

d #P1e8

e O7s1, #O7s6, #O8s6
e #O13e5b, #N8e3

With the sequence D k^ d we may compare D , which has regular

word-end after D; these examples are especially interesting, with their
suggestion that Pindar had intended to make a longer verse including
D at Wrst, but changed his mind. For example, if brevis in longo were
not established at the end of the preceding P1e7b (only at v. 59b

 : k ; in the other four repetitions the Wnal position is Wlled
with a long syllable), a longer verse, P1e7b e8 (e
e
D
d)
would emerge.

^e (w
): 1 example (O6s6)

This unique case will be discussed below along with the preceding
verse O6s5, which ends with an ibycean.
e2 (
w
w
): 1 example
N8e4

w


w
w
k
e
e2
 Normal D/e epinikia 421
The occurrence of this phrase in a D/e stanza may be an intrusion
from the other half , in which it is quite common.

sp (

) 3 examples
This phrase is employed at either the beginning or the end of the
verse. It should possibly be taken to be a substitute for e.
P1s3


w


w
k sp e
e
P9s2
w


w


k e
e sp
P1s2
w


ww

w


ww
ww


k e
d e
D sp
D/e is, by deWnition, totally made up of symmetrical phrases.
Nevertheless, there are in total six verses which include asymmetrical
phrases. They were collected in Part I, 7. 6, where I argued that so-
called expanded aeolics in the eighteen majors are not aeolic but
freer D/e.

hepta: 1 example

O13s6 x
w
ww


w


ww
ww

w
k hepta
e
D

This asymmetrical phrase is used as a transition from non-D/e to D/e
(see Part II, ad loc.). The initial anceps is link anceps as well as the
half-base of hepta; the following
w
ww
is, in a structural sense,
a substitute for D.

diomedean (ww
ww
w
x )
For the name diomedean see Itsumi, Enoplian in Tragedy; cf.
N10s7    

 H
 .
The sequence is one meaningful unit.
N10s1 ww
ww
w
x
w


ww
ww
k diom e
D
The Wrst single short is always short while the second is short only at
two repetitions out of 10 (vv. 1, 73); it is link anceps. Pindar does
not use the phrase ww
ww
w
( x ) elsewhere in the D/e odes,
but examples and kindred phrases are common in Euripides.
Its usage suggests that ww
ww
w
x occurs in places where
x
ww
ww
x is expected. For example, E. Hipp. 7558  76770:
ww
ww
w
w
w
w


ww
ww
w
w
ww
ww

422 Miscellanea
Snells notation of N10s1, ww
ww e u e
D, is inorganic and must
be rejected.
Both ww
ww
w
x in N10s1 and
x
ww
in N8s1 (below)
stand at the very beginning of an ode. This does not seem accidental.
Perhaps Pindar may have intended a surprise by starting a D/e ode in
this way.

ibyc : 1 example
O6s5

w


ww
ww
w
k
e
ibyc
Verse-end is guaranteed by hiatus (27, 56), brevis (34, 78) or both
(71). The following verse is also curious:
O6s6 w


w


w


ww
ww

k ^e
e
e
D

Snell proposes that pause intrudes into e (
kw
), which consists
of the Wnal position of s5 and the initial two of s6. But this idea
demolishes the fundamental notion of a verse and should be rejected.
It is better to accept that the phrase
ww
ww
w
is used in place
of D or D e after link anceps.

rdod
N8s1
x
ww


w

j rdod
e

40 h  I
; E K =   ,
This is the transmitted text of s1 and s1b. The metre is:

w
ww


w

j
ww
ww

k
The second position is always long at all the other repetitions. Thus
Snell deletes h and transfers  from the end to the begin-
ning of the line and changes it into the trisyllabic form I, with a
lacuna after . The intended metre is



ww


w

j
ww
ww

k
throughout all the repetitions. Its Wrst phrase is interpreted as D with
contracted initial double short.
But the contraction is highly unusual. Unlike true dactyls, D is
never contracted in any verse of the D/e odes. Perhaps it may be
compared with the second hemiepes (D) of the pentameter in the
elegiac couplet. As for the sense, h  I
and  
are natural and blameless. We do not need the bold metaphor
I  I
for the sake of the unparalleled metre. Rather it is
 Normal D/e epinikia 423
better to accept the transmitted text and the asymmetrical phrase

x
ww
as an intrusion from the other half.
N8 as a whole is unconventional. Our verse apart, e4 includes e2,
which is fairly common in the other half but totally unparalleled in
the D/e odes (see above). Of the less uncommon phrases, ^ D (s4), ^ d
e (e3),
d e (e6) are used. And link anceps is often short, especially
in e6, where it is always short throughout all the Wve repetitions.

ribyc
There is one apparent example:
O7e3
ww
www
ww
ww


ww
ww
k d ribyc
D
This must be examined along with the preceding verse:
O7e2
w


w


ww
ww

ja wwjb
O7e3
ww
wwjc w
ww
ww


ww
ww
k
Verse-end is needed somewhere (or nowhere), but colometry is most
uncertain. Word-end coincides at three places throughout all the Wve
repetitions (marked a, b, c). If we posit verse-end at either b or c, we
must acknowledge brevis in longo in all Wve repetitions.8 This is
possible and not unparalleled, but sounds artiWcial.
a b
14 $
E    , ! ',
33 r
  I
IA Pf K I
 ,
52  b E 
 E  .
71
F
 Ie

.  ! '
b d 
90 d
 IH d
d : K
d o  Ka 
b c
15 P
 Z
34 
b  
53 q b  .
72 
a 
91 P
E; 

c
15
 
H 

34 H f   

,
8
There is a diVerence between brevis at (b) and at (c). At (b) the Wnal syllable ends
with a consonant like ! ' in all Wve repetitions. In contrast, at (c), it ends with a
short vowel like Z in four repetitions out of Wve (v. 34  should be ).
Barrett Wnds that ending with a short vowel (which he calls SVE) is rare in Pindar; he
prefers (b) to (c).
424 Miscellanea
53  b d    .
72  K
d
 IH

91 d  
 Od  K IH
Strong sense-pause is observed after b in three repetitions (14, 33, 52)
and, if we do not divide between verses at a, adjective and noun are
united in two (33, 90). On the other hand, by dividing at a, a key
word, ! ' (14), and perhaps some others too, are emphasized at
the verse-beginning. The division at c coincides with sentence-end in
one repetition (53) but breaks a syntactical unit in others more or less
sharply. Z (15) before verse-end is especially awkward, but there
are some examples where a conjunction is separated from the fol-
lowing clause. In short, division at b is syntactically the least unnat-
ural.
From the metrical point of view, there are eight possibilities. Of
these permutations three are illogical, i.e. a b, a b c, none; Wve
remain:
1. a

w


w


ww
ww

j e
e
D

ww
ww
www
ww
ww


ww
ww
k ^ D ribyc
D
(or tel D
D)
2. b (the colometry I adopt, and perhaps Snell9)

w


w


ww
ww

w
kB e
e
De

ww
www
ww
ww


ww
ww
k d ribyc
D
(or dod D
D)
3. c (Turyns text)

w


w


ww
ww

ww
ww
w
kB e
e
D ibyc
w
ww
ww


ww
ww
k wD
D
4. a c

w


w


ww
ww

j e
e
D

ww
ww
w
kB tel
w
ww
ww


ww
ww
k wD
D
9
Snell starts a new verse after division (b) without indentation and numbers it 3.
But in his metrical chart, brevis in longo is not acknowledged, but two shorts ww are
left as they are.
 Normal D/e epinikia 425
5. b c

w


w


ww
ww

w
k e
e
D e

ww
w
kB dod
w
ww
ww


ww
ww
k wD
D
When division c is ignored, tribrach emerges (1, 2). This is very
strange in the D/e metre. Moreover, the tribrach is preceded and
followed by double-short movement. This is highly unusual even in
the other half. There are only three parallels, two of them in P5e:
P5e3
w
ww
www
ww
j rdod rdod
P5e6
ww
www
ww

w
j d rdod e
N6s3 www
ww
www
ww
ww
ww
j rdod ribyc ( rdod2d)
The analysis of the supposed tribrach in O7e is ambiguous. If we
suppose it to be equal to
w, as in the two verses of P5e cited above,
www
ww
ww
is a single phrase analysed as ribyc. If it is equal to
w
, the alternative colometry, which is in parenthesis in the chart
above, may be possible. The single merit of colometry 1 is that it
dispenses with brevis in longo throughout all repetitions.
The tribrach is eliminated when the verse is divided at c. Which-
ever colometry may be accepted among the relevant three (3, 4, 5), a
colon ending with
ww
w
emerges. This is unusual in the D/e
stanzas, but not without parallel. Certainly,
O6s5

w


ww
ww
w
k
e
ibyc
ends with it. But the division creates another oddity: the verse
following c starts with a link anceps that short in every repetition,
before D. This is unparalleled.
In conclusion: no colometry is obviously preferable to the others.
I adopt colometry 2 so as not to break the sentence units.
426 Miscellanea

B. TENDENCY CONCERNING BRIDGE AND CUT

Pindar tends to avoid word-end regularly at certain junctions between

metrical phrases. For example, as illustrated in Part I, 8, B, 1, word-end
is quite often absent between two aeolic phrases of which the Wrst
ends 2 ( . . . w
) and the second starts with full base (
x . . . ):
. . .
ww
w

x
ww
. . .
I group together the metrical patterns of this case (the regular
absence of word-end between two true longs) and call them Type A.
The same tendency is observed at certain places after link anceps
(Type B). For example, as illustrated in Part I, 6. D, word-end never
occurs after long link anceps when this anceps is explicit (i.e. the
position is Wlled by long at some repetitions but by short at others).
In other words, if cut occurs after explicit anceps, this anceps must be
short:
. . .
w
x  
w
. . . or . . .
w
x
 
ww
. . . or . . .
ww
x
 
w
-

. . . or . . .
ww
x 
ww
. . .
(Note that x represents a long syllable which Wlls explicit anceps. It
does not represent long implicit anceps although the same chart
would be applicable to it.)
There are diVerences in regularity. Of the cases grouped in Type B,
in that illustrated above (i.e. word-end after explicitly long link
anceps), bridge never occurs. There are 11 verses with explicit anceps
in mid-verse, found in a total of 79 repetitions, out of which long
anceps occurs at 30, and is never followed by word-end; cut occurs in
18 repetitions after anceps but only where this is short. However, the
situation of the anceps which is realized as long in every repetition
(long implicit anceps) is diVerent. The anceps is usually, but not
invariably, followed by bridge. There are six verses with long implicit
medial anceps, amounting to 74 repetitions, in 65 (87.8%) of which
word-end is absent (in other words bridge is observed) after these
ancipitia, however, cut exceptionally occurs in nine repetitions.10 As
for the representative case of Type A illustrated above, word-end

10
These cases are examined individually in Part I, 6. D.
 Tendency Concerning Bridge and Cut 427

between two aeolic phrases ( 2 ending followed by full base), word-

end is absent (i.e. bridge is observed) at the junction in 80 repetitions
out of 96 (83.3%) among 15 verses. There are 16 exceptions, and they
do not occur evenly among the 15 verses but are concentrated in six
(cf. List 5. 1). In other words, word-end is avoided without exception
at the junctions in as many as nine verses. This Wgure, nine verses out
of 15, enhances our sense of regularity, which cannot be ignored,
even if there is a counter-example, one verse in which cut occurs at
four repetitions out of eight (O1s1).
As stated in the Key to Terminology, I use the terms bridge and
cut in a looser manner than normal; my usage is simply descriptive,
with the object of collecting data for a further investigation. I exam-
ine whether word-end occurs regularly in a particular place and
count up the frequencies. I avoid the question why bridge occurs so
regularly there; important as it is, it is not my project to look for a
rule, or rules, and to explain its or their existence in Greek metre.
We may use the word rule for the limited cases where there are no
exceptions at all, like the bridge after long explicit anceps. But
perhaps we should not. First, the number of examples is small in
the eighteen majors. Contrast with Porsons Law, which is deduced
from c.30,000 stichic trimeters in tragedy. Secondly, the same ten-
dency is not always observed in similar cases; cf. the diVerence
between bridge after long explicit and implicit link anceps. Thus,
I use the word tendency instead of rule.
For a (strong) tendency to be observed concerning bridge in
Pindar (strictly, in the eighteen majors) the following three condi-
tions must be fulWlled:
(i) Examples of the same metrical pattern are not scarce.
(ii) Bridge occurs at the same point in a high proportions in these
examples.
(iii) There are some examples inwhich cut neveroccurs in any repetition.
(i) Suppose bridge were to occur at all the repetitions, or at most
repetitions (e.g. 7 out of 8), in a verse, but at the same time, the
metrical pattern were not found elsewhere except for this one verse.
Then we should not think this bridge represented a general tendency
in Pindar. Taking into account the variety of verses in the eighteen
majors, examples of the same pattern are not expected to be very
428 Miscellanea
frequent. I should tentatively suggest that there must be at least three
examples to allow us to assume the existence of a tendency.
(ii) How high must the proportion be so to make us feel that
bridge is regularly observed, even if there are a few exceptions? The
answer is necessarily subjective; but I suggest 75% as an answer if
all repetitions among all verses of the same pattern are added up. We
may not regard that 3 out of 4 as a high proportion in a verse by
itself. But if the frequencies of bridge in each of four verses are, say,
3/4, 6/7, 8/12, 4/4, making a total of 21/27 (77.8%), we shall say
that bridge is a tendency in this metrical pattern. These Wgures are
taken from an actual case: the junction between an aeolic phrase
and d.
. . .
ww
w

ww
or . . .
ww

ww

(iii) Even if the examples number more than three as under (i) and
the proportion is more than 75% as under (ii), we should hesitate to
acknowledge a tendency if there were one or more exceptions (i.e.
cut) at one or a few repetitions in every verse. Remember the junction
between two aeolic phrases discussed above: there are nine verses in
which cut never occurs at the junction. Thus, I should suggest that
there must be at least one verse in which cut never occurs at the
junction. In fact, there is no case that fulWls conditions (i) and (ii) but
fails (iii).
Based on these three criteria, I judge a tendency towards bridge to
occur in 11 patterns of Type A (bridge between two true longs) and in
three patterns of Type B (bridge after long anceps). They are all
collected below with reference to List 5 where the verses include the
pattern in question are given with the frequency of bridge and cut.
Then come three Wgures in square brackets. They represent (i)(iii)
above: the number of the examples in total, the frequency of bridge,
and the number of verses in which cut never occurs.

Type A (bridge between two true longs)

(a) between two aeolic phrases:
2 ending full base (List 5. 1) [15, 83.3%, 9]
 Tendency Concerning Bridge and Cut 429
Illustrated at the start of this section.
0 ending full base (List 5. 2) [5, 97.4%, 4]
e.g. . . .
w
ww

x
ww
w
. . . or . . .
w
ww

x
w
ww

...
(b) between d and an aeolic phrase:
d aeolic (full base or no base) (List 5. 7) [7, 82.1%, 2]
e.g.
ww

x
ww
w
. . . or
ww

w
ww
. . .
aeolic ( 2 or 0 ending) d (List 5. 8) [4, 77.8%, 1]
e.g. . . .
ww
w

ww
or . . .
ww

ww

(c) after x e followed by long at the verse-beginning:
x e aeolic (full base) (List 5. 11) [9, 81.3%, 4]
e.g. k x
w

x
ww
. . .
x e e/e2 (List 5. 15) [14, 81.8%, 2]
k x
w

w
. . .
(d) before e at verse-end:
aeolic ( 2 ending) e (List 5. 9) [15, 86.5%, 6]
. . .
ww
w

w
k
aeolic ( 0 ending) e (List 5. 10) [5, 76.0%, 2]
. . .
w
ww

w
k
aeolic ( 2 ending) e2 (List 5. 12) [9, 88.5%, 4]
. . .
ww
w

w
w
k
freer D/e phrase (e/e2/en) e (List 5. 14) [21, 81.4%, 5]
. . . [
w]
w

w
k
freer D/e phrase (d/D/D ) e (List 5. 16) [10, 73.2%, 6]
. . . [
ww]
ww

w
k
430 Miscellanea

Type B (bridge after long anceps)

after explicit link anceps [11, 100%, 11]
after implicit link anceps [9, 87.8%, 3]
These two are illustrated above.
after long half-base preceded by 0 ending of an aeolic phrase
(List 5. 3)
[3, 100%, 3]
e.g. . . .
w
ww
x
ww
w
. . . or . . .
w
ww
x
w
ww
. . .
 The All-but-One Rule 431

C . TH E AL L- B U T- O N E RUL E

Suppose a verse which includes an anceps position is repeated, say,

eight times. If this position is Wlled with a long syllable at all eight
repetitions, we shall judge the anceps to be regularized and accept the
regularization without any anxiety as it is. The next step is to ask
whether there is a metrical rule behind it. Or if this position is Wlled
with a long syllable at, say, three repetitions and with a short at the
other Wve repetitions, we shall again easily accept the situation,
thinking that the anceps is literally treated as an anceps. But in
contrast if the anceps is Wlled with a long (or short) at only one
repetition out of eight, we may feel uneasy: why is the anceps of this
particular verse Wlled with the one diVerent syllable against the
general tendency? Is this particular word including the long (or
short) anceps especially important, like a proper noun? Or, perhaps,
is there corruption in the transmitted text?
We not only feel uneasy about an irregularity caused by an appar-
ently insigniWcant word but also tend to suppose that the poet shared
the same anxiety. He should not have left the irregularity. Thus, for
example, Hermann in most cases proposes emendations to regularize
the responsion where exact responsion is breached by only one
repetition.11 However, as is demonstrated below, it fairly often hap-
pens that Pindar does breaks exact responsion by introducing a
diVerent pattern at only one repetition. In these cases the word in
question is not necessarily a key word; of course, the judgement
whether a word is important or not is to some extent subjective,
even apart from the consideration that every word in a poem is
indispensable and deliberately selected by the poet.
Anceps is not the only agent that breaks exact responsion between
verses. Breach also occurs by resolution. Anceps and resolution
concern a position of a verse and the syllables Wlling it. Apart from
positions, a regular pattern is also observed where word-end coin-
cides, or is avoided, after a particular position (bridge or cut).
11
Mommsens comment on non-regularized anceps in P10 is interesting: Anci-
pitis syllabae per totum carmen iuvenilis quaedam indulgentia et inconstantia regnat.
It betrays his feeling that anceps must be regularized in good poems and that
unregularized anceps shows immaturity. But, in fact, this is groundless.
432 Miscellanea
In section B above, I pick out 14 patterns where bridge tends to
occur. These three topics, anceps, resolution, and bridge in those 14
patterns, may be considered together; for each I count the number of
verses in which responsion is not exact, with the frequency of the
irregular cases. I limit calculation to verses repeated Wve or more
times because, the greater the number of repetitions, the more
uneasy we feel about the irregular case. In other words, we do not
call a case irregular if it is diVerent from just two or three others.
The stanzas which are repeated Wve or more times are listed
below according to number of repetitions (*means that the stanza
in question is monostrophic).
5 O2e P8e N2*
6 P6* I7s
7 I8*
8 O1s O9s P2s P10s P11s N3s
10 O10s O13s P8s
12 N4*

Table 1 represents the sum of all verses in which exact responsion is

broken. They are subdivided according to the number of repetitions
of each verse. Thus, for example, among the verses repeated eight
times, 17 have an anceps position which is not perfectly regularized
as either long or short.
Based on the same structure as Table 1, Table 2 represents the
number of verses in which exact responsion is broken at only one
repetition (All-but-One). Thus, in nine of the 17 verses repeated eight
times with an unregularized anceps a long (or short) syllable appears
at only one repetition against the general tendency of the other seven
(proper nouns are included).
 The All-but-One Rule 433
Table 1. Verses with inexact responsion
Locus of variation No. of repetitions

5 6 7 8 10 12 Total
Anceps 10 4 3 17 9 10 53
Resolution 2 3 2 4 9 0 20
Bridge 8 5 3 23 15 0 54
Total 20 12 8 44 33 10 127

Table 2. Verses with single inexact responsion (All-but-One)

Locus of variation No. of repetitions

5 6 7 8 10 12 Total

Anceps 0 2 3 9 3 2 19
Resolution 1 2 0 2 7 0 12
Bridge 3 4 1 10 5 0 23
Total 4 8 4 21 15 2 54

The ratio of All-but-One (Table 2/Table 1) is calculated in Table 3.

Table 3. Percentage of All-but-One

Locus of variation No. of repetitions

5 6 7 8 10 12 Total
Anceps 0 50 100 53 33 20 36
Resolution 50 67 0 50 78 0 60
Bridge 38 80 33 43 33 0 43
Total 20 67 50 48 45 20 43

The conclusion follows that All-but-One is not particularly rare.

434 Miscellanea

D. CONTINUOUS SHORT S YLLABLES

Pindar occasionally introduces a long sequence of continuous short

syllables in mid-verse. These sequences are usually 4 shorts long (24
examples), but there are much longer ones: 5 shorts (4 exx.), 6 shorts
(3 exx.), and even 7 shorts (1 ex.). These Wgures include examples in
which continuous shorts (i.e. resolution(s)) do not occur in every
repetitions (8 exx.). In drama, long sequences of shorts are often
found in dochmiac contexts and this usage culminates in the later
plays of Euripides, but dochmiacs are alien to Pindar. His typical
cases occur in a long single short movement such as e3, e5, e6. There
are two verses which contain two sequences each: N3e1b, P2s1. It is
striking that the examples are concentrated in particular stanza-
forms: P2s (4 exx.); P11s (3 exx.); N3s e (4 3 exx.); N7s (5
exx.). All of these belong to Class III. In the following list, all the
examples are arranged according to the number of continuous shorts
and their metrical contexts. # before a verse number means that
resolution occurs only once, and * means that resolution does not
occur at all the repetitions.
(a) Four shorts in single short movement
(i) included in e2:
O13s3 w
wrw


w
k w e2
e
#O13s4

wtw
x
w
k
e2 x e
P2s5

wrw
w
ww
k
e2 w d
P2s6

wrw
w
ww
ww
k
e2 w D
N3s3 www
ww
w

wrw
k gl e2
*N3e3
ww
w
w
ww

w
ww

wtw
k
dod w d rdod e2
(in 2 repetitions out of 4)
(ii) included in e3:
I8s7
w
wrw
w
ww
w

w
k e3 tel e
N7s5 w
ww

w
wrw
k w d e3
O1e2 w

ww

wrw
w
j ^ e d e3
 Continuous Short Syllables 435
(iii) included in e5 or e6:
N3s2
w
w
wrw
w
rw
k e5 e
O1s6
w
wrw
w
w
w
ww

w
w
k e5 w d e2
P2s1 rwrw
w
w
wrw
k e6
(iv) the Wnal part of en the following short anceps:
P11s3 rw
wr w
ww
k e2 w d
I8s3
w
w
ww

w
wr w
ww
w
wil e2 tel
N3e1b w rw
w
wr w
ww
w
w

w
k w e3 w d w e e
N3s5

w
w
wr w
ww

e3 w d
(v) the preceding short anceps en:
*P5e9 t x
ww
w

ww
w tw

w
k gl d w e e (in 3
repetitions out
of 4)
#P5e7b w tw

j we

N3e1b w rw
w
wr w
ww
w
w

w
w e3 w d w e e
(b) Four shorts in d:
#O10e10 w
wwt

wr
w

ww
k wd
eed
#N3s6

w
x
wwt
k
exd

(c) Four shorts in aeolic phrase ( short anceps):
N7s2
ww
wr w
w

ww
w

w
w
k dod w e dod e2
#N7s7
tww
w
rw
w
k tel e2
N7s4 w
w

wrw
ww
w

k w e wil 3
This case is included in (a)(iv) if the whole verse is analysed as x e e
hag.
(d) Five shorts (all are in aeolic phrases):
N6e2
wrww
w
j gl
P8s2
w
wwrw
k gl
P11s2b
w
wwrw

w
k gl e
*P11s4
wtww


wr
ww
k rdod
e d (in 4 repetitions
out of 8)
436 Miscellanea
(e) Six shorts:12
P5s4 wr rw
k ^e e
N7s6

ww
w
rwrw
w
k tel e3
P2s1 rwrw
w
w
wrw
k e6
(f ) Seven shorts:13
O1s8 w rwrw
w
w
w
k w e5
The successive short syllables are not always occupied by a single
word. Split resolution is frequent. Occasionally a long word is used
like 
 at v. 1 of P2s1. Impressive usage is cumulatively
found in N3 (s2) and N7 (s4, s5, s6); see Part II, ad locc. The type
  is not in Pindars repertory.

Addendum: Continuous Long Syllables

There are not many sequences of continuous longs is in the eighteen
majors (and possibly not in Pindars other poems either; see below).
The longest sequence is found in some of the repetitions in
P8e7:

u
u
ww
w
w

k
At three out of its Wve repetitions, both ancipitia are realized as long
so that six long syllables occur in succession. Of the sequences which
occur in every repetition (i.e. the sequences which consist of only
longs without anceps), that in O9e5 is the longest: Wve long syllables
in succession:

ww




j d sp sp
Others are shorter. There are in total six verses in which four long
syllables occur in succession, and nine others in which we Wnd a
sequence

x
(there is no example of
x

).

12
Outside the eighteen majors, there are two cases:
O4e9/10 rwr w
ww
w
w
w
k e w d w e2
Pae6s5 rwr w
ww
w rw

ww
j e w d w e d
13
Outside the eighteen majors, there is one case:
P7s1

w
wrwrw

k
e4

 Continuous Long Syllables 437
In the epic hexameter, successive longs are not rare (e.g.

a  N a ). Some dramatic anapaests contain only
long syllables. By contrast, the normal D/e may include three succes-
sive longs at most, the Wnal of D (or e) long anceps the initial of
D (or e). Pindar seems to have been much freer in non D/e metre. But
there are few examples of continuous longs as mentioned above.
Outside the eighteen majors there is a sequence of six (pure) long
syllables:
O4s4





j
And Wve continuous longs are used in
Pae2e3/4




ww

ww
ww
ww
w
j sp sp d ibyc
In Pindar, four or more long syllables never belong to one word.
 s
 (v.4, O4s4) is the nearest case. Perhaps such con-
jugated forms as  or  may have been avoided.
Thus, P11s1 is impressive where a single word repeatedly occupies
w


at verse-end (see Part II, ad loc.).
438 Miscellanea

E . V E R B A L A S S O NA N C E

Verbal assonance is the repetition of identical words, or words of

similar sound, at the same position in the same verse. Identical
repetition is conspicuous and easy to Wnd. For example, I occurs
three times at the end in this verse:
P2s5

wrw
w
ww
k
e2 w d
13    K  I
29 t. 
b
g K I
37 F f 
  I
In the next example, the word
E is the same and the subsequent
phrase is similar:
O9s4


ww
w


ww

k gl reiz
75 `N  : F
E  
94 @   F IH,
E  K 
 ,
In addition to the resemblance of sound, the grammatical form of
 and 
 is identical. Resemblances of sound are in
general, however, subtler and more delicate. For example, in this verse,
O1s7
w
w
w
w
ww

w
k e3 w d e
36 b (
, b  I
 ,
47 
 A  P H ,
not only do the sounds of I=P and
=H
resemble each other, but their metrical positions are also the same.
But grammatically, I=P are completely diVerent.
O2s8

w
rw
w
w
ww
k
e e2 w d
8 P 
  O
.
52 K   A  .
The resemblance of sound is less signiWcant, but undeniable. Then
how should we assess three other repetitions of this verse that start
with P- like P (8)?
16 h  
  
38 PA   d
 K  .
104 P

 I  
 Verbal Assonance 439
Their similarity does not go beyond P- (except that  in 38 is in the
same position as in 8 and 52). Certainly it is better not to widen the
deWnition of verbal assonance too much.
Semantic similarity is sometimes observable too:
P5s4 wr rw
k ^e e
46 
 , n 
97  
In this verse, verbal assonance is evident between
 (4) and

 (77). The initial four brevia are occupied by one word at
six repetitions out of eight.
I7s3/4

w
a
ww
w



ww
w
w

k
e tel gl 3
25/6

z
 b @ , a  IE
I.
42/3 NH. 
 a H
 .    . a a
 Y 
The underlined sentences are placed at the same position in the verse
and have related meaning. But this is not a case of verbal assonance.
Apart from the deWnition of verbal assonance, a word or word-
group with phonetic/grammatical/semantic similarities at the same
metrical position may be related to the creative process of the poet.
At the Wrst stage of creation, texts and music (metre) must have
simultaneously occurred to the poets mind, but gradually music
(metre) must have been Wxed before the texts at every repetition were
composed. I surmise that the process of creation will have been as
follows. Having composed a Wrst pattern-stanza, the poet devised
following stanzas to Wt the same melody, with, nevertheless, the possi-
bility of modifying the pattern-stanza here and there. We can imagine
that verbal assonance sometimes played a part in creating the new text.
Rudolf Fuhrer once made use of verbal assonance in an extreme
degree to prove the metrical responsion of seemingly non-respond-
ing texts in the word-order of the papyrus of Ba.17. I admire his
determination, but am not fully convinced to the point of accepting
all the metrical irregularities required. At least I am sceptical about
whether the same process can be extended to Pindar. Certainly verbal
association must have worked at some stage, but in the process of
improvement and reWnement of the poem, it disappeared, leaving
440 Miscellanea
only small traces. At least, Pindar did not lay stress on the similarities
of corresponding repetitions; the examples cited above are in the
minority. The majority of corresponding texts are very various,
phonetically, grammatically and semantically.
Thus, for example, Maas (Nachlese) proposed
 p at
O10. 110 ( 103 Sn.) in preference to   (the paradosis

  is metrically wrong) and Snell accepts it. The pres-

ence of
 in exactly the same metrical position in v. 44 ( 42
Sn.) is an argument for reading
 here as well, but, in view of
Pindars usual practice, not a very strong argument.
 Concentration of brevis in longo 441

F. CONCENTRATION OF B R EV I S I N LO N G O

Brevis in longo occurs occasionally with a remarakble frequency at a

particular verse-end. For example, eight repetitions out of 10 end
with brevis in O10s4. As illustrated above ad loc. in Part II, this
means that, if the ends of the other repetitions are modiWed, the verse
in question can be united with the following verse and that the united
verse will take on quite a diVerent form. There are even verses in
which all the repetitions end with brevis. The verse could be united
with the following if verse-end were not guaranteed by hiatus
(O10e9, I7e6).14
There are in total 24 verses in which a half or more of the
repetitions end with brevis. Of these 24 verses, nine are concentrated
in O10 (4 in O10s, 5 in O10e). In the table below these verses are
arranged according to the ratio of the number of breves in longo
against the total repetitions. When the ratio is the same, the verse
repeated more times comes Wrst.
verse repetitions brevis ratio
O10e9 5 5 1.00
I7e6 3 3 1.00
I8s8 7 6 0.86
O10s4 10 8 0.80
O10e2 5 4 0.80
P5s7 8 6 0.75
P5s9 8 6 0.75
O1e5 4 3 0.75
O2s2 10 7 0.70
P2s2 8 5 0.63
O10s2 10 6 0.60
O2e1 5 3 0.60
O10e3 5 3 0.60
O10e6 5 3 0.60
O10e7 5 3 0.60
I8s5c 7 4 0.57
I8s10 7 4 0.57

14
At v. 33 in I7e6, Barrett convincingly emends the text and unites e6 with e7 by
expulsion of brevis (I e E in place of `
 ). See Part II, ad loc.
442 Miscellanea
O10s1 10 5 0.50
O10s6 10 5 0.50
O13s5 10 5 0.50
O9s8 8 4 0.50
O9s9 8 4 0.50
N6s6 6 3 0.50
I7s2 6 3 0.50
 List 1. Aeolic Phrases 443

List 1. Aeolic Phrases

^ dod 1 gl 3 3 rdod 23 N4s8
O13s5 O9s2 O1e6b I7s1; e4
N7e5 P2s2, 7; e1a, wil 15
adon 2 I7s3/4 1b, e2, e3 O9s6/7
P10e2 hag 1 P5e2, 3 (2
), P5s2; e1
I7e2 P2s8 P5e6 P6s1/2, 4, 5
ar 2 P10s4 N4s3, 5, 6
hepta 12
O1e7 P11s4 N6s2
O9e8
P11e1 N3e2, 3 I8s1/2 (2
),
P2e7
N6s3 3, 4, 5a
dod 10 P8s3, 6; e2
N7e2
O1e1, 4, 7 P10s2a; e3, 4 wil 1 1
I7e7
O10s2 P11s2a P6s7/8
I8s5a (2
), 5b
P6s5 N4s2, 3 wil 2 4
(2
), 10
P8s4 I7s5a P2e8
reiz 9
N3e3 hepta 2 3 P5s8
N7s (2
), 3 O1e4
P8e1 P8s5
O9s3, 4, 5,
gl 49 P11s5 N3s7
6/7, 9
O1s1 N4s1 wil 3 2
O9s3, 4, 5, 8 hepta 2 N2s3 P10s5
O10s6 3 1 N4s4, 6 N7s4
P2s2; e1a, 1b, 2, P8e7 tel 29
3, 4, 5 hepta 3 2 O1e6a
P5s3; e2, 9 O9e8 O9s1, 2, 10; e4,
P6s1/2, 3, 6 P10e5 e6
P8s1, 2 P2s2, 4, 8
P8e3/4, 6 (2
) hipp 2 P5s7b
P10e1 P2e8 P10s6; e6
P11s2b; e4 N7s8 P11e3
N2s1, 3, ph 11 N3s1, 8; e4
s4 (2
) O1s1, 4 (2
)
N3s3, 4 O13s2, 5 N4s5
N4s4, 7 P5e4 N7s6, 7, 8
N6s1b, 2; e2, 8 P8e2, 3/4 I7s2, 3/4, 5a; e2,
N7s1; e3, e4, P10s1 3, 5
e5 P11e2 I8s3, 7
I7s5a; e1, e5 N2s4 tel 3 4
I8s4, 5c, 6 N3e2 N2s2
444 Miscellanea

List 2. Freer D/e Phrases

Each group is subdivided into subgroups in parentheses according to
whether the phrase in question is preceded or not by link anceps. Thus,
for example, three of Wve es in O2s3 are classiWed as (e) and two as ( x e):
O2s3

w
j
wrj
wyj

w
j tw
k
e e e
e e
The Wnal anceps is an exception, being attached to the preceding phrase; the
two es in O2s1 are classiWed into ( x e) and (e x ):
O2s1 w
w
j
w

k w e e

( x e x ) denotes e both preceded by link anceps and followed by link
anceps verse-end, as in:
O10e1b u
wr
k x e

^d 3 I7e6 N2s5 O1s2

O1e5 I8s6 N3s6 ( x D 4)
P6s4 (xd 24) ^D 4 O9e6
N6s5 O1s6, 7; e3, 5 (^ D 2) P2s3
d 55 O2s8 O10s1; e8 N6e3, 9
(d 26) O9s11; e1/2, 3, (^ D x 2) ^e 23
O1e2 4/5 O9e3 (^ e 22)
O9e5 O10e10 O13s1 O1s2, 9, 10; e2,
O10e1a, 4/5, P2s5 4, 7
D 9
10 P8e5 O2s2, 6/7;
(D 7)
P2e6 P10s4; e1 O10s4; e2
O10s3b; e2
P5s5, 10; e5, P11s3 P2e6
N6s5, 6a; e1, 5
e6, e9 N3s1, 5; e1a, P5s4, 6, 11; e1,
I8s9
P8s5; e5 1b, e3 5, 7a
P10s2b, s4; e3 N6s6a; e5 (xD 2)
P6s9
P11s4 N7s5 P2s6
P10s2b, 4, 5
N2s5 I7e7 P10s3
I8s9
^D 1
N3e1a (xdx 5) (^ e x 1)
N4s1 N6s4a
O9s11 N6s1a
N6s7; e1, 3 D 5
P10s3
N7e1 (D x 1)
P11e6
 List 2. Freer D/e Phrases 445
e 158 I8s5c, 8 ( x e2 11) sp 12
(e 92) (xe 51) O9e1/2 (sp 11)
O1s2, 7, 9(2 x ), O1s11; e1 O10e7 O2e4
10, e1, 3, 4, O2s1, 3(2 x ), O13s3, s4 O9e4(2 x ), 5
6b, 7 5, 8; e1, 4, 5 P2s5, 6 (2 x ), 7
O2s2(4 x ), 3 O9e8 P8s7 O10s3a, 3b
(3 x ), 4 O10s2, 3a P10s6 P8e7
(2 x ), 5 (2 x ), 4, 5; P11e5 P11s1
(3 x ), 6/7 e1a, 3(2 x ), I7s5b I7e6
(2 x ); e1 10 I8s10 ( x sp 1)
(2 x ), 2, 3 O13s2, 3, 4 ( x e2 x 3) P5s7a
(3 x ), 4 P2e5 O2e6
(2 x ), 5 P5s1, 2, 3, 9; e2, O9e1/2
O9s10, 11 4, 9 N3s8
O10s1(2 x ), 3a, P6s1/2, 7/8, 9
e3 13
3b, 4, 5; e3, P8s4, 7; e5
(e3 9)
10 P10e6
O1s3, 5, 7; e2
P2s3, 7(2 x ); P11s4; e6
O2e2
e4, 5, 6 N3s4, 6; e1b
N7s3, 5, 6
P5s1, 3, 4, 5, 6, 9 N6s7
I8s7
(2 x ), 10 N7s2, 4; e1
(3 x ); e1, 5, I7s2, 3/4; e1 ( x e3 3)
6, 7a, e8 I8s1/2 P5s11
(2 x ), 9 N3s5
(xex 3) N3e1b
P6s3 O10e1b
P10s2b; e3, 4, 6 P5e7b ( x e3 x 1)
P11s2b; e4 N3e1a N6e6/7
N3s1, 2, 4, 7; e3(aeol) 1
e2 38 P8s6
e1b, e4
N6s1b, 4b, s7; (e2 24) e5 3
e1 O1s6, 10, 11;
e1, 3, 5, 6a (e5 2)
N7e1, 2, 3, 4
O2s6/7, 8; e3 O1s6
I7s2
P2s8; e3 N3s2
I8s7
I8s8(2 x ), 10 P6s9 ( x e5 1)
(e x 12) P8e1 O1s8
O2s1, 6/7; e2, 5 P10s4 e6 1
O10e4/5 P11s3, 5; e3 P2s1
P6s6 N3s3; e3
P11e5 N7s1, 2, 7
N6s4a, 6a, 6b I8s3
446 Miscellanea

List 3. Identical Verses

Pure aeolic
gl 8 O9s8 O10s6 P8s1,2 N2s1 N4s7 N6e2,8
gl ph 2 O1s1 P8e3/4
gl rdod 3 P2e1a,1b,2
gl reiz 5 O9s3,4,5 N2s3 N4s4
gl wil 2 N6s2 I8s4
hepta 5 P2e7 P8s3 P10s2a P11s2a N4s2
ph 3 O1s4 P10s1 P11e2
rdod rdod 2 P5e3 I8s5b
tel 4 O9s1 O10e6 P5s7b I7e3
tel 3 4 N2s2 N4s8 I7s,e4
wil 2 O9s6/7 N4s6
Composite aeolic
gl e2 2 N3s3 N7s1
gl e 7 P2e4 P6s3 P11s2b P11e4 N6s1b N7e3,4
gl e x 2 P6s6 I8s5c
hepta 2 e2 2 P8e1 P11s5
rod e 2 O1e6b N7e2
tel e2 3 O1e6 P11e3 N7s7
x e dod 2 O1s2 P8s4
x e gl e 3 N3s4 P2e5 P5s3
x e ph 2 O13s2 P5e4
Freer D/e
^D x 2 O9e3 O13s1
^e D 2 O1e2 I8s9
^e d e 3 P2e6 P5e5 P10s2b
^e e 3 P5s4,6,e7
ee 2 O2s4 P5e8
e3 2 O1s3,5
x D 2 O9e6 N6e9
x Exe 3 O13s3,4 P11e5
x E 2 O1e7 I7s5b
x ee 2 O1s5 P5s1
x eee 2 O2e1 P5s9
x exd x 2 P11e6 N3s6
x e x 2 O1e1b P5e7b
 List 4. RSS in Ascending Order 447

List 4. RSS in Ascending Order

The Eighteen Majors D/e
stanza-form RSS Class stanza-form RSS
O9e 42.0% II N5s 34.3%
P8e 42.4% I I3/4e 35.6%
O9s 43.8% I I2s 36.5%
I7s 44.0% I O13e 37.2%
N2s 45.5% I I1e 37.6%
P10e 45.5% I N8s 37.9%
I7e 45.9% I I3/4s 38.2%
O2e 47.3% II N9s 38.5%
O10s 48.7% II N1s 38.9%
O13s15 49.0% II O3s 39.1%
N4s 49.5% I I6s 39.2%
O1e 50.0% III O6s 39.4%
P10s 50.0% III P1s 39.4%
P11e 50.0% III N5e 39.8%
P5s 51.8% II P4e 40.0%
P8s 52.1% I I6e 40.0%
P2e 52.4% I N8e 40.2%
O10e 52.4% II N11s 40.3%
P5e 52.6% III O3e 40.5%
I8s 52.7% III P3e 40.8%
O2s 53.0% II P4s 40.8%
N3s 53.1% III I2e 40.8%
P11s 53.2% III P9e 41.0%
N7s 53.3% III O7s 41.6%
P6s 53.6% I N1e 41.9%
O1s 53.7% III O13s68 42.0%
N3e 54.4% III P3s 42.2%
N6e 55.9% II N10s 42.5%
N6s 58.0% II O8s 42.5%
P2s 58.3% III P1e 42.6%
N7e 61.8% III P12s 42.7%
N10e 43.2%
The Four Minors O6e 43.3%
O5s 44.0% II I1s 43.4%
O5e 44.2% II P9s 43.7%
448 Miscellanea
O4s 45.4% II I5e 43.7%
O14s 48.1% III N11e 44.2%
P7e 48.4% II I5s 44.4%
P7s 50.0% II O8e 44.8%
O4e 53.7% II O7e 46.9%
 List 5. Bridge and Cut 449

List 5. Bridge and Cut

Examples in which cut is totally absent come Wrst; the rest are arranged
according to verse-number, with some exceptions in 9, 14, 15, 16.
In 3(b) and 4(b), only long ancipitia are counted; therefore the Wgures for
total repetitions are diVerent from those in 3(a) and 4(a). * means that the
anceps is Wlled by long at all the repetitions (i.e. implicit anceps); in those
cases these Wgures are equal.

1. Between two aeolic phrases: 2 ending and full base

bridge cut total
P2e8 wil 2 hipp 4 0 4
P6s1/2 gl wil 6 0 6
P6s5 dod wil 6 0 6
P8e3/4 gl ph 5 0 5
N2s4 gl gl 5 0 5
N2s4 gl ph 5 0 5
N6s2 gl wil 6 0 6
I7s3/4 tel gl 3 6 0 6
I8s4 gl wil 7 0 7
O1s1 gl ph 4 4 8
O9s2 tel gl 3 6 2 8
O13s5 ^ dod ph 8 2 10
P8e6 gl gl 3 2 5
N7s8 tel hipp 7 3 10
N7e5 gl gl 3 2 3 5

2. Between two aeolic phrases: 0 ending and full base

P8s6 hepta e3 (aeol) 10 0 10
P8e2 hepta ph 5 0 5
N3e2 rdod ph 4 0 4
N4s3 hepta wil 12 0 12
P2s2 rdod gl 7 1 8
450 Miscellanea

3. Between two aeolic phrases: 0 ending and half-base

(a) before the half-base
O9s6/7 wil reiz 1 7 8
O9e8 hepta hepta 3 1 3 4
N4s5 wil tel 6 6 12
N4s6 wil reiz 11 1 12
(b) after the half-base if it is realized as long
*O9s6/7 wil reiz 8 0 8
O9e8 hepta hepta 3 3 0 3
N4s5 wil tel 2 0 2

4. Between two aeolic phrases: 2 ending and half-base

(a) before the half-base
O1e4 dod reiz 4 0 4
P2s2 gl tel 8 0 8
N2s3 gl reiz 5 0 5
O9s3 gl reiz 5 3 8
O9s4 gl reiz 6 2 8
O9s5 gl reiz 5 3 8
N4s4 gl reiz 9 3 12
I7e5 gl tel 1 2 3
I7s5a gl tel 5 1 6
I7s5a tel hepta 3 3 6
(b) after the half-base if it is realized as long
I7s5a tel hepta 5 0 5
I7e5 gl tel 2 0 2
O9s3 gl reiz 5 2 7
*O9s4 gl reiz 5 3 8
*O9s5 gl reiz 6 2 8
*P2s2 gl tel 1 7 8
N2s3 gl reiz 0 2 2
*I7s5a gl tel 4 2 6

5. Between two aeolic phrases: 0 ending and no base

I8s5a wil rdod 7 0 7
P5e3 rdod rdod 3 1 4
I8s5a rdod rdod 1 6 7
I8s5b rdod rdod 3 4 7
 List 5. Bridge and Cut 451

6. Between two aeolic phrases: 2 ending and no base

P2e1a gl rdod 4 0 4
P2e1b gl rdod 4 0 4
O1e7 dod ar 2 2 4
P2e2 gl rdod 1 3 4
P2e3 gl rdod 3 1 4
I7e2 tel adon 1 2 3

7. Between d and aeolic

P10e1 d gl 4 0 4
I7e7 d rdod 3 0 3
P5e6 d rdod 2 2 4
P6s4 ^ d wil 5 1 6
P8s5 d wil 2 9 1 10
P10s4 d rdod 7 1 8
N3e3 d rdod 2 2 4

8. Between aeolic and d

P10e3 hepta d 4 0 4
P5e9 gl d 3 1 4
N4s1 hepta 2 d 8 4 12
I8s6 gl d 6 1 7

9. Between aeolic ( 2 ending) and e at the end of verse

O9s10 tel e 8 0 8
P2e5 gl e 4 0 4
P5s3 gl e 8 0 8
P11s2b gl e 8 0 8
P11e4 gl e 4 0 4
N6s1b gl e 6 0 6
P6s6 gl e x 6 0 6
P2e4 gl e 3 1 4
P6s3 gl e 5 1 6
P10e6 tel e 3 1 4
N3s4 gl e 7 1 8
N3e4 tel e 2 2 4
N7e3 gl e 3 2 5
N7e4 gl e 4 1 5
452 Miscellanea
I8s7 tel e 5 2 7
I8s5c gl e x 6 1 7

10. Between aeolic (zero ending) and e at the end of verse

O1e6b rdod e 4 0 4
N7e2 rdod e 5 0 5
P2s7 rdod e 6 2 8
P5e6 rdod e 3 1 4
P10e4 hepta e 1 3 4

11. Between x e aeolic (full base) at the beginning of verse

P2e5 x e
x gl 4 0 4
P6s1/2 x e wwwgl 6 0 6
P5e4 x e
wph 4 0 4
N7s4 x e
wwil 3 10 0 10
N3s4 x e
wgl 7 1 8
P5s3 x e
wgl 4 4 8
O13s2 x e
wph 5 5 10
P5s2 x e
wwil 7 1 8
P6s7/8 x e

wil 1 5 1 6

12. Between aeolic ( 2 ending) and e2 at the end of verse

O1e1 dod e2 4 0 4
O1e6a tel e2 4 0 4
P11e3 tel e2 4 0 4
P11s5 hepta 2 e2 8 0 8
N3s3 gl e2 7 1 8
N7s1 gl e2 9 1 10
N7s2 dod e2 5 3 8
N7s7 tel e2 9 1 10
P8e1 hepta 2 e2 4 1 5

13. Between aeolic (zero ending) and e2 at the end of verse

P2e3 rdod e2 2 2 4
P10s4 rdod e2 6 2 8
N3e3 rdod e2 1 3 4
 List 5. Bridge and Cut 453

14. Between two es at the end of verse

O2e1 ee 5 0 5
O2e3 ee 5 0 5
P5e8 ee 4 0 4
P5e9 ee 4 0 4
P5e7a ^e e 4 0 4
O1s9 ee 5 3 8
O2s2 ee 7 3 10
O2s3 ee 9 1 10
O2s4 ee 7 3 10
O2s5 ee 8 2 10
O10s1 ee 8 2 10
O10s5 ee 9 1 10
P2s7 ee 6 2 8
P5s1 ee 6 2 8
P5s9 ee 7 1 8
P5s10 ee 7 1 8
N3e1b ee 3 1 4
I7s2 ee 5 1 6
P5s4 ^e e 7 1 8
P5s6 ^e e 4 4 8
N3s2 e5 e 7 1 8

15. Between two es at the beginning of verse

P5s9 x ee 10 0 10
O1s11 x e e2 8 0 8
O1e1 x ee 2 2 4
O2s1 x ee 7 3 10
O2s3 x ee 9 1 10
O2s5 x ee 8 2 10
O2e1 x ee 4 1 5
O2e4 x ee 4 1 5
O2e5 x ee 3 2 5
O10s3a x ee 9 1 10
O10s5 x ee 9 1 10
O10e3 x ee 4 1 5
P5s1 x ee 6 2 8
O2s8 x e e2 7 3 10
454 Miscellanea

16. Between d (D, D ) and e

O1e3 de 4 0 4
P2e6 de 4 0 4
P5e5 de 4 0 4
P10e3 de 4 0 4
N6e1 De 3 0 3
P2s3 D e 8 0 8
O1s7 de 4 4 8
P10s2b de 3 5 8
N3s1 de 3 5 8
N7e1 de 4 1 5
 Bibliography

Barrett, W. S., Euripides: Hippolytos (Oxford, 1964).

Collected Papers Greek Lyric, Tragedy, and Textual Criticism: Col-
lected Papers, assembled and edited by M. L. West (Oxford, 2007).
Bergk14 Th. Bergk, Poetae Lyrici Graeci (Leipzig, 1843; 2nd edn. 1853; 3rd
edn., vol. i, 1866; 4th edn., vol. i, 1878).
Boeckh, A., ed. maior Pindari opera (Leipzig, 181121).
ed. minor Pindari carmina (Leipzig, 1825).
Bowra, C. M., An Alleged Anomaly in Pindars Metric, CQ 24 (1930),
17482.
The Metre of Pindar, Olympian II, CQ 30 (1936), 949.
OCT Pindari carmina (Oxford, 1935).
Braswell, B. K., Pythian Four A Commentary on the Fourth Pythian Ode of
Pindar (Berlin and New York, 1988).
Nemean One A Commentary on Pindar Nemean One (Fribourg,
1992).
Nemean Nine A Commentary on Pindar Nemean Nine (Berlin and
New York, 1998).
Burton, R. W. B., Pindars Pythian Odes (Oxford, 1962).
Bury, J. B., The Nemean Odes of Pindar (London, 1890).
The Isthmian Odes of Pindar (London, 1892).
Carey, C., A Commentary on Five Odes of Pindar: Pythian 2, Pythian 9,
Nemean 1, Nemean 7, Isthmian 8 (New York, 1981).
Chantraine, P., Gram. hom. Grammaire homerique (Paris, 1958).
Christ, W., BT Pindari carmina (Leipzig, 1869, 18972).
ed. maior Pindari carmina (Leipzig, 1896).
Dale, A. M., Collected Papers (Cambridge, 1969).
The Metrical Units of Greek Lyric Verse, CQ 46 (1950), 13848, ns 1
(1951), 2030, 11929.
LM The Lyric Metres of Greek Drama, 2nd edn. (Cambridge, 1968).
Dissen, L., Pindari carmina . . . ex rec. Boeckhii commentario . . . illustravit L.
Dissenius (Gotha and Erfurt, 1830).
Erbse, H., Beitrage zum Pindartext, Hermes, 88 (1960), 2333.
Fennell, C. A. M., Pindar: The Olympian and Pythian Odes, new edn.
(Cambridge, 1893).
Pindar: The Nemean and Isthmian Odes, new edn. (Cambridge, 1899).
456 Bibliography
Fuhrer, R., Beitrage zur Metrik und Textkritik der griechischen Lyriker: II a.
Text und Kolometrie von Bakchylides (c.17) II b. Zum Problem der Respon-
sionsfreiheiten bei Pindar und Bakchylides (NAWG 1976/5; Gottingen,
1976).
Gentili, B., Pindaro, le Pitiche (Milan, 1995).
Gerber, D. E., Emendations Emendations in Pindar 15131972 (Amster-
dam, 1976).
Olympian One Pindars Olympian One: A Commentary (Toronto,
1982).
Pindar, Nemean Six: A Commentary, HSPh 99 (1999), 3391.
Olympian Nine A Commentary on Pindar Olympian Nine (Hermes
Einzelschriften, 87; Stuttgart, 2002).
Gildersleeve, B. L., Pindar: The Olympian and Pythian Odes (New York,
1890).
Gunther, H.-Chr., Ein neuer metrischer Traktat und das Studium der pindar-
ischen Metrik in der Philologie der Palaologenzeit (Mnemosyne Suppl. 180;
Leiden, 1998).
Hartung, J. A., Pindars Werke (Leipzig, 18556).
Haslam, M. W., Stesichorean Metre. QUCC 17 (1974), 757.
Henry, W. B., Contracted Biceps in Pindar, ZPE 143 (2003), 1116.
Pindars Nemeans: A Selection (Munich and Leipzig, 2005).
Hermann, G., De dialecto De dialecto Pindari observationes, Programm
Leipzig, 1809 ( Opuscula, i. 24568), repr. in Heyne 1817, iii. 25075.
De metris Pindari, in Heyne, 17989, iii/1. 177356.
De metris poetarum Graecorum et Romanorum (Leipzig, 1796).
Elementa doctrinae metricae (Leipzig, 1816).
Emend. P. Emendationes Pindaricae, Programm Leipzig, 1834
( Opuscula, vii. 12973).
Nem. sextum Pindari Nemeorum carmen sextum, Programm
Leipzig, 1844 ( Opuscula, viii. 6875).
Notae Notae ad Pindarum, in Heyne 1817, iii. 23551.
Opuscula, 8 vols. (Leipzig, 182777).
Quinque Emendationes quinque carminum Olympiorum Pin-
dari, Programm Leipzig, 1848 ( Opuscula, viii. 11028).
Heyne, C. G., 17989 Pindari carmina, 3 vols. (Gottingen, 17989).
1817 Pindari carmina, new edn., 3 vols. (Leipzig, 1817).
Hohl, H., Responsionsfreiheiten bei Pindar (diss. Cologne, 1950).
Housman, A. E., Bacchylides Ode XVII, CR 12 (1898), 13440.
Classical Papers The Classical Papers of A. E. Housman (Cambridge,
1972).
Hummel, P., La Syntaxe de Pindare (Leuven and Paris, 1993).
 Bibliography 457
Irigoin, J., Histoire du texte de Pindare (Paris, 1952).
Itsumi, K., Choriambic Dimeter The Choriambic Dimeter of Euripi-
des, CQ, ns 32 (1982), 5974.
Enoplian in Tragedy, BICS 38 (19913), 24361.
Glyconic The Glyconic in Tragedy, CQ, ns 34 (1984), 6682.
Whats in a Line? Papyrus Formats and Hephaestionic Formulae, in
Hesperos: Studies in Ancient Greek Poetry Presented to M. L. West on his
Seventieth Birthday, ed. P. J. Finglass, C. Collard, and N. J. Richardson
(Oxford 2007), 30625.
Jurenka, H., Novae lectiones Pindaricae, WS 14 (1893), 134.
Kayser, C. L., JL review of Bergk1 and Schneidewin, Jahrbucher der Litera-
tur, 105 (1844), 97120.
Lectiones Lectiones Pindaricae (Heidelberg, 1840).
Liberman, G., Pindare, Pythiques (Paris, 2004).
Lloyd-Jones, H., GELT Greek Epic, Lyric, and Tragedy: The Academic Papers
of Sir Hugh Lloyd-Jones (Oxford, 1990).
Maas, P., Freiheiten Die neuen Responsionsfreiheiten bei Bakchylides
und Pindar, JPhV 39 (1913), 289320 (also publ. separately, Berlin, 1914).
Freiheiten II Die neuen Responsionsfreiheiten bei Bakchylides und
Pindar, zweites Stuck, JPhV 47 (1921), 1331 (also publ. separately,
Berlin, 1921).
GM Griechische Metrik (Leipzig, 1923; 2nd edn. 1929).
GM (1962) Greek Metre, trans. H. Lloyd-Jones (Oxford, 1962).
Kleine Schriften (Munich, 1973).
Nachlese Nachlese zu Pindar, JPhV 42 (1916), 1024 ( Kleine
Schriften, no. 6).
Merkelbach, R. Paonische Strophen bei Pindar und Bakchylides, ZPE 12
(1973), 4555.
Parker, L. P. E., Catalexis, CQ, ns 26 (1976), 1428.
Songs The Songs of Aristophanes (Oxford, 1997).
Trochee to Iamb, Iamb to Trochee, in Owls to Athens, Essays on
Classical Subjects Presented to Sir Kenneth Dover, ed. E. M. Craik (Oxford,
1990).
PfeijVer, I. L., Three Aeginetan Odes of Pindar. A Commentary on Nemean V,
Nemean III, and Pythian VIII (Mnemosyne Suppl. 197; Leiden, 1999).
Privitera, G. Aurelio, Pindaro: le Istmiche (1982).
Race, W. H., Pindar, 2 vols. (Loeb Classical Library; Cambridge, MA, 1997).
Radt, S. L., Pindars zweiter und sechster Paian (Amsterdam, 1958).
Rauchenstein, R., Commentationum Pindaricarum particula prima (Aarau,
1844).
Commentationum Pindaricarum particula altera (Aarau, 1845).
458 Bibliography
Rauchenstein, R., Zu Pindars Nemeen, Philologus, 13 (1858), 24563,
42142.
Rossi, L. E., Anceps, vocale, sillaba, elemento, RFIC 91 (1963), 5271.
Rutherford, I., Pindars Paeans: A Reading of the Fragments with a Survey of
the Genre (Oxford, 2001).
Schneidewin, F. G., rev. edn. of Dissen (Gotha, 1843).
Schroeder, O., BT13 Pindari carmina (Leipzig, 1900; 2nd edn. 1913; 3rd
edn. 1930).
ed. maior12 Pindari carmina (Leipzig, 1900; 2nd edn. 1923).
Pythien Pindars Pythien (Leipzig and Berlin, 1922).
Sicking, C. M. J., Griechische Verslehre (Munich, 1993).
Slater, W. J., Lexicon to Pindar (Berlin, 1969).
Snell, B., Griechische Metrik (Gottingen, 19623, 19824).
Pindarus, 5th edn. (Leipzig, 1971).
Stinton, T. C. W., Collected Papers Collected Papers on Greek Tragedy
(Oxford, 1990).
Pause and Period in the Lyrics of Greek Tragedy, CQ, ns 27 (1977),
2266 Collected Papers, 31061.
Stockert, W., KlangWguren und Wortresponsionen bei Pindar (diss. Vienna,
1969).
Theiler, W., Die Gliederung der griechischen Chorliedstrophe, MH 12
(1955), 181200.
Die zwei Zeitstufen in Pindars Stil und Vers (Schriften der Konigsberger
Gelehrten Gesellschaft, Geisteswiss. Kl., 17/4; Halle, 1941).
Thummer, E., Pindar: Die isthmischen Gedichte (Heidelberg, 1968).
Turyn, A., Pindari carmina (Oxford, 1952).
Verdenius, W. J., Commentaries on Pindar, Vol. 1: Olympian Odes 3, 7, 12, 14
(Mnemosyne Suppl. 98; Leiden, 1987).
Commentaries on Pindar. Vol. 2. Olympian Odes 1, 10, 11 (Mnemosyne
Suppl. 101; Leiden, 1988).
West, M. L., GM Greek Metre (Oxford, 1982).
Music Ancient Greek Music (Oxford, 1992).
Wilamowitz, GV U. von Wilamowitz-MoellendorV, Griechische Verskunst
(Berlin, 1921).
Pindaros (Berlin, 1922).
Willcock, M. M., Pindar: Victory Odes (Cambridge, 1995).
Young, D. C., Pindar Isthmian 7, Myth and Exempla (Leiden, 1971).
Three Odes of Pindar. A Literary Study of Pythian 11, Pythian 3, and
Olympian 7 (Mnemosyne Suppl. 91; Leiden, 1968).
 General Index

The items which are indicated in Table of Contents and Key to Terminology are
omitted. For the occurrences of each phrase (glyconic etc.), see List 1 and 2.

Aeschylus 34, 147; see also tragedy dochmiac, seeming 4, 12, 17, 45, 48,
inXuence of Pindar (?) (Cho. 317 and 147, 189, 379, 434
315 332) 40 n. 67
possibility of F  at Sept. 705 and eupolidean dicolon 26 n. 45, 252, 375
Suppl. 630 134, 136 Euripides see also tragedy 34, 40 n. 67,
Alcaeus 3, 68, 108; see also Lesbian 61 n. 102, 67 n. 114, 85, 172, 239,
ia gl in 70 LP 16 n. 29 259, 289, 297 n. 46, 389, 421, 434
Anacreon 6, 21, 34, 68, 82, 172, 245 catalectic cola in synartesis
Aristophanes 172, 252, 398 (Ba. 10510 1205)
glyconic starting at wwq 16 n. 28
(Ra. 1322) 5 n. 41
Hepahestion 6 n. 12, 133, 142 n. 3, 372,
bacchiac 28, 48, 205, 229, 234, 243, 266, 4036
308, 327, 357, 369
Bacchylides 8 n. 14, 108, 126, 229, 355 Ibycus 678, 108, 171, 174, 353, 402
3 2047
17 68 n. 115, 1689 lengthening of on at the end of verse,
18 5 n. 41, 46, 252 metrical 11 n. 20, 68 n. 115,
biceps 53 n. 88, 68, 182, 184, 194, 196, 2278, 269, 272, 299, 412
204, 302, 304, 305 n. 48, 305 n. 49, Lesbian (poets) 6, 16 n. 29, 21, 25, 34,
332, 392 n. 7 66, 185, 356, 402; see also Sappho
brevis in longo: and Alcaeus
or Wnal anceps 4 n. 89
not always a self-evident criterion 71, proper noun 24, 323, 39, 39 n. 63, 41,
73 n. 120, 347, 412 52, 65 n. 109, 105, 151, 191, 194,
213, 218, 253, 262, 275, 285, 297,
collected examples: 298, 331, 334, 340, 358, 385,
contact between two aeolic 392 n. 7, 431, 433
phrases 82 n. 125
e resolved 49 n. 81 Sappho 3, 64 n. 106, 68, 89, 108,
full base, exact responsion 32 n. 55 400 n. 10; see also Lesbian
full base, qx 32 n. 57 Simonides 110
half base 36 n. 60 acephalous dodrans (542P) 23, 204
rdod starting with www 9 n. 65 ibyc 3 (542P) 397
correption, epic 117, 130 n. 20, 249, Stesichorus 78, 108, 171, 174, 204, 402
262, 296, 303, 304, 306, 321, 337, 222b (responsion between single short
338, 351, 357, 361, 363, 364 and double) 68 n. 115
460 General Index
SVE (ending with a short vowel; named 63 n. 105, 68 n. 115, 80, 85, 99,
by Barrett) 195 n. 19, 231, 320, 108, 136, 195, 228, 234, 245, 259,
353, 423 n. 8 279, 285, 305, 327, 330, 359, 389,
427
tragic poets (tragedy) 10, 16, 22, 25, truncated phrase, doubly 26, 356
27 n. 47, 32, 34, 50, 52, 61 n. 102,
 Index Locorum

References are limited to (1) the Pindaric verses in theother half and (2) loci in
Part I.

O1s1 28, 32 n. 55, 34, 43, 623, O9s3 18, 32 n. 57, 33, 36 n. 60,
78, 80, 82 n. 125, 83, 82 n. 125, 86
92 n. 141 O9s4 86
O1s2 48 n. 79, 93, 94 O9s5 32 n. 55, 33, 36 n. 60,
O1s3 64, 96 n. 143, 99 82 n. 125, 86
O1s4 32 n. 55, 64 O9s6/7 15, 17, 32 n. 55, 36 n. 60,
O1s5 64, 96 n. 143, 99 38, 39 n. 63, 82 n. 125, 83
O1s6 19, 46, 61, 75, 95, 99, 101 O9s8 32 n. 57, 33
O1s7 46, 56, 61, 95, 98, 100, O9s9 36 n. 60, 75
101 O9s10 22 n. 38, 36 n. 60, 37
O1s8 45, 50, 51, 96 n. 143, 98 O9s11 22 n. 38, 57, 94
O1s9 49 n. 81, 96 n. 143, 97 O9e1/2 71, 79, 95, 100
O1s10 48, 96 n. 143 O9e3 37, 93
O1s11 54 n. 90, 96 n. 143, 99 O9e4 36 n. 60
O1e1 97 O9e6 66, 93
O1e2 26 n. 46, 99, 101 O9e7 65, 67, 69
O1e3 19, 95, 100 O9e8 36 n. 60, 74, 80, 82 n. 125
O1e4 36 n. 60, 78, 82 n. 125, O10s1 47, 49, 49 n. 81, 94, 97
87 n. 135, 92 O10s3a 49, 51, 51 n. 85, 56 n. 94,
O1e5 37, 47 n. 77, 94, 95, 101 57 n. 99, 59, 96 n. 143, 97,
O1e6a 17, 36 n. 60 97 n. 145
O1e6b 39 n. 65, 40 O10s3b 46, 93, 94
O1e7 27, 63, 80, 82 n. 125, 92 O10s4 51, 56 n. 94, 57 n. 99, 58,
O2s2 47, 49, 49 n. 81, 79 96 n. 143, 97 n. 145
O2s3 51, 49 n. 81, 56 n. 94, 57 n. O10s5 54 n. 90, 96 n. 143, 97
99, 58, 97 n. 145 O10s6 32 n. 57
O2s4 76 O10e1a 102
O2s5 49 n. 81 O10e1b 51, 76, 96 n. 143
O2s6/7 49 n. 81, 99 O10e2 48, 64 n. 107, 93
O2s8 79, 95 O10e3 49, 53 n. 88, 55, 57 n. 99,
O2e1 49 n. 81 58, 65 n. 109, 79, 97
O2e2 79, 99 O10e4/5 52, 71 n. 118, 94, 101
O2e3 71 O10e6 36 n. 60
O2e4 49 n. 81 O10e7 76, 96 n. 143, 99
O2e5 71 n. 118 O10e8 66, 76, 93
O4e9/10 50 n. 83 O10e9 65, 69
O9s19 22 O10e10 52, 55, 57 n. 99, 58, 94, 97,
O9s1 36 n. 60 101
O9s2 32 n. 57, 33, 36 n. 60, O13s1 37, 93
82 n. 125 O13s2 17, 32 n. 55, 54 n. 90
462 Index Locorum
O13s3 55, 57 n. 99, 58, 99, 96 P5e7a 49 n. 81, 76, 96 n. 143
n. 143, 97 n. 145 P5e7b 49 n. 81, 51, 71 n. 118, 76,
O13s4 49 n. 81, 54 n. 90, 57, 96 n. 96 n. 143
143, 97 n. 146 P5e8 76, 96 n. 143, 97
O13s5 32 n. 55, 423, 78, 82 n. 125 P5e9 34, 49 n. 81, 51, 87, 97, 99
P6s19 223
P2s1 19, 45, 50, 62 n. 104, 96 P6s1/2 74, 82, 84, 85
n. 143, 98 P6s3 32 n. 55, 40, 41, 42
P2s2 32 n. 57, 33, 36 n. 60, 39 P6s4 32 n. 55, 47, 86
n. 65, 74, 82 n. 125, 83, 85 P6s5 15, 32 n. 55, 82 n. 125
P2s3 66, 93, 94 P6s6 32 n. 55, 51
P2s4 15 n. 26, 36 n. 60, 61, 65, P6s7/8 32 n. 55
69, 79, 80 P6s9 48, 55 n. 93, 57, 96 n. 143,
P2s5 46, 95 97 n. 146
P2s6 46, 94, 95 P7s1 45 n. 75
P2s7 39 n. 65, 40, 97 P8s1 32 n. 55
P2s8 36 n. 60, 74, 91, 96 P8s2 32 n. 55, 41, 42, 50 n. 83
P2e1a4 32, 86 P8s3 36 n. 60
P2e1a 32 n. 55, 70, 78, 82 n. 125 P8s5 17, 32 n. 57, 33, 86
P2e1b 32 n. 55, 62, 78, 82 n. 125 P8s6 36 n. 60, 44 n. 71, 46, 81,
P2e2 32 n. 55, 39 n. 65, 82 n. 125 82 n. 125, 91
P2e3 32 n. 55, 39 n. 65, 82 n. P8s7 19, 21 n. 36, 46, 55 n. 93,
125, 90, 99 57, 96 n. 143, 97 n. 146, 99
P2e4 17, 32 n. 55, 40 P8e1 36 n. 60, 37, 90
P2e5 11, 17, 32, 32 n. 57, 54 n. 90 P8e2 32 n. 57, 33, 36 n. 60, 82
P2e6 26 n. 46, 48, 100 n. 125
P2e7 36 n. 60 P8e3/4 32 n. 55, 32 n. 57, 33, 62,
P2e8 32 n. 55, 78, 79, 80, 82 82 n. 125, 83
n. 125 P8e5 14, 94, 99
P5s1 96 n. 143, 97 P8e6 32 n. 55, 34, 78, 80, 82
P5s2 32 n. 55, 49 n. 81 n. 125
P5s3 32 n. 55 P8e7 29, 36 n. 60, 39
P5s4 50, 76, 96 n. 143 P10s1 32 n. 57, 33
P5s5 49 n. 81, 101 P10s2a 36 n. 60, 37, 39
P5s6 47, 76, 96 n. 143 P10s2b 26 n. 46, 48, 100, 101
P5s7a 11, 76, 96 n. 143 P10s3 54 n. 91, 57, 61, 78, 93
P5s7b 36 n. 60 P10s4 19, 26 n. 46, 57, 74, 78, 86,
P5s8 32 n. 55 88, 91, 94
P5s911 23 P10s5 32 n. 55, 48, 62, 92
P5s9 96 n. 143, 97 P10s6 21 n. 36, 36 n. 60, 40 n. 66,
P5s10 97, 101 56, 57, 61, 91
P5s11 48, 55, 57 n. 99, 58, 96 P10e1 32 n. 57, 33, 54 n. 90, 86
n. 143, 97 n. 145, 99 P10e2 75
P5e1 32 n. 55, 92 P10e3 14, 36 n. 60, 87, 100
P5e2 32 n. 55, 34, 62, 90 P10e4 36 n. 60
P5e3 39 n. 65, 79, 82 n. 125 P10e5 36 n. 60
P5e4 32 n. 55 P10e6 36 n. 60, 54 n. 90
P5e5 26 n. 46, 48, 100 P11s1 14, 40 n. 67, 65, 66 n. 110,
P5e6 39 n. 65, 79, 86 69
 Index Locorum 463
P11s2a 36 n. 60 N4s7 32 n. 55
P11s2b 32 n. 55, 42, 50 n. 83 N4s8 36 n. 60
P11s3 51, 95 N6s1a 11, 71, 76, 96 n. 143
P11s4 39, 40 n. 66, 40 n. 67, 41, N6s1b 32 n. 55, 71
56, 57 n. 99, 87, 90, 102 N6s2 32 n. 55, 82, 82 n. 125, 85
P11s5 16, 21 n. 37, 36 n. 60 N6s3 39 n. 65, 65, 69, 79, 80
P11e2 32 n. 55 N6s4a-s7 23
P11e3 36 n. 60 N6s4a 47, 52, 66, 71, 94, 101
P11e4 11, 32 n. 55 N6s4b 11, 71, 76, 96 n. 143
P11e5 51, 96 n. 143, 97 n. 145 N6s5 45, 47 n. 76, 93
P11e6 54 n. 90, 54 n. 91, 57 N6s6a 5, 15, 52, 57 n. 99, 93, 94,
101
N2s1 32 n. 55, 35, 62 N6s6b 11, 49 n. 81, 51, 76, 96 n.
N2s2 36 n. 60 143
N2s3 32 n. 57, 33, 36 n. 60, 82 n. N6s7 49 n. 81, 99, 101
125 N6e1 93, 94, 101
N2s4 11, 32 n. 55, 32 n. 57, 33, N6e2 32 n. 55, 40 n. 67, 41, 42,
62, 73, 80, 82 n. 125, 83, 84 50 n. 83
N2s5 54 n. 91, 57, 78, 94 N6e3 66, 93
N3s1 36 n. 60, 56, 78, 87, 95, 100, N6e4 65, 69
101 N6e5 93
N3s2 96 n. 143, 98 N6e6/7 53 n. 88, 71 n. 118, 96 n.
N3s3 32 n. 55 143, 99
N3s4 32 n. 55, 79 N6e8 25 n. 41, 31 n. 54
N3s5 19, 51, 95, 98 N6e9 93
N3s6 52, 54 n. 90, 57 N7s1 32 n. 55, 35, 62
N3s7 12, 17, 32 n. 55 N7s2 42, 74, 90, 90
N3s8 36 n. 60, 61, 91 N7s3 91
N3e1a 19, 55, 57, 79, 94, 95, 100 N7s4 32 n. 55, 39 n. 65
N3e1b 51, 75, 95, 97, 98, 99 N7s5 101
N3e2 156, 28, 32 n. 55, 63, 80, N7s6 36 n. 60, 50, 91
82 n. 125 N7s7 36 n. 60, 41, 79
N3e3 49 n. 81, 74, 78, 86, 87, 90, N7s8 32 n. 55, 33, 36 n. 60, 79,
95, 95 n. 142 82 n. 125
N3e4 15 n. 26, 36 n. 60, 61, 69, N7e1 100, 102
79, 80, 82 n. 125 N7e2 39 n. 65
N4s1 12, 16, 36, 36 n. 60, 37, 38, N7e3 32 n. 55, 42, 89
39 n. 63, 87 N7e4 32 n. 55, 40, 42
N4s2 36 n. 60, 38, 39 n. 63 N7e5 32 n. 55, 80, 82 n. 125
N4s3 32 n. 57, 33, 36, 36 n. 60,
82, 82 n. 125, 84 I7s1 36 n. 60
N4s46 86 I7s2 36 n. 60, 90, 97
N4s4 18, 32 n. 55, 36 n. 60, 78, I7s3/4 32 n. 55, 36 n. 60, 74, 82 n.
82 n. 125, 87 n. 135 125, 84
N4s5 17, 32, 32 n. 57, 36 n. 60, I7s5a 18, 32 n. 55, 36 n. 60, 73,
78, 82 n. 125 79, 82 n. 125
N4s6 17, 32, 32 n. 57, 36 n. 60, I7s5b 76, 96 n. 143, 99
39 n. 63, 38, 78, 82 n. 125, I7e1 32 n. 55, 56, 90
83 I7e2 36 n. 60, 82 n. 125
464 Index Locorum
I7e3 36 n. 60 I8s5b 80, 82 n. 125
I7e4 36 n. 60 I8s5c 32, 32 n. 57, 41, 52
I7e5 32, 32 n. 57, 36 n. 60, 82 n. I8s6 11, 15, 32, 32 n. 57, 87
125 I8s7 36 n. 60, 91, 96, 98,
I7e6 76 99
I7e7 86 I8s8 49, 50, 51, 79, 96 n. 143,
97
I8s1/2 32 n. 55, 49 n. 81, 74, 90 I8s9 48, 64 n. 107, 93
I8s3 32 n. 55, 36 n. 60, 51, 74, I8s10 54 n. 91, 91
91, 96
I8s4 32 n. 55, 82, 82 n. 125, 85 Pae6s5a 50 n. 84
I8s5a 32 n. 55, 35, 62, 73, 80, Pae6e10 30 n. 52
82 n. 125 Pae6e11 30 n. 52