Three Studies in

Athenian Demography

By M O G E N S H E R M A N H A N S E N

Historisk-filosofiske Meddelelser 56
Det Kongelige Danske Videnskabernes Selskab
The Royal Danish Academy of Sciences and Letters

Commissioner: Munksgaard ■Copenhagen 1988

 T he Royal Danish Academ y o f Sciences and Letters
publishes four monograph series, an Annual Report and, occasionally, special publications.
The format is governed by the requirements of the illustrations, which should comply with the
following measures.

Authorized Abbreviations
Historisk-filosofiske Meddelelser, 8° Hist. Fil.Medd.Dan. Vid.Selsk.
(printed area 175X 104 mm, 2700 units)
Historisk-filosofiske Skrifter, 4° Hist. Filos.Skr.Dan. Vid.Selsk.
(History, Philosophy, Philology, (printed area 2 columns,
Archaeology, Art History) each 199X77 mm, 2100 units)
Matematisk-fysiske Meddelelser, 8° Mat.Fys.Medd.Dan. Vid.Selsk.
(Mathematics, Physics, Chemistry, (printed area 180X 126 mm, 3360 units)
Astronomy, Geology)
Biologiske Skrifter, 4° Biol.Skr.Dan. Vid.Selsk.
(Botany, Zoology, Palaeontology, (printed area 2 columns,
General Biology) each 199x77 mm, 2100 units)
Oversigt, Annual Report, 8° Overs. Dan. Vid.Selsk.

The Academy invites original papers that contribute significantly to research carried on in
Denmark. Foreign contributions are accepted from temporary residents in Denmark, partici­
pants in a joint project involving Danish researchers, or partakers in discussion with Danish
contributors.

Instructions to Authors
Manuscripts from contributors who are not members of the Academy will be refereed by two
members of the Academy. Authors of accepted papers receive galley proof and page proof
which should be returned promptly to the editor. Minidiscs etc. may be accepted; contact the
editor in advance, giving technical specifications.
Alterations causing more than 15% proof charges will be charged to the author(s). 50 free
copies are supplied. Order form, quoting a special price for additional copies, accompanies the
page proof. Authors are urged to provide addresses for up to 20 journals which may receive
review copies.
Manuscripts not returned during the production of the book are not returned after printing.
Original photos and art work are returned when requested.

Manuscript
General. - Manuscripts and illustrations must comply with the details given above. The original
ms. and illustrations plus one clear copy of both should be sent to the undersigned editor.
NB: A ms. should not contain less than 32 printed pages. This applies also to the Mat. Fys.
Medd., where contributions to the history of science are welcome.
Language. - English is the preferred language. Danish, German and French mss. are accepted
and in special cases other languages. Where necessary, language revision must be carried out
before final acceptance.
Title. —Titles should be kept as short as possible and with an emphasis on words useful for
indexing and information retrieval.
 Three Studies in
Athenian Demography

By M O G E N S H E R M A N H A N S E N

Historisk-filosofiske M eddelelser 56
Det Kongelige Danske V idenskabernes Selskab
The Royal Danish Academy of Sciences and Letters

Commissioner: M unksgaard * C openhagen 1988

 Contents
I. Ephebic Inscriptions as Evidence for the Number of Athenian Citizens 336-22 3

II. A Note on the Growing Tendency to Underestimate the Population of Classical Attica 7

III. Athenian Population Losses 431-403. B.C. and the Number of Athenian Citizens in 431 B.C.

Synopsis
In recent years ancient historians have tended to underestimate the size of the population of
Attica in general and the number of Athenian citizens in particular. This trend is closely
connected with the new orthodoxy, that the Athenian economy was based on subsistence
agriculture and that the import of grain necessary to feed the population of Attica has been
much overrated. Against this prevailing tendency I advocate higher population figures. In
1985 I argued that the number of adult male citizens living in Attica in the 4th century B.C.
totalled some 30,000 rather than ca. 20,000 as often assumed.
The purpose of my three studies is to corroborate that conclusion.
In the first study I discuss a new roster ofephebes (nineteen-year-old recruits) of Kekro-
pis (one of the ten tribes) of ca. 332/1 which, compared with the other rosters we have,
indicates that a year class of ephebes in the 320s presumably numbered some 600 and that
many young Athenians, though fit for military service, did not serve as ephebes.
In the second study I investigate the population of Attica (citizens, metics and slaves
alike) and argue that, in the age of Demosthenes, Attica must have had a population of at
least 200,000-250,000 rather than the ca. 150,000 recently suggested.
The third study deals with the number of adult male Athenian citizens in the age of
Perikles. Around 400 B.C. there must have been at least some 25,000 citizens and, adding
up all the population losses suffered in the period 431-403, I conclude that there must have
been some 60,000 citizens in 432/1 B.C. and that this figure is perfectly compatible with
Thoukydides’ estimate of Athenian manpower at 2.13.6-8.
The higher population figures advocated in these studies help us to understand the
astonishing political participation in classical Athens (e.g. 6,000 citizens attending an
ordinary session of the people’s assembly) and the constant emphasis in our sources of how
much Athens depended on imported grain even in good years, not to speak of years of bad
crops.

MOGENS HERMAN HANSEN

lektor dr.phil.
Institut for klassisk Filologi
94, Njalsgade
DK-2300 Copenhagen S

© Det Kongelige Danske Videnskabernes Selskab 1988

Printed in Denmark by Special-Trykkeriet Viborg a-s
ISSN 0106-0481 ISBN 87-7304-189-0
HfM 56 3

I. Ephebic Inscriptions as Evidence for the N um ber

of Athenian Citizens 336-22

The recent publication of yet another roster of ephebes of Kekropis1is an

obvious occasion for me to pursue my studies of the number of citizens in
fourth-century Athens and to consolidate the views I stated in Demogra­
phy and Democracy.2 The new roster (of 332/1) combined with the informa­
tion we have about an unpublished roster of 333/23 and with a restudy of
the roster of 334/34 provides us with the numbers and demotics of almost
all the ephebes from Kekropis in the years 334/3, 333/2, and 332/1. The
evidence can be tabulated as follows:

Ephebes : 334/3 333/2 332/1 Total

Aixone (11?) 7 7 15 29
Melite (7) 11 4 9 24
Xypete (7) 2 5 3 10
Halai Aix. (6) 5 17 13 35
Athmonon (6?) ? 5 8 13+
Phlya (5?) ? 7 9 16+
Pithos (2) 2 6 4? 12
Sypalettos (2) p 0 1 1+
Trinemeia (2?) p 1 2 3+
Daidalidai (1) 1 0 0 1+
Epieikidai (1?) ? 0 1? 1+

Total 28 52 65? 145+

+ 14? = 42 = 159

(The numbers in brackets are the bouleutic quotas)

1 J.S . Traill, Demos and Trittys (Toronto 1986) 1-13: an ephebic inscription of Kekropis
from about 332/1 B.C.
2 M. H. Hansen, Demography and Democracy. The Number of Athenian Citizens in the Fourth
Century B.C. (Herning 1985) 47-50.
3 Cf. Hansen {supra n. 2) 109 note 237 and O. Reinmuth, The Ephebic Inscriptions of the
Fourth Century B.C. (Leiden 1971) no. 5.
4 In IG II2 1156 (= Reinmuth no. 2) the first 16 lines of Col. I are missing and the first
line of Col. II (with a demotic). The names lost in lines 1-16 plus the two broken names
4 HfM 56
This table prompts several observations on the nature of the ephebeia and
the use of ephebic numbers as evidence for the number of Athenian
citizens:
First, it is worth noting that the number of ephebes goes up from year
to year. Ca. 42 ephebes from Kekropis in 334/3 become 52 in 333/2 and
even 65 in 332/1. The two ephebic rosters from Leontis seem to reflect the
same development: in 333/2 the tribe had ca. 44 ephebes as against 62 in
324/3.3 Also, the rather small tribe Oineis had more ephebes in 330/29
than the probably larger tribe Erechtheis had in 333/2 (56? from Oineis
as against 48 from Erechtheis).6 The evidence we have for Pandionis, on
the other hand, is too fragmentary to be of any value.7 The ephebeia was
reshaped in 336/58 and one result of the reform may well have been an
increased participation so that the number of ephebes after a few years
was considerably higher than the number immediately after the reform.
Thus, as a rough guess we may assume that the total number of ephebes
may have been ca. 450-500 in the first years after the reform, but had
risen to some 600 or more a decade later. For demographic calculations it
is, of course, the higher figure that matters.
Second, now that we have rosters of ephebes from one tribe in three
consecutive years, the disproportion in several demes between the
number of ephebes and the bouleutic quota becomes even more appa­
rent. To illustrate my point I will discuss the figures for Xypete and
Halai Aixonides.
Xypete (with 2 ephebes in 334/3, 5 in 333/2 and 3 in 332/1) had seven
seats in the council of five hundred. There is no reason to suppose that
Xypete did not fill its quota.9 But in order to return, year in and year out,
7 citizens aged thirty or more, the nineteen-year-old Xypetaiones must,
on average, have numbered some 8-9,10 i.e. instead of the 2+5+3 = 10
ephebes actually attested in the three rosters, we should expect at least
25. The inference must be that many of the Xypetaiones who served in
the boule in the late 320s and onwards never served as ephebes, or, to put

preserved in lines 17-18 (= 16-17) cover the demes Athmonon, Phlya, Sypalettos,
Trinemeia and Epieikidai. There were undoubtedly some ephebes from the two fairly
large demes Athmonon and Phlya. If then one of the remaining three small demes was
unrepresented, the 18 lines missing will have contained 4 demotics (Athmonon, Phlya
and two small demes) and the names of 14 ephebes. Thus, the roster will have recorded
the names of42 ephebes.-The demotic missing at Col. 1.19 (= 18) must be Halaieis, cf.
Traill (supra n. 1) 8 note on line 18. The demotic missing at the top of Col. II must be
Meliteis, cf. Traill in Hesperia 38 (1969) 483 note on line 17.
HfM 56 5
it differently, that far from all citizens were enrolled as ephebes when
they came of age and had been inscribed in their deme. In addition to
those who were unfit for military service - at least 10% of all11 - there
must have been quite a few other young Athenians who became citizens
without having been through the ephebeia. A further conclusion is that it is
impossible to calculate the total number of citizens on the basis of the
number of ephebes. The population figures obtained by such calculations
will be much too small, and there is no way of telling whether the 500-600
ephebes constituted, say 50 % or 60 % or 80 % of a year class of Athe­
nians aged nineteen.
This conclusion, however, does not support the view that only citizens
belonging to the “hoplite class” were called up for ephebic service. Ca.
500-600 ephebes aged 19, plus ca. 10% unfit for military service, corres­
pond to some 16,500-20,000 adults aged 18-80+,12 and this figure is
incompatible both with the 9,000 hoplites assumed by some historians13
and with the ca. 14,500 assumed by others.14 Thus we can dismiss the
view that the ephebeia applied to citizens of hoplite status only. Further­
more, as has often been noted, Lykourgos says explicitly (1.76) and
Aristotle implies (Ath. Pol. 42.1 & 3) that the ephebeia, in principle, was
meant to be a training of all citizens.15 I do not doubt that citizens
5 Leontis 333/2: Reinmuth no. 9; 324/3: Reinmuth no. 15. The ephebes of 324/3 were
probably born in an intercalary year (344/3), cf. Hansen (supra n. 2) 48: but the
addition to the year of one month accounts for an increase of 4-5 ephebes only.
6 Oineis 330/29: Reinmuth no. 12; Erechtheis 333/2: Reinmuth no. 13, cf. SEG XXXI
162. On the relative size of the tribes cf. J. S. Traill, The Political Organization of Attica.
Hesperia Suppl. 14 (1975) 32.
7 Pandionis 333/2: Reinmuth no. 8. Pandionis 332/1: Reinmuth no. 10. On the fragmen­
tary state of the Pandionis rosters cf. Traill (supra n. 6) 32 with note 20.
8 Harp. s.v. Epikrates (= Lycurg. fr. V.3, Conomis).
9 Cf. Agora XV 31 lines 1-8 and Traill (supra n. 6) Table V II, Kekropis.
10 A life table of the model population adopted in Hansen (supra n. 2) 11-12 shows that
about 1/6 of the males die between 18 and 30, cf. A. J. Coale & P. Demeny, Regional
Model Life Tables and Stable Populations (Princeton 1966) 5.
11 Cf. Hansen (supra n. 2) 16-21, 49.
12 Cf. Hansen {supra n. 2) 12.
13 E.g. A. H . M. Jones, Athenian Democracy (Oxford 1957) 81; E. Ruschenbusch, ‘Zum letz-
ten Mai: Die Biirgerzahl Athens im 4. Jh. v. Chr.’, ZPE 54 (1984) 258.
14 E.g. A. W. Gomme, The Population o f Attica in the Fifth and Fourth Centuries B.C. (Oxford
1933) 8.
15 Cf. M. H. Hansen, Det Athenske Demokrati i 4. årh. f.Kr. 1. Staten, Folket, Forfatningen
(København 1978) 33-34, 88-92; E. Ruschenbusch, ‘Die soziale Herkunft der Epheben
um 333’, ZPE 35 (1979) 173-76.
6 HfM 56

belonging to the upper three census classes served as ephebes and as

councillors more frequently than the thetes, but the evidence we have
speaks strongly against the assumption that thetes were excluded from the
ephebeia.
The deme Halai Aixonides illustrates the reverse phenomenon. In
order to provide 6 councillors every year, the deme must have had an
average annual increment of 7-8 citizens coming of age and being in­
scribed as citizens. But the three years attested give a total of 35, and if
we assume that it took a few years for the ephebeia to become an estab­
lished institution, the sources indicate an average increase of more than
10 ephebes per year. Given that the ephebes constituted only a part of the
young citizens inscribed in the deme (cf. the discussion of Xypete supra)
the inference is that Halai must have been underrepresented in the coun­
cil of five hundred, and this assumption is corroborated by the reform of
307/6 when Halai was one of the six demes to have no less than 4 seats
added to its quota.16
A comparison between Halai and Xypete shows that, in the 330s, the
bouleutic quotas did no longer match the distribution of citizens over the
demes. All demes were (still) big enough to fill their quota (apart from a
few very small demes)17 but some demes had many more citizens than
required to fill the quota. Thus we can safely disregard all calculations
which attempt to estimate the number of citizens required to run the
boule, but pay no attention to the fact that some demes were “overpopu­
lated” in the sense that they had more citizens than necessary to fill the
quota.18 But this fact, now corroborated by the ephebic lists, is one of the
necessary conditions for the reform of 307/6 whereby the council of five
hundred was changed into a council of six hundred.

16 Cf. Traill (supra n. 6) 59.

17 Cf. Traill (supra n. 6) 58.
18 E.g. R. Osborne, Demos: The Discovery o f Classical Attika (Cambridge 1985) 43-44; E. Rus-
chenbusch, ‘Die soziale Zusammensetzung des Rates der 500 in Athen im 4. J H ’. ZPE
35 (1 9 7 9 ) 177-80.
HfM 56 7

II. A Note on the Growing Tendency to

U nderestim ate the Population of Classical Attica

A recent trend among British historians studying ancient Athens is to

assert that agriculture was the basis of the economy even in classical
Athens, that the typical Athenian citizen was the subsistence farmer, and
that the Athenians’ dependency on imported grain has been much over­
rated. The trend is exemplified by Peter Garnsey in his impressive article
‘Grain for Athens’ and, even more extremely, by Robin Osborne in his
two pioneer monographs: Demos and Classical Landscape with Figures}
The new approach is based on two lines of argument which both
question traditional views about ancient agriculture and ancient demo­
graphy. On the one hand it is argued that historians have underrated the
agricultural productivity of Attica but overrated a man’s annual con­
sumption of grain. Thus, Attica had the capacity to support a much
larger population than previously believed. On the other hand it has
become fashionable to accept much lower population figures (for citizens,
metics and slaves alike) than most historians did one or two generations
ago. The combined effect of these two factors leads to the view that in
normal years the Athenians’ demand for imported grain must have been
less urgent than suggested in almost all descriptions of the Athenian
economy. In Osborne’s opinion, Attica had the capacity to support a
population of around 150,000, and in the classical period it did have a
total population of around 150,000 (Landscape 46). The inference is (Land­
scape 99) that Attica could have been farmed in such a way as to feed the
whole population and that there may even have been a few years in the
fourth century when Athens did not need imported grain.
The revised view of the nutritive value of ancient agricultural products
seems well founded,2 but the new assessments of ancient agricultural

1 P. Garnsey, Grain for Athens, ‘Crux. Essays Presented to G. E. M. de Ste Croix on his 75th
Birthday’ (Exeter 1985) 62-75; R. Osborne, Demos: The Discovery o f Classical Attika (Cam­
bridge 1985); Idem, Classical Landscape with Figures (London 1987).
2 H. Forbes & L. Foxhall, ‘Sitometreia: the Role of Grain as Stable Food in Classical
Antiquity’, Chiron 12 (1982) 41-92.
8 HfM 56
productivity are questionable.3 For the sake of argument, however, I will
accept them in this article. Thus, I accept the more optimistic estimate of
per capita consumption advanced by Forbes and Foxhall as well as e.g.
Sandars’ views about a higher cereal production per hectare even in the
2nd millenium B.C.4 What I shall object to in this study is the growing
propensity to suggest impossibly low population figures whereby, in good
years, the need for imported grain almost disappears and the subsistence
farmer can be presented as the typical Athenian citizen. In this article I
will focus on Osborne’s two recent monographs and question his estimate
of the population of classical Attica as stated in Demos 42-46 and in
Landscape 46 where he offers the following figures: “the ancient literary
texts give enough information to enable an estimate of citizen numbers in
the fifth and fourth centuries to be made with some confidence: citizen
families probably accounted for some 60,000-80,000 people. The popula­
tion of resident foreigners fluctuated considerably, but probably never
exceeded about 20,000. The difficult calculation is that of the number of
slaves. Scholars have disagreed radically about slave numbers in Athens,
estimating anything from 20,000 upwards. What is interesting in the
context of the question of food supply is that on a conservative, but not
unreasonable, estimate of 50,000 slaves, Attica will have had a total
population of around 150,000.”
Osborne, concerned with agricultural productivity and food supply,
accordingly prefers population figures which comprise both sexes and all
ages. But the information to which he refers reports the number of adult
males only. Thus a discussion of his population figures must be opened
with a discussion of the model population on which he has based his
calculations.
Following Hopkins,5 Osborne (Demos 43) adopts a life expectancy of
ca. 25 years (as I do in D&D 11-12).6 On growth rate, however, Osborne
is elusive and his account muddled. When he offers his own calculations
(Demos 45) he adopts a percentage which corresponds to growth rate 0
(33,000 males of whom 20-21,000 would be over 18, i.e. 61-64% of all
males). But in n. 91 Osborne criticizes Patterson7 (44-45) for not allow­
ing an annual growth rate of more than 1 %, and he refers approvingly to
Snodgrass who in his inaugural lecture8 suggested a growth rate of ca.
4% in the archaic period. Now, adopting a life expectancy of ca. 25
years, adult males (18-80+) constitute ca. 61 % of all males if the popula­
tion is stationary, ca. 54 % if the annual growth rate is 1 %, and ca. 33 %
if it is 4 % per year.9 T hus, if we apply Snodgrass’ model to 4th century
HfM 56 9
Athens, 33,000 males would correspond to no more than ca. 11,000
citizens. I much appreciate Snodgrass’ views on population growth in the
archaic period, but his quantifications are impossible. A natural popula­
tion growth of 4 % per year is more than for Mexico in the 20th century.
By letter (of February 1986) Snodgrass has informed me that he accepts
the views on population grow th advanced in D&D (i.e. max. ca. 1 % and
0.5 % for rough calculations). Thus 4 % is out of the question, and even
2 % is far too rapid a growth, as Osborne seems to admit10 in JHS by
concurring with my criticism of Ruschenbusch, D&D 12-13. In conclu­
sion, Osborne and I agree on life expectancy (ca. 25 years) and we also
agree on population growth, if we adopt Osborne’s own calculations and
disregard his n. 91. Osborne’s calculations are in fact slightly more
pessimistic than mine (growth rate 0 rather than 0.5%) but this differ­
ence is insignificant, since we can make only very rough estimates. In the
following I will use the model population presented in D&D 11-12,
namely: Coale and Demeny, Model West, mortality level 4, growth rate
0.5%.
After this introduction on which model population to use for calcula­
tions of ancient Greek populations I turn to a discussion of Osborne’s
figures: 60,000-80,000 Athenians, max. 20,000 metics and ca. 50,000
slaves.
First the citizens. Ca. 60,000-80,000 Athenians (of both sexes) corres­
pond to ca. 30,000-40,000 males, of whom ca. 17,250-23,000 will have
been adult citizens aged 18-80+ (constituting 57.5 % of all males). But in
his review of D&D in JHS Osborne accepts that the sources relating to
fourth-century Athens point to a total of ca. 30,000 adult male citizens

3 J. E. Skydsgaard, ‘Transhumance in Ancient Greece’, PCPhS Suppl. 1988. Idem, ‘L ’ag-

ricoltura greca et romana: tradizioni a confronto’, Analecta Romana lnstituti Daniel 16 (1988) 7-
24.
4 G. Sandars, ‘Reassessing Ancient Populations’, BSA 79 (1984) 251-62.
5 M. K. Hopkins, ‘The Probable Age Structure of the Roman Population’, Population
Studies 20 (1966) 245-64.
6 M. H. Hansen, Demography & Democracy. The Number o f Athenian Citizens in the Fourth
Century B.C. (Herning 1985). = D&D
7 C. Patterson, Pericles’ Citizenship Law o f451-50 B.C. (New York 1981) 44-45.
8 A. M. Snodgrass, ‘Archaeology and the Rise of the Greek State’, Inaugural Lecture.
(Cambridge 1977) 13.
9 A. J. Coale & P. Demeny, Regional Model Life Tables and Stable Populations (Princeton
1966) 128.
10 R. Osborne, Review of M. H. Hansen: Demography and Democracy in JHS 107 (1987) 233.

2 T h re e S tu d ie s in A th e n ia n D e m o g ra p h y
10 HfM 56
living in Attica, and for the Periclean period all calculations (based on
Thuc. 2.13.6-8) point to a minimum of ca. 40,000 citizens and may allow
a maximum of 60,000 or even more; cf. my third study page 26. Now,
30.000- 40,000 adult male citizens correspond to some 52,000-69,000 male
Athenians of all ages and to a total population, not of 60,000-80,000, but
of at least 100,000-140,000 Athenians.
Next the metics. By contrast with the almost stationary citizen popula­
tion metic numbers must have fluctuated considerably according to the
economic situation (cf. Garnsey 79 with note 26). Peace and prosperity
may have caused the number of Athenian metics to grow rapidly in a few
years and, conversely, in a protracted economic crisis their numbers may
have dwindled again in less than a decade (cf. e.g. Xen. Vect. 2.1-7). On
the other hand, there can be no doubt that a maximum of ca. 20,000 is far
too pessimistic an estimate. When Demetrios of Phaleron conducted his
population census in a year around 315 he counted 21,000 Athenians and
10.000 metics living in Attica (Ath. Deipn. 272C). If both figures com­
prise adult males only (as they probably do cf. D&D 31-34) and if we use
the same method of calculation for metics as for citizens, there must have
been some 35,000 metics living in Attica when Demetrios conducted his
census, viz-, 10,000-^57.5X 100 = ca. 17,500 males of all ages X 2 = ca.
35.000 metics of both sexes and all ages). But, following Whitehead,11
one might object that women and children constituted a much smaller
fraction of metics than of citizens, since many metics may have stayed in
Attica for a few years only and without their family. Thus, a total of, say,
20.000- 25,000 is much likelier than my total of ca. 35,000. I have two
answers to Whitehead’s objections: first, Demetrios ascertained only the
number of metics living in Attica (D&D 31-32), i.e. short-term metics
(most of whom were males) are not included in the 10,000 actually
counted. Second, the tombstones set over metics (recording ethnics in­
stead of demotics) show a higher proportion of women than the tombsto­
nes commemorating citizens. The ratio is 2/5 metic women as against 1/3
female citizens. The resident metics seem to have had families just like
the Athenians, and ca. 35,000 in a year around 315 is not an inflated
figure but rather a minimum. First, short-term metics must be added to
the 10,000 (adult male) metics settled in Attica, and second the figure
10.000 probably included only metics of military age and fit for military
service. Thus, adult male metics must have numbered, not 10,000, but
rather some 12-14,000.
Demetrios’ population census, however, may have been conducted in a
 HfM 56 11
period when Attica was crammed with metics; we do not know. The only
source of information we have about the average number of metics in
fourth-century Athens is the tombstones mentioned above. They are an
important but neglected source of information for Attic demography.
The preserved private funerary monuments dated to the fourth century
B.C. record the names of some 1,800 citizens as against some 650 me­
tics.12 A priori one would assume that citizens put up funerary monu­
ments more frequently than metics. Furthermore, the tombstones give
evidence only of metics who resided and were buried in Attica. In order
to find the total number of metics we must add an unknown number of
short-term metics. Thus, the epigraphical evidence indicates that the
regular ratio between citizens and metics was at least 3:1, and according­
ly some 100,000-140,000 Athenians correspond to a population of some
33,000-46,000 resident metics. Summing up, both the population census
conducted ca. 315 and the tombstones point to a much higher number of
metics than the max. 20,000 estimated by Osborne.
Third the slaves. In the ancient Greek city states slaves were never
counted since they were neither taxable nor liable to military service. The
Athenians themselves did not know the number of slaves in Attica and
we shall never come to know the number either, no matter how many and
how valuable sources we may still recover. On the other hand, the Greeks
never refrained from producing rough estimates of the number of slaves.
We hear about 460,000 slaves in Corinth (Timaios, FGrHist 566 fr. 5),
470,000 slaves in Aigina (Arist. fr. 472, Rose) and 400,000 slaves in
Attica (Ktesikles, FGrHist 245 fr. 1) a figure which matches Hypereides’
estimate of 150,000 (adult male) slaves (Hyp. fr. 33, Sauppe). Several of
these estimates come from good sources but cannot, of course, be trusted.
For example, for Aigina to have 470,000 slaves, it must have had a
population density of about 5,500 per square kilometre of slaves alone!
So the numbers 400,000 slaves and 150,000 adult male slaves in Attica
are pure guesswork. They tell us nothing about how many slaves there
actually were in Attica but only how many there were presumed to have
been. Have these estimates then any value at all? This question is best

11 D. Whitehead, The Ideology o f the Athenian Metic (Cambridge 1977) 97-98 with note 185.
12 At Copenhagen University, Institute of Classics a group consisting of Lars Bjertrup,
Mogens Herman Hansen, Thomas Heine Nielsen, Lene Rubinstein and Torben Vester-
gaard has built up a data base of all Athenians recorded on private funerary monuments
ca. 400 B .C .-ca . 3 0 0 A .D .

2»
12 HfM 56
answered by a comparison with a still more fantastic figure from the
history of Greece, viz., Herodotos’ report that the Persian army was
1,700,000 strong. This is an impossible figure, but there is no reason to
doubt that Herodotos and many of his audience believed it. The infer­
ence is that the Greeks in general and Herodotos in particular did not
know the size of the Persian army; nor did they have a very secure grasp
of numerical magnitudes. On the other hand, all historians agree that
one may assume from Herodotos’ figure that Xerxes’ army was larger
than that of the Greeks. The figures 400,000 and 150,000 are equally
impossible but reveal the Athenians’ own estimate that there were more
slaves than free men, and there is no reason to doubt this impression. We
shall never know the number of slaves in Attica, but we may conclude
that if the number of free is set at X, then the number of slaves must be >
X .13 We must, however, allow for an important modification of this very
rough general assumption: like the number of metics, the number of
slaves must have fluctuated considerably. Many Athenian slaves were
bought and not bred. In a protracted crisis like the famine during the
320s, when, for example, a mining slave barely produced the amount of
silver he cost his master (cf. Dem. 42.20-21, Isager & Hansen 200), slaves
were undoubtedly not replaced and perhaps even allowed to run away
(cf. Thuc. 7.27.5). Given the short life expectancy in the ancient world,
the non-replacement of slaves will, in less than a decade, have had an
enormous effect on the total number of slaves living in Attica (cf. Xen.
Vect. 4.25). Summing up, for the sake of argument and to be on the safe
side, I will assume that if the number of free is set at X, the number of
slaves must be at least / 2 X. As argued above, the total free population
numbered some 133,000-186,000. Accordingly there must have been at
least some 66,000-93,000 slaves in Attica, during a protracted crisis
perhaps less than 66,000 and in a prosperous period probably more than
93,000.
Adding up citizens, metics and slaves, I conclude that in the fourth
century the number of persons living in Attica must have amounted to at
least some 200,000 and more probably 250,000, whereas in the Periclean
period the population of Attica must have totalled 300,000 or more.
These much higher population figures show that self-sufficiency was
impossible even in good years, and instead of Osborne’s rather extreme
position I prefer the more cautious view stated by Garnsey:14 “My own
(very tentative) calculations suggest that Athens in the fourth century
had to find grain for perhaps one-half of its resident population from
HfM 56 13
outside Attica, narrowly defined, in a normal year.” If about one-half of
the population in a normal year, and a much higher proportion in a bad
year had to buy their daily provisions in the market, it does not make
sense to speak about a subsistence economy. Instead of a territory which
in good years was almost self-sufficient, we must imagine an Attica which
in all years depended heavily on the import of grain and in bad years very
much so. And the typical Athenian citizen was not the subsistence farmer
who tended to stay in his deme of origin, but was rather like Chremes
whom Aristophanes describes as a citizen who attends the meetings of
the ekklesia and, on other occasions, carries his wine to the market in
order to buy flour (Ar. Eccl. 376ff, 815-22). I do not quarrel with “the new
orthodoxy which stresses the cellular self-sufficiency of the ancient
economy”,15 but Osborne seems to forget that classical Athens was the
exception and not the rule (ibidem). His account of classical Attica is
stimulating and often convincing, but there are some serious distortions
in his Classical Landscape with Figures. One of them is that there are too few
figures in the landscape.

13 S.Isager & M. H. Hansen, Aspects o f Athenian Society (Odense 1975) 14-16.

14 Cf. Garnsey (supra n. 1) 74.
15 Cf. Trade in the Ancient Economy ed. P. Garnsey, K. Hopkins and C. R. Whittaker (Cam­
bridge 1983) xi.
14 HfM 56

III. Athenian Population Losses 431-403 B.C. and

the N um ber of A thenian Citizens in 431 B.C

Historians who estimate the number of Athenian citizens in 431 base

their calculations on Thoukydides’ information about Athenian man­
power at the outbreak of the Peloponnesian war (2.13.6-8) supplemented
with the army strengths reported by him for e.g. the Megara campaign in
431 (2.31.2) and the battle of Delion in 424 (4.93.4-94.1). Very little
attention, on the other hand, is paid to all the information we have about
Athenian losses during the Peloponnesian war.1 Even Gomme, who de­
voted several studies to Thuc. 2.13.6-8, never gave a comprehensive
account of the population losses which Athens suffered in the period 431 -
403.2 Nevertheless, anyone who reads the eight books of Thoukydides
and the first two books of Xenophon’s Hellenika is astonished when on
every second or third page he learns about Athenian casualties, some of
them not too severe, but others almost catastrophic; and I suspect that
many readers have asked themselves the question I will discuss in this
study: are the rather low population figures accepted by most historians
compatible with the losses sustained by the Athenians during the
Peloponnesian war? In order to answer this question I will first attempt
to estimate the number of Athenians who succumbed to the plague, or
were killed in action, or died in the Sicilian quarries, or were starved to
death during the final siege of Athens.
1. According to Thuc. 3.87.3 plague mortality totalled no less than
4,400 hoplites, 300 knights and an unknown number of other Athenians.
Thoukydides states explicitly that by hoplites he means front-line troops
(ex xd)V xa^EUiv) and not citizens “of hoplite status”.3 Thus mortality
amounted to 4,700 out of 14,000 (13,000 hoplites + 1,000 knights cf.
Thuc. 2.13.6-8), i.e. about a third of all. There is no reason to suppose
that the plague affected hoplites and knights more severely than other
social groups. Accordingly, if we accept Gomme’s total of ca. 47,000
adult male Athenian citizens in 431, the plague which descended upon
Athens in 430/29, 429/8 and 427/6 must have caused the death of some
15,000 adult male Athenians.
2. T h e other m ajor disaster which struck A thens in the Peloponnesian
HfM 56 15
war was the annihilation of the forces sent to Sicily in 415-413. The losses
cannot be ascertained with any exactitude, and all I can present here is a
rough calculation of what I take to be a minimum. First the army. The
original Athenian contingent consisted of 1,500 hoplites and 700 thetes
(Thuc. 6.43.1). Later reinforcements included first 250 hippeis and 30
hippotoxotai (6.94.4); later again 1,200 more hoplites (7.20.2). Very few
escaped the disaster (7.87.6) and the total loss must have been some
3,000 Athenians. It is much more difficult to assess the number of Athe­
nians serving in the fleet. Of the first fleet, 100 triremes were Athenian
(6.43.1). Later on the Athenians sent a squadron of 10 Athenian triremes
commanded by Eurymedon (7.16.2) and a squadron of 60 Athenian
triremes under Demosthenes’ command (7.20.2). But of the 60 ships 10
triremes under Konon were sent to Naupaktos (7.31.5). Thus of the 73
triremes which arrived in Syracuse (7.42.1), 51 were Athenian, i.e. 50
sent with Demosthenes and one commanded by Eurymedon who had
joined Demosthenes with one of his 10 triremes. Consequently the Athe­
nian fleet lost in Sicily totalled 160 triremes (= 100+9+51). I exclude the
allied contingent which is of no consequence for my investigation. The
crew of a trireme was composed of 200 men, but the ships were often
undermanned.4 On the other hand, it is apparent from Thoukydides’

1 This article is a continuation of my three previous studies in Athenian demography:

‘The number of Athenian Hoplites in 431 B.C.’, SymbOslo 56 (1981) 19-32; ‘Demogra­
phic Reflections on the Number of Athenian Citizens 451-309 B.C.’, AJAH 7 (1982) 172-
89; Demography and Democracy (Herning 1985). The only serious attempt to add up all the
Athenian casualties during the Peloponnesian war is Barry Strauss’ seminal study:
Athens after the Peloponnesian War (New York 1986) 179-82; but his account does not
include all the evidence we have and his total (5,500 hoplite and 12,500 thetic casual­
ties) is, in my opinion, somewhat too low, cf. infra.
2 A. W. Gomme, ‘The Athenian Hoplite Force in 431 B.C.’, CQ 21 (1927) 142-50; The
Population o f Athens in the Fifth and Fourth Centuries B.C. (Oxford 1933); A Historical
Commentary on Thucydides II (Oxford 1956) 33-39; ‘The Population of Athens Again’, JHS
79 (1959) 61-68.
3 Cf. Gomme {supra n. 2) (1927) 149; (1933) 6; (1956) 388; (1959) 63-64. According to
A. H. M. Jones, Athenian Democracy (Oxford 1957) 165-66 the reference is to all hoplites of
military age. Strauss {supra n. 1) 75-76 discusses the meaning of taxis and, in my opinion
correctly, he sides with Gomme against Jones.
4M .A m it, Athens and the Sea. A Study in Athenian Sea-Power (Brussells 1965) 39-49;
J. S. Morrison, ‘Hyperesia in Naval Contexts in the Fifth and Fourth Centuries B.C.’,
JHS 104 (1984) 48-59; Hansen {supra n. 1) (1985) 22-24. On the crews of transport
vessels and the often undermanned squadrons cf. the important article by H. T. Wal­
ling, ’The Trireme and its Crew’, Studies... H.L. W. Nelson (Utrecht 1982) 463-82.
16 HfM 56
account that this fleet was the best manned and best equipped ever
launched (6.31.1) and that Athenian citizens must have constituted a
substantial part of the crews (e.g. 6.24.3; 31.3; 7.61.1, 3; 64.1; 8.1.2. etc.).
Of the first 100 triremes, however, 40 were transport vessels (6.43.1) and
so were some of the 60 triremes sent under Demosthenes (7.20.2). A
transport vessel probably had a crew of some 60-80 men only (cf. IG II2
1628.154-61). On the cautious assumption that, on average only 50 men
of a ship’s crew were Athenians, the total number of Athenians serving in
the fleet must have been ca. 8,000; and almost all were killed. Thus
estimating the total number of citizens lost in the Sicilian campaign, I
suggest that 10,000 is an absolute minimum.
3. Apart from the Sicilian disaster, the Athenians sustained several
other severe losses during the nineteen years of war from 431 to 421 and
again from 413 to 404. The first major defeat inflicted upon the Athe­
nians was at Spartolos in 429/8. Thoukydides estimates the losses at 430
citizens (2.79.7). At Delion in 424 the Athenians lost almost 1,000 citi­
zens (Thuc. 4.101.2) and adding those killed at Sikyon (4.101.4) we can
safely assume that casualties totalled more than 1,000 Athenians. The
engagement outside Amphipolis in 422 cost the Athenians 600 citizens
(Thuc. 5.11.2). The Sicilian disaster has been discussed above and I turn
to some of the more severe casualties in the Decelean war. In 411/0 the
Athenians lost 22 triremes off Eretria; of the men some were killed and
some taken prisoners (Thuc. 8.95.7). A loss of 500 citizens is, in my
opinion, a cautious guess. Next year (410/09) 100 hoplites and 300 other
Athenians were killed near Ephesos (Xen. Hell. 1.2.9). In 406 the Athe­
nians won the sea-battle of Arginoussai but suffered heavily. Because of
the storm 25 triremes were lost with all hands (Xen. Hell. 1.6.34). To
man the fleet the Athenians had mobilized both slaves and free, both
foreigners and citizens; even some of the knights had been drafted
(1.6.24). The number of citizens serving in the fleet must have been
considerable (cf. 1.7.8), and the loss of 1,000 citizens is, in my opinion,
too low rather than too high an estimate. A year later the Athenians
suffered the crushing defeat at Aigospotamoi. They were taken by sur­
prise and losses due to fighting may have been slight, but after the battle
Lysandros executed all the Athenians he had captured. According to
Plutarch, no less than 3,000 citizens were killed {Lys. 13.1; Aik. 37.4);
according to Pausanias (9.32.9) the total was 4,000. To sum up: in
addition to the 10,000 citizens lost in Sicily, the casualties in major
engagements must have totalled at least 7,000 citizens.
HfM 56 17
4. Casualties in major battles, however, do not exhaust the Athenian
losses. First, there were innumerable other battles, skirmishes and raids.
Occasionally the losses are indicated by our source. Thus 120 Athenian
hoplites were killed in a battle in Aitolia in 426 (Thuc. 3.98.4). In the
following year 50 Athenians were killed at Solygeia (Thuc. 4.44.6). At
Mantinea in 418 Athenian losses were light, i.e. no more than 200 citi­
zens (Thuc. 5.74.3).
Second, most of the major battles listed above were Athenian defeats.
When the Athenians won, losses were usually much slighter,3 but some
citizens must have lost their lives in the numerous successful campaigns,
e.g. the siege of Poteidaia 432-29, the subjugation of Mytilene in 428/7,
the Pylos campaign in 425, the attack on Megara in 424, the reconquest
of Torone in 422 and of Skione in 421, etc.
Third, in descriptions of sea-battles the historians tend to record only
the number of ships lost; only occasionally are we told that the men
escaped or were drowned or captured or killed. In most cases we have no
guide-lines on the number of men lost in a sea-battle. In several cases we
do not even have any information about the number of ships lost. The
battle of Arginoussai is exceptional in that we know that 25 triremes were
lost with all hands. A further complication is that we never know whether
Athenian citizens constituted a quarter or a half or an even larger propor­
tion of the crew of a trireme. Let me illustrate the problems by listing the
naval engagements in 412-410 reported in Thuc. 8 and in Xen Hell. 1.1.
(a) In a surprise attack from Speiraion a Peloponnesian squadron cap­
tured four Athenian triremes (Thuc. 8.20.1). (b) Three triremes were lost
off Chios. Of the men some were taken prisoner, some were killed and
some escaped (8.34.1). (c) The Athenians lost six triremes in a battle
near Syme (8.42.4). (d) An Athenian squadron of 32 ships was defeated
by the Chians (8.61.3). (e) In a battle near Eretria the Peloponnesians
captured twenty-two Athenian triremes and either killed or made prison­
ers of the crew (8.95.7). (f) The battle of Kynossema was an Athenian
victory, yet the Athenians lost fifteen ships (8.106.3). (g) Tn the Helles­
pont the first naval battle was a Peloponnesian victory (Xen. Hell. 1.1.1).
(h) Shortly afterwards, however, the Athenians won a major victory off
Abydos (1.1.7). (i) In the subsequent battle near Kyzikos the Athenians
annihilated the remaining Peloponnesian fleet (1.1.18). In my list above,

5 P. K r e n tz , ‘C a s u a ltie s in H o p lite B a ttle s ’, G R B S 26 (1 9 8 5 ) 13-20.

18 HfM 56
I have included only the 22 triremes lost at Eretria. All the other losses
are passed over in silence.
Fourth, there must have been many other battles or campaigns not
even mentioned by the sources we have. The best illustration of Athenian
losses that have not left any mark in the literary sources is a preserved
casualty list, which probably records Athenians killed in battle in 409
(Agora XVII 23). As reconstructed by Bradeen, the funerary monument
must have contained between 900 and 1,400 names. But the only Athe­
nian casualties in 409 reported by Xenophon are the 400 men lost at
Ephesos (Xen. Hell. 1.2.9 cf. supra). Thus in 409 some 500-900 men must
have been killed in other battles of which we have no information at all.
The discrepancy between actual casualties and casualties known to us
cannot have been of the same magnitude in all years. But if we list all the
battles, skirmishes, sieges and raids which the historians report without
mentioning the number of men killed, and if we keep in mind that many
other military operations are passed over in silence, we must assume
that, apart from the min. 7,000 Athenians killed in major battles, the
number of Athenians killed in action every year must have totalled sever­
al hundred. The Archidamian war lasted ten years, and the Decelean
war nine more years. To be on the safe side, I will assume a loss of only
200 Athenians per year (but 500 in 410/09, cf. the casualty list discussed
above) i.e. a total of some 4,000 citizens during the entire war.
5. Aigospotamoi was the last battle in the Peloponnesian war. But
further losses were inflicted upon the Athenians in the two following
years. In the winter of 405-404 Athens was besieged by Lysandros and
starved into surrender. According to the description given by Xenophon
(Hell. 2.2.11, 21; 2.3.41), Diodoros (13.107.4) andjustinus (5.8.1-3) citi­
zens must have died by the thousand and most Athenians would have
died if the siege had lasted much longer. No figure is stated in any source,
but some 3,000 adult males is probably a cautious guess. Furthermore,
1,500 Athenians were killed by the Thirty during the eight months they
ruled Athens (Arist. Ath. Pol. 35.4; Isoc. 20.11; 7.67; Aeschin. 3.235).
Many of these may have been Eleusinians (Xen. Hell. 2.4.8). Finally, in
the battle between the Spartans under Pausanias and the democrats
under Thrasyboulos first 30 and then 150 democrats were killed (Xen.
Hell. 2.4.34). To sum up, after the defeat at Aigospotamoi and until
democracy was restored in the autumn of 403, some 5,000 more Athe­
nians were starved to death, executed by the oligarchs or killed in the
civil war.
HfM 56 19
6. The number of Athenian citizens was reduced not only by war,
plague and famine, but also by emigration. During the Peloponnesian
war the Athenians sent out a considerable number of citizens either as
klerouchs or as colonists.6 The best attested klerouchy is Mytilene which,
after its surrender in 428/7, was settled with 2,700 Athenians (Thuc.
3.50.2). But klerouchs retained their Athenian citizenship. Some of the
Lesbian klerouchs may even have stayed in Athens as absentee land­
lords, and in any case, the Mytilenaians probably got their territory back
a few years later (Tod 63). To be on the safe side, I will disregard Lesbos
and assume that, eventually, the Mytilenaian klerouchs did not in the
long run diminish the total number of Athenian citizens living in Attica.
The Athenian colonists (epoikoi), on the other hand, were probably emi­
grants who left Attica for good and lost their citizen rights. An unknown
number of colonists were sent to Aigina in 431/0 (Thuc. 2.27.1; 8.69.3).
When Poteidaia was conquered in 430/29 the city with its territory was
handed over to 1,000 colonists (IG I2397; Thuc. 2.70.4; Diod. 12.46.7);
and the andrapodismos of Melos in 416/5 was followed by the sending out
of 500 Athenian settlers (Thuc. 5.116.4). Judging from the 1,000 colonists
to Poteidaia and the 500 to Melos I assume that 500 colonists sent to
Aigina in 431/0 is a very cautious estimate. In the same period the
Athenians probably founded other colonies as well,' but we have no
information. Thus, during the Peloponnesian war, 2,000 colonists sent
out from Athens with their families must be the minimum loss of popula­
tion due to emigration.
7. All the losses enumerated above add up to no less than 43,000
Athenian citizens. But these losses were sustained over a period of twen­
ty-eight years. Thus they were balanced by the number of young Athe­
nians who in the period 431-404 were inscribed in the demes. But from
this number we must subtract the “normal” mortality caused by factors
other than plague, warfare and emigration. Population changes during
the Peloponnesian war must be calculated by the year, and I suggest the
following method. For the sake of argument, I use Gomme’s figure as my
starting point, i.e. 47,000 adult male citizens in 431.8 If life expectancy at

6Jones (supra n. 3) 167-77.

7 Ph. Gauthier, ‘A propos des clérouquies athéniennes du Ve siècle’, Problèmes de la terre en
Grèce ancienne ed. M. I. Finley {Paris 1973) 163-78. Gauthier shows that colonists appa­
rently were more important than believed by most historians, who tend to concentrate
on klerouchies.
8 Gomme (1933) 26 estimates a total of 43,000 citizens aged 18-59, to which figure we
20 HfM 56
birth was ca. 25 years, and if the natural population growth was ca. 0.5-
1 % per year,9 males aged eighteen must have constituted some 3.3 % of
all adult males. Thus, the number of new citizens inscribed in 431/0 must
have totalled ca. 1,550. But in the years 430/29, 429/8 and 427/6 the
severe plague mortality must have affected boys as well as men, thus
diminishing the eighteen following year-classes to be inscribed as citi­
zens. Furthermore, many females of marriageable age must have died
from the plague, thus diminishing the number of births and, consequent­
ly, the number of young citizens to be inscribed from 412/1 onwards. On
the other hand, the severe losses in Sicily in 413 and the casualties during
the Decelean war will not have had any effect on the number of young
Athenians to be inscribed as citizens in the years 413-403. The demogra­
phic repercussions of the defeats which Athens suffered in the years 413-
403 will not have left their mark on the number of citizens coming of age
until the decade 395-385.
Summing up. The three major disasters were the plague (430-26), the
defeat in Sicily (413) and the defeat of Aigospotamoi followed by the
siege of Athens (405-404). For the period down to the Sicilian disaster I
have calculated the eighteen-year-olds year by year and rated a year class
at 3.3 % of all adult male Athenians alive in the preceding year. For the
period 412-403 on the other hand, I have taken a year class to be 3.3 % of
the year class of 413/2. This is undoubtedly too optimistic for the year
404/3 since the famine will have affected boys as well as men and reduced
the number of Athenians coming of age in this and the following years.
As to mortality, I have assumed that, in any year, 2.5% of all adult
males died “a normal death” i.e. their death was caused by anything but
war and plague. Thus, the annual growth rate is estimated at 3.3—2.5 =
0.8% , which is optimistic, but not impossible. It is worth remembering
what Thoukydides writes about Athens in 415: “the city had just reco­
vered from the plague and the years of continuous war, and a great many
of the young had grown to manhood” (6.26.2). In the three years of
plague, a “normal mortality” of 2.5 % per year is probably too pessimis­
tic since, according to Thoukydides, “during all this time there was no
serious outbreak of any of the usual kinds of illness” (2.51.1). A mortality
in this short period of 1.5-2.0%, however, would not affect my calcula­
tions significantly.
After this discussion of the various factors involved, the calculation of
must add an estimate of the number of citizens above 60, probably some 9-10 % of the
citizens aged 18-59, i.e. at least 4,000 citizens, cf. the following note.
 HfM 56 21
the Athenian citizen population during the Peloponnesian war is as
follows. First (a) I record mortality due to major disasters and defeats
reported in our sources;10 next (b) an average number of casualties
suffered in unknown campaigns and in battles for which our sources do
not report the losses. In most years the number is set at 200; in 410/9,
however, I estimate these losses at 500 on the basis of the presewed
fragments of the year’s casualty list. Third (c) I give the number of
Athenians sent out as colonists, and finally (d) I estimate the mortality
due to other reasons than war and plague at 2.5 % of all citizens alive in
the preceding year. The total mortality (E = a + b + c+ d ) is then counter­
balanced by the number of Athenians coming of age (F) and the resulting
total number of citizens is recorded in (G). The starting point is, as stated

9 The model population I use here and elsewhere is: A.J.Coale and P. Demeny, Regional
Model Life Tables and Stable Populations (Princeton 1966), Model West mortality level 4
(life expectancy at birth: 25.26 years) and growth rate 5.00 (an annual increase of
14%), cf Hansen (supra n. 1) (1985) 9-13. In this model the age distribution of adult
males is as follows:

Age % of all males % of all males 18-80+

18-19 3.85 6.7

18-59 52.47 91.3
18-80+ 57.47 100.0
19 1.92 3.3
20-39 31.44 54.7
20-44 37.05 64.5
20-49 41.77 72.7
20-59 48.62 84.6
30 1.57 2.7
40 1.21 2.1
40-49 10.33 18.0
50 0.86 1.5
50-59 6.85 11.9
59 0.52 0.9
60-80+ 5.00 8.7

Let me add that I might equally well have used, for example, mortality level 3 or growth
rate 0, or more cautiously, I might have calculated the age distributions for mortality
levels 2, 3, 4 and 5 at growth rates 0, 5.00 and 10.00, model west and model south. But
for the present purpose it makes little difference, so I restrict myself to tabulating the age
distribution for mortality level 4, growth rate 5.00, model west.
10 The plague mortality is estimated at Vs of 47,000 = 15,666 and this loss has been
distributed evenly over the three years of plague. Thus I print 5,200 in each of the years
430/29, 429/8 and 427/6.

3 T h re e S tu d ie s in A th e n ia n D e m o g ra p h y
22 HfM 56
above, the 47,000 adult male citizens calculated by Gomme on the basis
of Thuc. 2.13.6-8.

Year (a) (b) (c) (d) (E) (E) (G)

431/0 200 500 1,175 -1,875 + 1,550 46,675

430/9 5,200 plague 200 1,000 1,165 -7,565 + 1,540 40,650
429/8 5,200 plague; 430 Spartolos 200 1,015 -6,845 + 1,340 35,145
428/7 200 880 -1,080 + 1,160 35,225
427/6 5,200 plague 200 880 -6,280 + 1,160 30,105
426/5 200 755 -955 +995 30,145
425/4 200 755 -955 +995 30,185
424/3 1,000 Delion 200 755 -1,955 +995 29,225
423/2 200 730 -930 +965 29,260
422/1 600 Amphipolis 200 730 -1,530 +965 28,695
421/0 720 -720 +945 28,920
420/9 725 -725 +955 29,150
419/8 730 -730 +960 29,380
418/7 200 Mantinea 735 -935 +970 29,415
417/6 735 -735 +970 29,650
416/5 - 500 740 -1,240 +980 29,390
415/4 735 -735 +970 29,625
414/3 740 -740 +980 29,865
413/2 10,000 Sicily 200 745 -10,945 +985 19,905
412/1 200 500 -700 +985 20,190
411/0 500 Eretria 200 505 -1,205 +985 19,970
410/9 400 Ephesos, elsewhere 500 500 -1,400 +985 19,555
409/8 200 490 -690 +985 19,850
408/7 200 495 -695 +985 20,140
407/6 200 505 -705 +985 20,420
406/5 1,000 Arginoussai 200 510 -1,710 +985 19,695
405/4 3,000 Aigospotamoi; 3,000 siege 200 490 -6,690 +985 13,990
404/3 1,500 executed; 180 killed 200 350 -2,230 +985 12,745

To these 12-13,000 Athenians, however, we must add all the klerouchs

and other Athenians who lived abroad but, after Aigospotamoi, were
forced by Lykourgos to return to Attica so that Athens, during the subse­
quent siege, was crammed with people (Plut. Lys. 13.3; Xen. Mem. 2.8.1;
cf. e.g. IG II27180). We have no information about their numbers but
they may well have totalled several thousand. Thus, accepting Gomme’s
47,000 ad u lt male citizens in 431 and su btracting all the losses sustained
during the Peloponnesian war we are left with some 15,000 citizens in
403, and this figure is probably too high rather than too low. By contrast,
Gomme suggests, as a rough estimate only, that there were some 24,000
HfM 56 23

Athenians around 400 B.C.11 The discrepancy between Gomme’s esti­

mate and the calculation made above arises because Gomme used 19th-
century population figures as his models and also because he did not face
the demographic consequences of adding up all losses suffered by the
Athenians in the period 431-403. Thus, my first conclusion is a dilemma:
if we accept Gomme’s figure of 47,000 citizens in 431 we must reject his
estimate for 403 as much inflated. But if we accept his estimate for ca.
400, we must assume a considerably higher number of citizens in 431, viz.
some 60,000 adult male citizens of whom ca. 55,000 will have been of
military age (18-59 years).
What is the authority for Gomme’s 47,000 citizens in 431? As is well
known, his estimate is based on Perikles’ survey, in Thuc. 2.13.6-8, of the
army strength of Athens at the outbreak of the Peloponnesian war:
ojtXixag Se xpiaxiXioug xai pupiovg elvai av£u xa>v ev xolg tjjQovpioig
xai xdiv jtag’ ejiaX^iv E^axiaxiXicov xai pvpicov. (7) xoaoijxot yap
E<}>nXaaoov xo jiqcoxov ojioxe oi jioXipioi EofJaXoiEv, ajio xe xojv jiqe-
o(3uxdxoi)v xai xobv VEtoxaxajv, xai pEXoixcov oaoi ojtXIxat f|aav ... (8)
ijiJiEag 5 e ajtE^aivE Staxoaioug xai xiXioug |i)v LjutoxoS;6xaig,
E^axoaiong 6 e xai x^ioug xo^oxag, xai xpiriQEig xag jiXajipovg
xpiaxoaiag. Gomme’s interpretation of this passage is as follows: A field
army consisting of 13,000 infantry, 1,000 cavalry and ca. 500 in the forts,
i.e. a total of 14,500 Athenians of hoplite status and aged 20-49.12 Next, a
reserve of 16,000 men comprising the youngest (aged 18-19) and oldest
(aged 50-59) of the citizen hoplites (= 10,500 plus the metic hoplites (=

11 Gomme (1933) 26 estimates 22,000 citizens aged 18-59, to which figure we must add
some 2,000 citizens aged 60-80+, cf. supra notes 8 and 9.
12 It is a vexed question whether the field army did or did not include the year classes 40-
49. Gomme (supra n. 2) contended that the field army consisted of hoplites aged 20-49,
(1927) 142; (1933) 5-6; (1956) 34-37; (1959) 61. Jones (supra n. 3) 165 preferred to
explain the 16,000 in the reserve by assuming that the oldest comprised the year classes
40-59, and not just 50-59. The debate has focused on Socrates who served at Delion in
424 when he was in his mid forties. Did he volunteer? Was he called up exceptionally?
Or did men in their forties form a regular component of the Athenian hoplite army?
First, it is worth remembering, as Gomme points out, that Socrates served again two
years later at Amphipolis (PI. Apol. 28 E). Second, the problem is, in my opinion, settled
by Ar. Ach. 600-601 (never discussed in this context): optov JtoTaovg pirv av6pag ev xai?
xa^eotv veaviag 6’ olovg cru 6ia6e6eax6xag. Grizzled men were regularly seen in the
ranks, but men usually become grey-haired in their forties, not in their thirties. Similar­
ly, Demosthenes was forty-six when he served as a hoplite in the battle of Chaironeia
(Aeschin. 3.253; Plut. Dem. 20.2; Mor. 845F).

3*
24 HfM 56
5,500).13 Apart from 200 hippotoxotai and 1,600 toxotai Thoukydides does
not count the thetes (who rowed the 300 triremes), and Gomme, as a very
rough guess, suggests a total of 18,000 thetes Furthermore, Gomme
takes the 16,000 in the home guard to be a population figure which
included not only the youngest and oldest of hoplite status but also all
citizens of hoplite status aged 20-49 but unfit for military service (the
adynatoi, estimated at ca. 3,000).15 Similarly, we must assume that his
18.000 thetes is a population figure comprising all thetes aged 18-59. Now
14,500+10,500+18,000 = 43,000 aged 18-59 and when we add some
4.000 aged 60 or more, the total is 47,000 adult male Athenian citizens.
Gomme’s 43,000 Athenians of military age is one possible interpreta­
tion of Thuc. 2.13.6-8 but, as he admits, his calculations leave room for
doubt and, more important, they are based on one wrong and two very
weak assumptions. (1) According to Thoukydides the 16,000 in the home
guard is not a population figure, but an army figure. Consequently, the
total number of oldest, youngest and metic hoplites may be (e.g.) 20,000
or perhaps even 25,000. (2) There is no evidence that the oldest and the
youngest comprise only citizens o f hoplite census. The phrase oooi ojtkixai
r]oav goes only with pexoixcuv, and not with xcbv JiQeaPuxaxojv xai xa>v
VBCDXaxtDV. So the oldest and the youngest probably include the oldest
and youngest of the thetes.16 (3) The number of thetes is unknown. Gomme
estimates, rather arbitrarily, ca. 18,000. A priori, any estimate between,
say, 10,000 and 40,000 is equally possible.
For the sake of argument, let us accept Gomme’s ca. 10,000 Athenians
in the home guard but assume, as a strict interpretation of Thuc. 2.13.7
requires, that they were composed of the youngest (aged 18-19) and the
oldest (aged 50-59) fit for military service but recruited from all census
classes including the thetes. In the model population described above
(note 9) the year classes 18-19+50-59 constitute ca. 'A of all males of
military age (aged 18-59). Thus, 10,000 Athenians aged 18-19 and 50-59
correspond to ca. 50,000 Athenians aged 18-59. Again ca. 15,000 hoplites
aged 20-49 in the field army correspond to ca. 18,000 hoplites of military
age (18-59) and accordingly there will have been some 32,000 thetes to
man the fleet, to serve as light armed troops in the field and to assist in
defending the walls.
But in order to find the total number of Athenian citizens we must add
to the army figures an estimated number of Athenians above military age
and an estimated number of Athenians unfit for military service. The
home guard will undoubtedly have included many who were unfit for
HfM 56 25
service in the field army or in the navy; nevertheless at least 10 % of all
Athenians of military age must have been unfit for any kind of military
service.1' Thus, the total number of Athenians aged 18-59 must, on this
calculation, have been min. 55,000. Next, in the population model I have
adopted, men over 60 constitute ca. Zn of all adult males (18-80+), i.e.
we must add ca. 5,000 to the ca. 55,000 aged 18-59. In conclusion, this
sketchy calculation indicates that the number of male Athenian citizens
living in Attica, including citizens unfit for military service and above
military age, totalled some 60,000.
To sum up, calculations based on Thuc. 2.13.6-8 may give a total of ca.
47,000 Athenians (Gomme’s figure) or ca. 40,000 (as suggested by other
historians)18 or ca. 60,000 (as the above calculation indicates). 40,000 is
probably a minimum, whereas 60,000 is not even a maximum. The only
way of making a choice between 40,000, 47,000 and 60,000 is, in my
opinion, to start with the fourth century figures and then to move back­
wards.
The most reliable sources we have for Athenian demography relate to
the period ca. 350-322 and they point to an adult male citizen population
of no less than 30,000 Athenians living in Attica.19 For the first half of the
fourth century, we have no comparable information but are forced to rely
on two general lines of argument. (1) Even in the first years of the
restored democracy the institutions seem to have been run according to
the rules, which requires a minimum of ca. 25,000 citizens. Ca. 15,000

13 The number of metic hoplites is another moot point. Thoukydides tells us that the army
sent against Megara in 431 included 3,000 metics (2.31.2). Gomme (1933) 5 assumes
that there may have been some 2,500 other metics of hoplite census and military age.
Thus, the metics in the home guard amount to some 5,500 men, and Gomme takes the
remaining ca. 10,500 to be citizens of hoplite status. If, however, we apply the model
population printed in note 9 and calculate the youngest (18-19) and the oldest (50-59) of
the citizens on the basis of a field army numbering ca. 14.500 hoplites aged 20-49, the
result is a much lower number of citizen hoplites in the reserve, viz-
14,500-^72.7 x (6.7+11.9) = 3,710. On this calculation the number of youngest and
oldest in the reserve must be max. ca. 4,000 and the remaining ca. 12,000 must have
been metic hoplites. This is the line of argument adopted by R. P. Duncan-Jones in his
article ‘Metic Numbers in Periclean Athens’, Chiron 10 (1980) 101-09.
14 Gomme (supra n. 2) (1933) 26.
15 Gomme (supra n. 2) (1927) 144-46; (1933) 5; (1956) 35-36; (1959) 61-62.
16 For a full discussion of these two points cf. Hansen (supra n. 1) (1981) 19-24.
17 For a brief treatment of this neglected problem cf. Hansen (supra n. 1) (1985) 18-24.
18 C f e.g. C . P a tte rs o n , Pericles’ Citizenship L a w o f 451-50 B .C . (N e w Y ork 1981) 66-68.
19 Cf. Hansen (supra n. 1) (1985) 66-69.
26 HfM 56
adult male citizens is simply too few to fill the ekklesia, the boule and the
dikasteria.20 (2) Perikles’ citizenship law of 451 was reenacted in 403/2 and
thenceforth, as far as we know, rigorously enforced.21 In the 4th century
the demographic effect of naturalization was negligible, and thus the
factors affecting the number of citizens were mortality and emigration
balanced by fertility (but not by immigration, since immigrants became
metics and not citizens). Due to the slow natural growth of populations
before ca. 1700 and due to the fact that emigration of citizens was not
balanced by immigration, the number of Athenians living in Attica
cannot have grown very much. It must have been almost stationary, and
in some years perhaps even declining.22 When democracy was abolished
in 322/1 the number of adult male citizens totalled some 30,000. The
inference is that the number of citizens in ca. 400 must have been at least
ca. 25,000 and possibly ca. 30,000, i.e. twice as many as the estimate of
15,000 given above as a consequence of accepting Gomme’s 47,000 citi­
zens in 431.
So we are back at the dilemma outlined above: either the information
we have about Athenian population losses 431-403 is misleading or the
starting point, 47,000 adult male citizens, is wrong. It is in my opinion
hard to question the sources for the losses suffered during the Peloponne­
sian war. We may of course reduce the estimated plague mortality by
holding that Thoukydides’ phrase ex T(I>v xa^ecov is wrong or mislead­
ing, or that the thetes who rowed the ships suffered less from the plague
than the hoplites. We may also reduce the demographic effect of the
Sicilian disaster by assuming that, although Athenian citizens regularly
formed a considerable part of a ship’s crew, they constituted only a small
fraction of the men who rowed the ships sent to Sicily in 415-413. Also,
the 1,500 Athenians killed by the Thirty may be an exaggerated figure.
But I suspect that all attempts to question our sources are merely es­
capes, to avoid the conclusion that a total of 40-47,000 Athenians in 431
is incompatible with the losses attested by Thoukydides and Xenophon
and with the fourth century population figures.
Instead of calculating the Athenian manpower in 431 and then dis­
cussing the losses in the light of this calculation I recommend the reverse
method: (1) to assume that ca. 25,000 must be the minimum number of
adult male Athenians living in Attica ca. 400 B.C.; (2) to accept the
population losses reported by Thoukydides and Xenophon; (3) to find,
for the year 432/1, the num ber o f citizens that will fit a citizen population
of min. 25,000 in 403. The consequence is an estimate of min. 60,000
HfM 56 27
adult male Athenians in 432/1, as is apparent from the following revised
version of the table on page 22.

Year (a) (b) (c) (d) (E) (F) (G)

431/0 200 500 1,500 -2,200 + 1,980 59,780
430/9 6,600 plague 200 1,000 1,495 -9,295 + 1,975 52,460
429/8 6,600 plague; 430 Spartolos 200 1,310 -8,540 + 1,730 45,650
428/7 200 1,140 -1,340 + 1,505 45,815
427/6 6,600 plague 200 1,145 -7,945 + 1,510 39,380
426/5 200 985 -1,185 + 1,300 39,495
425/4 200 990 -1,190 + 1,305 39,610
424/3 1,000 Delion 200 990 -2,190 + 1,305 38,725
423/2 200 970 -1,170 + 1,280 38,835
422/1 600 Amphipolis 200 970 -1,770 + 1,280 38,345
421/0 960 -960 + 1,265 38,650
420/9 965 -965 + 1,275 38,960
419/8 975 -975 + 1,285 39,270
418/7 200 Mantinea 980 -1,180 + 1,295 39,385
417/6 985 -985 + 1,300 39,700
416/5 500 995 -1,495 + 1,310 39,515
415/4 990 -990 + 1,305 39,830
414/3 995 -995 + 1,315 40,150
413/2 10,000 Sicily 200 1,005 -11,205 + 1,325 30,270
412/1 200 755 -955 + 1,325 30,640
411/0 500 Eretria 200 765 -1,465 + 1,325 30,500
410/9 400 Ephesos, elsewhere 500 765 -1,665 + 1,325 30,160
409/8 200 755 -955 + 1,325 30,530
408/7 200 765 -965 + 1,325 30,890
407/6 200 770 -970 + 1,325 31,245
406/5 1,000 Arginoussai 200 780 -1,980 + 1,325 30,590
405/4 3,000 Aigospotamoi; 3,000 siege 200 765 -6,965 + 1,325 24,950
404/3 1,500 executed; 180 killed 200 625 -2,505 + 1,325 23,770

20 In the fourth century the ekklesia was regularly attended by no less than 6,000 citizens
(cf. M. H. Hansen, The Athenian Assembly in the Age o f Demosthenes [Oxford 1987] 14-19 &
125 with note 804); but it is unbelievable that the assembly was regularly attended by
ca. 2/5 of all citizens. Next, only citizens above thirty could serve in the council of five
hundred; but in the model population described in note 9 supra ca. 1/3 of all adult males
are under thirty. Thus, a total of 15,000 citizens corresponds to some 10,000 citizens
eligible for the boule which, again, is an impossibly low figure, cf. Hansen (supra n. 1)
(1985) 51-64.
21 Cf. Hansen (supra n. 1) (1982) 176-77 with notes 14-15 & 20-21.
22 Cf. Hansen (supra n. 1) (1982) & (1985) 8-9.1 would like to thank Theodore Buttrey and
Poul Christian Matthiessen for reading and commenting on these three studies.
28 HfM 56
If to the ca. 24,000 citizens left in 403 we add the klerouchs and the other
Athenians sent back to Athens by Lysandros, the number of Athenians
living in Attica ca. 400 will have been at least 25,000, or rather a few
thousand more. In conclusion: 30,000 Athenians in ca. 350-322 pre­
supposes a minimum of 25,000 Athenians in ca. 400, which again pre­
supposes some 60,000 Athenians in 431, if we accept the severe losses
reported by Thoukydides and Xenophon; and this figure is perfectly
compatible with Thoukydides’ account of Athenian manpower in 431.

S u b m itte d to th e A c a d e m y J a n u a r y 1988.
Published March 1988.
 Abstract, Summary. - An abstract in English is compulsory. It should count 10-15 lines, outline
main features, stress novel information and conclusions, and end with he author’s name, title,
and institutional and/or private postal address. - Papers in Danish may be provided with a
summary in another language by agreement between author and editor.
Typescript. - Page 1 should contain title, author’s name and the name of the Academy. Page
2: Abstract, author’s name and address. Page 3: Table of contents if necessary. Captions
should be delivered on separate sheets. Footnotes should be avoided if at all possible; if
indispensable, they, too, should be typed on separate sheets. Consult a recent issue of the series
for general layout.
Typewrite with double space throughout and leave a 4 cm margin right. Indicate desired
position of illustrations and tables with pencil in margin and repeat it in the galley proof.
Use three or fewer grades of heading unless more are indispensable. Avoid long headings.
Indicate clearly the hierarchy of headings.
Figures. - Please submit two copies of each graph, map, photograph, etc., all marked with the
author’s name. Whenever possible all figures will be placed within the text; the nature of the
illustrations will govern the editor’s choice of paper quality.
All figures, also line drawings, must be submitted as glossy, photographic prints suitable for
direct reproduction. Prints fitting the indicated printed area are preferred, but the final size is
the responsibility of the editor. The scale should be indicated in the caption or, preferably, on
the illustration itself.
Fold-out figures and tables should be avoided. Use distinct (but not dominant) capital
letters for the items in composite figures. For transfer lettering use simple, semi-bold typefaces.
The size of the smallest letters should not be less than 1.5 mm. Intricate tables are often more
easily reproduced from line-drawings or from technically perfect original computer or type
processor output.
References. - In general, the editor expects all references to be formally consistent and in
accordance with accepted practice within the particular field of research. Bibliographical
references should preferably be given as, e.g., Shergold 1975, 16, the latter figure indicating the
page number unless misunderstandable.

Correspondance
Manuscripts should be sent to the Editor, Det Kongelige Danske Videnskabernes Selskab,
H. C. Andersens Boulevard 35, D K-1553, Copenhagen V, Denmark (tlf. +45.1.11 32 40).
Questions concerning subscription to the series should be directed to the publishers.

Publisher
Munksgaard Export and Subscription Service
Nørre Søgade 35, D K -1370 Copenhagen K, Denmark

Editor: Erik Dal

© 1988. Det Kongelige Danske Videnskabernes Selskab. All rights reserved. No part of this
publication may be reproduced in any form without the written permission of the copyright
owner.
 Det Kongelige Danske Videnskabernes Selskab

Historisk-filosofiske Skrifter
Hist. Filos. Skr. Dan. Vid. Selsk.
Priser exd. moms / Prices abroad in Danish Crowns.
Vol.
10:3. O ldenburg , E velyn , and R o h w e d er , J ørgen : The Excavations at Tall Dar-
ük and at ‘Arab al-Mulk. (Publications of the Carlsberg Expedition to Phoeni­
cia 8). 1982 .............................................................................................................. 200.-
10:4. B u h l , M a r ie -L ouise : Sükäs VII. The Near Eastern Pottery and Objects of
Other Materials from the Upper Strata. (Publications 9). 1983 ......................... 200.-

11. S teensberg A x el : Hal og gård i Hejninge. En arkæologisk undersøgelse af to

sjællandske landsbytomter. 1986 .............................................................. ............ 200.—

12. L u n d J o h n : Sükäs V III. The Habitation Quarters (Publications of the Carls­

berg Expedition to Phoenicia 10). 1986 ................................................................. 400.—

Historisk-ßlosoßske Meddelelser
Hist. Filos. Medd. Dan. Vid. Selsk.
Vol.
48:1. H endriksen , H ans: Himachali Studies. I. Vocabulary. 1976.............................. 180 —
2. H endriksen , H ans : Himachali Studies. II. Texts. 1979 ...................................... 150.—
3. H endriksen , H ans: Himachali Studies. III. Grammar. 1986 ............................. 200.-

Vol.
52:1. Part I-II. M atthiessen , P o u l C hristian : The Limitation of Family Size in
Denmark. (Princeton European Fertility Project). 1985 ...................................... 180.—
2. K ølln , H erman : Der Bericht über den Dänenkönig in den St.-Wenzels-Bio-
graphien des 13. und 14.Jahrhunderts. 1986 ....................................................... 100-
3. M c K innon , A lastair : Dating Kierkegaard’s Battles with Fate. 1986 .............. 50.—

5 3 . A ndersen , L en e : Studies in Oracular Verses: Concordance to Delphic Re­

sponses in Hexameter. 1987 .................................................................................. 300.-

5 4 . P eter sen , A rne F riem uth : Why Children and Young Animals Play. A New
Theory of Play and its Role in Problem Solving. 1988 ......................................... 120.—

5 5 . O lsen , B irgit A n et t e : The Proto-Indoeuropean Instrument Noun Suffix

*tlom and its Variants. 1988.................................................................................... 100.-

56. H ansen , M ogens H erman : Three Studies in Athenian Demography. 1988....... 50.—

Printed in Denmark by Special-Trykkeriet Viborg a-s. ISSN 0106-0481. ISBN 87-7304-189-0

Id: 701
Forfatter: Hansen, Mogens Herman
Titel: Three Studies in Athenian Demography.
År: 1988
ISBN: 87-7304-189-0
Serietitel: Historisk-filosofiske Meddelelser
Serienr: H 56
SerienrFork: H
Sprogkode: