Techniques of Plysplit Braiding

M Wotoms Kerdebeblepelacneya
ad ERY Es sb
BRAIDING
‘el fi re
eh
Ply-split braiding is a thread

~ technique which is still
too little
known in Europe. This book is
the first in-depth investigation of
a technique which due to its ~
extremely localised distribution,
has only recently come on to the
textile map.
Peter Collingwood, as in his other

definitive works on rugs, sprang
and tablet weaving, has delved -
deeply into the technique of ply-
split braiding. On five visits to
India he has travelled to isolated
off-the-map villages and
homesteads in Rajasthan and
Gujurat to gather knowledge from
the few remaining practitioners —
-of this craft and also acquire a
bbe Cenelem-beleWbboeye) aes) kyomeo) I(rein (eyo)
of artefacts. :
The latter have been the
foundation for the present book
which contains detailed practical
instructions for the reader and
records these arrestingly beautiful
objects. The plates — both of
Indian work and the author’s own
innovative samples — illustrate the
range of designs and structures |
this simple method, which needs
practically no equipment, can
produce. Ply-split braiding will
become an exciting new addition
to the textile worker’s craft.
=t-
The Techniques of Ply-Split Braiding
(d. 1998)
friend and mentor
‘hie wechimalics-or
PLY-SPLIT
BRAIDING
<>
Peter Collingwood
Ne Pity, SA ate wh ¥y
ee C LTBKS
z BOOKS AND CRAFTS INC.
PETALUMA, CALIFORNIA
First published in United States of America in 1998
by Unicorn Books and Crafts, Inc.
1338 Ross Street, Petaluma, Ca 94954
Copyright © Peter Collingwood
All rights reserved. This publication may not be

reproduced, stored in a retrieval system or transmitted,
in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without
the prior permission of the publishers.
A catalogue record for this title is
available from the British Library.
ISBN 0-9625586-9-9
Typeset by Antony Gray

Printed and bound in Great Britain by
MPG Books Ltd, Bodmin, Cornwall
Acknowledgements
I thank the following: without question, helped dig the jeep out of
John Gillow, a real textile enthusiast, who the sand, acted as guides and smiled and
collected for me in India long before my smiled.
visits there, for which he gave expert advice. Mr Wazir, who took me to villages in
Noemi Speiser, for her intense interest and Kutch, Gujurat and Vankabhai; a rabari
encouragement in this project and a flow of himself, he helped us establish contact with
always helpful letters dealing with rabari women in these villages.
nomenclature and structure. The brothers Rathi in Jaiselmer and
Errol Pires, one of the few people I know Barmer, embroidery enthusiasts, who kept
who shares my deep passion for ply-splitting. an eye out for what I wanted.
His position as a teacher at the National Ann and Ralph Norman, whose free-
Institute of Design in Ahmedabad has drawn and computer-generated diagrams
enabled him to build up an _ enviable make such an impressive contribution to the
collection which he has always generously visual side of this book. Their ability to
allowed me to examine and photograph. convert each of my crude roughs into
Mr ‘Tak, incomparable guide and girth- something faultless and beautiful constantly
hunter; see many more details in the surprised and encouraged me.
Introduction to the Indian part of this book Paul Wakefield, who has joined us in girth-
(page 164 ff.). hunts, for his great photographic expertise.
Ishwar Singh, who gave unstinting help on He has given the objects that special ‘jump
my first visit to Jaiselmer and always off the page’ quality.
welcomed me warmly on subsequent visits; Antony Gray for his immense care and
see List of Makers. patience in the typesetting of this book.
All the other makers named-in that list who Marianne Cardale-Schrimpff, who gave
shared their knowledge, patiently explaining me special details and photographs of ply-
what to them was self-evident. My pledge to splitting in Colombia.
them that they would be named in this book My wife, Elizabeth, who was my sanity-
was a constant spur to its completion. preserving companion on all the Indian
All the anonymous makers whose work I trips, keeping a full and useful diary. She
saw or collected. The recording of their often thought I would never roll this
technical skill, ingenuity and innate sense of Sisyphean stone to the top of the hill; it is
design has been a driving force in writing this largely due to her frequent nudgings that it
book. has reached there and not turned into a
All the villagers who entertained us, fed us lifelong millstone.
Digitized by the Internet Archive
in 2022 with funding from
Kahle/Austin Foundation
https://archive.org/details/techniquesofplysO000coll
Contents
Introduction
13
CHAPTER ONE
Ply-Split Darning CHAPTER THREE (continued)
. Irregular splitting of parallel plied Ue Using a cord-making device 25

elements 15 D Making a long length of over-twisted
Regular splitting of parallel plied 2-ply 26
elements 15 4. Unusual methods 27
2.1 Using 2-ply cords 15
CHAPTER FOUR
Starting a Braid
2.4 Using cords with different plies 16
. Fixing the braiding cords 29
. Irregular splitting of parallel braided
elements 16 1.1 Fixing to a starting rod 29
1.2 Supporting the starting rod 30
1.3. Fixing to transverse cords 31
CHAPTER TWO
1.3a Three-selvage braid 32
Ply-Split Braiding
2. Starting mid-braid 32
. Distribution 18 ve Splitting implements 33
. History 18 Basic splitting manoeuvre 33
. Nomenclature 19 4.1 Rightward splitting 33
FP
re
WN General description 19 4.2 Leftward splitting 34
CHAPTER THREE CHAPTER FIVE
Non-Traditional Cord Preparation Explanations of Diagrams

Using hand or power drill 21 . Analytic or structural diagram 35
1.1 With one person 21 2. Working diagrams 36
1.2 With two people 23 . Design diagram 36
CHAPTER SIX
Plain Oblique Twining CHAPTER SIX (continued)
Introduction 37 rePAE Mounting a special colour

Method 37
(cont.) sequence on the starting rod. 54
Producing a special colour
Correcting mistakes 40 sequence with twined linkings
ees
Speed
Structure of POT 40 while making a braid.
4a. Selvage with unsplit cords 41 Twice-split cords
Number of cords 41 Transversely-placed
Longitudinally-placed
Character of POT 41
Obliquely placed
Variations in starting 42 Cord transposition
ee
Nano
Design in POT 42 8.2d Using multicoloured cords
8.1 Design with uninterrupted passage ‘Transverse stripes
of cords from selvage to selvage 42 Longitudinal stripes
8.1a Relationship between Combining one-colour and
starting colour sequence (SCS) multi-coloured cords
and
8.2e Dimensional possibilities
course colour sequence (CCS) 45
with POT
8.2 Design with interrupted passage Changing a braid’s long axis
of cords from selvage to selvage 4a. Forming a tube
Sa Twined linkings 4a From a flat braid
Longitudinally placed twined Working as a tube from the
linkings 46 start
Making a single colour border 46 Colour effects with a
Influencing central design 49 tubular braid
Transversely placed twined 8.2f Increasing the width of
linkings 49 a braid
As a repeated operation 49 Adding two cords centrally
As a single operation 50 Adding cords laterally
8.2b Slits 51 8.2g Decreasing width of a braid 68
Used as structural features 51
8.3 Variations in the structure
Using slits to alter colour
oF POT
sequence in the cords 52
8.3a Using a 2-ply cord
8.2C Oblique hairline stripes 53
Breaking the strict alternation 8.3b Increasing the amount
of dark and light cords, on of twist left in a cord
either course, by making between splittings
two dark or two light cords 8.3c Eccentric splitting
lie side by side. 53
9. Plain triaxial twining
CHAPTER SEVEN
Single-Course Oblique Twining CHAPTER SEVEN (continued)
1. Working a single-section braid 2.1le Variations in structure in

1.1 Practical details which cords run directly
from selvage to selvage 96
l.la Splitting cords centrally
Cords changing function
Pile Producing a firm fabric within a section 96
2 Features of the working process Changing number of sections
by inserting extra areas 97
t3 Structural features
Altering the width of sections 97
1.4 Distinguishing ply-split
SCOT from other types
Za Variations in structure in
which some cords do not
lige: Using two or more colours run directly from selvage
1.6 Cords changing function to selvage 98
along an oblique line Slits 98
2. Working in Sections Subdividing a braid into
narrower parts 98
2.1 Longitudinal sections
Altering the course of certain
MMe Braid with two longitudinal cords 99
sections worked on a V-fell 79 Twined linkings 100
Design using cords of two Longitudinally placed to
or more colours make a one-colour border 100
Design using cords of Longitudinally placed to
only one colour change the CCS 100
Design using two-coloured Obliquely placed twined
cords 83 linkings 101
2.1b Braid with two longitudinal Inversions 102
sections worked on an All cords involved in the
A-fell 84 InVErS1ONs 102
Design 84 Only some cords involved
DANG Moving from one type of in the inversions 103
fell to the other in a Inversions with cords of
two-section braid 85 many colours 105
Changing from aV- to an Analysis of designs with

A-fell 85 INVETSLONS 106
Changing from aA- to an 2.1g Other structural variations 106

V-fell 85 Selvage with unsplit cords 106
2.1d Braid with many longitudinal Adding cords 107
sections 87 Losing cords 108
Width of sections 88 Longitudinal cords 109
Methods of working 88 22 ‘Transverse sections with cords
Design 89 changing function along a
Design with warp divided in transverse line 110
to half dark, half light cords 90 2.2a Method 110
Design with warp of half dark,
2.2b Design 111
half light cords, arranged
symmetrically 91 2.2c Producing a straight
Design with colours arranged selvage 112
to give chevrons or zigzags 92 2.2d Variations 113
Design with variations 94
CHAPTER SEVEN (continued) CHAPTER EIGHT
2.3 Changing from transverse to Combining Plain with S- and Z-Course
longitudinal sections 114 Oblique Twining
2.4 Cords changing function 1. Basic method 127
eee ence pe 2. All the diamonds worked in
2.4a Giving diamond-shaped oblique twining 129
areas 115 : .
Matar cnianewounday. aie 2a eros’ of solid potus 129
Dividing the diamond 117 2.2 Diamonds not of solid colour 130
2.4b Changing from longitudinal 2.2a Dividing the diamond
sections to all-over transversely 131
diamonds 117 2.2b Dividing the diamond
2.4c Other examples of longitudinally 132
oblique boundaries 118 2.2c Dividing the diamond
3. Dimensional variations 119 obliquely 133
3.1 Changing the long axis of 2.2d Subdividing into smaller
the braid 120 diamonds 133
3.2 Working cylindrically 121 2.2e ere: into four
3.2a Starting with a flat braid = 121 sy: fee ms A se
3.2b Working as a cylinder os ha sae cae eats
from the start 124 2.3 Using Sate linkings in the
3.2c Design using longitudinal BEI ee AS i
sections 121 3. Only part of each diamond
3.2d Design using transverse orca in cola Se bd
sections 122 4. Other uses of obliquely twining
Design 123 cords combined with POT 142
3.2e Sections with oblique 4.1 Outlining cords 142
boundaries 124 4.2 Free-running cords 143
Continuous spirals 124 4.3 Asa border 143
AEN Soe 128 5. A structure with regular
4. Treating cords abnormally 125 longitudinally twining cords
4.1 Eccentric splitting of cords 125 on a POT ground 144
4.2 Leaving more than a quarter-twist
between successive splittings 126
4.3 Combining cords of different
thickness 126
CHAPTER NINE CHAPTER TWELVE
Two-Layered Oblique Interlacing Traditional Cord Preparation
. Method 145 Introduction 164
. Structure 148 Le Materials used 164
. Design 149 1.1 Goat hair 164
Bh
WN. Joining the two layers of interlacing 151 1.2 -Gotton 164
4.1 At the starting edge 151 1.3 Other materials 164
4.2 At the sides 152 Preparation of the fibres 166
4.2a Without added cords 152 Spinning 167
4.2b With added cords 153 3.1 Device used (dhera) 167
Used as linked cords 153 3.2 Spinning the single yarn 167
Used as a ply-split border 153
3.3 Making the 2-ply yarn 168
Using POT 153
Using SCOT 153
3.4 Making the 4-ply cords 168
. Recognition of ply-split TLOI 154 3.4a As separate cords 168
Converting two-colour into Variations in working

one-colour cords 154 methods 170
. Using three colours 155 3.4b As connected cords 171

Analysis 173
CHAPTER TEN
CHAPTER THIRTEEN
2/2 Twill Oblique Twining
Ply-Splitting in India
. With longitudinal ridges 156
. With transverse ridges 157 . Introduction 174
. Equipment used in all ply-split
braiding 174
CHAPTER ELEVEN . Tangs and lhoums 174
Twined Linking 3.1 Short tang 175
. Introduction 158 3.1a Cords used 175
. Two types 158 3.1b Starting end 176
. Uses of R/R twined linking 160 3.1c Attachment to the saddle 176
Single course oblique
2.1 As the sole structural element 160
twining (SCOT) 177
2.2 As the means of changing the Plain oblique twining (POT) 179
courses of dark and light cords 160
Combined plain and S- and
2.3 As a textural effect 161 Z-course oblique twining 179
2.4 Other uses 161 ‘Two-layered oblique
lacing (TLOI) 179
. Uses of R/L twined linkings 162
Plain oblique twining with
. Incomplete linking 162 longitudinal twining cords 180
. Two-layered interlinking 162 2/2 twill oblique twining 180
Combining ply-split
braiding with darning 180
CHAPTER THIRTEEN (continued) CHAPTER THIRTEEN (continued)
3.1d General working methods 180 3.6b Creating yarn branches 186
3.1e Structures used 177 a aesne side branches
186
3.1d Other features in OE eee
short tangs 180 By coiling yarn around a core 187
Slits 180 3.6c Using extra yarn 187
Braided cords 181 4. Gorbandhs 188
Cords covered with knots 181 4.1 Contruction 188
ore hh 5. Godiya and sariya, ankle and
oa knee bands 190
Embroidery 181
Thafts 181 6. Muhra or moorka 191
191
3.2 Long tangs 181 7. Hardi
191
3.3 Short lhoums 182 8. Bheer
3.4 Long Ihoums 182 9, (Cheekah 193
Structures used 182 9.1 Open-mesh cheekahs 193
POT 182 9.1a Made from one cord 193
Plain combined with S- and 9.1b Made from 3-ply cotton 194
eee So ene oe 9.1¢ Made with 8-strand braids 195
Mixed techniques 4183 9.2 Closely made cheekahs 195
3.5 Fringes on tangs and lhoums 183 9.2a AE SCOT 196

3.6 Jhoums 185 9.2b Using POT 197
3.6a Covering a wrapped 9.2c Using twined linkings 197
strip of cloth 186 9.2d Flat-bottomed type 197
With weft wrapping 186 10. Mukhiarna 198
With interlacing weft 186
With looping weft 186 CHAPTER FOURTEEN
Ply-Split Braiding in Colombia
200
APPENDIX
Makers of Camel Girths
202
GLOSSARY
205
BIBLIOGRAPHY
207
LIST OF SUPPLIERS
208
Introduction
This attempts to be a different sort of textile book,

being intended for both the craft worker and the
ethnologist. For the former it provides explicit
working instructions, for the latter detailed
descriptions of traditional methods and artefacts
found on field trips to India.
The author has always found this bringing
together of the craft and research disciplines a fertile
field for understanding and enlightenment. The
craft worker can only gain from familiarity with the
traditional material, which exhibits ideas and
refinements, both technical and aesthetic, evolved
over centuries. The research worker can only gain by
himself trying out the techniques in a practical way
and so understanding fully the problems involved
and the ingenuity of the solutions reached.
These two strands have been separated for
convenience, the practical preceding the ethno-
logical, though the reverse order may seem equally
logical. Figure 1 Basic plan of a ply-split darning
and b ply-split braiding.
Every type of fabric has its own specific method of

interworking its constituent eléments into a coherent
whole. In knitting, it is interlooping with a single
element; in weaving, it is interlacing with two sets of
elements. The method in ply-splitting can be defined
as follows. Previously prepared units in parallel
arrangement are joined by elements running more
or less at right angles, which do not interlace with
them, but split them in sequence (after Speiser,
whose Manual of Braiding, 1983, 1988 and 1991,
gives the best overall summary of ply-splitting,
placing it in context with all other types of braiding).
In its simplest form the units vary widely, being
anything that can be pierced, like twisted bunches of
INTRODUCTION *+ 13
straw, split rattan or braided fibres. This two cords cross, one splits the other. This
technique is therefore merely splitting. But if technique is ply-split braiding.
the parallel units are carefully plied cords and Of the two techniques, ply-split darning is
the splitting elements pass between these by far the commoner and is known world-
plies, the technique becomes ply-splitting. wide, having many references in textile
Such plied cords can be split in two very literature. For these reasons it will only
different ways. be dealt with in summary. The main part of
They can lie side by side, parallel to the this book concerns the relatively ‘new’
selvage of the fabric produced, and be split technique of ply-split braiding, covering the
by other elements which run from selvage to several structures it surprisingly yields and
selvage at right angles to them, as in Fig. 1a, their immense pattern-making potential in
so that there are two sets of elements bearing both two and three dimensions.
a warp/weft relationship to each other. This It should be noted that as an operation,
technique is ply-split darning; in some not a fabric-making technique, ply-splitting
carefully made examples the resulting is of course an integral part of the splicing of
structure 1s warp-twining. ropes where it is used to join ropes end to end
Or the cords can lie obliquely in relation to and make terminal or mid-length loops. A
the selvage as in a braid, see Fig. 1b; they run sailor’s pocket knife often includes a spike for
obliquely until they reach one selvage and forcing an opening in a rope, i.e. splitting it,
then change direction and run obliquely to and sometimes a loop for pulling a ply
the opposite selvage. At every point where through such a split.
14 * THE TECHNIQUES OF PLY-SPLIT BRAIDING

CHAPTER ONE
Ply-Spht Darning
The following are a few examples of ply-split communication; Collingwood, Maker’s Hand,
darning arranged in a simple classification. 1987/98)
+ Belts and bands on _ clothing of
1. IRREGULAR SPLITTING OF Sarakatsani nomads of Greece. Made of
PARALLEL PLIED ELEMENTS black goat hair sometimes 10 cords wide.
This is a common way of making a strap Joyce Smith, ‘Decorative Techniques of the
which has no cut ends. Around two stakes, Sarakatsani’, Weaver’s fFournal, Vol. VIII,
the required distance apart, a cord of 2- or Number 1, Issue 33, 1984)
more ply is wrapped a sufficient number of + Camel girths from Gujurat. Made of
times to give the desired width. Starting at black and white goat hair to give simple
one end, all these cords are pierced at right patterns which are altered by occasionally
angles by a needle carrying a thread, often leaving no twist between splittings. These are
finer than the cords, which gradually works to rarely seen, ply-split braiding being a far
the other end. The needle does not follow any commoner technique for camel girths.
special path through the cords, splitting them (Author’s observation)
haphazardly. The straps securing the roof
spokes of a yurt to the side frame are made in * Guatemalan support for back strap
this way, one end being looped to aid tying. weaving. (Virginia Harvey, ‘Split-Ply
Twining, An Update’, Interweave White
Paper, 1981)
2. REGULAR SPLITTING OF PARALLEL
PLIED ELEMENTS * Indian yoke halter for oxen = jotar.
(Eberhard Fischer and Haku Shah, Rural
In this group the splitting element takes an
Craftsmen and their Work, NID, Ahmedabad,
exact and carefully repeated “path through
1970)
the plies of the parallel cords. It leaves the
same amount of twist in each cord between + Indian pouch for sling (also described in
successive splittings and so leads to a perfect above volume).
warp-twined structure.
The above three are made from one single
2.1 Using 2-ply cords cord which is initially arranged as a narrow
<> Mexican headband. Made of 2-ply sisal, warp and then begins splitting itself at one
right half S-, left half Z-plied. end and then the other, the long floats in
between adding to the width of the warp. In
<> Botswana beer strainer. Made of two stiff this way an oval or diamond-shape is made,
grasses plied together and worked to give a suitable for their use. The jotars are made ina
conical shape, sometimes with stripes. The great variety of ingenious designs, the single
technique’s local name is ‘ukucusha’. (Private cord often taking a bewildering path.
PLY-SPLIT DARNING « 15
2.2 Using 3-ply cords change accomplished by unplying the cords
<* Nepalese headband = namlo. Made from and re-plying them so each has a different
S- and Z-plied nettle-fibre cords which colour arrangement; see Plate 1, right. This
alternate across the warp. The splitting cord trick has no known counterpart in tablet
goes across under one ply of each cord then weaving so is a further distinguishing feature.
back under two. A special retrograde looping (Author’s observation)
of this cord at both selvages gives a neat edge. It follows that any warp-twined design
(Susi Dunsmore, Nepalese Textiles, British normally produced by turning tablets in only
Museum Press, 1993; Collingwood, Maker’s one direction — 1e. with no _ twining
Hand, 1987/98) reversals —can be faithfully reproduced by
ply-split darning appropriately made cords.
+ Zulu flat mat. Made from grasses, plied S
and Z to make wide stripes. The 2-ply 2.4 Using cords with different plies
splitting element takes an ingenious path A horse harness known as ‘shiri-gai’ is found
through the 3-ply cords and so avoids floats in the Shoso-in treasury at Nara in Japan and
over two wefts. J. W. Grossert, Zulu Crafts, could date from the 8th century. This
Pietermaritzburg, 1978; Collingwood, Maker’s consists of two flat tubular straps which at
Hand, 1987/98) one point pierce each other at right angles. A
nineteenth-century example of the same
2.3 Using 4-ply cords
type, examined by Noemi Speiser, showed
+ Indian camel girths. Made from black and
the combined use of 2-, 3- and 4-ply cords
white goat hair. and was definitely ply-split darned. (Manual
This 4-strand warp-twining has to be of Braiding, 1991) The origin of such a
distinguished from the almost identical complex piece was probably a professional
structure produced by tablet weaving, as
workshop; it was not made by an individual
both methods can produce similar patterns; for his own use as are the majority of other
see Plate 1, left and centre. In both cases
ply-split objects.
there is a combination of S- and Z-plied
cords set parallel to each other. In ply-
splitting the cords exist from the start and so 3. IRREGULAR SPLITTING OF
the S-plied cords will consist of Z-plied PARALLEL BRAIDED ELEMENTS
elements and the Z-plied cords of S-plied + Zulu belt of grass fibre. (Grossert, Zulu
elements. But in tablet weaving the cords are Crafts, 1978)
made as the weaving proceeds, so both the
+ Soles of espadrilles.
resulting S- and Z-plied cords will consist of
Both of these are made from a 3-strand braid
elements with the same ply. Apart from this
stood on edge, folded to give the required
crucial and diagnostic distinction, tablet
shape, then pierced at irregular intervals by
weaving usually shows a twining reversal
needle and thread to give a solid fabric.
point in the cords, ply-splitting never can;
see middle of Plate 1, centre. + Mexican belts and hatbands made of
horsehair.
* Indian tubular belts. Made from cotton so
that one side of the flattened tube is + Tibetan sling straps made of wool.
patterned, the other side plain. The Both are made from 8-strand tubular braids
retrograde looping of the splitting element stitched edge to edge with a thin thread. The
gives the impression of another cord at each juxtaposition of their various two colour
edge. These sometimes have a pattern patterns give the impression of a wide

complex braid. (Collingwood, Maker’s braiding to darning and back again for no
Hand, 1987/98) obvious reason; see Plate 94.
In a rare Indian example, after the ply-split
In the few places where ply-split braiding is a braid had been made, new cords were
traditional technique, it is often combined stretched longitudinally in front of it. They
with ply-split darning. For instance, cords were ply-split darned with a weft which also
for braiding may be _ initially bunched pierced the braid at each selvage. This
together within a ply-split darned tube as in produced a ply-split darned layer
the starting loop of a camel neck decoration superimposed on a ply-split braided layer,
in India or in the Colombian crupper. Indian the two joined at their selvages to make a
camel girths occasionally switch from flattened tube.
PLY-SPLIT DARNING «+ 17
CHAPTER TWO
Ply-Spht Braiding
1. DISTRIBUTION on ten months field research carried out in

Ply-Split braiding has a remarkably, indeed 1974 and 1975. Though unpublished, its
puzzlingly, limited distribution. This may contents made a large contribution to her and
simply be due to its relatively late arrival in Betsy Quick’s joint production, Ply Split
the consciousness of textile academics and, Camel Girths of West India, a 1982 publication
as happened with tablet weaving, once its of the then Museum of Cultural History
existence and characteristics become better (now the Fowler Museum), University of
known it will be more widely recognised. It California, Los Angeles. The latter is the most
can also be attributed to the few practical complete description of the technique to
uses of the typical ply-split artefact, a thick date, based largely on information gleaned
heavy, usually narrow, fabric. By its nature from Ishwar Singh in Jaiselmer, Rajasthan.
the technique cannot be adapted for fine
fabrics; by its slowness it is not suitable for 2. HISTORY
making large fabrics like rugs. It is not surprising that there are no early
Today the technique is found almost written records for such humble artefacts as
exclusively in the Thar Desert, a region camel girths and trappings and that their
which includes parts of the states of existence in former times has to be argued
Rajasthan and Gujurat in NW India and from paintings and photographs. A possible
neighbouring areas of Afghanistan and exception is found in Ain-1-Akbar, a survey
Pakistan. It has one other occurrence right of the kingdom of the great Mogul ruler
across the globe among the Guajiro Indians Akbar (1542-1605), made by Abul Fazel.
in Columbia, South America. This was noted (Translated from the Persian by H.
by Marianne Cardale-Schrimpff in her Blochmann, Calcutta, 1873) This mentions
exhaustive doctoral thesis “Techniques of camel girths with their local names and even
Handweaving and Allied Arts in Colombia’ prices, but there is no way of discovering
of 1972 (unpublished), and hers is the first whether these girths were ply-split.
known description of the technique. Painters, both Indian and those visiting
Four years later, Virginia Harvey’s Split-Ply from the West, have depicted camels, some
Twining appeared as Monograph 1 in the with clearly shown girths. A popular
“Threads in Action’ series. This was based on Rajasthani folk story in which a princess is
the analysis of several camel girths and dealt rescued by her camel-riding lover has led to a
with one of the structures and its non- much copied painting of this event. The
traditional possibilities. In the same year, camel girths show the zigzag pattern which is
1976, Judith Stein submitted a MA thesis so typical of ply-splitting that its presence
entitled “The Camel Belts of Western India, can fairly safely be taken as diagnostic of the
Technical Analysis and Classification’, based technique. A version in the Palace Library,

Jodhpur is dated 1820. An 1868 painting, wool worked with two sections of the single-
“The Midday Sun’ by John Griffiths, shows a course oblique twining (SCOT) technique.
street scene with camels wearing typically
ply-split girths. More recently in 1930, the 3. NOMENCLATURE
American Hubert Stowitts (an ex-dancing
No accepted terminology exists for some of
partner of Anna Pavlova!,) painted in Jaipur
the structures obtained with ply-split
‘His Highness the Maharajah’s Camel’, an
braiding. In these circumstances the choice
animal weighed down with ply-split
of new words and phrases becomes an
decorations.
agonising problem, knowing that a bad
‘The zigzag patterning on the girth holding
choice can cast its shadow of inaccuracy or
the saddle on fighting and hunting elephants,
inaptness well into the future. Also the
clearly shown in one or two mid-eighteenth-
desires for conciseness and for accuracy are
century Rajasthani paintings, suggest that it
usually in conflict.
was also ply-split, but no such artefacts have
Iam well aware that fault can be found with
been seen.
my choices, but correspondence over almost
Maharajahs were often keen amateur
a decade, chiefly with Noemi Speiser, attest
photographers and some palace museums in
to a long and difficult search with the
Rajasthan hold albums of their work. A
reluctant discarding of many alternatives.
thorough search through these would
One incidental advantage of two of the
probably locate early pictures of camels,
descriptive phrases, single-course oblique
especially those decorated for weddings and
twining and plain oblique twining, is that
official occasions. An example is one dated
they can be abbreviated in writing and
1908 in the Lallgarh Palace Museum,
speech to SCOT and POT.
Bikaner; it shows two camels with girths
exhibiting the zigzag pattern. Rajasthan,
divided as it was into small kingdoms and not 4. GENERAL DESCRIPTION
having much economically to offer the As already pointed out, ply-split braids form
occupying British, was largely left to its own a subdivision of general braids, that is,
devices. So the chances of any of its non- fabrics through which a single set of elements
exotic crafts being recorded or brought home move obliquely on two diagonal courses. On
as souvenirs in the period of the Raj were slim. reaching the edge of the fabric on one
However, a photograph of a pair of beautiful course, each element is reflected back into
girths, purchased in Bikaner, appeared in the the fabric along the other course, as in Fig.
first edition of Tablet Weaving (Mabel Peach, 1b. This represents a generalised diagram
Dryad Handicrafts, Leicester, 1921), being only indicating the course of the six elements
mistakenly thought examples of the latter involved, not what happens when one
technique. These passed from the Dryad crosses another. But it is how two crossing
collection to the Leicester Museum and are elements interact which crucially determines
now unfortunately lost. the braid’s structure. If one simply passes
Despite this understandably sparse record over: the other and both continue on their
from India, a ply-split braid recently found in paths in some regular over/under sequence,
excavations at the fortress of Masada, Israel, the structure is oblique interlacing. What
which was overrun by the Romans in aD 73, distinguishes ply-split braiding is that, at
shows that the technique itself has long been every crossing point, one of the elements
in existence. This fragment is made of 2-ply actually pierces the other, the whole length of
PLY-SPLIT BRAIDING + 19
the former actually passing between the plies maintains its position without any effort on
of the latter. the worker’s part. The twist in the as yet
As the elements most often used are 4-ply unworked cords holds the last splitting cord
cords, the split cord has two of its plies lying securely in place.
over, two under, the splitting cord. At this A disadvantage of the ply-splitting
point in the fabric, the latter is completely technique is that it can only be used with
concealed, hidden by these plies, two of relatively thick elements; there is a lower limit
which are visible on the front, two on the beyond which a thread is too fine to be split
back of the braid. with ease. This is in strong contrast with
It will be found that whenever two cords other fabric techniques, like weaving or
meet on their oblique courses through the knitting, which can employ threads from the
fabric, the worker has to make only two thickest to the finest.
decisions. He has naturally to decide which At any specific point in a ply-split braid,
of the two cords is the one to be split, for that the same split cord is visible on both the front
is the one which will be visible at that point in and the back of the fabric; so these two faces
his design. But he must also decide how are invariably identical in structure. Note
much twist is to be left in the split cord that this is in strong contradistinction to
between this and its previous splitting. oblique interlacing, where an _ element
Several quite distinct structures and their passing ‘over’ on the front is seen to pass
resulting patterns all derive from the correct ‘under’ on the back. Using two-colour cords
ordering of these two easily controlied does not alter this fact, though the two sides
variables. It is surprising that such an will differ in colour. The only and rare
incredibly simple operation as ply-splitting, exception to this rule of identical structure
requiring the minimum of equipment, can occurs when a cord either from choice (for
have the many, often complex, results design purposes) or from necessity (having
described and illustrated in this book. an odd number of plies) is not split centrally;
An advantage over other types of braiding the number of visible plies on the two sides
is that the fell, the growing edge of the fabric, then differs.
20 > THE TECHNIQUES OF PLY-SPLIT BRAIDING

CHAPTER THREE
Non-Traditional Cord Preparation
Ply-Split braiding is traditionally worked It has first to be given extra twist in its ply
with 2-, 3- or 4-ply cords. The latter are by direction, i.e. in S direction if S-plied, and
far the commonest and in fact are implicit in then be allowed to twist back on itself to form
the usual definitions of the structures the cord in one of the following ways.
produced. So, unless otherwise stated, cord
made up of four, usually 2-ply, elements 1. USING HAND OR POWER DRILL
plied together will be considered the norm in
1.1 With one person
this book. It will further be assumed that the
final plying of the cords is in the Z direction. Set a thick darning needle at an angle in a
This again follows tradition and will be seen block of wood, clamped to a table top, as in
in most of the diagrams and illustrations, but Fig. 2. Fix a cup-hook, or length of wire
is of course not essential to the technique. So similarly bent, into the drill chuck. Measure a
the cord has a formula of Z/2S/4Z. length of yarn about twice the distance
In ply-splitting the plies of such a cord are between the extended hands. Tie an
repeatedly forced open by the implement overhand knot at each end. Opening the ply
used and by the passage of another cord. It is near each knot, slip one end over the needle
essential to the firmness of the fabric being and the other over the hook, as in Fig. 2a.
made that the cord split in this way Keeping this stretched yarn taut, begin to
immediately re-plies. This means that the work the drill so that the ply direction in the
cord must be designed to have a high degree yarn is increased. So, assuming it is S-plied,
of twist in its final Z-plying. It helps if it is it gradually acquires more and more S-twists
also made from a fibre of natural springiness. per centimetre. The yarn will as a result
As such properties are unlikely to be found in slowly shorten and the worker must walk
commercially available yarn, the worker will forwards. At a point at first discovered by
have to make his own cords. Suitable non- trial and error and subsequently known by
traditional methods will now be described to counting the turns of the handle (hand drill)
produce such cords; traditional methods are or marking in some way the amount of
forward movement (power drill), stop this
described later in Chapter 12.
The worker can begin from scratch, over-twisting. As a rough guide, it is the point
where any further twisting will cause the 2-
spinning Z-twist singles, plying two of them
ply yarn to kink and snarl.
to make an S-twist 2-ply and then plying four
Now grip the yarn at its halfway point with
of the latter to make a Z-twist 4-ply cord. But
one hand, say, the left, and with the right
it is more likely he will construct his cord
bring the drill to the needle in the block and
from some existing yarn, perhaps a 2-ply S-
slip the end of yarn from the hook on to the
twist carpet wool, or a cotton or linen of
needle. See Fig. 2b. The yarn now lies in a
suitable thickness.
NON-TRADITIONAL CORD PREPARATION * 21

doubled form, its two knotted ends
on the needle and its loop held by the
left hand
Place the drill so that its hook
engages with the two yarns halfway
between the needle and the loop held
in the left hand, as in Fig. 2c. Then
place this loop over the needle. The
yarn now lies, folded in four,
stretching from drill to needle, as in
—_—
— — .
Fig. 2d.
F After making sure that the whole
y,s ae
U
oe EIS
rar
a ora — length of yarn is equally tensioned in
(4 ;
ea this new configuration, work the drill
so that the hook revolves in the
opposite direction to that used
before. The four over-twisted yarns
will easily and evenly combine into a
4-ply cord.
Put a short length of fine varn
through the needle’s eye so, as the
cord is slid off the needle, this will go
through its elements; then tie it to
prevent them unplying. A similar yarn
is tied through the two loops on the
drill hook before that end of the cord
Figure 2 Stages in making a 4-ply cord by one person. is released. This is made easier if the
end of the hook is hammered flat and
a small hole drilled there, as suggested
in Fig. 2d, to take the fine yarn as did
the needle’s eye. The two secured
ends will then look as in Fig. 9b.
Alternatively the end with two loops
can be put directly on the starting rod
which is to hold the cords while ply-
splitting; as in Fig. 9a, Chapter 4.

When released the cord will assume its own degree
of twist. If the drill has been over-turned in the last
stage, the cord will untwist, if under-turned it will
twist up a little. This final plying twist has to be such
that the cord if unplied at any point re-plies
immediately of its own accord. It is likely that the
first cord made, especially if of a hitherto untried
yarn, will fall short of this ideal. But once the cord is
made it cannot be altered as the final plying twist is
directly related to the degree of initial over-twisting.
So if it is too softly plied, more over-twisting is
needed in the first stage when the next cord is made;
if impossibly stiff, then less over-twisting.
The above method will make a cord of one colour
about 60 to 70 cm long, depending on the worker’s
arm span. It will have two loops at one end and a
loop and two cut ends at the other; as in Fig. 9a and
b in Chapter 4. A series of such individually made
cords provides a good set of elements on which to
make samples, small projects and develop designs.
‘To make a 4-ply cord with two of its constituents
dark, two light — as used in some techniques — begin
with two half-lengths, one of dark, one of light,
knotted together, as in Fig. 3a. In the final stage of
plying, arrange them carefully so that like colours lie
side by side in the cord. If four quarter-lengths are
knotted together, a four-colour cord can be made.
To make a thicker cord, begin the process with a
doubled or trebled length of yarn, not a single.
1.2 With two people

The length of cord produced by one person is
limited by the distance he can stretch as he transfers
one end of the yarn from drill to needle. But two
people — a drill worker and an assistant — can make a
cord of any length, only limited by the space
available.
The over-twisting is exactly as described above,
recording in some way the amount inserted. The
assistant then takes the end off the hook, looping the
yarn itself around the hook instead and walks towards
the distant needle; see arrow in Fig. 3b. Naturally this
action also draws the drill worker towards the needle
as he carefully keeps the doubled yarn tight. The
assistant places the end he is carrying over the needle,
then walks back to the drill. Again he takes the yarn
NON-TRADITIONAL CORD PREPARATION +23

Poy VO
OTTT TTT a
Figure 3 a Making a two-colour cord; b—c stages in making a 4-ply cord

by two people; d using two sticks to ensure correct lay of the threads
from the hook, looping the doubled yarn over in the final plying as the top in rope making.
the hook, and again walks to the needle where About 10 cm from the hook in the drill, he
he puts the loop he is carrying over the needle, places two sticks or even his fingers through
the drill worker having to move forwards as the four constituent yarns in a cross
before. See Fig. 3c. formation, separating them as shown in Fig.
The final stage is as described above, with 3d. As the drill is turned, the 4-ply cord
the drill being turned in the opposite begins to form between the hook and these
direction and the two ends of the resulting 4- sticks. The assistant then gradually slides
ply cord being secured with ties. A series of them, at a speed in harmony with the drills
long cords made in this way can be used side turning, towards the needle where he
by side to make some very long fabric. Or removes them, leaving in their wake a
they can be cut into shorter lengths, each perfectly laid 4-ply cord.
with secured ends, which are doubled over A drill at each end of the yarn naturally
the starting rod. See Fig 9c. speeds up both the over-twisting and plying
There is a tendency for the plying of this stages.
greater length to lead to slight faults in the It is possible for one person to make such a
laying of the plies, which should be avoided long cord by acting, as it were, as his own
especially when making two-colour cords. assistant. After over-twisting he lays the drill
The assistant can overcome these by acting on the floor or a table, carries the yarn end,

thena yarn loop, to the needle, dragging the drill to ensure the four elements ply in the
drill as he does so. He then picks up the drill correct order. Alternatively an assistant with
for the plying stage. two sticks can control them as described
above. The forming cord will begin to
2. USING A CORD-MAKING DEVICE lengthen so the worker moves backwards;
There are several cord-making devices finally it shortens slightly.
specially developed for ply-splitting; see List It is a benefit of this method that if the cord
of Suppliers. The use of the type shown in is found to be unsuitable, e.g. too soft, the
Plate 2 will be described. mistake can still be rectified. Work the drill in
’ The twisting head has four hooks on it the opposite direction to undo the final
which turn in unison because they mesh with plying. When the four elements are running
a central gear wheel whose protruding axle straight and separate from the twister to the
fits into the drill chuck. The yarn to be used is far hooks, put on the catch. Then work the
stretched between these four hooks and a set drill to increase their twist and carry on as
of similarly spaced hooks on a block of wood normal. Remember to add these extra turns
at the far end of the rope walk. As Fig. 4 when making subsequent cords.
shows, two lengths are needed. One takes a Because at its two ends the cord is attached
COUTSe Age le cee and tne-other Cy.5,.4, D: to hooks, it has to be bound or taped before
Some form of catch, fixed to the drill, is fixed cutting it free. The long cord produced can
on to the head to prevent it from rotating. So be cut into shorter lengths, each with its ends
when the drill is worked the head stays still but secured.
the four hooks spin round, simultaneously Of course, all four hooks need not be used;
adding twist to the four lengths of yarn. 2- and 3-ply cords can be made with equal
When sufficient over-twist has been put in, ease. To make cords with two dark, two light
the catch is released from the head and the plies, the upper yarn in Fig. 4 would be dark,
drill is worked in the same direction as before. the lower light. A one-colour cord of twice
The whole head now revolves (the hooks the above thickness can be made by taking a
remaining stationary in relation to it) and single length of yarn around the hooks in the
plies the yarns into a 4-ply cord. following way. Starting at A, it goes round 1,
As this final plying is done, pull hard on the 2, By 2non Gs 3, 450,451 andiback to Ay
LEZ Z III IIA

SE = 15 JS
~J
Figure 4 Cord-making device and fixing board at far end.
NON-TRADITIONAL CORD PREPARATION *¢ 25

A four-hook twister, worked in a completely
different way, based on a Tibetan yak-hair spinning
device, has very recently (1998) been developed; see
Ann Norman in List of Suppliers.
3. MAKING A LONG LENGTH OF OVER-

TWISTED 2-PLY
Another approach is to make a very long length of
over-twisted 2-ply yarn, wound on a bobbin, and
then ply it using the traditional continuous-cord
method. The length is made as follows.
Fix a length of threaded rod, A, in Fig. 5, into the
chuck of a drill, B. On to this slip a bobbin, C, then a
crosspiece of wood or metal, D, long enough to
protrude beyond the flange of the bobbin. Fasten
them in position with a nut. Attach the yarn to the
shaft of the bobbin.
The principle is to point the bobbin at the distant
yarn source, then work the drill in the appropriate
direction so that the crosspiece, catching the yarn,
increases its twist with each revolution. When
sufficient twist has been added, stop the drill and
turn it through a right-angle. Then start it again, Figure 5 Making long length of over-
revolving in the opposite direction, and let it wind twisted 2-ply.
the over-twisted yarn on to the shaft of the bobbin
meanwhile walking towards the yarn source at the Figure 6 Converting over-twisted 2-ply
into 4-ply cords, using traditional
relevant speed. Realign the bobbin and draw out continuous-cord method.
more yarn and repeat the process. This presupposes
either an assistant at the yarn source paying it out or
gripping it as required, or some simple friction
device which has the same effect.

When sufficient of this over-twisted
yarn had been prepared, wind it an
even number of times around two
posts, A and B, their distance apart
being close to the length of the final
cords. See Fig. 6. The winding must
be carefully done so yarn does not
override itself at either post.
Lift the two top loops from B, put
them on the hook of a drill and ply
them into a cord; meanwhile a helper
must hold the yarn at A to prevent
any slipping. Tie some thread to
secure the two loops at the end of this
cord, then make another and so on.
By this neat method, a series of
cords are made, all exactly the same
length and all ending with loops; no
knots are involved, except in the first
and last cord made. Each cord is
connected to its two neighbours at
post A in the way shown in Fig. 7;
this is of course the end where the
braiding starts, post A, or a
Figure 7 How continuously made cords are related to the substituted narrower rod, being the
starting rod. starting rod. Alternatively a cord can
replace A and be tied to itself into a
tight ring, from which the cords now
radiate, ready to be braided as the
base of a bag, bowl or basket. A warp
of dark and light cords or of cords
with 2 dark/2 light plies can also be
made by this method.
4. UNUSUAL METHODS
Fig. 8 shows two unusual methods of
making 4-ply cords suitable for ply-
splitting. In Fig. 8a, the rod splitting
two 2-ply yarns is forced along their
length (arrow) separating their plies
which then immediately reform on
the other side of the rod into a 4-ply
cord. In Fig. 8b, from a long 4-ply
cord, two plies are tied and pulled
out from opposite sides. These
immediately twist up into two 4-ply
NON-TRADITIONAL CORD PREPARATION °27

cords. They can be lengthened at will
(arrows) or added to by repeating the
process elsewhere along the original cord’s
length. So a series of cords can be made all
connected at one end, each finishing with
two blind loops. A similar operation will
make four 2-ply cords.
Whatever method he uses the worker will

quickly find how crucial correctly made
cords are to the ply-splitting technique; no
amount of tightening and pulling will
persuade softly plied cords to yield a firm
fabric.
28 » THE TECHNIQUES OF PLY-SPLIT BRAIDING

CHAPTER FOUR
Starting a Braid
. 1. FIXING THE BRAIDING CORDS

For convenience in working, cords
made as described in the last chapter
need to be fixed in sequence in some
way.
1.1. Fixing to a starting rod

A common way is to attach the cords
to a narrow rod, the starting rod. It can
be a narrow dowel, a stiff wire, a
jacquard lingo, a steel rod, anything
that is a little longer than the width of
the proposed braid and stiff and
narrow. It can be left in place when the
piece is finished or replaced by a cord.
There are several ways of attaching
cords to the starting rod.
Figure 9 Three ways cords can be attached to the
a_ Slip the two loops at one end of an starting rod.
individually made cord over the rod,
as in Fig. 9a
b Double a cord over the rod and fix

it in position by passing one half
through the other, as in Fig. 9b. In
other words, the cord is just hung over
the rod and then secured by what in
reality is the first row of ply-splitting.
How this row is worked depends on
the structure being used.
c Double a cord, pass the rod
through between the plies at its central
point, leave a half-twist and pass the
rod through again; see arrow in Fig.
10a and the resulting attachment in Figure 10 a—b Stages in producing the attachment
Fig. 10b and 9c. in Fig. 9c; for c—-d see over
STARTING A BRAID + 29
d A newly developed way is shown in Fig. 10c.
The rod enters the cord as in the above method but
with the working end going to the right. Leave a
half-twist as before, then pass the rod through but
in the reverse direction; see arrow in Fig. 10c and
the final result in Fig. 10d.
For the last three methods, individually made
cords of double the required length can be used.
Alternatively, lengths can be cut from very long
cords made as described in the last chapter, with
their ends secured in some way to prevent unplying;
e.g. by wrapping with yarn or taping as in Fig. 9c, or
Figure 10 c—d another attachment
method. even dipping in quick-drying glue.
NOTE
+ the initial spacing of the cords on the rod is not
important because the braid will take up its own
width after a few rows of ply-splitting.
* that method b is the only possible way of fixing
cords to a buckle, ring or other closed system, if this
is wanted in place of a starting rod.
+ that cords made by the continuous-cord method,

described at 3 in last chapter, are of course already
fixed on a rod.
+ that methods a and b tend to place the cords

more closely on the rod.
1.2 Supporting the starting rod

‘Though the Indian girth-makers work on the move,
holding everything in their hands, it is easier to work
sitting at a table. To a clamp at the far edge of the
table attach two lengths of Texsolv, long enough
nearly to reach the near edge. Slip the ends of the
starting rod into appropriately placed openings in
4
the Texsolv, so it is lying transversely on the table at
y a convenient position for the hands to work below it,
SSS
as in Fig. 11. Attached thus the braid can be easily

turned over, a frequent occurrence in ply-splitting.
INN
REE
In the absence of Texsolv, two doubled cords with
overhand knots every 10 cm serve well; they are also
| SSS
useful if the starting rod is too large for the holes in
‘Texsolv.
Figure 11 Supporting the starting rod Other methods are to catch the ends of the rod
with Texsolv attached to a fixed point. behind two clamps fixed to the table’s near edge, or

to tie a cord from one end of the rod to the other and
loop this over a clamp, similarly placed. Both methods
allow the braid to be turned over.
If using Texsolv or doubled cords, move the rod to
more distant positions as the braid grows in length.
Eventually it will have to be rolled up or a loop of braid
pinned to itself.
1.3 Fixing to transverse cords

~ A neat starting edge, useful for some projects, involves
splitting transverse cords with the braiding cords. So
there is in fact a narrow strip of ply-split darning
before the ply-split braiding begins. Fig. 12 shows two
methods assuming two 4-ply transverse cords are
used, conveniently stretched between clamps. In Fig.
12a start with a long braiding cord above the transverse Figure 12 Two types of starting edge
elements. Split the latter and pull a loop of the braiding with working cords splitting transverse
cord down through it to the required length. Split it cords.
again, leaving a quarter-twist between these splittings
and pull another loop down and so on. Wrap or tape
the midpoint of each loop and cut mid-wrapping to
give the required free-hanging cords for braiding.
In Fig. 12b start with the braiding cord below the
transverse elements. Splitting the latter, pull a small
BAW,
Senet ue
1)
loop of braiding cord up through it. Leaving a quarter- rz
twist, repeat, making the second loop pass through

and so lock the first and so on. Lock the final loop by
passing the transverse cords through it. Both methods
could of course be done with shorter lengths, allowing
colour changes.
Pack the splitting cords as close as_ possible,
otherwise this starting edge will prove longer than the
width of the braid that follows.
When working, hold this edge in a bulldog clip, tied to
a clamp on the far edge of the table, or pin it to a narrow
wood strip, held between doubled cords, as in the above
method, so that it can be turned back and forth.
A much simpler method is to have a single trans-
verse cord and split it with individual braiding cords.
Each is pulled through to its midpoint and the two
ends allowed to hang downwards. A quarter-twist is
left between each splitting of the transverse cord.
STARTING A BRAID « 31
1.3a Three-selvage braid
See Chapter 6 for a starting edge with a structure just
like the selvage, suitable for the POT technique.
2. STARTING MID-BRAID
In this method the ply-splitting is begun without the
cords being fixed in any way and at some point away
from their ends, perhaps at their midpoints. So it is
the first few rows which actually fix the cords
together. Work can then go in both directions from
this point.
3. SPLITTING IMPLEMENTS
Though the opening of one cord and the pulling
through of another is possible using only the fingers,
most workers employ some type of needle, hook or
pliers for this basic operation. Plate 3 shows a
selection of possible implements. From left to right
they are as follows.
a Needle, modelled on those used in India, whittled

out of wood.
b_ A length of heavy wire, bent as shown and with its

loop filed to a sharp edge, works almost as well.
c A two-ended needle, as used in Colombia, made

from a strip of wood or brass.
d Latchet hooks. Though there is obviously a

limiting thickness of cord beyond which they will not
function, their great advantage is that they give so
little disturbance to the plies of the split cord.
e A gadget used in the making of rag rugs. This is

excellent as the grip is spring loaded.
f Haemostats, obtainable from anglers shops. They

come in many sizes, some straight, some curved at
the tip, but all with a good serrated grip.
g A ‘gripfid’, originally invented for rope work by
Stuart Grainger, is ideal for dealing with thick cords.
See list at end of book for suppliers of some of these.
For simplicity, in all the descriptions in this book the

implement used will be referred to as a needle

4. BASIC SPLITTING MANOEUVRE
As in all braiding techniques, two elements
lying on opposite courses meet and are
interworked. Here the interworking takes
the form of ply-splitting. There are only
two possibilities. Either the splitting is
rightward, the right-hand cord lying on the
Z course being split by its neighbour to the
left lying on the S course, as in Fig.13a. Or
~ itis leftward , the left-hand cord being split
by the right-hand as in Fig. 13b. Figure 13 The only two ways cords can be split:
It is easier for a right-handed worker to a a rightward splitting,
hold and manipulate the needle in his right b a leftward splitting.
hand, so it is the former, rightward,
splitting which is here regarded as the
basic manoeuvre; it is now described in
detail.
4.1 Rightward Splitting

Hold the needle in the right hand. Between
the left thumb and fingers grip the cord to
be split, A in Fig. 14a, 2 or 3 cm below the
starting rod (or fell in subsequent rows).
Moving the left thumb to the right, roll the
cord so that the short length between
fingers and rod is completely unplied, as in
Fig. 14a. With a movement upwards and to
the left, push the needle centrally through
the four plies thus separated, stopping
when the eye is clear of the cord on its far
side. Two of the plies now lie over the
inserted needle, two under it, as shown in
Fig. 14b.
With the left hand put the end, or a loop,
of cord B, the next cord to the left, through
the eye of the needle, as in Fig. 14c. Now
withdraw the needle back through cord A,
dragging cord B with it; see arrow.
Remove cord B from the needle’s eye
and pull the rest of its length through the
split cord A. Give a firm pull to B and a
strong twist to A to re-ply it tightly and the
splitting is complete, as in Fig. 14d.
Through constant repetition the two
hands soon become accustomed to their
roles in this simple operation. Figure 14 Details of making a rightward splitting.
STARTING A BRAID « 33
Figure 15 Different ways of making a 4.2 Leftward Splitting
leftward splitting.
Almost all ply-splitting techniques involve the use of
both right- and leftward splittings. The latter can be
worked in several ways.
1 The commonest way is simply to turn the braid

Ny over. The splitting then becomes rightward and the
hands carry out their accustomed movements. It
a
will be assumed in this book that this is the method
ie
]i
2! p
xt, used.
2 Ambidextrous workers can make the hands

thy
G. os
change roles. The right unplies the cord B, as in Fig.

15a, the needle, now in the left hand, is pushed
upwards to the right to reach cord A and to pull it
down to the left. It is just the mirror image of the
rightward splitting described above.
3 Using the two-ended needle, the latter, held in

the right hand, is pushed downwards to the left
through cord B, which is unplied by the left hand.
Cord A is put in its eye, as in Fig. 15b. The needle
then carries on in the same direction pulling A
through B. The needle’s movement is thus
continuously in one direction, see arrow.
Of course an ordinary large-eyed needle would
also do this. But the two-ended needle has the
advantage that it can be used for both leftward
splitting as just described and for rightward, used as
in Fig. 14.
4 A tubular form cannot be turned over to convert

left- to rightward splitting. So methods 2 or 3 can be
used or perhaps more easily the piece can be turned
upside down, when the splitting is rightward, as in
Bigmisc:

CHAPTER FIVE
Explanations of Diagrams
Usually a ply-split fabric is so extremely close

and dense that very little of its internal
structure can be made out from the surface.
‘To present this structure and the manner of
achieving it in clear diagrams, reality has to
be greatly distorted.
1. ANALYTIC OR STRUCTURAL DIAGRAM

A very expanded representation of a finished
braid indicates which cord splits which to
produce a particular structure or design.
The cords are shown lying on widely-
spaced S- and Z- courses, forming a regular
and open grid. At each meeting point, the
cord which is split — and therefore visible — is
shown crossing over the splitting cord on the
countered course. See Fig. 16a.’[o emphasise
this and to enable cords of two colours to be
distinguished, the crossing point has either a
dark rectangle (for a split dark cord) or a light
rectangle (for a split hght cord). It will be
understood that these rectangles are the only
parts visible when the cords are pulled tight
in the actual braid, so they give an extended
impression of the design being made;
compare Fig. 16a with 17a.
The degree of twist left in a cord between
successive splittings can be indicated by fine
angled lines. For a 4-ply cord, one line means
that the very common quarter-twist is left,
two a half-twist and four a whole twist, as in
Fig. 16b.
The angle of the line also indicates the
direction of ply of the cord, usually Z, as in
the diagrams. If no such line is shown, it is Figure 16 a—b An analytic diagram;
always assumed a quarter-twist is left c—e stages in a working diagram.
EXPLANATIONS OF DIAGRAMS « 35
2. WORKING DIAGRAMS The numbering of cords in working
A series of representations shows the diagrams is sometimes right to left,
working process, stage by stage, enabling sometimes left to right, depending on the
the reader to reproduce the technique in method of work. This should not lead to
question. problems because these numbers will only be
The necessary splittings are indicated by of assistance the first time a technique is tried
arrows, as in Fig. 16c and 16d; sometimes out. Once it is learnt, they quickly become
backed up by words, e.g. cord 2 through 1; superfluous.
cord 4 through 3 and 1; or more concisely, It will be noticed that as the fell in a
De 14 oe 5 le working diagram moves downwards, an
‘These are normal rightward splittings but analytic diagram is gradually formed above it.
the splitting in Fig. 16e is leftward. It can be
worked in any of the ways described earlier. 3. DESIGN DIAGRAM
However in this book it will be assumed that Sometimes, to illustrate designs, a more
the method of turning the braid over to realistic convention is adopted which
convert leftward to rightward splitting will be concentrates on the surface patterning.
used. So the move shown is written: turn Though not showing structure or method of
over, 2 > 4. working, it should give enough information
Thus whenever leftward splittings are to allow the experienced to reproduce the
seen, the worker must turn the braid over to pattern shown; thus Fig. 17a is the design
convert them to rightward splittings. This diagram for the braid in Fig. 16a. Fig. 17b
repeated turning of the work back and forth shows graphically the relationship between
is at first confusing, because the design and an analytic diagram (top) and a design
its mirror image are seen alternately on the diagram (lower).
braid whereas the paper design being
followed only shows one of these.
>
OY
S O S
SII
SRE
Ye
.
Figure 17 a A design diagram;
rire
b how this relates to an analytic diagram.

CHAPTER SIX
Plain Oblique Twining
1. INTRODUCTION
In plain oblique twining (POT), each cord lying on its
oblique course alternately splits and is split by the
cords it meets on the opposite course. In the
commonest form the minimum amount of twist, one
quarter-twist, is left in the cord between successive
splittings. In this close form of POT, a cord is only
seen when it is itself split, two of its plies showing on
the front, two on the back of the braid. When it is the
splitter, it is completely hidden, being covered by two
plies lying over it, two under it. The result is that each
cord is visible at one meeting place and concealed at its
next and so on. It is this alternation, visually akin to the
surface texture of plain oblique interlacing, which
after much indecision led to the choice of terminology
— plain oblique twining.
2. METHOD
The technique is best understood by making a small
sample.
Double four cords over a starting rod, to which they
will be fixed by the first row; see Fig. 18a. The eight
free-hanging cords are numbered 1 to 8, right to left,
which is the usual direction of working. These
numbers soon become superfluous as a cord is always yao ee
split by its immediate neighbour to the left.
ROW 1 LONGROW FIG. 18b

Pull cord 2 through 1, as shown in the diagram. The
arrow and rectangle indicate this. Then pull 4 through
3, 6 through 5 and 8 through 7 to end the row. This can
be written as 2>1;4>3;6>5;8> 7. O @ O
To attach the cords tightly to the rod and to Do 825.9683. 4-1 2
minimise the size of the starting loops if the rod is
finally removed, take all the splitting cords in one hand Figure 18 a Start; b Row 1 of POT:
and the split in the other and pull them strongly apart. for 18c see over
PLAIN OBLIQUE TWINING + 37

ROW 2 SHORT ROW FIG. 18c
It is the essence of the POT structure that row by
row the cords change roles. A cord which is a
splitter in Row 1 is itself split in Row 2 and vice
versa.
This change can be seen beginning in Fig. 18c
where the arrowed cords, which were split in row
1 are all splitting the adjacent cord to the left in
Figure 18 (continued) c Row 2 of POT Row 2.
To establish the oblique movement of the
cords, the outermost cord on each side, that is
cord 2 and 7, take no part in this short row. They
can be looped back out of the way as suggested in
the diagram.
As explained in Chapter 5, the leftward
splittings in Row 2 imply that either the worker’s
hands must change roles or he must turn the
braid over. The latter more common practice will
always be followed in this and later descriptions.
When it is turned over, work again begins at
the right. So split cord 8 (a splitting cord in last
row) with the adjacent cord to the left, 5 (a split
cord in last row), then split 6 with 3 and so on,
always taking the next cord in sequence and
splitting it with its neighbour to the left. This can
be written; 5 > 38; 3 > 6; 1=4. Rememberito
tighten this row by pulling the split cords against
the splitting.
NOTE
* that the last cord pulled through in Row 1 is
the first cord to be split in Row 2, cord 8 in this
instance. So hold on to it as the braid is turned
over and dispel any doubt as how to start the
second row.
* that because the outermost cords are not

involved in this row, there are only three splittings,
compared with four in the first row. This
alternation of long and short rows is a natural
feature of any form of oblique interworking
performed on a transverse fell. The total number
of splittings in these two rows, which make up a
cycle of the technique, is always one less than the
number of elements being used — an instance of a
basic law of braiding. So here 4 + 3 = 7, for a braid
of 8 cords.

ROW 3 LONG ROW FIG. 18d
As Fig. 18d shows, the splittings for the third row
are all rightward, so turn the braid back to its
original position. The two outermost cords ignored
in the last, short, row are brought back into the
braid to make it a long row. The row starts on the
right as usual, each cord being split by its neighbour
to the left. So split 2 with 4 and then 1 with 6.
The latter splitting is the first time that a cord
split in a previous row is split again. Cord 1 was
split in the first row and is now split again. So it is
the first time that the degree of twist between two
successive splittings of a cord can be controlled to a
quarter-twist, an essential feature of the close form
of this structure. It can best be done as follows.
Unply cord 1 to such an extent that it can be Figure 18 (continued)
plainly seen that the same two plies are uppermost, e Row 4 of POT.
between the left-hand fingers and where the cord is
visible in row 1.’ These two plies are shown heavily
shaded and unshaded in Fig. 19a. Now allow the
cord to re-ply for a quarter turn, which naturally
changes the uppermost pair, as in Fig. 19b. What
was the right-hand ply (heavily shaded) is now on
the left, and is joined by a ply previously hidden
(ightly shaded). Push the needle through the cord
under these two plies, pick up the next cord to the
left, 6, and draw it through in the normal way. Do
exactly the same with the next two cords and so on.
NOTE
* it is only as a cord is unplied in the way just
described that its two plies, visible in the row before
the last (i.e. in Row 1), take on their final and
distinctive parallel alignment.
ROW 4 SHORT ROW FIG. 18e

Turn the braid over to convert the leftward
splittings in Fig. 18e into rightward. Then work
exactly as for Row 2, remembering that the twist
in all split cords must be controlled as described
above. So unply a cord until no twist is visible
between the fingers and where the cord is seen in
Row 2, then let it re-ply for a quarter turn before
passing the needle through. Remember this is a
short row, so the outermost cords are not involved.
Repeat Rows 3 and 4 which together make up the
Figure 19 Leaving a % twist between
complete cycle. successive splittings.
PLAIN OBLIQUE TWINING « 39

3. CORRECTING MISTAKES
ODP OLD RS JPR
If a splitting has to be undone, overtwist
the splitting cord so it becomes almost KRWF Sy
ee LAMEN | isy
lkme aNiNip
rodlike and so can be pushed back
through the cord(s) it originally split. ~\ Ci SO WI Lx Ui
The plies of a cord are disturbed if it is Ap ip i ‘ ((l )
pe, OX °™VVP
YIP <X\ OS
“Hf
gripped inadequately in an attempt to
K
AMK
ey QS SS
SK
pull it through. P
If a mistake is discovered some rows
back, the splittings can be undone EX GF Qoitr
selectively to create a triangular opening : ‘) 7 NQY ipSS OH
with the faulty splitting at its apex. The
fault rectified, the triangle is recon-
structed from the apex downwards with
ever lengthening rows, turning over
between each as usual. Stop when its fell
is level with that of the rest of the braid.
Undoing whole rows is_ seldom
necessary — a great saving of time,
especially in a wide braid.
4. STRUCTURE OF POT
Fig. 20a shows the structure in a highly
diagrammatic way, each ply of every
cord being visible. It shows plainly the Figure 20 a Highly diagramatic view of POT;
quarter-twist left in a cord between b more realistic view.
successive places it is split. A result is
that the two plies which here come to the
surface are constantly changing in a
regular sequence. Each ply appears at
two successive points, first as the right-
hand of the pair then as the left-hand.
This circumstance is not so obvious in
the more realistic diagram, Fig. 20b, and
is completely obscured in the fairly
lifelike Fig. 21. See also Plate 4 onwards.
Here the cords lie close together and the
same two plies apparently surface every &
time. The quarter-twists carefully left in
a cord are hidden by the plies of the cords
it splits. In the emphasised cord, the
dotted lines show what is happening to
these concealed plies and the numbers
show how each appears twice on the
front and of course twice on the back. Figure 21 Lifelike view of lower end of braid in POT.

4a. Selvage with unsplit cords
Though, worked in this way, POT gives a strong
neat edge, it can be further strengthened by leaving
unsplit cords at the selvages which then acquire a
wrapped or rolled appearance.
After a normal short row, work a row in which the
outer pair of cords on each side which should be
involved are left untouched; see Fig. 22a where
cords 1 and 2 and 7 and 8 are the untouched cords.
- In the next row include the outermost cord on
each side. So cord 8 is brought in and is split by 6,
and cord 1 is brought in and splits 3, as in Fig.
22b. These cords must consistently either pass over
or under the adjacent cord. Here it is over at one
selvage, under at the other. Repeating these two
rows, always ignoring the outer pairs in one row and
bringing in the outermost cord in the next, produces
the selvages seen in Fig. 22c.
5. NUMBER OF CORDS
The above description assumes there is an even
number of cords. It will be found that with an odd
number all rows are exactly the same, there being no
long or short rows. Depending how the first row was
worked, all the rows either begin like a long and end
like a short row, or begin like a short and end like a
long.
6. CHARACTER OF POT
As the work proceeds, the typical surface texture will
become evident, superficially resembling plain Figure 22 Making a selvage with unsplit
oblique interlacing. The way each cord is alternately cords.
split and splitting, plus the ‘quarter-twist left
between these splittings, makes this the most firm
and stable structure produced by ply-split braiding,
so it is ideal for any free-standing 3-D project. If well
made with a tightly spun stiff fibre, like goat hair, it is
almost impossible to shift one of the plies on the
surface. This made its analysis difficult when the
structure was first encountered.
At any point its thickness is about double that of a
cord, being made up of a splitting cord plus the plies
over and under it. If worked tightly the cords lie on
courses which run at 45°, so the grid adopted for the
diagrams is a realistic one.

7. VARIATIONS IN STARTING
The other methods of attaching cords to the
starting rod, described in Chapter 4, can be used. A
different method which dispenses with a rod gives a
perfect starting edge with no loops or irregularities.
It is structurally the same as the lengthways
selvages.
The slightly awkward beginning is best done by
laying the cords on a table and pulling cords one by
one, as far as their midpoints, through their
neighbours to the right, as in a short row; see Fig.
23a. Alternatively they can be hung over a fine cord
as in the diagram, and this cord removed later.
Then either turn over carefully, or use leftward
splitting, to make a long row ensuring that only a
quarter-twist is left between these and the first
row’s splittings. Pull split and splitting cords firmly
apart. Turn back for the next short row and so on.
After about 4 rows there is enough braid to be
gripped by a clip or pinned to a wood strip and the
braid can be supported normally. The starting edge
Figure 23 Making a perfect starting edge has the typical appearance of the right and left
without the use of a rod.
selvage, into which it continues neatly and without
interruption at the braid’s upper corners, making it
the most satisfying of all the starting methods. See
Fig. 23b.
8. DESIGN IN POT
8.1 Design with uninterrupted passage of cords

from selvage to selvage
So far cords of only one colour have been con-
sidered; once two or more colours are used the
typical patterns to be described appear.
NOTE
* that the course taken by these cords and the
resulting designs are identical visually to those
found in plain oblique interlacing. It is only the
interworking of crossing cords which is different,
i.e. one splits the other, instead of passing over or
under it.
The simplest arrangement on the starting rod of an

equal number of dark and light cords is to place all
of one colour to the right, all of the other to the left.

‘Then as row follows row, there is a gradually
increasing number of dark with light
splittings at the centre. Eventually these
stretch from selvage to selvage, then they
decrease and a simple repeating pattern is
seen to emerge. Down the centre of the braid
runs a series of diamonds where dark and
light cords meet, their visible plies giving a
small chequered pattern. These chequered
diamonds are flanked by triangles alternately
Figure 24 a-c Three designs resulting from
different colour arrangements at the start of a
16-cord braid. For c see over.
7 SOs
SE SSSO 7 ©
YW
REA wg,&
x QS
\N
OOO ON
\
BODO \
USS Ox
YY
>> 4
xs<2 OCVG
>»
VCAQ
CX? OQ QR
L <>
Cx &
Ve
<s
OX
SSSSLy y
Ssse
ROP
<
22
with a starting colour sequence (SCS) of 8 design ingredients but on a smaller scale. This
light, 8 dark. is the hound’s-tooth pattern, familiar to
If the same 16 cords are arranged 4 light, 8 weavers, turned through 45°, and is seen
dark, 4 light at the start, the pattern becomes centrally at the bottom of Plate 6, right.
symmetrical. There are now alternately light In the ultimate division of dark and light
and dark diamonds down the centre, flanked cords, they appear alternately on both
by chequered diamonds, with light and dark courses giving fine hairline patterns; these
triangles at the sides, as in Fig. 24b. See are considered later as a separate entity.
similar braid in Plate 4, right; on the left is a There is no need for symmetry in the SCS,
piece showing how different the design looks nor does there have to be an equal number of
if there are more dark than light cords. the two colours; Fig. 24c shows both these
A further division of the colours into 2 light, options. Naturally more than two colours
4 dark, 4 light, 4 dark, 2 light gives the same can be used.
Cc
> oo.
OoXe4 Sy
ae \
66 “xSSs
CSe RC2
cS
SOc™
CO”
ORR,
DORRSKS, OR
SP K S
’y Q - os S

Relationship between starting colour sequence
(SCS) and course colour sequence (CCS)
When designing a braid it is important to realise that
the way colours are ordered on the starting rod
differs from the way they appear on either of the
braid’s courses. This is because of every pair of
adjacent cords fixed to the starting rod, one is made
to run on the S-, one on the Z-course, by the manner
in which the first row of POT is made.
In Fig. 25, the 8 cords are numbered as they in lie
in sequence on both courses and down both
selvages. So the course colour sequence (CCS) is
12345678. Tracing these up to the starting rod, the
starting colour sequence (SCS) is seen to be
18273645. This looks confusing but has its own
logic, the first half of the numbers written to the
right are then interspersed with the second half
written to the left.
Applying this knowledge it is easy to work out the
correct SCS to produce any CCS. If, for example, a
CCS of ABCDAABB was wanted, write 1 to 8
under it thus —
ABCDAABB
12345678 -
then read off the colours above in the sequence,
18273645, giving a SCS of ABBBCADA.
Figure 25 Showing relationship between

CCS and SCS.

oe
8.2 Design with interrupted passage of cords from
selvage to selvage
XX
The type of design already described will repeat
faithfully and endlessly as the CCS of the cords runs
unchanged down the whole length of the braid. A new
set of possibilities arises if that sequence is altered
ee-
during the process of working. There are two ways of
eR OR
doing this; by introducing twined linkings or by
producing slits in the braid.
wews
8.2a Twined linkings
As explained in detail in Chapter 11, a twined
linking between two cords makes them swap
courses. So in this context it is always worked
=PSS
oe
@ SE between a dark and light cord, thus altering the
colour sequence of both courses. Twined linkings
‘a
SSPERK
SKS
can be placed longitudinally or transversely in a
OK
braid.
Longitudinally placed twined linkings

Making a single-colour border
Fig. 26 shows part of a braid with a dark border. To
Figure 26 Placing twined linkings preserve this border, whenever a light cord from the
longitudinally to make a single-colour border. central design approaches within a set distance of
the selvage, it is twine linked with a dark cord. It is
thereby deflected back towards the centre and never
reaches the edge. The twined linking is represented
as usual with a circle.
As this only happens on alternate rows the extra
bulk of the twined linking does not build up
excessively. The twined linkings can be of the R/R or
R/L type, each giving a slightly different toothed
appearance. Naturally no twined linking is needed
where a dark cord from the centre reaches the
border; it can run on its course all the way to the
selvage; see Plate 5, centre. But if the border were of
a third colour, both dark and light cords would have
to twine link with it.

penne
Becca
dteorcenstonentonts
BeHlstotatetatatatatatatata
sscronattetotntaee stone
SX
OK
KORO ates
enncee
8 WY SOECECCCSC OS:
C
ReCCCOS OS
ROO
WLM
RRR
MMMM MMMM
| E
OKO AORKOEORKAOKAOROAORKOKEG |
RR RRO. ||
|
EIEN
WOOO OOOO CEI
2|SORES
EEE ERI
SS SEES |
PLAIN OBLIQUE TWINING
+ 47
BOR OOCOC
OOO OPP ae:
ee
Figure 28 Figure 29 Designs created by transversely

at the top A simpler way of representing Fig 27; placed twined linkings.
lower part Further design possibilities.

Influencing central design cord turns through a right angle. The result,
Longitudinally placed twined linkings in the as seen below the arrow, is that the chequered
central area can keep the dark and light cords diamonds (outlined by a wavy line) still
separate so that they form distinct length- appear but the central area remains light as at
ways stripes. See the top of Fig. 27 and of the start. So the braid will show a succession
Plate 6, right. of light diamonds down the centre, flanked
Starting with such a central dark stripe and by chequered diamonds, edged by dark
letting two dark cords on each side move into triangles. See Plate 6, left. As always the
the light area normally and preventing the twined linkings will give a slight variation to
next two by means of twined linkings, and the surface texture across the centre of each
continuing in this way, produces the effect chequered diamond.
shown in the main part of Fig. 27 and upper Fig. 29b shows an elaboration of this
part of Plate 6, right. This diagram can be simple idea in which 2 dark cords on each
represented in the simpler way seen in the side are deflected from the rest by twined
upper half of Fig. 28, where dark and light linkings at the level of the arrow. They then
cords are shown as solid or dotted single make thin lines which cross and re-cross in
lines. There is no indication of which cord the light central area. Flanking this area are
splits which, it being assumed the structure is small chequered diamonds (again outlined)
POT throughout. The rest of the twined with twined linkings at their mid-lines; see
linking diagrams will be of this type. Plate 7, right. The variation in Fig. 29c
By working in the reverse way and letting begins in a similar way, but at the top arrow
some cords from the sides move across the only the outer 4 dark with light crossings are
braid and others be stopped, various patterns twine linked, and again at the lower arrow. A
can be produced in the central diamond, see wavy line again outlines the chequered areas,
bottom of Fig. 28, where there is a 2 dark, 2 see Plate 7, left.
light colour sequence on both courses; see
also bottom of Plate 6, right. NOTE
+ in the plates referred to there are more
Transversely placed twined linkings cords than in the diagrams.
As a repeated operation * itis only when drawing out such diagrams
that the ingenuity and accuracy of traditional
Starting with 8 dark, 16 light, 8 dark cords as
examples is fully realised. A twined linking
at the top of Fig. 29a, the cords will move
has only to be misplaced by one row, or
until at the level of the arrow every crossing is
shifted one crossing to right or left, for the
of dark with light. Continued thus, the
whole design to be lost.
typical design already seen in Fig. 24b would
* how in these examples the light cords
appear. If, however, at this point, twined
never reach the dark border so no twined
linkings are placed all across, as shown, every
linkings are needed there.

As a single operation
Different designs using the same total number of dark and
light cords can be worked, one after the other, down the
length of a braid. But the cords’ colour sequence at the
start of one design is likely to be different from that
required for the design preceding and following it. Luckily
the sudden switch from one CCS to another can be easily
achieved by using twined linkings, precisely placed in a
single transverse row. This switch can at first appear
baffling and is best understood by following an example.
Fig. 30, and Plate 8 which follows it exactly, starts at the
top with 12 dark and 12 light cords. These if worked
normally will give the familiar pattern of chequered
diamonds flanked by dark and light triangles. To change
the colour sequence, work must stop halfway through
such a diamond when there are dark/light crossings all
across.
At such a point, a in Fig. 30, the 6 right-hand crossings
have been twine linked. As the design below shows, the
colour sequence has changed to 6 dark, 6 light which gives
a series of dark and light diamonds running down the
centre. The 6 twined linkings show as a slight alteration to
the chequered pattern.
At the next change, b in Fig. 30, the work has again
stopped when every crossing is of dark with light cords
and the outer three on each side have been twine linked.
This splits up the colours still further into a 3 dark, 3 light
sequence, giving more and smaller diamonds in the
design.
At c in Fig. 30, where all crossings are once more dark
with light, every third one has been twine linked making
the cords alternately dark and light on both courses and so
giving a hairline pattern.
At the final stage, d in Fig.30, twine linking every other
crossing re-establishes the original colour sequence of 12
dark, 12 light, seen at the foot of the diagram and Plate 8.
An important practical result of these subtle
manoeuvres is that the cords can be set up in some very
simple colour sequence on the starting rod, the worker
knowing he can transform it at will into any sequence he
desires.
Of course twine linkings can be placed irregularly, e.g.
at b in Fig. 30 there could be 3 at the left and 4 at the right.
Figure 30 Use of a single set of The drawback is that it; will take many : more apie
rows than:
transverse twined linkings to alter normal before all crossings are again dark with light, if
colour sequence in the cords. another colour sequence is wanted.

8.2b Slits
A slit in plain oblique twining results from the failure
of two cords on opposite courses to cross at their
meeting point. So in Fig. 31, row 1, the light cord A
(on Z course) should be split by B (on S course) at
the point marked X. But both cords, turning
through a right angle, begin to move on the opposite
course in row 2. The same happens with cords C and
D which fail to cross at Y and so on.
Gradually a slit develops extending downwards as
shown. While producing this split, the worker is in
effect making two unconnected braids each with its =
WD
of
NO
own two selvages and its own sequence of long and

short rows. See Plate 9, right, for the start of a long
central slit. This situation continues until two Figure 31 Introducing a slit.
relevant cords are allowed to cross (as in row 7 in the

diagram), thereby closing the slit and joining the two
braids together. In the same way, and with no slower
rate of working, many slits can be made across the Figure 32 Staggered slits producing an
width. open structure: a method of work;
b actual appearance.
Used as structural features
A slit can be used as a decorative device or as an
essential element in a 3-D construction.
If slits are alternated between two positions a very
stable open meshwork is produced. It is found
traditionally in an anti-fly fringe worn by horses in
India, and is there made from 2-ply cords; see Plates
118 and 119. Using 4-ply cords, start with a number
divisible by four; see Fig. 32a.
Row 1 — a normal long row.

Row 2 — a short row, but only ply splitting every
alternate pair, beginning with the first available pair
at the start of the row. Here and subsequently it is
essential to leave a half-twist between successive
splittings of every cord.
Rows 3 and 4 are the same as rows 1 and 2. These
four rows produce a series of four-cord braids
separated by slits.
Row 5 — a normal long row.
Row 6 — a short row, ply splitting alternate pairs, but
using the pairs omitted above. So ignore the first
pair and start with the second.
Rows 7 and 8 are the same as rows 5 and 6.

These four rows close the original slits and make
new braids, one less in number than previously,
separated by newly sited slits. Each is formed by two
cords from the right-hand and two cords from the
left-hand braid above. The two cords which are
inevitably left over at each selvage are interworked in
the long rows and make thin two-cord braids, as seen
at the right of the diagram.
Repeat these 8 rows. It will be seen that the long
rows are the same throughout; it is how the splittings
are placed in the short rows which control the
position of the slits.
Although Fig. 32a is accurate as a working diagram,
the finished braid has pronounced almost circular
holes, as in Fig. 36b and Plate 9, left. In Fig. 32a,
cords marked with dashes and others similarly placed
are shown taking a curved path, longer than the cords
they cross. But in POT every cord takes as straight a
path as possible, as it is pulled tightly up against the
previous row. This forces the narrow braids apart and
gives the structure its characteristic openness.
Leaving only a quarter-twist between successive
splittings leads to slightly smaller openings.
Using slits to alter colour sequence in the cords

For a slit to effect the cords’ colour sequence it has to
be so placed as to prevent a dark and a light cord
from crossing; this happens to cord A and B in Fig.
31. Turning through a right angle, they swap
courses. So a twined linking achieves this effect by
making two cords cross twice, a slit by preventing
them crossing at all. Comparing the two methods,
visually a slit gives a much cleaner boundary
between dark and light areas, but physically it can
weaken the structure, gaping if prolonged. See the
borders of the central dark and light diamonds in
Plate 10, right, and compare with the dark triangle at
the top of Plate 6, right.
The two methods can be combined so that a
colour boundary made by a slit is occasionally
bridged by a twined linking. Plate 10, left, shows
these two methods alternating at the left border and
elsewhere; compare with the borders in Plate 5
which have only twined linkings.

8.2c Oblique hairline stripes The above implies that the direction of the
If dark and light cords alternate on both hairlines can be altered by any manoeuvre
courses, fine oblique lines, hairlines, appear, which makes a cord change from splitting
the braid lacking any areas of solid colour. cords of one colour to splitting cords of the
‘These lines — analogous to the hairlines in other. Three such manoeuvres are now
woven colour and weave effects — are one described.
cord wide. They result from the eye reading
as one continuous line a cord’s visible plies Breaking the strict alternation of dark and
and those of the cords it splits. They are light cords, on either course, by making two
therefore aligned with the course of cords dark or two light cords lie side by side.
which split their own colour. S and Z In Fig. 33c, a change in the dark/light
hairlines are therefore possible, the latter alternation of the cords on the Z course makes
being used exclusively in Plate 11, left. two light cords (bracketed) lie together. As
In Fig. 33a, the cords on the S course are all cords A, B, C, D, on the S course cross these
splitting cords of their own colour giving S two, they all change from splitting their own
hairlines, while in Fig. 33b, which is the back colour to splitting the other. In this way an
view of the same band, cords on the Z course oblique dividing line between areas of S and Z
do this giving Z hairlines. hairlines is produced. It is a light dividing line
Note that S hairlines always appear as Z because formed by two adjacent light cords;
hairlines and vice versa when a braid is it could have been formed by two adjacent
turned over. dark cords giving a dark dividing line. The
existence of these two types gives subtlety to
the designs, which can incorporate one or
both types, running on one or both courses.
The cord colour sequences needed for
such designs can be achieved in several ways.
Figure 33 Producing oblique hairline stripes:

a S hairlines,
b Z hairlines,
c one method of changing direction of hairlines.

Mounting a special colour sequence on the starting rod
In Fig. 34, dark and light cords are doubled over the rod
with one single light cord added at the left. This gives the
alternating triangles of S and Z hairlines shown, separated
by a zigzagging light dividing line, shown by a heavily
outlined cord. The addition of another doubled dark cord
at the right would make the dividing line dark. Extending
the braid by mirror-imaging it to the left produces the
succession of diamonds seen in Plate 11, right. The top of
Fig. 35 shows a more complex SCS which gives rhomb-
oids of S and Z hairlines, separated by both dark and light
dividing lines. Patterns like this will repeat exactly as long
as desired. See Plate 7, right, at bottom.
Producing a special colour sequence with twined linkings

while making a braid
‘Twined linkings applied to a braid with an equal number
of dark and light cords can set their colour sequence so
that hairline patterns appear, such as those shown in
miniature on the right of Fig. 36. Once set, the patterns
keep on repeating automatically. The exact placing of
these twined linkings depends on what happened before
in the braid. Here it is assumed that the braid started with
6 light, 12 dark, 6 light cords, see Fig. 36, and has
reached the mid line of the two chequered diamonds,
when all crossings are dark with light, a row arrowed at
both sides of this and all the other diagrams.
Figure 34 S and Z hairline design
a The main part of the diagram indicates how the
resulting from a special SCS.
twined linkings, shown by circles, must be placed to
produce the corresponding designs shown at the right.
‘These are now described.
Figure 35 Design diagram for S and
Z hairline pattern. b_Hairlines in one direction all across. Twine link either
the 6 crossings shown or the 6 omitted. See Plate 11, left.
SO R2O<SOXS
enYX OK
0YAy Oe
AK XK
OOLy c Diamond of S on a background of Z hairlines. Place
Rog GK Log ogioh the twined linkings asymmetrically as shown; the
VSO
Oe SC Se CS Oe SY
et Ce
oo Ne yiKy KK ymWN
Oe ee diamond will be bounded by dark and light dividing
e, Y)
\ \
%,
Ly Ay Ky AK yA
lines. See Plate 12, left.
\
ROE Y
RN d Smaller diamonds. Place the twined linkings in the
short row before, or after, the dark/light crossing row.
KBKIRY
UY Y Ge YUY OO
Y GY
Which row is chosen decides where dark and light
MK JK Jr ix YX YS >)
dividing lines appear.
QZOQP ORLY?
REO OeRK ee ERY
em
e Diamond outlined only by dark lines. Twine link as
SRR III shown in the row before the dark/light crossing row. This

ne
We y
Figure 36 Using twined linkin hairline patterns: a starting position;

nC
b—g corre ct placings of twined XE= shown at right.

So
—

and the two remaining designs need a
twined linking of the first and last
crossing in a long row when this
happens to be dark with light.
f Oblique stripes. Place twined
linkings in one half of the braid in the
dark/light crossing row and in the
other half in the next row.
g Triangles. Place just the four

twined linkings shown and_ the
triangles will be divided by light lines,
as in Plate 12, right. Adding one to
the outermost crossings in the next
row (dotted circles) will make these
dividing lines dark.
To produce such designs but

starting from some other point than
that in Fig. 36a, just look at the colour
sequence of the cords at the bottom
edges of Figs 36b to g and place
twined linkings to produce this
sequence. As long as all crossings are
dark with light, this will always be
possible.
Twice-Split Cords
Upsetting the normal splitting/split
sequence of cords is a neat way of
controlling hairline designs and is a
manoeuvre which can be worked in
transverse, longitudinal and oblique
lines.
Transversely-placed
‘The twice-split cords, outlined in Fig.
37a, are produced as follows.
After the normal Row 1, do not
turn the braid over. Work the next
row, 2, with the usual rightward
splittings. All cords on the Z course
are thus split twice with the usual
quarter-twist left between these Figure 37 Using twice-split cords to contro! hairlines:
a above placed transversely to give zigzags;
splittings. Now turn the braid over b facing placed longitudinally and transversely to give
and continue normally. concentric diamonds.

This converts the Z hairlines above
the twice-split cords into S hairlines
below them. In the method described,
the hairlines will meet head on at their
change of direction; so repeating this
makes the hairlines zigzag con-
tinuously down the length of the
braid, as in Plate 13, right. If in Row 2
the splittings are alternately right-
and leftward, the hairlines meet
differently.
A transverse line of twice-split cords
could be used in such designs as c and
d in Fig. 36 and would automatically
produce the mirror image of what
My came before it; see Plate 13, left.
OY
Longitudinally-placed
‘Twice-split cords placed one above
the other give a longitudinal junction
between S and Z hairlines; see top
third of Fig. 37b. Note that one half of
each row has rightward, one half
leftward splittings. These can be done
without turning the braid over, the
hands changing roles mid-row. Take
care to stack the twice-split cords
accurately one above the other.
A neat alternative for the non-
“aS
% / ambidextrous is to start with a half-
row running from the centre to the left
selvage, as in Fig. 38a. All subsequent
rows, even though not strictly on one
level, can now be worked straight
across with rightward splittings,
turning the braid over in the normal
way; see Rows 1 and 2 in Fig. 38b-—c.
The only slight complication is at the
x S vy centre where in one row a cord is left
unsplit (arrowed in Row 1), and in the
next row a cord is split twice (outlined
in Row 2)
Combining longitudinal with
transverse lines of twice-split cords,
as in the lower part of Fig. 37b, gives
a design of concentric diamonds,

which can be extended widthways; see Plate
ce PO
6 Scns
Gof
Seo aor
eat
Oogo bpgwd
= e BastEYP
Ro
SS
rr
14 , right. Fig. 39 shows how a different SCS AR HY
ee) 8 gt
128
en)
0
places the conce ntric diamonds dramatically

2 oO &
Soo
a
OOF an}
|
en
, oO 4
RARBRAY
Lo
O O — toe) A method of working with twice-split
©8
2533o2co ce 2 (e,o)—=za wT & Ti =>
eee

Be ©[O, -@, © ©2-0° ce Obliquely placed
alive
adjacent cords of the same colour; compare

Fig. 41 with Fig. 34. But they have the
advantage that a motif can be placed
anywhere at will, unrelated to the SCS; see
the diamond in Fig. 42 which is outlined by
light cords.
oe re 41 _ Using twice-split cords, placed

obliquely, to give alternating triangles of S and Z
hairlines.

oterteets
8.2d Using multicoloured cords
All the techniques described so far have used
Se
cords consisting of either 4 dark or 4 light
plies. Some little explored possibilities in
POT stem from the use of cords each of
which has both dark and light plies. For
SEES
example, cords with 2 dark followed by 2
light plies (made as described in Chapter 3)
ive some interesting although small-scale
SOR:
effects.
When such a cord is worked in POT with
JRA
indicates, the emphasised cords show 2 light
plies (arrowed on both courses), then a dark
and a light together, then 2 dark, then a light
and a dark together; the sequence then
repeats. Fig. 44b shows how this can be
their plys’ four different appearances

combine in some meaningful way. Three
examples are now described.
S ».SS
Figure 44 How multicoloured cords appear in POT.
VE Y;
\
SRK
RRS
SOD UR
OY
SSSESS
SSNSS
b

Transverse stripes © ad ©
Starting with 16 cords, work a long row ©)
splitting cords so that the 2 dark plies are
uppermost; see Row 1 in Fig. 45a. Turn
the braid over and in Row 2 split so that
a light ply lies above a dark. Turn again
and in Row 3 split so that a dark ply lies
above a light. Turn again. In Row 4 the
dark plies must be uppermost as it is
made, giving the visible light plies in the “c
view from the front in Fig. 45a. KOKON
Continuing in this way for the 8 rows
shown produces the stripes, seen at the
top of Plate 15, right.
The first 3) tows *set the” cords;
thereafter, leaving a quarter-twist
between splittings automatically brings
the correct selection of plies to the
surface for subsequent rows. There is
one exception at the left selvage, where
a half-twist must be left as each short
row is started, in order that the cords
show the correct plies and agree with
others in that row. This is indicated by
two fine oblique lines across the cords;
see Fig. 45a.
Longitudinal stripes Figure 45 Using multicoloured cords to give transverse

: d longitudinal stripes.
A repeat of only 4 rows is needed for Spee eee oy enue,
the stripes shown in Fig. 45b and
bottom of Plate 15, right.
In Row 1, a long row, the visible plies
are either both light or both dark. In
Row 2, they show all the four possible
colour arrangements. Row 3 is as
shown, Row 4 copies Row 2 exactly. As
with the last example, the first 3 rows
set the cords correctly and the design
appears automatically afterwards. The
one exception is that Row 2 and Row 3
must begin by leaving a _half-twist,
indicated by the two fine lines in the
diagram. The interlocking branches
between the stripes give it an Escher-
like character, seen in the bottom half
of Plate 15, right.

Combining one-colour and multicoloured cords
This represents a large field for experimentation. It
is best to work out a design on a grid first, though
interesting chance discoveries may be made if cords
are set in some unusual way and then worked to see
what appears. Plate 15 left and Fig. 46 show a simple
use of this idea. Start with 6 dark cords at the centre
and 2 at each side with 2 dark and 2 light plies. Then
working as in the diagram will produce the neat light
diamond on a dark background. In the two arrowed
places no twist is left between successive splittings.
Other possibilities arise when using cords with 3
dark/1 light plies or 3 light/1 dark plies and when
using cords with 3 or even 4 differently coloured
plies.
Modu Ramji with black gorbandh of his

making, at Kakrala, February 1989.
Figure 46 Combining one-coloured and

multicoloured cords.

8.2e Dimensional possibilities with POT
Plain oblique twining is ideal for 2- and 3-D fibre
structures. If the right cords are used, such
structures can combine some of the stiffness of
basketry with the flexibility and surface interest of
textiles. So free-standing objects can be made which
are yet soft to the touch.
Changing a braid’s long axis

The way the cords lie on oblique 45° courses makes
it easy to turn the axis of a braid through a right
angle. It is worked as follows.
Decrease the number of splittings in an ordered
way so the braid acquires an oblique fell, finally
tapering to a point, as in Fig. 47a. Turn the braid so
the point is now uppermost as in Fig. 47b. Starting
at the top, work downwards, increasing the splittings
by including one more cord from the oblique fell in
eachirowe lhe sequence will be =a); 1 > 3:3.>2,
4>1;1>5,2> 4; and so on. turning over at each
row in the usual way. Eventually all the cords will be
engaged and the braid can continue normally, as in
Fig. 47c, its long axis having swung neatly through a
right angle with no interruption to the ply-split
structure.
There are now many dimensional possibilities
depending on whether subsequent oblique fells are
made parallel to or at right angles to each other.
Making them consistently parallel to each other
gives the zigzag braid seen in Fig. 48a. Making them
at right angles to each other and as close together as
possible gives a braid which begins to lie over itself,
spiralling into a stack with concertina-like properties;
see Fig. 48b. Combining these two options forces the
braid into almost any configuration, only limited by
the right angles its long axis must move through at
each change of direction.
With the rows of splittings decreasing to give a V-
fell, two half-width braids branch out sideways, as in
Fig. 48c. If they decrease to give an A-fell, these little
braids cross each other or angle forwards together,
as in Fig. 48d. If the braid has hairlines, they will
change direction at each of the braid’s corners, i.e.
along the dotted lines. Figure 47 Turning braid’s long axis
through a right angle.

OS
58
6s
<<
Se
NeotsSS
WeickARS SOT
PLY
KK P55
SS
Maes
Th
we
XSne
L
eeaetee
OFO
MUL
KX
px XJ
RY
RS
ne
em
Figure 48 Various results of continuously turning long axis of braid through a right angle.

Forming a tube
From a flat braid
A flat braid can be converted into a tube in three
different ways.
1 By joining the selvages. Stop after a short row. There

is a free cord hanging at each selvage. Fold the braid so Figure 49 Ply-Splitting upwards when
these selvages meet and make the left-hand free cord working a cylinder.
split the right-hand. The tube has now begun.
Normally the next row would be worked from the
back; this is now impossible. So turn the braid upside-
down and the row can now be worked with the usual
rightward splittings, the cords being pulled upwards
rather than downwards, as in Fig. 49. For the following
Figure 50 Dividing cords into two
row turn the braid back and so on. layers to start a cylinder.
Because each row is now a complete ring of splittings,
it does not matter where on the circumference a row
starts. Any cord is pulled through its neighbour to the
right and knotted at the end to act as a marker. The work
then continues all the way round until the marker is
reached and that row is finished. This means that every
row Is similar, involving every cord; there is not the usual
alternation of long and short rows. Obviously the tube
will be half the width of the flat braid it sprang from.
2 Dividing the cords into two layers. After a long row

separate the cords into two layers so that one pair is
kept to the front, the next pushed to the back. In Fig.
50a, the back layer cords have a dotted outline.
Now start a row using only the front layer cords.
Carry on round the left selvage, turning the braid over,
and then use the back layer cords. The 6 splittings
which make up this row are numbered in sequence in
Fig. 50b. Turn the braid upside down and work the
next row and so on.
3 Superimposing two layers. Work the braid so there is

a split down the centre. When there are two half-width
braids several centimetres long, cross one over the
other and begin to work these two superimposed layers
as a tube.
NOTE
* that a tubular braid must have an even number of
cords, otherwise there will be an odd cord left over at
the end of a circular row. A flat braid can have either an
even or an odd number.

Working as a tube from the start
Fix the cords initially over a stretched length
of twine, not a rigid rod. Start the first row as
shown in Fig. 23a. At its end release the twine
and curve the cords round so that the first
and last cord can be interworked. Turn
upside down and work the second row. This
will be a little difficult to handle until 2 or 3
rows fix the cords firmly together. See bottle
and jacket shape in Plate 16.
A single tube can branch into two. Pinch
the fell of the tube so that two sides touch at
some point. Start the next row by making
two cords cross from one side to the other.
Then the rows continue clockwise around
the two new circumferences.
Colour effects with a tubular braid

It is important to realise that the elements in
any obliquely interworked tube run in a
continuous spiral, half on the S-, half on the
Z-course. They do not repeatedly change
course as they do at the selvages of a flat
braid. The designs produced by cords of two
different colours naturally reflect this fact.
If the cords at the start have some simple
colour sequence like 8 dark, 8 light, 8 dark, 8
light, these will spiral down the tube, crossing
and recrossing, repeatedly giving dark, light
and chequered diamonds. There will not be
the triangles found at the sides of a flat braid.
If the starting sequence is 2 dark, 2 light all
the way round and then in the first row every
other splitting is a twined linking, all the dark
cords will run on one course, all the light on
the other. So every crossing from then on will
of necessity be of dark with light. From this
any colour sequence can be derived with the
Figure 51 Increasing width of braid by adding
aid of twined linkings.
cord centrally.

8.2f Increasing the width of a braid
Adding two cords centrally

If a single cord is added to the body of a braid, it
will upset the split/splitting sequence in the
structure. So the minimum that can be added is
two cords. This is neatly done by introducing the
centre of a long cord and letting its two ends be the
two cords, one lying on each course, as in Fig. 51a.
Work the row to the spot where the new cord is
to enter. Here the two cords, A and B in Fig. 51a,
which would normally cross fail to do so, but
engage with the new cord (shaded) as shown. In
the next row seen in Fig. 51b, cords A and B
cross and the two ends of the introduced cord
begin to move on opposite courses. There will be
some distortion of the braid as suggested in the
diagram. But due to the tightness of the
structure, it is often hard to spot where a new
cord has been inserted. The widening of the neck
of the bottle in Plate 16 was produced in this way.
On the other hand, the technique is a way of
introducing a cord of new colour or character
which is meant to be visible.
Adding cords laterally

Cords can be added singly or in pairs at the selvage
of a braid. After a short row split a new cord and
through it pull A, the right-hand free cord; then
split B and pull through another new cord. See Figure 52 Increasing width of braid by adding
Fig. 52a where the new cords are shaded. cords laterally.
Now if only one cord is to be added at each
side, continue with another short row, cords A
and B again being the free cords, followed by a 8.2g Decreasing width of a braid
long row and so on. The unused end of each new There are no ways of decreasing the
cord can be knotted or left as fringe. width of a braid as neat as those for
But if two cords are to be added at each side, increasing it. A flat braid can be tapered
bend over the new cords and make them the first by occasionally dropping off a cord
and last cords of the next (long) row, engaging either at one or both selvages, giving a
with A and B, as in Fig. 52b. Adding more and free-hanging fringe.
more cords in this way gives the gradually In a tubular braid, an adjacent pair of
widening braid an oblique selvage. cords (crossed in the last row) can be
By using the latter method, a braid can start pushed back into the central space and
with a point and then slowly widen until it excluded from the next row. Pulling the
reaches the required width; see Fig. 52c and the cords tightly in the latter will prevent any
top of Plate 17. obvious break in the surface.

8.3 Variations in the structure of POT
8.3a Using a 2-ply cord

All that has been described so far for a 4-ply
cord is applicable to any cord with an even
number of plies, all of which can be split
centrally and will give comparable
structures. Using 4-ply cords, the two visible
plies lying above a splitting cord are parallel
to each other and to the parent cord, no
matter in which direction the cord is plied.
Hence the visual similarity to an extended
plain oblique interlacement. But with 2-ply
cords, the single visible plies pass over only
one splitting cord and then under the next.
This constant movement forces them out of
line with the parent cord, tilting them
towards the horizontal or the vertical.
This tilt is controlled by two factors, the
cord’s ply direction and the direction of the
course it is lying on. Fig. 53a takes two cords
in isolation, one S- and one Z-plied, and
shows how their plies lie nearly horizontally
or vertically in accordance with these factors.
So in a braid made of. Z-plied cords, the
latter lie at 45°, but their visible plies tend to
the vertical on the S-course and to the
horizontal on the Z-course. The net result is
the open texture seen in Fig. 53b. This is
quite different from that produced by 4-ply
cords, see surface of bottle in Plate 16 and
also triangles at top of Plate 18.
D.DQQY An interesting development is to make a
braid with half S- and half Z-plied cords.
Down its centre will run a succession of
diamonds where the two types of cord cross,
one having vertical plies on both courses, as
in Fig. 53c, the next horizontal plies, as in
Fig. 53d. In these diamonds the cords bed in
closely, contrasting with side triangles
showing the open texture noted above.
Moreover, the plies in the diamonds run
together to give subtle horizontal and vertical
Figure 53 Using 2-ply cords plied in S and Z ridges. If the S-ply cords are dark and the Z-
direction:
a how plies appear on the two courses of a braid;
ply light, as suggested in Figs 53c—d, the
b—d how they appear when they cross. ridges immediately stand out as dark and
PLAIN OBLIQUE TWINING «+ 69

light stripes which reverse direction on the split eccentrically by choice. So with a 4-ply,
back in an initially baffling way. See Plate 18. the splitting cord will pass between 3 and 1
So these striped diamonds replace the plies, making the two faces even more
chequered diamonds produced with 4-ply dissimilar.
cords plied in the same direction. These textural differences suggest designs
It is unlikely that both a S- and Z-plied with contrasting areas of close and spaced
version of the same yarn will be found from plies on the same side of the braid. In Plate
which Z and S cords could be made. But if a 19, left, rectangles of 1/3 and the normal 2/2
doubled length of either a Z or S 2-ply yarn, splittings alternate.
knotted to itself, is taken as the starting
element, this loop can be _ over-twisted
9. PLAIN TRIAXIAL TWINING
successfully in either direction. To make the
stripes show well, do not ply the cords The useful basket structure of triaxial
excessively tight. interlacing has its ply-split counterpart; see
‘Two-ply cords each made with a dark and a Plate, 19, right.
light ply can easily give horizontal or vertical Start by taking 6 cords, split each centrally
stripes all across a braid. If in the first row’s and pull another through to make the
splitting the dark plies are uppermost, hexagonal ring seen left of centre in Fig. 54.
horizontal stripes will appear; if in the first Now ply-split at each of the crossing points
row, a dark and a light ply are alternately marked X, in such a way that the alternating
uppermost, vertical stripes will appear. splitting/split sequence is maintained in every
cord as shown. As this is by nature an open
8.3b Increasing the amount of twist left ina structure, probably more than a quarter-twist
cord between splittings will be left between successive splittings.
The rule in all the descriptions so far has been The three axes, vertical, S- and Z-oblique,
for the minimum amount of twist — a quarter- are now established. Continue by adding new
twist — to be left between successive splittings cords parallel to and outside each of the six
of each 4-ply cord, giving the strong closely- fells in turn. The dotted cords, 1 and 2 in
made fabric typical of its traditional use. Any diagram, show the start of this process. Using
departure from this rule markedly alters the cords of different colours or with 2 dark/2
fabric’s character. The more twists left light plies will lead to a new range of designs.
between successive splittings of cords, the
more the latter can slide about and the bigger
become the diamond-shaped openings they
frame. So the structure becomes netlike and
flexible but still retains enough stability to
have practical uses. See Plate 115.
8.3c Eccentric splitting

Cords with an odd number of plies must be
split eccentrically. So with the commonly
available 3-ply cord, the splitting cord will
pass between 2 and 1 plies giving the braid
two dissimilar faces, one showing close, the
other spaced-out plies.
Cords with an even number of plies can be Figure 54 Plain triaxial twining.

CHAPTER SEVEN
Single-Course Oblique Twining
In single-course oblique twining (SCOT), as each

cord moves on its oblique course, it is split and
therefore is visible, or splits and therefore is
concealed, at more than one successive crossing
point with cords on the opposite course. ‘The
structure is best understood by working a small
sample in the simplest form of the technique.
1. WORKING A SINGLE-SECTION BRAID

In strong contrast with POT the cords on the two
courses in a SCOT braid are not inclined at 45° to
the horizontal. The visible cords le more steeply
(between 50° and 65°) than the concealed (35° to
45°); see the detailed explanation later. As a result
there 1s no convenient transverse line in the structure
which can be used as a starting point. The following
method, though imperfect, is the better of the two
options available and will be used throughout this
chapter.
Attach three 4-ply cords to the starting rod in any
of the ways described earlier, giving six hanging
ends, numbered as in Fig. 55a.
Starting at the right side, split with the needle all
the cords insequenceyie: 6,5, 4, 3, 2; close under
the starting rod, so that two plies of each lie over the
needle, two under. Stop short of the left-hand
outermost cord, 1, and put it into the needle’s eye.
Now withdraw the needle, pulling this cord, 1,
through the centres of the other five cords to the
right. Here it is swung down to become the new
right-hand outermost cord.
‘The splitting cord now lies under the starting rod
and almost parallel to it, completely hidden by the
plies of the five cords it has split. This movement,
Figure 55 Stages in working a braid with
one section of SCOT: a start; b first row;
shown in Fig. 55b, can be written as 1 > 2, 3, 4, 5, 6.
for c and d continued over
SINGLE-COURSE OBLIQUE TWINING © 71

Note that a formula of this type always shows the
sequence of cords as they are encountered by the
splitting cord, not as they are initially split by the
needle.
Repeat this manoeuvre exactly for the second row,
shown in Fig. 55c, which can be written as 2 > 3, 4, 5,
6, 1. But it is a basic characteristic of this structure
that some twist is left in a cord between successive
splittings. In the case of a 4-ply cord, as here, it is
usually a quarter-twist which is left. This now applies
to cords 3 — 6, which were split in the first row, but not
to cord 1, just pulled through. So with the needle split
cord 1 as close as possible to its emerging point; then
cords 6, 5, 4, 3, leaving the vital quarter-twist in each
between this and the first row’s splitting, shown by
the small oblique lines in Fig. 55c. Put the left-hand
outermost cord, 2, in the needle’s eye and draw it as
before through the other five cords to the right, where
it becomes the new right-hand outermost cord.
Continuing in this way, always pulling the left-
hand cord through the other five, the passage of the
splitting cord will quickly become more and more
oblique until it, and therefore the braid’s fell, soon
acquires a stable unchanging angle, as in Fig. 55d.
It is especially important when dealing with this
technique to realise that diagrams such as Fig. 55
and the majority in this chapter are only analytical.
‘They merely show which cord splits which at their
point of crossing. The 45° grid employed is a pure
convention and bears little relation to the actual
angles at which visible and concealed cords lie in
reality. These angles vary considerably and so make Figure 55 (continued ) c second row;
it difficult to produce a grid which is both realistic d third row.
and applicable to all cases.

1.1. Practical details 1.2 Features of the working process
‘These can be summarised as follows:
1.1a Splitting cords centrally
* one cord is pulled through, i.e. splits, many.
Leaving a quarter-twist between * the fell is oblique and parallel to the last splitting
successive splittings of a cord is much cord.
easier with single-course than with * once started, every row is the same and involves
POT as nothing obscures the view of every cord. Compare with the long and short rows
the cord concerned. in eOr
Unply the cord with the left hand * there is no need in this single-section braid to
until the plies run almost straight turn the work over, as all the splittings are
' between the fingers and the fell, as in rightward.
Fig. 56. In order to leave the required
quarter-twist, the needle must pass
under the two nearer plies, one of
which lies over, one under, the last
splitting cord. So when the unplying
brings these two uppermost, shaded
for clarity in Fig. 56, pass the needle
under them and over the other two
plies as shown.
1.1b Producing a firm fabric

Production of a firm fabric depends
both on using cords of a sufficiently
tight ply and on a few practical points.
When the needle has split all the
required cords, hold them in the left
hand and push the needle hard up
against the fell, compressing the work
done so far.
After the splitting cord is pulled
through, give it a tug and keep hold of
ie ifaethe mignee Nand: sit. Asesthe
consistency of this tug, row by row,
which determines the straightness of
the braid’s edge. Then take each of the
split cords in turn and pull it away
from the fell with the left hand while
strongly twisting it in its ply direction.
This is to replace any of the Z-twist
lost by the passage of the needle and
the cord and is especially necessary if
the cords have less than the ideal Figure 56 _ Inserting needle to leave quarter-twist between
amount of twist. successive splittings of cords.
SINGLE-COURSE OBLIQUE TWINING « 73

1.3 Structural features 3 All cords are forced close together by the
‘The small braid made according to the above working method. The visible cords on the Z-
instructions is shown naturalistically in Fig. course will therefore lie compressed, side by
57a, and in greatly opened-out form in Fig. side, with no space between them. But each
58. In these the following structural features concealed, splitting, cord is separated from
can be seen. the next by the two plies which move
between the front and back; see Fig. 58. Thus
1 As each cord takes its zigzag selvage-to-
the S-course cords are spaced out, but the Z-
selvage course, characteristic of a braid, it is
course cords are not. The only way such a
split and therefore visible on the Z-course,
difference in spacing can be accommodated
but is splitting and therefore concealed on
in a braid is for the two courses to run at
the S-course.
different angles; as at b in Fig. 57 and not at
2 Because of the quarter-twist left between the expected angles at c. As shown at the
splittings, each ply of a visible cord passes over bottom of Fig. 57a, the Z-course visible
two splitting cords and then under the next cords lie more steeply than the S-course
two, and so on; see the shaded ply in Fig. 58. concealed cords, the courses crossing each
As aresult the surface, back and front, is made other approximately at right angles. This
up of oblique two-span floats. A visible cord is important feature is not peculiar to ply-
thus structurally related to the concealed splitting but is found with any braid in which
cords as is a warp cord to the weft in the 4- the structure allows elements to lie more
a
strand warp-twining seen in simple tablet- closely on one course than they do when
weaving. lying on the other.
Figure 57 a Realistic view of braid with one section of SCOT,

showing different angles of visible and concealed cords;

4 b and c angles of cords.
4 From Fig. 58 it is clear that the concealed not be so if this structure was made by
cords lie in their normal tightly-plied state, as twisting free-hanging elements, nor by
they pass from selvage to selvage. This would splitting plied cords.
na
Ss
——s
SEIS
i—NS
eZ
ey
——
MM
GUI),
/Kl CL
ry
ey
cy
Figure 58 Greatly expanded view of braid with one section of SCOT.
SINGLE-COURSE OBLIQUE TWINING + 75

5 When each piece is turned over, the visible 1.4 Distinguishing ply-split SCOT
cords, unfortunately for descriptive purposes, will from other types
lie on the S- not the Z-course, but naturally remain SCOT produced by ply-splitting has
Z-plied. The plies are therefore tilted differently on to be distinguished from that structure
the two sides which as a result look markedly produced by various other methods,
dissimilar; see Fig. 59. On the S-course side, such as the use of free-hanging
longitudinal ridges appear due to the surface plies elements worked by hand, bobbin-
of one cord coinciding with and visually running wound elements worked on the maru-
into the surface plies of adjacent cords, as in Fig. dai and kara-kumi-dai and looped
59a. On the Z-course side, the surface plies run elements held on the fingers and
into each other in a similar way to produce less hands. These non-ply-split methods
distinct oblique ridges, as in Fig. 59b. will be referred to as hand-produced
In neither case does a ridge correspond with the oblique twining because in all of them
course of a cord. Failure to realise this can make the it is the hands which insert the twining
analysis of a braid difficult. The use of these ridges twist during the actual interworking
with one- and two-colour cords is described later. process, whereas the essence of ply-
splitting is that the twist is pre-
fabricated, existing in the cords before
any work begins. It is important to
make this distinction because the two
types resemble each other closely and
have structural features such as
longitudinal sections and inversions in
common.
The following are four crucial
differences, the first being most
important.
1 Inhand-produced oblique twining,

the twining elements are _ twisted
around the concealed elements during
Figure 59 Plies of visible cords on S- and Z- courses the working process. So when, at the
forming longitudinal and oblique ridges. junction between sections or at the
selvage, they themselves become the
concealed elements, they lie as several,
roughly parallel but quite separate,
elements not as a tightly plied cord as is
the case with ply-split braiding. So
forcing apart the twining elements and
inspecting the normally concealed
elements should give an immediate
decision whether ply-splitting is the
method or not.
2 Inthe hand-produced method the

twining of a pair of elements can
change at will from being in the S- to

being in the Z-direction, a switch which often
happens between sections or at an inversion. This is
an absolute impossibility with the plied cords
prepared in advance for ply-splitting.
3 4-strand oblique twining is easy and common

with ply-splitting, though 2- and 3-strand are
equally possible; 2-strand is the only feasible type
with hand-produced oblique twining, 3- or 4-
strand being almost impossible to handle.
4 ‘The fell of the braid, if present, gives an instant

diagnosis. With ply-splitting there will be plied
cords beneath the last concealed element at the fell,
holding it firmly in place, whereas with hand-
produced oblique twining the elements used in
twining will hang separately beneath the last con-
cealed element which will tend to sag downwards.
1.5 Using two or more colours

Starting with half the cords dark, half light, a
pattern of alternating dark and light rhomboids
eventually appears, as in Fig. 60 and Plate 20, right.
This is an inevitable result of each cord being
visible on one course, concealed on the other. A
cord can be thought of as zigzagging methodically
through the braid, as the cord emphasised in the
diagram, with the structure only allowing the ‘zigs’
to be seen. With smaller groups of dark and light
cords, the rhomboids will be narrower, reaching the
extreme of a rhomboid only one cord wide.
Rhomboids of any width are therefore the basic
design units in all braids made by SCOT.
Note that due to the method of working the braid,
the SCS and the CCS are identical, both being 4
dark, 4 light in Fig. 60. Compare with POT. Figure 60 How cords of two colours
appear as visible rhomboids in a one-
section braid.
SINGLE-COURSE OBLIQUE TWINING - “ee

1.6 Cords changing function along an oblique line
An interesting effect is obtained if cords, while
maintaining their selvage-to-selvage course, change
from being visible to being concealed, and vice
versa, the change occurring along an oblique line
parallel to the S-fell used above.
At any point, turn the piece over so it looks as in
Fig 36a, chem work as Tollows! 2 ale 3515-25
AS 1525550 Sal; 23.5, 43 6 Sl, 2, 354,55. This builds
the triangle outlined in Fig. 61b which has a new S-
fell at its lower edge. Work can continue normally on
this fell or alternatively the piece can now be turned
back and another triangle made in exactly the same
way. A constant repetition of these triangles gives a
characteristic curvature to the braid, not hinted at in
the diagram. This is seen in a one-colour braid in
Plate 20, left. Using two or more colours, as in Plate
20, centre, emphasises the curving paths taken by
individual cords.
Figure 61 Cords changing function along oblique line.

2. WORKING IN SECTIONS
‘The method of working described so far, drawing
the cord at one selvage through all the other cords to
the opposite selvage, is obviously not convenient for
wider braids. Moreover if it were used, the braid
would tend to twist. So a system is employed of
working in sections, each of which involves only a
limited number of splittings. The sections can be
longitudinal, transverse or oblique in relation to the
long axis of the braid
2.1 Longitudinal sections

A braid with two longitudinal sections can be
worked on a V- or an A-fell
2.1a_ Braid with two longitudinal sections worked

on a V-fell
Double over four cords and fix them to a starting
rod to give 8 cords, numbered 1-8 in Fig. 62a.
Divide them into right- and left-hand groups of
equal size, 1.e. 4 cords each, as in the diagram.
1 Pull the left-hand outermost cord through the

other cords in that group (2, 3, 4), close under the
starting rod, and deposit it in the centre where it
now becomes the innermost cord of the right-hand
group. This rightward splitting, seen in Fig. 62b,
can be written as 1 > 2, 3, 4.
2 ‘The mirror-image movement is now required,

i.e. the right-hand outermost cord, 8, must pierce
the cords in its group, which are now 7, 6, 5 and 1,
as in Fig. 62c
To make this into the accustomed rightward
splitting on an S-fell, turn the braid over. The
required movement then becomes almost identical
to that in 1 above, the left-hand outermost cord
being pulled through to the centre.
3 ‘Turn the piece back and pull cord 2 through

cords 3, 4 and 8 into the centre. Remember to leave
a quarter-twist in cords 3 and 4 between this and
their splitting in the first row, as shown in Fig. 62d.
4 ‘Turn the piece over again and pull cord 7 through

6,5, 1 and 2 into the centre, leaving the vital quarter-
twist in cords 6, 5 and 1, as shown in Fig. 62e.
Figure 62 Stages in working a braid with
By now the V-shaped fell will be evident with four
two longitudinal sections on a V-fell.

cords emerging from its S-fell, four from its Z-fell;
see bottom of Plate 21, right. Work continues by
alternating the movements in the last two rows
described above, which are now considered in detail.
Working on an S-fell, the needle, starting at the

centre of the braid, travels upwards and to the left,
splitting all in the left-hand group except the
outermost. This is put in the needle’s eye and then
drawn down into the centre. So one cord splits three.
If the correct face of the braid is uppermost, the
first cord the needle splits is always the cord pulled
through in the last row. Knowing this immediately
tells the worker whether a braid is the right side up for
the next row. This cord is split close to its emerging
point, the other two leaving the necessary quarter-
twist. There are now three cords emerging from the
left-hand fell, five from the right-hand, as in Fig. 62b.
When the piece has been turned over the same
movement is repeated, this time pulling one cord
through four. There are now four cords once more
emerging from each fell, as in Fig. 62c. The cycle is
then repeated. Note how work is always done on an
S-fell.
Fig. 63 shows a braid made from 8 cords
according to the above description. It has two
obvious longitudinal sections, the left-hand
consisting of visible cords on the Z-course, the
right-hand of those on the S-course. So working on
two fells gives a two-sectioned braid.
One cord has been differentiated in this diagram
and ain Plate 21, right, to clarify the structure.
Following it down, it is first visible lying steeply on
the S-course. When it reaches the right selvage, it
turns through approximately 90° and changes
function from being split to splitting. So it is
concealed, lying more flatly on the Z-course as it
passes towards the centre. Here it reappears as a
visible split cord, still on the Z-course but at a
steeper angle, and runs to the left selvage. Again it
turns through a right angle and becomes a splitting
concealed cord running more flatly until it emerges
at the bottom of the braid in the centre. Every other
Figure 63 Braid with two sections,
cord is taking a similar path. The more expanded showing how one cord appears and
view in Fig. 64 shows what is happening to disappears on its selvage-to-selvage
individual cords within the braid. course.

NOTE The number of crossings in a cycle is
* it will be obvious that working in two always one less than the total number of
sections contradicts the name of the structure, elements making up a braid. So here there
single-course oblique twining, because there are 7 crossings, 3 in the first row, 4 in the
are twining cords on both courses. But the second, in a braid of 8 cords. This means that
name was chosen as it describes the structure one section, the left-hand in Figs. 63 and 64,
in its simplest form from which all others are is slightly narrower than the other and that
developed. the dividing line between the sections lies to
The changes of angle assumed by a cord, the left of centre. In order to make a
both at the selvage and at any junction structurally symmetrical braid it is necessary
between sections, is an inescapable charac- to start with an odd number of elements.
teristic of the structure. As explained above, it Starting with nine, 5 on the left, 4 on the
is the only way the sudden transition from right, will give two sections of equal width,
crammed (split) to spaced-out (splitting) each having four crossings.
cords can be accommodated.
aN
|SS
\\ = cf i. S« WA
\ \ Be = INN
\ Gi EXK ==") \
YA == \
A yor i We CONN
Vi)
Oy
ZION
Y PSA RN
OL Pome OY
Mp Ved ba oO \
> = \¥
WAiY aS =a);
= il)
ee a i‘Wisely NOG
See
EN WSO
Lg HD"iyae= iNw Ge ON
S/O WRYS_
NY SN
ey
key LT
HLL tTMPSS
PRN SSWN
WY & NS2P)
SSE) XG
alls, f/ iy 1, aN AY IS SN S SS > aR
ten
, ew Me NSA WAS
o, ky ve oS Figure 64 Greatly expanded
a eg a. = view of braid with two sections.
oF a
SINGLE-COURSE OBLIQUE TWINING ° 81
Design using cords of two or more colours
It was seen in Fig. 60 and Plate 20, right, that
with a braid of one section, a pattern of
repeating rhomboids was produced. In the
same way, each of the two sections made on
a V-fell will show rhomboids, but they will be
angled in opposition to each other, following
the line of the visible cords. The final pattern
will depend on how the rhomboids in the
sections relate to each other. Some examples
show the possibilities.
With a warp of 8 cords dark, 8 cords light,
a dark rhomboid will meet a light at the mid
line, as in Fig. 65a. With a symmetrical warp
of 4 cords light, 8 cords dark, 4 cords light,
the rhomboids will be half the size and match
centrally to give a transverse zigzag or
N> chevron, pointing towards the starting rod,
as in Fig. 65b, a very popular motf in
traditional ply-splitting. Any sort of colour
arrangement however haphazard will always
give a repeating pattern.
Design using cords of only one colour

Even if the braid is made with cords of one
colour, the two sections are differentiated
visually due to the angling of the cords and
their plies. The right-hand section in Fig. 63
will show longitudinal ridges, see Fig. 59a;
the left-hand, the less distinct oblique ridges,
see Fig. 59b. The prominence of these ridges
is increased the more tightly the braid is
worked; a point can be reached when they so
dominate the surface that the eye cannot
pick out an individual cord. See Plate 21,
centre.
It will be found that the section where the
course angle is opposite to the cord’s plying
angle has the longitudinal ridges. So with
normal Z-plied cords, this will apply to
sections with visible cords on the S-course,
i.e. the right-hand section in Fig. 63. This rule
immediately suggests using both S- and Z-
plied cords in a one-colour braid and so
having rhomboids, not just sections, with the
Figure 65 Two ways of arranging colours at start
of a two-section braid and the resulting design. two types of ridges.

Design using two-coloured cords left, shows such a braid with cords arranges
The above textural’ ridges can be 4S-, 4Z-, 4S-, 4Z-, 4S-, 4Z-plied at the start
dramatically highlighted by converting them and worked on a V-fell. The complex pattern
into coloured ridges, as in Fig. 66a and b. which results completely masks the cords
‘This requires cords made with dark, light, simple oblique placing in two sections. In
dark, light plies, as seen at the bottom of Fig. Plate 23, right, a third colour has been added.
66b. They are always split so that the needle These linear patterns can be altered in
goes under a dark and light ply from one several ways.
cord, then under a light and dark from the 1 Using cords with 2 dark/2 light plies. This
next cord, and so on alternately, as at the doubles the scale because two adjacent
bottom of Fig. 66b. Plate 21, left, shows the ridges will be dark, the next two light. Plate
result; longitudinal and oblique colour ridges 22, right, shows a 3-section braid worked in
of astonishingly mathematical precision. this way.
Those interested in the bewildering
geometry underlying this effect will discover 2 Using cords with 3 dark/1 light plied
that in both sections the angles between the cords, which gives ridges of unequal width.
colour ridges and the course of the visible 3 Using cords with plies of four different
cords is identical; see angles marked in Fig. colours, combined with S- and Z-plying.
66a-b. As with the monochrome version, the Plate 23, left, shows the sort of complex result
back of a section with longitudinal ridges this can yield even with a 2-section braid.
shows oblique ridges and vice versa. Note that the braids in Plates 22 and 23 are
Using such cords in S- and Z-plied form one-sided, the back lacking a coherent pattern.
greatly increases the possibilities. Plate 22,
Figure 66 Using two-coloured cords to emphasise

the longitudinal and oblique ridges existing on the
two courses.

2.1b Braid with two longitudinal sections NOTE
worked on an A-fell <> how each cord pulled through is swung
As usual work is done on S-fells, which here down to become the first cord split by the
implies that splitting cords move from the needle in the next row on that side.
centre to the selvage. + again the splittings (4 in first row, 3 in
Start as before with 8 cords divided into second) add up to one less than the number
two equal groups, as in Fig. 67a. Pull the of cords involved.
innermost cord (4) of the left-hand group
through the four cords of the right-hand Design
group (5-8). Then swing it down so that it As the visible cords always lie at right angles
becomes the outermost cord of that group. to the fell, a braid worked in two colours on
This can be abbreviated as 4 > 5, 6, 7, 8. an A-fell could look like that in Fig. 68, i.e.
Now perform the mirror-image movement, the inversion of that in Fig. 65b. Hence all
i.e. 5 > 3, 2, 1, first turning the braid over so designs for V-fells will appear upside-down; a
that this is also done on an S-fell. These two simple zigzag for instance will point towards
movements are seen in Fig. 67b. Repeating the fell, not the starting rod, as the diagram
these soon gives the A-fell seen in Fig. 67c. shows.
and at the bottom of Plate 24.
Figure 67 above Stages in working a braid with

two longitudinal sections on an A-fell.
Figure 68 right How two colours can appear when

working on an A-fell.

2.1¢ Moving from one type of fell to the
other in a two-section braid
Changing from a V- to an A-fell

‘To accomplish this, a triangle has to be built
up on either side of the mid line. One of
several possible methods is as follows. Start
at the left-hand side and without turning the
piece over between movements, work thus:
DS 255339 Ase ae ee ise gives tne
triangle outlined with dotted lines in Fig. 69a
and forms one side of the desired A-fell. Now
turn the piece over and do exactly the same
movements with cords 5—8, making the other
triangle which completes the A-fell. Working
on this fell in the normal way will produce a
sudden change in function of all cords, as
shown in Fig. 69b.
Changing from an A- to a V-fell

Here a diamond must be built up centrally to
produce the change and again one of several
possible methods is described.
Work as follows, turning the piece over
between, each=movement: 45, 6.07. 3;
Sey 2e8 ls SO > Ones Os, OZ lee ens
7>1, 2; 1>8. This produces the diamond
outlined in Fig. 69c. Working on this new V-
fell gives an immediate change in function in
all cords, as shown in Fig. 69d.
In the diagram for the two above methods,
the oblique line along which the cords
changed function from split to splitting is
emphasised. Here there is a visible change in
the otherwise level surface. The two cords
parallel to this line stand proud of the cords
at right angles to it; they have been shaded in
the diagrams.
By constantly changing the fell, diamond
shapes can be made to run down the centre
of the braid. Alternating the above
instructions is one way of achieving this. It
will show a textural effect if cords of one
Figure 69 a—b Changing from V- to A-fell; c—d colour are used, as in Plate 24; cords of two
changing from A- to V-fell. colours naturally increase the possibilities.
SINGLE-COURSE OBLIQUE TWINING «+ 85

2.1d Braid with many longitudinal sections follows:24 = 23322..21-and 165,17 > 45,14,
The principles described above can be £3 andiS:.9 =). 0.6 sands baoee eles 7 ay
applied to wider braids consisting of any There are now 4 cords emerging from each
number, odd or even, of longitudinal sections. fell as at the start. Turn the piece back and
Fig. 70 shows a 24-cord braid with 6 sections, repeat the cycle
worked on 3 V-fells. At the very start no fells
NOTE
or sections exist; they have to be created by
+ that it is only the very first movement,
the manner in which cords are grouped and
pulling the extreme left-hand cord through
worked in the first cycle of movements. This
the three others on that fell which does not
cycle is shown in Fig. 70a, where the arrows of
consist of four splittings. The left-hand
splitting cords are numbered to indicate the
section will therefore be narrower than the
sequence in which they move. After a few
others, as séen-in-Figs7 I.
cycles, 3 V-fells appear and soon acquire their
+ the above shows the total number of
final and _ stable obliquity, shown
crossings is 23, one less than the number of
diagrammatically in Fig. 70b.
the cords.
Four cords emerge from each fell and the
* how a particular cord, like the shaded one
cords have been numbered 1 to 24. The cycle
in Fig. 71, is visible in one section, concealed
in detail is as follows, always performed on S-
in the next, reappears in the next and so on, as
fells.
it lies on its selvage-to-selvage course. This
Starting at the left, .{>253, 4; S$ > 9; 10;
has a profound effect on the pattern of a braid
CV 26 eo 155.1 9.202 giving the result
made with two or more colours.
seen in Fig. 70c. There are now 3 cords
* that, working as described on the S-fells,
emerging from the leftmost fell, 5 from the
from left to right, on both sides of the braid,
rightmost, all the others having 4.
the last cord pulled through on one side, e.g.
‘Turn the piece over and work on the newly
cord 16 in Fig. 70c, is always the first one the
presented S-fells, starting again at the left, as
needle splits when the braid is turned over.
UUYyuyWGy
b Y
NZ
Figure 71 How acord is alternately visible and

concealed as it moves through a 6-section braid. Figure 72 Sections of unequal width.

Width of sections
As explained above the sections are not all
equally wide, even if starting with the same
number of cords on each fell. But a braid can
be worked with fells of very different lengths,
either arranged symmetrically, as in Fig. 72a,
or asymmetrically, as in Fig. 72b, giving
sections of varying width as shown. In either
case, the S-fells are worked on either side of
the braid as described above.
The narrowest possible sections are the
result of working with only 2 cords emerging
from each fell. To have ‘sections’ any
narrower reduces the structure to POT. With
very narrow sections, the cords do run at
approximately 45°. Changing from split to
splitting so frequently, they have no time to
acquire a new, and theoretically correct,
obliquity.
Methods of working
So far only one method has been described,
that of always working on an S-fell and
turning the piece over repeatedly to make
this possible. But of course the braid can be
kept in one position, the S-fell worked as
above then, the hands changing roles, the Z-
fell worked.
A third method employs the long needle
with a central eye, encountered in Colombia.
Use this as a normal needle for the S-fell so it
moves up, splitting cords 5 to 2, collects cord
1 and comes down again, as in Fig. 73a. For
Figure 73 Using a needle with central hole for the Z-fell, do not turn the braid over, but
working the S- and Z-fells without turning the braid
over.
point the needle, still in the right hand,
downwards to the left and split cords 9 to 6,
as in Fig. 73b. Either now, or before the
splitting, insert cord 10 in the needle’s eye.
Then push the needle in the same
downwards direction, drawing this
outermost cord through the other cords. So
here the needle’s movement is all in one
direction.

si es in many-sectioned e results whe ids, all started with a
SCOT are numerous. They depend on the warp of 6 light, 12 dark, 6 light are worked
interaction between the CCS in the cords and with 2, 4, 6 an and 12 section
sections respectively. Fig.
the size and number of the sections in the 74a was worked on an V-fell, b on an M-fell, c
s which are infinitely variable. and d on zigzag fells. A few basic possibilities
Figs 74a—d highlight th ion, showing re now described in detail.
Design with warp divided into half dark, Fig. 75b shows the effect when there are 8
half light cords sections of equal width, numbered 1 to 8. It
Imagine a set of cords, the right-hand half will be seen that cords on a particular course
dark, the left-hand half light, as at the top of only appear in every other section, that is in
Fig. 75a. Worked as a braid, the cords will the sections where they are split. In the
run as shown, taking oblique selvage-to- diagram, cords on the Z-course consistently
Selvage “courses. “But ‘the “presence” “of appear only in the odd-numbered sections.
longitudinal sections will severely limit the As already emphasised the rhomboid is the
visibility of the simple underlying grid in a basic unit of design in this structure; and it
way both precise and pleasing. will be seen in Fig. 75b that when either the
Figure 75 8-section
braid in two colours:
a general plan of
course taken by the
cords; b how their
appearance is
influenced by the
sections.

dark or light cords appear in a section it is resolved visually into the way rhomboids in
always as an elongated rhomboid. Almost all one section match or mismatch those in
the illustrations in this chapter show this. adjacent sections.
Down both sides of the braid in Fig. 75b, Once again notice that in Fig. 74 and
where cords of like colour cross, there are the subsequent diagrams the cords are drawn
expected dark and light triangles, unaffected for convenience on a consistent 45° grid. It
except in texture by the presence of the cannot be emphasised enough that in reality
sections. But the central diamond-shaped they always and inevitably lie more steeply
areas, where dark cords on one course cross when visible than when concealed.
light cords on the other, are forced by the
sections to appear as strong longitudinal Design with warp of half dark, half light
stripes. See Plate 25, right and centre, for two cords, arranged symmetrically
examples with 6 and 12 sections. Placing a mirror down the side of Fig. 75b
With more sections in the width, these will show the appearance of a braid with
stripes will naturally be narrower and more cords arranged symmetrically. The braid in
numerous. With sections of unequal width, Plate 25, left, started in this way with 10
the stripes will vary accordingly. So the light, 20 dark, 10 light cords and was worked
sections always have a dominating influence in 14 sections. Here there are both side
on the appearance of the underlying grid of triangles and central diamonds of solid
cords. colour and between these lie striped
Looking at the sections individually, notice diamonds. Compare Plate 25, left, with
how every one shows the same repeating Plate 4, right. Notice how the more sections
sequence of dark and light rhomboids. They there are, the more the underlying grid of
are of equal size but sited differently down cords is revealed
the length of each section. The design can be
SINGLE-COURSE OBLIQUE TWINING « 91

ER
OSSER
BO
Ber
SUS eh
oe
Shee
=
ose
a5
eS
ais
=
ye
a
tH
—
ee
s
lay)
o =2= Sn
§8
SBxYELES
oun
Ona
osie
Fe6
se,
i
peveeere te
q.g°
aD
%4%&
segkguws
me! RON
we
se
GP
Se
r
oe
Se
8bt5
eee ee
eon it EO
ET SeA eee
Sth pS S
=)
PP
tel By
Boao
2)
al Be
co) D
Ty
Wh
areeee
oe
ee)
ae Py
a
4
Ey
fe)SY
Oe
ae
iets
Onn
UPR
vBo
¢‘
oe
ies
we
2
fe a
Ge
Oo (Sem
Boles
ESEeA
SEsesesze
O
a
IOS
0
ORE
2
cD)
DU yw
OO
on
1S) eae 7
ane
ee
eae
A
Sal
a
Oe
tae
gpg
STE
BST
SH
an,
=
ess oe
i
ee
eS ie
ae
ee Ew YD
eo)
GH
.:
60
Be a
ofco)2 E ow
aad
-
Mona
ov
oe
&
5
lot)
NE DMer OsaA
ROBO SSO BES
fe) Sag
Ceres qa
yyao
9=
1S) S [=
°
Byes
a SS Fr Sis
ess
Sop
ESS
S
4 Oo
ar) acy
=
a ep
ayo
8,
AYFHg
Gand
HES
a 2) oO
om q& BaES ore)
ale) og
a oO
Me 00
o
Bae Ay
HERe ee we )
== i)
eOnk
Aer
B Oa fl
ete)
OY
is
0.4
IA
©BO
lo)ey 3 =
Q,
: ett
Oa
Ya 565
S8eeo
°8 RY)
sof
Q,
4S)
°=
>
All this does not happen by chance but is repeat of 8 cords. So the simplest case is for
governed by a simple rule which relates the the sections to have a width of 8 crossings (or
width of the sections to the size of the colour 7 or 9) as in Fig. 76a and 77b. But the section’s
repeat in the cords, the Course Colour width could be 2 x 8 = 16 crossings (or 15 or
Sequence. The rule is that the number of 17), or 3 x 8 = 24 crossings (or 23 or 25) and
crossings in a section must be a multiple of the so on, giving bigger and bigger zigzags.
number of cords in the colour repeat — or that The plus or minus one in the rule is easily
number plus or minus one. In the above explained. In Fig. 77b, there are three
example, there are 4 dark and 4 light cords vertical lines in the structure, marked by
_ running on both courses, giving a colour arrows, where every crossing is either dark
Figure 77 a Starting a 4-section braid to give transverse zigzags; b later stage.
SINGLE-COURSE OBLIQUE TWINING «+ 93

with dark or light with light. In any other
vertical line dark with light crossings will
be found. This implies that the colour
design would be unaffected if any of
these crossings were rightward instead
of leftward splittings or vice versa,
because the same colour would still be
visible at that point. But such a change in
the splittings would naturally increase or
decrease the width of a section by just
one crossing.
Fig. 76b shows the zigzag pattern is
unchanged if, following this rule, the
splittings are changed to give sections 7,
9, 7, 8 crossings wide. On the other hand,
Fig. 76c shows how the rhomboids mis-
match and the continuous zigzag is lost if
the width of the sections is varied by more
or less than one crossing.
NOTE
+ how in all these examples the
crossings still add up to 31, one less than
the total number of cords.
Design with variations

1 ‘This pattern can be varied endlessly,
depending on the number of sections in
the braid and the width of the zigzags. At
on extreme a CCS of 1 dark, 1 light will
give zigzags only one cord wide. The
colours will then meet whatever the
section’s width, the one exception to the
above, rule, see Fig. 7s. Plate 26, left,
shows a braid with 16 sections each only
two crossings wide.
2 Zigzags of different widths. A CCS

of 3 dark, 1 light, 2 dark, 2 light will give
thick and thin zigzags with sections 8
crossings wide.
3 Zigzags of more than two colours.

Plate 27, right, shows an example with
five colours arranged in an 8-cord CCS.
The sections are 8 crossings wide, so all
colours meet perfectly to give zigzags.

Plate 27, left, shows another example
with a 10-cord CCS and sections 10
crossings wide.
4 Using sections half the width

required by the rule. On a large scale
this gives the elegant design in Plate
28, right, where 4 sections, each 10
crossings wide, are worked with cords
having a 20-cord CCS of 10 dark, 10
light.
5 Using more than two colours which
only occasionally form zigzags. So far
rhomboids of the same colour have met
to form zigzags. But with a large colour
repeat, it can be arranged that, though
the basic pattern of meeting rhomboids
remains, like colour meets like only
occasionally, giving a less rigid design.
This is seen in the five-colour braid in
Plate 28, left, and the plan of its
coloured rhomboids in Fig. 79. Many
other possibilities exist, some of great
subtlety but all easy of execution once
the colours have been correctly
arranged at the start.
Figure 79 How the rhomboids lie in a 4-section braid

worked with cords of five colours.

CCI WMYY
o a0SBSSr See
a Bente
SSeS
= BS So Do2 Seer
26feSPOosFaeoy
ERTS
ON 7) 5
Oe
we:
Soe D osteo
ey ~
Se Gaon
SSR oagSseeoFR, eo420 Pe Seas ™
eSode
ica
oe So PKS OO we 2D yO =a
eS yok.
eebot eee
eco ys
‘a a Ge
Lp
°
acak
Le
Oo $a
ye) OK
SB) srt SEP GO)

Ss iB
AE
56 —_
ee
ie noa eee
SKS
eleGeo 28 . i=l
OOo
l
ge
ic
_
SS BOLa & ye:

9.x agp a,
G
ee
Bo
osee
Oo
IES D aie
ae Ky o
=
Cs! a OW
wo2
St De = ey Oo: 5s i VG
:
Skuaee& Bbee£24
l
OiRsS etFg Je 5 eg @ge
6 Po MV°lvrer
ay = glo
a 43Gee
a4Bad
298 3S50
seh Ee
OO : mEeho & QO,3
oeed a
~~
I Og—
a) se
EOS) ict eee = 8
Grae gS&sp
ab 0) Oeey
x#}xHoF
E
5
2 6 cgce.
22 G5
2 moo
ue © 2 age
o7E4S-6397 (are
oe
orooe Sees0
Oo oeICFP
OiseCARS
ry a
OvaA HO yvawio
ae)ao 4°2 aAGESA
Pelee
So) 6 a, ms2x 2
gee SO | ee6O45
OcessO.0M20 o) eloeeELSE
Te §ongj
Ss @)
5.8) bh
%
RH Sal 8 eo eo spcs fe
rs =
ee)
6 OS =
oea
Cee Fey c=
&
ee
oe
iu
FY
a oan,
R
a,
a
See VW 1S) yw = el Ow
=
—
6
= vod
SQ =
7S} 68g 8 we gr
ODE
eos
Senn) o aS
SE s 3
Oe
One
Dy2
Ss osSS
o oo sd q¢
FRAN
sv
9
oO ou ao a 9 40) ®
ee2aay
é
Sh
atat
en
Os=
es
:
Ji
ie a5 ®5
had SO
oc
ee
Lesooc
= Lo
So
a
ol SaSE
wswogo2e 29
oro Soo
ges
ers 53% 2 25
23
eo
o@ Se
soe
°
KS) £
56
8
Gne
2
= aD oO2 2Qa
Changing number of sections by inserting extra
areas
The top of Fig. 81a shows a braid, 8 sections
wide, worked on four V-fells. At the level
marked by the upper dotted line, the normal
working stopped and two diamonds (shaded)
were inserted as shown, using the method
shown in Figs 69c and 69d. They reduce the
sections to four and work continues on the two
V-fells so formed. Further on, work again ceases
and at the middle dotted line a diamond is
inserted centrally to produce the two sections
seen at the bottom, worked on one large V-fell.
The more detailed Fig. 81b shows the effect
such changes have on a design of narrow
Figure 82 Altering width zigzags. Plate 30 also shows this and how the
of sections.
whole process can be reversed back to the 8
sections.
Altering the width of sections

The longitudinal boundary between two
sections can be easily shifted to right or left. Fig.
82 shows an 8-cord braid which up to the level
of the dotted line was worked on a V-fell, giving
two sections 3 and 4 crossings wide.
After the last normal operation which drew
cord 1 into the centre, do not turn the piece over
but pull cord 2, then cord 3 through as shown.
Now turn the piece over and work normally to
produce a braid with two sections, 2 and 5
crossings wide. So the boundary between the
two sections has been shifted to the left.

2.1f Variations in structure in which some two braids, each two. sections wide.
cords do not run directly from selvage Repeating the above cycle will lengthen these
to selvage narrow braids and the slit between them; see
bracketed middle section of Fig. 83b. At any
Slits
moment the narrow braids can be reunited
Subdividing a braid into narrower parts by letting two central cords cross in the
Fig. 83a shows a 4-section braid being normal manner.
worked on two V-fells, with a slit just begun At the bottom of Fig. 83b, two more slits
centrally. On the left-hand S-fell, cord 1 is have been made, so now each of the original
brought in normally, but instead of cord 8 four sections has become a separate braid.
being the splitting cord for the right-hand S- Such narrow braids can be twisted or inter-
fell, it is cord 9 which performs this task. worked in some way before they reunite.
When the piece is turned over, cord 16 works Using the colour arrangement in Fig. 83
normally but cord 8 moves on the other fell, each of these braids is a solid colour. See
as shown. As no cord has crossed the mid bottom of Plate 23, left, and Plate 95, right,
line, there are now in effect the beginnings of for examples of slits.
Figure 83 a Making a central slit to

divide a braid into two; b adding two
more slits to divide braid into four.

Altering the course of certain cords adjacent central section, as at left of diagram,
By using slits, a design of two or more or in the opposite direction, as at right. The
colours can have a one-colour border. Fig. method chosen will naturally affect the type of
84 shows a braid with a dark and light zigzag fell to be used: a V-fell in the former case, a W-
pattern in the centre, edged with a dark fell in the latter. A wider border could be two
border. It will be seen that where the dark sections wide. There is obviously a practical
central cords meet this border they carry on limit to the length of such border slits, so they
to the selvage. But where light central cords are often combined with twined linkings.
meet it, there is a slit to deflect the light cords Slits give a very clean boundary between
inwards, keeping them out of the border. colours; see Plate 31, right, and compare with
‘The border cords can either be split at the Plate 31, left, where twined linkings were used.
Figure 84 Using slits to

make a border of one colour.

Twined linkings
Twined linkings can be _ placed
longitudinally and obliquely in a
SCOT braid but, unlike in POT, not
transversely, there being no transverse
line in the structure for them to lie on.
They are usually of the R/R type.
Longitudinally placed twined
linkings achieve the same effect as the
slits just described, their advantage
being they can be used repeatedly
without weakening the structure.
Their disadvantage is twofold in that
they do not give such a clean visual
boundary between the two colours
involved and that the added bulk of
their double splittings can lead to
stiffness and distortion of the braid.
Longitudinally placed to make a one-

colour border
As with slits, twined linkings can
deflect colours used in the centre of a
braid so they never interrupt a one-
colour border. So two twined linkings
could replace each slit in Fig. 84. Plate
31, left, shows how linkings of the R/R
type, if carefully made, can form a
decorative toothed junction between
dark and light cords.
Longitudinally placed to change the

CCS
This can best be illustrated by an
example. Working a two-section braid
q \ which starts with a CCS of 8 dark, 8
KROES
light gives rhomboids of the two
OOS
colours, as in Fig. 65a. But by twine
linking precisely in the four places
indicated by the circles in the upper
COCO SEE
4
part of Fig. 85, the CCS is converted to
4 dark, 4 light, giving the symmetrical
zigzag shown below and in centre of
Plate 32.The lower part of the diagram
Figure 85 Using longitudinally placed twined linkings to shows two further possibilities. Twine
alter the CCS of cords in a 2-section braid. linking where the circles are placed will

give the CCS of 2 dark, 2 light; seen below.
Twine linking where there are dotted
circles will give a CCS of 1 dark, 1 light, as
at bottom of Plate 32.
Obliquely placed twined linkings

Plate 33, left, shows a traditional use of
obliquely placed twined linkings giving
neat arrowhead motifs. Fig. 86 is the
corresponding diagram and makes clear
the design is the familiar zigzag but with
the colours reversed mid-zigzag, giving
dark arrow shapes on a light ground. It is
worked as follows on a braid with a SCS of
4 light, 8 dark, 4 light cords.
Start with the colours on the V-fell as at
the bottom ofFig. 86; then 1 > 2, 3, 4, 5, 6,
7,8 and 8 > 1. So light cord 1 twine links
with the last dark cord it meets, cord 8.
‘Turn the piece over and make an exactly
similar movement on the new S-fell. So a
cord again twine links with the fourth cord
of the opposite colour.
Turn it back and again the outermost
light cord 2 is drawn in and twine links
with the last dark cord it meets, which will
be cord 7. Turn it over and continue this
sequence, placing the twined linkings as
indicated by circles in the diagram. Plate
90, right, shows similar arrowheads.
Using more than two colours, as in Plate
33, centre, increases the possibilities.
Plate 33, right, shows another design in
which twined linkings move obliquely
right across a 1-section braid. Each cord
twine links twice as it is drawn across from
left to right selvage on the single S-fell.
NOTE
<+ Fig. 86 is drawn ona more realistic grid
in order that it may approximate more
closely to Plate 33, left.
<> the above two designs are the ply-split
analogues of those found in the obliquely
interlaced Assomption sashes. Examining Figure 86 Using obliquely placed twined linkings to
the latter may suggest other designs. give arrowhead design.
SINGLE-COURSE OBLIQUE TWINING +101

Inversions
An inversion is a mirror-imaging of the structure
about a transverse axis. Using one-colour cords,
inversions give subtle patterns of curving cords.
But these movements are more apparent and
therefore easier to understand with cords of two
or more colours, as in the following descriptions.
All cords involved in the inversions

The inversion is accomplished by moving
suddenly from a V- to an A-fell, or vice versa,
without any intermediate steps. The braid in
Fig. 87a has a SCS of 3 dark, 6 light, 3 dark,
giving the single zigzag seen at the top; it is
worked on a V-fell.
‘To make the inversion, the last cord pulled
through to the centre on the S-fell, shaded in the
diagram, continues on through all the other
cords to the right of the centre, until it reaches
the right selvage, see arrow. Then a similar
movement is made with the last cord pulled
through on the other fell so it reaches the left
selvage. After these two atypical splittings, work
proceeds normally on the A-fell so formed.
Naturally the manoeuvre leads to some
looseness where the split cords swing round
from one course to the other, especially at the
selvages. More than one quarter-twist can be left
here between splittings, as the diagram suggests,
but as the splitting cords run nearer the
horizontal than shown in the diagram this may
not be necessary.
Coloured cords will show up this inversion as
an elongated oval. Because of the moment
chosen to start the inversion in relation to the
placing of the colours on the fells, the oval
produced in Fig. 87a and Plate 34, right, will be
dark with a light centre. A braid with more
colours will give narrow concentric ovals as in
Plate 34, left. Plate 35, right, shows the
unexpected and pleasing result of such an
inversion using cords with dark, light, dark,
SS light plies. Such a single inversion if made on a
wide braid distorts it into a convex shape; see
Plate 88, centre.
Figure 87 a Producing single inversion in a
2-section braid; b reversing the process. ‘To reverse the process when working on an A-

fell, the last cord pulled from the centre to the Plate 34, right, shows the inversions in Figs
selvage is immediately pulled back through 87a and b following one after the other.
the cords it has just split and finishes back in In a similar way a wider braid of several
the centre; see the shaded cord in Fig. 87b. A sections can show an all-over pattern of
similar movement is made on the other side alternating ovals. Fig. 88 gives’ the
and the V-fell is produced. The resulting movements required for this.
central slit is usually avoided by ply-splitting
the two central cords before the inversion; Only some cords involved in the inversions
see small arrow in Fig. 87b, also in Fig. 88. Inverting only some of the cords and
Leaving two quarter-twists between the two allowing others to carry on normally has
consecutive splittings of the outermost cord, many possibilities. This can be a neat way of
as indicated in the diagram, stops any loose both increasing the number of sections and
plies at the selvage but gives a bulge here altering the CCS, as shown in Fig. 89.
even if the splitting cord is pulled extra Starting with cords which have a CCS of 4
tightly as it returns to the centre. The visual dark, 4 light, a 16-cord braid can be worked
effect of this inversion is of two half-ovals. on aV-fell giving the usual zigzag, seen at the
Figure 88 All-over design of small inversions.
SINGLE-COURSE OBLIQUE TWINING °103

top of the diagram. Stop at the point where 16 > 15, 14, 13 and 9, a normal move; 8 > 7,
the coloured cords are as in the diagram, 6, 5 and 1, to complete the oval.
numbered 1 to 16; it is only the central 8 There are now two V-fells, see dotted line,
which are to be inverted. Then make the and the CCS has become 2 dark, 2 light.
following moves, illustrated by the rightward Such a manoeuvre can now be repeated on a
pointing arrows: 1 > 2, 3, 4, a normal move; smaller scale. If the process is then reversed,
9>/10, 11, 12, to create the right- half of the heart shapes appear. Plate 35, left, shows this
oval.Turn the piece over, then following complete progression on a braid twice the
the leftward pointing arrows, work thus: width of that in Fig. 89.
Figure 89 Only involving

some of the cords in

Inversions with cords of many colours
The above shows only some of the
many possibilities offered by
inversions. They find their most
dramatic and astonishing expression
when worked with cords of many
colours arranged in some symmetrical
way. A much used Indian CCS for this
technique is 2 red, 1 green, 1 white, 2
blacks =i yellow, | green, ie. “Live
colours in an 8-cord sequence. This
sequence implies that any structural
procedure has eight possible colour
ways depending on the eight different
ways the colours can lie on the fells at
its start. See Plate 34, left, and Plate
89. The working of such seemingly
complex designs is in fact aided by the
many colours because at any moment
it will be all the cords of one particular
colour which are moved similarly.
There is a tendency for the crossing
of two cords at the very centre of an
oval to be concealed by the cords
curving around it. Working a twined
linking at this point will preserve this
spot of colour and also lock these two
cords into position.
The pleasing designs in Plate 36,
right, and 36, left, are based on
analyses of Peruvian fabrics originally
made by the oblique twining of free-
hanging ends (Speiser, 1982). Plates
37 and 88 show other possibilities.
Figure 90 Design with several inversions showing fells

used in the different stages of its production.

Analysis of designs with inversions 2.1g Other structural variations
The analysis of inversions in a finished braid
Selvage with unsplit cords
is initially confusing. The basic fact to
remember is that a visible cord lies at right The selvage can be strengthened and given a
angles to the concealed cords splitting it and rolled appearance by making cords curve
therefore to the fell used at that point. So the around it without being split. For example, in
curving of a cord from the S- to the Z-course Fig. 91, the next move would be 1 > 3, 4, 5,
implies a sudden change of the splitting cord i.e. cord 1 passes over (or it could be under)
from working on a Z- to working on a S-fell. cord 2, omitting to split it. As the diagram
It is the change of fells and the atypical shows, the result is an over one/under one
movement of the splitting cords causing interlacement of unsplit cords at the left
them which must be recognised. edge.
As an example, Fig. 90 shows in The right edge has been worked differently
diagrammatic form four ovals within a larger in that two cords are passed over (or under)
one. At the top, the zigzags show the work and not split, giving an over two/under two
was done on a V-fell, labelled 1, at the side. interlacement.
This continues, see bracket, to the centre of
the upper small oval. Here the central 6 cords
curve on to the opposite course implying a
change of fell, giving the W-fell, labelled 2.
Work on the four sections so formed
extends to the midpoint of the design where
every cord curves. This means all fells
change, giving an M-fell, 3. These four
sections carry on to the centre of the lower
small oval, where once again it is the central 6
cords which curve. So the fell alters to an A-
fell, 4, which gives the inverted zigzag
completing the design.
3
Figure 91 Making a selvage of unsplit cords.

Adding cords Between sections. A_ single cord,
Cords can be added merely to increase the introduced at the centre, immediately
width or to introduce new colours. Fig. 92 appears in both sections as in Fig. 92b.
shows some methods and assumes the added Alternatively it can be pulled out to the edge
cord is of a new colour. and then into the centre, as in Fig. 92c.
In midsection. A single cord, separately At the edge. As with POT, a doubled over
made so possessing a blind lop at one end, cord can be introduced at one or both edges,
can be threaded on the splitting cord as in Fig. 92d.
somewhere in its passage across a section; see If a tube is being worked, a cord or cords
left side of Fig. 92a. A doubled-over cord can can easily be brought in from the central
be introduced in the same way, being twice space or dropped back into it.
split as shown on the right of the diagram.
>
Figure 92 Adding cords, a in midsection, b between sections, c and d at edges.

Losing cords needle then takes its normal splitting route to
Cords can never be lost from a braid as the edge, drops off this cord, picks up the
unobtrusively as they can be added. Working outermost cord and draws it back into the
on a \V-fell, the outermost cord can at centre, as in Fig. 93b. So there are two cords
intervals be abandoned and so not drawn lying together in the same opening of each
into the centre; see Fig. 93a. Another way is split cord. It is safe to cut the protruding
to thread a central cord, pulled through in cords flush with the edge.
the last cycle, into the needle’s eye. The
Figure 93 Two ways of losing cords.

Longitudinal cords
An extra cord, or cords, can be worked
longitudinally, running down the braid at the
junction between two sections. See top of
Plate 61.
Working on a V-fell, as in Fig. 94a, the
needle first splits this longitudinal cord then
the others in the left-hand group before
picking up the outermost in the normal way.
Working on an A-fell, as in Fig. 94b, the
needle splits the cords in the right-hand
group normally then the longitudinal cord
before picking up the correct cord from the
opposite group. In both cases, turn the braid
over and repeat, using the same spilt in the
longitudinal cord.
Turn back and repeat the above two rows,
but leaving a quarter-twist between this and
the last split in the longitudinal cord.
Such a cord will show even if of the same
colour as the surrounding structure as it
stands proud of it, making a visible and
palpable ridge.
Figure 94 Using longitudinally twining cords, a on a V-fell, b on an A-fell.

2.2 Transverse sections with cords changing and 5 are dropped, being used no more in
function along a transverse line this section.
‘Though working in transverse sections is rare The subsequent splittings are all
traditionally, it is a technique with several (one > four), so the next is 7 > 6, 4, 2, 1; drop
possibilities, though not as many as possessed 7 and 1;9 > 8, 6, 4, 2; drop 9 and 2; and so on,
by the longitudinally sectioned type. as in Fig. 95b. When the left edge is reached,
16> 14,12,10 andi 12.446 1m. Bie 5c,
2.2a Method
One section has now been completed.
1 Start with 8 cords looped over a starting
rod, giving 16 cords — numbered from right NOTE
to left in Figs. 95a to 95c. Begin at the right * there is now a perfect transverse lower
with 3 > 2, 1 and 5 > 4, 2, 1, 3, which gives edge to this section, analogous to the
the triangle seen in Fig. 95a. Assuming the junction between longitudinal sections.
sections are to be four crossings wide, this * this is all done without turning the piece
width has now been established and cords 3 over, so makes for quick working.
11 10 10
Figure 95 Working transverse sections: a-—c stages in producing first section; d three completed sections.

Now tighten the structure by pulling all the 2.2b Design
splitting cords away from the split cords. The effect of transverse sections on design can
2 ‘Turn the piece over and proceed in easily be demonstrated by comparing Fig. 75b
exactly the same way from right to left. Two with Fig. 96. In both the cords (half dark, half
quarter-twists can be left in the first split light) take an identical course, but in the
cord as it was also split at the very end of the former the central diamonds consist of
first section; see arrows in Fig. 95d. This longitudinal, in the latter of transverse stripes.
makes the second section.
Repeating 1 and 2 gives transverse sections
four crossings wide; three are shown in Fig.
95d; see Plate 38, left (top), and 38, right.
A marked feature is that the selvage will not
be straight but will have an irregular outline
which keeps step with the sections as the
diagram shows and corresponds to the
irregular starting edge found with longi-
tudinal sections. This irregularity is probably
why the structure is little used traditionally.
‘There are two exceptions: in a cylindrical
form where each transverse section is a
continuous selvage-less ring; and combined
with a corrective longitudinal section at each
side, see later.
The structure is of course identical with
the longitudinal variety but turned through a
right angle. So just as the visible cords lie
more steeply than the concealed cords in
longitudinal sections, they lie more flatly
with transverse sections — lying in fact at
about 30° to the horizontal. This is seen in
Fig. 95d, where the grid used approaches
reality; but the subsequent diagrams are
drawn on the usual 45° grid. This flat angling
of the cords means the braid,will be wider
than a longitudinally sectioned one made
with the same number of cords.
Because the crossings are _ aligned
transversely, the structure could be worked
in transverse rows, always pulling one cord
through one and having alternate long and
short rows. So it would be similar to the
method for POT, except that here several
rows would be done before the piece was
turned over. However, it is more efficient to
pull one cord through many, as described
Figure 96 Effect of transverse sections on design
above. of a two-colour braid.

Figure 97 a How a transversely sectioned braid In a similar way the familiar zigzag pattern
can have a straight selvage; b—c method of will be turned through a right angle and so
production.
run down the length of the braid instead of
across it, as the top of Fig. 98 and Plate 39
show. The same rule applies; a CCS of four
cords, say 2 dark, 2 light, is needed if the
sections are four crossings wide. But to
centre the zigzags in the braid, the colours
are arranged in an asymmetrical way at the
start, as Fig. 98 shows.
2.2c Producing a straight selvage

The inherently irregular selvage is easily
overcome. By working atypically at the end
of each section, the cords at the left selvage
are all made to lie on the Z-course, those on
the right all on the S-course; so in effect
these cords make a longitudinal section at
both sides; see Fig. 97a. The method is as
follows:
1 At the left side of the work, make a small
thiangie thus: 4. 5* 26 3.550) eee ots
as shown in Fig. 97b.
2 ‘Turn the piece over so now this triangle
is at the right side and start pulling one cord
through four all across. So the first move is
7 > 6, 4, 2, 1, as the arrowed cord shows in
Fig. 97c. Continue in this way until only one
cord, 16, remains to be pulled through. Then
16)>:14.512. 10-8514 54125 1085) 2e= 10;
8; 10 > 8, making the triangle below the level
of the transverse fell seen in Fig. 97c.
3 ‘Turn the piece over and this triangle is at
the right as in 2 above. Ignoring the two
lowest cords, pull one through four all
across, ending exactly as described above.
So once the initial small triangle is made,
every section is worked in the same way,
ending with the fell shown in Fig. 97c. The
technique gives a pleasant textural pattern
when worked in one colour. The all-black
camel girths in this structure, having two
longitudinal sections each side, are master-
pieces of restrained design. Combining it
with longitudinal zigzags has the effect seen
in the bottom half of Plate 39.

PPE ESE
OS
ORROOKROROKS
KIKI
WOCCCCCCOS
SEOCCOCESC.
©
OXOKOKOKONK)
ne Statatata ates
a RN A
VE. OOO COL GGG

Ze CK XY Z, SEER XK MK MK vb O
re 98 /eft Using twined
ings to alter the cords’
Many of the variations listed above for longitudinal

sections can be applied to transverse sections. Three will
Figure 99 below Cords
changing from visible to
1 Using cords with dark, light, dark, light colour i a transversely
concealed in n
e. The fine lines already seen in Plate 21, left, sectioned braid.
verse lin
right, where each section is 10 crossings wide and the
non-straight edge contributes to the design.
2 ‘Twined linkings. If while making zigzags four twined
GS
NS
linkings are placed as in Fig. 98, almost all the cords
S
S
moving on the Z-course become light and those on the
e
-course dark, as shown. So the CCS is now similar to
that in Fig. 96 and if the work is continued that design
will begin to appear, starting at the level of the arrows in
that diagram.
a transverse s hange runs dow

the centre of the braid and is indicated by a heavy line. >>
2.3 Changing from transverse to longitudinal the number of sections produced depends
sections on the number of triangles built up.
Stop at the end of a transverse section, as in Fig. ‘To change back, triangles have again to
100a, and then build up the triangles shown be worked to re-establish a transverse fell.
below in Fig. 100b. Then work normally on the
NOTE
three S-fells, turn over and work on the new S-
+ how in Plate 38, left, the braid narrows
fells and so on, producing five longitudinal
slightly and the edge straightens where the
sections as at bottom of Plate 38, left. If zigzags
longitudinal sections begin.
are being worked they will suddenly change from
lying longitudinally to transversally. Obviously
Figure 100 Changing from transverse to longitudinal sections.

2.4 .-Cords changing function along oblique lines
Although in the majority of the above techniques
cords have changed function from splitting to split,
and vice versa, along longitudinal and transverse
axes, giving sections similarly aligned, there have
been’ instances of this change occurring along
oblique lines. The latter were in these cases merely a
means of achieving some other end; but they can
become an important or even main feature of a
design, as the following descriptions show.
2.4a Diamonds
The diamond in Fig. 10la is bounded by four
oblique lines. Consider the one emphasised in the
diagram. The visible cords on the Z-course, making
the actual diamond, are naturally packed closely side
by side; but where they cross the oblique boundary
and become concealed, they are forced apart by the > a
ix
plies of the cords they are splitting. This is plainly = : ba)
iL
seen in the model in Plate 40, left, where the
boundary has been opened out. These conflicting
RSV 2
Ny
Y
us
Ev
spacings can only be reconciled by letting the visible

cords fan out slightly as they approach the
boundary, which itself is forced into a curve; see the
similar boundary in Fig. 101b.
There is a similar condition at all four boundaries,
so what can be drawn as a simple diamond becomes
in reality the curved shape shown in exaggerated
form in Fig. 101b and seen in Plate 40, right.
NOTE
+ how both the cords of the diamond and the
background are implicated.
Nature of oblique boundary

Along an oblique boundary where a number of split
cords suddenly start splitting, there is a very obvious
groove. Fig. 101c is a section through A-A in Fig.
101a and the arrow points to this groove equally
visible on both sides. It has the effect of making the
cord adjacent and parallel to the boundary appear to
stand proud of the cords splitting it. The groove
emphasises the boundary, especially if all cords are of Figure 101 Cords changing function
along oblique lines making a diamond:
one colour. The longer the oblique line, the more the
a design diagram;
cords crossing it are diverted from their regular course b how cords are curved;
and so the more curved that boundary becomes. c cross-section through A-A in a above.
SINGLE-COURSE OBLIQUE TWINING 115

Dividing the diamond
The cords can change
function from splitting to
split within a diamond in
various ways. Fig. 102
shows how this effects the
curv-ature of the boundary.
Note how the background
cords must be angled to
ensure there is a change of
function at every boundary.
Figure 102 Curved edges of diamonds divided in various ways.

2.4b.. Changing from longitudinal sections to over and make these diamonds, each by
all-over diamonds pulling one cord through three, three times
Starting with the three S-fells of a braid of six over. Turn the piece back and the situation is
sections, as in Fig. 103a, make the following again as in Fig. 103a. Repeat these two
moves, without turning the piece over: 1 > 2, stages. This transition is sometimes seen
3; 2 >3 (to make the left-hand half-diamond); traditionally as the only form of decoration
697555935 > 758,954> 738;9 (to:make the on an all-black camel-girth. All-over
left-hand diamond); 12 > 13, 14, 15; 11 > 13, diamonds can be the basic structure of a
14, 15; 10> 13, 14, 15 (to make the right- SCOT braid, completely replacing
hand diamond); 17>18; 16>18, 17 (to longitudinal or transverse sections.
make the right-hand half-diamond).
2.4c Other examples of oblique boundaries
This gives the situation in Fig. 103b where
the newly made shapes are outlined with Fig. 104 shows the design diagram for the
dots. There are now three A-fells ready to white braid in Plate 41, right. It is essentially
receive three new diamonds. Turn the piece a braid with 16 cords made in two sections,
CK YSOKOWKOSKOOHK
ORK YS
OOK OOKOKROKOK
©
YY
3
RES KENSS ee
o
PLO
2 6%)
SX
8
Ke
3 4
¢.ecececansccn
Secon
Figure 103 Stages in changing from longitudinal sections to all-over diamonds.

each of which is divided obliquely on the Z diagonal (heavy
lines) leading to pleasant curves. The distortions produced
show well on a one-colour braid; using two or more colours
produces unexpected patterns, as Plate 41, left shows.
Fig. 105a and Plate 42, right, show another such design, the
heavy lines in the diagram emphasising the oblique
boundaries. It is part of a repeating pattern, called horolava (=
crab), found on a Colombian crupper (M. Cardale-Schrimpff,
1972). Repeated trials usually reveal an economical way of
producing such an atypical design; Fig. 105b is one such
solution.
Start with 8 dark, 16 light, 8 dark cords, and work in four
sections to reach the M-fell at the top of the diagram. Then
work the areas in the numbered sequence. For the central
areas 1, 4 and 7, the piece must be repeatedly turned over, this
not being necessary for the other areas. After completing the
nine areas the cords are back in their starting position.
Plate 42, left, shows a rather similar design derived from one
Figure 104 2-section braid found on girths in Rajasthan; see Plate 90, left. It has only two
with cords changing function longitudinal sections which are divided by transverse and long
along oblique lines. oblique boundaries.
°
°
°
2,
°
2 Z
¢ e
oni
te ~
te
e
Figure 105 a Diagram of design from Colombia with cords changing function along oblique lines;
b sequence of working.

3. DIMENSIONAL VARIATIONS
3.1. Changing the long axis of the braid

The long axis of a SCOT braid can be
angled in a similar way to that used for
POT? braids, but with one basic and
important difference. In the latter the
angle whether to right or left is always a
right angle, because cords on both
courses run at 45°. But here, because of
the very different angle taken by the
concealed and visible courses, the axis
will be changed either through an obtuse
angle of about 120° or an acute angle of
about 60°, depending whether the angle
is parallel to the splitting or split cords;
see Fig. 106.
Starting with a single-section braid of 6
cords, work on the S-fell in the usual way.
If the change-of-axis line is to be parallel
to the concealed splitting cords, stop at
any point, i.e. at the level of the dotted line
in Fig. 106a, where cord 6 has just split
cords 1 to 5. Cord 6 has now to split these
cords in reverse order, i.e. 5 to 1. Turn the
piece over to make this (and subsequent
ones) a rightward splitting on an S-fell.
The next move is for 5 to split cords 4, 3,
2; 1 and 6, and so on. The result is an
obtuse angle to the left in relation to the
braid’s original orientation. See first angle
in Plate 43, top.
To make the change-of-axis line
parallel to the split cords in-Fig. 106b,
first create a Z-fell by building the
outlined triangle, pulling cords through
one less cord in each row. Then turn the
piece over and work as follows: 5 > 6;
it ROO At 0s 2s Fs Osea es
3, 4, 5, 6. This builds another triangle
analogous to the one just made; work can
now continue normally, pulling one cord
yoy
ean
or
ae
me
tar
eee
ame
ake
ee
Ones
eee
2aSe
through five on the new S-fell. The result
is an acute angle to the right in relation to
the braid’s original orientation. See Figure 106 Changing long axis of the braid,
second angle in Plate 43, top. a to give obtuse angle, b to give acute angle.

NOTE
+ how neatly the direction of the cords on both
courses is not interrupted by either of these angles.
<+ that two quarter-twists can be left at the point of
each angle where the same cord is split twice in
succession, as indicated in the diagrams.
+ that with these changes of axis, it is helpful to
secure the braid with a clip, tied to a fixed point.
After each angle, reposition the clip so the part
being worked is correctly aligned.
Figure 107 Various ways of using the

two types of angles in Fig. 106 to make
shaped braids.
Figure 108 Right-angled

change to long axis of the braid,
with cords changing function at
¥ the angles.
Repeating one or other of the angles will give steep or flat

zigzag shapes, as in Figs 107a and 107b. Alternating two
acute with two obtuse will give the flat serpentine shape in
Fig. 107c and Plate 43 bottom. Many other shapes are
possible, none of which will have quite the mathematical
exactness of these diagrams.
Cords can change function from concealed to visible,
and vice versa, along the oblique change-of-axis line,
leading inevitably to distortion of the courses and a
visually less tidy angle. The result approaches a right-
angled change to the braid’s long axis; see the less
diagrammatic Fig. 108.To accomplish this, simply follow
the instructions for the acute angle when the cords are as
in Fig. 106a and for the obtuse angle when they are as in
Fig. 106b. See the last two angles in Plate 43, top.

3.2 Working cylindrically 3.2b Working as a cylinder from the start
Attach the cords to a stretched cord, not a
3.2a Starting with a flat braid
rod. Build up triangles to make V-fells and
A flat braid with an even number of then join them into a cylinder as just
longitudinal sections can be easily converted described, tying the starting cord to itself to
into a’cylinder. With the 8-section braid in make a ring.
Fig. 109, the next normal move would be
1 > 2,3, 4. But to start a cylinder the braid is 3.2c Design using longitudinal sections
curved round on itself so that the extreme There being no selvages, any colour
right-hand cord, 32, can be pulled through sequence in the cords on either course will
cords, 1, 25,3,.4. The other three S-fells are always stay on that course spiralling round
worked normally. To work the Z-fells it is best the cylinder. If the S-course has only dark
to turn the cylinder upside down and slant cords and the Z-course only light, the
the needle, still in the right hand, downwards longitudinal sections will be alternately dark
instead of upwards. Continue working the and light for the length of the cylinder.
four S-fells and Z-fells alternately.
NOTE
*+ how in every move one cord is pulled
through four. So every section is exactly four
crossings wide; and the total number of
crossings in a cycle equals the number of
cords involved. So it differs from the same
cords worked as a flat braid, where one
section of necessity would be narrower,
having only three crossings.
Figure 109 Flat braid about to become cylindrical.

3.2d Design using transverse sections
‘Transverse sections are well suited to a
cylindrical structure as there are no selvage
problems to overcome. The following way of
starting is a littke awkward but it gives a
perfect and firm upper rim to the cylinder.
First Section
1 Through a cord’s central point, arrowed
in Fig. 110a—c, pull another cord until its
central point is reached, as in Fig. 110a.
2 Pull a new cord as far as its central point
through the above two cords, as in Fig. 110b.
‘These manoeuvres can be done flat on a
table or each cord can be temporarily hung
over a stretched cord, as suggested in the
diagrams.
Figure 110 a-—c Stages in starting a cylindrical

braid with transverse sections; d finishing first
section; e finishing second section.

3 Continue introducing new cords in this section will eventually reach the starting
way until one cord has been pulled through point as in the centre of Fig. 110e and the
four, assuming the transverse sections are to triangular gap is filled in a similar way.
be four crossings wide, as in Fig. 110c. The third section is like the first, the fourth
4 Now as each new cord is added, pull it like the second, and so on.
through only four cords, as shown by
arrowed cords in extreme left-hand part of NOTE
Fig. 110 d. Gradually a narrow section is built * the working of all sections, after the first,
up. can begin anywhere on the circumference
5 When a sufficient number of cords has and then finish at the same spot.
been added, remove the stretched cord. Fold
the section round on itself to make a ring, the Design
growing fell, now on the right in Fig. 106d, If the cords on one course are all dark and on
meeting the start on the left. the other all light, then the transverse
Ge Now i> 95 7165/5083 12 > 65) 75 S353 S73. 83 sections will be alternately of these two
4 > 8; thus completing this first transverse colours; see top of Plate 44. There is an easy
section as one continuous ring of fabric. way to arrange this. As the first section is
made, introduce a light and dark cord
Second Section
alternately, giving a twined linking to, say, the
This can be worked in several ways; the
dark but not the light. The rim of the cylinder
hands can change roles and _ leftward
will then be as at the top of Fig. 111a with all
splittings be made or the cylinder can be
the dark cords ready to run on the Z-course,
turned upside down as already suggested.
all the light cords on the S-course.
However worked, the growing edge of this
Figure 111 Stages in starting a cylindrical braid

with spiralling sections of different colours.
SINGLE-COURSE OBLIQUE TWINING #123

3.2e Sections with oblique boundaries
Continuous spirals
Sections running spirally round a cylinder
can be made as follows: starting with dark
and light cords arranged as just described,
build up the triangles shown in Fig. 111a. In
one triangle the Z-course cords are split and
therefore visible, in the next the S-course
cords are split and therefore visible. Because
all the Z-course cords are dark, all the S-
KS
course light, these triangles are alternately
y
Y dark and light, as the diagram shows. Four
Va cords now emerge from every fell all round
SCO)KB
the cylinder. For this design to fit into the
circumference, there must be an even
number of triangles.
Now fit diamonds into each of the A-fells
by pulling one cord through four, four times
Figure 112 Changing the direction of
for each diamond; see arrowed cords in Fig.
spiralling sections of a cylindrical braid.
111b. To make the dark diamonds, pull light
cords through dark in the sequence shown by
arrows 1 to 4.’Io make the alternating light
diamonds, pull dark cords through light,
shown by arrowed cords 5 to 8.
Continue working, making more diamonds
for the new A-fells.
As the top of Fig. 112 shows, the spiral will
be on the S diagonal if the coloured
diamonds are arranged in the way indicated.
‘To reverse the spiral’s direction, put a dark
diamond where there should be a light; see
the outlined diamonds at the centre of the
diagram. Then carry on placing diamonds of
the correct colour for the Z diagonal spiral
seen at the bottom of the diagram.
In reality the spiral sections will not have
the equal width shown in Fig. 112, even
though they are all four crossings wide. The
cords on a course parallel to the spiral section
they are creating, i.e. the light cords in the top
half of Fig. 112, seen in more detail in Fig.
113, can lie closely side by side; there is
nothing to hold them apart. The method of
Figure 113 Showing the varying width of work will therefore force them together.
the spiralling section s of a cylind rical braid. making a dense narrow section. But the dark

cords on the opposite course lie at right 4. TREATING CORDS ABNORMALLY
angles to the section they are creating,
running straight across it; see centre of Fig. 4.1 Eccentric splitting of cords
113. They are spaced out where they emerge As in POT, cords can be split eccentrically:
from and enter the adjacent narrow sections. by choice if they have an even number of
The result is a broad section in which cords plies, of necessity if they have an odd
lie slightly apart. This difference in width of number. With the normally used 4-ply, this
sections becomes very obvious where the means a splitting cord passes between one
spiral’s direction is reversed and each section and three plies of the split cord. So on one
changes from wide to narrow, or vice versa. face of the braid each cord shows single plies,
See bottom of Plate 44. on the other face each cord always has three
The above spirals consist of alternating plies visible, each one passing over three
sections of S- and Z-course cords. But splitting cords, giving a rich appearance of
coloured stripes can also spiral around the thick cords lying side by side. See the section
cylinder if it is thought of as just a single, through a visible cord in Fig. 114.
endlessly wide, transverse section. So the One-colour cords show these two very
visible, split, cords of different colours will different textures to advantage, if they both
make continuous spiral stripes, parallel to appear as sections or areas on the same side
their course, while the splitting cords on the of a braid. For instance longitudinal sections
opposite course will never be seen. To can show the two types alternately; see top
accomplish this, make a transverse section of part of Plate 45, right. In the same way,
convenient width, as described above, transverse sections and obliquely set areas
followed by another and another, but always can use the two textures; see lower part of
with rightward splittings. Plate 45, right. Working such a braid make
sure that, on every S-fell on both sides of the
Obliquely set areas braid, the needle goes under one ply of a
A series of shapes like the diamond in Fig. cord. This is much easier to do accurately
101a works well on a cylinder. They show up than passing it under three plies.
subtly if all cords are the same colour, as in As described above, the cords on one
Plate 114, top, or more obviously if the cords course are consistently split 1/3, those on the
are dark on one course, light on the other, as other 3/1, so a section shows either one
in Plate 44 and Plate 114, bottom. For the texture or the other. But there is also the
latter, start with the cords as at the top of Fig. possibility of changing the type of splitting
111a, then make a line of triangles all of one within a section, 1.e. the needle as it moves up
colour followed by a line of diamonds of the the S-fell splits some cords 1/3, some 3/1. In
other colour to fit into the A-fells so formed, this way a large obliquely set area can be
followed by another line of diamonds of the halved longitudinally or transversely, or
original colour and so on.
Figure 114 Cross-section through an eccentrically

split cord.

quartered. See the diamonds in Plate 45, left. normally concealed, is made apparent. See
Commercial cord and rope is usually 3-ply Plate 88, centre. If cords with 2 dark, 2 light
so its use in ply-splitting depends on plies are used and the needle always passes
eccentric splitting, with designs of 1/2 under 2 dark in one row and under 2 light in
splitting contrasted with 2/1. the next, then this course is even more
obvious, showing as oblique dark and light
4.2 Leaving more than a quarter-twist
stripes. These combine into flat-angled
between successive splittings
zigzags if the braid has more than one section;
Leaving two quarter-twists between see top of Plate 46, right. At the centre of this
splittings gives a very different surface braid, an extra one-colour cord has been
texture from the normal. Two plies of a split introduced in one section to point out the
cord cover one splitting cord, then its other steep course of the covering cords.
two plies cover the adjacent splitting cord
and so on. This means that a 4-ply cord is in 4.3. Combining cords of different thickness
effect behaving as if it were a 2-ply. A cord much thicker than the rest will stand
There are two main results. The splitting out as a ridge in the sections where it is split,
cords are separated more than usual because showing equally on both faces of the braid.
four, instead of two, plies of the split cord So it is possible to make a one-colour braid in
separate each adjacent pair. This forces them which the design comes solely from the
to lie at an even flatter angle than usual. contrast between thick and thin cords,
Secondly, the absence of any plies passing leading to more subtle versions of the two-
over two adjacent splitting cords leads to colour designs described in this chapter.
grooves running between and parallel to these Plate 46, left, shows some inversions worked
splitting cords. So the course of the latter, using 4 cords much thicker than the rest.

CHAPTER EIGHT
Combining Plain with S- and Z-Course Oblique Twining
As often happens in textiles, combining two POT have reached their midpoints. Now
structures is a fruitful field for new designs. complete these three areas but as quite
In a typical POT braid with an equal number separate entities. This implies fewer splittings
of dark and light cords, there is a rigid in each row until the situation below the
repeating pattern, consisting of dark and arrows in Fig. 116a is reached. There are now
light diamonds where like colours cross and two A-fells with dark and light cords
chequered diamonds where unlike colours emerging symmetrically as shown. It is how
cross; see Fig. 115 and Plate 4, right, in these cords are worked in oblique twining,
Chapter Six. filling in the diamond areas outlined with
It is by replacing these chequered dia- dashes, which determines the final design.
monds, partly or completely, by areas of The principle of such infilling is now
oblique twining on S- and Z-courses that a described and then the innumerable
truly astonishing new range of designs variations are summarised.
becomes possible.
POO OPOO SOO)
1. BASIC METHOD
The unusual method of work results from
SSELREG S
the fact that POT is normally worked on a
transverse fell, but SCOT on an oblique fell.
At the level of the upper arrows in Fig. 115
and in the more detailed Fig. 1 16a, the solid-
colour diamond and triangles worked in
cxBook
=e<0, Sr 2 XQ
S
SS ee
>
SY S
S KEELER
BOOS Soo
Figure 115 Repeating pattern of a POT braid with
coloured cords arranged symmetrically.
LEEESSIS
CR
ASRRR SSSR.
ANAS
KANNAN
COMBINING PLAIN WITH S- AND Z-COURSE OBLIQUE TWINING * 127

Figure 116 Stages incom bining plain with S- and Z-course oblique twining.
2. ALL THE DIAMONDS WORKED IN flanked by dark triangles (where dark cords
OBLIQUE TWINING cross dark cords). Again there will be two A-
fells, but the cords emerging from their sides
2.1 Diamonds of solid colour will be of opposite colour to those in Fig.
At the stage reached in Fig. 116a, the right- 116a. Again diamonds are worked in oblique
hand S-fell is worked as follows: cord 6 > 7 twining to fill in these A-fells, exactly as
foals > 7 to. 1247-10 12= and*so on; described above, but this time the right-hand
making the light diamond seen below in Fig. will be dark, the left-hand light.
116b, consisting of six split and therefore Continuing in this way, a design of oblique
visible light cords. When similar movements stripes, alternately dark and light, will appear,
are done at the left-hand S-fell, a dark as shown in Fig. 117b. They appear because a
diamond is produced as shown. So a light dark diamond, worked in POT, and the
and dark diamond have replaced what would adjacent dark diamonds, worked in oblique
have been two chequered diamonds in POT. twining, though differing in structure, are
The next stage is to work the central read by the eye as a single dark design unit.
diamond and two side triangles, outlined with Similarly light diamonds worked in the two
dashes in Fig. 116b, in POT. So the transverse structures appear as a single light unit.
rows in this technique will gradually increase The dark and light diamonds worked in
in length, from a single splitting at the apex of POT are fixed unalterable elements in the
each area, until the latter’s midpoint is design. But the colour of the intermediate
reached (lower arrows in Fig. 116b). The diamonds, marked X in Fig. 117a, can be
areas will then contract, exactly as in Fig. either dark or lght in any _ desired
116a, but giving this time a central light arrangement, depending on how the oblique
diamond (where light cords cross light cords) twining is carried out. So referring back to
Figure 117 a Repeating pattern of POT braid; b—e designs resulting from combining it with
S- and Z-course oblique twined diamonds of solid colour.
COMBINING PLAIN WITH S- AND Z-COURSE OBLIQUE TWINING + 129

Fig. 116a, the right-hand diamond could be detailed Fig. 118. With an asymmetrical
worked thus, after turning the piece over; design like Fig. 117e, the whole braid is
122640:13 3S Oto 119
> 'G-to 1 sandss0.08; distorted; see its realisation in Plate 47, left.
producing a dark instead of a light diamond.
2.2 Diamonds not of solid colour
In a similar way, the left-hand diamond could
be light instead of dark. This means that each The possibilities multiply greatly if the
diamond can be of either colour, making obliquely twined diamonds are not of one
possible the designs in Fig. 117b to e. solid colour, but each has both dark and light
Such braids will not look mechanically cords visible. With an A-fell, as in Fig. 116a,
perfect as the diagrams suggest. This is giving 6 dark and 6 light cords, the diamond
because the diamonds of oblique twining will have 36 crossing points, each of which
elongate in the direction of the split cords, can show either a dark or a light cord
due to the latter cramming together. But the depending how the cords are split. The
POT diamonds and triangles more or less diamonds in Fig. 118 have 64 such crossing
retain their strict geometric shape. As a points. The worker can therefore think of the
result the selvages are seldom straight; see diamond as a blank space or grid subdivided,
Plate 47, right, which follows’ the as in Figs 119a and 119b, and however he fills
symmetrical Fig. 117d and the more it in to make his own dark/light design, that
design can be produced by oblique twining.
Some such designs are now considered.
Figure 118 /eft Design diagram for Fig. 117d.
Figure 119 below a The grid for a 12-cord and

b a 16-cord braid, to be filled in with S- and Z-
course oblique twining.
~
aS
<
LS
S
OS,
SOA

2.2a Dividing the diamond transversely
Referring to Fig. 120a, work the oblique
twining as follows: 6> 7-12; 5> 7-11;
4 > 7-10; and so on, giving the diamond a
light upper half, as shown above the arrows.
Turn thie piece over and by working 8 > 1;
9>2, 1; 10 > 3-1; and so on, produce the
dark lower half of the diamond.
NOTE
. *% that the diamond can never be divided
into two exact halves. The way described
gives a slightly larger upper part.
If the piece is turned over before the start of

the above manoeuvres, the diamond will
have a dark upper and a light lower half.
‘These two possible ways of arranging the
colours are utilised in the following designs.
Fig. 121a shows a braid with the dark and
light diamonds produced by POT, the fixed
elements in the design; the intervening
diamonds to be filled in with oblique twining
are marked X. If every such diamond has a
light upper half, as in Fig. 120a, the design
will read as the large triangles in Fig. 121b,
shown on a larger scale in Fig. 122. If the
diamonds in one row have a light upper part,
those in the next row a dark, the design in
Fig. 121c will appear. An alternation of the
two types of diamond in every row produces b
the designs in Fig. 121. In all these designs,
the centres of two rows of obliquely twined
diamonds are marked by arrows.
Plate 48, right, follows Fig. 121b exactly.
The top of Plate 48, left, follows Fig. 121c
and the bottom of Fig. 121d. Plate 49,
right and left, only uses the centre half of
Fig. 121c, between dark borders and with
two aberrant twining cords outlining the
POT diamonds. Plate 49, left, also has spots,
two cords wide, in each triangular area
Figure 120 Dividing the obliquely twined diamond

a transversely, b longitudinally and
c obliquely. Cc

2.2b Dividing the diamond longitudinally 8 > 4-1 and so on, ending with 1 > 12. As
If the obliquely twined diamonds are divided shown this gives a dark left half and a light
longitudinally down their centres, the designs right half. The colours would be reversed if
in Fig. 121b-e are produced, but turned the piece were turned over before the above
through a right angle. Such a division is movements. The top of Plate 61, shows the
worked thus, turning the piece over after each design in Fig. 121c turned through a right
move to make all the splittings rightward (see angle with the addition of extra longitudinal
Fig. 120b): 6>7-12; 7>5-1; 5> 8-12; twining cords
Figure 121 a Basic repeating pattern, b—e designs resulting

from dividing the obliquely twined diamonds transversely.
OSs
Zs
§
PRX
MEIER
x
AN
PIII RR <A
SK
NN
IN
Figure 122 Design diagram for Fig. 121b.

ZA
RX
;
2.2c Dividing the diamond obliquely

Following Fig. 120c, work thus: 6 > 7-12; Figure 123 above Diagram for design with
5 > 7-12; 4 > 7-12. Then turn the piece over obliquely twined diamonds divided obliquely.
and continue with 7 > 3-1; 8 > 3-1; 9 > 3-1
and so on. The colours are reversed if the piece
is turned over before the start. Fig. 123 shows Figure 124 below Subdividing the obliquely twined
a design using both of these possibilities. diamond into smaller diamonds.
So far these diamonds have been divided into

only two areas. Further subdivision greatly
increases the range of possible designs.
2.2d Subdividing into smaller diamonds

Fig. 124 shows a large diamond subdivided
into four smaller ones, each of which is
simple to work. Subdividing all the
appropriate diamonds in this way gives the
design in Fig. 125, which has the same motif
in both dark and light cords but the dark is at
right angles to the light, an Escher-like effect
seen in Plate 50, right. These four small
diamonds can be further subdivided
transversely as in the bottom of Plate 50, left,
or both transversely and longitudinally as in
top of that plate. Figs 130c and 130d are the
diagrams for these variations. In Plate 51,
centre, the diamond is subdivided into nine.

2.2e Subdividing into four triangles opposite colour. Fig. 126a and b show two
A motif much used on Indian girths is ways the centre can be worked, the former
shown in Plate 51, top and bottom. The being completely symmetrical. Fig. 127a
diamond is subdivided into four triangles and b are the corresponding analytic
each of which abuts a POT diamond of the diagrams. The four areas concerned can be
e"
WS
Ly
LL
CA
»
x
YS
Ky
ys
FTE
SOON N
SSNZ
SS
% \ NIZA
M)* O, S
NG
SORROW VSS KSKK
Figure 126 Designs resulting from two ways of subdividing the obliquely twined diamonds
into triangles.

produced by several working methods. For 2.2f A basic structural feature
instance, Fig. 127a can begin with 8 > 9; The boundaries between an obliquely twined
7 >9; 6>9 and so on, ending with 2 > 9; diamond and the surrounding POT
then. alter “turning sover, .10> S911 > 3: diamonds, although straight in a diagram, are
12 >8, ending with 16>8. These moves in reality forced into a curve. The effect has
are indicated by arrowed cords in Fig. 128. already been seen in a more pronounced way
Continuing thus, moving from the outer in the SCOT chapter. As there, it is the result
edge of the diamond to the centre, will make of the concealed, splitting, cords being spaced
the top half of the diamond. out (see the cross-section in Fig. 129a), while
Then work from the centre outwards, the visible, split, cords are crammed closely
turning after every move, thus: 12> 4; side by side (as in Fig. 129b). But here it is
A> 1S ande Wise le aloe > 1A: modified by the fact that all the cords in POT
4> 14; 3 > 14 and so on. This will make the are spaced equally and to a degree in between
lower half. Other methods can be discovered the two above cases; see Fig. 129c.
which involve pulling one cord through So whenever cords cross the boundary
many; they may be quicker but are not so easy between an obliquely twined diamond and a
to grasp. See Plate 51, top and bottom, and POT diamond, they will either become
Plate 52, right and left, for this motif worked slightly more or slightly less spaced out. This
in different scales. forces the boundary into a curve which is
always concave towards the area of more
Naturally any of the above treatments of the closely spaced cords. Applying this to the
obliquely twined diamond can be combined motif just described, the obliquely twined
in one braid, making for more possibilities. diamonds will have four concave sides and so
See lower part of Plate 17. the POT diamonds will have convex sides.
Figure 127 Analytic diagrams for diamonds in Fig. 126 divided into four triangles.
COMBINING PLAIN WITH S- AND Z-COURSE OBLIQUE TWINING »* 135

Figure 128 above First stage in dividing diamond into four
triangles.
Figure 129 right Spacing of a splitting cords,

b split cords in SCOT and c of both split and splitting cords in POT.
These slight curves rob this design and the others in this
chapter of their geometric rigidity in a way both subtle and
pleasing. Though the curves’ direction is predictable, their
degree is not and seems to depend somewhat on the
sequence of moves used in working the braid. For instance,
the four sides of the POT diamond in the present design
sometimes curve outwards slightly and sometimes to such a
degree that the area approaches a circle, as in Plate 52, left.
The ways of subdividing the obliquely twined diamond are

literally endless and with large diamonds can be really
complex. A few are shown in diagrammatic form in Fig.
130 and in reality in Plates 50, left, and 53 and 54. Four
diamonds treated as in Fig. 130g surrounding a dark POT
diamond make a convincing star; see Plate 55 where,
however, the star is light on a dark background. Some can
be seen as isolated figures, others join visually with
adjacent POT one-coloured diamonds to produce overall
designs, as already seen. No instructions are given because
the diagrams should furnish the worker with enough
information to go on. It is also important to realise that for
any design there are usually several possible working
sequences. In all of these designs the eye can trace the
alternating dark and light diamonds of POT. This feature,
which is an unalterable constant in the methods described
so far, is lost in the following designs.

g
Figure 130 Other ways of subdividing the obliquely twined diamond.
2.3 Using twined linkings in the obliquely they are of the R/L type, the same cords are
twined diamonds visible below as above the line, but running
The use of twined linkings in the obliquely on the opposite course, as in Fig. 131b. As
twined diamonds adds a whole new range of the first gives a two-colour, the second a one-
possibilities. This is because if all the cords in colour diamond, this is obviously a case
these diamonds are twine linked, there is a where the type of twine linking chosen is
significant change to the basic plan seen in crucial to the design.
Fig. 121a.This results in the vertical stacking Fig. 132b—e show some simple motifs, the
of POT diamonds of the same colour, as in twined linkings always being indicated by
Fig. 132a, instead of the two colours transverse dashes. In b and c, the twined
alternately. So the basic underlying colour linkings are R/R and the _ two-colour
plan is altered. diamonds are arranged in different ways.
Consider a transverse line of twine linkings The top of Plate 56, right, follows Fig. 132b
bisecting the diamond. If they are of the R/R and shows the distortion already referred to,
type, the cords of the opposite colour will the triangles on the diagram becoming
appear below the line, running on the same arrowheads. The bottom of Plate 56, right,
course as those above, as in Fig. 131a. But if follows Fig. 132c.

In Fig. 132d—e, the R/L twined linkings cords. In Fig. 132e large dark and light
give one-colour diamonds which §are diamonds alternate, the result of some one-
arranged in various ways, shown also in more colour diamonds being light, some dark. By
detail in top and bottom of Fig. 133. In Fig. making the one-colour diamonds smaller as
132d all the one-colour diamonds are dark so in Fig. 134, light and dark longitudinal
there is a series of small light diamonds, one stripes are obtained; see Plate 56, left, top
above the other on a dark ground, giving the and bottom. Starting as for the motif in Fig.
mistaken impression of more dark than light 126 and then making R/L twined linkings at
Figure 131 above Transverse line of a R/R and b R/L twined linkings across centre of obliquely twined diamond.
Figure 132 below a Basic repeating pattern; b—c designs resulting from R/R d-—e from R/L twined linkings.
WI
KAMA KA SS S
NSS
YA NSANAS %

TZ TSR
YiS$ the diamond’s mid line gives

the effect in Fig. 135 and in
Plate 57, right. Note how the
POT diamonds become
almost circular.
Leaving only a quarter-twist
Os
OeSOK between the two splittings
which constitute an R/L
twined linking gives some
Os WyQZ
RR y
SLR surface irregularity but
XD SS ensures that the splitting cord
SSSis well concealed at this point.
It is easy to introduce variety

EID into the above motifs by
making one or more cords in
COG
NR the obliquely twined diamond
VALLI AAD SOOO NAG behave differently from the
NSF AFFER NASA
rest. This has already been
seen in the outlining cords in
i
Plate 49. A further develop-
4
I?Vy
GY, ment makes such split cords
SC take a curving path. For
SALONS
OS Ws
QS
TOS
INXS
CR S
RRR
SKS
ZAIN
NX
DANIK
ZA Ns Figure 133 top /eft Design diagram
ZO ARN for patterns in Fig. 132d—e.
SON
Figure 134 middle left Making
longitudinal stripes with small
diamonds crossed with R/L twined
GAA.
linkings.
OD ide DS
»| Figure 135 /ower left Using R/L

OX
SN
PROXCX
twined linkings across centre of
DOOD YS diamonds, worked as in Fig. 127.

COS
yk i FOX.
LES
KOKO ee
“G KSA
GN

example in the top part of Fig. 136, a single Working a row of R/L twined linkings
dark cord (heavily outlined) from each when halfway through a motif, such as in Fig.
diamond is split instead of splitting. It is 130a, neatly inverts the design; see Fig. 137
therefore plainly visible and takes the path and Plate 58, left.
shown, curving around the light cords. Plate
58, right, shows this design with a colour It will be realised that the all-dark and the all-
reversal. In the bottom half of Fig. 136 both light diamonds in the above examples do not
dark and light cords behave in this way; again need to be in POT. Whatever structure is
heavily outlined. The very elegant design in used it will still give the same one-colour
_ Plate 57, left, follows this idea but uses such diamond and the visual dark/light pattern
cords in pairs. will remain unaltered. So the diamond could
be worked in SCOT or even in TLOI, the
former giving a quite different texture
depending how the sections are arranged.
Plate 59, left, shows light diamonds in the
centre alternately worked in the POT and
SCOT structure.
3. ONLY PART OF EACH DIAMOND

WORKED IN OBLIQUE TWINING
In the above examples, all of the diamond
which would have been chequered in POT
was worked in dark and light oblique
twining. If, however, the latter only occurs
in some parts of the diamond and the
remaining parts are in POT and so still show
the chequered effect, a three-tone design
becomes possible. In Plate 59, right, half of
each diamond is chequered, half obliquely
<A
twined; there are outlining cords of another
YY colour. The larger the diamond, the more
SKK scope for this three-tone effect; see Fig. 138
and Plate 60, right.
Figure 137 top /eft Inverting a design by working

R/L twined linkings across centre of diamond.
Figure 138 /ower left Three-tone effect resulting

from not completely covering the diamond with
obliquely twined cords.

ROSS +
oe erate)
axe :
gag
OFa
io Bg £2529ae)
TZ,
—_
8 ae 26
iO
Be
ee
as is a & yn
3G oa
ae Lo
ee) OO
mo
BB
ah
is
Bo
RK soe om
O 0d >=
oe
Qa's >o
tes ase ne:
= HH
weass re
sa s x
ease
ie)
sac OA, a00

Hee Seasq
SES SER oEM
ES
69 S| amo bh
=o
9 Su
Ss fee es
Belpre
oe e
ges See
are
= ee
= Sar GO°P
Poe P ooo
Se ee OO
es 2 ee fe
-n of Ye 5_ 3
ee bee OBO
=
-
o> oO &
a Ao
o°
& ad G26
Ww
S Ao2 oo
Ss y Pag
Be:hota!
ra)
Is
ee
SD
s/o 8 D5
83% as°u
Bes
8 oOo eto
ieee FE3
Ww A
4.2 Free-running cords 4.3 As a border
Cords can twine obliquely across a POT A POT braid can have an obliquely twined
ground independently of any pattern in the border. In Fig. 141, at the end of the row
latter. The simplest application is in a one- being worked, the outermost cord is pulled
colour braid where such cords only add a through three cords, as shown. In the next
texture of ridges, visible from both sides. But row, after turning the braid over, the outer-
they can be of a different colour and run at most cord is pulled through four cords, thus
will on the S- or Z-course, as in Fig. 140 and making a single obliquely twined section at
Plate 62, right. each side. Starting with an odd number of
_ The pairs of dark cords moving outwards cords would enable both sections to be
and inwards in Figs 27 and 28 in Chapter Six equally wide. Using twined linkings these
could be in oblique twining. They would then sections could be a different colour from
appear as solid lines, not a sequence of spots, those used in the central area.
thus strengthening the design.
Figure 140 Obliquely twining cords running freely Figure 141 Making an obliquely twined border to
across a POT braid. a POT braid.

5. A STRUCTURE WITH REGULAR needed at one side for symmetry. If their
LONGITUDINALLY TWINING CORDS course is changed repeatedly, the twining
ON A POT GROUND cords come to lie longitudinally in the braid
Fig. 142a shows a structure which could be as shown. The method of work is as follows.
considered as existing in its own right or as The total number of cords must be a
being simply a development of that in Fig. multiple of 3, plusl1. Knot the lower ends of
140. There is a twining cord for every two the cords which will twine to avoid con-
POT cords, making a structural unit of three fusion. Work the rows transversely, always
cords, bracketed at the top; an extra cord is pulling a POT cord through a twining cord
and another POT cord, as shown by the
arrowed cord in Fig. 142a. The odd number
of cords means there will not be the normal
alternation of long and short rows; instead a
cord will always be left over at the end of a
row and will be the first one to be split when
the braid is turned over and the next row
begun. A quarter-twist is left in all cords
between successive splittings.
If only one colour is used, the result is a
braid with an interesting and _ easily
recognised surface texture. When two
colours are used, these twining cords can
stand out from the POT background, as at
the top of Plate 62, left. The use of twined
linkings leads to designs in which cords
switch from one function to the other. At the
centre of the braid illustrated, twined linkings
were placed obliquely, working from the
edge inwards on both sides, as indicated in
Fig. 142b. Cords which were twining
longitudinally then became part of the POT
ground and moved obliquely in the normal
way, and cords which were in the ground
began to twine. Another set of such twined
linkings restored the cords to their previous
state.
Figure 142 a Structure applying longitudinally

twining cords toa POT ground; b combined with
twined linkings.

CHAPTER NINE
Two-Layered Oblique Interlacing
A two-layered oblique interlacing (TLOD,

with endless design possibilities resulting
from the interchange of elements between
the layers, 1s a surprising structure to be
produced by ply-splitting. Unlike most of the
previously described techniques, it requires
cords made of 2 dark/2 light plies, running in
pairs, as in Fig. 143. Chapter Three
describes the preparation of such cords. If
using a simple drill, follow Fig. 3a in that
chapter; if using the special twister, follow
Fig. 4 with the upper two yarns dark, the
lower two light. These cords are always split
so that either the two dark or the two light
plies are uppermost, the choice naturally
determining the colour visible at that point in
the braid.
1. METHOD
The method can best be understood by
making the simple design in Fig. 144, and
Plate 63, top — an inverted dark triangle on a
light background.
‘Take 8 such two-colour cords and either
Uy
double them over the starting rod as in Fig.
143a, or split them centrally with the rod as
AVM
in Fig. 143b. This gives 16 free-hanging
cords for the braid.
Row 1 This is worked exactly as a long row

in POT. So starting at the right, split the
outermost cord and pull its neighbour to the
left through this split. Then work the next
pair of cords similarly and continue to the
Figure 143 a—b Attaching two-colour cords to the
end of the row, where the penultimate cord is
starting rod; c—d how they appear after the first
row of TLOI. split by the left-hand outermost cord.
TWO-LAYERED OBLIQUE INTERLACING + 145

Now these splittings are not haphazard and pulling them apart. All the light plies will now
must go between the light and dark plies of be visible as at the top of Fig. 144, Row 1. In
each cord. According to the design in Fig. Plate 63, top, the 4 left-hand cords are fixed
144, the two lhght plies should be uppermost, to the rod as in Fig. 143a, the 4 right-hand as
lying over the splitting cord, and so be visible in Fig. 143b.
from the side being worked. This naturally
NOTE
makes the two dark plies lie under this cord
+ how these methods seal the two layers
and so be visible at the back of the braid.
together at the starting edge. See the end of
Fig. 143c—d show this splitting in detail,
this chapter for a method which allows this to
with the light plies lying over the splitting
remain open.
cord. In the case of c it is this row which
actually attaches the doubled cord to the rod. Row 2. The second row, according to Fig.
Tighten this row by taking the split cords in 144, must also have light plies uppermost
one hand and the splitting in the other and at all the splittings. It can be worked from
od oe os ad o © oe ©
x):
Va WS 9, @ Row
.gy |
PY
FW
DO
ON
| co
Figure 144 Analytic diagram for inverted triangle design.

the “same side with leftward splittings. Row 7 ‘Turn the braid back. In the first two
Alternatively and more commonly the braid splittings bring light plies uppermost, in the
can be turned over, when the first row just central 4 bring dark plies uppermost. In
made will appear dark and must be followed neither case leave any twist in these cords, as
by a short row with dark plies uppermost in they were showing the same colours when
every Splitting which is now rightward. last split in Row 5. For the last two splittings,
bring light plies uppermost, but here leave
Row 3 ‘Turn the braid back and work this
a half-twist, as both of these cords were
long row exactly as Row 1. But as each cord
showing dark in Row 5.
must again have its light plies uppermost to
agree with the design, it is essential to
Following the diagram in this way, leaving a
eliminate al] twist between these splittings
half or no twist as the colour in the design
and those of the same cords in Row 1.
indicates, the triangle will gradually taper
Row 4 ‘Turn the braid over and repeat Row until it reaches its point in Row 10. The next
2, again removing all twist between these two rows, though showing light all across,
splittings and those in Row 2. will still contain half-twists to return the plies
of the cords, altered for the triangle, to their
By continuing to work in this way, repeating original position. See the double lines in the
Rows 3 and 4, two identical, but completely diagram. So not until Row 13 will no twist be
separate, layers of oblique interlacing are left in all the cords between successive
produced. A light layer is superimposed on, splittings.
but is unconnected to, a dark layer, a fact The result of the above is an inverted dark
which can be demonstrated by sliding a rod triangle on a light background on one face of
in between them. So this is the braiding the braid, and an inverted light triangle on a
analogue of plain-weave double cloth or dark ground on the other; see Plate 63, top
pocket cloth. and bottom. Working in this way a motif,
however complex, will show identically on
Row 5 However, the design in Fig. 144
the two faces, but with colours reversed.
shows that in Row 5 dark plies begin to
appear, starting the triangular motif. So for
It will be seen that the working method is
the 6 central splittings, bring dark plies
very similar to that of POT, with the same
uppermost; in other words, leave a half-twist
alternation of long and short rows. The
in these cords between these splittings and
crucial difference is that here either no twist
those in Row 3. The half-twist, or two
or a half-twist is left in cords between
quarter-twists, is indicated by two oblique
successive splittings, whereas in POT it is
lines across these cords in Fig. 144. The first
always a quarter-twist which is left. Another
and last splittings in this row continue to
practical difference is that the two colours
show light plies.
make it so much easier to select the correct
Row 6 When the braid is turned over for path for the splitting needle. An error is
this row, the base of the triangle just worked immediately apparent.
will appear as a light line. To continue the The two oblique lines in Fig. 144,
triangle, bring light plies uppermost in the 5 indicating where a half-twist is left in a cord,
central splittings, leaving a half-twist in these are usually omitted as it is always obvious
cords between these and the splittings in from the colour of a cord where it should have
Row 4. In the first and last splittings bring a half and where there should have been no
the dark plies uppermost. twist between splittings.

Fig. 145 gives a highly diagrammatic represent-
ation of the structure. In reality the light elements
are closely interlaced and completely hide the
We
Wy )
similarly interlaced dark elements. It is the 4 ff
mathematically exact placing of visible light over Be LN
invisible dark plies and vice versa which makes any SS SY S
design perfectly reversible. iy=
(KS . (KY
There will always be a potential space or pocket y ~
between the two layers of oblique interlacing. So y Ye p 5
the dark triangle in Fig. 144 is quite separate from ti SS [KR
the light triangle beneath it, visible from the other ai SA LMP SS
side. Fig. 146, an oblique section through Fig. 144, Fiquren|45. Highly disqrmunatie aentat
from arrow to arrow, shows this in an exaggerated TLOI.
manner, the curving lines indicating the pockets.
But once the triangle is finished, its pocket is sealed
on its three boundaries by plies moving between the
two surfaces of the braid.
The method of work implies that the carefully
prepared 4-ply cords attached to the starting rod
are almost completely unplied by the end of a braid.
In fact, the dark and light plies of any cord travel
through a braid as parallel pairs of elements, lying
one above the other, as in Fig. 146. It is only at
colour interchanges that the necessary half-twists
leave a reminder of the ply’s initial state as a 4-ply
cord. So the two-colour cords used in this tech-
nique are just a remarkably ingenious way of
temporarily keeping together, as an easily handled
unit, those dark and light plies which are to shadow
each other throughout the structure.
Of course the cord format also helps to keep the
fell straight and firm when working, as in all ply-
split braiding techniques.
Now if a 4-ply cord is untwisted, e.g. by inserting
a rod between the dark and light plies and forcing it
down to the lower end, these plies return to the
highly twisted state they were in during the cord’s Figure 146 Cross-section through A-A in
making. The present technique to a large degree Fig. 144.

has the same effect. One result is that the over-
twisted 2-ply elements produced in this way tend to
interlace unevenly, giving a distorted surface,
especially in large areas of one colour. So the latter
look markedly different from the smooth regular
one-colour areas produced by POT. Smaller motifs
with their many colour changes and therefore many
half-twists left in the cords show this effect less. All
this implies that cords for this technique do not need
to be prepared with such a high degree of twist.
The cords on S- and Z-courses lie at 45° to the
transverse axis of the braid, so diagrams such as Fig.
144 can be used without distorting reality.
3. DESIGN
Designs can be easily represented by filling in the
squares on a grid; so Fig. 147a is the design for the
inverted triangle in Fig. 144. Once the principle is
understood, such designs contain all the information
necessary for making that motif. It is just a matter of
reading the design, line by transverse line, and
splitting each cord to bring the appropriate colour to
the surface at that point. Initially there may be some
confusion because in one row the braid looks like the
design being followed and in the next, the braid
having been turned over, the colours are reversed.
So a little mental gear-changing is necessary.
Remember, when drawing out a grid, that the
number of squares in a long and short row must
add up to one less than the total number of cords to
be used. So the grid in Fig. 147a has 8 + 7 = 15
squares, one less than the 16 cords in use.
This alternation of odd and even numbers of
crossings (squares in the grid) has a useful side-
effect. It means that whether a motif starts with an
odd or even number of squares, it can always be
centred in the braid by being started in the
appropriate row.
Fig. 147b shows a design on a grid with 12 + 11 =
23 squares in a long and short row and so suitable
for a braid with 24 cords. Reading from the top, the
first two rows are the same colour all across. In the
third row, the 4th and 9th splitting must bring the
opposite colour to the surface to start the two
Figure 147 a Design diagram for Fig. 144; diamonds; so here a half-twist is left. Similarly with
b diagrams for motifs on a wider braid. the 3rd, 4th, 8th and 9th splitting in the fourth row,

and so on. So at each splitting the needle is deciding
which of the two colours is to be visible at that point,
either dictated by a pre-drawn design, as here, or at
the whim of the worker.
The oblique stripes in these diamonds work well,
being parallel to the S- or Z-courses. They will have a
smooth outline even if, as here, they are only one cord
wide; see Plate 64, right. On the other hand, transverse
and longitudinal stripes, seen at the centre of Fig. 147b
and in Plate 64, left, will always have the irregular or
stepped edges shown. They should not be narrower
than those in the outer square in Fig. 147b. Lines as
narrow as those forming the inner square will only look
convincing on one side of the braid, appearing as
disconnected spots on the other side. The different
characters of these colour boundaries can add interest
to a simple design, as in Plate 65, right.
‘The small repeating pattern at the foot of Fig. 147b
resolves itself into a simple sequence. Long rows
alternately have all dark and all light plies uppermost.
The intervening rows are all the same having dark and
light plies alternately uppermost all the way across; see
Plate:-65, left.
Such simple geometric designs give only a hint of
the endless possibilities of this, the most flexible,
technique offered by ply-splitting. Any design that
can be drafted on the grid can be faithfully converted
into a ply-split braid. So obviously the more cords
used, and therefore the wider the grid, the more scope
for elaborate patterning.
Plates 66 and 67 show more complex geometric
designs, one duplicating a hairline pattern found in
POT. Plates 68 to 70 show that men, women, animals,
trees and birds have been represented with varying
degrees of realism by the Indian girth makers. The
technique’s flexibility also makes letters and digits
easy to produce. Some Indian girths bear the maker’s
name, as the name Bhaga Ram in Plate 70, left. The
maker’s village and occasionally even the girth’s price
in rupees is included. Fig. 148 shows a large motif of a
peacock on a tree growing from the water pot a
woman is carrying on her head; it is adapted from a
Rajasthan girth and needs 36 cords.
Plates $id, 97, left, .and 101; tight, i, Chapter
Twelve show complete artefacts from India in this
Figure 148 Large motif based on those
technique.
on Indian camel girths.

4. JOINING THE TWO LAYERS OF All diagrams in Fig. 150 begin with this
INTERLACING open starting edge. The arrowed splittings in
Fig. 150a are the ones which must have all
4.1 At the starting edge their twist eliminated.
The two methods in Fig. 143 will join the two
layers together at the starting edge; the
following method will allow them to stay
separate.
Centre the cords over a stretched line, as in
Fig. 149a. Split each at its midpoint with its
neighbour to the left, keeping the light plies
uppermost. This is a short row. In the
following long row, it is essential that all twist
is eliminated between these splittings and
those in the first row just worked; see Fig.
149b. Then carry on normally. This com-
pletely separates the dark and light plies
giving a scalloped upper edge to both the
unconnected layers when the stretched line is
removed.
MYL
, Mp
‘y
Zi)
=)
D>
AZ
‘ffs
Figure 149 Starting edge of a braid with its two Figure 150 Ways of joining the two layers at the
layers unjoined, a first row, b second row. selvage.

4.2 At the sides If the two layers are joined at only one selvage by
Following the above practical any of these methods, the opening at the other
descriptions, the two layers of oblique selvage allows access to a usable pocket; see the
interlacing will be unconnected at the hundred-rupee note in Plate 72, left.
two selvages, as in Plate 63. So there is a
pocket between these layers stretching
inwards as far as the first colour change
in the design. If this state is not wanted,
there are several ways of sealing the two
layers at the selvage.
4.2a Without added cords

‘The simplest method is to leave a half-
twist in the cords at the first and last
splitting of every long row, as in Fig.
150a. The spots of the opposite colour
thus produced build up into a very
narrow longitudinal stripe, as seen in
Fig. 148 and the top section of Plate
71. Fig. 150b shows an edge with
more obvious striping; once it is
started, leave a half-twist in the first
and last two splittings in a long row
and in the first and last in a short row.
See the next section down in Plate 71
and 68, left. Many other types of edge
stripes are possible, e.g. that in Plate
70, right.
Another way to approach the matter
is consciously to create a regular
succession of small shaped pockets, as
in Fig. 150c. The two layers are then
separate but only as far as the inner
boundaries of the triangular areas; see
next section down in Plate 71 and
Plate 72, right.
If only a quarter-twist is left in the
cords at the first and last splitting of
every long row, the needle passes
under a dark and a light ply and gives a
decorative two-colour sealed edge,
seen in the lowest section of Plate 71.
When these cords are again split in the
next long row, a three-quarter-twist
. Z Figure 151 Making a border of one-colour cords which
has to be left in order to bring the interact with the two-colour cords:
correct colour to the surface. a by linking; b by working in POT; ce in SCOT.

4.2b With added cords twist is left in the split cords in the border.
Used as linked cords See the centre section of Plate 73, left. If the
Begin with three or more one-colour cords twined linkings are only made in every other
attached to the starting rod on either side of row, alternating with slits, as at the top of Fig.
the central two-colour cords. They are not 151b, their bulk will not build up excessively.
involved in any ply-splitting but each in turn
is linked, not twine-linked, with the first and Using SCOT
last cord worked in a long row of the design, Here there are more variations as the border
as in Fig. 151a. They wrap around each other can have one or more sections. The good-
in the manner shown. They can be inserted looking border in Fig. 151c has two sections
quite loosely then periodically the slack and is easier to work if there is an odd
pulled up to make a strong rounded edge, as number of interlacing cords. All twined
in Plate 73, right. This can be an addition to linkings are of the R/L type.
an already closed edge, as in Fig. 151a. As a preliminary move, pull the left outer
border cord through the next two, then turn
Used as a ply-split border
the braid over. Now pull the right inner
Here one-colour cords, added at each side, border cord through the two nearest border
are worked in POT or SCOT to make a cords; see the right-hand arrow in Fig. 15Ic.
narrow border, twine-linked to the two- Work the interlacing cords normally, starting
colour cords in the central design area. with an R/L twined linking of the first
Examples follow, both needing five added interlacing cord with the nearest border cord;
cords. see second arrow from the right in the
diagram. Finally pull the left outer border
Using POT
cord through the next two and turn the braid
In Fig. 151b a short row has just been worked over.
and the long row begins thus: 2 > 1; 4 > 3; Repeat the above movements for every
6> 5, 326) Ge. a” R/K twined linking row. The arrows in Fig. 151c show these
between cord 5 and 6); 8 > 7; 10 > 9 and so moves and thus the appearance of the braid
on with a similar twined linking between a at the end of any row before turning over.
border and interlacing cord at the left. The Using the more compact R/L twined linking
short row is normal. Naturally a quarter- reduces the tendency for their longitudinal
placing to distort the fell. See upper section
Of Plater73; lett:
The added border cords can be of a third
colour, as in the diagrams and Plates, or the
same colour as the one used in the central area.
By combining methods 1 and 2 above, a
bag can easily be made. It will be open at the
top (the starting edge of the braid), sealed at
the sides using the methods in Figs 150 and
151 and sealed at the bottom by the unsplit
cords. These can make a fringe or be hidden
inside if the bag is turned inside-out.
Figure 152 a Cross-section of braid in Fig. 151b—c;
b making a pocket open at one side; A transverse section through Fig. 151b orc
c apocket only accessible from one side. is seen in Fig. 152a. So there is a two-layered

interlaced braid sealed at its edges by a narrow them completely. Then tape or tie the four
one-layer, ply-split, braid. But if the right light plies together and the four dark plies
border is worked for some distance as a together at the level of the rod. Cut the plies
separate braid unconnected by twined linkings below these ties and they will immediately re-
to the interlacing cords, the section will be as ply into a dark and a light cord. Using
in Fig. 152b. So there is a pocket which opens crossed sticks, as in Fig. 4d (Chapter Three),
against the side of the border braid. and moving them up the cords ensures the
If, however, the border cords on the right plies lie correctly.
are connected to the interlacing cords by Obviously the reverse can be done,
incomplete linkings (see Chapter Eleven), converting one-colour into two-colour cords.
then, as Fig. 152c shows, the pocket’s This conversion, simple though slow, means
opening will only be accessible from one side a braid at any desired point can switch
of the braid, being invisible from the other; between those structures needing two-colour
see lower section of Plate 73, left, with a coin and those needing one-colour cords.
in the pocket.
Sox x :
5. RECOGNITION OF PLY-SPLIT TLOI
As ‘TLOI can be produced in other ways
(using free-hanging cords, double interlaced
sprang, etc.), it is important to be able to
recognise the variety produced by ply-
splitting 4-ply cords. An obvious pointer is the
presence of such cords emerging at the lower
end of a braid or visible at the beginning. Also
other ply-split structures may be associated
with it, such as the borders just described.
In the absence of these clues, if one cord is
carefully followed through a complex design,
it will be found that the half-twists at each
colour change are consistently in the same
direction. This is the direction of ply of the
cord and so will be Z if the 2-ply elements are
seen to be S-plied; see Fig. 146. Occasionally
an Indian girth maker will at some point unply
his dark/light cords, carefully re-ply them as
all dark and all light cords and then continue
working in either POT or SCOT, an obvious
indication of the method used previously.
‘This transformation can be done as follows.
Converting two-colour into one-colour cords

Put a rod through two adjacent two-colour
cords just under the braid’s fell so that it
passes under 4 plies of the same colour, two
from one cord, two from the other. Force it
down to the cords’ lower ends, unplying Figure 153 Using three colours in TLOI.

6. Using three colours crossing and so the design becomes that in
So far in this chapter it has been assumed the lower third of Fig. 153. Twined linkings
that all cords in a braid were made of the placed transversely across these diamonds
same two colours. Possibilities in design have the same effect. See Plate 74, with
multiply even if only one more colour is detail from Plate 97, top.
added with, for example, some cords having ‘Twined linkings can be used in other ways,
2A/2B plies and some 2A/2C plies. e.g. to take all colour B to the right side of the
With a warp of 6 x 2A/2B; 12 x 2A/2C;6 x braid, all C to the left; or to rearrange B and
2A/2B, a braid can be made with one layer C to give hairline patterns. All such twined
showing solid colour A and the other quite linkings are invisible on one side, the front
separate layer showing the familiar pattern of side, of the braid but very obvious on the
diamonds and triangles in colours B and C. reverse.
In the upper third of Fig. 153, the A layer is Though made in exactly the same way as
just visible behind the B and C layer. By other twined linkings, each of the above is in
interchanging layers, areas of A can be reality a two-layered linking. What precedes
placed on this background of B and C. If and follows it removes any element of
such areas exactly cover the chequered twining. Fig. 154 shows that structurally an
diamonds where B crosses C, the design A with A linking lies directly over, and is
becomes that in the centre of Fig. 153. unconnected with, a B with C linking; thus
Placing twined linkings longitudinally the separation of the two layers of
down the centre of the A diamonds (dots in surrounding interlacing is preserved.
the diagram) prevents colours B and C from
Figure 154 Highly diagrammatic view of twined

linkings when using three colours.
TWO-LAYERED OBLIQUE INTERLACING «+ 155

CHAPTER TEN
2/2 Twill Oblique Twining
This structure can be defined as oblique each side, as at the bottom of Fig. 155.
twining where each cord is alternately visible Then 1 >2; 4>5 (Gn same split as used
for two crossings and then concealed for two previously), 6; 8>9 (Gn same split as
crossings with countered cords, its path before), 10.
being set over by one cord from its Turn the piece over and then 12> 11(in
neighbours. same split as before), 8; 9 > 7 (in same split as
The twill ridges so formed can be before) 4: 5 > 3 (in same split as before ), 1.
transverse or longitudinal as with 2/2 oblique ‘Turn over and repeat this cycle.
interlacing. Structurally each cord splits two So the needle in each upward movement
adjacent cords and is itself split by two splits one cord in the normal way, leaving a
adjacent cords, both of which pass through quarter-twist between this and any previous
the same opening. This draws all cords closer splits, and then splits another cord in the
together, giving a braid thicker and narrower same opening as its previous split, i.e. leaving
than a plain obliquely twined braid with the no twist. As a result the same two plies of a
same number of cords. cord stay uppermost in the two successive
splittings of that cord, giving a surface
texture identical to 2/2 twill oblique
1. WITH LONGITUDINAL RIDGES interlacing with paired elements; see the
The method is as follows starting with a structure in the diagram. It is difficult to see if
zigzag fell with two cords emerging from exactly a quarter-twist has been left as the
Figure 155 /eft Working diagram for 2/2 twill

oblique twining with longitudinal ridges.
Figure 156 above Method of fixing to starting rod

to obtain correct starting colour sequence.

cord concerned is hidden by two cords the diagram. The working of every row is the
between this splitting point and its previous same. So turn the piece over and repeat the
appearance in the braid, but as long as the above movements exactly to make Row 2. It
two plies lie parallel side by side in the latter, will be found in this second row (and in all
the amount of twist is not crucial. subsequent rows) that the first cord split was
If the cords start with the symmetrical the last one split in the previous row. So leave
colour arrangement shown, the transverse two quarter-twists between these two
zigzags pattern in Fig. 155 and at the top of splittings to give a good selvage, as the
Plate 75 emerges. This colour arrangement is diagram indicates. Longitudinal zigzags
easily achieved if the cords are fixed to the appear, as in the centre of Plate 75, if the
- starting rod by the splitting method. Split a braid is started with the colours shown, an
light cord only once, then a dark cord twice arrangement achieved as described above.
in the normal way, then give the light cord its So in both cases the zigzags run at right
second split; see Fig. 156. Do this all across. angles to the ridges and in fact somewhat
‘Then of each dark pair lying adjacent to each obscure their presence.
other pull the right-hand cord through the If the starting sequence were of the type 2
left-hand and the cords are arranged as at the light, 4 dark, 4 light, 4 dark, 2 light, the
bottom of Fig. 155. diagonal stripes seen at the bottom of Plate
75 are produced. This colour sequence can
be derived from that used above by twine
2. WITH TRANSVERSE RIDGES linking in a row where all the crossings are
Starting with a transverse fell as at the dark with light, e.g. the point arrowed in Fig.
bottom of Fig. 157, work across thus: 2 > 1; 157. Twine linking at two adjacent crossing
4 > 3,1 Gnsame split as before); 6 > 5, 3 Gn points, not at the next two, and so on, if
same split as before); 8 > 7, 5 (in same split started correctly, effects the change neatly.
as before); and so on. So in all except the first Other designs derived from colour and
move, the needle splits two cords, the first in weave designs in 2/2 twill weaving can be
the same split as before, the second in a new worked out.
split, a quarter-twist beyond its previous In a similar way, ply-split structures can be
split. This gives a row like Row 1 at the top of based on other twills, giving wider ridges or
ridges of unequal width.
Figure 157 Working diagram for 2/2 twill

oblique twining with transverse ridges.
2/2 TWILL OBLIQUE TWINING * 157

CHAPTER ELEVEN
Twined Linking
1. INTRODUCTION
In a twined linking, when two cords
on opposite courses meet, one splits
the other in the normal way, then
immediately the same two cords are
involved in a further splitting, both
splittings taking place in the same row.
This manoeuvre has the effect of
suddenly turning the oblique course of
both cords through a right angle. As
Figs 158a and 159a show, the light
cord after the twined linking continues
on the S-course previously held by the
dark cord; and the dark cord continues
on the Z-course previously held by the
light; in other words the two cords
swap courses.
So the effect on the movement of the
cords is identical to that in simple
linking, but here it is made more stable
by the two ply-splittings involved.
2. TWO TYPES
‘There are two types. In Fig. 158a, the
dark cord splits the light, then
immediately the light splits the dark.
‘These two Rightward splittings make
this a R/R twined linking. In Fig. 159a,
the dark splits the light and then splits
it again, leaving a quarter- or half-twist
between these two splittings. The
Rightward then Leftward splitting
Figure 158 R/R twined linking: a top left open and
makes this a R/L twined linking. It
b top right pulled tight; c Jower left its working diagram;
d lower middle its symbol; e Jower right symbol indicating must be pointed out that R/R and R/L
either R/R or R/L type. refer to the appearance of these twined

OY/,
W/p
linkings on the front of the braid as it is

worked; on the back the appearance will be
hy
YOY
d e
4
\ a the opposite. Both types effect the two cord’s
courses in the same way, but they differ
visually and in their degree of stability.
Once two cords are joined in a R/R twined
linking, they are tightly locked together, as
Figure 159 R/Ltwined linking: a top /eft open in Fig. 158b, making it the more stable type.
and b top middle pulled tight; c top right leaving
Fig. 158c shows the working diagram and
only a % twist between the two splittings; d its
working diagram; e its symbol. Fig. 158d the simple symbol which will be
used in this book.
When two cords are joined by a R/L twined
linking and pulled tight, as in Fig. 159b, the
dark cord can easily slide back and forth
through the two splits in the light, so it is only
stabilised by the normal structure in rows
preceding and following it. Fig. 159d shows
the working diagram and Fig. 159e the
symbol for this. Fig. 159c shows a common
variation where only a quarter-twist is left
between the two splittings.
It is often immaterial which type is used
and the twined linking can then be
represented by the simple circle in Fig. 158e.
‘Though both types involve two splittings, the
R/R type is the bulkier of the two and if
frequently used its cords have to be pulled
really tight to avoid distortion of the fabric.
TWINED LINKING + 159

2. USES OF R/R TWINED LINKING
2.1 As the sole structural element

One example is the large meshed netting made from
rope, found in children’s playgrounds, where a
conventionally knotted net would be unpleasantly
lumpy. The size of the mesh results from maybe 30
cm of rope being left between adjacent twined
linkings.
At the other extreme, a stable yet stretchy fabric
can be made when only a quarter- or half-twist in
each cord is left between adjacent linkings. It is done
as follows. Make a row of R/R twined linkings, Row
1 in Fig. 160, using all the cords. Then without
turning the piece over, make another row excluding
the outer cord at each side. This is the shorter Row 2
in the diagram. Repeat these long and short rows
alternately. At the first split of each twined linking
leave a quarter-twist between this and the previous
splitting; but at the second split leave more, maybe
half or even a whole twist, because as the diagram
shows the previous splitting of this cord is some way
back in the structure. The fine oblique lines in the
diagram show these degrees of twist.
On the side as worked, the fabric has a distinctive
and pleasing texture, dominated by the twice split
cords on the Z-course. See Plate 116. But if the Figure 160 Structure made entirely of
R/R twined linkings.
piece had been turned over between rows and R/R
twined linkings always made, then viewed from
either side the rows would be alternately R/R and
L/L and the surface appearance quite different.
Cords can work in pairs, as in Fig. 161, where
cords A and B pass through the same split in cords C
and D, and then C and D pass through A and B
similarly. The result is a flat and very secure twined
linking; the cords of a pair can form a loop, as
shown. See Plate 112.
2.2 As the means of changing the courses of dark

and light cords
In the above examples, the structure consisted of
twined linkings throughout and it was assumed all
the cords were of the same colour. But it is when
individual twined linkings are inserted into a two-
colour piece which is otherwise in POT or SCOT Figure 161 Twined linking of paired
that they find their most ingenious use. cords.

As the aim is to alter the course of cords in the
interest of design, a twined linking is only placed
where a dark cord on one course meets a light on the
other course. This switching of their courses — a
deceptively simple but subtle manoeuvre — can
produce a profound change in the colour sequence
of the cords on both courses and so in the design
being worked. Such twined linkings can be placed in
longitudinal, transverse or oblique lines; these can
be repeated within a design or just occur once to set
a new colour sequence at the start of a design. See
many examples in Chapters Six to Nine.
2.3 As a textural effect

One or two rows of twined linkings make an effective
line of different texture in a one-colour fabric made
in POT or SCOT. See Plate 114, top.
2.4 Other uses

There are two rightward splittings in a normal R/R
twined linking. This number can be increased either
to enlarge the size of the mesh in the type of open-
work fabric described above or to make a narrow
braid from only two cords. In both cases it must be
decided how much twist to leave between the
splittings. When making the narrow braid, if a half-
twist is left between adjacent splittings, the result is a
4-strand 2-ridge twill braid, as at the top of Fig. 162.
If no twist is left, the 8 plies involved reform
strangely into two separate 4-ply cords, one being
S-, one Z-ply, as at the bottom of Fig. 162. The
structure becomes far more intricate if a quarter-
twist is left, showing how even the simplest ply-split
operation can lead to hidden layers of complexity.
A length of handmade rope usually has an integral
loop at one end. If this is pushed through the plies, as
in Fig. 163, then the whole length of rope threaded
through the loop, as shown by the arrow, a R/R
twined linking is made giving a secure noose at the
rope’s end.
Figure 162 Making braid with repeated
R/R twined linkings of two cords.
TWINED LINKING « 161

3. USES OF R/L TWINED LINKINGS
Because of the instability of this type, it is not
suitable as the sole structural element in a fabric. In
fact, the only known occurrence of its use in this role
is in open-work bags from Peru, maybe dated before
AD 1400. Pairs of 2-ply elements work around the
bag, being linked to the previous row with R/L
twined linkings. (Mary Frame; correspondence)
It is most useful in connection with SCOT as the
means of switching courses of dark and light cords.
The fact that the split cord, light in Fig. 159a and
159b, completely hides the dark splitting cord when
Figure 163 Making secure loop at end of
the linking is compressed between the preceding and rope with a R/R twined linking.
following rows becomes an essential feature of the
designs produced; see examples in Chapter Seven.
Because it is less bulky than the R/R type, it finds a
use where many linkings are stacked one above the
other, e.g. between a border and a central area, and
does not lead to any distortion of the fell.
4. INCOMPLETE LINKING
A simple form of linking between two cords which
involves no twining is shown in Fig. 164. There is
only one splitting. It can be used between the border
and a pocket in TLOI.
5. TWO-LAYERED LINKING
By eliminating all twist between two successive R/R
twined linkings, a configuration arises containing no
element of twining, so it is simply linking. This
happens when the manoeuvre is introduced in TLOI
to exchange colours on one side of the fabric; see
centre of Fig. 154 in Chapter Nine. But if used
consistently it can also be the basis of the following
structure, which exists in its own right. is
If a loosely made R/R twined linking is opened out

it will appear as in Fig. 165. It consists of two quite pry,
MM
/
separate linkings, each involving two plies from
cord A, two from cord B. It will be seen there is a Z
linking superimposed on a S linking.
Now if a fabric is made, row by row, entirely of R/R
Figure 164 Incomplete linking.
twined linkings, and no twist is left between any two
successive splittings of cords, it will always be the
same two plies linking in the front and the same two
behind. So two quite separate layers of linking will be

produced simultaneously, as in Fig. 166. Note how
the plying of each cord, seen at the bottom, is
completely undone by this process. The diagram
shows at the top arrow that a row of R/R twined
linking gives a row of Z linking in front of a row of S
linking. When the fabric is turned over and the next
pt,pi) ~&
CESa
* row of R/R twined linkings worked from that side, it
fea
will add a row of S linking in front and Z linking
behind. Both layers will therefore consist of alternate
tip
rows of S and Z linking.
The technique can be applied with cords of one
Figure 165 Opened-out R/R twined linking.
colour but it is easier and more interesting to use
cords having 2 dark, 2 light plies, as in Figs 165 and
Figure 166 Two-layered linking made from 166. In this way a dark and a light linked fabric will
repeated rows of R/R twined linking, leaving be produced, one lying over the other. A structure
no twist between successive splittings.
which has the stretch properties possessed by all
linked fabrics has several applications.
The two layers can be joined at the selvages by
leaving a half-twist in the cords involved in the first
and last twined linking in a long row, thus producing
a tube.
If at any point in a row a half-twist is left in the
cord as a twined linking is made, the colours
interchange between the back and front layer. So
two-colour designs are easily worked.
If the fabric is never turned over between rows of R/
R twined linking, one layer will show only S linking,
the other only Z linking, giving a different surface
texture.
By placing R/R and R/L twined linkings in the
same row, variations of linking normally associated
with the sprang method can arise, such as hole
designs.
TWINED LINKING * 163

CHAPTER TWELVE
Traditional Cord Preparation
INTRODUCTION their even greater dustiness. My inspection

This part of the book contains details of all the and photography followed, eager faces being
ply-split artefacts I found on five trips to reluctantly excluded from the shots.
Gujurat and Rajasthan between 1987 and We soon adopted a code phrase, ‘Does he
1996. An attempt at classifying each want to get rid of this one?’ meaning this is the
functional type - e.g., girths, bags — is one I want. Bargaining would then begin
provisional as on each visit I found new between Mr Tak and the owner, who by now
examples or structures. So Iam conscious this had appeared. Mr Tak said he found this
is not a complete listing; further fieldwork, if difficult because his sympathies were divided.
carried out soon, before insects consume the He both wanted the villagers, his countrymen,
evidence, might yield a rich harvest. This is to get a decent price and for me, his employer,
especially so in the areas of the Thar desert on not to be overcharged. Doubts as to a tang’s
either side of the India—Pakistan border, from availability were often dispelled when the
which the military usually excludes visitors. owner was told its picture and his name would
That these trips produced what they did is appear in a book. Sometimes, thinking they
largely due to the efforts of our driver, might understand numbers in English, Mr
interpreter and friend, the late Mr ‘Tak of ‘Tak would stretch my mental arithmetic by
Jodhpur. He was raised in a village, joined saying, “He is asking one and half times two
the air force and with an enquiring mind oh oh rupees. Will you pay that?’
educated himself especially about the ability Occasionally negotiations would reach an
of villagers to survive on ‘what nature gave impasse, the owner not agreeing to Mr Tak’s
them in these desert areas’. He proved an idea of a fair price. He would suddenly stand
invaluable bridge between us Europeans and up and say, ‘We go.’ I would be in a panic, not
the inhabitants of the isolated Rajasthani wanting to leave behind some excellent tang
villages and farms we visited. and in fact being willing to pay much more
He would take us to a village, begin talking for it. In the white Ambassador car, Mr Tak
to some older men about the crops or the drove away slowly, his eyes fixed on the rear-
rains, then bring the conversation round to view mirror. Then a sudden stop as he saw
‘tangs’, the girths I was seeking. Soon young the owner running after us, waving the tang
bystanders would be sent scampering off to and shouting that he accepted our price. This
fetch pieces from their homes. We would ruse never failed to work or to scare me.
probably by then be sitting in the shade on The only real disappointments came when
charpoys, perhaps drinking tea, but refusing a woman showed us a beautiful tang but her
the buttermilk which Mr ‘Tak drank with husband being away she had not the
pleasure. If we were lucky, tangs would arrive, authority to sell it; or an absent man had
being banged on the dusty ground to reduce locked in a chest tangs which his family

always described in terms that made their 1.2 Cotton
non-visiblity hard to bear. Machine-spun cotton is available and of
Mr'Tak’s involvement in these ‘tang hunts’ course can be dyed as required. Its lack of
was such that he began to take it personally if springiness is compensated for by an ex-
a village revealed nothing or if some tang I tremely tight plying of the cords which can
really wanted could not be purchased. Also lead to the girth being unpleasantly stiff and
memorable was his way, when hungry him- twisting on its long axis. However, that cotton is
self, of saying to our host of the moment, not just a poor substitute for goat hair is shown
‘Madam is hungry’ — the prelude to the by a strong tradition of multicolour cotton
inevitable and delicious millet chapati and girths, especially in eastern Rajasthan; see later.
vegetables, eaten outside or lounging in the A preference for cotton is sometimes expressed
village guest house. due to its resistance to pests as compared with
goat hair. Against this, cotton absorbs dirt in
The most common use of ply-splitting in contact with the camel’s body, whereas goat
India is in the making of camel girths, so it is hair wears clean. In addition, cotton’s dyed
the cord preparation for these artefacts colours fade faster than does black goat hair.
which is now described. Because it is the Presumably due to the relative scarcity of
girth-maker himself who always prepares his white goat hair, girths are sometimes seen
own cords, he can ensure they will possess with the white cords made of cotton, the
the physical properties he knows are black of goat hair, this mix of materials
essential. These in turn affect the strength apparently presenting no technical problems.
and durability of the finished girth, which in A few makers say black goat hair is stronger
all probability will be for his own use. For than white, and this is the reason for the
these reasons he works slowly and with existence of so many all-black tangs. Whereas
almost exaggerated care. goat-hair 4-ply cords always twist in the Z
direction, cotton cords sometimes twist in the
1. MATERIALS USED S direction because of the initial twist
direction of the commercial cotton yarn from
1.1 Goat hair which they are constructed.
The cord must be tightly plied and if possible
1.3 Other materials
made from a fibre of natural springiness. The
latter quality, together with its strength and Other materials are so rare that only one
easy availability from local herds, probably example of each of the following has been seen.
explains why black and white goat hair is the aak — a coarse fibre extracted from the stem
most used material. As the white is only rarely of the aak bush, found especially in the
dyed, colours are limited to a natural black — western part of the Thar desert. Its poisonous
fading to brown — and a natural white — nature ensures that the fibre resists insect
changing over time to cream. The contrast attack. Plate 82, left, shows a tang bought at
between the two is often further lessened in Mandawas, south of Jodphur, in 1994, with
an old girth by its being stored in roof rafters, 2-ply cords, dyed pink and green. Ishwar
where the smoke gradually changes it into Singh of Jaiselmer knows the long and
two shades of brown. As there are many more complex process of extracting this fibre.
black than white goats, if one colour is to
predominate perhaps as a border it is with jute —a tang seen at Roicha, south-west of
very few exceptions the black. Jodhpur, in 1996, had 2-ply jute cords used
in conjunction with white cotton cords.
TRADITIONAL CORD PREPARATION «+ 165

camel hair-—a tang bought at Bandhani, east of
Bhuj, had 4-ply camel-hair cords together with black
goat-hair cords.
The preparation of goat-hair cords with the usual

formula of Z-2S-4Z will now be described
2. PREPARATION OF THE FIBRES

LE
The fibres when cleaned need to be in a compact b —————
_—
portable form because they are often spun while the

spinner is walking around, herding his animals. The
solution evolved converts the fibres into a long thick
Us
~~ = wear =zTEE
ZZ
aS
om
roving formed into a ring which the spinner, if right-

handed, carries over his left arm. It is called a
beardhee (= knot) in Rajasthan and a vindhi (=
ring) in Kutch, Gujurat.
The method of its production is well documented
in Quick and Stein (1982), though it should be
noted that their informant, Ishwar Singh, is left-
handed. Such a detailed description is omitted here
as it is not specific to ply-split artefacts. The fibre,
either goat or camel hair, to be spun for any heavy
textile such as a rug is prepared in this way. The
following summary, with references to Fig.167,
outlines the process.
a The knotted end of a long over-twisted roving is

allowed to double back on itself, making a short 2-
ply length with a final blind loop.
b_ The rest of the roving is wrapped continuously

around the knees (X in diagram) of the squatting
Figure 167 Stages in making a beardhee.
maker, in a circular manner.
c Removed, it is allowed to twist back on itself as a

multi-ply rope of roving.
d_ Bent into a ring, this rope is secured by passing

the blind loop through this rope’s end loops and then
the knot through the blind loop; see arrow in
diagram.
e Any extra roving is wrapped tightly round this

ring.
See Plate 76 for two examples and 77 and 78 for

their use.

3. SPINNING rotates the hanging whorl (in a clockwise
direction seen from above) and spins a
3.1 Device used (dhera) convenient length. He winds this temporarily
Most western drop spindles have a central around his left hand. He places a small twig
vertical shaft, the spindle itself, which pierces or thorn on the whorl and then winds the
a heavy disc, the whorl. The spun yarn is yarn from his hand around the whorl and
wound around the spindle, while the twig. Finally he catches the yarn around both
momentum for the spinning comes from the ends of the twig to prevent it unwinding; see
whorl acting as a flywheel. It is therefore Fig. 168. He continues thus, occasionally
difficult to call the spinning device used in repositioning the twig nearer the surface of
India a spindle as it often consists only of a the ball of spun thread as it increases in size.
whorl, called a dhera. The latter is a cylinder In spinning the hands are held together as a
of hardwood, about 20 cm or more long by unit, the gap between them bridged by a
2.5 cm in diameter, thicker at its centre; see finger from the left hand; see spinners in
examples in plate 77. Suspended at its
midpoint it is spun round horizontally. Its
weight, being at least twice that of most
European spindles, gives it great momentum;
this is essential for producing the high degree
of twist required in some stages of cord
making. The whorl also acts as the core
around which the yarn is wound. The latter
naturally adds weight and improves the
momentum. Being at the whorl’s centre of
gravity it has no tendency to upset its balance.
Sometimes a vertical stick or metal rod
passes through a hole in the whorl’s centre,
see Plate 77, top, and the spun yarn is led up Figure 168 Dhera with spun yarn
to a hook at its tip, as in European spindles, attached by means of small stick.
but the yarn is still wound on round the
whorl. More often this stick is dispensed with
and, as in the following description, a small
twig pushed into the already spun threads
prevents them unwinding.
This device is used in the three stages of
producing a cord. First, spinning a single
yarn, then plying two of these into a 2-ply
yarn, then plying four of the latter to make
the final 4-strand cord used in ply-splitting.
3.2 Spinning the single yarn

Pulling some fibres from the beardhee
carried on his left arm, the spinner twists
them with his fingers into a fine Z-twist
single yarn. When it is long enough he ties it Figure 169 Forming ball as yarn is
round the centre of the whorl. He then wound on to dhera.
TRADITIONAL CORD PREPARATION « 167

Plate 78. In this way drafting can be exactly 3.3. Making the 2-ply yarn
controlled even though the spinner may be The spinner takes one end of yarn from the
moving about. The spinning is slow and centre of the ball and one from the outside,
carried out with care; it results in a worsted ties them to the whorl and rotating the latter
thread of great consistency. in an anticlockwise direction makes a 2-ply
The yarn is wound around the whorl with yarn, putting in a high degree of S-twist. The
the same meticulousness; it follows such an method is similar to that described above and
exact course that the finished ball has an ends with a ball the same size but of 2-ply. A
almost mechanical perfection. The yarn is ball may be made first from the two ends of
laid in the same vertical line each time by the the singles ball and then twist added as
left hand, see large arrow in Fig. 169. But described.
because the right hand slowly rotates the
whorl, as shown by the small arrow, the yarn’s There are two distinct methods of combining
course is slightly shifted each lap it makes. It four 2-ply yarns into a cord. In both, the 2-
crosses the previously laid lap twice, once at ply is folded back on itself twice to provide
the top, once at the bottom; it is a course the four components of the cord, but in one
similar to that used winding a turban. method each cord is a separate entity, in the
A full-sized ball may weigh 450 gm, so to other each cord is connected to its neighbour
reduce the weight on the spinner’s hands the to right and left.
whorl may be pulled out towards the end of
the process, the ball itself becoming the 3.4 Making the 4-ply cords
whorl; see Plate 78, top left. The whorl is in
any case slid out when the ball is complete 3.4a As separate cords
(the vertical hooked stick, if present, being The required length of cords naturally
first withdrawn), leaving a perfectly made depends on what is to be made, but it will be
ball of Z-twist single yarn, ready for the next in the region of 3 to 4 metres. Assuming he
stage. See bottom of Plate 77. has an assistant, the spinner sits at this
distance from a peg in the ground or a hook
in a wall. From the ball beside him, he leads
the 2-ply yarn around a stick, A in Fig. 170,
held by the assistant, and another, B, held by
himself; then around A again and finally ties
it to B; see the configuration in the diagram.
‘The assistant moves towards the peg (arrow)
drawing yarn from the ball as the four
strands of 2-ply lengthen between him and
the spinner. Reaching the peg, he puts on it
the two loops from his stick.
The spinner now cuts the yarn coming
from the ball and unties the end from his
stick. He removes the stick letting his left
index finger take its place. He takes the free
end, which was tied to the stick, across then
under the two yarns lying over his index
Figure 170 Arrangement of 2-ply yarn around two finger, as in Fig. 171a. Then he ties it with a
sticks at start of making a separate 4-ply cord. reef (square) knot to the other free end which

came from the ball. In this way he neatly
secures both the loop and the two cut ends, as
in Fig. 171b.
Rolling this knotted end between the palms
of his hands, the spinner plies the four
strands in the Z direction as far as their twist
warrants. The assistant removes the looped
end from the peg. Note that the peg end of
the cord has two blind loops, the looped end,
and the spinner’s end has one loop and a
knot, the knotted end.
He repeats this process until a sufficient
number of lightly plied cords have been
made. He then joins them end to end by
passing the knotted end through the looped
end and tying a bowline, as in Fig. 172a. The
cords are now in one continuous length
which he winds into a ball, starting with a
knotted end. He attaches the final looped end
around the whorl with a lark’s-head knot, as
in Fig. 172b.
Rotating the whorl clockwise, he adds a high
degree of extra Z twist to the cords, building
Figure 171 a How the spinner secures the near up a ball on the whorl as done previously. He
end of a newly made cord; b the knot thus then removes the whorl and immerses the ball
produced. in water, soaking it thoroughly. Finally he
stretches the length of connected cords tightly
between trees or some convenient posts and
lets it dry. This wet treatment stabilises the
cords which when dry have no tendency to
shed their extra twist.
The over-twisting of the cord takes the
plying of its four elements well beyond the
point which their degree of twist would
produce. Notice that each extra turn of the
cord’s over-twist robs the 2-ply strands of
one turn of their twist, so as the cord
becomes tighter the strands become softer
and more flexible. This is a_ subtle
procedure; for to reach the required twist in
the cord by grossly over-twisting its four
elements and then allowing them to twist
back on themselves would lead to a cord of
rope-like stiffness. The over-twisted cord
Figure 172 a Join between lengths of 4-ply cords; when made as above is unstable and so
b fastening cord to dhera for over-twisting. needs the wet treatment to set it.
TRADITIONAL CORD PREPARATION °« 169

Variations in working methods
The above description is based on demonstrations
given by Ishwar Singh of Jaiselmer in 1987, but
many variations have been either seen or described.
Some others have been inferred from analysis of
finished girths. These are now listed, with the
informant, when known.
W
WAYe a
After over-twisting, the cords are soaked in water ~N WY

“a
s
then re-twisted and dried as separate cords. When RAS

ws
ESAS
dry each cord is held at its ends by two men who Sy SAY
stretch it and add yet more twist, while a third man CxSSww
ON, 8
RAY
squeezes the cord in a wet cloth and runs this up and »y WE
y: AQAY
down its length. Then the cord is again dried under RA 8
AN
Be
N
Crs aN.
tension. This illustrates the great care a girth-maker OS
NO. S
N Ws.AS,
will take to produce suitable cords. . !
iS
We
Amana Ram, near Pali village, Rajasthan, 1987 wy
ANG
“ARN
Girths are commonly seen with cords which taper to
a single loop at their lower end, see Fig. 173b. This
SN
W SYN
AA
AS
neat arrangement, which makes it very easy to pass ANE
|
J
the cord through the eye of the needle used in ply- Sa

S
AN
WS
splitting, is achieved as follows. When a cord is being
made, the spinner does not knot the two free ends BS
SS
but arranges them so they are shorter than the loop at ee.
‘SN
his end; see Fig. 173a. He unplies these two ends
slightly and pulls out some fibres to taper them. Then
as the cord is given its twist between the spinner’s
Figure 173 a Shortening and tapering
hands, these two ends become completely buried in the two free ends of 2-ply to produce
its centre and the cord appears as shown. b a tapered end to the cord.
Khema Ram Machra, at Ghevra village,
north of fodhpur, 1987; and Deepa Ram, Figure 174 Arrangement of dark and
at Kabrakala, also north of fodhpur, 1989 light cords around sticks when making
two-colour cords.
The assistant may dispense with the stick in any of
these methods and simply take a loop in either hand
as he moves away from the spinner.
For TLOI, the cords must each consist of two

strands of black and two strands of white goat hair.
So the spinner has a ball of each colour beside him
and leads them both over his assistant’s stick and ties
them both to his own stick; see Fig. 174. When the
black and white loops have been placed over the peg
by the assistant, the spinner cuts and unties the yarns
at his end. Holding the two free ends of black in one
hand and of white in the other, he opens his arms to
separate the colours cleanly down their length. Then

When these last few centimetres form into a
4-ply cord, the two colours are naturally
muddled and not clear cut.
In all the above methods it is the end of the

cord with two loops, the end which went over
the peg, which is slipped on to the starting rod
before ply-splitting begins, as in Fig. 176a.
3.4b As connected cords

‘Though frequently found on both old and new
girths and also described by makers, this
Figure 175 Making connected cords. method was only seen once; it was demon-
strated by Pakistani refugees in Jhura-Sodha
Camp, north of Bhuj, Kutch, Gujurat, in 1990.
he ties them together, using a hitch of one of
Several men took part, at least three or four
the plies around the other three, as seen in
of whom were essential. They drove two pegs
Fig. 176b. He then twists them in his hands in
into the ground, A and B in Fig. 175, their
the usual way. Because of this separating
distance apart giving the approximate length
movement he hopes the black and white
of the final cords. The over-twisted 2-ply
strands will lie as pairs as in Fig. 176b.
yarn was tied to peg A, then carried
Ishwar Singh, faiselmer, 1987
continuously around the two pegs, being
Another way of securing the four cut ends is built up carefully at each peg so there was no
to double back one and let it twist up with overlapping, see Plate 79, top. Two journeys
another. Done twice this gives two end loops. around the pegs gave the required four
Figure 176 How cods are attached to starting rod: a and b with separately made cords;
c and d with connected cords.
TRADITIONAL CORD PREPARATION « 171

elements for one cord, so the process The result of this process was that each
stopped when an even number of cords lay cord had two blind loops at its free lower end
around B, equal to half the number of cords while its loops at the other end around the rod
to be produced. The end of the yarn was then were connected with those of the adjacent
tied to A. cords; see Fig. 176c. The two lower loops
While this was happening several sticks tend to separate during the working process
were cut with a natural hook at one end. The and on finished girths they are sometimes
top two loops were taken from peg B and put seen still tied together with fine cotton, as in
on a hooked stick; see Fig. 175. The shaft of Fig. 176c, to prevent this.
the stick was then rolled between the palms If black and white cords are wanted in
of the hands by one man to give Z-twist to some sequence on the starting rod, the yarns
the four 2-ply yarns, while another man at are wound in this sequence around the two
peg A made sure the yarn did not slip or shift; pegs.To avoid knots of both yarns at the start
see Plate 79, bottom and finish of each section of one colour, the
The high degree of twist in the cord so yarn not being used, say, the black, is held
formed was preserved by putting the stick under tension while the white is carried
under the foot or having a friend hold it, round the pegs. When the white section is
while similar over-twist was given to the next completed, the black is passed right over the
two loops from B on another stick, and so on. white loops and its section is made, while the
When three or four such over-twisted cords white is held tight. These jumps of yarn,
were on sticks, held by helpers, they were all alternately black and white, are a sure sign
transferred to a single stick. By rolling this that the connected cord method was used.
between the hands in the reverse direction to This is useful in analysis if the tell-tale two
that previously used, S-twist was given to the loops at the end of the cord are worn away or
collection of cords which readily plied hidden in some decorative finish.
together making a thick stable rope. In this To make cords with two strands black, two
ingenious way most of the over-twist in the white, both yarns are tied to peg A and they
cords was secured and the stick could be are taken round the two pegs together. The
removed. two loops taken from peg B for plying on a
‘This process was repeated until all the yarn hooked stick will always be a black and a
had been plied into cords and these in turn white. At the end of the process, the loops
had been re-plied as described. Peg A was pass over A, now the starting rod, as in Fig.
taken from the ground; it still had the loops 176d, and each cord has a black and a white
of 2-ply around it and the two knotted ends. blind loop at its free end(described by Pokar
Peg B was also taken up and threaded Ram, at Doli Kalan, west of Jodhpur, 1989).
through all the loops which were previously Connected cords ending in two blind loops
on sticks. So between A and B were now thick have two advantages. They cannot fray, as
ropes of re-plied cords. would a cut end, during the long ply-splitting
Everything was dipped in water and process, and they have no bulky knot which
thoroughly wetted. Then with one peg on may be awkward to pull through. A tie
high and a weight tied to the lower peg, the connecting the two loops, as described
cords were stretched vertically and left to above, overcomes their tendency to separate.
dry. Peg B was then removed and the ropes
separated out into individual cords. They
then hung from peg A which acted as the
starting rod and ply-splitting could begin.

4. ANALYSIS referring to the upper end (attached to the
When analysing a girth, the methods starting rod) and the lower free end, and
described above can usually be identified by looking for the following details.
Separate cords
One-colour cords Upper end two loops of 2-ply yarn over starting rod. Fig. 176a.
Lower end either one loop and two cut ends knotted together,
as in Fig. 171b, or a tapered end with two shorter
ends hidden in a loop, as in Fig. 173b.
‘Two-colour cords Upper end two loops of yarn, one black, one white. Fig. 176b.
Lower end either four cut ends secured in some way or two
loops made by splicing.
Double-length cords Upper end 4-ply cord looped over rod or buckle.
Connected cord
One colour cords Upper end one loop over rod plus two yarns, one connected to
cord to right, one to cord to left. Fig. 176c.
Lower end two loops, maybe joined with fine tie.
Two colour cords Upper end two yarns (one black, one white) joined over rod to
cord to right, two yarns to cord to left. Fig. 176d.
Lower end two loops, one black, one white.
NOTE
* the 2-ply yarns passing over the starting + with connected cords there should be a
rod are usually so tightly packed that they visible knot where the first and last yarn is
cross and override each other, making it attached to the rod or where a new colour is
difficult to trace the four components of any introduced; with separate cords such knots
particular cord. are absent.
TRADITIONAL CORD PREPARATION * 173

CHAPTER THIRTEEN
Ply-Sphitting in India
1. INTRODUCTION eye near its pointed end. The selection in

Ply-Splitting in north-west India has been or Plate 80 indicates the variety encountered.
still is carried out by the following castes: the From right to left they are:
Raika or Rabari (sheep and camel herders), + a rough needle made in seconds by
the Rajput, some Suni Moslems, around Khema Ram Machra for an impromptu
Jaiselmer and Barmer, and occasionally the demonstration at Ghevra.
Vishnoi. + aneedle with a gradual taper made by
The following objects seen or collected in Rana Ram, Jaiselmer.
north-west India and surrounding areas are
+ two needles made by Ishwar Singh,
made entirely or mainly by ply-split braiding:
Jaiselmer, and by Kharna Bhima, Mamuara,
Tang a girth used to attach the saddle to the Kutch; both with a shoulder and a sudden
camel’s back, used in pairs. change to a flattened cross-section; suitable
Lhoum a decorative piece hanging from the for techniques like POT or TLOI in which
saddle only one cord is split.
Gorbandh a decorative camel necklace. * a needle made from a ram’s horn by
Pokar Ram, Doli Kalan.
Godiya and Sariya ankle and knee bands
for decorating camel’s legs. + two of the less commonly used hooks,
from Suna and Sipla, villages near Jaiselmer.
Muhra bridle for camel’s head.
Hardi enlargement on the reins near their A needle seen near Barmer had a metal
attachment to camel’s nose. working end riveted into a wooden handle.
Sometimes a needle which has snapped at its
Bheer narrow strap for tying goods to camel
weakest point, near the eye, finds a new use as
Cheekah a carrying bag. the starting rod on which the cords are fixed.
Mukhiarna anti-fly fringe tied to horse’s
head. 3. TANGS AND LHOUMS
These will now be described in detail and in ‘These are both long narrow fabrics and share
the above order. common structures and designs, so at first
sight they seem very similar. But their
2. EQUIPMENT USED IN PLY-SPLIT differing functions, a tang to cinch the
BRAIDING saddle, a lhoum to act merely as decoration,
Apart from his hands, all the maker needs is a are reflected in small details which allow four
needle, called a gunthani or khili or suiya, types to be distinguished. Fig. 177 shows
which he fashions himself. It is usually these and their relation to the saddle and the
wooden and about 20 cm long, with a large camel’s belly.
174 + THE TECHNIQUES OF PLY-SPLIT BRAIDING

A short tang — has a buckle or stick at one end, a fringe
at the other. Length 200-300 cm. Fig. 177a.
A long tang — has two buckles fixed into the structure, a
fringe at both ends. Length about 350 cm. Fig. 177b.
A short lhoum— has a buckle or stick at one end, a
fringe at other, so it is like a short tang except in its
length which is 120 to 150 cm. Fig. 177c.
A long lhoum — has a fringe at both ends, no buckle or
stick. Length 320 to 400 cm. Fig. 177d.
A short tang and lhoum are often combined to provide a
hanging piece on both sides of the saddle.
3.1 SHORT TANG

The primary function of a short tang is to hold the
heavy wood and metal saddle in position on the camel’s
back. Used in pairs the girths are attached at each side to
the horizontal bar forming the lower edge of the saddle.
As a much shorter girth would do this, it is clear that the
extra length which hangs down on one side, swinging to
the camel’s slow pace, shows it has a secondary function
as pure decoration. See Plate 8la—d.
Length — between 200 and 300 cm
Breadth — normally between 7 and 10 cm, but can be
narrower and as wide as 18 cm.
3.1a Cords used

As described in Chapter Twelve, the only materials used
are goat hair, cotton and aak. The quality of goat hair
spinning varies greatly, old tangs having very fine
smooth cords, new having thick, hairy cords. Rarely,
white goat hair is dyed and then usually bright orange or
yellow. The old tang in Plate 82, right, suggests this
dyeing may have been commoner in earlier times. Some
tangs made for sale at cattle fairs and those used for
decorating the Border Force camels are of cotton. Also
some of the most skilfully made tangs use this material.
The only tang seen made of 2-ply aak cords is a clumsy
unappealing artefact, its saving grace being its resistance
to insect attack, see Plate 82, left.
The number of cords used is between 32 and 64, 40
being a very popular number.
Figure 177 Arrangement on camel of

a short tang, b long tang, ¢ short |houm and d long Ihoum.
PLY-SPLITTING IN INDIA + 175

3.1b Starting end
At one end of a short tang, the starting end in
making, the 2-ply elements of the cords may still
pass around the wooden starting rod, either as in
Fig. 176a and 176b in Chapter Twelve if the cords
were separately made, or as in Fig. 176c and 176d if
they were continuously made.
The rod may be roughly whittled from a stick or
more carefully made with a knob at each end to
prevent the threads slipping off, see Plate 83, right,
and 87, left. Commonly the stick has been slipped out
(snapped in two if it had knobs) and the end loops of
the cords sewn to a brass or iron buckle of
characteristic shape; see Plates 82, left, 83, centre and
left, and 84, top and bottom, left. The sewing is
usually strengthened with a covering strip of leather,
often showing elaborate cut-out patterns; see Plate
83, centre, and Plates 92 and 93. Rags wrapped
around the buckle before the sewing protect the
tang’s cords from being cut by the metal. Where these
threads are eventually worn through, the whole tang
is looped through the buckle and sewn to itself. The
degree of wear on the buckle sometimes suggests it
has been fixed to a succession of tangs. Letting the
owner cut off and keep such a buckle can sometimes
be the only way of acquiring a worn-out but
technically interesting tang.
Where cut lengths of 4-ply commercial-cotton cord
are used, they will be seen lying around the starting
rod or buckle, which can therefore be present from
the start. It is also possible to use the continuous cord
method incorporating the buckle from the beginning;
the latter replacing post A in Fig. 175. The ball of 2-
ply over-twisted yarn has, naturally, to be small
enough to pass through the buckle as the yarn is
wound between buckle and post B.
3.1¢ Attachment to the saddle

The tang is attached to the saddle at some point in
its length which allows a sufficient gap between the
buckle and the saddle on the other side for a
tightening strap; see left of Fig 177a and Plate 84,
top and bottom, left. It can be passed over the saddle
bar below which a wooden or metal skewer (one is
visible in Plate 90, left) pierces its two thicknesses, as Figure 178 Attachment of tang to saddle
in Fig. 178a. Small slits produced during the tang’s a with a pin and b with a tie.

making are sometimes said to be for this purpose and
would naturally lessen damage to the fabric. A less
destructive method is to pass only a loop of the tang
over the bar and tie a cord through the loop and across
the tang, as in Fig. 178b.
‘The special shape of the buckle means that a leather
strap about 2 cm wide can exert an even pull over the
full breadth of the tang. In the absence of a buckle, it is
replaced by a tied or spliced loop of thick cord; see Plate
83, right. Occasionally the rod is removed and a loop of
cord passed through the end loops of the braided cords.
3.1d General working methods

When work starts on a short tang the cords are 7 to 9
‘arms’ long, that is about 4 metres. This unmanageable
length is shortened as follows.
Each cord is strongly over-twisted in the Z direction,
folded in two somewhere near its halfway point and its
end pushed through its own plies and brought down
again; see Fig. 179a. Then this configuration is allowed
to twist back on itself in the S direction, as in Fig. 179b.
‘The shortened cord is of course thicker at its lower end
=
SSN
AAAAASAARAAAAAARAARARN
Sy
and needs a bigger opening to pass through in the ply-
LT) SS
splitting.
A girth is often made on the move; the maker therefore
throws the cords over his left shoulder to prevent even
ANSAAARAAAA
LISSTyy
Ss
aos these shortened cords from dragging on the ground as he
S -*'
a
NN
AAARRSAARE walks about. In the ply-splitting itself the maker takes
BAN’ elaborate care to produce a close, even, texture. So after
each operation he gives a tug to the splitting cord and a
restorative Z-twist to each split cord or cords. As the
Qitarrrece
FARRAR girth grows in length, the maker rolls it around the
&S= 4
starting rod, securing the roll with a cord; see Plate 84,
top and bottom, right.
Figure 179 Shortening the cords at
start of work. 3.1e Structures used
Single-course oblique twining (SCOT)
See Chapter Seven for details of features mentioned
below and Plates 82-90.
The most commonly seen tang is all black, worked in
SCOT with longitudinal sections, often 10 in number
but there may be only 2 or up to 12. The total number of
cords in tangs of this type is between 30 and 48, 40
being a common number.
Indicative of the scarcity of white goat hair are tangs
with few white but many black cords. Plate 85 shows on
PLY-SPLITTING IN INDIA * 177

the left one with only 4 white and 32 black surface. Plate 88, right, shows smaller
cords worked in 12 sections; on the right, one inversions. Plate 88, left, shows a unique tang
with 10 white and 22 black cords worked in found in a village near Bhuj, Kutch, Gujurat,
16 sections, each only two splittings wide. worked mostly in white cotton with only 8
However a large number have an equal black goat-hair cords. The design consists
number of black and white cords, often 20 of entirely of repeated central inversions,
each, though the presence of a black border emphasised by 4 much thicker cords, another
can mean there are a few more black. Here unusual feature.
the pattern of transverse zigzags is so However, the most complex use of
prevalent that its representation on old wall inversions is seen in a group of old cotton
paintings is usually taken as evidence of ply- tangs. They all have exactly the same
splitting. See Plate 81a; also details in Plates sequence of five colours in the cords,
26, 27, 31 and 82, right. The zigzag can be suggesting a common origin, usually said to
anything from one cord wide upwards. Its be the village of Ghevra, north of Jodhpur;
popularity probably rests on the fact that they are also said to be the work of women,
once the two colours are correctly set at the which is unusual. See Plates 89 and 34, left.
start, the bold design emerges automatically; The sequence of green, 2 red, green, white, 2
also, the sections being wide, the work goes black, yellow, whether reached intuitively or
relatively quickly. The pattern can exist on by intent, has great potential for design. The
different scales and with or without a border, placing of white and yellow cords, which
as the above plates show. stand out, allow them to make independent
If the colours are not set in an ordered way, over-patterns; the green always flanking the
an all-over haphazard pattern appears. Plate reds enhances both colours; the two reds and
86 shows two such, the left-hand being two blacks can make slightly more solid areas.
unusual in that the white cords outnumber They are technically the most advanced of
the black and that a half-twist, not the any tangs worked in this structure and are a
customary quarter-twist, has been left type no longer produced.
between successive splitting of a cord. All the foregoing are worked with
The unit of design in all single-course longitudinal sections. Tangs with transverse
oblique twined braids is a rhomboid, formed sections are much less common. The
by cords of one colour appearing together in irregular edge this technique gives is usually
a section. The zigzag pattern is just a special avoided by having a single longitudinal
case where the rhomboids in adjacent section as a border. This combination can
sections are made to tally. This does not produce a texture of great subtlety,
happen with two other common ways of especially in an all-black tang.
arranging an equal number of two colours, Rarely, this structure is worked not in strict
see Plate 87, which also demonstrates how longitudinal or transverse sections of visible
narrow sections make the colours appear as S- and Z-course cords but in areas of these
very elongated rhomboids. The cotton tang in with oblique boundaries. This gives rise to
Plate 28, left, has rhomboids of five different curves both in the design and at the edges, as
colours, making a large pattern repeat. in Plate 90, left. Only two examples of this
The use of inversions, though fairly particular type were seen, both south of
common in old examples, is rare nowadays. Jodhpur, near Mandawas. Occasionally the
Plate 88, centre, shows an example with large basic all-over texture is of small diamond-
central inversions and with a half-twist left shaped sections or areas.
between splittings giving a characteristic Twined linkings are used as a single

operation to alter colour sequences, as in large single motifs to small repeated motifs.
Plate 90, centre, or repeatedly, as in the See Plate 81c, a fine old girth with dyed goat
frequently seen arrowhead pattern in Plate hair outlining the motifs, and Plate 93; also
90, right. Solid colour borders also details in Plates 48-61.
necessitate repeated twined linkings between ‘There may be a few isolated twining cords
these borders and the main part of the tang, on a plain oblique twined ground of another,
as seen in Plate 31, left, and in Plate 81a. or even the same, colour, but more usually
there are an equal number of black and white
Plain oblique twining (POT) goat-hair cords or of two colours of cotton
See Chapter Six. cords. A few of the cotton examples are
Short tangs made entirely in this structure outstanding; their actual source near
are not common; those seen are old. This Jaiselmer has not been found.
could be because, if used simply, it gives the Almost every tang of this type encountered
familiar but uninteresting arrangement of will show some hitherto unknown variation,
diamonds and triangles, with maybe some so generalisations and classification are
twined linkings to alter the scale; see Plate 91, impossible. Only three present-day makers
right. Also the skill to use it in the more have been found: Ishwar Singh’s brother,
complex ways, seen in Plates 91, left, 92 and Dunga Singh, a maker in That near Pokaran,
81b, appears to be lost. These fine designs and Kamu Kahn at Mungolai near Lawan.
are of two types: extended geometric motifs
depending on the adroit placing of twined Two-layered oblique interlacing (TLOI)
linkings and/or slits in the course of the work, See Chapter Nine.
and small often square areas of intricate Because of its simple working method and
hairline patterns. The two types often unending design possibilities, this is one of
alternate down the length of the tang. The the most popular techniques, both in the
few known examples of this type show that past and today. Such a tang usually has a
they represent a high point in skill and border which seals the two layers together
creativity, especially appealing to those and often has cross-stripes which separate
interested in structure-related design. the motifs. The latter are occasionally
Plate 11 shows two tangs, both unusually geometric, as in Plates 81d and 64-67, but
narrow, which have endlessly repeating are more often representational. Animals
hairline patterns. (camels, horses, deer, scorpions, elephants),
Though tangs made exclusively in this humans (women carrying water pots, men
structure are rare, it is common to find it on camels, acrobats, soldiers with guns) and
occurring in other tangs, almost as if the objects (temples, trees, even planes and jeeps
maker was resting for a while from the more ) are all depicted with varying degrees of skill
demanding work preceding and following it. and realism. See Plates 68, 69 and 70, right.
‘These motifs are presented formally down
Combined plain and S- and Z-course oblique the length of the tang, boxed in by the
twining borders and cross-stripes. This placing of
See Chapter Eight. non-symmetrical motifs like animals needs
Technically this is the most interesting considerable judgement, remembering that
group and it illustrates the seemingly limit- the work is done row by transverse row from
less inventiveness of the makers. ‘The group the top downwards and without any
contains geometric designs which are usually preliminary planning on paper. For example,
bilaterally symmetrical and vary in scale from the tip of a deer’s antlers must be sited

correctly for the rest of the animal to stand Plain oblique twining with longitudinal
centrally in its box. twining cords
The flexibility of the technique encourages The twining cords can be introduced in two
some makers to include their own name, ways. Either they run longitudinally at close
village and, in one case, the cost of the tang in intervals through a POT structure to give an
rupees. Plate 70, left, shows the name Bhaga all-over effect, only seen on two examples
Ram (unfortunately in reverse) and another from Bithu and Mandewas (see Fig. 142) or
old tang collected has an inscription running they are spaced out on a POT or SCOT
down its whole length. ‘Village Kalali ground moving freely or outlining areas, see
Bankarli Pargana Pali Marwar’ can be made Plates 60, left, and 62, right. Plate 61 is a
out, meaning ‘made in village Kalali and detail reproduced from the tang in Plate 81c.
Bankarli, in Paragana area of the Pali district
of Marwar’. Only rarely does a tang show 2/2 twill oblique twining
anything approaching a naturalistic scene. See Chapter Ten.
‘The two-layered structure was sometimes Again a very rare structure, but giving a
utilised to make a pocket, opening at the side; tang of great solidity and thickness. Only an
see Plate 72, left. An informant (Deepa Ram, all-black one was collected but one was seen
see List of Makers) said gold coins were in black/white goat hair, having a pattern of
hidden here and so avoided detection when a small transverse zigzags.
camel train was raided by dacoits. Ishwar
Singh graded this type of tang according to Combining ply-split braiding with darning
the maker’s ability to avoid repetition of In a few tangs the normal braiding stops and
motifs down its length, a good tang having a the technique suddenly changes to ply-split
succession of quite different motifs. darning, an outer cord becoming the splitting
‘weft’. After a few centimetres, the braiding is
For convenience the above techniques have resumed. The design of longitudinal stripes
been described separately, but very often a thus produced will vary according to how the
tang includes two or more, maybe colours are arranged at the moment of
progressing from simple to complex. So a change; compare the three areas in Plate 94.
tang may, for example, begin with a section ‘This was only seen in Kutch, where tangs
of zigzags in SCOT, then switch suddenly to made completely from ply-split darning of
the diamonds and triangles typical of POT, 2-ply cords are not uncommon.
and then to one of the many possible
combinations of those two structures. 3.1d Other features in short tangs
In a few tangs, even the transition to TLOI
Short tangs can contain other features which
was made possible by unplying solid colour
do not affect their function and so are purely
cords and re-plying them to make cords with for decoration.
two black, two white elements. This tour-de-
force is rare. But the transition can also be Slits
accomplished by linking a new warp, with Suddenly dividing the full width into several
the required type of cords, into the starting narrow braids by means of slits is quite
loops of a finished tang, the rod having been common in SCOT tangs, especially in black
removed, and then working from here in the tangs which may have no other form of
reverse direction. decoration. The slit is always sited in line with
a junction between sections so these are not
disturbed or altered. In Plate 95, right, five

slits, each preceded by a constrictor knot, inserted near the final fringe, usually in
split a 12-section black tang into six narrow transverse rows; see Plate 68, right, 69, right,
2-section braids which later recombine. A 6- and 96, right and left. Presumably put in at
section braid with two slits is a more the time of working, they show a succession
common sight. of neat loops on one side; on the other the
‘Two short slits are sometimes said to be for cut ends of each loop spread out into a thick
the insertion of the piercing stick which fixes close pile. As the latter is probably thought of
the tang to the saddle on one side. The stick as the front side, this shows that the
goes through a slit, through part of the tang interlaced tangs, on which they often
behind and then back through the other slit. appear, are meant to be seen from the side
~ This would obviously save wear. with black motifs on a white ground.
Tufts can also be inserted so that they
Braided cords appear as pile on both sides, only held in
‘The ply-splitting can stop and the cords be place by the closeness of the surrounding
divided into groups, which are each gripped structure. Cotton tufts and bobbles can be
by a constrictor knot and then worked as a merely sewn on after a tang is made.
tubular braid, usually of 8 strands. Each
braid finishes with another knot and the ply- 3.2 LONG TANGS
splitting continues. See Plate 95, left. A long tang has a fringe at both ends
equidistant from which two buckles are
Cords covered with knots incorporated in the structure. So its central
As above, the cords are divided into groups. portion passes under the camel’s belly, a strap
Then each group is completely covered with from each buckle secures the saddle and
a succession of constrictor knots, often beyond each buckle there is a decorative
alternately black and white. hanging part. It therefore both holds the
saddle in position and provides two side
Beads decorations. Perhaps for the latter reason it is
It is very easy at any moment in ply-splitting usually in a technique which lends itself to
to thread a bead on to a cord and slide it up to complex patterning, such as combined plain
the fell, where the next row fixes it into the and S- and Z-course oblique twining and
structure; see Plate 95, right. Incorporated in TLOI; see:Plate:97 None has been seen in
this way, a bead shows equally well on both simple SCOT or only in black.
sides of the tang. The position of the buckles is obviously
crucial. They are always close to 150 cms
Mirrors
apart, the total length of the tang being about
Small circular mirrors are occasionally sewn 350 cms. The buckles, maybe padded with
on to the part of the tang which hangs down, rags, are fixed in several ways.
but often only the circle of entrapping stitches
a After the tang is made, it is passed
remain.
through the buckle and sewn to itself.
Embroidery
b_ At the correct point in the making, each
Embroidery on a girth was only seen once; it cord in turn encircles the buckle completely
showed equally on both sides. and is held in place by the following rows of
ply-splitting.
Tufts
Tufts consisting of many strands of fine c At the correct point, half the cords, say
coloured cotton thread are sometimes the ones pulled through in the last row, are

passed through the buckle and half under. There are two types. One uses for
The following rows hold it tightly in position. decoration the normal two-colour patterns
worked in SCOT or POT or these structures
A long tang has to be made from separate combined; see Plate 99. The other type is
cords, around 56 in number. As there is no worked in only one colour in SCOT and
starting rod, there is no obvious point for the relies for decoration on added elements, like
ply-splitting to begin. Ishwar Singh says it is cotton tufts, mirrors, white shells or buttons;
usual to start at a buckle and work the shorter see Plate 100 for examples from Gujurat, one
distance to the fringe, keeping the unused pair made of coarse red cotton, one pair of
cords rolled up into a ball. Then work in the black goat hair.
reverse direction from the buckle to the far As such a lhoum takes no strain, is not
ends eCare: thas sso mbes takenoithatcany subject to chafing and moreover is only used
representational motifs are the right way up on special occasions, like weddings, old
on both hanging portions. Sometimes a examples, as in Plate 99, left, can still be
slight break in the texture near the tang’s found in very good condition. Their decorative
midpoint shows this was the starting point, use is emphasised by the ever-present and
from which work went in both directions. exuberant fringes seen in the plates.
‘The tang in Plate 97, left, is unusual, being
the only known instance of TLOI using three 3.4 Long lhoums
colours.See detail in Plate 74. Plate 97, right, A long lhoum can be 320 to 440 cms long; it
shows a highly patterned tang decorated with has a fringe at each end, but no attached
two-strand twining in cotton around the buckles. It passes under the camel’s belly and
familiar diamond pattern. See detail in Plate is fastened to the saddle at each side, but not
98, right. Plate 98, left, shows a motif, under tension as with a tang, its two fringed
worked in cotton tufts at the end of another decorative ends hanging down at the sides.
long tang, visible on both sides. It can have two wooden rods incorporated
It is a matter of doubt whether some long into the structure, nicely carved as they are
pieces which have two wooden rods in place permanent fixtures. They assist its attachment
of two buckles are to be regarded as to the saddle and are sited equidistant from
functional tang or decorative lhoum. Here the lhoum’s midpoint, about 150 cms apart.
they are described as the latter. ‘Two examples of another type were seen: this
was laid over the camel’s back, the hoop of the
Though lhoums are structurally similar to
saddle being accommodated in a long central
tangs, their sole function is decoration, a fact
slit, partly seen in Plate 9, right. The lhoum in
sometimes reflected in the extremely fine
Plate 102, right, may be of this type or just a
cords used in their making. In Kutch they are
clever way of saving on white goat hair.
called jaba.
3.3 Short Ilhoums Structures used
POT
A short lhoum is about 120 to 150 cms long,
with a buckle or wooden rod at one end and a Whereas the lhoum in Plate 102, right, shows
decorative fringe at the other. It is often used only the basic diamond and triangle pattern,
in conjunction with a short tang, attached to that shown in detail in Plate 7, left, has a
the saddle, so that there is a hanging repeating pattern controlled by transversely
decoration on both sides. They are placed twined linkings. The fine example in
sometimes made in pairs to hang on one side Plate 101, centre, is unusual in that, except
of the saddle. See Plates 99 and 100. for the difference in the two hairline patterns,

its design is mirror imaged exactly about its length, 440 cms, is partly accounted for by
midpoint. It is that rare type relying only on the extensive fringes which contribute at least
geometric and hairline patterns. 150 cms. It shows the darkening typical of
work stored in a smoky roof which
Plain combined with S- and Z-course oblique unfortunately converts the original black and
twining white into two shades of brown.
Plate 101, left, shows an outstanding long
Ihoum in this technique. Every motif is 3.5 Fringes on tangs and lhoums
different and perfectly worked, ranging from ‘Though some tangs and lhoums end in a set
a large single star to subtly curved repeating of unworked cords, the majority have an
patterns, some never seen elsewhere. Only a elaborately made fringe finishing with
small section is in simple POT and is bobbles, called jhoum or phuumk. This final
probably the starting point. The many exuberant flourish employs a great many
transitions from one motif to the next, thread techniques, but as very few are ply-
involving the exact placing of twined split they will only be summarised. Rarely, a
linkings, are accomplished smoothly. This is tang will have no fringe at all but be worked
a tour-de-force, the work of an inventive and to the very end of the cords, usually giving a
skilful maker. See details in Plates 55 and 57. rather untidy tapered finish.
After the last row of ply-splitting the cords
TLOI are gathered into groups, normally of eight
If figures are used in this technique, then care cords, and each group is secured with a
has to be taken that they are the right way up constrictor knot. So a 40-cord tang has five
when the lhoum is in use. The fine Ihoum in such knotted groups.
Plate 101, right, has a pleasing succession of The identical constrictor knot is found on
very simple geometric designs. A clever shift every ply-split object encountered. If tightly
in the border stripes, at the two points it tied with a firmly made cord it is practically
would be attached to the saddle bars, so that impossible to undo.
its back surface becomes the front, ensures It is made as follows. Grip the group of
that the borders appear unaltered through- cords in the left hand with the thumb
out the whole lhoum when on the camel; see uppermost. Select one of the cords for the
detail. in scentre’ of Plate» 66,. left.. This knot and take it, the working end, around
thoughtful refinement is seen fairly often. the thumb and other cords two or three
times. Each of these wrappings must be
Mixed techniques above the previous one, i.e. nearer the tang
The lhoum being entirely decorative provides itself; see Fig. i80a. Slide the thumb out
a good platform for the maker to show off his from beneath the wrapping cords and insert
skill in different techniques. The lhoum in the working cord downwards in its place, see
Plate 102, left, proves this and is a real arrow. Now tighten the wrappings as firmly
technical anthology. It has the SCOT zigzag as possible, starting with the lowest and
patterns with large and small inversions, a working upwards. This produces a length of
square of hairline patterns in POT and a excess cord above the tightened wrappings.
good length of these structures combined. In Pulling from below on the working cord —
addition, at its midpoint there is an obvious easily identified, being now shorter than the
link with a new warp of two-coloured cords, rest — eliminates this excess and finishes the
allowing the other half of the Ihoum to show a Enot.. sce, iss. Ls0b..l here: is\.a. neat
succession of motifs in TLOI. Its great alternative method of tying, demonstrated

to me by Pokar Ram in 1969 (Collingwood, 1990).
The 8 cords emerging from each constrictor knot
are worked as a four-ridge tubular braid as follows.
Having divided the cords into a right- and left-hand
group of four, take the top cord on the right, cord 8
in Fig. 181a, and pass it down through the middle of
the left-hand group, i.e. between cords 2 and 3.
Then return it to the right where it now becomes the
lowest cord on that side. Then do a similar
movement with the top cord on the left, as in Fig.
SS
Z
Gg
Z 181b. Continue repeating these two movements.
a Ti A one-colour tang can obviously only give one-
colour braids, but in most cases there will be an equal
b number (or because of a border stripe, a near equal
number) of black and white cords. These are usually
arranged at the start of the braid in one of the three
Cetey ways shown in Fig. 182, giving vertical stripes or

wide or narrow chevrons. Sometimes the worker
shows his skill by using all three types symmetrically
as in Plate 103, right, at the top.
Figure 180 a Method of tying the Such a braid made with the two-colour cords from
constrictor knot and b the result. a TLOI tang looks messy. So these cords are often
carefully unplied and then re-plied as all-black and
all-white cords, the transition being neatly hidden
within a constrictor knot.
After a certain distance, each braid has another
knot or knots below which two half-size braids
emerge. They are still 8-strand but now each strand
is two 2-ply yarns, i.e. half a cord, so the original
cords have been unplied. These smaller braids, now
10 in number, usually start and end with their own
knots. Below there may be a further subdivision,
with the emergence of 20 braids each made of eight
strands of 2-ply. With each change of size, the braid
can change its patterning. Plate 103, left, shows a
fringe made exactly as described. Occasionally there
is still one more stage with the appearance of 40
four-strand braids.
Such a complete mathematical progression is not
common but the first two stages are often seen. With
this branching out of braids as a basic plan, there are
many variations.
Figure 181 Making an 8-strand tubular

1 A row of black constrictor knots may be followed
braid. immediately by a row of white, this alternation
continuing a few times to give strong transverse

stripes contrasting with the vertical effect of the
braids themselves. See Plate 103, left.
2 Braids may be interconnected by dividing and

re-forming. If repeated, this gives a lattice held in
place by the constrictor knots, as at bottom of Plate
103, left. Sometimes the outermost braids are joined
so the whole takes on a ring-like form.
3 Braids can be completely or partly covered with

a close wrapping of fine cotton yarn in changing
colours, as in Plate 101, right.
4 ‘The constrictor knots can be covered with dense

tufts of cotton, short lengths being tied in place with
fine cords.
5 A cloth strip can be wrapped tightly round each

constrictor knot and then covered with simple
looping, each few circumferential rows in a different
colour. This makes a near spherical enlargement of
extreme solidity.
6 It is easy to link a new set of cords into the end

loops of a continuously made set. In this way the
braids of a tang can suddenly change colour from
being, say, black/white to being all black, the switch
lying hidden under a constrictor knot. See Plate 104,
left.
7 Occasionally, instead of using the tubular braids

described, the tang is finished with narrow ply-split
braids. They can be in SCOT comprising just one or
two sections, or one cord can work spirally splitting
the other cords in turn, making a tube of ply-split
darning. Plate 104, right, shows an example where
Figure 182 Various ways of using two the narrow braids are worked in TLOI, each needing
colours in an 8-strand tubular braid. only 8 cords.
3.6 Jhoums
The effect of the dividing braids is to widen the
fringe from top to bottom. The jhoums (or balls or
bobbles) which often decorate each final subdivision
of the braids contribute greatly to this, giving a
heavy spreading mass at their lower ends.
The jhoums are made in several ways, many of
which depend on adding yarn or other material.
Some ways are now described.

3.6a Covering a wrapped strip of cloth
With weft wrapping

A strip of cotton cloth about 2 cms wide is wrapped
tightly around the end of a braid to make a cylindrical
swelling. Some yarn on a needle is repeatedly passed
down between the braid and cotton and then up over
the cotton, making a series of evenly spaced vertical
stitches, as in Fig. 183. These mould the cotton into a
more spherical shape. They then act as a warp around
which goat-hair yarn can work in a weft-wrapping
technique (2/1 soumak), spiralling around the sphere
from top to bottom until it is completely covered and
tightly compressed. If one weft is used it is usually
black goat hair, giving an all-black jhoum. But if two Figure 183 Cloth encircling end of a
braid with stitches to form foundation of
or even three wefts spiral round in parallel, the effect
subsequent weft wrapping or interlacing.
is of transverse stripes.
With interlacing weft

Less commonly the weft interlaces with the warp,
made as above. In this case the latter must have an
odd number of threads. Sometimes two wefts are
used alternately, giving vertical pick-and-pick stripes.
With looping weft

The cloth can be covered with a yarn worked in
simple looping, the first row engaging with some
stitches through the cotton. This is a suitable
technique as the number of loops can easily be
increased or decreased to fit the shape of the jhoum.
See Plate 97, right.
3.6b Creating yarn branches

By pulling side branches from 2-ply yarn
If a tightly spun 2-ply yarn, such as exists fourfold in
the ply-splitting cords, has its two plies pulled out
sideways, see arrows in Fig. 184a, the loops so
formed will immediately twist back on themselves to
make two side branches of 2-ply, each ending in a
blind loop, as shown.
If the branches are long enough, they can then each
be made to produce side branches, and so on, until
there is a whole bush of short branches stemming
from one 2-ply yarn. If there are, say, six original
2-ply yarns, then the mass of branches produced is
considerable and makes an impressive bobble at the
end of the braid. See Plate 99, left.

Figure 184 a Creating yarn branches from tightly twisted 2-ply; b—c ways of making loops at end of a branch.
Its spread can be further increased by 3.6c Using extra yarn

carefully dividing each end loop into two or A popular and less structured way is to attach
more loops, as in Fig. 184b-—c, or by inserting a 2- or 3-ply goat-hair or cotton yarn to the
a tuft of yarn of some other colour into each end of the braid and so grossly overtwist it
end loop, as in Plate 100, right. Another way that it ‘snarls’ into a tight mass. Doing this
is to open the end loops and let them coil like several times with care can create a hard
a spring, as in Fig. 184c; this gives the jhoum heavy spherical bobble; see Plate 101, left.
a more solid outline.
By coiling yarn around a core -

A more carefully structured branching
FSS
=
produces regular stiff rods of yarn, as in Fig. CEE

ae)
185a. Beginning at the top end of a rod, two
2-ply elements are wrapped around
themselves as a core, extra twist being
et =~
constantly added to maintain the rod’s ve
pe Wfat=a
2)
me iy. CSEyDil
ani amr eee
rH
rigidity. As the work continues downwards, \ a
DF gy
DADsy e
the wrappings can be tightened against each a
other by pulling the core. At the bottom of the

rod, the wrapping thread becomes the core
for the next rod. In this way, a whole bunch of
rods, say 4-8, can be made from one 2-ply.
Joining the two ends of this yarn, as in Fig.
185a, makes a stalk attaching the bunch to the
end of the braid. Sometimes an extra yarn ry (
"
4 Y
ip
catches into a few of each rod’s wrappings, as . \ tM
in Fig. 185b, and is then pulled tight,
b AS
converting it into a series of knuckles of yarn.
The jhoum then has a firm compact surface,
hiding the network of extra cords within. Figure 185 a Coiling 2-ply yarn around a core to
make rods of yarn: b curving rod with an extra yarn.

4. GORBANDHS 4.1 Construction
‘Though gorbandhs (or neck bands) vary The starting loop is the one structural feature
greatly in detail, they all share some basic of gorbandhs which shows little or no
features in common. See Plates 105, 106 and variation. The correct number of separately
107; made cords are held or tied with a constrictor
‘They have a loop at the upper end which knot at their midpoint. Then the two ends of
slips over the front horn of the saddle; below a long cord, or those coming from the knot,
this there is a widening meshwork of narrow work outwards from this point spirally,
braids which in use spreads out decoratively splitting the outermost cords of the bundle
between the saddle and the base of the one at a time. A quarter-twist is left in each
camel’s neck, perhaps embellished with cord between successive splittings. This
shells, buttons, mirrors and cotton tufts. It produces a closely made sheath in ply-split
includes near its lower end a large central slit darning, tightly enclosing the rest of the
for the camel’s head to go through, and cords, see Plates 105-107. Occasionally the
always ends in an elaborate tassel which sheath is worked in 4 sections of SCOT.
hangs centrally under the camel’s neck. Other variations include a double loop.
Gorbandhs are entirely decorative and are When the sheath is long enough and its two
only worn on special occasions. Nevertheless ends are secured with constrictor knots, it is
because they employ similar cords and curved into a horseshoe-shaped loop so that
structures to those used in the functional all its constituent cords can be simultaneously
tangs, they possess great sturdiness. As a braided.
result, a large gorbandh, measuring 150 cms The initial rows are usually worked in POT
in length and loaded with shells, can weigh as it is the structure which most efficiently
up to 4 kg. This strength and limited use spreads the cords out, albeit a little untidily;
explains why many still survive, even though see Plate 106, left at top. Thereafter that
their owners no longer have camels. structure can be used throughout or it may
‘They are normally made in only one colour, change to SCOT in two or more sections or
black goat hair being the commonest material. quite frequently to 2/2 twill oblique twining.
Red cotton is often used in Gujurat, see Plate From the loop onwards, whatever technique
107; cotton gorbandhs in white, in green (Plate is used, any conformity of construction ends;
105, left) and in blue have also been seen. The every gorbandh is different. However, all the
example in Plate 106, right, is unusual in variations contribute to one end — to make a
combining three colours of cotton; however, meshwork below the loop which is as
they are not used to produce patterning, extensive and impressive as possible. This
merely narrow braids of solid colour. From always involves subdividing the broad braid
this it can be seen that ply-split braiding is not produced below the loop. Sometimes the
used here for its pattern-making potential, as resulting narrow braids separate and join
happens with tangs and lhoums, but simply as making staggered slits and it is left to the
a strong vehicle for carrying decoration. An weight of added shells and border tassels to
exception, seen at Wankaner, started with a spread these out into diamond-shaped
loop of red and yellow cotton cords which were openings when the gorbandh is in position on
used to give patterning in SCOT. The the camel.
frequently seen unembellished gorbandhs, as More commonly the narrow braids are
in Plate 105, left, and 106, left, though forced apart more positively. One method is
appealing to Western eyes, are said to lack the to divide them so that a straight longitudinal
essential decorations. braid alternates with an even narrower

oblique one. The latter joins the former to left and to
right. Between these points of attachment, the
oblique braid is worked for more rows than are the
longitudinal braids, hence it takes a curving zigzag
course, pushing the latter apart. See Plates 106, right
and left, and a more elaborate version in Fig. 186a.
A more extreme way is to work some added cords
as a transverse or oblique braid. Where it meets a
longitudinal braid its cords split the latter’s cords as
in ply-split darning. In this way, the spacing out of
the longitudinal braids can be exactly controlled.
Fig. 186b shows such a braid lying both transversely
and obliquely and illustrates how it has to curve
through a 180° at the outer border and then work its
way across in the reverse direction. Alternatively,
when it reaches the border the braid can continue
onwards, ending in a jhoum. See Plate 105, right
and left.
Added cords have already been mentioned. Many
gorbandhs have such because more cords are
needed for the main part than exist in the starting
loop. They can be added in various ways.
a_ AA transverse cord joining the two ends of the

loop has new cords fixed to it.
b_ Cords are hitched to the loop with a lark’s-head

TT
eetestene
99-0, 297008
eplestetete
Ose,
ere% x
ealegseteten:
6205
S50?
knot.
eat
a OMe R559
c Added cords for transverse or oblique braids

spring from the side of longitudinal braids, as in Fig.
Figure 186 Longitudinal braids forced 187; usually a pair is attached to successive edge
apart: a by longer curving braids, b by cords, as at the bottom of the diagram.
transverse and oblique added braids.
The outer edges of a gorbandh, in reality its lower
edge when worn, are always decorated in some way.
Braids with jhoums of different types are common
and can be the ends of transverse or oblique braids.
Another edging is seen in Fig. 188. It is based on the
principle already encountered, that of a braid forced
into a curve because it is longer than the braid it
intermittently joins. Here at its outermost point it has
a structural right-angle bend, emphasised by a
protruding loop. See Plate 106, right, for this feature,
but without the protruding loop. As Plates 105-107
show, it is usual for cords to be gradually lost to the
gorbandh in the form of edging braids. So the parts
which finally pass on either side of the camel’s neck

and meet under it are often reduced to two simple
narrow braids. These are caught together with a
decorative encircling knot, or a wrapping yarn, or
even a reversion to ply/split darning, as in the initial
loop. Whatever method is used it is followed by a
jhoum often of great size and complexity. Juaoums
on gorbandhs can have a similar construction to
those found on tangs and lhoums. Probably the
most extreme are those on the red cotton
gorbandhs from Gujurat. See Plate 107. Here the
jhoum is a flattened disc at least 12 cms across,
covered with multicolour cotton wrapping and
with an edging of added cords. The latter are
wrapped and end with the typical branches, each
with a red tuft in the final loop. Other more severe
terminal jhoums using no extra yarns are Figure 187 Attaching transverse to
longitudinal braids.
reminiscent of traditional nautical knot-work.
A ply-split gorbandh showing some of the above
features is made by Baluchi tribes near the Figure 188 Edging used on some
Afghanistan—Pakistan border; its black goat- gorbandhs.
hair foundation is almost entirely covered with
appliqued fabric, tassels and shells. (Textiles of
Baluchistan, M. G. Konieczny, British Museum,
1979)
5. GODIYA AND SARIYA, ANKLE AND KNEE

BANDS
A well-dressed camel, in addition to a gorbandh,
wears decorative bands above the knees and at the
ankles; see Plate 108. They are simply made. The
starting loop of about 8 cords is tightly wrapped,
its two ends brought together and wrapped as one
for a short distance. The emerging 16 or so cords
are then worked in SCOT in 4 narrow sections. At
the far end, all the cords are wrapped, making a
firm cylinder. This finishes 1n a knob large enough
to be a tight fit in the starting loop.
The outer surface as worn is covered with ranks
of white shells or buttons, interspersed with many-
coloured cotton tufts and occasionally mirrors.
The lower edge may have added cords wrapped in
coloured cottons.
The only examples collected were made of the
same coarse red cotton found in the lhoums and
gorbandhs from Kutch, Gujurat, so presumably
are also from this area.

6. MUHRA OR MOORKA the wider of the two braids which emerge
Bridles often show great ingenuity in their beyond. Each of these braids finishes with
making, for instance, a type encountered in a constrictor knot and then 5 free cords
Kutch, Gujurat, made from a single intricately which are twisted together into a rope. The
manoeuvred length of rope. Only three ply-split two ropes hold the bridle in place, passing
bridles were seen and collected. Two were highly under the camel’s neck and being tied to
decorated and made from several sets of cords. loop F.
But the one in Plate 109 only employs one set of It will be understood that the seemingly
ten cords and the sequence of work to produce it over-complex way that B and D are worked
_ was at first puzzling. Perhaps it was the need to in opposite directions is actually an
withstand hard wear which led to a method extremely ingenious method of making
devoid of knots, joins or added elements, except these two parts of the bridle without adding
as decoration. Analysis shows the method was new cords.
probably as follows; see Fig. 189.
The bridle use 10 black goat-hair cords of the 7. HARDI
usual 4-ply type and was worked almost Near where the reins are attached to the
entirely in POT. camel’s nose-peg, there is always a
Starting where shown, section A (the nose cylindrical thickening, the hardi. It is 10 to
band) was made, long cords being left above the 15 cm long and made in some fabric
start line. As elsewhere in Fig. 189, the arrows technique. Its purpose is to add weight and
show the direction of work and a small zigzag apparently its presence helps keep the
the start of a new section. The band then divided camel under control. Plate 110 shows the
into two. Section B, now only 5 cords wide, was only ply-slit braided one seen. It is SCOT
made and then pierced 5 of the cords above the worked cylindrically, over some _ core
start line (those which were to become D). It material, with seven transverse sections,
then joined the other 5 to make section C. using 2-ply goat-hair cords.
Section D was then made, as long as B, these
being the parts to pass under the camel’s jaw. 8. BHEER
Section D then joined up with the 5 remaining The bheer is a long all-purpose strap used
cords of A to start section E. After a short for tying goods and fodder to the camel’s
distance, 5 cords split off and then rejoined to back during long journeys. Also when tied
make the fastening loop F. Section E then to the saddle it helps the driver to mount
continued to the end. the camel without first making it kneel
Its edge was decorated with ’a zigzag made down. See Plate 111.
from 5 newly added cords. This is carefully Its finished length, from the initial loop
graded, the zigzags getting larger then smaller. to its tapered end, is about 5 metres. It is
Each has a right-angled change to its long axis, usually made of black goat-hair cords,
made as in Fig. 188, a feature also found in loop between 8 and 13 in number. These are
F. Where this border ends most of the 5 cords separately made, each being about 13
are buried in E and cut off short. metres long. The two cut ends of the 2-ply
In the complex interworking of C and E at used may be spliced together at some point
their final meeting point, the POT structure is so that both ends of each cord have two
lost and there are roughly rectangular areas blind loops, a seeming impossibility.
where about 5 cords split another 5, as in ply- These cords are joined at their midpoint,
split darning. All 20 ends of the cords are not either being worked as a square braid or as a
needed and many cut ends are seen here and in single section of oblique twining. Bending

=
SSlVR, 7S.
Seam, oe
* Figure 189 Sequence of work to produce a muhra.

&O
9
GS
LG
Ww

this braid round on itself makes the initial loop from analysis of a recently made example is
from which work proceeds with all the cords as follows. See Plate 112.
to make a braid in SCOT, two sections wide.
a Wind the cord around two sticks, fixed
Sometimes white cords are included
about 50 cms apart, beginning by slipping the
symmetrically to produce a zigzag pattern.
looped end over one stick. After an even
The method described obviously gives an
number of journeys round the sticks (26 in the
even number of working cords, said to be
bag examined) the free end will be back over
unlucky by one informant. For an odd
the stick that initially took the looped end.
number, a half-length cord is added after the
loop is made. It is attached with a constrictor b Now pass the free end through itself,
knot encircling the two ends of the loop and splitting it twice in the way used for attaching
so adds strength at this point. cords to a starting rod; see right side of Fig.
‘The tapered end comes naturally because 190a. As the diagram shows this leaves a
the cords are bound to become slightly small loop. Taking the cords one at a time off
uneven in length and each is worked to its the stick, split them similarly with the free
final loop. end. Finally pass the free end through the
9. CHEEKAH
Several types of cheekah (or bag, or pot-
holder) are found. They use the same ply-split
structures and are usually worked with the
same 4-ply goat-hair cords seen in the camel-
related objects already described and are
another example of the camel drivers’ skill.
The majority function as pot holders,
designed to support a pot or bowl, hung up
out of reach of children or animals. For this
use the open-mesh structure some exhibit and
the inevitable hole at the bottom, where the
making begins, offer no drawback. Raika
herding camels use it for a metal pot, mugra,
which can hold camel’s milk, their only food.
A much rarer type has the hole filled in and is
used for carrying flour by raika‘herding goats.
9.1 Open-mesh cheekahs

9.1a Made from one cord
This, the most frequently seen cheekah, is
suitable as a general carryall, being very
stretchable in all directions. It is ingeniously
made from one single 4-ply goat-hair cord,
which for one bag examined must have been
at least 25 metres long. As with other
separately made cords, one end has two blind
loops and the other end has a loop and a Figure 190 Two stages in making an open-mesh
knot. The method of construction deduced cheekah from one long cord.

blind loops at the cord’s end. All the cords are 9.1b Made from 3-ply cotton
now off the sticks. A similar open mesh was found on a cheekah
made from commercial 3-ply cotton cords,
c Bring the free end round and pass it
only one example being seen. See Plate 113.
through the small loop, as in Fig. 190b. Now
Sixty-four cords are initially attached to a
pull it very tight, making a ring from which
stiff ring, similar to that worn on the head
the looped cords radiate. Inevitably there is a
when carrying water pots, which is to be the
central hole. This action also tightens the
base of the cheekah.
small loop.
These cords then work in pairs, ply-
d_ As shown in the diagram, now split the splitting, as shown at the top of Fig. 191, in
free end and pull a looped cord through, one set of crossings and then in the simpler
leave a half-twist, pull the next loop through, way shown at the bottom in the next set. The
and so on. This continues spirally round the diagram shows the splittings as seen by the
base of the bag three or four times. It may worker, 1.e. with the bag upside down. The
include one revolution in which the free end splitting of the 3-ply cord is of course
splits all the looped cords instead of the eccentric; the two plies are always towards
above. Finally the plies of the free end are the outside. But this fact is not obvious as the
knotted together and the knot pushed inside. actual ply-splitting is completely hidden by
stitched-on mirrors, suggested by circles of
e ‘The bag proper is now begun with rows
dashes in Fig. 191.
of R/R twined linkings, using the two
The red, blue and _ =yellow’ cotton
elements of a looped cord as the unit, see Fig.
surrounding these mirrors, that which wraps
161. The pair of looped cords twine-linked in
the side loops of the upper type of crossing
one row are separated in the next, so
and the starting ring, as well as the elaborate
diamond-shaped openings are made; see
Plate 111. Their size depends on the length
of unworked cord left between the rows.
f Atthe top of the bag, the looped cords are

well suited to take a drawstring or carrying
cord.
Normally in ply-splitting, the small degree of

unplying involved in each splitting of a cord Ul
forces it to twist on its long axis, below the
fell, during the working process. The twisting
is usually of no consequence, but here
because the cords are joined in pairs this
adjustment to the unplying effect of the
splitting is impossible. Instead the cords
become progressively more tightly plied
towards the closed ends of the loops. This
Was
probably explains why the method of
working with loops, otherwise so convenient
with no danger of cords unplying at their
lower ends, is limited to bags with relatively Figure 191 Detail of work in cheekah made from
few splittings. 3-ply cotton.

finishing at the bag’s top, all combine to Fig. 192. Its free end is passed through that
make this a very decorative object. The loop and pulled extremely tight, densely
wrapping of the side loops is in fact a packing together the tops of the radiating
structural necessity. Without it a cord cords to form a ring. This end then works
could be completely pulled through itself, spirally, splitting each cord in turn, as in ply-
capsizing the loop. split darning. Iwo quarter-twists are usually
‘The meshwork being on a small scale and left in a cord between each penetration, as in
the top being permanently closed, there is a Fig. 192. The cord continues its spiral
side slit extending halfway down to allow course, splitting each cord 4 to 8 times. It is
access to the interior. consistently pulled tight. As a result the cords
are firmly interconnected and make a
9.1¢ Made with 8-strand braids
curved, near-solid, base. The cord end is
A bizarre bag of enormous size is included finally tucked inside or hangs outside, maybe
heres it micasures. 2 metres from top. to decorated with a bead.
bottom. The elements used are 8-strand
tubular braids, not plied cords. Twenty-four
of these white cotton braids work in pairs in a
manner similar to that used in the last
example. They always cross in a way similar
to that in the top of Fig. 191. At each
crossing, the braid is usually split with 4 of its
strands over, 4 under, the splitting braid.
The fact that sometimes a strand is itself split
shows this was the method of construction
which could be called ‘braid-split braiding’.
It rules out the alternative method in which
the braids would be made during the process.
At one point two pairs fail to cross, leaving an
opening for introducing some unknown
object.
QU:
9.2 Closely made cheekahs
S
"
( (EL
V's
Cheekahs with a close texture are made in Vy
WM) ‘
three different ways. However, they all start {isss fy ‘
in a similar fashion which is now described. iy iN ‘) NK
A set of 4-ply cords all of the same length is Ns *) « \Ny
needed. They are made as connected cords s E E =

t. = i ys $
(see Cord Preparation, page 171) suspended
Neopod
v: i FSS
pectNN
oh
|
not from a rod but from a heavy cord, which SN SSS SN SSS NN SS
ns == \\=— WV. <== NNé
was 6-ply in the only example possible to yyy B
analyse. See Fig. 192, top. There may be as "S
many as 240 such cords; in any case, it is a
number carefully calculated to suit the
designs which follow.
The heavy cord, being made from three Figure 192 Detail of start of closely-made
yarns doubled over, has a loop at one end; see cheekah.

9.2a Using SCOT more decorative bands, B and D, areas
The ply-split braiding of the main part now of visible cords on both courses make subtle
begins. So the cords from lying closely side triangles and rhomboids; see Plate 114, top.
by side have suddenly to change direction Such bands are always separated by a single
and lie alternately on an S- and Z-course; see row of R/R twined linkings, as shown. As
top of Fig. 193. only one colour, black, is used, the function
The SCOT is worked in transverse sections of the latter can only be to make a slight visual
in which either all the S-course or all the Z- break and maybe also to re-establish an equal
course cords are split and therefore visible; spacing of the cords on both courses.
see bands A, C and E in Fig. 193. In the Bands A, C and E being just transverse
sections of SCOT, the usual distortions of the
courses (away from the 45° direction in the
diagram) naturally occur. The visible cords lie
at a very flat angle and the concealed at a very
steep. As a result their constituent plies give a
surface texture of transverse ridges in A and
near vertical ones in C and E. In bands B and
D the boundaries of the shapes are curved.
After some sequence of the above bands,
the cheekah ends. The cords may then be
grouped in eights, each of which is secured
with a constrictor knot and then worked as a
tubular braid. Alternatively the SCOT
structure continues in the form of narrow 2-
section braids, as at the bottom of Fig. 193.
stacate
e XS @ After about 5 to 10 cms, its cords too become
OXOKO 8-strand tubular braids.
‘The 30 or so braids produced in this way
KKK need a terminal loop to take a drawstring or
hanging cord. The cords can be simply
carried over the drawstring and fastened to
the braid or divided into two sets which are
spliced into each other.
Though working a cylindrical form offers
great scope for two-colour patterning, only
one cheekah of this type (made from black and
white goat hair) has been seen; see Plate 114,
bottom. Maybe the inclusion of the scarcer
white goat hair is thought inappropriate for
such an object, which, unlike a tang or lhoum,
KX
SG, is solely utilitarian. However, the severe
beauty of the black structural patterning is
extremely satisfying.
The only added decoration seen is that of
sewn-on beads or rings of two- or three-
Figure 193 Analytic diagram for SCOT cheekah. colour chain stitching at the base.

9.2b ‘Using POT 9.2c Using twined linkings
Only one example of this type was seen, old Only two examples of this type were seen
and damaged; see Plate 115. It began with and collected, Plate 116 showing the finer.
only 72 finely made cords. After four circuits After the usual start, as at the top of Fig.
of ply-split darning, in which a quarter-twist 194, the darning cord is itself split by all the
was leit in each cord between splittings, a 126 cords in line for another two circuits.
new, separately made, cord was threaded on Thereafter the whole cheekah is made with
each original cord and held in place by the R/R twined linkings, all worked in the same
subsequent twine linking of the latter; see way, row by row. In each of these twined
Fig. 194. This doubling of the number of linkings, at the first rightward splitting a
cords gave a more curved outline to the start quarter-twist is left between this and the
of the cheekah and also a smaller hole at the previous splitting, at the second a whole twist
base. Then after some narrow transverse is left. The result is the pleasing close texture
sections of SCOT and twined linkings, the seen in Plate 116, dominated by the curving
main part was made in POT, but leaving a visible plies of the cords on the Z-course,
whole twist between adjacent splittings. This which yet has some give in it. In the coarser
gave a small-scale tight mesh with the same example seen, it is a quarter- or half-twist
two plies of each cord always coming which is left, exactly as in Fig. 160.
uppermost at a splitting.
9.2d Flat-bottomed type
This type of cheekah, seen in villages north
of Jaiselmer, is used by goat herds on long
journeys. Its flat base and the closing of the
central hole with a leather patch make it
suitable for carrying flour in a goat-proof
manner. See Plate 117.
Starting in the usual way, two or more
circuits of ply-split darning are worked
through the 160 cords radiating from the
central hole. Then the cords are doubled
over, either singly or in pairs, and the
\)
‘ (iy
i
i,A\ |a 2 (iy)i (\it)
\ splitting cord threaded through their end
Ss \\ ||
1]
>
7 ft
Ilo\ fi
yi
\3 tliM ‘
\) \)
lige
<9
y ‘4 ((]
fi)
? loops. There is no discernible system to the
placing of these doubled cords but after
! i)
ag
sg \) ) ry
about 2 or 3 circuits, all cords have been
treated in this way. So now the ply-split
darning splits 320 cords on each circuit. This
LDA
increase in number enables a flat disc to be
AS shi produced, the base of the cheekah; the
=y;
SesWe central hole is later closed with leather.
The sides are made with transverse
sections of SCOT, a section of visible S-
course cords alternating with a section of
visible Z-course cords. Naturally, as all cords
Figure 194 Attaching extra cords at the start of a are now in the form of loops, each splitting is
POT cheekah. of two (looped) cords by two (looped) cords,
PLY-SPLITTING IN INDIA * 197

mS
x<aes
as shown in Fig. 195a. A half-twist is always left
$0,
between adjacent splittings. Pulling on the splitting
AN
K cords when a section is completed enables the
cheekah’s shape to be accurately controlled.
Ko
After the final band, adjacent loops are linked into
oe oS.
each other as shown in Fig. 195b, an ingenious way of
WV)
XX
“s“ts FY
neutralising any difference in length between S- and
NAY Z-course cords. A carrying cord is passed through
the double loops so formed, as shown in diagram.
we sate
10. MUKHIARNA
rare
os Mote
A mukhiarna is a fringe worn over a horse’s eyes
as protection against flies. It is a well-thought-out
arte-fact, combining ply-split darning and braiding;
A\ OK the details vary little in all the examples seen. See the
selection in Plates 118 and 119.
The material used is always cotton — 2/py for the
<*
oe braiding, 4-ply for the transverse supporting cords.
The 2-ply used in old examples is the finest cord seen
in any ply-split braiding and must necessitate a
special, probably metal, needle combined with good
eyesight for its working. The structure, which is
found nowhere else, is a mesh produced by staggered
slits in POT; see Fig. 32 in Chapter Six.
The top edge of a mukhiarna consists of 2 or 3
transverse 4-ply cords, through which the braiding
cords are darned in one of two ways. In the way
shown in 196a, loops from a ball of fine 2-ply cord,
held beneath the transverse cords, are pulled up
through them. Each loop is fixed by going around the
following loop, the final loop being caught around the
transverse cords themselves as shown. The loops
below are all secured at their midpoints and then cut
to give closely spaced cords, essential for the braiding
of a fine mukhiarna.
For coarser or thicker mukhiarnas, a ball of over-
twisted yarn is needed. Loops are darned downwards
Figure 195 a Detail of SCOT pairs of through the transverse cords, as in Fig. 196b, and are
cords on a flat-bottomed cheekah; b how allowed to twist back on themselves to make a set of
loops are linked around a carrying string. 2-ply braiding cords, each ending in a blind loop.
In both methods a quarter-twist is left between
adjacent splittings of the transverse cords. The arrows
at the top show the direction of work.
The hanging cords are all braided to give the
structure shown in Fig. 32, producing a rectangular
panel below which they are darned through another

28-0r 5 tralisverse’ cords. “i nis completes the
construction of the simpler types; below hangs the
long fringe of cords each ending in a loop making it
unlikely to fray; see Plate 118, right. But in the finer,
more complex, types the braiding continues further,
forming quite extensive triangular areas; see Plate
118, left, and 119, right. The long cords emerging
from these are plied in threes; some decorative tuft
of cotton is often included in their terminal knots.
See Plate 118, left.
The braiding cords are always all of one colour
which may be different from that of the transverse
cords. Finer examples are decorated with sewn-on
beads, the grid of the braided mesh acting as a guide
to the beads’ diagonal placing; see Plate 119, right.
Cotton tufts, sequins, even mirrors may also be
sewn on.
SsSESS
COMES
iA Extra cords are often looped around the upper
transverse cords and braided upwards to make a
narrow strip, ending in a tassel, which lies between
the horse’s ears; see Plate 119, right.
In Jaipur, makers of the simpler type still exist
(1996). Two muslims, one called Imtiyaz Kahn,
produce mukhiarnas (and girths) which are sold ina
saddlery shop near the Wind Palace. They use an
iron needle, called suiya. But the whereabouts of the
highly skilled makers of the fine mukhiarnas has not
been traced, though Pakistan has been suggested.
Figure 196 Ways of darning the working

cords through three transverse cords at
start of making a mukhiarna.

CHAPTER FOURTEEN
Ply-Split Braiding in Colombia
At present the decorated crupper made by to be the underside, the needle always
the Guajiro Indians living on the Colombia passing under one of the plies of the 3-ply
-Venezuela border is the only recorded cords in the section being worked. See Plate
example of active ply-split braiding outside 120, top. So the right side has a closer
that found in the Thar desert of India and appearance, with each ply passing over two
Pakistan. Though the two instances are half a splitting cords.
world apart they share some similarities. From photographs it seems that the
Normally a crupper is just a leather strap splitting cord is put in the needle’s eye before
which lies along the backbone of a horse or the latter splits the cords in a section. The
donkey, from the root of the tail to the saddle, needle is then pushed through, dragging the
preventing the latter from sliding forwards. splitting cord behind it. So the needle always
Here it is replaced by a piece of ply-split makes a single movement in one direction,
braiding shaped so that it widens from the unlike the Indian method where a needle
saddle backwards, ending with tassels and splits cords in one direction, then receives the
fringes covering the root of the tail. The cord, then moves back with the latter in the
following description is based on the analysis opposite direction.
made by Marianne Cardale-Schrimpff in her Among the many patterns used it is
exhaustive but unfortunately unpublished interesting to find the transverse zigzag
thesis on Colombian textiles of 1972. She pattern so popular in India, as seen in
also mentions a donkey bridle made by the Plate 120, top. Others include interlocking
Guajiro in a similar way, but it is hard to diamonds involving many inversions and the
decide whether this is ply-split darning or unusual pattern in Fig. 105a, called
braiding. ‘horolava’, the crab. In all of these, new
A collection of 3-ply woollen cords is cords are graduaily added, once the braiding
assembled, correct in number for the proper begins, using the method in Fig. 92b.
proposed pattern. At their central point they This both makes possible the expanding
are joined into a flat band by ply-split shape and the introduction of new colours.
darning. When about 10 cm long, this band ‘The crupper finishes after about 30 cm of
is slid into a leather tube and the latter folded braiding. There is then a tightly worked area
over, bringing all the cords together. This of doubtful technique to which is attached a
leather loop is the means of attachment to the leather strap which passes under the
saddle. From hereon the technique is SCOT, donkey’s tail. The end of each hanging cord
worked, as was the darning, with a long is knotted to another short length and then all
wooden or bone needle with a central hole; the plies combed out to make a thick fringe.
see Plate 120, top. The surprising occurrence of ply-split
The work is done consistently from what is braiding in one isolated spot in Colombia, so

far from its plentiful presence in north-west and not a traditional Colombian product at
India, can be regarded in at least three all. Again, the fact that its existence was only
different ways. It could be a completely thrown up here by a study of extraordinary
unrelated development, similar to the thoroughness could be an argument for the
conjectured invention of weaving in several possibility of other such studies revealing its
different places in the world. It could also be existence elsewhere. For this to be a likely
argued that as donkeys and horses are relative happening, this intriguing technique has to
newcomers to South America, coming with become part of the textile researcher’s
the Spaniards some four centuries ago, this vocabulary. That it should become so is just
donkey-related textile may also be an import one of the aims of this book.
PLY-SPLIT BRAIDING IN COLOMBIAS « 201

APPENDIX
Makers of Camel Girths
This list is the result of research visits to India tangs; he grades tangs aesthetically by the
beginning in 1987. It includes makers met in maker’s ability to create new motifs down the
person and those named by people I met. They whole length, i.e. with no repeats. He knows the
were all men aged between forty and seventy. long but seldom used process of preparing
‘They are arranged in the chronological order fibre from the aak bush and has made tangs
of my meeting them. If I photographed them from this material. On other visits he has
on slide film, it is noted in brackets. shown me the work of his brother Dunga
Singh Bhati, who also uses SCOT/POT
Khema Ram Machra, seen 6 February 1987
techniques very well. (Photographs)
at village Ghevra, north of Jodhpur. Demon-
strated SCOT, on end of tang I had just Girdharilal Darzai, seen early February 1987
bought, using a gunthani he quickly made from at village Dabla, south of Jaiselmer. Sold me a
a twig on the spot. Also showed how he made beautiful TLOI tang, which he untied from a
the tapered end to a cord. (Photographs of him saddle. Brass buckle, pierced leather
working) reinforcing, animal and people motifs.
Ishwar Singh Bhati, seen many times in Rana Ram, seen with a group of other camel
Jaiselmer from 8 February 1987 onwards. drivers on outskirts of Jaiselmer in February
The man who furnished the authors Betsy 1987. Demonstrated neat way of tying off end
Quick and Judith Stein with the information for of 4-ply cord, see Fig. 171 in Chapter Twelve.
their book, Ply-Split Camel Girths of West India, Gave me a gunthani he had made; see Plate 80.
published in 1982, and for Judith Stein’s MA (Photographs)
thesis of 1976. This was during their study visit
Hanwant Ram, shepherd caste, seen 21
in 1974-5. He won a National Award for
January 1989 at his dhani, near village Pali.
Master Craftspersons in 1986. He told of a visit
Sold me short tang with longitudinal and
to Delhi to record on film the complete process
transverse sections of SCOT and ending with
for an Australian investigator; this film has
POT. With him was Amana Ram, who
never been traced.
explained cord making.
On my first visit, he showed me the complete
process of cord making from singles spinning Hukma Ram, seen 22 January 1989 at village
to final 4-ply cord. He made and gave me a Mogra, off Jodhpur to Pali road. Showed
hardwood dhera and sold me a completed TLOI girth and black gorbandh he had made.
TLOI tang and one just begun. He concen- (Photographs)
trates on this technique and experiments with
it, making small bags and short wall hangings Sona Ram, seen 22 January 1989 at village
for sale to tourists. Mukanpura, near Rohet, on Jodhpur to Pali
He calls POT structure ‘paka’ = strong, and road. Saw tang and beautifully spun balls of
SCOT structure ‘kacha’ = weak, and a twined camel-hair yarn; also black goat hair wound on
linking ‘a join’. He appreciates fineness in old a dhera. Described his cord-making technique.
(Photographs)

Bhanwar Lal, Rao caste, saw a tang by him, spinning goat hair. Described cord making;
24 January 1989, at village Bithu, south of five arms’ length of cord for tangs, seven for
Jodhpur on Jalore road. It was in SCOT, with Ihoums. Ended each cord with a taper by
longitudinal extra twining cords, as in Fig. burying the two ends, ‘to make it easier to put
142a in Chapter Eight, but in black/white, through the gunthani’s eye’. Told purpose of a
showing no coherent pattern. When we refused pocket in TLOI tangs was to hide gold coins so
tea, his wife asked, ‘Is it because I am low they would be safely hidden from dacoits
caste?’ attacking a camel train. When photographed
said, ‘For once in my life I will be in a
Mohamed Amin and Umar Din, seen 25
photograph.’ (Photographs)
January 1989 at village That, near Pokaran.
Both are tang makers; demonstrated the Pokar Ram, Vishnoi caste, seen 8 February
methods of making both separated cords and 1989 at village Doli Kalan, off Jodhpur to
continuous cords. Was sold a newly made and Balotra road. Sold me almost completed TLOI
excellent POT/SCOT tang but not sure who tang made of goat hair and white cotton; also a
made it. Leather-reinforced buckle, good gunthani made from a deer horn. He worked
fringe rows left to right but was right-handed.
Described continuous cord making. Showed
Kamu Khan, seen 28 January 1989 at his
ingenious way of tying the ubiquitous
dhani, at Mungolai, near village Lawan,
constrictor knot, with obvious disbelief that I
Pokaran, on Jodhpur to Jaiselmer road. Active
did not know such a simple operation.
tang maker who was making a SCOT tang with
(Photographs)
zigzag stripes when visited; very willing to talk
about his work. He could also make bheer 8 Modu Ramji, Dewasi caste, seen 8 February
cords wide and simple open-mesh cheekah. 1989 at village Kakrala, off Jodhpur to Balotra
Showed several other tangs including a long road. Showed two fine TLOI tangs, with many
POT/SCOT one with inversions. Described animals and people as small motifs running
his method of producing cords nine ‘arms’ like a strip cartoon down their length, one with
long. After wetting the cord, he stretched it two names written on it; also an all-white
from his house to a distant tree. Showed balls of gorbandh made of cotton so insects would not
finely spun black and white goat hair; he eat it; and an open-mesh cheekah and a black
appreciated what he was currently using was gorbandh. He is teaching younger men the
much coarser. craft; saw one coarse piece made by a pupil.
Visited his dhani again on 19 December (Photographs)
1995 and was instantly recognised by his wife,
Takim Singhji. His grandson, Karan Singh
who got out the photographs we had sent him.
Parihar, seen 10 February 1989, sold me his
But he was away with his goats till sundown, so
grandfather’s beautiful old long lhoum, with
we could not see his recent tangs which were
fine elaborate fringes, in excellent condition
kept locked in a chest. (Photographs)
though the pattern obscured by a general
Tulsa Ram, seen 1 February 1989 at village darkening by smoke. Technique changes from
Dhivera, on Bikaner to Ganganager road; sold SCOT and POT to TLOI near the centre by
me a finely made POT/SCOT tang, the work linking 2-colour cords on to starting loops.
of his father, who was a little surprised when he Said, laughing, ‘You have worked some magic
found out. on me. Many people wanted to buy this and I
did not mean to sell’
Deepa Ram, Vishnoi caste, Bolan subcaste,
seen 3 February 1989 at village Khabra Kala, Bheru Singh, son of Narphat Singh, seen 26
between Cherai and Ossiyan, north of Jodhpur. October 1990 at village Narnadi, off Jodhpur
Had been a maker of tangs and gorbandhs; still to Barmer road. Bought tang in POT/SCOT
MAKERS OF CAMEL GIRTHS * 203

(its extra width said to be for the comfort of the Singh Tanwar On 2 February 1994 visited his
camel) and one beginning with full-width village, Kelawa, off road to Ossiyan. He was
inversion, half-twists left between every away (and again on a visit in 1996) but was
splitting. Saw another with inversions, and a shown his box of tools dragged to the guest
good black gorbandh with extra cords lark’s- house by two of his relatives. Inside were
headed on to starting loop. (Photographs) spindles, dheras, buckles, twisting stick. After
some argument over its price, I was presented
Jala Ram Bidar, Vishnoi caste, working in
with a beautiful hardwood dhera as a gift. As
Doli region. Not seen but reported on 29
well as making tangs, he had a collection of
October 1990 as still being active.
‘about thirty’, but as so often happens they were
Kama Ram, seen 2 November 1990 at village locked away in his absence and I never saw
Khandi, west of Rohet. Shown a cotton and them. Examples of his artistry were around in
two goat-hair SCOT tangs. Gave a detailed the wonderfully patterned charpoy stringing
account of continuous cord making, using a and in the fading floor paintings. (Photographs)
stick to twist the 4-ply cords as lifted from the
Gordhan Singh, seen 22 December 1995 at
stake in ground. Sold me beautifully wound
village Kanod, north of Jaiselmer to Pokaran
ball of singles goat hair. Used a gunthani which
road. Showed me old wreck of a TLOI tang in
he called a ‘khili; made of bamboo.
red and yellow cotton he had made.
(Photographs of him spinning)
Gunga Singh Bhati, Rajput caste, from
Bahadru Ram, seen 3 November 1990 at
village Damodra, between Jaiselmer and Sam.
village Buckhala, on Jodhpur to Jaipur road.
On 22 December 1995 at Hamira bought a
Sold me a black SCOT tang with transverse
beautiful bold TLOI made by this man, now
sections made by him. (Photographs)
dead. (Photographs)
Bhaga and Labu Ram, makers of two tangs
Rewat Ramji, seen 23 December 1995 at
seen 7 November 1990 at tea stall at village
dhani in Jaiselmer area. Sold me a POT tang
Bachharau, on Barmer to Mount Abu road.
with hairline patterns and outlining cords.
Both had maker’s name written in TLOI
(Photograph)
technique. Bought by self and friend.
Shera Ram Godara, Vishnoi caste. On 26
Kharna Bhima, seen 8 January 1994 at village
December 1995 visited his dhani in Lunewas
Mamuara, near Bhuj, Kutch, Gujurat. Sold
area. He is active tang maker but was away. I
me a pair of short lhoums in TLOI made of
was shown SCOT tang of his making; also
orange-dyed and black goat hair, very coarse.
strange little keyrings; given one by his son.
Knew meanings of motifs. Said he could make
(Photographs)
one in fifteen days if did nothing else. Made
cords eight feet long (four times distance from Getah Ram and Aatam Ram, active tang
elbow to finger tips). Knew about SCOT; tied makers from village Sam, but usually found at
the constrictor knot by wrapping cord round Sam Sands in evening when tourists take camel
his thumb plus the cords and tucking final end rides; last seen 21 December 1995. Skilful;
through where his thumb had been. Called a Aatam, a fine spinner, sold me a_ half-
Ihoum a ‘jaba’ and the actual end tassels completed SCOT tang, see Plate 81, left.
“‘phuumk’. (Photographs) (Photographs)
Deva Ram On 2 February 1994, at dhani near Jumi Imtiyaz Kahn, Muslim maker of cotton
Netra, west of Jodhpur to Nagaur road, bought tangs and mukhiarna, sold at Amar Seat Works,
a black bheer from this man’s son, who said his Hawa Mahal Bazaar, Jaipur. Said by
father made it. shopowner to use a metal needle called a suiya.

Glossary
Braid. A fabric, usually longer than broad, Fell. The growing edge of a braid where the
through which elements pass obliquely, taking ply-splitting takes place, adding to the braid’s
a zigzag path if a flat braid, a helical path if a length row by row. It is transverse in POT and
tubular braid. Where elements on opposite TLOI, and oblique in SCOT. An oblique fell
courses cross, they can be interworked in can either be an S-fell or a Z-fell, depending
several ways, e.g. interlacing with or wrapping on its angle. One of each can form a V-fell or an
about each other; but in ply-split braids one A-fell, two of each a W-fell or an M-fell, more
splits the other. Note a braid is a more specific than this a zigzag fell.
definition than a band, which, though often of
Interlacing. The over/under passage of one
similar shape, can be worked in any technique
element through others set at right angles.
with a warp/weft orientation of elements.
Inversion. The mirror-imaging of a structure
Concealed. Applied to a cord which is hidden
or a motif about a trdnsverse axis.
from view by the plies of the cord it is splitting
at that point. See Visible. Lhoum. Decorative braid attached to camel
saddle. See Chapter Thirteen.
Cord. The plied element which is the working
unit in ply-splitting. Can have any number of Longitudinal. In the direction of the length of
plies, odd or even, though four is the most the braid being made, so parallel to its selvage.
common.
Needle. General term applied for convenience
Course. An abstract expression indicating the to whatever device is used for splitting a cord
oblique path on which parallel elements lie as and pulling another through.
they pass through a braid, crossing the
Plain Oblique Twining. The structure in
elements on the opposite course. Can be
which each cord, as it moves on its oblique
specified as either an S-course or a Z-course.
course, alternately splits (so is concealed) and
Course Colour Sequence. The repeat of the is split (so is visible) at each successive crossing
colour sequence of the cords as they lie on the point with cords on the opposite course.
S- or Z-course; identical to the sequence read Abbreviated to POT. See Chapter Six.
at either selvage. Abbreviated to CCS.
Repeat. The amount of braid worked until the
Crossing. The point where two cords on cords return to their starting position; almost
opposite courses meet and one splits the other. always consists of several cycles.
Row. In POT and TLOI the ply-splitting of

Cycle. A sequence of manoeuvres continued
cords in pairs, worked transversely from
until the first is again repeated.
selvage to selvage. Long and short rows
Dark and light. The two yarns used in the alternate, the two making a cycle. In SCOT the
description of most techniques. So a braid can pulling of one cord through many on an
be said to start with 10 dark, 20 light, 10 dark oblique fell.
cords, or each cord to have 2 dark, 2 light plies.
GLOSSARY + 205
Single-Course Oblique Twining. The structure Twined linking. After one cord splits another
in which each cord, as it moves on its oblique the same two cords are involved in another
course, splits (so is concealed) or is split (so is splitting, both splittings being in the same row.
visible) at more than one successive crossing If both splittings are Rightward, this is an R/R
point with cords on the opposite course. twined linking. If a Rightward is followed by a
Abbreviated to SCOT. See Chapter Seven. Leftward, this is an R/L twined linking. Both
types have the same effect of making the two
Section. A longitudinal or transverse portion
cords swap courses. See Chapter Eleven.
of a braid where all cords on one course are
visible (and on the other course, concealed), Twining. The structure in which two or more
this state being reversed as they pass to the elements twist around each other and in so
adjacent section and at the selvages. Only doing enclose a third element, usually running
found in SCOT, where longitudinal sections at right angles to them. Twined cords can lie at
are far commoner than transverse. an angle to the selvage in oblique twining or
parallel to it in warp twining or at right angles
Starting Colour Sequence. The colour
to it in weft twining.
sequence of the cords at the beginning of a
braid, i.e. on the starting rod. Identical to the Two-Layered Oblique Interlacing. The
CCS in SCOT, but not in POT. Abbreviated to structure in which two separate obliquely
SCS. interlaced braids, usually of different colour, lie
one above the other, the two layers being
Tang. Literally, a belt, but in this context a
produced simultaneously by ply-splitting two-
shortened form of unt ka tang, a camel girth.
colour cords. Design comes from the
See Chapter Thirteen.
interchange of elements between the two
Transverse. Straight across the braid, so at layers. Abbreviated to TLOI. See Chapter
right angles to its selvage. Nine.
2/2 twill oblique twining. The structure in Visible. Applied to a cord which is split and so
which each cord alternately splits (so is can be seen from both sides of the braid at that
concealed) for two crossings and then is split point. See Concealed.
(so is visible) for the next two crossings with
cords on the opposite course, its path being set
over by one cord from its neighbours. See
Chapter Ten.

Bibliography
Peter Collingwood, The Maker’s Hand, 1987/ Betsy Quick, ‘The Camel Belts of North-
98, Lark Books and Interweave Press, USA, Western India: A Socio-cultural Study’,
also called Textile and Weaving Structures in unpublished MA thesis for University of
Batsford edition, UK. Ten pages about ply-split California, Berkeley. (Not seen)
braiding with monochrome and colour photos
Betsy Quick and Judith A. Stein, Ply-Split
of Indian examples.
Camel Girths from West India, 1982. Pamphlet
Peter Collingwood, Rug Weaving Techniques, Series Vol. 1, Number 7, Fowler Museum,
Beyond the Basics, 1990, Interweave Press, University of California, Los Angeles.
USA. Illustrated description of Indian way of Excellent well-illustrated description based on
tying a constrictor knot. the authors’ theses. Describes the three main
structures of ply-split braiding. Good
Peter Collingwood, ‘Ply-Split Braiding’,
bibliography and details of girths brought back
article in Weavers 29, Fall 1995, XRX Inc.,
for the museum collection.
Sioux Falls, SD, 57104, USA. Introduces the
subject and describes POT in detail. Marianne Vere Cardale-Schrimpff, Tech-
niques of Hand Weaving and Allied Crafts in
Peter Collingwood, ‘Ply-Split Braiding, Part
Colombia, 1972. Regrettably unpublished
IV’, article in Weavers 32, Summer 1996.
doctoral dissertation for Oxford University.
Detailed instructions for TLOI. Published as
Very detailed descriptions of all types of thread
above.
work produced by nine Indian tribes in
Virginia Isham Harvey, Split-ply Twining, Colombia, including the Guajiro who employ
Threads in Action, Monograph 1, 1976, HTH SCOT; many illustrations.
Publishers, Santa Ana, Ca, 92705, USA.
Noemi Speiser, The Manual of Braiding, first
Description of SCOT with illustrations of
edition 1983; third 1991. Self-published by
freeform interpretations.
author at Ziefenerstrasse 25, CH 4424,
Virginia Isham Harvey, ‘Split-ply Twining, Arboldswil, Switzerland. The most complete
An Update’, article in Interweave White Paper, overview of the subject, including succinct
1981. Interweave Press, Loveland, CO, 80537, account of ply-splitting.
USA. Describes some bridles, a Tunisian
Judith Allyn Stein, ‘Camel Belts of Western
muzzle and a Guatemalan back-strap made by
India, Technical Analysis and Classification’,
ply-split darning.
1976, unpublished MA thesis for University of
M. G. Konieczny, The Textiles of Baluchistan, California, Berkeley. Detailed description of
1979, Museum of Mankind, London. girth making, based on information given by
Photographs of ply-split gorbandhs made by Ishwar Singh to the author and Betsy Quick.
Baluchi tribes on Afghanistan—Pakistan border.
BIBLIOGRAPHY + 207
List of Suppliers
Fred Aldous, PO Box 135, 37 Lever Street, Ann Norman, Sagaman, Aston Road,
Manchester 1, M60 1UX, UK. Tel 0161 236 Bampton, OX18 2AL, UK. Tei/fax 01993-
2477 and 4224; fax 0161 236 6075; e-mail 850823; e-mail address <sagaman @
address <101547.1355 @ Compuserve. com> compuserve.com> A professional ropemaker
Supplier of latchet hooks in wooden handles willing to assist with the making of, or actually
and the ‘rag rugger’, a gadget designed for rag- to provide, cords of any size suitable for ply-
rug making but well-suited to ply-splitting. Its splitting. Sells the ‘Gripfid’, an ingenious tool
pointed end makes splitting easy and its spring- originally designed by Stuart Grainger for rope
loaded pincer grip holds a cord firmly as it is splicing but excellent for ply-splitting heavy
pulled through cords. Made of brass with a Turk’s-head knot as
handle, it comes in two sizes suitable for cords
Bradshaw Machinery, PO Box 102, Deer
up to 10mm and 16mm in diameter.
Isle, Maine, 04627, USA. Tel/fax (207) 348-
Has measured plans for the making of the
2921. Maker and supplier of a sturdy four-
Tibetan twister and occasionally a limited
hook twister, designed for attachment to a
supply of these devices.
power tool; can be made with 6 hooks.
Leanda, Unit 1/W2, Scott’s Yard, Ber Street,
Anna Crutchley, Frater Studio, 6b Priory
Norwich, NR1 3HA, UK. Tel. 01603-763340
Road, Cambridge, CBS 8HT, UK. Tel 01223-
(workshop); tel/fax 01603-434707 (admin-
327685. Supplier of a neat four-hook twister,
istration); e-mail address <leanda @ netcom.
called a ‘cord spinner’. A finger or thumb is
co.uk> Agent for an excellent four-hook twister
used as the brake in the initial over-twisting
made as integral part of a hand drill
stage.
Linda Hendrickson, 140 SE 39th Avenue, Note The ingenious principle on which all the
Portland OR, 97214, USA. Tel (503) 239- above twisters are based was the original
5016; e-mail address <lindahendrickson @ invention of Ralph Nellist, who made his first
cnnw.net> Supplier of ply-splitting equip- (wooden) model in 1989 but is no longer
ment, including kits with ready-made cords. producing them.
Homecrafts Direct, PO Box 38, Leicester,

LEP9BU; UK. Tel.0116 2513139; fax0116
251 5015; e-mail <post @ speccrafts. demon.
co.uk> Supplier of the same two gadgets as
Aldous, see above.

q
<7
g WN
BIE
SUED
Left, ply-split darned strap from Manubhai, dealer, Anmedabad, November 1990; black/white goat hair.
Centre, Anatolian tablet weave from London dealer; black/white goat hair.
Right, ply-split darned tubular belt, gift from Errol Pires, NID, Anmedabad; dyed commercial cotton. See text page 16.
Cord-making device for constructing 4-ply cords. See text page 25.
w Selection of ‘needles’ for ply-splitting, as described in text, p 32.
POT, basic diamond and triangle pattern. Left, tang acquired from camel driver, Sam Sands, January 1994. See text page
44. Right, |houm collected by John Gillow, Cambridge, 1990s.
Below left: Camel with two
tangs showing zigzag pattern,
Sam Sands, November 1990.
Below right: Wall painting
showing two tangs with typical
zigzag pattern, Red Fort,
Jhodpur, January 1994. See
text page 19.
5 POT, twined linkings at border. Lhoum acquired from Manubhai, dealer, Anmedabad, 1989. See text page 46.
6 POT. Left, transversely placed twined linkings. Lhoum acquired from Manubhai, dealer, Anmedabad, 1989. See text page
49. Right, longitudinally placed twined linkings. Author’s sample, wool. See text pages 44 and 49.
jet
PODLEE®
O04?
SL SSqesee
= ~s PtCA&bIhae
poo Seeeene & 99D ERD
4 se85 SHO
- *
7 POT, transversely placed twined linkings. See text page 49.
Left, tang acquired from UK dealer. Right, |houm acquired from Manubhai, dealer, Ahmedabad, 1989.
8 POT, transversely placed twined linkings, exactly as in Fig 30. Author’s sample, wool. See text page 50.
9
9 POT. Left, open structure produced by staggered slits. Author’s sample, cotton. See text page 52.
Right, |ong central slit. Lhoum collected by Girdhar Ram, Jaiselmer, 1996. See text page 51.
10 POT. Left, slits combined with twined linkings to alter colour sequence in cords. Right, slits used to achieve same effect.
Both on tang acquired at Dabla, January 1994. See text page 52.
11 POT, oblique hairline stripes. Left, tang acquired at Dabla, January 1994. Right, tang acquired at Shiv, 1987. See text
paaqes 53 and 54.
10
11
LS ie
Agere ag
dd
oS
aeedsdada?
eeeoeet m
betonse SAANa
12
12 POT, oblique hairline patterns. Left, tang from Dabla, 28 January 1994.
Right, |houm acquired from Manubhai, dealer, Anmedabad, 1989. See text pages 54 and 56.
13 POT, transversely placed twice-split cords to control hairline stripes.
Left. tang acauired at Dabla. Januarv 1994. Right. tana acquired at Bouha. north of Jaiselmer 1994 See text nane 5&7
Above: Tang with arrowhead pattern on camel belonging to the
recorders of Rao cast, near Jetpur, January 1989.
Right: POT and TLOI tangs on camel at desert fair, Jaiselmer,
1987.
POT, transversely and longitudinally placed twice-split cords.

Left, author’s sample, wool. Right, tang acquired from Sabu Kahn, west of Devicot, January 1994. See text page 58.
15
POT, multicoloured cords. Left, combined with solid-colour cords giving diamond pattern.
Right, giving transverse and longitudinal stripes. Author's samples, wool. See text pages 62-3.
POT, worked cylindrically to give a bottle and jacket shape. Made by author 1985, using 2-ply Indian cord.
See text page 67.
17
it/ POT, top, pointed start of braid by adding new cords laterally; ower section, combining with
oblique twining in different ways. Author’s sample, wool. See text page 68.
18 POT, combining light Z- with dark S-plied cords to give striped central diamonds; scale changed
in lower half by four twined linkings. Author’s sample, cotton. See text pages 69 and 70.
= [oe]
Gina
SUR
\
im : 3 Z Cee
|
SS!
SX
20
POT. Left, plain triaxial twining. Author's sample, 2-ply Indian cord. Right, combining areas of 2/2 and 1/3 splittings.
Author’s sample, wool. See text page 70.
SCOT, single-section braids. Left, cords changing function along oblique fell, using one colour. Centre, the same,
using three colours. Right, rhomboids of colour. Author’s samples, wool. See text page 77-8.
21
21 SCOT, two longitudinal sections.

Left, ridges at centre showing as colour stripes by using cords with dark/light/dark/light plies. See text page 83.
Centre, one-colour braid to show longitudinal and oblique surface ridges. See text page 82.
Right, one dark cord to clarify structure. See text page 80. Author’s samples, wool.
22 SCOT. Left, 2-section braid using S- and Z-plied cords with dark/light/dark/light plies.
Right, 3-section braid using cords with 2 dark/2 light plies. Author's samples, wool. See text page 83.
23 SCOT, 2-section braids. Left, S- and Z-plied cords each with the same four differently coloured plies. Right, using S- ar
Z-plied cords with dark/light/dark/light plies, introducing a third colour. Author’s samples, wool. See text page 83.
N2
23
a
eZ
A
CK ss ‘ G & s x ONT.
CU]
a
24
SCOT, 2-section braid, central diamonds produced by constantly changing between A- and V-fells.
Author's sample, wool. See text page 85.
SCOT. Left, 14-section braid, with cord colours symmetrically arranged. Tang acquired from John Gillow, 1980s.
See text page 91.
Centre, 12-section braid, half cords dark, half light. Tang acquired at Gagani, January 1989. See text page 91.
Right, 6-section braid, with similar cords. Tang acquired at Kharabera, January 1994. See text page 91.
SCOT, zigzag pattern. Left, on a 16-braid tang made of black goat hair and white cotton. Acquired from its maker,
Bheru Singh, at Narnadi, October 1990. See text page 94.
Right, on a 4-section braid. Tang acquired by John Gillow. See text page 92.
28
SCOT, zigzag pattern. See text page 94—5. Left, on a 6-section cotton braid using five colours. One of a pair, acquired
at Roopniwas, Nawalgarh, January 1996. Right, on a 6-section cotton braid using five colours in a more complex
sequence. Acquired from Thakur Rao Sher Singhi, at Kalyanpur, Rajasthan, October 1990.
SCOT, rhomboids not giving zigzags. See text page 95. Left, 4-section braid in five colours. Cotton tang acquired from
Lalji, dealer, in Jnodpur, 1987. Right, 4-section braid in black and white. Cotton tang bought at Amar Seat Works, Hawa
Mahal Bazaar, Jaipur, made by Jumi Imtiyaz Kahn, 1996.
29 SCOT, cords changing function within a section. See
text page 96. Left, 2-section braid with split cords
forming stripes and spots. Right, 4-section braid with
single cords being split instead of splitting in
alternate sections. Both author’s samples, wool. Left: SCOT tangs on camel at desert fair, Jaiselmer, 1987.
30 SCOT, changing number of sections from 8 to 2 and Right: Red- and white-cotton tang, SCOT and POT, on
then back to 8. Author’s sample. See text page 97. camel at desert fair, Jaiselmer, 1987.
30
31a SCOT, one-colour borders. See text page 99-100.
Left, using twined linkings between black and white
cords. Tang acquired from Rathi, Barmer, November
1990. Right, using slits between black and white
cords. Tang collected in India, early 1980s. Left: SCOT tang attached to unused saddle with wooden
32 SCOT, twined linkings altering CCS; see Fig 85. pin at Godewas Vishnoi village, February 1990.
Author’s sample, wool. See text page 100-1. Right: Gorbandh on camel at desert fair, Jaiselmer, 1987.
33 SCOT, obliquely placed twined linkings. See text page 101. Left, 2-section braid, with symmetrically place twined
linkings giving arrowhead design. Centre, as left, but using three colours. Right, single-section braid with twined linkings
going across width. Author’s samples, wool.
34 SCOT, inversions. See text page 102-3. Left, 6-section braid, with inversions giving narrow concentric ovals. Tang from
Kalyanpur, for details see caption to Plate 27. Right, 2-section braid with repeated inversions, giving ovals. Author’s
sample, wool.
35 SCOT, inversions. Left, altering the number of sections by only inverting some cords. See text page 104.
Right, 2-section braid, effect of dark/light/dark/light plied cords on a single inversion. See text page 102.
Author’s samples, wool.
36 SCOT, inversions. See text page 105. Right and /eft, braids based on designs found in Peruvian fabrics worked with
frae hanaing ends (Sneiser 1982). Author’s samples woo!l.
COmaCD)
© ow
o KR
39
SCOT, inversions; 2-section braid with large central

inversions, based on tang acquired in Kutch;
see Plate 88, left. Author's sample, wool. See text
page 105.
38 SCOT, transverse sections. See text page 111-14.
Left, cotton braid with transition to longitudinal
sections.
Right, wool braid with dark/light/dark/light plied
cords, giving sections with transverse and oblique
stripes. Author’s samples.
9 SCOT. See text page 112.
Top, transverse sections producing zigzags running
down length instead of across braid.
Bottom, addition of two longitudinal sections to
straighten selvages. Author’s sample, wool.
SCOT, cords changing function along oblique lines. See text page 115.
Right, giving diamond. Left, model to show curving of cords at
diamond’s boundaries. Author’s samples, cotton.
Al SCOT, cords changing function along oblique lines. See text page 117-18.
Right, cotton 2-section braid, with cords changing function on the Z
diagonal. Left, similar but using two colours of wool. Author’s samples.
42 SCOT, cords changing function along oblique lines. See text page 118. ;
Left, 2-section braid based on tang seen at Mandawas; see similar in Left: Muhra, in Plate 109, on camel in
Plate 90, left. Right, 2-section braid based on design on Colombian Lahti, January 1989.
ply-split crupper. Author's samples, wool. Right: Mukhiarna on horse at Samode.
y
44
43 SCOT, changing iong axis of braid. See text page 119. Top, obtuse, acute and right-angled change to
long axis of single-section cotton braid. Bottom, serpentine shape in wool produced by alternating two
acute with two obtuse angles. Author’s samples.
14 SCOT, working cylindrically; long tube showing transverse and oblique sections and obliquely set
areas. Author’s sample, wool. See text pages 123-4.
SCOT, eccentric splitting of cords. See text page 125. Left, diamonds
with cords split 1/3 and 3/1 on both courses. Right, longitudinal and
transverse sections, and obliquely set areas of cords split 1/3 on
Above: Deepa Ram spinning goat hair Z-course and 3/1 on S-course. Author’s samples, cotton.
at Khabra Kala, February 1989. 46 SCOT, variations. See text page 126. Left, inversions, incorporating four
Right: Khema Ram Machra making very thick cotton cords. Right, 3-section wool braid with two dark/two
1eedle for impromptu demonstration, light plied cords leaving half-twist between successive splittings, with
ahevra, 1987. added black cord to show course of a split cord. Author’s samples.
.
ZT\
<
\<
Z
« &W
SS
48
Combined plain with S- and Z-course oblique twining; one-colour twined diamonds. See text page 130.
Left, giving asymmetrical pattern as in Fig 117e. Author’s samples, wool. Right; giving symmetrical pattern as in Fig 117d.
Combined plain with S- and Z-course oblique twining; twined diamonds divided transversely. See text page 131.
Left, top, pattern as in Fig 121c; bottom as in Fig 121d. Acquired from dealer in Barmer, November 1990.
Right, triangle shapes as in Fig 121b. Acquired at Dabla, February 1997.
Y
Combined plain with S- and Z-course oblique
twining; twined diamonds divided transversely.
See text page 132.
Right, shape found in centre of Fig 121c combined
with outlining cords. Tang acquired at Suna, January
1989.
Left, similar with spots two cords wide in the half
diamonds. Tang acquired at Dabla, January 1994.
0 Combined plain with S- and Z-course oblique
twining; twined diamonds further subdivided.
See text page 133.
Right, into 4 small diamonds. Tang from dealer at
Barmer, November 1990.
Left, top and bottom, into 8 triangular areas. Tang
from Joss Graham, London, 1987.
Plain combined with S- and Z-course oblique twining;
centre, twined diamond divided into 9 small
diamonds; top and bottom, divided into 4 triangles.
Tang acquired at Suna, January 1989.
See text pages 133 and 135. 51
Plain combined with S- and Z-course oblique twining. See text page 135.
Left, popular 4-triangle division of twined diamond. Fine old Ihoum from the
Left: /shwar Singh Bhatti making TLOI collection of Marianne Straub.
tang in his house, Jaiselmer, 1987. Right, same motif on smaller scale. Lhoum acquired from Sam, 1995.
Right: /shwar Singh and Mr Tak 53 Plain combined with S- and Z-course oblique twining; twined diamonds
discussing an old cotton gorbandh, subdivided to give four different large motifs. Tang collected by Girdhar
December 1995. Ram at Bouha, January 1994. See text page 136.
53
OS Ge
54 Plain combined with S- and Z-course oblique twining; twined diamonds
divided in different ways. See text page 136. Left, to make large-scale lattice
pattern with colour reversal. Tang collected by Girdhar Ram, December 1995.
Right, to give concentric diamonds. Tang acquired from Ishwar Singh,
January 1994, made by Dunga, his brother.
99 Plain combined with S- and Z-course oblique twining; four diamonds divided
so they combine with POT diamond to make 8-point star. Lhoum acquired at
Sam, 1995. See text page 136.
36 Plain combined with S- and Z-course oblique twining. See text pages 137-8. Left: Mr Tak thinking at
Left, using R/L twined linkings to give longitudinal stripes. Old tang acquired Abhamaniyu Rathi’s Embroidery
at Kanod, December 1995. House, Jaiselmer, January 1984.
Right, using R/R twined linkings to divide the diamonds transversely, giving Right: Mr Tak drinking at Shiv,
two different patterns. Tang acquired at Dholera, January 1994. February 1989.
‘\s
SAN
» > e 2 ‘.
“aw 4 m
a &oc
*
59
Plain combined with S- and Z-course oblique twining. Left, using R/L twined linkings combined with curving twined
cords. See text page 141. Lhoum acquired at Sam, 1995. Right, using R/L twined linkings. See text page 135.
Left, using R/L twined linkings to invert a design mid-diamond. Tang acquired from Joss Graham, London, 1987.
Right, pattern with curving twined cords as in Fig 136. Tang acquired at Dabla, January 1994.
Left, to show central white diamonds worked alternately in POT and SCOT. Tang acquired from Lalji, dealer, in Jodhpur,
January 1994. Right, only upper part of each diamond worked in oblique twining. Tang acquired from Tulsa Ram, the
maker’s son, at Dhivera, February 1989.
Plain combined with S- and Z-course oblique twining. Left, outlining diamonds with cords of two colours. Tang acquired
at Dabla, February 1987. See text page 142.
Right, only part of diamond worked in oblique twining. Author’s wool sample, copied from tang in collection of Hiroko
Iwatate, Tokyo, 1997. See text page 141.
61 Plain combined with S- and Z-course oblique twining; outlining cords in a third colour. Author’s wool sample, based on
part of tang from Kanod, acquired December 1995. See text pages 132 and 142.
Plain combined with S- and Z-course oblique twining. Left, structure with
regularly spaced longitudinally twining cords. Author’s wool sample,
based structurally on tang acquired at Mandawas, October 1990. See tex
Left: Getah Ram making black SCOT page 144. Right, free-running cords twining on a one-colour ground. Tang
tang at Sam Sands, December 1995. acquired at Sam Sands, November 1990. See text page 143.
Right: Pokar Ram making TLOI/ tang 63 TLOI. top and bottom, two sides of braid with inverted triangle motif
at Doli Kalam, February 1989. exactly as described in text on pages 145 to 147. Author’s sample, wool.
LO
oO
WS
& SIZES
ENT
66
TLOI. See text page 150. Left, to show irregular transverse and longitudinal stripes. Tang acquired at Suna, January 1994.
Right, to show clean oblique stripes. Tang collected by Girdhar Ram, January 1994.
vl TLOI. See text page 150. Left, small repeating spot pattern. Right, combining colour boundaries of all types.
Both from tang acquired from Kalu Ram at Lahti, January 1989.
wv TLOI geometric motifs. See text page 150. Left, with shift in border where tang hangs over saddle bar. Lhoum acquired from
Lecklaj, Barmer, November 1990. Right, with narrow border. Old tang acquired from Sabu Kahn, Devicot, January 1994.
67
©
e
es
Sit
ty
saa,
hig
enAy.
ee ae FED hi
Cee < ‘ oe ceoe eeee é ey x
’ a a ee er ee eA
rt ee ee
ee ee ee 2
hg Z - 4
1 eRDe OS we ew ee 6 ke le ee
TLOI. See text page 150. Left, complex geometric motifs. Cotton tang acquired from Girdhar Ram, 27 January 1994.
Right, geometric motifs, including area with hairline pattern. Tang acquired from Abhamaniyu Rathi, Jaiselmer,
December 1995.
TLOI. See text page 150. Left, representational motifs. Tang made by Gunga Singh Bhatti of Damodra, acquired at
Harmira, December 1995. Right, similar, plus transverse rows of cotton tufts. Tang acquired at Suna, January 1994.
POSbanecs:
PP
is
Me
%
~—
*4
7
e
_«&
e
ad
oS
—*
~~
+ 3
<
re ;
nthe
nohatetam
70
We TLOI. See text page 150. Left, peacock on house. Newly made tang bought from its maker, Ishwar Singh
Bhatti, Jaiselmer, 1987. Right, motifs, including camel and two friendly soldiers, plus cotton tufts. Tang
acquired at Bouha, January 1994.
7
A) TLOI. See text page 150. Left, unfortunately reverse of maker’s name, Bhaga Ram. Tang acquired at
Bachharau, November 1990. Right, mixing geometric with representational motifs. Tang made by Gunga
Singh Bhatti of Damodra, acquired at Harmira, December 1995.
71 TLOI, four ways of sealing the two
layers at selvages. Author’s sample,
wool. See text pages 152-4.
72 TLOI. See text page 152.
Left, pocket with a hundred-rupee
note. Lhoum acquired from John
Gillow.
Right, sealing selvages with triangular
pockets. Unfinished tang acquired
from its maker, Pokar Ram, at Doli
Kalan, February 1989; white is cotton
73 TLOI, Right, extra cords linking with
TLOI cords to seal selvages.
Left upper and middle section, using
SCOT and POT to make
ply-split border; lower section, coin in
pocket only accessible from one side
of braid. Both author’s samples, wool.
See text pages 153-4.
71
74 TLOI, two designs in black, white and
orange dyed goat hair. Tang acquired
from Manubhai, dealer, in Anmedabac
1989. See text page 155.
75 2/2 Twill oblique twining; from top
downwards; longitudinal ridges giving
transverse zigzags; transverse ridges
giving longitudinal zigzags; diagonal
stripes. Author's sample, wool.
74 See text page 157.
NSuw
TOI
SZ
>Ms
Zi
NY
76
77
we Black and white goat-hair bheardee. Acquired at Dhaneti, near Bhuj, Kutch, January 1994. See text page 166.
From top, dheras made by Karan Singh Tanwar at Kelawa, February 1994; Nathu Ram at Lawan, Pokaran,
December 1995; Ishwar Singh Bhatti, 1987; ball of singles white goat hair, spun by Kama Ram at Khandi,
1992. See text pages 167-8.
as
U7<u
/
78 Spinners of goat hair. See
text page 167.
Top left, Jnat camel driver
using ball as whorl.
Top right, using very large
bheardee.
Bottom, close-up of
spinner’s hands, Dhaneti,
Kutch, January 1994.
Making connected cords.
See text page2 171-2.
Top, winding 2-ply goat
hair around two stakes in
the ground.
Bottom, twisting two loops
on hooked stick. Jhura-
Sodha Camp for Pakistan
refugees, north of Kutch,
Gujurat, November 1990.
79
80 Above Ply-splitting needles. Right to left; made by Knema Ram Machra at Ghevra, 1987; Rama Ram at Jaiselmer,
1987; Ishwar Singh at Jaiselmer, for author, 1987; Kharna Bhima at Mamura, Kutch, January 1994; Pokar Ram at Doli
Kalan, February 1989; two hooks from Sipia, November 1990, and Suna, January 1994. All wood except Pokar Ram's,
made from deer horn. See text page 174.
Above: Kharna Bhima, maker, at Mamura, Kutch, January 1994.

Right: Aatam Ram and author, holding tang made by Aatam, at Sam
Sands, January 1996.
81
ee
etek, te ae
Maat a ae
S
2 = s
@
SS EEO
S28> gee
PSS
O2 0
gts
YO a @
So
7-7
ORES Seo
6BORN OG Nog
ae
ee s etay
e)es,
Co
Os oe
© EG 16
1996. See text page 175ff.

(3) Oo SY) ep) eo =hd
~o <0
E Oo &<5
Se a
TS
©
0 : Oo So se)
2
oligo G
fCetyct
sE S50 22
e o, OS 5
Oke
a a eee yes
—
G=eB eELO=)
~5,Or5.0co
20a
Ono)
cCOC
oO =
r
Looa7 Oost
31
82
Right: Tulsa Ram holding SCOT/POT tang he made, Dhivera,

February 1989.
83
82 Left, rare SCOT tang made of 2-ply aak fibre. Acquired at Mandawas, January 1994. See text pages
165 and 175. Right, very old SCOT tang with natural and dyed goat hairs. Acquired from John Gillow,
Cambridge, 1980s; exactly similar tang seen at Umgiriji-ka-Math temple, Kelawa Kala, north of
Jodhpur, said to be made at Ghevra. See text page 175.
83 Starting ends. See text page 176.
Left, buckle on cotton tang. Acquired in Shekavarti district, near Nawalgarh, January 1994.
Centre, buckle on tang. Also from Dabla, January 1994.
Right, wooden stick on tang. Acquired at Dabla, January 1994.
84 Top left, buckles fixing tang to saddle, at desert fair, Jaiselmer, 1987.
Top right, Knarna Bhima making TLOI tang, at Mamura, Kutch, January 1994.
Bottom left, buckle fixing tang to saddle, at Sam Sands, Jaiselmer, 1996.
Bottom right, Aatam Ram making SCOT tang, Sam Sands, January 1996.
See text pages 176 and 177.
85 = Left, SCOT tang with 12 sections and only 4 white cords. Acquired from Asu Lal, dealer,
at Chohatan, February 1989. Right, SCOT tang with 16 sections.
See text page 177-8.
86 Two SCOT tangs with haphazard arrangement of colours. See text page 178.
Left, acquired from its maker, Bheru Singh, at Narnadi, October 1990.
Right, acquired at Sikapura, January 1989.
ary
87 Two common ways of arranging
equal amounts of black and white
goat hair at start. See text page 178.
Left, SCOT tang from John Gillow,
Cambridge. Right, SCOT tang
acquired at Gagani, January 1989.
88 Inversions. See text page 178.
Left, SCOT cotton tang. Acquired
near Bhuj, Kutch, Gujurat, January
1994. Centre, SCOT tang. Acquired
from its maker, Bheru Singh, at
Narnadi, October 1990.
Right, SCOT tang. Acquired at
Gagani, January 1989.
89 SCOT multicolour cotton tang with
inversions. Acquired from Thakur
Rao Sher Singhi at Kalyanpur,
October 1990. See text page 178.
90 Left, SCOT tang with cords changing
function along oblique lines, with
securing peg. Acquired at Dholera,
Janaury 1994.
Centre, SCOT tang with pattern
changed by twined linkings. Acquired 87
from John Gillow, Cambridge.
Right, SCOT tang with arrowhead
design produced by twined linkings.
Acquired from Asu Lal, dealer at
Chohatan, February 1989.
ae
Oe ow
=“
Lex
Lio
tien
se
parees
(SS
PPP:
—
4¢.
33h
os.
TPES,
3>>7.
Sa
aS
a
amcor
93
92
91 Left, POT tang with hairline and other designs. 92 POT tang. Collected by Girdhar Ram, Jaiselmer,
Acquired from Sabu Kahn at Devicot, January 1994 January 1994. See text page 179.
Right, old POT tang with pattern changed by twined 93 Tang with combined plain and S- and Z-course
linkings. Acquired from John Gillow. oblique twining. Collected by Girdhar Ram,
See text page 179. Jaiselmer, January 1994. See text page 179.
94 Tang combining sections of
ply-split darning with ply-split
braiding. Collected by Mr Wazir,
Bhuj, Kutch, Gujurat, 1980s.
See text page 180.
95 Left, SCOT tang with section of braided
cords. Acquired at Shiv, February 1987.
See text page 181.
Right, all-black SCOT tang with beads
and slits. Collected by John Gillow. See
text page 180.
96 Two tangs decorated with cotton tufts.
See text page 181.
Left, acquired at Sarecha, January 1989.
Right, acquired at Suna, January 1989.
95
96
97
97 Two long tangs. See text page 181-2. Left, TLOI in three colours. Acquired from Manubhai, dealer, in Ahmedabad,
1989. Right, combined plain with S- and Z-course oblique twining. Acquired from John Gillow, 1980s.
98 Details from two long tangs. See text page 182. Left, plain combined with S- and Z-course oblique twining, with motif in
cotton tufts. Acquired from Lecklaj, dealer, in Barmer, January 1990.
Right, showing twined cotton decoration. Complete tang shown in Plate 97, right.
4,
x &% 4
ils OX,
Y
Ch
.
* ”wf
99
99 Short Ihoums. See text page 182. Left, a pair in SCOT. Acquired at Sarecha, January 1989.
Right centre, same technique, acquired from Hanswa.
Far right, combined plain with S- and Z-course oblique twining, faded red and green cotton. Acquired from Abhamaniyu
Rathi, Jaiselmer, February 1987.
100 Short lhoums. See text page 182.
Left pair, SCOT in red cotton, decorated with buttons, cotton tufts and mirrors. Acquired from UK dealer, 1980s.
Right pair, SCOT in black goat hair, decorated with white shells, cotton tufts and mirrors. Acquired from Mr Wazir, Bhuj,
Kutch, Gujurat, January 1994.
100
SS
settee SSNS
($e%e
>
101
eae
Poeee
=
eal
=
ee
Zl
we
F
Ke
a
SoS
Dead
7
i
SEP)
SE
S$S33333
35°
>>
e
P >
%
ee r
me4a73
101 Long Ihoums. See text pages 182-3. Left, combined

plain with S- and Z-course oblique twining. Acquired
from camel driver at Sam, 1985.
Centre, POT. Acquired from Manubhai, dealer, at
Ahmedabad, 1989.
Right, TLOI. Acquired from Lecklaj, dealer, at
Barmer, November 1990.
102 Long lhoums. Left, SCOT, POT and TLOI. Made by
Tikam Singhi, acquired from his grandson, Karam
Singh Parihar, at Bhadka, February 1989. Right,
POT with large central split to fit over saddle.
Acquired from John Gillow, Cambridge, 1980s. 102
104
103 Fringes. See text page 184. Left, 8-strand braids becoming
smaller and smaller, on a tang acquired at Dabla, February
1987. Right, 8-strand tubular braids in three different colour
ways, on a tang acquired at Bouha, 1994. Left: Hands of Modu Ramji making a SCOT
104 Fringes. See text page 185. Left, 8-strand braids changing tang, Kakrala, February 1989.
colour where new cords are attached, on a tang collected by Right: Rajput men, who sold me two multi-
John Gillow, 1980s. Right, narrow TLO! braids, on a tang from coloured cotton tangs said to be from Ghevra,
Suna, acquired January 1994. in guest house near Jhanwar, October 1990.
105
105 Gorbandhs. See text pages 188-90. Left, POT at top changing to SCOT, green cotton.
Acquired from John Gillow, late 1970s.
Right, SCOT, black goat hair, decorated with shells and cotton tufts.
Acquired at Sarecha, January 1989.
:
A
¥
A
FW
7%.
aVaasM%
va¥
Coe
ao i,
-~
. wataval sar,
Vavaa¥ avYat,¥ Avs
ry
+>
.
106
106 Gorbandhs. See text pages 188-90. Left, 2/2 twill oblique twining, black goat hair. Collected by
Hiroko Ilwatate, Tokyo, 1970s.
Right, SCOT, three colours of cotton, decorated with white buttons. Acquired at Suna, January 1989.
107 Gorbandh; SCOT, red cotton,
decorated with white buttons,
tufts and mirrors. Acquired
from Manubhai, dealer,
Ahmedabad, 1990. See text
page 188-90.
108 Godiya and Sariya. Top,
sariya, ankle band, SCOT, red
cotton with buttons and tufts.
Bottom, godiya, knee band,
similar plus mirrors. Acquired
from Manubhai, dealer,
Ahmedabad, 1990. See text
page 190.
109 Muhra; POT, black goat hair.
Acquired from Kalu Ram at
Lahti, January 1989. See text
pages 191-2.
108
109
110 Hardi; SCOT in transverse
sections, black goat hair.
Acquired at Hamira,
December 1995.
See text page 191.
111 Bheer; 2-section SCOT in
black goat hair, one end
looped, other end tapered.
Acquired from son of the
maker, Deva Ram, near
Netra, February 1993.
See text page 191.
112 Cheekah; open-mesh type,
twined linkings of pairs of
cords. Acquired at Hanswa,
January 1994.
See text page 193.
113 Cheekah; open-mesh type,
widely spaced POT areas,
3-ply white cotton,
decorated with mirrors and
yellow, red and blue cotton
wrapping. Acquired from
Manubhai, dealer, at
~ Amhedabad, 1990.
See text page 194.
112
114
114 Cheekah; closely made
type, SCOT in transverse
sections, with diamond
and triangle shapes.
Top, black example.
Acquired near Chohatan,
February 1989.
Bottom, black and white
example. Gujari
collection, Anmedabad.
See text pages 125 and
196.
type; POT with whole
twist left between
successive splittings of a
cord. Acquired Rajasthan,
1990s.
See text page 197.
type, only using twined
linkings. Acquired at
Hamira, December 1995.
See text page 197.
117
118
117 Two cheekahs; closely made

119
type with flat bottom; SCOT in
transverse sections with
leather patch to close central
hole. One acquired from
Hanuman Singh, near Kanod,
December 1995; other from
Bouha, January 1994.
118 Mukhiarna. Left, yellow cotton
with tufts and added tufts at
end of fringes. Acquired from
Abhamaniyu Rathi, Jaiselmer,
January 1994. Right, brown
cotton. Acquired from Lalji,
dealer, Jodhpur, 1987.
119 Mukhiarna. Left, modern, bright
yellow. Made by Jumi Imtiyaz
Kahn for Amar Seat Works,
Hawa Mahal Bazaar, Jaipur;
acquired 1994. Right, fine
example, red and yellow cotton
decorated with tufts and beads
in 5 colours. Acquired at
Kalyanpur, January 1989; said
to have been made in Pakistan.
120 Columbia crupper. Jop, detail showing the typical
zigzag pattern and the use of a two-ended needle.
Bottom, Guajiro Indian making crupper.
Photographs by Marianne Cardale-Schrimpff,
Bogota, around 1970; see bibliography and
text page 200.
yw!
%
a
f= 5
: oa
Uy
ay
:
Indianapolis
Marion County
Public Library
Renew by Phone
269-5222
Renew on the Web
www.imcpl.org
For general Library information

please call 269-1700.
Ions onOxo)Ibbevcanveleremeu-bbelaemerelelon
Ethel Mairet, Barbara Sawyer and
Alastair Morton. He set up his own
workshop in 1952 and began to
produce handmade rugs that were
sold at Liberty’s and Heal’s and
r.debleyicce Ml onanrete Colecmrole (ois(wrt ele!
Feauh)Coteus w Com oF:tmmetotesolantews Beyetelovem-tae
schools since 1954 and held yearly
workshops in the United States since
1962. His work has been shown all
over the world and he was the first
living weaver to be given an exhibition
at the Victoria and Albert Museum in
London. He has received many
awards for his work, including an
OBE in 1974.
Peter Collingwood published his

much acclaimed The Techniques of
Rug Weaving in 1968 (now in an
eleventh edition in paperback). He
subsequently published The Technique
of Sprang (1974), The Technique of
Tablet Weaving (1982, paperback
1996), Textile and Weaving Structures
(1987), reprinted as The Maker’s Hand
(1998), and Rug Weaving Techniques,
Beyond the Basics (1991).
Unicorn Books and Crafts, Inc.

Petaliege. Ca
© 94954
ISBN 0- 9625586- 9- 9
cat 7 "b2556b9-"
| 5869
Illustrations on front cover, left:

Multicoloured cotton tang.acquired:at
Jodhpur Right: Aatam Ram with the author
: te ©
INDIANAPOLIS-MARION COUNTY P.L.
Above: Hol y man with very old tang near Ossiyan, ‘Above: Man examining the author’s samples at
: | A
Rajasthan : - Dangiyawas, Rajasthan
- Below: Detail of long tang acquired from UK dealer ~ Below: Rings made by, the author
*
—
=
ey =
—=
ae
=
———
ae
—
—
a
—
—=
—=
;——_—}
——s
—
ee
——s
—
—
—
ans
————
—=
——-
==
—
9¢G-HO —_
—
—
aaa
—
—_

Techniques of Plysplit Braiding

Uploaded by

Copyright:

Available Formats

Techniques of Plysplit Braiding

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Techniques of Plysplit Braiding

Uploaded by

Copyright:

Available Formats

M Wotoms Kerdebeblepelacneya

Ply-split braiding is a thread

Peter Collingwood, as in his other

z BOOKS AND CRAFTS INC.

Copyright © Peter Collingwood

All rights reserved. This publication may not be

Typeset by Antony Gray

Ply-Split Darning CHAPTER THREE (continued)

. Irregular splitting of parallel plied Ue Using a cord-making device 25

CHAPTER THREE CHAPTER FIVE

Non-Traditional Cord Preparation Explanations of Diagrams

Plain Oblique Twining CHAPTER SIX (continued)

Introduction 37 rePAE Mounting a special colour

Single-Course Oblique Twining CHAPTER SEVEN (continued)

1. Working a single-section braid 2.1le Variations in structure in

Changing from aV- to an Analysis of designs with

Changing from aA- to an 2.1g Other structural variations 106

. Recognition of ply-split TLOI 154 3.4a As separate cords 168

Converting two-colour into Variations in working

. Using three colours 155 3.4b As connected cords 171

. Introduction 158 3.1a Cords used 175

. Two types 158 3.1b Starting end 176

3.5 Fringes on tangs and lhoums 183 9.2a AE SCOT 196

This attempts to be a different sort of textile book,

Every type of fabric has its own specific method of

14 * THE TECHNIQUES OF PLY-SPLIT BRAIDING

16 * THE TECHNIQUES OF PLY-SPLIT BRAIDING

1. DISTRIBUTION on ten months field research carried out in

18 * THE TECHNIQUES OF PLY-SPLIT BRAIDING

20 > THE TECHNIQUES OF PLY-SPLIT BRAIDING

Non-Traditional Cord Preparation

NON-TRADITIONAL CORD PREPARATION * 21

22 * THE TECHNIQUES OF PLY-SPLIT BRAIDING

1.2 With two people

NON-TRADITIONAL CORD PREPARATION +23

Figure 3 a Making a two-colour cord; b—c stages in making a 4-ply cord

24 * THE TECHNIQUES OF PLY-SPLIT BRAIDING

LEZ Z III IIA

Figure 4 Cord-making device and fixing board at far end.

NON-TRADITIONAL CORD PREPARATION *¢ 25

3. MAKING A LONG LENGTH OF OVER-

26 > THE TECHNIQUES OF PLY-SPLIT BRAIDING

NON-TRADITIONAL CORD PREPARATION °27

Whatever method he uses the worker will

28 » THE TECHNIQUES OF PLY-SPLIT BRAIDING

. 1. FIXING THE BRAIDING CORDS

1.1. Fixing to a starting rod

b Double a cord over the rod and fix

+ that cords made by the continuous-cord method,

+ that methods a and b tend to place the cords

1.2 Supporting the starting rod

as in Fig. 11. Attached thus the braid can be easily

30 > THE TECHNIQUES OF PLY-SPLIT BRAIDING

1.3 Fixing to transverse cords

loop of braiding cord up through it. Leaving a quarter- rz

twist, repeat, making the second loop pass through

a Needle, modelled on those used in India, whittled

b_ A length of heavy wire, bent as shown and with its

c A two-ended needle, as used in Colombia, made

d Latchet hooks. Though there is obviously a