flyj L.Wnv
^
Cop. S.
3»
HATTEl^N -^ DESIGN
LEVVISFD/^
LIST OF BOOKS BY THE LATE LEWIS F. DAY
IN PRINT ENAMELLING.
ART IN NEEDLEWORK: A BOOK ABOUT EMBROIDERY. Fourth Edition.
ALPHABETS OLD & NEW. Third Edition.
LETTERING IN ORNAMENT. Second Edition.
PATTERN DESIGN. Third Impression.
MOOT POINTS.
OUT OF PRINT
WINDOWS: A BOOK ABOUT STAINED GLASS.
Third Edition. NATURE AND ORNAMENT.
ORNAMENT AND ITS APPLICATION.
HATTERN ^ DESIGN
A BQDK FOR STUDENTS TREATING IN A PR AC TICAL WAY OF THE AN ATOMY PLANNING 8>l EVOLUTION OF REPEAT ED OI^NAMENT BY
LEWIS FDA/
AUTHOR. OF NATURE IN ORNAMENT ALPHABETS LETTER-iNGi ORNAMENT
LONDON. B.TBATSFORDLtd
FirsI Impression printed 1 903 Second ., ,, 1915
Third ,, ., 1923
Tipografia Sociale del Cav. Carlo Sironi - Milano, Via Goffredo Mameli, I 5 (Printed in ItalvJ
PREFACE.
A MAX has a right, I suppose, to pull down the building he once put up, and to raise another in its place. If he should see fit to use sometimes the ver\' stones which belonged to it, he would only be stealing from himself. I have done something very much like that.
In the course of the last fifteen years the times have changed, and with them the standpoint of students and teachers of design; and, though my point of view has not altered, my outlook has widened with experience. When it came to the revision of " The Anatomy of Pattern " with a view to a fifth edition, it seemed to me I had done all I could do to it, that it was past mending, and that the simplest thing would be to start afresh.
The present volume, however, though it covers the ground of the former one, and answers much the same purpose, is not the same, but really a new book upon the foundations of the old one.
It contains, indeed, all that was in the other, but otherwise expressed. Here and there an explanation or description, which, by revision after revision, had been reduced to the fewest and plainest words I could find, has been allowed to stand. So with the illustrations, the greater number of them are new. Such of the old diagrams as were essential to the purpose of the book have been drawn again, not merely on the larger scale allowed by the page, but in a simpler and more self-explanatory way
VI PREFACE.
It will be seen from them and from the table of contents that " Pattern Design " covers much more ground than "The Anatomy of Pattern ". But it does not go beyond its subject. The appearance, since the original publication of my little books, of a number of similar volumes each attempting to embrace more than the one before it, has firmly convinced n.e that the better plan is to confine oneself to a definite subject, and to treat it thoroughly. The last word, of course, is never said so long as there is life left in it.
I know very well that knowledge gained in practice can be only very partially conveyed in words; but something of the experience of five and thirty years and more in practical pattern design is surely communicable; and, for what it is \\orth, I give it here.
LEWIS F. DAY.
1st September 1Q03.
" Pattern Design " is a book which represents so completely my father's mind on the subject with which it deals and which treats, moreover, so largely of principles unaffected by changes of time and fashion, that it has seemed wisest in issuing a new edition to leave his work practically as it left his hands. The changes now made are limited, therefore, to the correction of a few verbal errors.
RUTH DAY. 15 Tavitox Street, Gordon Square, W.C. July 1915-
CONTENTS.
<::-iAPTEK PAGE
I. What pattern is i
II. The square . . . n
III. The triangle 19
IV. The octagon - . . 24
V. The circle 28
VI. The evolution of pattern 40
VII. Practical pattern planning 54
VIII. The " drop repeat " 60
IX. Smaller repeats ... ..... .81
X. Sundry scaffoldings 98
XI. The tl^rnover 119
XII. The turn-round 126
XIII Pattern planning in relation to technique . 131
XIV. How TO set about design 157
XV. To prove a pattern i9'^>
XVI. Colour 201
XVII. Borders 212
XVIII Pattern not strictly repeating .... 240
XIX. Expedients in practical design .... 249
XX. The invention of pattern 261
Index 265
LIST OF ILLUSTRATIONS.
NC PACE
1. Peach blossoms on the ice — Japanese .... 4
2. Predetermined gaps in pattern — L. F. D. . . . 6
3. Balance of ornament enough for a panel but not
for a repeated pattern l. f. d: . . . . 7
4. Wall pattern in which the lines of the scroll
are quite evenly balanced l. f. d. . . . 9
5. Stripe — breaks giving cross-lines ii
6. Zig-zag — -points giving cross-lines ii
7. Stripes — ornament giving slanting lines . . . 12
8. Stripes — ornament giving upright lines ... 12
9. Square lattice. Chf.couer 12
10. Interlacing 12
11. Interlacing strips of different colour . . . 13
12. Close plaiting 13
13. Looser plaiting 13
14. Open plaiting 14
15. Plaited strips of different width 14
16. Diapers of squares and groups of squares ... 14
17. Diaper, of which the unit of repeat is a group of
three squares. Diaper, of which the unit is a
GROUP of five squares 1 3
lb. Diaper, of which the unit of repeat is a group of
EIGHT squares — tAlHAMBRESQUE TILE MOSAIC . . 1 5
19. Fret pattern on the lines of a square l.\ttice . i6
20. Diamond diaper i6
21. Zig-zag diaper built on diamond lines . . . . i6
X LIST OF ILLUSTRATIONS.
NO. PAGE
22. Diamond diaper i6
23. Frets, etc., on the lines of a square lattice . . 17
24. Triangle. Equilateral triangle. Diamond . . 19 2^. Diaper built on a triangular plan — Alhambresque
TILE mosaic 20
26. Hexagonal and star-shaped diapers built on
triangles 20
27. Diapers built up of hexagons 21
28. Lattice and the lines on which it is built — Indian
PIERCED stone 21
29. Lattice and its triangular b.\sis — -Indian pierced
STONE 22
30. Lattice and its triangular basis — -Indian pierced
STONE 23
31. Octagon and the lines on which it is built . . 24
32. Different views of same pattern 25
33. Pattern from cane-bottomed chair .... 25
34. Tile work from the Alhambra 26
35. Simple and more complicated trellis lines . . 27
36. Pentagon diaper and its skeleton . . . . . 27
37. Zig-zags developing into wave lines .... 28
38. Relation of octagon to circle diaper .... 29
39. Straight-lined and curvilinear varieties of the
same pattern — Alhambresque tile mosaic . . 29
40. Diaper of circles planned on square lines . . 30
41. Diaper of circles planned on diamond lines . . 30
42. Diaper of circles planned on lines of diagram 40,
BUT radius relatively WIDER 30
43. 44. Diaper of circles pi..\nned on lines of di.\gr.\m
41, but radius relatively wider . . . .30,31
45. Double diaper of circles planned on lines of
diagram 40, but 'radius relatively wider still . 3i
46. Diaper of interlacing circles 32
47. Diaper of circles interlacing with square trellis . 32
48. Diaper of scales and derivatives from it . . -33
LIST OF ILLUSTRATIONS. xi
NO. PAGE
49. Scales tx.rned about to make a flowing dl\per . 34
50. Showing relation of scale to diamond, ogee, and
hexagon 55
51. Wave lines, ogee diaper, and interlacing ogees —
THE last giving HEXAGONAL SHAPES ■ . ■ • 3,5
52. Diapers constructed on circular lines — Gothic
tracery 3'3
53. A DIAPER and its CONSTRUCTIONAL LINES GOTHIC
TRACERY 3';>
54. Tracery pattern and its constructional lines —
Gothic 37
55. Diaper and its constructional lines — Gothic tracery 3S
56. Straight-lined and curvilinear versions of diaper
built on the same lines as 6l 4o
57-62. Diagrams showing the various ways in which
THE same diaper MIGHT POSSIBLY BE CONSTRUCTED .4I-44
63. counterchange diaper and its construction
Alhambresque tile mosaic 45
64. Lattice pattern — Arab woodwork 45
65. Diagram showing construction of 64 . . . . 46
66. Diagram explaining the composition of the unit
in patterns 64 and 65 46
67. CoUNTERCHANGE PATTERNS CONSTRUCTED AS 66 ARAB . 47
68. Key to the construction of 69 47
69. Roman mosaic pavement pattern 48
70. Counterchange pattern — Alhambresque tile mosaic . 49
71. Lattice pattern, possibly derived from tessel.\ted
work — Arab woodwork 50
72. 73. Lines on which lattice 71 might have been built 51
74. Lattice and the lines of its construction — Arab
woodwork 52
75. Lattice and its construction — Arab woodwork . . 53
76. Curvilinear development of diagram 26 . . . ^5
77. Diagram showing the relation of equilateral
triangles and hex.\gons to square repeat . . 56
58
XII LIST OF ILLUSTRATIONS.
NO.
78. Diagram showing how octagoxal diapers will re-
peat ON rectangular lines
79. Lattice pattern • built on triangular lines — Indian-
pierced STONE
So. Diagrams showing how a pattern may be designed
TO hang in two ways 61
81. Diagram to illustrate the " drop repeat " , . 62
82. Diagrams to illustrate the way repeats must fit . 62
83. Diagrams to illustrate the difference of measure
WHEN A square REPEAT IS TURNED ROUND TO WORK
AS A DROP 63
84. Diagram to show that a pattern may be contained
equally within square and various diamond lines 64
85. Two plans upon which the same design might have
BEEN built 66
86. Wall-paper design by Walter Crane .... 68
87. Diagram showing the mechanical relation of
various plans .69
88. 89. Diagrams showing the division of a square re-
peat into three parts and their transposition
to form a wide diamond 69
90. Brick or masonry pattern — L. F. D 70
91. Diagrams of the brick plan and its relation to a
" drop repeat " 7i
92. Hexagon plan — its relation to diamond ... 72
93. Counteracting lines of design — the repeat a double
square • 73
94. Diamond forms resulting from counteracting lines
THE repeat a double SQUARE 73
95. Diagram of a step plan which does not gi\'e a
" DROP repeat " 74
96. Diagram of a pattern which drops one-third of its
DEPTH 75
97. A PATTERN WHICH DROPS ONE-QUARTER OF THE DEPTH.
OF REPEAT — Persian 75
LIST OF ILLI'STRATIOXS.
NO.
98. A STEP PATTERN WHICH IS NOT A DROP REPEAT- — TlI.ES
IX LOW RELIEF. L. F. D
99. A PATTERN WHICH DROPS ONE-THIRD OR TWO-TIHRDS OF
ITS DEPTH — Tiles. L. F. D
100, loi. " False drop " patterns ....
102. A fret designed on ogee and square LINES
103. Diagrams showing drop and step repeats within
the width of the material
104. Diagram showing the plan of a " drop repeat "
measuring only two-thirds of the width of the material
105. Diagram showing the plan of a " drop repeat "
measuring only two-fifths of the width of the material
106. A drop design schemed on a diamond measuring
TWO-THIRDS OF THE WIDTH OF THE MATERIAL .
107. A drop repeat measuring 7 inches by 14 INCHES
PRINTED FROM A BLOCK MEASURING 21 INCHES BY I4 INCHES
108. Diagram of a " drop repeat " on diamond lines, the
diamonds not filled all in the same way
109. Pierced work with geometric divisions enclosing
ORNAMENT WHICH DOES NOT REPEAT ByZANTINE
carved and PIERCED STONE. VENICE
no. Diagram of a three-spot repeat III. Diagram of a four-spot repeat 1 1-2. Diagram of five-spot repeat
113. Diagram of a six-spot repeat .
114, 115. Diagrams of seven-spot repeat 116, 117. Diagrams of eight-spot repeat
118. A CEILING pattern DESIGNED ON THE PLAN OF A SIX
repeat — L. F. D
119. Diagrams showing the evolution of iiS
120. Diagram showing the distribution of flowers only
on the plan of a six-spot repeat, the SCROLL
designed independently . . ...
XIII
PAGE
/ / 78. 79
80
Si
-spot
82
S3
86
87
88 88 88
89 90
91
02 93
94
XIV LIST OF ILLUSTRATIONS.
NO, . PAGE
121. Diagram showing the distribution of open spaces
of ground on the plan of a five-spot repeat . 95
122. Diagram showing the further development of the
PRINCIPLE OF distribution ALREADY EXEMPLIFIED
in diagrams iio to ii9 96
123. Diagrams illustrating another principle on which
SPOTS, ETC., may be DISTRIBUTED .... 97
124. Turnover pattern, revealing the vertical line on
WHICH IT IS reversed — Cretonne. L. F. D. . 99
125. Design almost but not quite self-contained within
THE width of the STUFF VELVET. L. F. D. . lOO
126. Diagram showing the construction and develop-
ment OF COUNTERCHANGE PATTERNS . . . . lOI
127. counterchange pattern constructed on the lines
of diagram i26 alhambresque .... loi
128. 129. Ribbon and feather pattern and the lines of its
construction i02
130. Velvet pattern — ^Late Gothic 103
131. Pattern in which wave-lines divert the eve from
vertical — Printed cotton. L. F. D. . . . 104
132. 133. Diagram showing the scaffolding of a pattern
and the lines naturally resulting from it . 105
134. Old print — Dutch 106
135. Woven pattern — Cret.\n 107
136. Pattern in which the horizontal line is deliber-
ately marked 108
137. Diagrams of scaffoldings 109
138. Silk brocade on the lines of diagram b. 137 . .110
139. Velvet on the lines of diagram c. 137— Italian
sixteenth century Ill
140. Pattern in which the diagonal line was an after-
thought— Curtain material. L. F. D. . . 112
141. Pattern in which the starting-point was a diagonal
wave — Dress fabric. L. F. D. . . . .113
142. Diagonal pattern and its scaffolding — Fifteenth
century 115
LIST OF ILLUSTRATIONS. xv
so. PAGE
143. Silk pattern and its scaffoldixg— Sicilian . .116
144. The planning of a design — French. Louis Seize . 117
145. " Turnover " pattern — Byzantine 120
T46. " Turnover " pattern — Gothic 121
147. Quasi-" turnover " pattern — Wall-paper. L. F. D. 122
148. Turnover design with central strip not turned
OVER 123
149. " Turnover " and " drop " pattern — Stairc.\se wall-
paper. L. F. D 124
150. Not a turnover pattern — Printed linen. L. F. D. . 125
151. Diagram showing a square repeat which turns
ROUND 126
152. Diagram showing a square repeat which turns
round and drops 1 27
153. Turn-round pattern— Floor tiling. L. F. D. . . 128
154. Lattice and the lines into which it may be resolved
— Arab 129
155. Diagram to show a half-turn of the block in-
printing AND the reversal OF ALTERNATE STRIPS OF the material IN THE HANGING— CeILING PAPER.
L. F. D 133
156. Diagram of ceiling pattern (in effect 42 inches
ACROSS) ON THE PRINCIPLE OF DIAGRAM I55^PaPER.
L F. D 13^
157. Diagram showing a pattern (in effect 84 inches
wide) produced by 2 BLOCKS, EACH ONLY 21 INCHES
BY 21 INCHES 1 35
158. Six-inch TILE DESIGNED TO BE FIXED BRicKwisE — L. F. D. 137
159. Narrow turnover pattern adapted to weaving — -
L. F. D 138
160. Narrow woven pattern — Damask. Seventeenth
century 139
161. Wall-paper pattern 1014 inches wide by which no
economy is effected — L. F. D. . . . .140
162. Diagram of woven pattern, "turnover" and "single" 141
XVI LIST OF ILLUSTRATIONS.
NO. PAGE
163. Diagrams showing the correspondence in depth of
repeat between the parts of a design . . . j42
164. Diagram showing the relation of diaper repeat to
REPEAT OF side AND BOTTOM BORDERS . . . I43
163. Diagram showing portions of design answering
practically to borders i45
166, 167. Diagrams of a iable damask design and the
LINES ON which IT IS PLANNED .... I46, I47
168, 169. Diagrams of the centre part and lengthening
PIECE of a SQU-^RE tablecloth 1 50
170. Diagram of the centre part of a long tablecloth,
showing the introduction of lengthening pieces 151
171. Diagram showing change of colour in the weft
threads 133
172. Diagram showing change of colour in the warp
threads 154
173. Diagram showing revers.\l of design in chenille
WEAVING 153
174. Would-be free pattern falling into the lines of
THE brick repeat MODERN FRENCH . . . 1 57
173. Diagram showing how a recurring feature marks
the plan of a pattern 1 58
176. Painted diaper in which the " repe.\t " does not
WORK — English Gothic 159
177. Diagram showing the geometric recurrence of
features not geometrically distributed . . l6o
178. Pattern falling into stripes — Old table damask . 161
179. Featureless " allover " pattern — Ceiling paper.
L. F. D 162
180. Features of the pattern disguised, but not quite
LOST, in " allover " EFFECT WaLL-PAPER. L. F. D. I63
181 Pattern with marked governing lines — Wall-paper
founded upon an old stuff. l. f. d. . . 165
182. Pattern in which the lines of recurrence are
purposely lost — Wall-paper — L. F. D. . . i66
LIST OF ILLUSTRATIONS. xvn
NO PAGE
183. "Free" dress pattern deliberately planned — L.F.D. 167
184. Diagrams showing development of a " counter-
change " border pattern 168
185. 186. A geometric diaper and diagrams showing the
STAGES OF its DESIGN — FLOORCLOTH. L. F. D. . 1 69
187. Diagram showing a false start in design . . .170
188, 189. Diagrams showing the fair start of a design
and its development 171, 172
190, 191. Diagrams showing start of a drop repeat and
its development 173, 174
192. Wall-paper pattern — L. F. D 175
193. Diagram showing " brickwork " plan .... 177
194. Floral pattern in which the distribution of the
flowers was the first step of the design
L. F. D. and B. W 178
195-199- Diagrams showing the stages of a design . 179-182
200. Portion of the finished design developed in
diagrams 195 TO 199 — Cretonne. L. F. D. . 183
201, 202. Diagrams showing the stages of a pattern 184, 185
203. Finished design developed in diagrams 201, 202 —
L. F. D 1S7
204. Simple pattern and its after elaboration — Discharge
print — l. f. d 189
205. Diagrams showing the development of an Arabesque
DESIGN — Italian sixteenth century . . .190
206. Arabesque pattern bursting out into grotesques —
Embossed leather paper. L. F. D. . . . 193
207. Diagram showing afterthoughts as to treatment . 194
208. Diagrams showing how much of a repeat may with
advantage be roughly set out to begin with . 197
209. Diagrams to show the proving of a pattern . . 19S
210. Diagram to show how a pattern designed on
diamond lines may be proved on other lines . 199
211. 212. Diagrams showing the effect of different
colour schemes upon the same designs . 202, 203
xvin LIST OF ILLUSTRATIONS.
•NO. PAGE
213. Design in which the smaller detail, evenly dis-
tributed, GIVES at a distance SOMETHING OF THE
effect of a tint ^'\'all- paper. l. f. d. . . 2o4
214. Forms of a design blurred by colour in accidental
PATCHES — End paper. Old Dutch or German . 205 213. Diagram showing the systematic departure from
THE system of COUNTERCHANGED COLOUR ALHAM- bresque tile mosaic 206
216. Diagram showing the systematic disturbance of
THE uniformity OF A GEOMETRIC DIAPER BY MEANS
OF COLOUR — Alhambresque tile mosaic . . . 207
217. Pattern showing the systematic disturbance of
geometric form by variation in colour alham- bresque tile mosaic 208
215. Diagram showing the effect of different colour
TREATMENTS OF IDENTICALLY THE SAME FORMS.
L. F. D 210
219. Diagrams of flowing and broken borders . . 213
220. Cross lines in a border 213
221. " Stops " in a border 214
222. Running i ine and fret borders 214
223. Fret border — Greek 215
224. Fret facing both ways . . ■ 216
225. Interrupted fret facing both ways . . . .216
226. Broken Chinese fret 216
227. Mexican frets 216
228. [Masonry frets from the temple at Mitla in Mexico 217
229. Greek fret, including stop 217
230. Fret satisfactorily broken 218
231. Wave satisfactorily broken 218
232. two ways of arriving at the same fret . . .2x8
233. Fret and its equivalent in flowing lines . . 218
234. Development of wave scroll 218
235 Wave lines 219
230. Zig-zag and double Zig-zags 219
NO.
237- 238. 239. 240.
241. 242.
243- 244.
245- 246. 247. 24S. 249.
250.
251- 252.
253-
254- 255- 256.
257-
25S,
260.
261. 262. 263. 264.
265. 266.
LIST OF ILLUSTRATIONS.
From chf.vron to \va\'e
Twist — the curvilinear equivalent to diagram 236 Plaits of three and four strands guilloches and their construction
Twist of six strands
Interlacing border — Keltic .... Strap pattern on the lines of diagram 241
Chain patterns
Border of ornamentai links .... Wave-strap derived from zig-zag . Wave-strap derived from wave line Wave-strap derived from circles . Bent strap showing transition from square to lines
the
of a
Spirals added to a wave line Consecutive spirals giving a wave line Continuous scroll, possibly derived from
branching spiral scroll .... Effect of breaking across the running line
SCROLL
Effect of filling in spandrils in colour .
Leaves designed to fit spaces
Parallelism in border design ....
Borders with a strong and a weak side .
239. Border plan equivalent to drop repeat and dia
GRAM showing ITS RELATION TO A WAVE PATTERN
Borders planned on principle shown in diagram
233
Border planned on zig-zag lines . Counterchange border on zig-zag lines Intermittent border pattern .... Block borders equivalent to the ch'ecouer pattern
IN all-over design
Alternating features
Counterchanging geometric borders
XIX
PAGE 219
219 220 220 220 2 21 221 222 222 222
223 223
223
224 224 224 225 225
226
226
227 227 227
228
XX LIST OF ILLUSTRATIONS.
NO. PACK
267. Block border 229
268. Block or panel border 229
269. Continuous lines as it were disappearing behind
UPRIGHT features 229
270. Oriental use of diapered interspaces . . . 230
271. Cresting pattern 231
272. Cresting with CO shaped scroll 231
273. 00 shaped scrolls on zig-zag plan .... 231
274. 00 shaped scrolls, on zig-zag plan, crossed by them-
SELVES 231
275. Border design turned over in each direction . 232
276. Border growing alternately from either side . 232
277. Fringe border . . 233
278. Diagrams showing the plan of the border in
relation to its DIRECTION 233
279. Frame design, the pattern turned over but other-
wise NOT repeating 234
280. The border in relation to the corner . . . 235
281. Breaks in the border 237
282. Book cover in which border and filling are
inseparable 242
283. Central ornament growing out into border . . 243
284. Jacobean panelling — more or less a geometric
diaper 244
285. A more or less geometric diaper, resolving itself
into a border — Roman pavement .... 246
I. WHAT PATTERN IS.
Pattern not understood — The meaning of the word— Comes of repe- tition, and is closely connected with manufacture — -Has alvvaj's a geometric basis — Use and necessity of system in design — Lines ine- vitable, and must not be left to chance.
To readers of a book upon the subject, no apology for pattern is necessary. Modest as may be its pretensions to artistic consideration, it covers ground enough to command attention. It is here and there and everywhere about us. There is too much of it by more than half — and more than half of it is of such a kind as to make the discriminating wish they could do without it altogether. Still, there it is; and there is no escape from it.
If folk knew a little more about it, realised what was and what was not within the control of the desigrier, under- stood how pattern came to be, and something of its scope and purpose, as well as of the processes through which a design must pass before ever it comes (for their momentary delight or lasting annoj'^ance) to be produced, they would be less at its mercy. For the difficulty of designing is by no means in proportion to the importance of the field of design; and in the case of repeated pattern, with which we have mostly to do — even those of us who are not concerned with trade or manufacture -the invention it requires is in inverse ratio to the free scope afforded. It is easier, as WilHam Morris confessed, to design a big hand-made carpet, in which the artist is free to do very much as he hkes, than to plan a small repeating pattern to the \\idth of Wilton pile or common Kidder-
2 PATTERN DESIGN.
minster. The art of pattern design consists not in spreading yourself over a wide field, but in expressing yourself within given bounds.
The very strictness of such bounds is a challenge to inven- tion. In the realm of appUed design manufacture is autocrat, and the machine is taskmaster. Let who can rebel against their authority. For those who cannot — and they are the great majorit^^ — revolt is futile. We are all of us, artists no less than the rest of the world, dependent upon manufacture; and those of the title who stand aloof from it give ground for the accusation, commonly brought against artists, of being at best unpractical and wrong-headed. Their sense of fair- ness is at fault, too, in blaming manufacture because it falls short of art, while they stand by and refuse a helping hand to the makers of things which will be made, and must be made, and made by machinery too, whether they like it or whether they do not. It rests with those who have some faculty of design (their name is not legion) to come to the aid of manufacture, which, without help from art, is given over to the ugliness which they deplore.
Pattern, it seems plain, and repeated pattern, conforming to the conditions of manufacture and even to mechanical production, is a consideration of importance, not merely to manufacturers and others engaged in industries into which art may possibly enter, but to all whose comfort and well- being depends in any degree upon the beauty and fitness of their surroundings.
The word " pattern " is here used in a somewhat technical sense— not, as the dictionary has it, to mean " a specimen " nor yet " a shape or model for imitation, " but ornament and especially ornament in repetition. Pattern is, in fact, the natural outgrowth of repetition; and in ever}'^ case the Unes of its construction may be traced; they pronounce them- selves, indeed, with geometric precision. Geometric pattern grew, of course, out of primitive methods of workmanship.
WHAT PATTERN IS. 3
No mechanism so simple but it gives rise to it. To plait, to net, to weave, or in any way mechanically to make, is to produce pattern. The coarser the work, the more plainly is this apparent — as, for example, in the mesh of a coarse canvas; but, though refinement of workmanship may be carried to the point at which, as in the finest satin or the most sumptuous velvet, warp and weft are not perceptible to the naked eye, the web is always there, and forms always a pattern. The pride of the mechanist is to efface such evidence of structure. To the artist it adds an interest; and, far from desiring to obliterate it, he prefers frankly to confess it, and to make the best of the texture or pattern which a process may give. He regards it as a source of inspiration even, which to neglect would seem to him wasteful of artistic opportunity.
It is to his determination to make the best of whatever may naturally come of any way of working that we owe much of the simplest and most satisfactory, if not absolutely the most beautiful, pattemwork.
So infalUbly does the repetition of simple units, resulting no less from elementary processes of handwork than from mechanical production, end in pattern, that wherever there is ordered repetition there it is. Take any form you please and repeat it at regular intervals, and, as surely as recurrent sounds give rhythm or cadence, whether you want it or not, you have pattern. It is so in nature, even in the case of forms neither identical nor yet recurring at set intervals. The daisies make a pattern on the lawn, the pebbles on the path, the dead leaves in the lane; the branches of the trees above, the naked twigs against the sky, the clouds that mottle the blue heavens by day, the stars that diaper their depths by night, all make perpetual pattern. The grain of wood, the veining of marble, the speckling of granite, fall so obviously into pattern that they have been accepted in place of intelUgent design. The surface of the sea is rippled.
4 PATTERN DESIGN.
as the sandy shore is ribbed, with wind-woven device, the rock? are covered with shelliish clustering into pattern. Your footprints, as you walk, make a pattern on the pattern of the dewy grass; your breath upon the window-pane crystallises into pattern.
Technically speaking, however, we understand by pattern not merely the recurrence of similar forms, but their reciu"- rence at regular intervals. The Japanese rendering of
1. PEACH-BLOSSOMS ON THE ICE — JAPANESE.
peach-blossoms on the surface of thin ice is undeniably ornamental. It may be regarded as part of a pattern, but, to be complete, it should repeat, which here it does not.
It must not be inferred from the casual occurrence of what is called pattern that there is anything casual about design: the very name denies that it is so.
The artist's hand does not crawl aimlessly over the paper and trail behind it flowers of the imagination. There is scope in ornament for all the fancy of a fertile brain; but
WHAT PATTERN IS. 5
design is no mere overflow of a brimming imagination; it is cunningly built up on lines necessary to its consistency, laboriously, it might be said, were it not that to the artist such labour is delight. Whoever finds it irksome may be sure his bent is not in the direction of applied design.
The main hues on which repeated ornament is built are so few and simple that they can quite easily be traced. Just as the man of science divides the animal world into families and classes, so may the man of art classify pattern according to its structure. He is able, no less than the scientist, to show the affinity between groups of design to all first appearances dissimilar, and to lay bare the very skeletons upon which all possible pattern is framed.
The idea of setting out to design a pattern without regard to its logical construction is contrary to reason. It is all very well to protest that art is free of laws: they govern it none the less. And the pattern designer is bound to reckon with the dry bones of design. With regard to the unit of his design he is free; he may, if he will, throw taste to the winds; but when the pattern comes to be repeated, the very order of its repetition reveals the skeleton; it was in the cup- board all the while.
This insistence upon the geometric basis of design may seem Hke dogmatism; and all dogma cuts two ways, irri- tating the student into opposition where it does not con- vince him; but experience will prove to him that the way to avoid the appearance of formality is not to set to work at haphazard. Suppose one were to begin without any thought of formal distribution and to design, let us say, a scroll, in itself as graceful as might be. A series of such scrolls, side by side, would show lines not in the least con- templated by the draughtsman, and in all probability as inelegant as they were unexpected. Who has not suffered in his time from wallpaper or other patterns in which certain ill-defined but awkward stripes would thrust themselves
6 PATTERN DESIGN.
upon his attention? And to the designer himself one of his strangest experiences is the trick a seemingly quite innocent pattern will play upon him in repetition. A design, for example, which appears to be quite evenly dis- tributed will run, when hung, into lines which slant in such a way as to give the impression that the walls are not true, or that the paper has been hung askew.
In a pattern in which patches of the ground are left bare, the gaps are by no means accidental. They are most
2. PREDETERMINED GAPS IN A PATTERN.
deliberately planned — and from the very beginning — or there is no knowing what havoc they might play in repetition.
Amateurs will tell you (and a painter is an amateur when first he tries his hand at pattern) that the lines which are so distressing in incompetent pattern are the result of mathematical planning. That is not merely false, but, as every practical designer knows, the very opposite of the truth. There is no more radical mistake than to suppose that the awkward stripes which come out for the first time when a pattern is repeated are the result of the designer's having worked upon the obtrusive lines :they are the natural
WHAT PATTERN IS.
3. BALANCE OF ORNAMENT ENOUGH FOR A PANEL BUT NOT FOR A REPEATING PATTERN.
8 PATTERN DESIGN.
and inevitable result of not working upon any lines at all. If you work \nthout a system the only safety is in insignifi- cance. A pattern may be comparatively featureless; and, so long as there is in it no feature pronounced enough to dis- tinguish itself, lack of order may perhaps pass unnoticed But it is hardl}- worth while going out of the way to secure an end so insignificant.
A design of any character has usually in it features which, when it is repeated, stand prominently out from the rest. To these the eye is irresistibly drawn; and, not merely so, but the lines they take in relation one to another insist on being seen. It is barely possible that, in the event of such lines not having been taken into consideration by the designer, they should fall together in the happiest con- ceivable way. More likely they will look awry.
The balance which in a single composition satisfies the eye is not enough when it comes to repetition. The shoulders of the mantling, for example, on page 7, one rising above the lion's back, the other falling below it, would in repeated scrollwork almost certainly give the impression of being out of the level. The only way to be sure that scroUs in repetition will balance is to begin by disposing them, as in the wall- pattern opposite, quite symmetrically.
The designer of experience runs no unnecessary risk. Accepting some sort of geometric plan as the basis of his design, and appreciating at their worth the severity and strength resulting from it, and the sense of scale it gives, he makes sure of lines deliberately fulfilling the purposes of decoration. He will counteract a tendency to stripes in one direction by features which direct attention otherwards; he will so clothe a doubtful Une that there shall be no fear of its asserting itself, as in its nakedness it might. The lines he leaves in his design were chosen for their strength and steadi- ness. Such lines as reveal themselves are the lines upon which it was built, by no means unforeseen.
WHAT PATTERN IS.
4. WALL PATTERN IN WHICH THE LINES OF THE SCROLL ARE QUITE EVENLY BALANCED.
lo PATTERN DESIGN.
If lines left to chance reveal themselves, as they are apt to do, in sequence not to be endured, what else was to be excepted? Only by a miracle could they happen to fall precisely as art would have them. The best of players makes sometimes a happy tiuke in design; but he does not reckon upon such luck.
The point is this: it is, practically speaking, inevitable that lines shall in the end assert themselves in repeated pattern; if the artist does not arrange for them in his design, they fall as maj' happen; it is therefore the merest precaution of common-sense on his part to lay them down from the beginning, to make them the framework upon which his pattern is built, the skeleton of his design.
A practical designer has not, as a rule, much difficulty in tracing the bones of a design, amply as they rnay be wrapped in foliation or other disguising detail. To lay them bare enough, howercr, to demonstrate the anatomy of pattern recourse nmst be had to dissection.
II. THE SQUARE.
Geometry the basis of all pattern — Breaks in the simple stripe give cross-lines — Hence the lattice and the checquer, on which a vast variety of pattern is built.
It will be as well, before proceeding to dissect design apparently far removed from the geometric, to show the lines which of themselves make pattern. They prove to be the basis of all pattern.
The simplest of all patterns is the stripe — a series of parallel lines in one direction. But the Umits of the mere stripe are soon reached; for any break in the repeated line or any deviation from the straight gives, by its regular recurrence, other lines in the cross direction. Gaps in a series of broken horizontal Hues (5) give vertical lines; and in the same way the points of the zig-zags mark the upright (6). Any recurrent feature between the lines gives, again, lines across, slanting or
5. STRIPE BREAKS GIVK CROSS- 0. ZIG-Z.A.G POINTS GIVE CROSS-
LINES. LINES.
PATTERN DESIGN,
|
J> |
s> |
j> |
9 |
|
|
S) |
i> |
i> |
||
|
y> |
5> |
9 |
||
|
>) |
3 |
D> |
9 |
|
|
S> |
5> |
9 |
||
|
3 |
3 |
3 |
||
|
S> |
3 |
3) |
^ |
STRIPES — ORNAMENT GIVES SLANTING LINES.
|
9 |
9 |
9 |
|
9 |
9 |
3 |
|
9 |
9 |
9 |
|
9 |
9 |
9 |
|
3 |
9 |
9 |
|
9 |
9 |
3 |
|
9 |
9 |
9 |
8. STRIPES — ORNAMENT GIVES UPRIGHT LINES.
at right angles to them as the case may be (7 and 8). We arrive in effect, as the primitive basket weaver must have arrived in fact, at cross-hnes; and upon these a vast amount of varied pattern is built up, the simplest forms of which are the lattice and the checquer (9) . They must also have been the earliest evolved by the basket plaiter. Grasses all of one colour naturally showed the lines of interweaving, the lattice (10). Grasses alternately light and dark in colour asserted their checquered masses of colour, and gave the chess-board pattern. Strips of different colours led also to more intricate pattern (11); and the width of the coloured strips had only to be varied to give all manner of plaids and tartans.
Something of the variety of pattern resulting from the closer or looser plaiting of equal strips is indicated below — as
I J
9. SQUARE LATTICE.
CHECQUER.
10. INTERLACING.
THE SQUARE.
13
II. INTERLACING STRIPS OF DIFFERENT COLOUR.
well as the further variety which comes of strips of unequal width (12 to 15).
On the regular network of cross-lines a vast number of patterns not necessarily regular may be built up, many of them suggestive of plaiting if not actually suggested by it.
Taken singly, and filled in alternately light and dark, the squares give only the chess-board pattern; taken in groups of two, or alternately of two and one (16), they begin to show possibilities in the way of upright or diagonal patterns. Grouped in threes or fives (17) they give already independent
d
DTTxrn
n
□
iz
ml
□
d
LD
u
|
: h |
1 |
n |
1 1 |
1 1 |
1 1 |
1 |
L 1 3 3 |
12. CLOSE PLAITING.
13. LOOSER PLAITING.
14
PATTERN DESIGN.
|
c |
n |
m |
-j |
||||
|
1 |
|||||||
|
1 |
1 |
||||||
|
" LJ |
LJ ■ |
- |
|
I f 1 — i — |
14. OPEN PLAITING.
15. PLAITED STRIPS OF DIFFERENT WIDTH.
units of design. The unit of the Arab diaper opposite CiS) is clearly a group of eight squares.
Working upon the lines of the lattice, we arrive, without in any way departing from them, but simply by intermitting some of them, at something in the nature of a key or fret pattern (19). The elaborate Japanese fret on page 17 is built upon that plan, upon which it will be seen (same page) all manner of interlacing and free diapers may be schemed.
There are two ways of setting about design of this sort, both of which amount practically to the same thing. The one is to rule square Unes and rub out parts of them, the other is to
16. DIAPERS OF SQUARES AND GROUPS OF SQUARES.
THE SQUARE.
15
17. GROUPS OF THREE SQUARES.
GROUPS OF FIVE SQUARES.
work \vith a pen or a brush upon faintly ruled paper — an exercise childish enough for the kindergarten, but by no means to be despised by the artist whose soul is not above pattern. It is wonderful what a vast and varied range of pattern is to be built upon the simple scaffolding of square lines.
The chess-board has only to be turned partly round to give us diamonds— but as they are still square, and it is only the point of view that is altered, we have as yet (as in the case of a diagonal stripe, which is still only a stripe) no
18. THE UNIT OF REPEAT A GROUP OF EIGHT SQUARES.
i6
PATTERN DESIGN.
IQ. FRET PATTERN ON THE LINES OF SQUARE LATTICE.
dift'erent plan, but only a difference of effect — so great, however, as to be worth noting, and of quite exceptional importance when it comes to the construction of freer pattern.
By altering the angle at which lines cross, we get at once not only a series of new shapes (20 to 22) but a variety of the diamond which for clearness' sake it will be convenient to distinguish as the diamond — not in itself so satisfactory a form as the square, but invaluable in connection with cross- hnes in a third direction.
The plan of the rectangular lattice is, however, as a matter of fact, the basis of the great part of our pattern design. There are obvious adx^antages in the use for example, of
♦♦♦♦♦♦ ^W^ ^N^
♦♦♦♦♦♦ WWW y^ mtH www w^^
20. DIAMOND DIAPER. 21. ZIG-ZAG BUILT ON 22. DIAMOND DIAPER. DIAMOND LINES.
THE SQUARE.
17
Lrj"q_rrqjj~i
Lcrq^qjjq. rb]Jiriq_rrbjjir
c
II
n
_D
rb-crh d
1-Fi-S
?
H
?
d
_n
_n
IZTL.OJ-' U HlX]
e
FTP
n p_rlin_pLrh
d
?
U'
^J-1
Lr
ri
Lr
^
LC n
?
d
n
r
^
23. FRETS, ETC., ON LINES OF SQUARE LATTICE.
1 8 PATTERN DESIGN.
square tiles and right-sided blocks for printing and so forth; but, over and above convenience, it seems to come more naturally to us to think out a design on square Hues than on anv other.
III. THE TRIANGLE.
The square lattice crossed by diagonal lines gives the triangle — Hence the diamond — And out of that the hexagon, the star, and other geometric units familiar in Arab diaper.
The introduction of a third series of cross-lines makes quite a new and most significant departure. Cross a square lattice by a series of diagonal lines bisecting the right angles, cutting the squares in half that is to say, and the halves give us a new- form to work upon — the triangle (24). And so it is if we cross in the same way a lattice of elongated diamond shape. But if we start with a lattice of a certain proportion; if, that is to say, the two sharp angles of the diamonds are together equal to one of the blunter angles, then, when they are bisected by a third series, the halves of the diamond prove to be equi- lateral triangles. That being so, or, to put it another way, our
— ' — — I
24. TRIANGLE. EQUILATERAL TRIANGLE.
DIAMOND.
PATTERN DESIGN.
25. DIAPER BUILT ON TRIANGULAR PLAN.
diamond being composed of equilateral triangles, new possi- bilities of design appear upon the horizon.
The equilateral triangle is the basis of an infinity of geometric pattern. The Arabs (or the Byzantine Greeks re- sponsible for their art) made infinite use of it, building up their intricate patterns upon it (25) as Western nations built upon the square. From the triangle they derive not only the diamond, which is composed of two triangles, but other in- genious combinations, such as the unit of the diaper above which is composed of seven triangles or of a central triangle and three diamonds radiating from it. Six triangles form a hexagon (20), and six other triangles ranged round that result in a star.
l6. DIAPERS BUILT ON TRIANGLES. HEXAGON = SIX TRIANGLES. STAR = HEXAGON H SIX TRIANGLES.
THE TRIANGLE.
21
27. DIAPERS BUILT UP OF HEXAGONS.
The hexagon itself is a unit which makes a perfect repeat. The stars ma}' be so arranged, point to point, as to leave only hexagonal intervals between. In the diaper on page 20 the central hexagons round which the stars are
28. INDIAN LATTICE AND THE LINES ON WHICH IT IS BUILT,
PATTERN DESIGN.
29. IXDIAX LATTICE AND ITS TRIANGULAR BASIS.
built range with the hexagonal intervals. The result is a starry pattern, composed of hexagons and triangles, in which the cross-lines in the three directions are very plainly marked.
Three hexagons together give a figure (27), commonly employed in Arab ornament, which repents (it will be seen) either as a close-fitting diaper or with hexagonal intervals between. The figure itself is plainly related to the triangle — the sides of which might easily be bent into zig-zags giving the other nine angles which go to make it.
"" The friendly way in which triangles, hexagons, stars and other shapes compounded of the triangle, unite to give com- plex and ingenious variety of pattern, accounts for the per- sistent use of such units in Byzantine floor patterns and Moresque tile-work. It will be seen that the intricate Indian
THE TRIANGLE.
23
30. INDIAN LATTICE AND ITS TRIANGULAR BASIS.
window lattices illustrated (28, 29, 30), resolve themselves (apart from the rosettes or stars inhabiting the central hexagons) into shapes either formed by lines crossing in the three directions or built up of equilateral triangles, in the ways already explained.
nxrpTixi.. ,0., B...
31. OCTAGON AND THE- LINES ON WHICH IT IS BUILT.
IV. THE OCTAGON.
Four series of lines give the octagon — Not the unit of a coraplete pattern, but the basis of some radiating patterns — More complicated cross-lines, giving sixteen and eighteen sided figures, result in more elaborate pattern, but involve no new principle— Pentagon pattern really built on simple trellis lines.
A FOURTH series of cross-lines naturally gives new shapes, but no longer shapes which of themselves form a compact repeating pattern (31). The little squares at a are plainly formed on the trelUs shown in dotted line. Cross that with a wider trelHs, b, and you get the octagons, c. But, pro- portion them as you will, there remains always a series of square or oblong or diamond-shaped gaps between; and the two forms are together necessary to make a coherent pattern. The two octagon diapers opposite (32) are of course identically the same, presented only from different points of view.
The lines of a double trellis, one crossing the other, cannot therefore be said to give a new geometric pattern unit; but they give new Unes on which a vast number of radiating patterns are built, from the comparatively plain interlacing of a cane-bottomed chair (33) to the ornamental diapering on the lower walls of the Alhambra (34). II is upon lines such as these that the bewildering patterns in which a race of mathe-
THE OCTAGON.
32. DIFFERENT VIEWS OF SAME PATTERN.
maticians delighted are based. We are inclined to wonder sometimes how human ingenuity comes to grasp the thread of such intricate patternwork, much less invent it. Some- thing of the wonder ceases when the tangled lines of its con- struction are unravelled.
The double lattice which gives the eight pointed figures in diagram 34 might be crossed by a similar lattice giving sixteen pointed figures, and that again by itself, giving figures of twice as many points, without the introduction of any new principle of design. The lines would merely be elaborated; they would resolve themselves into cross-lines not in four or eight, but in sixteen or thirty-two, directions — that is all.
33. PATTERN FROM CANE-BOTTOMED CHAIR.
26
i
f^:^
PATTERN DESIGN
i
pK' -^
mi :.:..... _ ,
.np^
1#
\^,:.,i
34. TILE WORK FROM THE ALHAMBRA.
Practically, we have now exhausted the plans on which straight-Uned patterns can be schemed. It is within the bounds of ingenuity to put together right-lined figures, such as the pentagon, in a way to mystify even the expert and to suggest a new discovery in pattern planning. Such a pattern is the cunningly counterchanged pentagon diaper opposite (36) . The star shapes within the larger pentagons and the smaller stars between, with their pentagonal centres suggesting other pentagonal shapes (in point of fact not complete) , confuse the Hues on which the pattern is really built. But, artfully as
THE OCTAGON.
27
35. SIMPLE AND MORE COMPLICATED TRELLIS LINES.
they are disguised, they prove to be famihar hnes involving no new plan. The pattern might be built indeed upon five series of lines in the direction of the five sides of the pentagon; but a network of such lines would be involved to a quite perplexing degree. The lines indicated on the diagram are more likely those on which the artist worked; and they are intimately related, as will be seen, to the simple trellis of cross-lines above.
36. PENTAGON DIAPER AND ITS SKELETON.
37- ZIG-ZAGS DEVELOPING INTO WAVE LINES.
V. THE CIRCLE.
The circle gives no new plan but only curvilinear versions of the fore- going— The wave a rounded zig-zag — The honeycomb compressed circles — Segments of circles give scale pattern, a curviUnear v-ariation upon diamond — The ogee— The circle itself a scaffolding for design.
A MOST important element in geometric pattern is the circle: with it curviUnear design begins at once to flow more freely.
But (as in the case of the pentagon) the circle gives us no new plan to work on; it must itself be planned upon one or other of the systems already described; it must be struck, that is to say, from centres corresponding to the points of inter- section given by a lattice of straight lines.
Curvilinear pattern is in its simplest form plainly only a suaver variety of rectilinear design. Flowing patterns can often be deduced from angular, and vice versa. The priority of either is open to dispute, but hardly worth disputing.
Long before geometric principles were formulated in the mind of man he practised them intuitively. As to the use of the circle in ornament, we need not ascribe it to geometry, nor trace it back to the sun's disc and symbolism, nor yet to conscious imitation. The primeval artist had but to pick up
THE CIRCLE.
29
38. RELATION OF OCTAGON TO CIRCLE DIAPER.
the nearest dry twig and indent the damp earth with it, and lo! a diaper of circular forms. Or, again, he might begin to scratch zig-zags, and, as his hand flowed on, they might de- velop into waved lines (37). Wave or zig-zag lines fall na- turally into stripes: it is not the plan of the pattern, but only the detail, that differs. The wave is in fact a zig-zag blunted at the points, the zig-zag an angular form of wave. A net- work of straight staves gives, as seen in slight perspective (look at any common hurdle), distinctly waving lines.
39. STRAIGHT-LINED AND CURVILINEAR VARIETIES OF THE SAME PATTERN.
30
PATTERN DESIGN,
40.
DIAPER OF CIRCLES PLANNED ON SQUARE LINES.
41.
DIAPER OF CIRCLES PLANNED ON DIAMOND LINES.
Again, round off the corners of the hexagon or octagon, and 3'ou have straightway a circle (38). Indeed, at a Uttle distance, the Hnes of a sixteen-sided figure round themselves, to all appearance, and give the effect of a circle. The reduc- tion of the circle to hexagonal shape is practically effected in the honeycomb. The busy bee, if one may so far throw doubt upon his proverbial forethought, works blindly in a circle, and the shape of his cells is simply the result of gravitation.
42. PLANNED ON LINES OF 40, BUT RADIUS RELATIVELY WIDER.
43. PLANNED ON LINES OF 4I, BUT RADIUS RELATIVELY-WIDER.
THE CIRCLE.
31
(ViViVO () [) [
44. PLANNED ON LINES OF 4I, BUT RADIUS RELATIVELY WIDER.
Cylinder crowded together crush themselves into hexagonal prisms. There is not a question of design : it is a matter of plasticity and weight.
It is clear on the face of it that we have in illustration (39) not so much two patterns as straight-lined and cur\'ed va- rieties of the same thing.
45. DOUBLE DIAPER OF CIRCLES PLANNED ON LINES OF 40, BUT RADIUS RELATIVELY WIDER STILL.
32
PATTERN DESIGN.
46. DIAPER OF INTERLACING CIRCLES.
Circles closeh' packed to form a diaper show by the shape of the interspaces the plan on which they are put together. Arranged on the square (40) they show between them a four- sided space, on a diamond (41) a three-sided.
Larger circles struck from the same points, or circles of the same size struck from similarly arranged but closer points,
ai,d:tN_/JriN_zHN_/ti
-J -^J
47. DIAPER OF CIRCLES INTERLACING WITH SQUARE TRELLIS.
THE CIRCLE.
33
48. DIAPER OF SCALES AND DERIVATIVES FROM IT.
give, it will be seen (42, 43, 44, 45), more intricate-looking diapers. And infinite variations may be played upon the same tunes. The dotted lines in the diagrams fully explain the construction of the patterns, and show that no new principle is involved in the planning of them.
In the pattern (46) on page 32, the larger circles are struck from the points given by equi-distant lines crossing at right angles, and the smaller circles from points midway on the lines between these. The combination of straight with curved lines (47) helps only to show more plainly than ever the scaffolding on which the pattern was built. Other lines (square
34
PATTERN DESIGN.
49. SCALES TURNED ABOUT TO MAKE FLOWING DIAPER.
always) on which it might have been constructed are indicated By dotted Unes.
As with the circle so with its segments and its compounds (the trefoil, quatrefoil, and so forth); they give new forms, to be arranged always on the old plans — the quatrefoil naturally upon square lines, the trefoil or sexfoil upon the lines of the triangle. The segments of the circle give us the scale pattern (48), derived it might be from the scales of a fish or from the plumage of a bird's neck, but, practically speaking, only a translation of the diamond into curved lines.
Regarding the scales only as curvilinear diamonds, we are free to turn them about, as neither scales nor feathers would naturally grow, and to produce a flowing diaper (49) in which occurs a form, compounded of four scales, which itself may be regarded as a version either of the diamond or of the hexagon. All this is more plainly shown in diagram 50. The flowing shape occurs again in yet another diagram (51) together with the waved lines out of which it is composed. The inter-
THE CIRCLE.
35
50. SHOWING RELATION OF SCALE TO DIAMOND AND OGEE, AND OF OGEE TO DIAMOND AND HEXAGON.
XX>'
51. WAVE LINES, OGEE DIAPER, AND INTERLACING OGEES, GIVING HEXAGONAL SHAPES.
36
PATTERN DESIGN.
52. GOTHIC TRACERY DIAPERS CONSTRUCTED ON CIRCULAR LINES.
lacing ot these waved lines gives a six-sided figure, the lines of which only want straightening to be recognised at once as the familiar hexagon. It is not surprising that in the fifteenth and sixteenth centuries pattern commonly took the graceful lines of the ogee. The designer fell into them as naturally as Mr. ^^'egg dropped into poetry.
53. GOTHIC TRACERY DIAPER AND ITS CONSTRUCTIONAL LINES.
THE CIRCLE.
37
The circle deserves further to be considered because it is itself the scaffolding, or at all events forms part of the scaffolding, upon which a great number of more or less geometric patterns have been devised.
The simple Gothic diaper (52) to the left on page 36
54. GOTHIC TRACERY AND ITS CONSTRUCTIONAL LINES.
shows not only circles but forms into which the segments of smaller circles enter. One circle, it seems, begets others.
The tracery pattern next to it (52) is constructed by the help of small circles, themselves arranged on a circular (or hexagonal) plan, though what they give in the result is a sort of vertical wave pattern.
PATTERN DESIGN
55. GOTHIC TRACERY DIAPER AND ITS CONSTRUCTIONAL LINES.
The rather more elaborate design at the bottom of the same page (53) repeats upon the Hnes of a hexagon, the points of which correspond with the centres of star-shapes. But the six of these enclose another star and the hexagon is seen to be a compound unit — of which the component diamond repeats as a drop. The small circles drawn within these dia- monds at A, give by the mere effacement of a portion of them the twisting shapes at b, which only remain to be subdivided as at c, and the skeleton is complete. It is once more on in- tersecting circles that the pattern (54) on page 37 is set out;
THE CIRCLE. 39
and the points of intersection give, as there shown, the points of the starry rosettes.
Yet another Gothic tracer}- pattern is gi\'en (55) , planned this time on the lines of the double square. One half of this is a turnover of the other. It works as a drop repeat, and shows plainly that at an early stage the circle entered into its construction. Here again the stages by which it might possibly have been reached are indicated; but that is not to say with any certainty the designer may not have approached it from another direction.
The limited variety of skeleton upon wliich pattern is built, is nowhere more plainly shown than in the way in which, in the maze of design, we find ourselves, no matter on what path we set out, arriving over and over again at prec- isely the same point.
56. STRAIGHT-LINED AND CURVILINEAR VERSIONS OF DIAPER.
VI. THE EVOLUTION OF PATTERN.
Various starting points for the same pattern — Six ways in which it might have been evolved — The construction of sundry geometric diapers — Influence of material upon design — Some complex lattices.
It is not safe to pretend to say with authority, the way in which a given pattern was evolved; there are usually several ways in which it might have come about. Close up the waved lines in diagram 51 and they give you an ogee diaper. Open out the ogee diaper and it gives you waved lines. The starting point of the interlacing pattern might equally well have been waved lines, the ogee, or the idea of a net.
In diagram 56 (straight and curved lined versions of the same thing) the dotted Unes may be taken to indicate the way the larger unit was built up of four smaller units forming in themselves a repeat; but the smaller unit would result equally from, as it were, crossing the larger pattern b}/ itself. The dotted diaper is the same as that in solid lines. Together they give a smaller diaper which may or may not have been its origin.
There are at least half-a-dozen ways in which a simple
THE EVOLUTION OF PATTERN. 41
57. BEGINNING WITH RECTANGULAR TRELLIS.
star pattern such as that above might possibly have been arrived at: —
1. By beginning with a diamond lattice and occupying the spaces with four-pointed stars (57).
2. By arranging either star-shapes or diamonds point to
58. BEGINNING WITH FOUR-POINTED STARS.
42
PATTERN DESIGN.
59. BEGINNING WITH CROSS BANDS OF DIAMONDS.
60. BEGINNING WITH OCTAGONS AND STARS.
THE EVOLUTION OF PATTERN.
43
point, and drawing diagonal lines across, between the stars, or through the diamonds (58).
3. By starting on diagonal lines, crossing a row of dia- monds by a similar row in the cross direction, and steadying the diamonds by giving them a backbone (59).
4. By beginning with a diaper of octagons and foiir-
61. BEGINNING WITH ZIG-ZAG LINES.
pointed stars, crossing that by itself, and adding hnes to steady the effect (60).
5. By beginning with zig-zag lines, crossing them by similar zig-zag lines, and crossing the pattern thus produced by itself (61).
6. By starting with the eight-sided unit — by no mccins necessarily arrived at on the Hnes of diagonal zig-zags — and crossing it by itself (62).
In the two last-mentioned diagrams the long diagonal
44 PATTERN DESIGN.
backbone lines are for the sake of clearness omitted, and the unit of a repeat common in Oriental diaper-work is emphati- call}' expressed. A very similar unit occurs in the diaper opposite (63), which may be built up, as will be seen, in the simplest possible way, on the basis of the square. The squares have only to be crossed by X lines alternately in transverse direction, and the framework of the design is there. Another
62. BEGINNING WITH EIGHT-SIDED UNIT.
simple way of producing a very complex result (64) is shown on page 46 (65) , where the design is resolved into a compara- tively simple square pattern crossed by itself. The construc- tion of the Arab lattice pattern (64), shown in single line to the left of diagram 65, is better explained by diagram 66, from which it will be seen that the unit is merely a square enclosing a small diamond, the sides of which diamond are successively continued to the corners of the square.
THE EVOLUTION OF PATTERN.
45
63. COUNTERCHAXGE DIAPER AND ITS CONSTRUCTION.
The same diagram will help to explain the construction of the patterns shown at 67 and 70.
Given the base lines to the left of diagram 66, and the desire to counterchange the colour, what is the designer to do? It is easy to make the figures in one direction dark and in the other light; but there remain the small intermediate diamonds which can obviously be neither one nor the other. By effacing the diamond, however, and joining the loose ends of the lines to the left of the diagram in the way shown to the right of it, the difficulty is at once overcome and we get the diapers (67) on page 47.
64. ARAB LATTICE PATTERN.
46
PATTERN DESIGN.
65. CONSTRUCTION OF ARAB LATTICE 64.
A similar difficulty is got over much in the same way in another Alhambresque tile pattern (70). The pointed pro- jections at the sides of the oblong shapes corresponding abso- lutely to indentations at the ends of them, the parts fit toge- ther perfectly, except for the small square spaces between. By dividing these into four triangular parts, alternately light and dark, the patterns are made to counterchange. And the expedient is just what would occur to an artist building up his pattern, as the Moors did, out of shaped pieces of tile.
The kind of kev or swastika seen in diagram 66 occurs
66. DIAGRAM EXPLAINING THE COMPOSITION OF THE UNIT IN PATTERNS 64 AND 65.
THE EVOLUTION OF PATTERN.
47
67. ARAB COUNTERCHANGE PATTERNS CONSTRUCTED AS 66.
also in diagram 68 which is in some sort a key to the con- struction of the Roman pavement pattern 69. The most likely way of setting about such a design would be to divide up a small square with swastika lines (as shown to the right of diagram 68) and to reverse the unit in either direction as shown b}^ the arrows to the left — where the dotted lines of the diamond are given. The two together (nearer the centre)
|
/ / |
^"-.^ |
^'' |
\ \ |
|||||||
|
1 / |
\ |
i |
\ \ |
|||||||
|
/ |
1 |
\ \ \ \ \ |
L |
\ |
1 ■■ |
|||||
|
i ^^"-. |
||||||||||
|
/ v~~:; |
t ^.^-' |
\ \ \ \ \ \ |
/ / / / / / |
^ |
||||||
|
1 |
1 n |
1 |
r |
\ n |
||||||
|
y'^ |
1 |
68. KEY TO CONSTRUCTION OF ROMAN PAVEMENT PATTERN 69.
: :
■ ■ ■■■■••■•■
••'..
•••
«•••■ * \ • s •«««■•■•■••>■•••■•■«■••■■••
• * ■
• ■««a«aaaaaa«a«*t •«%*•••■• a • ma ••aaaaaa«a»aaaa«*aaaaBaa
a • * m * m ^ ^ « •
• •
A
.*••■•••■■ ^ *. •••t»aaa \ .* ••at*»«i« / % aaaaaaaaa
; • / \ : : V : s •* ♦ I •
• •••a ! * ^ * ••aaaaaaaaaaaaaaaaaaaaoaa t ^ a a aa»«<
69. KOMAN MOSAIC PAVEMENT PATTERN.
THE EVOLUTIOX OF PATTERN.
49
IMIZI
70. ALHAMBRESQUE COUXTERCHAXGE PATTERN.
give, in the space of four squares, the complete compound unit, which repeats on the hues of the larger square.
A very broad hint as to the lines on which a designer actually went to work is sometimes given by the nature of the work on which he was engaged Working in tesserse, a mosaicist would naturall}- start with Hues — which, somehow, thin as they may be, never look mean in a pavement. Work- ing with triangular-shaped blocks he would as naturally fit together the parts of his pattern puzzle-wise.
Speaking as a practical pattern designer, and one who finds it most amusing to devise merely geometric pattern, I am strongly disposed to believe that the elaboration of Oriental pattemwork (which resolves itself at last into a net- work of lines not easily to be disentangled) comes of the practice of building up designs out of little triangular-faced pieces of marble, glass, or tile. '
The intricate Hnes develop themselves as the artist pro- ceeds; and, having got them, he goes on to emphasise them. Carvers and others translate the tile pattern into line, as in the natural course of their craft they must, and in that way
50 PATTERN DESIGN.
we get the cunningly intersecting line work of pierced lattices and so forth (71). Designing upon the lines themselves he would get caught in the meshes of his own pattern, and lose the sequence of line so difficult to keep in view.
It dazzles one to think of the plan of no more elaborate a lattice than that below. The simpHfied unit of the design is shown in dotted lines. It might equally have been built up upon the lines of either of the diagrams opposite (72-73), octagons point to point, with four-pointed star-shapes
71. LATTICE PATTERN, POSSIBLY DERIVED FROM TESSELATED WORK.
between. And in this particular case it is not clear that any great advantage would have been gained by building up the pattern; but where the component shapes are triangles or compounds of triangles, it is not only easier to play with them than with the involved lines of a complicated lattice, but much more fun to do so.
The complexity of the lattice (74) on page 52 is less puz- zhng when one realises the plan of it, a squat diamond, which may be divided into two equilateral triangles a and b (one a
THE EVOLUTION OF PATTERN.
51
72. LINES OX WHICH LATTICE 7 1 MIGHT HAVE BEEN BUILT.
turnover of the other) either of which may be subdivided, as the dotted lines show, into equal equilateral triangles and the corresponding diamonds. Given those lines it is a com- paratively simple thing to build up the lattice.
73. LINES ON WHICH LATTICE 7! MIGHT HAVE BEEN BUILT.
7-1. ARAB LATTICE AND THE LINES OF ITS CONSTRUCTION.
THE EVOLUTION OF PATTERN.
53
i3<^^^V^^^^^/l ij^^^^^^^^^ci ii^^^^V^^^i^C^
WmmWm
73. ARAB LATTICE AND ITS CONSTRUCTION.
One more instance. The repeat of the pattern above (75) (equivalent to the tile which would go to make it, or the block from which it could be printed) is indicated in the skeleton to the right, a square; but the component unit is all there in the triangle which forms only a fourth part of it. And of the twelve parts of which (as the dotted lines show) that is made up, all but one are repeated three or four times over, so that it takes figures of only four different shapes to make it — presumptive evidence at least that that is the way in which the design came about, not perhaps in this particular instance of the lattice, but in the case of some pattern which was its prototype.
VII. PRACTICAL PATTERN PLANNING.
Possible and practicable lines of pattern construction — Lines often fixed for the designer — Conditions of production affect plan — Trian- gular plan, Oriental — Rectangvilar plan, Western — Relation of oneplan to the other — of triangular and octagonal repeat to tectangular — Pos- sibilities of the diamond — Design regulated by proportions of repeat.
The lines of the square or of the parallelogram, of the diamond or of the triangle, are naturally, as a glance at our illustrations \\dll show, conspicuous in geometric design; and, even where they do not make up the pattern altogether, the}/ constantly make part of it. The design opposite (76) is, for example, in its main lines only a translation into waving lines of the simple hexagon and star pattern on page 20. But these same square, diamond, and triangular lines underlie also repeated pattern of the freest kind. And it is because they are the basis of all repeated pattern, that it behoves the designer to acquaint himself with certain simple geometric principles, as indispensable to him as a knowledge of superficial anatomy to the figure-draughtsman.
The fact is that to many arts, or to proficiency in them, and certainly to proficiency in pattern design, there goes a modicum of science — without which the merely practical con- ditions imposed by the necessity of repetition, and especially of repetition within a given area, are hardly to be overcome.
The art of the pattern designer is, not merely in devising pretty combinations of form, but in scheming them upon lines not of his choice at all, mapped out for him, on the contrary, by the conditions of his work, by no means always those which he would have chosen for himself as the most promising.
PRACTICAL PATTERN PLANNING.
55
76. CURVILINEAR DEVELOPMENT OF DIAGRAM 26.
His task is to get beautiful results out of no matter what unpromising conditions. Then indeed he may claim to be an artist.
I have been at some pains to lay down the lines on which pattern may possibly be constructed, but the possible lines are not in all cases practicable. Conditions of production have to be taken into account, and they affect not merely the character of design (that is a subject discussed at length in another volume) but its plan also.
A necessary preliminary to design is the determining of the lines on which it shall be distributed — to plan it, that is to say. The possible lines are few, and the more clearly the artist reahses what thev are, the easier it will be for him to determine which of them are available, and the one it is expedient to adopt. Mechanical conditions or practical con- siderations mav so limit his choice that he has no alternative, and it is mere waste of time to do anything but proceed at once upon the inevitable course.
56 PATTERN DESIGN.
The system upon which of old the mathematicallv-minded Oriental craftsman built up, out of the simplest units, elabo- rate schemes of ornament, encouraged the use of triangular lines as the basis of his design. The more practical and expeditious habit of the Western manufacturer leads him to work more often upon rectangular lines, and compels the designer to abandon the triangular basis, except in so far as triangular units can be made to conform to rectangular repetition (see below). The designer for manufacture, there- fore, is restricted as the handworker is not. He works, how- ever, on the old lines still — manufacture following constantly in the footsteps of handicraft.
We put down early methods of design to tradition. But traditions grew out of ways of working; and we lind ourselves to-day using expedients of design which, if they had not re- sulted from the simple contrivances of elementary handicraft, would most certainly have been evolved out of the more complex conditions of modern manufacture. The square Hnes, for example, given us by the most rudimentary form of tapestry are equally imposed by the power loom.
\/ ^ v
V ^\ /\ /\ /2-
\/ \/ \/ \/ \\
77. DIAGRAM SHOWING RELATION OF EQUILATERAL TRIANGLES AND HEXAGONS TO SQUARE REPEAT.
PRACTICAL PATTERN PLANNING.
It is on square lines that we have mainly to work, and our design has to be considered in relation to the rect- angular repeat which the conditions of to-day determine.
The possibilities of working upon other plans are limited. To adapt, for example, the equilateral triangle or octagon to a square repeat measuring let us say 21 inches either way, is possible only on a scale which makes it not often worth doing.
Take 21 inches as the base of an equilateral triangle. It will be found to measure from base to apex 18 inches. The difference between 21 and 18 is 3, the greatest common measure of both. This gives us an equilateral triangle of 3 inches from point to point as the largest which will repeat precisely within an area of 21 inches by 21.
The accompanying diagram (77), though it represents only half that area (21 inches by 10 1/2), explains the situa- tion and proves the point. It shows also the possibilities of adapting a diamond equal to two equilateral triangles to the rectangular space. A diamond measur- ing 10Y2 inches by 6 inches repeats in the width 3^ times (which would work out only as a "drop" pattern-presently to be discussed, page 60). To repeat on the square, or as a straight-running pattern, these diamonds must needs be quartered and reduced to 5 ^/4 by 3 inches.
ixx.
<:-5cat(?21iixbe->
78. DIAGRAM SHOW- ING HOW OCTAGONAL DIAPERS WILT. REPEAT ON RECTANGULAR LINES.
58
PATTERN DESIGN.
79. INDIAN LATTICE PATTERN BUILT ON TRIANGULAR LINES.
PRACTICAL PATTERN PLANNING. 59
The hexagon, which is a multiple of the equilateral triangle, adapts itself no more readily to the square (78). Hexagons half the width of the square, though they would repeat lengthwise in it, would not be equal-sided, but of the elongated form shown at the top of the diagram. True hexagons would not fill the square, but only a space of 21 by 18 inches.
Equilateral hexagons of 6 inches from side to side would repeat 3^ times in the width (and work therefore only as a " drop "). Like the triangles, they would need to be reduced to 3 inches wide before they would repeat in horizontal order. The designer is frequently asked by inexperienced people to adapt designs to proportions which put them quite out of the question. If he is not well aware of the possibilities, and especially the impossibilities, of so doing, he is likely to waste valuable time over a task which was from the first hopeless. Few persons would realise, until failure had taught them, how proportionately small a triangle or hexagon it is which lends itself to a square repeat.
Let any one try and make the lattice pattern on page 58 repeat on rectangular hnes of given dimensions, and he will realise, as no verbal explanation can possibly prove to him, how difficult it is to think of it, even, as built up on anything but the triangular lines which are in great measure respon- sible for it.
VIII. THE " DROP " REPEAT.
Scope given by drop repeat — Designed on diamond lines — And on the square^ — Geometrically same result — Practically different patterns — Opportunity of carrjdng pattern beyond width of stuff — Brick or masonry plan — Octagonal plan — -Step pattern — False drop.
The mystery of the drop repeat is more easily explained than how it came to be a mystery at all. The root of the trouble in designing it is perhaps in the fact that the inexperienced ^^•ill not take the trouble to set out repeats enough of their pattern to show how the lines of it will come. Designers of experience do that as a matter of course — because of their experience.
The pattern of a woven or printed stuff must naturally foUow on throughout the length and across the breadth of the piece— the top edge of the design must, that is to say, join on to the bottom edge, and the one side on to the other. But, whilst it is obvious that the pattern must follow in a conti- nuous line throughout the length of the stuff, it is not a matter of necessitv that it should be designed to take the same level when the strips come to be sewn together or hung upon the wall. They have to tally — that is all. The pattern may just as well be schemed to " drop " in the making up or hanging.
It is quite possible to design a pattern ^^'hich shall repeat both on level lines and as a drop. The diagram opposite (80) shows that very plainly. It was drawn by Professor Beresford Pita to explain how some wall-papers of his design could be hung either way.
The planning of a drop repeat is in realit}- a verj^ simple
THE DROP REPEAT.
6i
matter, how simple may be seen in the diagram overleaf {8i), in which the upright lines mark the width of the stuff, and the squares the limits of the repeat. It will be seen that the central feature in stripe a does not in stripe b range with it, but falls miduay between two repeats: it " drops, " in fact, one half the depth of the repeat. In the third stripe, which drops again in the same way, the feature finds once more its level; in the fourth it drops again, rights itself once more in
"Blockl
'<BlocK'-^
Same 'Block-OS • but- hun^-as-drop
80. DIAGRAMS SHOWING HOW A PATTERN MAY BE DESIGNED TO HANG IN TWO WAYS.
the fifth, and so on to the end. A further effect of the drop is seen in the direction of the stem; the wave, instead of repeating itself, seems to take the opposite line, and not to follow but to be turned over or reversed.
It will be clear from this what new scope is given by the " drop " pattern.
And what applies to the strip of material applies no less to the units of a repeat within the width of the stuff. A pattern, for example, half the width of the material may drop
62
PATTERN DESIGN.
Stripe A Stripe B
8l. DIAGRAM TO ILLUSTRATE THE "DROP REPEAT. "
within its area so that, in the stuff as it hangs, the double pattern does not drop. A drop pattern one-third the width of the stuff, would hang as a drop again.
Referring once more to diagram 8i, it will be seen that, though the pattern is built upon the square, lines drawn from centre to centre of a given feature in it form a diamond; and this diamond equally with the square, contains all the parts of the pattern, and may, just as well as the square, be regarded as the unit of repeat.
The difficulty which the inexperienced have in scheming a drop pattern, would be considerably diminished if they would only accustom themselves to think of it as a question
\2^c^
82. DIAGRAM TO ILLUSTRATE THE WAY REPEATS MUST FIT.
THE DROP REPEAT.
63
of filling a diamond instead of a square. That is all it is — designing, in fact, within a trellis of diagonal, instead of vertical and horizontal, lines.
The diamond is merely a square or parallelogram turned part way round. In designing" upon the one plan an artist would probably proceed to do differently from what he would have done upon the other; but the problem is the same. The opposite sides of either pattern have to tally. The lines ending at a b (in diagram 82) must be taken up at c D, or vice versa. In the same way whatever portion of the design extends bevond the margin b d or a c must recur again ziithin
■^'
m
//i:^
<^
V
m .^=^
|
/ |
||
|
/V; |
'') |
|
|
-1 ,•■■ |
83. DIAGRAM TO ILLUSTRATE DIFFERENCE OF MEASURE WHEN A SQUARE REPEAT IS TURNED ROUND TO WORK AS A DROP.
the margin on the opposite side, no matter whether of square or diamond. The pattern has only to join on and tit.
It should be observed, however, that if a pattern designed upon square lines is turned part way round, though it repeats as a drop upon the lines of the diamond, the measure of the repeat (as manufacturers reckon it. from top to bottom and from side to side) is no longer the same. If, that is to say, the squares in a of diagram 83 measure 5 in. each way, the diamonds in b (though it is the same pattern drawn to the same scale) measure 7 in. from point to point; and that is the trade way of reckoning, with which it is advisable to fall in.
WIDTH'
84. DIAGRAM TO SHOW A PATTERN MAY BE CONTAINED EQUALLY WITHIN SQUARE AND VARIOUS DIAMOND LINES.
THE DROP REPEAT. 65
It is sometimes quite worth while, bearing this danger always in mind, to start a design {S^, c), not the full width of the stuff, but on a diamond seven-fifths of its width. This would work out as a drop pattern of the right width twice as long as it is wide; but the start would suggest a pattern which could never ha\'e occurred to any one working within the narrower strip. The expedient is useful especially in designing floor or ceihng patterns in which the direction of the design is not meant to be obvious.
It would be difficult to overestimate the value of the drop repeat, or of the diamond plan upon which it is commonly but not invariably devised.
Mechanically speaking, there is no reason why the design should drop just half its depth. It might drop any distance; and there are occasions when a pattern which drops a third is extremely useful. But if the drop were very slight, say only one-sixth of its depth, it would take six repeats before the design righted itself; and, moreover, the recur- rence of any pronounced feature in it would be apt to mark a diagonal line. A stepped pattern has naturally a tendency that way. One great use, indeed, of the diamond plan is, that it minimises the danger of horizontal stripes, likely always to occur in a pattern repeating on the same level. In the case of a drop of one-half, the stripes go, as it were, alternately up and down, and give zig-zag lines, if any, or perhaps the lines of a trellis, to which there is little or no objection.
Another condition materiall}- affecting design is the area of the repeat, the proportions of which are determined often by conditions quite beyond the control of the artist.
It was shown in diagram 81 that a drop pattern designed upon square lines was contained also within the Hues of a diamond. In the same way a pattern designed within the lines of a diamond is contained within the Hues of a rect- angular figure working as a drop. Many drop patterns are designed upon the lines of the diamond. They may be
85. TWO PLANS UPON WHICH THE SAME DESIGN MIGHT HAVE BEEN
BUILT.
THE DROP REPEAT. 67
designed equally within the lines of a diamond two sides of which run parallel with the width of the fabric. It will be seen in diagram 84 how the pattern there given in skeleton is contained equally within the square, the squat diamond, and the diamond formed by upright and slanting lines. It will repeat, that is to say, as a square pattern occupying the width of the material, as a diamond, and as a pattern of the width of the material cut as it were upon the slant.
In theor}^ the design might have been started upon any one of these plans. In practice such a pattern would have been more likely to have resulted from working upon the lines of the diamond: as a point of fact it did result from it.
The design opposite (85), actually planned upon the diamond, might possibly have come about upon rectangular lines ; it would certainly not have resulted from working upon the diagonal lines shown in diagram 84.
Such a pattern, on the other hand, as Com. Walter Crane's (86) was clearly built upon the upright lines given by the width of the repeat (shown in dots) and hues across (from left to right) meeting them at the points emphasised by the puff balls of the dandelion. The sweep of marguerites, plainly the leading feature of the design and perhaps the start of it, falls very comfortably within the slanting shape, which seems almost to have suggested the composition of the flowers, evidentl}' planned to take their graceful line, and afterwards provided with stalks. *
Mechanically it all amounts to precisely the same thing. You have but to snip off the two opposite corners of the square to the left of diagram 87, and shift them to a position beyond the lines of the square, and they give you the obUque shape. You have but to snip off the four corners and arrange the pieces on either side of the remaining hexagon, as shown to the right of the diagram, and they give you the squat
* Having made this assertion, I thought it as well to ask Com. Crane's authority for making it, and he tells me I am quite right.
68
PATTERN DESIGN.
86. WALLPAPER DESIGN BV WALTER CRANE.
diamond. Artistically it makes all the difference in the world to the designer upon which plan he sets to work. Either one of them would encourage him to do something which the others would not. His design is materially influenced by the shape he sets himself to fill. It would never occvu" to him, for instance, to stretch a wreath of flowers across a width of space which he did not see before him. And the idea of extending a design far beyond the wddth of the material in which it is to
THE DROP REPEAT.
69
n
87. DIAGRAM SHOWING MECHANICAL RELATION OF VARIOUS PLANS.
be executed, may be set down as directly due to working on the lines of the diamond. A designer does not, except in certain deliberately formal patterns, keep his design within the lines upon which it repeats. But he has them always in view; and he does not stray from them so far that it ceases to make a difference what lines he works on. The advantage of setting out a drop design upon the plan of the diamond is, that the simplicity of the four straight lines enables him to keep more clearly in view than the other lines upon which the drop is worked, the ultimate relation of the parts of his design, and the order in which they will recur. Perhaps the most conspicuous advantage of the drop repeat is that it enables one to perform the apparently impossible feat of designing a pattern twice the width of given material, which yet works out perfectly as a repeat within its limits.
Working on the lines of the diamond, it is easy to do this. You have only to subdivide the area of your square repeat as here shown (88), (it might just as well have been a parallelo- gram as a square) so that two smaller divisions a and v together equal the larger ^. Then if you transpose the smaller parts a and v so that together with •^, they form a squat diamond twice the width of the original square, you
|
e |
A V |
|||
|
V |
||||
|
8S. |
Dl.\G |
RAM SHOWING |
DIVISION |
|
|
OF SQUARE |
REPEAT INTO |
|||
|
THRE |
E PARTS |
So. Diagram showing transposition of
PARTS OF SQUARE TO FORM WIDE DIAMOND.
70
PATTERN DESIGN.
have the repeat of a design which amounts to, mechanically, the same thing as a square repeat of half that width. In the case of materials which can be dropped one-half their depth in hanging or in making up this is clearly a great gain.
The advantage, it may be argued, is only apparent: what is put into one strip is, as it were, taken out of the other; but in the case of pattern appearance must be allowed to count for a great deal. It is for want of knowing things like this, the common property of trade designers, that, genius not-
90. BRICK OR M.\SONRV PATTERN.
withstanding, artists inexperienced in practical work fall short even of the trade standard of efficiency.
The skeleton given b}^ the upright marginal lines of the fabric and parallel lines in a diagonal direction across it, is plainly helpful in the design of a diagonal stripe. The angle of inclination determines the depth of the drop.
An all-over pattern may also be designed within those Hues; and they encourage greater freedom than rectangular or diamond lines; but it is not easy on such a scaffolding to balance the parts of a design; and if there are emphatic features in it they are liable to come out awkwardly in repetition.
THE DROP REPEAT.
71
m.j47
\^==^
91. DIAGRAMS OF BRICK PLAN
AND ITS RELATION TO
DROP REPEAT.
Another ver\' useful stepped plan on which to scheme especially patterns which take diagonal lines is the brick or masonry- basis — which also works out as a drop.
In the tile pattern 90 the masonry lines form part of the design, and materially influence the lines of its growth.
It will be seen from the diagrams on this page (91) that a pattern designed upon brick lines and one upon the lines of the ordi- nary drop may amount to precisely the same thing,
72
PATTERN DESIGN.
though either plan would naturally affect to some extent the growth of the pattern. (The diamond lines in the lower of the two diagrams show plans upon which theoretically it might have been, but practically would not have been, con- structed.) The diagonal lines may wave as freely as you please within the four sides of the brick, at the same time that the rigid skeleton of brickwork enables you to distribute your flowers or other free-growing features in strict order.
The adoption of the brick plan leads sometimes to con- fusion as to the dimensions of the repeat. The brick pattern above would not answer to the description of " a drop-repeat twice as wide as it is deep. " As a unit of those dimensions, it does not drop in the technical sense. It drops in fact as a repeat twice as deep as it is wide, or as a right-angled diamond. If, therefore, a drop-repeat oj given dimensions were speci- fied, and the artist were to send in a design planned on a brick of those proportions, he might have it thrown back upon his hands, as not being to size.
92. HEXAGON PLAN — ITS RELATION TO DIAMOND.
THE DROP REPEAT.
73
|
^ |
^ |
II |
=11 |
— |
||||
|
^ |
^ |
II |
II |
— |
II |
|||
|
S! |
. |
^ |
II |
II |
^ |
93. COUNTERACTING LINES OF DESIGN — REPEAT A DOUBLE SQUARE.
The hexagon, again, does not drop in the orthodox man- ner, though it amounts to the same thing as a diamond which does drop. Diagram 92 will show how, if the dotted portions of the hexagon were cut off, and attached again in the position of the solid black triangles, the result would be a diamond.
But it is not convenient to design upon the hexagon. It gives you no scope which the diamond does not; and it does not help you to avoid a too horizontal arrangement of fea- tures, as the diamond does. It may be at times convenient to prove a design planned upon the diamond on the lines of the hexagon.
A drop design upon vertical and horizontal Hues, say upon the square or other rectangular step, does not afford the oppor- tunities given by the diamond plan; but it has compensating advantages of its own, especially where it is desirable to give
•V
-4
/k
\^
\f
y^
y\
4
/x.
A^
^
^S.
AS.
94. DIAMOND FORMS RESULTING FROM COUNTERACTING LINES- REPEAT A DOUBLE SQUARE.
74
PATTERN DESIGN.
an upright tendency to the pattern, and more especially still when the depth of the repeat happens to measure (as in some manufactures — tiles for example— it is convenient it should) precisely twice its width. The unit of repeat being in this case a double square, gives us the trellis on which the chessboard pattern is built. A checquer pattern (illustra- tion 93) works as a drop-repeat one square wide and two squares deep.
The unit of a double square planned to step half-way, is most convenient in the case where it is desired to preserve square lines of construction in the design and yet to avoid an\- tendency it might have in one direction or the other.
95. DIAGRAM OF STEP PLAN WHICH DOES NOT GIVE A " DROP REPEAT. "
THE DROP REPEAT.
75
V _!V^ ^r
K-2.! iNCHLb-^/ ^^^V"
96. DIAGRAM OF PATTERN WHICH DROPS ONE-THIRD OF ITS DEPTH.
This is effectually done by counteracting the vertical tendency of the lines in one division of the repeat by horizontal lines in
97. PATTERN WHICH DROPS ONE-QUARTER OF THE DEPTH OF REPEAT.
76
PATTERN DESIGN.
%
m
^sSa^l^tf-
uS. STEP PATTERN WHICH IS NOT A DROP REPEAT.
the other (93). Counteracting diagonal lines give, in the same wav, diamonds (94). *
Further uses of the step, in heu of the diamond, will be apparent when it comes to the discussion of freer patterns designed on its lines.
There is one form of step pattern which does not really constitute what is understood by a drop repeat. The second strip in diagram 95 drops slightly; but the third reverts again to the level of the first (or, if it can be said to drop, it takes a step out of all proportion to the last) . As a consequence any such features as the flowers at the top of the repeat would, in recurrence, give a sort of zig-zag line.
True, a drop pattern may recover itself in the third repeat; but only on the condition that it drops just half its depth. A drop of one-third its depth recovers itself only in the fourth
* Another variety of the drop pattern in which the unit is not merely of the proportions of a double square, but is built deliberately upon the two squares counterchanged, is shown on page loi.
THE DROP REPEAT.
77
99.. PATTERN WHICH DROPS ONE-THIRD OR TWO -THIRDS OF 115 DEPTH UNIT OF^REPEAT SUBDIVIDED AS IN 98.
strip (96); a drop of one-fourth its depth, only in the fifth (97), and so on. In a drop-repeat, properh^ so-called, each successive strip drops, and drops alwa3's the same distance. It does not jump up and down. The repeat in diagram 95 is really two strips wide; it does not drop; and there is no mechanical reason why the two flowers, any more than the stems (indicated in the lower part of it), should be repeated.
The peacock feather tile (98) is not a regular drop pattern; it drops in the second row two-thirds of its depth; but in the third it starts afresh on a level with the first.
The tile pattern above (99), on the other hand, is designed to drop regularly two-thirds of its depth, and would recover its level naturally in the fourth row, as would a pattern de- signed to drop only one-third of its depth. As a matter of fact, though it drops two-thirds if we work from left to right, from right to left it drops only one-third. That may read as if it were impossible, but if vou work it out on paper you will find it is so.
Patterns of which diamonds, or equivalent ogee shapes, are the basis (100), have always an air of being drop patterns;
78
PATTERN DESIGN.
but they do not really work on that plan unless the pattern in the diamonds or ogees also drops. But this sort of " false drop " plan is useful. The framework of severe lines steadies the ornament, which yet may be varied; and there is perhaps a charm of unexpectedness in the result. One starts some- times with the idea of a drop pattern (loi) which develops, nevertheless, into a pattern which works only on square lines. There is no harm in this, so long as it is not necessary that the pattern should step in the working — which it may be.
It has been shown how no new principle is involved in designing on waved lines. They are but another version of the straight-lined skeleton, and amount to the same thing, except that their curves give the designer a lead which he is often wise to take.
Working upon the lines of the diamond, of which the ogee shape is the curvilinear equivalent, he would not so easily have arrived at the fret pattern (102) overleaf. Starting
FALSE DROP
THE DROP REPEAT.
79
FALSE DROP PATTERN.
with opposite wave lines, filling each ogee with a square, and just making pointed encroachments upon this, he arrived almost inevitably at what he did.
Another very useful form of " false drop " is where the unit of design occupying, let us say, a diamond or ogee, is turned over in the dropping. A sprig pattern, for example, in which the sprigs are alternately reversed — one row of flowers turning from left to right the other from right to left — may with advantage be planned on the diamond; but the fact that the sprig in the diamond which drops is not a repeat of the first, but the reverse of it, removes it from the category of drop patterns proper. The unit is now the double diamond; and that no longer drops in repetition.
This may appear to the reader a mere verbal quibble not worth discussing; but it has a very practical bearing upon design. If, for example, the unit of design occupying diamond B in diagram 104 (see page 82) were reversed in diamond c, it would naturally be reversed also in the two quarters at a a, .
8o
PATTERN DESIGN.
I02. FRET DESIGNED OX OGEE AND SQUARE LINES.
and would not join on to the half unit at a facing the same wav as at b, and the pattern would not work.
Every fresh skeleton plan is a boon to the designer; for, working upon any fixed proportions (such as the conditions of any manifacture are sure to lay down for us), we fall in- evitably into certain grooves of design; and all opportunity of varying them is to the good. It is because they offer each its own particular hues of construction (by which design can- not but be influenced) that it is worth the designer's while to puzzle over the various plans upon which pattern may pos- sibly be built.
A
.."^
m^^-
1 \ 1 T ; T ;
lOS. DIAGRAMS SHOWING DROP AND STEP REPEATS WITHIN THE WIDTH OF MATERIAL.
IX. SMALLER REPEATS.
Width of repeat divisible into width of material — Repeat two-thirds or two-fifths of width of material— F"ull width repeat seeming smaller — Variety in apparent imiformity — Weavers' ways of doing it — Same principle applied to larger design — Method and haphazard — More complicated svstem — Other plans for disguising precise order of small repeats.
As a rule the designer is anxious to get the most out of the space he has to deal with. The use of the drop, it has been explained, enables him to go even beyond the width of his material. But it is not always that he wants the whole width allowed. There are reasons of economy and use (economy of design no less than of manufacture) which make it necessar^^ at times, and especiallj^ in certain classes of design, that several repeats of the design should occur in the width of the stuff. If the repeat is on horizontal Unes it must clearly be contained exactly twice, or three or four or more times, in the width;
82
PATTERN DESIGN.
104. DIAGRAM SHOWING PLAN OF DROP REPEAT TWO-THIRDS OF THE WIDTH OF MATERIAL,
otherwise, when the material comes to be joined up, the design will not match, without cutting the stuff to waste.
A drop-repeat within the width of the material does not, it should be mentioned, entail a corresponding drop in joining or hanging. Suppose the material in diagram A (103) to be wallpaper 21 inches wide, and the repeat to be only 7 inches wide, and drop just a third or two-thirds of its length. In that case the paper will hang not as a drop but on level lines. The step pattern b, on the other hand, which recovers itself
105. DIAGRAM SHOWING PLAN OF DROP REPEAT TWO-FIFTHS OF THE WIDTH OF THE MATERIAL.
S.MALLER REPEATS.
83
106. DROP DESIGN SCHEMED OX A DIAMOND TWO-THIRDS OF THE WIDTH OF THE MATERIAL.
84 PATTERN DESIGN.
after the first repeat, will hang in corresponding fashion, dropping in one strip 7 inches, and in the next recovering itself again.
The " drop " offers yet further possibilities in design, and makes possible a repeat measuring, not merely a half, a third, a quarter, but two-thirds of the width of the material, or two-fifths of it, or two-sevenths, and so forth.
If, as in diagrams 104 and 105, you divide the area of pos- sible repeat vertically into three, or into five, a pattern desi- gned on a diamond the width of two divisions will work as a drop. All that is necessary is that the half diamond A on the one side and the two quarter diamonds a a on the other should join on. They form, in fact, together the unit of the repeating pattern — supposing, that is to say, diamonds a, B, c in the one case (104), and a, b, c, d. e in the other (105), to be filled in the same way.
The design (106) on page 83 shows a 21 inch material of which the design is schemed on a diamond 14 inches by 21 inches.
Ceiling pattern 107 w^orks on a similar plan; but, as it happens to turn over within the diamond, it works also as a drop-repeat 7 inches wide b}^ 14 inches long. The block from which it is printed measures 21 inches b}' 14 inches.
The repeat, however, in such patterns is not dependent upon the filling of the diamonds all alike. So long as in diagram 104 the half diamond a and the two quarter diamonds a a make one complete diamond, and the two half diamonds c c another, the three diamonds (a, b, and c) may be occupied each with a separate figure. The design (108) on page 86 is planned upon the system of diamond divisions measuring two-thirds the width of the material. The dia- monds (a), of which only half would occur in the width of the stuff, are occupied by sprays of foliage, and through the zig-zag space (b) between (equivalent to the other two diamonds) winds a separate growth.
SMALLER REPEATS.
There are manv occasions on which it is advisable to reduce pattern to a scale far less than the mechanical con-
^■^ rS^^yi.
107. DROP REPEAT 7 INCHES BY I4 INCHES PRINTED FROM A 21 INCHES BY I4 INCHES.
ditions would allow. In that case it may nevertheless be well to take advantage of those conditions in order to get
86
PATTERN DESIGN.
variety, which, though not perhaps immediately apparent, is always pleasing when it is discovered. What is in effect quite a small repeat may, in point of fact, occupy the full width of a ^^ide material.
An expedient that is often useful is to set out the lines of your pattern as if for a small repeat, and within those lines allow yourself all possible libert3^ For example, you may devise a small sprig pattern, and then amuse yourself by playing variations upon it, so as to suggest perhaps, even in mechanically produced pattern, something of the freedom of handiwork — at all events avoiding the mechanical effect of too obvious repetition. In the Byzantine piercing (109) on page 87 a pattern of interlacing bands is diversified by filling the geometric spaces with sprigs as it were accidentally dis-
-A"'--.
v:>\a±eruit— >v
108. DIAGRAM OF DROP REPEAT ON DIAMOND LINES, THE DIAMONDS NOT FILLED ALL IN THE SAME WAY.
SMALLER REPEATS.
87
109. BYZANTINE PIERCED WORK WITH GEOMETRIC DIVISIONS ENCLOSING ORNAMENT WHICH DOES NOT REPEAT.
persed. In this particular instance the foKated ornament does not repeat at all. But it might very well have done SO. There are, however, two dangers in playing any little game Hke this; the one, that you may get confused as to the particular units which must join; and the other that, failing system in the variation, the changes may be sufficient to throw the design out of balance, and allovs' certain units to assert themselves detrimentally. It is consequently well worth the pains of any one engaged in designing small repeats to work out the various plans upon which sprays and so forth may be schemed, so as not to recur quite obviously, and yet to fall surely into satisfactory lines.
Weavers have, in fact, perfected a system by which the danger of apparent lines in small repeats is minimised. Some of these arrangements * give, it will be seen, diagonal lines, others afford a ready means of avoiding them.
How they arose out of the necessities of weaving is not
* " Sateeas ,, they are technically called.
88
PATTERN DESIGN.
<3
£>
Z>
<3
£>
^ : <i?
V : Z'>
o; <3
■V; » : • ■
OP
<?i
6
V
c>
V
5
<e7
r-'
a
Q
V
Cb
Q
Q?
lS
<e7
V
a
c^
<o
^
a
^
o
no. DIAGRAM OF THREE-SPOT III. DIAGRAM OF FOUR-SPOT
REPEAT. REPEAT.
here the question. Nor is it necessary to go into the matter of " ends, " " counts, " " picks, " " treads, " and other tech- nicalities famiHar enough to the expert in weaving, and to those who are not, more puzzling than explanatory. But, as they may be helpful to designers of no matter what kind of
|
A |
G |
■4 |
a: |
||||||
|
» |
|||||||||
|
4 |
' A : |
||||||||
|
¥ |
•., ;*•: |
||||||||
|
» |
•*;;■;; |
||||||||
|
■: V |
6;^ |
(b |
^ : |
||||||
|
l 6^ ': Cb ': G ■ G |
Cb |
Cb: A i ^ ! |
112. DIAGRAMS OF FIVE-SPOT REPEAT.
SMALLER REPEATS.
Sg
pattern so long as it repeats, it is worth while giving them for what they may be worth.
The designer begins by dividing his repeat into squares 3, 4, 5, 6, 7, or 8 each way, as shown in the corner of each diagram. He has then to occupy ",, ■], 5, 6, j, or 8 of these squares in such a way that in any row of squares, from top to bottom or from side to side, one of them, and one onh^ is inhabited.
|
•>; |
(Si (S A .-A |
ft ft |
|
; ;*•: |
||
|
a]" v |
||
|
•: ^ i ^ •A |
(? ; (? ft ; ft |
(b ft) ft i |
|
:■ ^ |
O : ^ (b; & ft ■: ft ft ift |
c^ - ft |
113. DIAGRAM OF SIX-SPOT REPEAT.
How this works out in the repeat — whether it takes a stripe or not, and what stripe, is shown by the repetition of this group of squares in outline.
In the diagrams, the heart-shapes, it will be seen, face all ways about, to show\ how, at the option of the artist, the spray or .wha.tever it happens to be may be varied.
The application of such a principle as this to design on a •comparatively large scale as in the ceiling paper on page 92
go
PATTERN DESIGN.
|
"T'T- : " ; i iA |
<2 |
|
|
A: : : : : ! |
||
|
•: ^ |
<59 |
: ::■.:: A
■■:■-■■:■■ Vff,r r' ...^....^.. ..............
■'ryi ■;':■■■;».[■■
c:^
'^
G^
'^^
&
^
CP
(^
^
G^
<5?
Ch
^
c?
^
^
c^
^
(h
^
G^
^
^
c?
^
o
c^
^
0^
^
&
S)
c?
^
a
^
<D
o
Q
C?
O
G^
114, 115. DIAGRAMS OF SEVEN-SPOT REPEAT.
S:\IALLER REPEATS.
91
|
"';^;"T • ■ i • v;" : : i |
& : C> : o ; a i c= ; ^ : |
|
: 6 : "^ |
|
|
: C> : c3 \ ^ |
|
: • : :• ;^; ; |
^ : |
<3 : |
|
'iki'";'";'"^"'^'";'":'" |
||
|
O : A: <3 : c? : |
Il6, 117. DIAGRAMS OF EIGHT-SPOT REPEAT.
PA*rTERN DESIGN.
IIO. CEILING PATTERN DESIGNED ON THE PLAN OF SIX-SPOT
REPEAT.
SMALLER REPEATS.
93
is CNplained by the diagrams which follow (119). In the first is shown the occupation of six squares by forms not yet care- fully considered. In the second these begin to take leaf shape extending somewhat beyond the boundary lines. With the more careful drawing of these leaves and the breaking them up into feathery composite foliage, the design takes its final shape.
A farther application of the idea is shown in yet another diagram (120) in which only the main features of the design are distributed systematically. The position of such heavier and more emphatic masses determined, it is safe to sketch in the more delicate connecting scrollwork quite freely.
119. DIAGRAMS SHOWING EVOLUTION OF PATTERN NO. II
Similarly the squares (one in each row) may be reserved, not for the pattern, but as spaces free from ornament (121), places of rest, where the eye can appreciate the quality of plain material. The diagram insists, for emphasis' sake, upon the squareness of the spaces left, but in a finished design the scaffolding lines would of course not be there.
An alternative to the more systematic^manner — and one which appeals to the ungovernable frame of artistic mind — is to begin \vith sprays or whatever they may be on the margin of the repeat, and work gradually to the centre.
PATTERN DESIGN,
2. I 3 1 1-1 5
rka
I20. DIAGRAM SHOWING DISTRIBUTION OF FLOWERS ONLY ON THE PLAN OF A SIX-SPOT REPEAT.
trusting to the guidance of artistic instinct. That seems, perhaps, the readiest way; but it is in the end the longest — if ever it leads to an\i;hing but disappointment.
The full possibilities of the systematic principle, are in- dicated in the last of this series of diagrams (122), designed to show the successive stages by which, first, six of the squares are occupied with leaf forms; then, in a similar way, six other squares with spirals; then others in succession with flowers, stars and butterflies. The result is not a pleasing pattern — that was not aimed at — but an unmistakable chart
SMALLER REPEATS.
95
121. DIAGRAM SHOWING DISTRIBUTION OF OPEN SPACES IN THE GROUND ON THE PLAN OF A FIVE-SPOT REPEAT.
of the steps by which the designer may proceed to fill out his design. It should be useful also as an indication of the way in which, employing always the same or similar spra\s or whatever they may be throughout his design, he may \'arv their colour. Let the five features represent five tones of colour, and the monotony of a single spray of ornament would be vastly reheved. Let the sprays be further slightly (more or less accidentally) varied in design, and an element of mystery would be introduced which seldom fails to add to the charm of pattern.
96
PATTERN DESIGN.
Two other plans upon whicli sprays, &c., ma^^' conve- nientl}" be distributed are worth showing (diagram 123).
Mark on the sides of a square central points, and from these to the corners draw parallel lines obliquely across. That will give you a centre square and eight parts of corresponding squares. Complete the four squares which want least to make them perfect, and you will have a cruciform unit of five divisions, no one of which is in a \'ertical or horizontal line with another.
Or, again, mark on the sides of a scpiare two points a and b, dividing them into three equal parts, and from a draw oblique lines to the corners, and from h to h lines parallel with them. That will give you four complete squares and twelve portions of corresponding squares. Complete as in the last case, the four of these which are most nearly perfect, and two of the half squares not opposite one to the other, and \-ou will have a unit of ten divisions no one of which is in a
|
• ! ! ^ \ |
■""T"": \"'W""' \<i\ i"""i '■ ■■"T"T' t^i : \"\ rr\^ |
|
|
'.':'. \ '• |
\ \^\ \ \ |
\ i^i « ; |
|
#1 'M^\ \ |
^1 ijrf)^>i^ci ' |
122.
DIAGRAM SHOWING FURTHER DEVELOPMENT OF THE PRINCIPLE OF DISTRIBUTION ALREADY EXPLAINED.
SMALLER REPEATS.
97
123. DIAGRAMS ILLUSTRATING ANOTHER PRINCIPLE ON WHICH SPOTS MAY BE DISTRIBUTED.
vertical or horizontal line with another. The result is in either case a square lattice askew. It is shown in the diagrams above both in repeat and in relation to the width of the material.
Any pattern occupying these squares \\ould, if it followed the slope of the lattice, take slanting lines, and little or nothing would be gained. But in an upright spray, more especialh- if there were in it a marked vertical line, as, for example, in a fleur de lis, the upright tendency of the diaper would con- trast with the lines of the plan, and the order of repetition would not be too apparent. Remove the trelhs of scaffolding and it would take one some time to make out the precise order in which the diaper was sprinkled about.
The value of systems like these is just that. It makes the order of an obvious repeat less obvious.
X. SUNDRY SCAFFOLDINGS.
Importance of variety of plan — Area of pattern not confined to area of repeat — Excursions compensated bv incuisions — Tines thiis dis- guised— Wave-lines, turned over, result in ogee — Wave-lines result from working within narrow upright lines — Uprightness of narrow repeats counteracted by lines across — Diagonal wave-lines to connect features forming horizontal band — Designs obviously based upon slanting and horizontal lines — Wave-line from side to side of broad repeat — Scaffolding of an old Louis XVI. pattern.
Inevitably as repeated patterns fall into the lines of the square, the diamond, or perhaps the triangle, those were not by any means invariably the lines on which the designer set to work. Reference has been already made to some possible scaffoldings; others remain for consideration; and, in view of the vital way the lines on which a man works influence his design (one plan suggesting what another does not so much as allow), it is important that he should have the widest possible base of operations.
Beginners, bv the way, seem always to be unnecessarily bothered by the shape of their repeat — square, oblong, dia- mond, or whatever it may be. I have seen it somewhere stated, for their guidance, that they need not confine the lines of their design to it. Indeed they need not. It would be safe to say that, except in mere diaper, they must on no account do so. If they do, the hne of the repeat, not crossed by ornament at all, will assert itself, very probably in a way that is anything but desirable.
Even in a comparatively open pattern with plenty of plain ground, as in the printed stuff on page 99, a marked vertical
SUNDRY SCAFFOLDINGS.
99
124. TURNOVER PATTERN. REVEALING THE VERTICAL LINE ON WHICH IT IS REVERSED.
line results from keeping the pattern entirely within the width of the repeat. In textile design it is sometimes thought ad\dsable purposely to confine the pattern in this way, so that it may have the appearance of completeness when made up in furniture or upholstery. As a rule it is expedient, even where the design is mairdy contained within the width of the stuff, to block the gap in the ground which would occur where two strips join by carrying comparatively insignificant, but sufficiently substantial, portions of the pattern across it.
PATTERN DESIGN.
125. DESIGN ALMOST BUT NOT QUITE SELF-CONTAINED WITHIN THE WIDTH OF THE STUFF.
SUNDRY SCAFFOLDINGS.
126. DIAGRAM SHOWING CONSTRUCTION AND DEVELOPMENT OF COUNTERCHANGE PATTERNS.
A test by- ^^•hic]l to judge the competence of a pattern- designer is the wav he manages to give you in his designs features extending far bevond the hmits of his repeat, obtain- ing by that means a bolder scale and a freer line than are otherwise to be got. There is no great art in thus exceeding the limits of the repeat. One has onlv to remember that excess on one side of it must be compensated on the other. It is a question of addition and subtraction.
This is very plainly shown in those geometric patterns of which the scaffolding forms part of the design.
Given a checf^uer of black and white, anv inroad of the
17. ALHAMBRESQUE COUNTERCHANGE PATTERN CONSTRUCTED ON THE LINES OF DIAGRAM 1 26.
PATTERN DESIGN.
{<-AAm4fM
128. DIAGRAM SHOWING CONSTRUCTION AND DEVELOPMENT OF RIBBON AND FEATHER PATTERN, DIAGRAM 120.
black into the white has onlv to be followed bv a correspond- ing inroad of the white into the black, and you have a well- balanced pattern.
129. RIBBON AND FEATHER PATTERN CONSTRUCTED ON LINES OF DIAGRAM 128.
SUNDRY SCAFFOLDINGS.
103
I ^O. LATE GOTHIC VELVET PATTERN.
PATTERN DESIGN.
125. DESIGN ALMOST BUT NOT QUITE SELF-CONTAINED WITHIN THE WIDTH OF THE STUFF.
SUNDRY SCAFFOLDINGS.
126. DIAGRAM SHOWING CONSTRUCTION AND DEVELOPMENT OF COUNTERCHANGE PATTERNS.
A test b\' whicli to judge the competence of a pattern- designer is the wav he manages to give yon in his designs features extending far bevond the Umits of his repeat, obtain- ing by that means a bolder scale and a freer line than are otherwise to be got. There is no great art in thus exceeding the limits of the repeat. One has only to remember that excess on one side of it must be compensated on the other. It is a question of addition and subtraction.
This is very plainly shown in those geometric patterns of which the scaffolding forms part of the design.
Given a checquer of black and white, any inroad of the
127. ALHAMBRESQUE COUNTERCHANGE PATTERN CONSTRUCTED ON THE LINES OF DIAGRAM 1 26.
* 1 • I
i t I • I
• ; f A
it
PATTERN DESIGN,
128. DIAGRAM SHOWING CONSTRUCTION AND DEVELOPMENT OF RIBBON AND FEATHER PATTERN, DIAGRAM 129.
black into the white has only to be followed by a correspond- ing inroad of the white into the black, and you have a well- balanced pattern.
129. RIBBON AND FEATHER PATTERN CONSTRUCTED ON LINES OF DIAGRAM 128.
I . 1
i I i
I I
SUNDRY SCAFFOLDINGS.
103
130. LATE GOTHIC VELVET PATTERN.
I04
PATTERN DESIGN.
131. PATTERN IN WHICH WAVE-LINES DIVERT THE EYE FROM VERTICAL.
That explains itself at a glance in diagram 126. But more intricate-looking patterns come about in precisely the same way. The inroads of the black into the white diamonds at a have only to be compensated by identical incursions of the white into the black diamonds at b, and you have the unit which gives the very satisfactory counterchange (127) at the bottom of page loi. The two (black) bites out of one square are paid for by two (white) bites out of the other.
Practically the only \\a.y to avoid the lines of open space
SUNDRY SCAFFOLDINGS.
105
&ubdlvit>too
132. DIAGRAM OF SCAFFOLDING AND THE LINES OF A PATTERN RESULTING FROM IT.
133. ANOTHER PATTERN BUILT ON THE SAME LINES AS I32.
io6
PATTERN DESIGN.
134. OLD DUTCH I'RIXT.
which result from keeping the unit of design within the hnes of the repeat is, to cover so Httle of the ground with it, to leave so much space about it, that it resolves itself into some sort of a sprig or spot pattern.
They ma\- be disguised b\' designing within, not squares
SUNDRY SCAFFOLDINGS.
107
or diamonds, but some sucli broken geometric scaffolding as would be given by say four of the shapes in diagram 127 (which would themselves repeat on the lines of the diamond) or \vithin some equally broken but less regular shape. For my own part, I have never thought that worth doing. It might, however, be worth while to design a narrow flowing pattern, in which it was desired to a\-oid anvthing like a vertical direction within the Hues of a zig-zag (128) or of the slanting " herringbone " (128) which results from continuing the dia- gonal hues until they meet at right angles. It is long unbroken lines in oiie direction which are so apt to assert them.selves.
In an old Itahan velvet (129) the above-mentioned dia- gonal brick lines have been adopted as the plan of a peacock's feather pattern tied together by ribbons, which mark, not precisely the brick, but a flowing-lined variant upon it.
fik fiW^. V. o^A ^^.,% /%
135. CRETAN WOVEN PATTERN.
136- PATTERN IX WHICH THE HORIZOXTAI- LINE IS DELIBERATELY
MARKED.
SUNDRY SCAFFOLDINGS.
log
A vast number of excellent patterns liave been frankly built upon the ogee, which curved variety of the diamond seems, wherever it is emplo3''ed, to command acknowledgment.
The late Gothic patterns of which that on page 103 is a type, seem to be the result of simply opening out an ogee pattern, dropping the ogee shapes, that is to say, some dis- tance apart, so as to give zig-zag bands between. Tliese are still to be traced in the design illustrated, though the ogee shapes are no longer intact.
Not every pattern in which the ogee occurs, was neces- sarily designed upon its lines. One may start with a scroll, and the turning over of the wave-lines gives at once the ogee.
The wave-line itself comes (though it may seem like a con- tradiction in terms to say so) of working upon narrow upright lines, or between them. There is no readier means of counter- acting the too upright tendency of a long narrow repeat than by lines or bands waving fromside to side of the confined space.
Even when the space is not narrow, as in ti;e design (131) on page 104, it may be convenient to anticipate any possibility of vertical lines by carr\ang the eye alternately from left to right and from right to left.
A scaffolding which leads to new developments in design is to be got by means of a trellis which divides the rect-
ocxya
V i Y ; V i V ;
pddCDm
( y y V X
PPDO
B.
137. DIAGRAMS OF SCAFFOLDINGS.
PATTERN DESIGN,
I_^8. SILK BROCADE ON THE LINES OF DIAGRAM B.
angular area of repeat into six parts, grouping them as (132) on page 105. If these hues are repeated a broad space reveals itself between the smaller oblongs which, w^hen it comes to occuppng it with pattern, results almost inevitably in a wave- line as does the zig-zag chain of parallelograms between.
Yet another diagram (133) on the same lines shows that even when the square lines of the plan are insisted upon, something like wave-lines result.
Another obvious means of counteracting the uprightness of \ery narrow repeats is to cross the upright lines, and
SUNDRY SCAFFOT.DINGS.
139. SIXTEENTH CENTURY VELVET ON THE LINES OF DIAGRAM C.
112
PATTERN DESIGN.
140. PATTERN IN WHICH THE DIAGONAL LINE WAS AN AFTERTHOUGHT.
perhaps the waved hnes within them, by features which give a pronounced band either in the horizontal or in the diagonal direction. The tendency of the narrow turnover in the Dutch print (134) on page 106 is effectually overcome by the pro- nounced horizontal line of birds, though the direction of the stalks into which they develop helps also in the same way. In the Cretan weaving (135) on page 107, though the wave-
SUNDRY SCAFFOLDINGS.
lines are not actually broken by the flowers, they form in repetition compact bands, which go far to stop their upward tendency. One seems to read in that case very plainly the genesis of the design — a narrow repeat dictated b\- the loom ; wave-lines, to take from its straightness; emphatic bands of flowers, to stop the upward direction of the pattern; and further breaks in the colour of the wave-lines, with the same object. The plan might be described as a trellis consisting of upright wa\c-lines and straight lines crossing them horizontallv.
It is true that horizontal lines of the kind here shown result, whether the designer will or no, from the repetition of any feature which nearly takes up the full width of the repeat.
,4*
-;iU^g*3fS
141. PATTERN IN WHICH THE ST.\RTING- POINT WAS A DIAGONAL
WAVE.
114 PATTERN DESIGN.
It was very likelv that which gave the hint to wea\^ers; but they were not slow to take it, and to turn it to very deliberate and constant use.
No designer will doubt for a moment that the long leaves in the pattern (136) on page 108 are an artifice by which to stop the flow of blossom and sprays, and to steady the effect.
A natural thing to do in a narrow turnover pattern (with a view to interrupting its straight-up direction) is to plant, as in diagram a, 137, a prominent feature, occupying nearly, if not quite, the full width of the stuff, alternately in the centre of the strip and centring with the joint between the two strips. (This applies, of course, just as much to the repeats which recur several times in a single width of the material.)
Diagram b (137) brings us to something like the plan of the fifteenth century pattern (138) on page no, and may be resolved into a diamond scaffolding. But, if the strips or the repeats are narro\v, and there is a fair amount of space between the alternate bands of features, flowers let us call them, an}' lattice of stalks connecting them, whether on diamond or ogee lines, would be too long-drawn-out for beauty. A single Hne from flower to flower would be much more satis- factory—from which results (whether we mean it or not) a diagonal stripe — more or less ingeniously to be disguised as in the fifteenth century silk on page iio.
Diagram c (137) explains the genesis of some of the most sumptuous patterns of sixteenth century brocaded velvets (139). Their starting-point seems to have been a huge conven- tional flower or pine-apple, occupying nearly, if not quite, the entire width of the material, recurring, of course, at intervals, with one broad waving stem from flower to flower, not, it is clear, running behind the flowers in a continuous sweep, but appearing rather to stop against the flower below it: at all events the flow of the line is not continuous. The puzzling thing at first about these handsome patterns is that you don't
SUNDRY SCAFFOLDINGS.
142. FIFTEENTH CENTURY PATTERN AND ITS SCAFFOLDING.
ii6
PATTERN DESIGN.
follow the logic of the design. I am inclined to think there is none; that the designer did not bother himself about the repeat; that he trusted to the bigness of the pattern, the sequence of which one can seldom see, and takes too readily for granted.
I can speak quite positively as to the process by which my own design (140) on page 112 came about. The starting- point was there the flowers; the diagonal growth was, not precisely an accident, but an afterthought.
' (Ml
-^ ' Mi)
143. SICILIAN SILK PATTICRX AND ITS SC.\FFOLDING.
In the pattern (141) on page 113, on the other hand, it is clear that the start was a diagonal line waving gaily from corner to corner of the repeat, and that the sprays of flowers were put there to steady the effect. For the diagonal stripe came at an early date not merely to be accepted but to be insisted upon as an acceptable feature in design — which to unprejudiced eyes it still is.
There seems to me no room for doubt that patterns such
SUNDRY SCAFFOLDINGS.
144. PLAN OF A FRENXH DESIGN OF THE LOUIS SEIZE PERIOD.
iiS PATTERN DESIGN.
as that (142) on page 115 were deliberately planned on the horizontal and slanting linos indicated in the lower part of the diagram; or that the Sicilian silk (14^)) was built up on the same sort of scaffolding.
The scaffolding of a design by a French designer of the period of i.ouis Seize (144) is interesting and instructive. The heavier of the vertical lines give the width of the material (rather more than two widths are sliown), the finer of them stand for the pencil marks which the artist ruled for his guidance. The horizontal help lines mark the length of the repeat and its subdivision into four parts, two of which give the drop. It will be seen that he has divided the width of the material also into four parts, two of which (2 and 3) are reserved for the central features of the design, ^^•hilst the other two (i and 4) confine the hanging wreaths (which frame the central features) within (^asily manageable areas. It is plain that the scaffolding lines assist him in carrying these wreaths from one width of the material to the other. You feel, in fact, that without this scaffolding he would not easily have arrived at a composition which even those who have no sympathy with the style of it must admit to be exceptionally graceful.
It is well worth while working out for oneself plans of this kind, as a means of compelling the invention out of the ruts sure to be worn by continu.ill\- working on the same lines.
XT. THE TURNOVER.
A weaver's device — Doubles width of pattern — Exact turnover not desirable where conditions do not make it necessary — Balance must be preserved — Use of doubling over in border d. sign — Suited to stencilling and pouncing.
To the practice of folding or doubling over in the vertical direction, may be traced a large class of bi-symmetrical designs. Merc doubling makes a sort of pattern; and some of the steadiest and most satisfactory designs rely to a large extent for their symmetry and steadiness upon the reversing of their lines.
To the weaver the " turnover " (145) is a veritable god- send, enabling him, without increase of cost or trouble, to double the width of his pattern. It does not even involve the cutting of more cards; it is simph' a question of the gear of the loom.
So obvious is the advantage of the " turnover " to the weaver that the device might well have originated with him. But that is a point upon which it is useless to speculate. A man has only to double a sheet of paper and he can with one action of the knife cut out the two halves of what when it is opened out is a bi-svmmetrical pattern.
Once invented the " turnover " proves the easiest and simplest means of doubling without more ado the width of a pattern.
Apart from the fresh facilities afforded b}- it for broader pattern planning, and the m.uch larger scale of design which
PATTERN DESIGN.
it makes possible (observe how very narrow is the strip turned over in what is in effect a bold Gothic tapestry, 146), designers generally, even though they may have no technical grounds for so doing, will " turn over " the lines of a design, partly perhaps with the idea of economising draughtsmanship, but chiefly with a view to the value of the steadiness of effect to be obtained by that means. They permit themselves, how- ever, in that case (or they lay themselves open to the charge of rather niggardly invention), consider- able variety of detail within those stead}^ lines. When rigidly exact repetition is no part of the con- ditions imposed by manufacture, it is almost incumbent upon the designer to assert his freedom, and not, for example, to suggest that his printed pattern is woven. He does well to avoid making one side of his design a mere reflection, as it were, of the other; and in particular its too mechanical turnover at the axis (147).
The absolutely strict turnover of any but the most rigid pattern, especially when the main stem is its axis, is so unsatisfactory that weavers often arrange their looms so that
145. BYZANTINE TURNOVER " PATTERN.
THE TURNOVER.
there is a central space of some inches (S) in which there is no turning over (148).
There is no occasion or excuse for the objectionable mechanism when, as in printing, the conditions do not compel it.
146. GOTHIC " TURNOVER " PATTERN.
Although it adds greatly to the interest of a pattern in which the main lines are reversed to introduce, if the con- ditions allow it, variety of detail, it is not safe to take Uberties with the Hnes themselves or with the proportions of the opposite masses, else the balance of the design, which it is most important to preserve, may be lost. It is essential, too.
PATTERN DESIGN.
147. QUASI " TURNOVER " PATTERN.
THE TURNOVER.
123
that, for example, anv two opposite features should be pre- cisely opposite, and that their branches, curving from the central stem or towards it, should turn, like the spirals on page 124 (149), on precisely the same level. Inaccuracy in either of these respects, though it may pass in a dramng for artistic freedom, is almost sure, in repetition on the wall, to give the impression that it is out of the straight. The eye expects a level; and it is strange how slight a deviation from it will produce an unfortunate effect. And so with any departure from the upright. lu what concerns the equi-
ttirnooet
£> '^-^-^ turnover
1 48. TURNOVER DESIGN WITH CENTRAL STRIP NOT TURNED OVER.
124
PATTERN DESIGN.
149.
TURNOVER AND DROP PATTERN".
librium of a pattern it is impossible to be too mechanically exact.
A common device in design is to turn over the iinif of design as in the diagram opposite (150), but that does not constitute what is known as a turnover repeat unless it turns over on the same level.
The " turnover " is nowhere more valuable than in border design. It is a most useful means of stopping the flow of the pattern, and of giving the lines across. \N'hich go so far towards the stabihty continually demanded in a border. Such lines
THE TURNOVER.
125
^ ) ^r ^- "^^ ' ^t ^'- -^r ' -^rf ^^- -^ ' ^(
150. NOT A TURNOVER REPEAT.
may be expressed or understood: the counterpoise of parts suggests the axial line even when it is not put down.
A sprig or other j)attern for wall decoration may just as easily be turned over as not. A stencil or a pounce has only to be turned face to the wall to give the design in reverse.
Q^i^
wpr^
or r
151. DIAGRAM SHOWING SQUARE REPEAT WHICH TURNS ROUND.
XII. THE TURN-ROUND.
Unit of design may be turned part way round — Unit of 6 by 6 inches results in repeat of 12 by 12 inches — Works either on the straight or as a drop— For radiating pattern a triangle half the size of smaller square suffices for unit — Fold and fold again— Arab lattice pattern dissected.
Ix designing for tiles and such like the condition of con- tinuity obvious in the case of woven pattern no longer exists, and possibiUties occur which are denied to the weaver. The repeat of a 6 inch tile, or of the two or more 6 inch tiles which go to make the complete pattern, need no longer be always in the one direction. The designer is free to devise a unit which has to be turned completely- round in repetition, or half-way round, or three-quarters of the way; he can, conse- quently, out of a 6 inch unit get a design which will not repeat on a straightforward trellis in less than four times its area.
In this way the repeats above (151), supposing them to be 12 inches square, could be got out of a unit onW 6 inches
THE TURN-ROUND.
127
^ff^'.:=^>^.
152. DIAGRAM SHOWING SQUARE REPEAT WHICH TURN'S ROUND AND DROPS.
square, provided it could be turned round (as a tile could be) in the way above described.
In the case of a pattern repeating on horizontal lines, the design might extend {151) beyond the lines of the repeat. In the case of one that stepped (152), it would be necessary to keep within the four square lines.
If that were so, the pattern could without difficult}^ be schemed to work, not only as a drop, but on the straight also; and, as a matter of practice, many tile patterns are so designed.
In a tile pattern such as that (153) overleaf, which ra- diates instead of following round, assuming the squares to measure 6 inches, and the pattern 12 inches across, the unit of repeat (except for the interlacing of the lines, which is no part of its construction) reduces itself to a triangle half the size of the square -or rather, that being itself a " turnover " again, to one a quarter of its size.
The building up of such a repeat on diamond lines is on the face of it apparent.
It is an Eastern practice (I have been told by Sir Caspar Purdon Clarke) to design on the lines of a sheet of paper folded in parallel lines, and folded again in lines at right angles to those, and then again in the diagonal direction — a
128 PATTERN DESIGN.
practice which one ought ahnost to have divined from the nature of the patterns resulting from it.
The Arab lattice opposite (154) is just such a pattern. Or it might be built (on the lines very similar to those shown in diagram 38) of octagons, the centre of which is marked c, and four pointed stars, of which the centre is marked by four dots; or on the zig-zag Unes which give those shapes. It repeats also on the lines of a rectangular diamond, the points of which occur at c; or of a parallelogram a b which drops half its length; and as a is only the reverse of b it works also as a turnover pattern. Further than that, b is actually
153. TURX-ROUND PATTERN.
^^msmm^Mmi^MMmi
154. ARAB LATTICE AND THE LINES INTO WHICH IT MAY BE RESOLVED.
I30 PATTERX DESIGN.
the same unit as a, merely turned part way round until what in A was the top is in b the side of the square. The design would therefore work as a square tile of the dimensions of A. Patterns of somewhat similar construction, even more plainly to be set out upon the lines given by folding and folding again, are shown on pages 50, 52, and 53, all of them typical Arab lattices.
XIII. PATTERN PLANNING IN RELATION TO TECHNIQUE.
Dimensions of design determined by conditions of manufacture — Possibilities in block printing — Limitations in weaving — Narrow re- peat a condition of the loom — The "turnover" — A space of "sin- gle " — Borders — Table damask — The lengthening piece — Difficulties resulting — Conditions affectmg colour — Change of shuttle — Its use and danger — Carpet weaving — " Planted " colours — Chenille — Cha- racteristics of stvle accounted for bv technique.
It has been shown how the pattern designer is practically compelled to design, not precisely on square lines, but on the lines of a parallelogram. And not only that. The distance of the lines apart is almost certainly laid down for him. It is a parallelogram certainly of restricted size, and possibly of arbitrary proportions, with which he has to do. Without uniformity in the width of stuffs — silks, velvets, carpets, chintzes, or whatever they may be — it would be difficult to estimate off-hand their relative cost; and estimating is a matter of everyday necessity. Without stock sizes of tiles, the price b}-^ the yard, and the cost of fixing them, would not be easy to settle.
The width of stuffs is determined, if not by mechanism, by custom and convenience. The length of a woven pattern is restricted by considerations of economy, and that of a printed one bv the girth of the roller, or the size of the block it is convenient to handle; so that in a vast number of cases a designer has to work within conditions which fix for him, not only the size, but the proportions, of his design. It
132 PATTERN DESIGN.
resolves itself into his working within the lines, say, of a parallelogram 30 by 15 inches for printed cotton; or 21 by 21 inches (at most) for wall-paper; for tiles, wdthin a scjuare mesh of lines 6 or 8 inches apart. And he is free only within such limits. Theoretically, it is true that a design for wall-paper may be spread over an area involving any nmnber of blocks; as a matter of fact, it is not. The designer is occasionally allowed in the case of sumptuous papers, and of certain single prints, a repeat of 42 inches long; but patterns spread over a larger area than that would cost more to produce than paper- hangings are usually worth. And, over and above the com- mercial consideration (which is in itself enough to prevent that kind of extravagance), it is a point of craftsmanship not to waste labour. It is the test of a designer's capacity that he should not ask for further facilities, but make the most of what the conditions offer him.
The mechanical conditions of block printing permit certain extensions of plan which roller printing does not. It is possible with a single block 21 by 21 inches to print either a radiating or a turning-round pattern which in the hanging shows a repeat measuring 42 inches each way.
Imagine the square lines in diagram 155 to be 21 inches apart. Tlie unit contained in one of the divisions A stands for what the block will print. The printer has onlv after printing one impression (A) to give the block a twist round l)efore printing the next (<) to get the result shown at the bottom of the diagram, which represents also the width of the paper. As yet, however, we have only half a pattern. It remains with the paperhanger to set that right. He hangs ever\7 other strip as it were upside down (V < ) and the com- plete pattern results on the wall.
The design given on a smaller scale on page 134, in which also the repeat is actually 42 inches across, is got out of a single (21 inch) block in the same way.
Further it is possible by means of two 21 inch blocks to
>•. ^
^%-
i
155. DIAGRAM TO SHOW A HALF TURN OF THE BLOCK IN PRINTING
A CEILING PAPER, AND THE REVERSAL OF ALTERNATE
STRIPS IN THE HANGING.
134
PATTERN DESIGN.
print a pattern of which the repeat works on a rectangular diamond measuring 84 inches from point to point. In this case, however, the design must radiate, and not turn round — or, when the alternate strips came to be hung (as to complete the pattern they would need to be) opposite ways about, the design \\ould not run on.
The diagram opposite (157) shows four widths of paper. In the strip to the left A and r, represent the prints from the two 21 inch blocks, ::; and < prints from the same blocks t^visted round. In the second strip av > :: a strip precisely
MnVS^ '■^fC3<>^^ 5i(^Y5)"^^
; PI)
•G
150. DIAGRAM OF CEILING PATTERN (iN EFFECT, 42 INCHES across) on the PRINCIPLE OF DIAGRAM 1 55.
strip i
/I
©Vfl
i \ ; -T (T)
rv^iB
137. DIAGRAM SHOWING A PATTERN IN EFFECT 84 INCHES WIDE PRODUCED BY 2 BLOCKS EACH ONLY 21 INCHES BY 21 INCHES.
136 PATTERN DESIGN.
similar to AB w< upside down, the hanging is so schemed tha't a is on a level with A and v with b. The third strip is hung the same way up as the first, but so as to drop 42 inches below that. The fourth strip is again the same way up as the second, but so as to drop 42 inches below that.
All this would be difficult to follow in print, but for the diagram. With that to refer to, it is easy enough.
Patterns of this character are not wanted in wall decora- tion; but for ceilings they give not only a sufficient scale, but just the lines which are most serviceable.
These devices by no means exhaust the possibilities in the way of cunning contrivance. But the block printer does not look kindly on designs which ask of him a little extra care — and as for the paperhanger, he is persuaded that the use of his brains is no part of his business. Indeed that scheme last explained is already too intricate to have been put into practice, which it might easily have been if only the paper- stainer could have depended upon the goodwill of the paperhanger.
The designer of wall tiling has every reason for scheming his repeat to work on the brick system (158).
The material for which a man is designing settles, in a measure at least," both the dimension and the proportions of his pattern. Thus, for a printed fabric the roller commonly allows him an area twice as wide as it is deep. For wall- paper the block allows him at the most a square of definite dimensions, except that he may on occ-isioii be free to use two blocks. For a woven fabric the loom gives him a consider- able length of pattern not greatly restricted by expense, but usually only a narrow width, precisel}^ fixed according to the loom, and affords him very hkely the opportunity of doubling the width of his design by turning it over. So uniformly are these conditions so, that an experienced designer can often tell, from the proportions and scale of a design, the kind of manufacture for which it was made. The copyist, on the
PLANNING IN RELATION TO TECHNIQUE. 137
138. SIX-INCH TILE DESIGNED TO BE FIXED BRICKWISE.
other hand, who finds a pattern which has apparently been overlooked, and thinks to appropriate it to his own use, dis- covers perhaps, before he has done with it, that there was good reason why it had not already been annexed— inasmuch as it depends upon proportions which the machine, to the requirements of which he desires to adapt it, will not permit him to preserve. Even among a designer's own happiest thoughts there will be some which (if he works only in one material) must, for much the same reason, be stillborn. A new set of conditions start a man off in quite a new vein. In the design on page 138 the width of the material is
159- NARROW TURNOVER PATTERN ADAPTED TO WEAVING.
PLANNING IN RELATION TO TECHNIQUE.
139
indicated in the central strip, where the background is filled in. If that were wall-paper, 21 inches wide, it would take no less than four full-sized blocks to print it — which would not be worth while. If it were a woven stuff, the long repeat, though adding to the expense of production, would not be very much against it. As a matter of fact patterns of that relative length often occur in textiles. The one on this page is again a turnover (160).
The narrow pattern overleaf (161) is a wall-paper design which is only 10 14 inches wide. In wall-paper printing there is no economy in this as there would be in cotton printing — but artisti- cally there may be very good reasons for using sometimes only half the width the block allows.
The weaver adopts the long and nar- row repeat all the more readily that he has a handy means of counteracting its too upright tendency. The cross stripes which form the pattern of an Indian dhurri or an African blanket represent the handloom weaver's simplest means of changing colour — that is, by changing his shuttle. In more elaborate pattern he has the same facility, and can always cross his upright strips b}' bands of colour carrying the eye in the other direction. And this scheme of banding extends through much of the early weaving, affecting also the form of the design.
1(30. NARROW WOVEN DAMASK PATTERN.
I40
PATTERN DESIGN.
l6l. WALL-PAPER PATTERN- IQI^ INXHES WIDE BV WHICH NO ECONOMY IS EFFECTED.
PLANNING IN RELATION TO TECHNIQUE.
141
A Byzantine or Sicilian weaver of old was the more inclined to make use of the horizontal lines suggested by the shuttle, because he had no fear of their asserting themselves. In fact he was in the habit of insisting upon them, for he valued the stripe as a means of marking the folds and show- ing the fulness of a hanging. They do that so effectually that a flat wall-design in horizontal stripes seems to want folds, and to suggest tliat it was borrowed from a textile. Many a pattern borrowed from an old stuff — by its stripes you shall know it — is far from satisfactory as a wall-paper.
The proportions allowed for the repeat naturally affect the character of the design. You cannot without considerable allowance in the way of length indulge in boldly flowing
ttirpouer > single -^ — tCvrnover
162. DIAGRAM OF WOVEN P.\TTERN, " TURNOVER " AND " SINGLE. "
142
PATTERN DESIGN.
Ttirr>ou«r-><
■"■■ wm
Tttrnooct- isinj)!*. Ttirnoocr
163. DIAGRAMS SHOWING CORRESPONDENXE IN DEPTH OF REPEAT BETWEEN THE PARTS OF A DESIGN.
scrollwork; nor, where the width is narrow, avoid a certain upright tendency in the growth of pattern — counteract it as you may by cross bands.
The weaver's custom of reserving in the centre of a turn- over pattern a space in which the design is not reversed has been already mentioned (page 120). By that means the stiff- ness of a definite upright line, the formality of mere reversal, and the obtrusiveness of what is after all a mechanical device, are avoided. A loom may be so harnessed, and commonly is so harnessed (162), as to allow the designer a space up the centre of his curtain (or of the repeat of it) in which he is free to do as he pleases so long as this central part of his design joins on at the sides to the two broad \\ings which make up the main portion of his design. The same thing applies equally to the design of a border. In diagram 162 the border on one side is a turnover of that on the other; but
PLANNING IN RELATION TO TECHNIQUE.
143
the turnover might equally well be within the border itself. Filling and border pattern, that is to say, niay alike be turned over; and in each may be reserved a central strip of what is technically termed " single. " The \ndth of that portion of the design is a question of arrangement and partly of cost. An important consideration to be borne in mind is that the introduction of any proportion of single design is at the cost of possible increase in width.
The device of turning over gives one, for example, double the width otherwise allowed say for a border. Instead of a " free " pattern 9 inches wide it allows a bisymmetrical one of
164. DIAGRAM SHOWING RELATION OF DIAPER REPEAT TO REPEAT OF SIDE AND BOTTOM BORDERS.
144 PATTERN DESIGN.
i8 inches. But it is only so much of the width as is turned over that is doubled. If, then, you reserve let us sav j inches in the centre for " single, " the extent of your border would not be i8 inches but 15 — the sum that is to say of 3 inches (single) and twice 6 inches (turned over). The technicjue of wea\'ing has here, it will be seen, considerable bearing not merely upon design but upon its plan.
Single and turnover portions of a design must naturally correspond in length. They need not of necessity be equal as at A (163), but it is practically convenient to make them so. The design b could of course be woven; but, if the repeat of the turnover measured 9 inches, a manufacturer would not allow 18 inches for the single. In the same way the repeat of a border must naturally correspond at the side of a curtain with the depth of a filling pattern, and at the bottom with its width (162, 164) — and if part of the filling is single, the corresponding portion of the border also may be.
In the case of a narrow and not very important border it may be shorter than the filling as long as its length is divisible into that — a 9-inch filling may ha\-e a border of 41/2 inches, or 3 inches, or lYo. And so where the border is the main feature and the filling a mere diaper (164), that may measure only half or a third or a quarter of its length. And were it mechanically possible to weave border and filling the repeats of which measured respectively say 7 and 9 inches, it is doubtful if artistically it would be worth while: the simple thing to do is commonly the right one. It is some- times desirable to make a curtain or other pattern complete in itself with start and finish (165). For working purposes these may be regarded as borders, and must conform to the conditions regulating border design.
Further complexities occur in the design of table Hnen. The conditions seem at first sight to allow great freedom to the artist. He has only to design a square or oblong cloth each quarter of which is a turnover of the other, and in the
PLANNING IN RELATION TO TECHNIOUE.
145
bortl<?rk— 3 repct^te oP Filling
■^ I border.
^
165. DIAGRAM SHOWING PORTIONS OF DESIGN ANSWERING PRACTICALLY TO BORDERS.
centre he may have a space where there is no repeat. That is charming in theory. In practice his task is not so simple. It is compUcated by the necessity of arranging some means of lengthening the cloth to suit tables of various dimensions. And it resolves itself into his having to design a lengthening piece (usually of 9 or 18 inches) which must be so schemed that it can be inserted once, twice, thrice, or an^' number of times, to make a cloth of anv length. The scope which the nianu- facturer gives with one hand he thus takes back with the
146
PATTERN DESIGN.
^
>
18 in.
•>
— >
i
pieces
-I
<r
tSio
tuTOOuer
pieces
i
£iT)C)t€
centre oF
5>qti^re
clotb.
5ii39t6
t
turnover
r
Border I
3 i 4
5 i 6
7 ! 3 i
Border
9 ;io
166. DIAGRAM SHOWING HOW A TABLE DAMASK DESIGN MAY BE PLANNED.
PLANNING IN RELATION TO TECHNIQUE.
147
167. DIAGRAM OF TABLE DAMASK DESIGN PLANNED ON THE LINES OF DIAGRAM 1 66.
148 PATTERN DESIGN.
other. To such an extent is his freedom restricted that the artist is inclined at first to think his possibiHties are narrowed to httle more than the extension and finishing off of the design for a lengthening piece.
Where it is not desired in any way to acknowledge the centre of the cloth the problem may be resolved into the design of an 18-inch repeat (reversed or not in the centre) merely finished off at the edges or cut short by the border.
Where it is desired to give importance to the centre of the cloth the loom can be arranged so that there is no turning over there; but if the end portions are turned over it is difficult to scheme a growing pattern in which the stems do not grow two wa3's. Again if the artist is disposed to take advantage of the area allowed him to get good sweeping lines in his main design, his ardour is damped by the reflection that he must somehow combine them with the comparatively re- strained lines which are all that is possible in the lengthening piece. This affects the border in particular very seriously. Try to introduce into a sweeping scroll design a yard long a lengthening piece of half a yard, and you will realise the impossibility of it.
Counsels of safety are: to confine oneself in the main design to lines such as can be repeated in the lengthening piece: to allot spaces at least in the design to sprigs, sprays, or disconnected diaper: to avoid, like the Arabs or their imita- tors of the Renaissance, growth so natural as to be hurt when it rs suddenly doubled back or made to grow two ways. To take full advantage of the apparent opportunity of design afforded by the dimensions of an ample tablecloth, and at the same time to preserve something like logical growth, is what any but an experienced damask designer will find it difficult to do.
The accompanying diagrams (166, 167) may be of use to the beginner. The first of these is divided, it will be seen, into ten divisions each measuring 9 inches (tablecloths are always measured by quarter yards), two of which are given to the
PLANNING IN RELATION TO TECHNIQUE. 149
border and two to the single piece up the centre, which leaves two for the turnover piece between.
The plan sho^^s three quarters of a square " ten-quarter " cloth (a smaller size is " eight-quarter "), and, above, to the right, one quarter of a cloth into which two lengthening pieces are introduced.
The corresponding diagram (167) shows the beginnings of a pattern planned on similar lines, but with the two lengthen- ing pieces inserted, one above and one below the centre.
An 18-inch border practically represents that portion of the cloth which may be presumed to fall over, and the central six quarters the portion which will lie flat on the table. Any extra border v.ithin that space is reckoned as part of the filling; anv part of the filHng which extends beyond the six-quarter area is reckoned as border. The lengthening piece or pieces need not be introduced as shown in the diagrams above; they may come in the centre of the cloth.
The plan more usually adopted by damask designers is to halve the design, open it out, and let in the lengthening pieces. Diagram 168 (overleaf) represents a square which might stand either for an eight-quarter cloth or the centre portion of a larger one. Below it (169) is the lengthening piece, and opposite (170) the result of opening out the square and letting in two lengthening pieces.
The changing of the weaver's shuttle, responsible for the stripes in a dhurri, gives scope to the designer of more sump- tuous and less simple fabrics. There can be no more colours in a stuff than there are threads of different colour in its make. But each group of threads may be brought to the surface at the option of the designer — and, if for any group or groups of threads he prefers to use instead of a single colour alternating bands of different colours, he can do so — and if these particular colours do not come often to the surface, he can get as it were jewels of extra colour without calHng attention to its occurrence in bands — but
I50
PATTERN DESIGN.
•^ ttvrneoi^ev
J ^ J
'9^
•tf'e*
w
si
0
(ft
c
I
t
1 68. DIAGRAM OF CENTRE P.\RT OF SQUARE TABLECLOTH.
169. DIAGRAM OF LENGTHENING PIECE TO CORRESPOND WITH
ABOVE.
PLANNING IN RELATION TO TECHNIQUE.
151
tno
?tG\^S(^;
:^K
170. DIAGRAM OF CEXTRK PART OF LON'G TABLECLOTH SHOWING INTRODUCTION OF LENGTHENING PIECES.
152 PATTERN DESIGN.
it takes some ingenuity to do that. The stripes have a per- sistent way of asserting themselves. Successfully to divert attention from the mechanism imderlying such a distribution of colour is within the scope only of an expert designer. His task is easier if he is free to gradate the various colours so that they die one into another or into the ground; but even with flat colours a man who knows his trade can effectively disguise the means employed to variety.
The kind of variation possible is illustrated in diagram 171, where the strawberry blossoms are successively of three different tints, indicated in black, in dots, and in diagonal lines, and the changes of the shuttle are very plainly shown in the bands at the side which may represent the selvedge.
What one weaver does with the weft another does with the warp. The carpet designer, working for a material of which the warp comes always to the surface, does by the arrangement of his warp threads in bands what another weaver does by changing the shuttle. In a " five-frame " carpet five series of warp threads are brought to the surface and give a design in five colours, but if in one of them (or it may be two or even three " frames ") the threads instead of being all of one colour are arranged as it were in ribbons of different colours, these various colours can just as easily be brought to the surface as threads all of one colour.
According to the number of stripes in which the threads of a " frame " are arranged is the number of the colours to be got out of it. But, as in the case of the changing shuttle only the colour of that one shuttle could possibly occur in the line across which it was shot, so in the case of the warp threads no one colour in a given frame can cross the path of another — it occurs only in the line of the underlying stripe.
Diagram 172 shows one frame of the warp divided into six stripes which give only three colours. They might just as easily have been six, and the\^ might each of them have been gradated from light to dark or from warm to cold. That
PLANNING IN RELATION TO TECHNIQUE.
^53
would have made with the other four frames ten colours in all — as it is we have seven, onl}' four of \\'hich the artist is free to use as he ^^dll.
There is no real difficulty in scheming a pattern to meet such conditions as these. And even in the more complicated case where two frames or more are thus divided into stripes all that the designer has to do is to make sure that his " planted " colours, to use the trade term, do not exceed their
mm^:^'^.
a*3
171. DIAGRAM SHOWING CHANGE OF COLOUR IN THE WEFT THREADS.
154
PATTERN DESIGN.
172. DIAGRAM SHOWING CHANGE OF COLOUR IN WARP THREADS.
bounds. This he can easily do either by ruhng guide Unes on his drawing or by the use of a strip of paper painted with each colour in its order and proportion which he can move about as a gauge. With a clear head he should have no great difficulty in keeping one colour out of the way of another.
PLANNING IN RELATION TO TECHNIQUE.
155
He is not ahvays careful to make his details correspond precisely with the colour stripe, and by the occasional over- lapping of the form by a colour not apparently belonging to it, or vice versa, the appearance of abrupt transition is avoided. It looks as if mistakes in gauging had resulted occasionally in happy effects of confusion, and that the device had since been emplo\^ed deliberatel}'.
It will be seen (172) that it is mainly in the flower centres that the planted colours are used — the mass of the flower itself carries the eye far beyond the spots of colour, w-hich might otherwise run into stripes. The idea is, of course, that in the confusion of flower, leaf, ground, and outhne colours, the order of these jewels of bright colour shall not be too apparent.
i
1^1
73. DIAGRAM SHOWING REVERSAL OF DESIGN IN CHEXII.LE WEAVING.
156 PATTERN DESIGN.
In the diagram they are purposely insisted upon, and the foliage is barely indicated. In the woven fabric the form of the pronounced foliage would help very much to give that mystery of effect which is at timics so valuable. The com- plicated mechanism necessary to the fretjuent changing of the shuttle in powerloom weaving leads in many modern fabrics to the use of a number of warps, any one of which can be brought to the surface wherever the colour of it enters into the design. If yet more colours are wanted, they may as already explained (page 152) be " planted. " Each additional ^arp adds natur- ally to the heaviness of the stuff.
A very exceptional facility is afforded by the process of chenille weaving. The design may extend right across the curtain — and the repeats need not follow one above the other in the usual way. Each alternate one may, if it is desired, be reversed. The repeat of the design (173) on page 155 would in the ordinary way include two groups of llowers (a, b), and there would be no economy in making one the reverse of the other; but in chenille weaving there would; and the repeat is comprised in the unit ^, of which b is the reverse.
The characteri.stic lines of time-honoured patterns are for the most part the direct result of the restrictions under which the designer was working. Fashion has had her say in the matter no doubt — it is a wicked way she has — but, though certain lines of design may have become associated in our minds with a particular period or country, it will be found, I think, that there was alwa\'S some technical or practical reason why in the first instance they were adopted. Appropriate pattern lines do not come of themselves — growth and fitness go together.
Pattern design has always been and \\ill always be con- siderably affected by considerations which never occur to the uninitiated.
174- WOULD-BE FREE PATTERN FALLING INTO THE LINES OF THE BRICK REPEAT.
XIV. HOW TO SET ABOUT DESIGN.
Free patterns planned on formal lines — Features recur at intervals determined by unit of repeat — Planning the only way to avoid unfo- reseen effects — Means of disguising formal lines — Necessity for system ■ — Genesis of counterchange border — of geometric diaper — How not to do it — Detail not to be determined too soon — -Genesis of conventional floral pattern starting with the masses — of a drop pattern — of a pat- tern starting with line — of a floral pattern starting with distribution of flowers — of a velvet pattern starting with severe lines — " Inha- bited " pattern — Evolution of Italian arabesque pilaster — Animal form in pattern — Starting at a venture — and from an idea — -After- thoughts.
(tEOMETKic patterns have as a rule much less reticence in exposing the lines of their construction than others. You see more plainly in them the various plans of construction upon which such stress has been laid. The freest and loosest of patterns will be found, however, to repeat as geometrically as the severest, and on precisely the same lines: it is for that reason so much stress has been laid upon geometry. A flowing pattern does not Bow so freely as might be supposed. Mark any recurrent feature in it — and four such features will
158
PATTERN DESIGN.
give you points from which may be drawn the four straight Unes which mark the square, or parallelogram, or diamond upon which the repeat works. It ma}- be doubted whether the ^?/^7J-/-pictorial French wall-paper (174) on page 157 was planned upon the lines of the brick,* but it falls into them, and the masts of the ships practically give the vertical divisions of the plan.
Each and every feature in a design recurs at intervals determined by tlie proportions of its unit. Let your uidt be a square, for example (175), and, in a cluster of four squares, any given detail will mark by its recurrence the proportion of the square, no matter whereabouts in the square it may occur. The diagrams below show this.
1 I 1 1 I 1 1 • •
• ^ ^
• •
: — -i ; — I ■ — 9 »
-- • • , ; ;
75. DIAGRAM SHOWING HOW RECURRING FEATURE MARKS THE PLAN OF A PATTERN.
The recurrence of the details of the pattern is a certainty. It is as well to make certain of the sequence in which they shall recur. Any reliance upon haphazard at the beginning is sure to give trouble in the end. Happy-go-lucky arrange- ments seldom work out happily; there is no reason whv they should.
A painter may, and often does, go jauntily about his work and put in a diaper upon a screen behind his figures without taking the pains to plan it ; but the further he goes the wider he gets of accuracy*, and the more plainly his carelessness is revealed. In the diaper opposite (176), for example, the " repeat " does not repeat. This matters nothing in a painting. It even gives the painter an opportunity of
* See page 71.
HOW TO SET ABOUT DESIGN.
159
adapting the pattern to his pictorial needs. In a design for practical purposes it would matter everything. It would be in fact not a design but orJy a suggestion for one.
A designer, like other artists, trusts largely to his instinct; and rightly relies upon it for artistic prompt- ing throughout his work; but it will not supply the place of order, to which in the nature of things he is pledged. He is free only within the limits of his reheat — practically a right-angled space, or a diamond of given dimensions.
Suppose it to be a square. Within the four sides of that he may do as he likes. He may sprinkle sprigs about in the most admired disorder. There may be no more geometric relation between them than between the six black spots in the central square over- leaf (177); but, where there is no geometric relation be- tween the members of the group, it is not easy to anti- cipate, as a designer should, what will be the effect of
the group itself when it comes to be repeated. It will be seen that in repetition the spots fall into irregular Hues with awkward gaps between— just the kind of line which comes by accident, and might easily have been avoided by careful contrivance (see page 5).
176. PAINTED GOTHIC DIAPER IN
WHICH THE " REPEAT " DOES
NOT WORK.
r6o
PATTERN DESIGN.
For want of more systematic planning the pretty damask pattern opposite (178) falls into stripes which, it seems to me, the artist did not foresee. They are comparatively harmless there — and would be equally so in a table damask — but in a wall pattern, for example, they might assume distressing prominence.
A stripe is by no means necessarily to be avoided in design — and it is in obedience rather to the prejudices of a timid public than to their own artistic instinct, that designers avoid frank lines. Artists know how useful they are. But they should be the lines that play their part in the pattern; and, to do that, they must be well considered; not left to chance: the chances are all against a happy fluke. One way out of the difficulty is boldly to insist upon the stripe and make a feature of it. Another is to cover the ground with
^
^
^
^
s
«*
•
^
€
•s>
^
^
*
^
s. ^
^
^
177. DIAGRAM SHOWING GEOMETRIC RECURRENCE OF FEATURES NOT GEOMETRICALLY DISTRIBUTED.
HOW TO SET ABOUT DESIGN.
i6i
OLD LINEN DAAL\SK PATTl.RN FALLING INTO STRIPES.
uniform pattern in whicli is no break and no feature more prominent than another— in the manner of the daisy pattern overleaf (179) — and that is an exceedingly difficult thing to do. But the way to do it is, not gaily to scatter daisies about upon the paper, but to plan them (see page 89 et seq.), and, even then, the uniform covering of the ground involves an amount of experiment and reconsideration, which is in itself enough to explain the comparative rarity of such patterns. Unpre- tending thev mav be; but there are occasions when what is wanted is simplicitv \erging even upon insignificance, and where yet obvious geometric forms would not do. Hence
1 62 PATTERN DESIGN.
the need for all-overish ornament — pattern which is meant to break a surface or a colour and not much more. Even then it is not a bad plan to introduce into it features such as the circular groups of flowers opposite (i8o), and the little flowers in upright pairs betw'een, which, though in a measure lost in the even distribution of detail, may be relied upon to assert themselves, if anything catches the e^'e at all; and these are planned, of course, with a view to their effect in repetition. The difficulty and danger of design is lessened in a pattern in which there are suclt points of emphasis (however
HOW TO SET ABOUT DESIGN.
163
180. FEATURES DISGUISED BUT NOT QUITE LOST IX " ALLOVER "
EFFECT.
slight), features balanced one against the other, supporting it may be or counteracting one another, and yet producing an effect of even weight; or in a pattern in which there are marked governing lines, whether symmetric and plainly re- vealed, as in illustration iSi, or flowing as it were freely, and partly lost in scrollwork or leafage. To lose the lines of recurrence altogether, as in illustration 182, is not easy.
Insistence upon the necessity of governing lines in pattern must not be taken to imply that they must always be in- sisted upon or that they may not assert themselves too strongly. There is a point at which they are an annoyance.
It may be expedient to subdue them — even to efface them at times. The necessary subjection may be effected in various ways. They may be arbitrarily interrupted. They may be overpowered by detail, not perhaps very significant in itself. Two or more schemes of design may be interwoven,
1 64 PATTERN DESIGN.
the one asserting itself here the other there, and each calling attention from the other. The lines themselves may be so ingeniously interlaced that it is hard to disentangle them. Some of them ma3'' be traced merely in outline, hardly strong enough to hold its own against more substantial features, or in a colour having more affinity with the ground than with the ornament generally.
But the most usual way of disguising the skeleton is, taking the hint from nature, to clothe it with something in the way of fohation — by which the bare constructional lines are as effectually hidden as the branches of a tree by its leaves. By this means the spirals of a scroll can be made to assert themselves as much or as little as occasion may demand. Only if the curves are not well-considered it is hopeless to try and make up for that by foliation, to dis- guise bad lines by leafage. A broken-backed scroll betrays itself beneath it all. There is no disguising its native in- firmitv. Pattern is vertebrate; and in a scroll the spinal cord is very plainly pronounced.
As to whether it is better to reveal or to disguise the construction of pattern, to insist upon it or to call attention away from it, that is a question to be answered partly accord- ing to the temperament of the designer, partly by the circum- stances of the particular case. Either plan is best upon occasion. But it is a point upon which the artist should in every case make up his mind at once. He should know what he is going to do, and do it deliberately.
Referring to the popular prejudice against anything like formality in design and especially against anything which " you can count, " as they say, the public has a riglit to call the tune it pays for, and will no doubt get what it wants. If it will have nothing of severity or restraint in pattern, so much the worse for design. If, however, any student of ornament should feel that waj^ so much the worse for him, or for his chances of success in this direction. His wiser
l8l. DESIGN IX WHICH MARKED GOVERNIN'G LINES STEADY THE EFFECT.
1 66
PATTERN DESIGN.
182. WALL-PAPER PATTERN IN WHICH THE LINES OF RECURRENCE ARE PURPOSELY LOST.
course would be to turn his attention to some branch of art for which he has more aptitude: he lacks the instinct of pattern design. A wilful world will have its way. An artist should know that, in sacrificing everything in the nature of formality, we renounce much of the dignity which belongs
HOW TO SET ABOUT DESIGN.
167
to the best in whatever form of art. The finest of old pattern work is invariably formal and owes to formalitv something of its noble character.
Apart from that, it simplifies, as was said above, the problem of design, to accept the recurrence in it of a feature more or less plainly marked. And it is not altogether a matter of choice. In any design not absolutely all-overish one feature, or some features, must be more emphatic than the rest. Over-emphasis is provided against by points of lesser emphasis, to balance them, and points of lesser weight again to balance these perhaps. By the careful balancing of parts, it is possible, if not easy, to draw off attention from any formal plan. Indeed to such purpose has the art of hiding art been exercised in this respect, that the advocates of " go as you please, " seeing in some good patterns no evidence of construction, are not to be persuaded that they were ever built upon a plan. They may take the word of a designer for it that they were.
The dress pattern below (183) is of the class called free.
l8j. " FREE " DRESS PATTERN' DELIBERATELY PLANNED.
1 68
PATTERN DESIGN.
But it was as deliberately set out as if the geometric con- struction were conspicuous in it. If you detect no formal lines in it, it is not because the plant was allowed to trail as accident would have it— there was nothing free and easy about its disposition — but because the lines of growth were from the first schemed with a view to seeming freedom, and the details were so plotted as to divert attention from the svstem upon which they are distributed. The system is there.
If you would avoid the unforeseen in your completed work — and the unforeseen reveals itself often in the most unsatisfactory manner — system is essential.
A practical designer does not idly let the pencil in his hand meander about upon a sheet of paper, in the vague hope that something may come of it. He starts ^vith a definite notion of some sort — a happy thought, an image in his mind perhaps, or, if not that, the idea at least of the sort of thing he wants, the thought of certain lines or masses, or the combination of the two, which promise when repeated to make pattern.
184.
DIAGRAMS SHOWING DEVELOPMENT OF A BORDER PATTERN.
COUNTERCHANGE
185. GEOMETRIC DIAPER PLANNED AS IN DIAGRAMS BELOW.
|
A |
B |
A |
|
B |
A |
B |
|
A |
B |
A |
A
|
B |
B B |
rx. |
186. DIAGRAMS SHOWING STAGES IN THE DESIGN OF A GEOMETRIC
DIAPER.
PATTERN DESIGN
1S7. DIAGRAM SHOWING A FALSE START IX DESIGN.
The lines upon which a design is planned need not, it has been explained, form any part of the pattern. But, if thev do, it is easier to trace the steps by which it came to its effect.
Take the simplest of patterns, a border (184) in which the repetition is only in the horizontal direction, and begin with a wave line down its centre dividing it equally into two halves, the one white the other black (a). Following the lines of the wave on the one side, and of the margin on the other, we arrive in the simplest way at a sort of double wave giving a white enclosure in the black space and a black one in the
HOW TO SET ABOUT DESIGN.
T71
white (b). To turn these into flowers (c) and to give them stalks to connect them with the waved hne is an obvious thing to do; and so we arrive, almost before we know it, at a complete and consequent counterchange pattern.
The genesis of a geometric diaper (185) is scarcely more difficult to trace.
The initial idea worked out on page i6q (186) was a flooring pattern, planned upon the square therefore, or rather, as it happened, the double square working as a drop. The double square and the desirabihty of retaining the square form, suggest an equal-sided unit, merely turned about, to
188. DIAGRAM SHOWING THE START OF A DESIGN.
172
PATTERN DESIGN.
1 89. DIAGRAM SHOWING DEVELOPMENT OF l{
mark the double square, and (in the flooring) to prevent the effect of lines in one direction. But though it was advisable to retain the square form it was not desirable that it should assert itself too prominently. The pronounced additions to, square a in the second of the smaller diagrams (186) effectually prevent any such danger; and, repeated in the cross direction in squares b, they give in the fourth diagram already' a co- herent pattern. But it is empty, and the proportion of light and dark is not what was wanted. A central disc of black upon the white puts that right, and the continuation of the curved lines in the direction of the disc does awav with the
HOW TO SET ABOUT DESIGN.
173
disconnected look of the various parts. The completion of the design (185) is then only a matter of detail. The square divisions are kept, and remain a feature in the design; but attention is diverted from them by the wave-lines crossing the lattice, which give yet more emphatic features, and take the eye from them.
One sets about the design of a pattern of whicli the lines of construction form no visible part in much the same way — with a definite idea, and on definite lines, but uever with any definite detail, such for example as a natural spray of flowers.
190. DIAGRAM SHOWING .START OF A DROP REPEAT.
174
PATTERN DESIGN.
191. DIAGRAM SHOWING DEVELOPMENT OF 190.
Painters unpractised in design afsume sometimes that they have only to repeat at given intervals no matter what studv from nature, and make good the connection between the repetitions of it, and the trick is done. It is not quite so easy as that. Let any one try and connect the isolated details (187) on page 170.
The natural lines of a flower, determined by no thought of repetition, are scarcely likely to bear repetition verj^ well, and the difficulty of working up to nature, and comprehend- ing such naturalistic details in any satisfactory scheme of
HOW TO SET ABOUT DESIGN.
175.
192. WALL-PAPER PATTERN.
176 PATTERN DESIGN.
composition, is extreme. If anything results that wav which goes for ornament, it is by accident and not design. Em- phatically that is not the way to set about it. A designer makes his flowers grow his wa}'.
He starts, never with detail, but with one or other of the two important factors in design, line or mass — whichever, according to his aim, naturally takes precedence. In the case of a scroll, he will first get right the sweep of the lines, before beginning to clothe them; in the case of a floral pattern, he will mxore likely dispose his flower masses in the order in which they should come, leaving lines of growth and foliage for after-consideration.
It m' st not be supposed that defects of construction are to be made good by clothing or disguised by foliage. No one worth deceiving is deceived that way; and any one disposed to scamp prc'hminary work should know that in the end it does not even save labour. Starting with the idea of a symmetrical design in which the flowers and buds shall be the prominent features — the designer starts naturalh' with what he desires should first be seen — he begins bv planting somewhere about the centre of the repeat, say, a heart-shaped mass (i) diagram i88. That perhaps suggests to him at either side a smaller bud-shape (2), near enough to the margin to group with its repeat, and so be useful in taking the eye from the joint, designed to balance the heart-shape, but not compete with it in mass. These forms repeated suggest, as a means of breaking the plain space below, features of interme- diate size and different shape from either (3). A still remaining vacant space or belt of ground between these and the heart- shape below, determines the introduction of a pair of smaller buds (4), which in repetition give groups of four, valuable if only for variety's sake. A space of still too open ground suggests additional budlets (5), far enough apart to appear singly in contrast to the pairs about them. These points of interest determined indicate of themselves the lines to connect.
HOW TO SET ABOUT DESIGN.
correct and counteract them. The order in which the}^ suc- cessively occurred is given by the letters a b c d.
The designer may or may not, in planting such features on the ground, have somewhere at the back of his mind an idea as to the way they shall eventually be con- nected; but the connecting lines must in the end be de- termined by the necessity of accounting for those masses as they stand. Supposing them to be flowers, they must grow in some coherent way. Lines and masses once determined, the next process is to give thera more specific shape, and to modify them to some extent in so doing (189); to evolve perhaps out of the heart-shape a con- ventional flo\\er, out of the smaller shapes husks with berries. The lines become connected stalks, clothed in the end with foliage, the scale of which is fixed by the spaces to be occupied, and the character by that of the flowers.
Invention, it will be seen, is here progressive. Each advance enables the designer to see further ahead, as when, in climbing, you reach another ridge of hill. To a man in the vein, one move suggests the next — he may not have known what he was going to do, but, one step made, he feels the next must be just so,
193. DIAGRAM SHOWING BRICK- WORK " PLAN.
PATTERN DESIGN.
194. FLORAL PATTERN IN WHICH THE DISTRIBUTION OF THE FLOWERS WAS THE FIRST STEP OF THE DESIGN.
and no otherwise. What is done pledges him to something furtlier.
The process of designing a drop pattern is set forth in diagram 190, in which much the same forms as before are purposely employed.
In this case it is more than ever necessary to repeat each form, as soon as determined, in the outer spaces round about the central square (containing the unit of design, but not the parts of the pattern in their entirety). The cone-shaped feature (2), for instance, not only oversteps the line, but grows from a stalk which trails over from the side. That much settled, the balance and the lines to the artist's satisfaction, lie can safely go on to the details — in this instance, as it
HOW TO SET ABOUT DESIGN.
179
happens (191), \'ery difterent indeed from the last — from which it will be understood how little the planning or first roughing-out pledges one to any definite character of detail. Fither of these tw'o rough first suggestions might just as well have been carried out after the manner of the other. The completion of a pattern \ery similar in detail to iqi is shown opposite to it (19^).
The main point to bear in mind is that there must be liarmony between the detail and the way it is planned. Comparatively natural flowers nuist grow in a comparatively natural way (194). Forms m.ore deliberately ornamental (192) demand correspondingly formal lines to accompany them. It is in the precise relation of the two that the taste of the artist is shown.
The evolution of a design beginning with line instead of mass, is shown in the diagram illustrating the brickNvork plan (193) in which the consecutive steps were: the wa^■e-line across
c
/M
%M^
C
195. DIAGRAM SHOWING FIRST STAGE OF A DF.SIGN.
PATTERN DESIGN.
196. DIAGRAM SHOWING SECOND STAGE OF DESIGN (195).
the brick; the continuatiou of the line across the other bricks, to see how it would come; the placing of the flower spots to steady the effect; their connection ^^■ith the main stem; and the final fiHing out with foliage.
Whether it is better to start with line or mass depends upon what you want to do. It is as well to begin N\ith what is most important. If you want grace of line, that must be your first thouglit. Line deserves prominence in design only on the condition of its being beautiful. That is where the designer conies in.
In the case of deUberately fioral design (194), the more convenient plan is to begin by distributing the flowers (assum- ing they are to be at all prominent), settling where they shall occur, their size and shape, grouping them here into bunches, there breaking the ground with isolated blossoms or smaller buds, but considering them always as so many colour patches. The main stems from which they grow may then be thought
HO^^' TO SET ABOUT DESIGN.
i8i
-^o-"-
. -C,
^S
^
d">o
r
V Q, P'
u.
197. DIAGRAM SHOWING THIRD STAGE OF DESIGN (195) •
r^c
o VN
o-
0,0
A.
McQ
c ^
/^o-
198. DIAGRAM SHOWING FOURTH STAGE OF DESIGN (l95)'
1 82
PATTERN DESIGN.
of, and finally the foliage which is to occupy the space between the flowers, avoiding them it may be or backing them.
TliC stems of flowers (which must be natural in proportion to the naturalness of the flowers and leaves) are a standing" difficulty in design. You must have them but 3^ou do not as a rule want them to be marked; and they have a way of marking themselves very determinedly. The wary artist in planning his design bears in mind from the first the necessity of something hke natural growth in a natural flower, but still lie starts with the flower masses — imless of course the flowers form no important part in his scheme; in which case he begins with tb.e foliage, if that is more important; but flowers insist as a rule upon being the first consideration. Absolutely natural growth is rarely possible in pattern, even were it to be desired. It takes beautiful lines but seldom quite the lines wanted in a given pattern. It is expedient in such case to
199. DIAGRAM SHOWING FIFTH STAGE OF DESIGN (iQs).
HOW TO SET ABOUT DESIGN.
183
200. PORTION OF DESIGN I95 FINISHED.
disguise or lose the line of growth in foliage — much as it is lost in nature.
The development of a fairly complicated floral, but not too naturalistic, pattern, is traced in diagrams 195 to 200. The initial idea was a free-growing pattern in which flowers of relatively large size should be supported by smaller ones, of different colour for variety's sake — a double growth that is to say. That would give also an opportunity for variety in the colour of the leaves. Naturally one growth would be more prominent than the other.
The first thing to do, having settled that it should be a drop pattern (the dimensions of the printer's roller settled that in should be twice as wide as it was deep), was to plant the more important flowers in place, as at a (195). A central group of three large flowers (i) and two small buds (2), when repeated as at b, suggested the placing of further flowers (3) between, rather nearer to the side edges. These repeated
i84
PATTERN DESIGN.
20I. DIAGRAM SHOWING FIRST AND SECOND STAGES OF A VELVET
PATTERN .
as at c, there seemed to be sufficient of this sort, remembering there were others to come. The number and position of these others (naturally of a different shape) was determined (196) by the ground left bare. They are wliat the vacant spaces seemed to call for: a group of three (i), to stop the downward gap; a pair (2), to break the joining line; three separate flowers (3), to fill the vacant spaces in the centre.
The next consideration was the order in which the flowers should grow, and first the larger ones. A solid line from A to A (diagram 197) shows a stem which runs through and joins on satisfactorily at the sides; it had only to be repeated in the upper part of the drawing to suggest the
HOW TO SET ABOUT DESIGN.
185
202. DIAGRAM SHOWING THIRD AND FOURTH STAGES OF A VELVET
PATTERN.
more or less contrasting (dotted) lines connecting the flowers with it.
The growth of the larger flowers accounted for, that of the smaller (given in dotted lines in diagram 198) had to be schemed, which left only the leafage to be blotted in (diagram 190) and the map of the pattern was there.
In carrying such a sketch further (a portion only of the finished cretonne is given on page 183) the shapes of the forms might well want considerable modification, something would have to be erased or added, but the groundwork would be all there, the plotting done, and the designer free to follow the promptings of his artistic impulse. A design of this kind is
i86 PATTERN DESIGN.
easier to manage if the distinction between the separate growths is made clear by the use of very pronouncedly different colours, such as red and yellow for the flowers, blue and green for the leaves- even though there should be no in- tention of any such sharp contrast in the final effect. Some such guide is almost necessary, to enable the designer to keep the threads of his design separate. Indeed, in the case of a complicated design of any kind, and especially where there are two or more separate elements in it, it is not a bad plan, even thougli it is to be eventually in monochrome, to use different tints in plotting it out. It reduces the very serious danger of confusion to a minimum.
The design of which the genesis is next given starts neither from stem lines nor from flower masses. The idea was to get a broad pattern, bold but not too bold, in three shades of colour, light, dark, and middle tint, the kind of relation which is so effective in old velvets, where the glossy satin ground, the dense rich pile, and the intermediate uncut, ribbed surface known as " terry, " give three very distinct stages of colour, and lead, almost naturally as it seems, to a characteristically rich sobriety of effect.
Thinking still of velvet and the softening effect of the outline in terry; it was only natural to determine upon the middle tint for the outline.
The first thing to be settled was the main lines the design should take. It was as well, as a bold effect was wanted, to make them very bold, they could always be refined and softened. That being so, there could be no better plan than waving bands which in opposition give the ever satis- factory ogee shape (201). But as it was not a geometric pattern that was desired, these broad bands had forthwith to be broken in some way; which was very simply done by treating them as bands of foliation, twisting about, and to some extent disguising the too plainly geometric planning. This was a means of getting, too, some life into the lines. It
203. FINISHED DESIGN OF VELVET PATTERN.
1 88 PATTERN DESIGN.
was high time by now to think of the pattern in mass as dark upon a light groimd (diagram 202), and to sketch in not merely the turning over of the foliage but the serration of its outhne. The broad bands began on this to disappear, but the lines were still stiff, and the masses of light and dark in too crude contrast. That was corrected by the introduction of dark foliage into the ground space d, ^^■hich very