^Binac^iiK 
 
 
 r *»we«pW tf V;, 
 
 '■"""iroW'^ j/' 
 
 ,;•■. 
 
 
 
 ::^*! &, SN* ; 
 
 '■ . ■■ t.-,'^ ,,; 1 
 
 m™««*»^;; 
 
 
 . - *f I ■ 
 
 0<i>***<5 
 
 • . 
 
 ■^ 
 
 'V, v ..-' ', 
 
 
 .<#»-. 
 
 
 
 .>»*>.. 
 
 
 
 V- 
 
 ; ■ 
 
 ■'•-' 
 
 Ill 1 
 
 \£k 
 
 
 THE 
 
 DECORATIVE 
 
 ART • OF 
 
 JAPAN 
 
 '"•-'IJi^lcJ* 1 ' it 
 
 ^immi^0»** 
 
 •to.^Jui**'^ 
 
 F. T. PI.G.GQT.T 
 
 
 
 
 WM2 
 
 
 ■Rv? 
 
 &*&* 
 
 
 v "Swo^ >' \ 
 
 
 
 **<>■ <vi II 
 
 Ndi fu riP£l 
 
 
 "WjCUTii 
 
 '^W^ '/ 
 
 
 
 ■ f^ 
 
 
 
 ■■>■:■ 
 
 ■■{$; 
 
 ;J -1 ■■■.«? 
 
 i 
 
 
 
 
 
 pflr - 1*.. T 
 
 
 ?*j«*i« 
 
 
 ^»<«>»i" l * rt -L>»J*^*« 
 
 MgMg** 
 
 
 tL>"? , ?* l ' 
 
 •^J "\j 
 
 p) 
 
 
 
 HWttfc-'^kw:/' ■ 
 
 
 r<sj-i4 1 
 
 
 ■- w/^Wr: 
 
 j'- 1 ""' 111 "'-''*;^ "" 
 
 "'■'■*V- ; iHt:r''' ' ;t ■..■ 
 
 ttWWl^**" '"*' 
 
 
 ' ' ' ^ifc ''**-iw» 
 
 HMMHriMM ■ 
 
 
 
 
 H ft 
 
 rij 
 
 I 
 
 f ^»A 
 
 £/ 
 
 / "\ 
 
 
 

 i 
 
 i 
 
 (Ttecoratix/ecArt 
 
 iOYU 
 
 STERL1 N G 
 AND FRANCiNE 
 
 CLAR1C 
 APT INSTITUTE 
 L1BRART 
 
 I 
 
£//!\M\£r< 
 
STUDIES IN THE 
 
 DECORATIVE ART 
 
 OF 
 
 JAPAN 
 
 ^ 
 
 SIR FRANCISi PIGGOTT 
 
 (IIIKF JUSTICE OF HONGKONG 
 
 LONDON : P.. T, BATSFORD, 94 HIGH H( >LB< )RN 
 
 1 9 I 
 

 PRESS OF 
 
 THE BOX OF CURIOS PRINTING & PUBLISHING CO. 
 
 YOKOHAMA 
 
Digitized by the Internet Archive 
 
 in 2012 with funding from 
 
 Sterling and Francine Clark Art Institute Library 
 
 http://archive.org/details/studiesindecorat1910pigg 
 
The three fragments of brocade, which serve as a triple frontis- 
 piece, are given as typical examples of tlie art of design in its 
 most elaborate expression, when colour and the sheen of silk lend 
 their aid to highten the effect. They illustrate also the great 
 dexterity of draughtsmanship to which Japanese artists, sometimes 
 very humble ones, often attain. 
 
Plate 1. 
 Frontispiece — A. 
 
 HKOCAUIi 
 
Plate 2. 
 Frontispiece — B. 
 
 BROCADE WITH EMBKOIDERY 
 
Plate 3. 
 Frontispiece — ('. 
 
 BKOCADIi 
 
I have not attempted in these pages anything approaching an exhaustive 
 examination of the decorative art of Japan. The area it covers is so extensive 
 that I doubt if it is possible to do more than touch lightly upon its manifold 
 phases ; for even the most diligent student there always come surprises, and the 
 end of his labours ever recedes into the distance. The art knows no limitations, 
 and a hundred books would not be sufficient to set it forth in all its infinite 
 variety. Studies, somewhat intermittent I fear, seem to me to be all that is 
 possible. 
 
 The decoration of the temples has been treated as a distinct subject ; for 
 although it touches the art of the people at many points, and on one side of it 
 reveals the historic influence of the Buddhist religion as it came to Japan through 
 China from India, yet stands apart as an art specially Japanese, inspired by the 
 national purposes to which it was put. 
 
 I am conscious that the method I have adopted for reproducing the designs 
 is open to much criticism, for it is limited only to their general effect — very 
 imperfect sketches of light and shade, leaving the student or the architect to 
 work out the ground-plan, or scheme of construction, for himself. 
 
 The examples of the more popular and better-known side of the art are for 
 the most part tracings from lacquer boxes. It is here that the impossibility of 
 touching more than the fringe of the subject makes itself greatly felt ; and I 
 have been compelled to limit myself to an endeavour to illustrate the broad 
 principles on which the Japanese designers work. 
 
 A special chapter is devoted to the evolution of the " key "-border in the 
 East. The conclusions arrived at in it are not at all in accordance with the 
 tradition which assigns its origin to Greece. 
 
 In 1892, I gave a lecture at the Royal Institution on that bastard decora- 
 tion, then so prevalent in England, for which "Japanesque" seemed the most 
 appropriate name ; and in the following year some of the designs figured in this 
 book formed the subject of a series of articles in the "Builder." I trust that 
 the idea of preserving in more permanent form the designs, as well as some of 
 the opinions expressed on those two occasions, will not be thought presumptuous. 
 
 F. T. P. 
 
 Hongkong, 
 
 March, 1910. 
 
STUDIES 
 
 IN THE 
 
 DECORATIVE ART 
 
 OF 
 
 JAPAN 
 
PAGE 
 
 The Art of the Temples ------- 17 
 
 The Decoration of Flat Surfaces - - - - 21 
 
 Wave and Cloud Forms ------- 32 
 
 Lattice Work ----------39 
 
 The Decoration of the Japanese - - - - 49 
 
 The Use of the Circular Form ----- 62 
 
 The Pakwa Diapers and the Key Borders - 83 
 
 HP 
 
 -St 
 
THE ART OF 
 THe TEMPLES 
 
 The Nikko and the Shiba Temples as I knew them twenty years ago have 
 passed away; Nikko the beautiful has decked herself anew, Shiba has been 
 destroyed by fire. These temples were the home of one branch of the art of 
 Japan, studied in the days gone by, of which I have attempted to reproduce an 
 echo. 
 
 I must introduce those of my readers who are unfamiliar with them to these 
 splendid shrines, and to the genius which pervades them, inspiring this art, 
 which for pure decoration stands unrivalled. 
 
 The famous avenue of cryptomerias, winding like a lazy dragon across the 
 plain from Utsonomiya, ends at the entrance to the long village street of Nikko : 
 the Daiya-gawa bustling along through the rocks and boulders which are its bed, 
 in autumn a mountain torrent tumbling in the new channels made last winter, 
 recklessly cleaving new ones for the floods to come : from the bridge one of those 
 landscape pictures which make the beauty of Japan, the mountains of Chiuzenji 
 in the distance, and the red lacquer bridge for foreground : paths winding up 
 the hillside in the shade of the giant trees, past priests' houses, past the 
 Mangwanji with its sweeping roof, and its garden whence comes the mellow 
 boom of the great bell telling the passing hour: a broad roadway in the sunshine- 
 leading to the massive stone torii : a flight of steps, a gorgeous portico, and then 
 a glitter of gold and colour, shrines, cryptomerias, here dark groves of gloom, 
 there the glinting rays of the sun : groups of soberly-clad peasants reverently 
 curious: paths leading upwards through the mighty trees to the unseen hills, a 
 world of its own, dwelt in by the priests and the spirits of the dead Shoguns : 
 unique, which leaves the mind dazed, the eyes dazzled, with so little definite 
 impression on either that the hotel verandah comes almost as a welcome relief. 
 
 But the temples lure you back, and back again, until definite impressions: 
 begin to formulate in the brain, and the eye finds signs by which to know where 
 pathways lead, to recognise one shrine from another, and tokens which as the 
 days go by become familiar, and in the after-years translate themselves into 
 imperishable memories of greatest beauty. 
 
18 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 The great white gate with its storied roof and heavily-carved columns holds 
 yon first, and compels men of all creeds and no creed to realise that it is the 
 Beautiful Gate of the Temple : makes them admit the meaning, if only as a 
 practical necessity, of the hehraic injunction as to removal of shoes. Those who 
 rebel grow fewer year by year, and many now willingly render western homage, 
 entering bareheaded. Within, where the gloom of the groves seems to have 
 found a permanent habitation, there is sanctuary from the sun. Then gradual- 
 ly, as the eye gets accustomed, through the shadows appears such a wealth of 
 splendid decoration, that it is practically impossible to decipher, much less to 
 remember any of it. Heyond the temple the shrine ; and beyond the shrine 
 flights of stone steps in massive masonry of cunning setting, leading into the 
 woods upon the hill, where in their perfect calm is the tomb of Iyeyasu, first of 
 the great dynasty of Tokugawa Shoguns. This is Nikko of the Hills. 
 
 The other picture differs in all its essential details. Vast courtyards of 
 silence, full of stone and bronze lanterns, votive offerings of the daimyos to the 
 memory of their chief, set in a little park, an oasis of peace in the heart of the 
 busy city, from which the sounds of traffic come in a subdued, almost reverent, 
 murmur, as of a far-away world. A latticed screen shuts out still more that 
 strenuous world, through the gates in which, heavy with wreathing dragons, you 
 pass to the temples and the shrines of other Shoguns of the famous dynasty, and 
 beyond them, as at Nikko, to their quiet tombs. This was Shiba of the Plains. 
 
 Never endowed by nature with allurements to the time-tied traveller, Shiba 
 at no time attracted so many visitors as Nikko, and so had few funds to draw 
 upon to prevent its buildings from falling upon evil days. Many travellers pass 
 Tokyo by in their rapid rush to the well-known sights; it has almost fallen out 
 of the beaten track, so that the cunning curio dealers have departed to spread 
 their wares in more likely places. Perhaps not a tenth of the crowds who went 
 to Nikko had ever seen, many probably had never heard of, Shiba. To them 
 the news of its destruction meant nothing; but to those who had passed hours 
 of delight in the spacious courtyards, who had felt the charm of its temples and 
 yielded to the fascination of the place, it meant almost a personal disaster, so 
 greatly was it held in affectionate remembrance I saw, after many years, the 
 Shiba Temples before their passing, and noted sadly the progress of decay; but 
 the charm still hung round them; only the jangling of bells as the tramcars 
 passed, jarred upon the ear, and disturbed the city's murmur. The spirit of the 
 place still lived and spoke with its old grace, awakening memories, more easily 
 indeed than the spirit of Nikko could emerge through the brilliant coating of 
 fresh colour, glistening lacquer, and new gold. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 19 
 
 It is no imaginary spirit that I have conjured up in calling these temples to 
 the mind. The gradual transition from the sumptuous decoration of the temple 
 to the perfect peace of the tomb typified the passing of the prince from the 
 restless splendour of the palace to the peace of the world beyond. All that 
 art-labour could do enhanced the gorgeousness of the temple. Screens and 
 sliding panels painted or carved by the most famous artists of the day stood in 
 array in its halls ; colours in marvellous harmonies, or splendidly discordant, 
 lavish expenditure of gold, wealth of design, lacquer beyond compare, adorned 
 them. But art in her more suggestive and mystical mood, found in the quiet 
 sea and in the clouds, the most appropriate ornament for the gateway to the 
 other life. In the graveyard even the sacred symbols, which hold so large a 
 place in the decoration of the temples, give way to emblems suggestive only of 
 repose. 
 
 This, as it seemed to me, was the spirit of the place, the genius which guid- 
 ed the hands of the artists who gave of their best without grudge of time, and 
 who realised, as [ think no other artists have ever realised, the place which 
 decoration properly holds in the scheme of existence. To appeal to the imagina- 
 tion, to fill the mind with suggestion, to alternate playful fancy with more 
 serious thought, to infuse into ornament a meaning, grave or gay as the occasion 
 may demand, that those who pass by shall carry with them some little recollec- 
 tion of that meaning, and of the story which the priest and artist had to tell. 
 
 That beyond the symbolism which is at the foundations of much of their 
 art, beyond that graceful imagination which made play with all that nature ever 
 created and pressed it into service, there were a multitude of ether influences at 
 work is abundantly apparent. What those influences were which lay bevond the 
 common stock of eastern life and Buddhist religion, and what the manner of 
 their coming to Japan, are questions which I do not propose to discuss. It is 
 evident that the arts of Europe, of many periods and many countries, reached 
 the temple designers and influenced their work, in a way which the early trading; 
 settlements do not altogether account for. But the extraneous search for the 
 origin of things is not altogether profitable, and I propose only to study the 
 things themselves. 
 
 It is, however, inevitable that the historic relationship between Greece and 
 India, and so between Greece and China, should be referred to in the course of 
 these studies: for art, history's most accurate registrar, has traced upon the- 
 ancient walls of both a common design; and from this fact many conclusions 
 may be drawn as to the origin of things, at a time when the only known 
 exchanges were the crowded bazaars and the silent desert. I may premise at 
 
20 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 once that these conclusions are not altogether in accord with those which are 
 current among the learned ; but for the moment, 1 do not propose to dogmatise 
 further than this, that the art of the Japanese temples developed in the temples 
 themselves. Apart from the well-known symbols, certain fundamental and 
 elementary designs were used here, as they were used in all other branches of 
 Japanese decorative art — they were the indispensable formulas; but on them, 
 and often irrespective of them, a separate and distinct art arose, engendered 
 simply by the desire to beautify the shrines of the great dead. I have found it 
 nowhere else in Japanese art in the simplicity of use as in the temples; it was 
 the like link between art and a religion which put the dead in the highest place, 
 and it seems to have been deliberately excluded from the ordinary adornments 
 of every-day life. But in the temples, every square yard, foot, or inch of 
 available space, seen and unseen, as the Japanese way is, was literally smothered 
 with ornament: carved, painted, inlaid, in high relief, and low relief, in woven 
 silk and chased metal, produced by every means which human skill and loving 
 labour could devise for the making of things so beautiful that they should be 
 worthy of a place in a shrine of worship. 
 
THE DECORATION 
 OF FLAT SURFACES 
 
 The temple decorators seldom allowed the smallest space of wall or floor to 
 go without its appropriate ornament : loved rather to devote to it the greatest 
 possible amount of labour. Time was not of the essence of their contract, which 
 was life-long, the living wage being guaranteed practically in perpetuity. Fail- 
 ing powers did not terminate the agreement, for the arts and handicrafts were 
 hereditary, and the son was ready to take the father's place. If art is long, life 
 in old Japan was even longer. 
 
 Thus the art-idea which inspired their work came to be an abhorrence of 
 plain flat surfaces, and they adopted the diaper as a means of decorating them. 
 The diaper does not seem to be an universal form of surface decoration, and other 
 forms are apt to be mistaken for it. There is the reiterated design, where some 
 ornament is arranged methodically over the surface, but without any coherence. 
 Of this many of the Persian designs given by Owen Jones in his " Grammar of 
 Ornament" (pi. 44) are typical. If we substitute " sprinkling" for methodical 
 arrangement, this form of decoration was and is much used by the Japanese, 
 though not I think in their temple work. Then again there is tile-work, which, 
 as it was practised by the Moors, was a species of mosaic, each element of the 
 design being composed of individual tiles of small size. The modern style of 
 this work, in which large uniform tiles are used, at its best, limits its designs to 
 one tile, so that the work becomes the means by which the design is reiterated. 
 But a more or less continuous design is often adopted running over a large area, 
 several tiles being used, and so tile-work approaches the diaper; but its chief 
 characteristic is wanting, the treatment of the surface as a whole, either by 
 interlacing lines repeating at regular intervals the principle figure which gave 
 the diaper its special character, or by a series of crossing lines which divided the 
 surface into a number of similar spaces in which line or floral ornaments were 
 placed. Diapers prevail in almost every form of Japanese decorative art, but 
 tessellated work is I believe unknown. 
 
 The most primitive diaper is composed of interlaced lines in a manner 
 suggested by straw-plaiting, and known to us as the " basket-work " pattern. 
 This will be seen in its simplest form in fig. 12 (pi. 6). A more elaborate form 
 
22 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 occurs in fig. 1 (pi. 4). This interlacing, though it has the appearance of sim- 
 plicity, is complicated in a manner peculiarly Japanese. There are eleven 
 threads, plaited in two groups of three and one of five, the length of the longest 
 being eleven "thread-breadths." The difficulty in at once following the con- 
 struction of the design arises from the fact that the groups of five short threads 
 are opposite both the medium and long groups, which have together six, so that 
 the last of the long threads has to be continued into the first of the opposite 
 medium group. 
 
 But the charm of the temple diaper-work is due not merely to such dexter- 
 ous intricacy of design, of which we shall see many examples, but to the manner 
 in which they were wrought. In the large majority of examples of line diapers 
 which I have seen the lines are deeply cut with a triangular gouge, with the 
 result that they scatter the light in all directions, the effect becoming more and 
 more bewildering as the designs increase in intricacy; so bewildering indeed, 
 that in the sunlight a gold panel in a temple door corruscates, and the design is 
 almost lost in the glitter of light. 
 
 Sometimes these line diapers are left without further ornament, as in 
 many of the examples given in the Diaper plates (4 to 8); but they are often 
 used as a background on which an independent design is placed. A very 
 charming effect frequently used in the Nikko temples is a white diaper with 
 small brown wood ornaments set on it, as in fig. 3 (pi. 4), and fig. 2o (pi. 8). 
 
 I have been much puzzled how best to convey the effect of these diapers as 
 they actually appear. A mere outline of the design would give no effect at all, 
 and would hardly explain it. I have therefore made the illustrations take the 
 form of sketches of light and shade; and I must ask my readers to imagine, in 
 spite of imperfect drawing, the triangular cutting of the lines, one side being in 
 light and the other in shadow. 
 
 One other point should be noticed. The carver-designer has one element 
 more to work with than the painter. In addition to varying length and breadth 
 of line, straight and curved, and angle of inclination, he has thickness of his 
 material, wood, with which he gets real instead of artificial depth, and therefore 
 he can make great play with real shadows. Such a design as fig. 21 (pi. 8), 
 depends for its effect mainly on varying thicknesses of wood, both for the cross 
 pieces and the central ornaments. We shall come across this effect again in the 
 lattices, plates 13 to 15. 
 
 Reverting to the diapers, the next in order of simplicity is one which is 
 formed by dividing the surface into lozenges, each containing a series of smaller 
 ones — a linear adaptation of the Chinese "nest" of boxes; terminating, as in 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 23 
 
 fig. 23 (pi. 8), with a single dot, or as in fig. 3 (pi. 4), with four small lozenges 
 arranged lozensre-wise. This latter idea becomes a substantive design in the 
 alternate divisions of fig. 18 (pi. 7). 
 
 The "nest" is frequently used to break up surfaces by grooving: in a 
 simple manner in the ornaments at the lozenge-points in fig. 16 (pi. 7) and fig. 
 20 (pi. 7), and in a more elaborate manner in the hexagons in fig. 19 (pi. 7). In 
 fig. 17 of the same plate, the idea of the grooves, though it appears to be the 
 same, is in reality quite different. Here the suggestion is of broad bands of silk 
 surrounding the central flower, looped and knotted at the points of the central 
 hexagon, and the grooves are the folds of the silk. 
 
 But many of the more complicated diapers, and there is a long series of 
 them to be specially studied presently, are based on the religious symbols of the 
 East, formed of elementary arrangements of lines. The priest, past-master in 
 art and craft, saw in them a means of reminding the curious, who came to 
 admire without a pause for worship, of the eternal verities to which these 
 symbols gave formal expression ; and art, having once pressed them into service, 
 evolved from them designs ever growing in intricacy. This too served the pur- 
 pose of the priest ; for the involutions of the design arrested attention, and then 
 the symbol gradually evolved itself from the maze of lines, compelling notice 
 from the mere fact that at first it lay concealed. 
 
 Although the more complete study of these symbolic diapers must be defer- 
 red for the present, a few examples of them are given now, because they hold so 
 large a place in the surface decoration of the temples, and in order to shew their 
 general effect and the artistic methods in which they were treated. Figs. 8, 9 
 and 11 (pi. 4), are three examples of diapers based on the Pahwa symbol, the 
 component elements being raised pieces of wood, triangular in 8 and 11, and flat 
 in 9. Pig. 7 in the same plate is based on the Svastika, as also figs. 6 and 10, 
 and fig. 2 (pi. 4). Their construction is complicated, and the explanation of it 
 must be deferred ; it should be noted at once, however, that there is a continuity 
 •of line in them which differentiates them from the more simple "nest" of 
 lozenges already referred to (fig. 3, pi. 4). 
 
 The manx-like form on which figs. 4 and 5 of plate 4 are based at first sight 
 also seems to be symbolic. It was however suggested by the form 
 which results from the intervening spaces between a series of YiT 
 hexagons spaced at regular intervals, which is shewn in a Per- 
 sian diaper given by Owen Jones, (pi. 44, fig. 15); or by y 
 •diagonal plaiting crossed by a perpendicular thread, as in the Chinese diaper 
 •also figured by Owen Jones (pi. 59, fig. 34) 
 
24 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 In fig. 4 all the elements are turned in the same direction ; but in fig. 5, 
 every alternate row is inverted, the properties of the hexagon 
 rendering this possible. These two designs have some affinity 
 with the " nest "-work of fig. 3 ; there is, however, this difference 
 between them : that whereas in the " nest," the grooved figures 
 diminish from the outside, in figs. 4 and 5 they work outwards from the central 
 figure. 
 
 But while the priest availed himself of symbols for the double purpose of 
 inventing religious memoria technica and of creating ingenious artistic devices, 
 he was not slow to see and use the beautiful decorative forms which nature 
 suggested ; and here he found a very treasure-house replete with beauty and 
 suggestion of graceful form. 
 
 In adopting nature's suggestions he did of course what all other decorators 
 have done, and he devised, inevitably, a iour-petalled figure, used in fig. 18 (pi. 
 7), and in fig. 25 (pi. 8), much resembling, if not identical with those which had 
 developed in other systems of art. Indeed he almost touched, though how far 
 independently is another matter, some of the classical forms and scroll-work of 
 the West, of this I shall give some examples hereafter. 
 
 But there is a distinctive character in this branch of eastern art, which 
 arises from the different floral forms to which the artists had recourse, rather 
 than from any peculiar method of treatment. The pseony and the chrysanthe- 
 mum are largely used, for they are flowers to which special attention is given in 
 the East. Of these figs. 10 and 20 (pi. 7) are typical. In fig. 12 (pi. 6), the 
 more rigid Chinese form based on the pseony will be seen, of which there are 
 many variations. Another very common floral form, with three petals and three 
 sepals, is used in fig. 14 (pi. 6), and figs. 17 and 10 (pi. 7) ; but I am not very 
 clear what flower is its original. 
 
 It will probably surprise many of my readers when I say that in fig. 13 
 (pi. 6), we have another design based on natural suggestion, for its rigid lines 
 seem rather to place it among those produced by mechanical means of rule and 
 compass. It is known as the " pine-design ; *''' and a reference to plate 20, where 
 a whole pine tree is treated conventionally, will make its evolution easy to under- 
 stand. It belongs to the class of diapers which I have characterised as " light- 
 distributors " ; and when the pine needles are cut with the triangular gouge to 
 which I have already referred, it is the most dazzling of them all. Fig. 15 
 (pi. 6), is a strip of the same design cut diagonally across ; but its pine character 
 has vanished, giving place to a star-like form, which is emphasised by the 
 omission of every alternate spine of the original design, the space being left flat 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 25 
 
 and coloured. In fig. 14 of the same plate, this star-form is treated as an inde- 
 pendent unit of design. 
 
 The introduction of floral forms into the spaces of the diapers, as in figs. 10 
 (pi. 5) and 12 (pi. 6), led to the creation of designs for surface decoration much 
 less formal in their nature than the symbolic diapers ; but in the fact that the 
 surface was broken up into regular figures, into each of which were placed one 
 or more recurrent devices, they fall naturally to be considered with the diapers. 
 Fig. 16 and 20 (pi. 7), are taken from designs similar in their general nature to 
 those in figs. 10 and 12 ; the only variation being in the subsidiary ornaments at 
 the lozenge-points, which truncates them, and somewhat alters their character. 
 
 The hexagon, nature's mechanical figure, is also very frequently used, and 
 is treated like the lozenge ; that is to say, the design is composed of a series of 
 hexagons like a honeycomb, each of which is filled with a floral form, as in fig. 
 17 (pi. 7): or one of them only, which is made the central figure, and is surround- 
 ed by a series of six hexagons containing a pure line design, as fig. 14 (pi. 6) and 
 fig. 19 (pi. 7). 
 
 The pine-design (fig. 13, pi. 6), also naturally works into a honey-comb of 
 hexagons; but it will be noticed that when the central point is taken in conjunc- 
 tion with the centres of the surrounding hexagons, a large 6-rayed star appears, 
 constructed on the same principle as the star-form in the fillet of the design, fig. 
 15 (pi. 6). 
 
 Fig. 24 is a lozenge diaper in red laquer cut in low relief. I cannot say 
 with certainty to what suggestion this very charming design should be attribu- 
 ted. It seems probable however that it is based on one of the folded papers 
 which in Japan always accompany gifts. 
 
 Fig. 22 comes from a pillar in Shiba of brown lacquer; it is cut in very 
 deep relief. In fig. 25 the floral figure is enveloped by seven zig-zag lines in 
 outline, a form of decoration with which we shall become familiar when we come 
 to study the lattices. 
 
Fig. 1 
 
 Fi 2 . 2 
 
 Fig. 4 
 
 Plate 4 
 
 Diapers A 
 
Fig. 6 
 
 Fig. 7 
 
 r 
 
 1 • « I II I ., ,■ I 
 
 Fig. 10 
 
 Fig. 9 
 
 Plate 5 
 
 Diapers B 
 
 Fig. 11 
 
Fig. 12. 
 
 Fig. 14. 
 
 Fig. 13. 
 
 Fig. 15. 
 
 Plate 6. 
 
 Diapers C. 
 
Fig. 18 
 
 Plate 
 
 Diapers D 
 
Fig. 21 
 
 Fig. 22 
 
 Fig. 24 
 
 Fig. 23 
 
 Plate 8 
 
 Fig. 25 
 Diapers E 
 
WAVE AND 
 CLOUD FORMS 
 
 But the temple decorator had not yet exhausted his sources of inspiration. 
 He had used the symbols of religion to the full ; he had pressed all nature's 
 most beautiful handiwork into art-service ; but there were still the waves of the 
 sea, in restless motion or repose, and the ever-changing clouds, to suggest curves 
 of beauty to his artistic imagination. In these he saw an expressive symbolism 
 of life and death, which fitted in with his scheme of giving to art a deeper pur- 
 pose. 
 
 The consideration of the wave designs follows that of the diapers in natural 
 sequence, as they were also largely used for surface decoration. I think I am 
 right in saying that they are, as well as the cloud designs, peculiar to the East ; 
 they certainly received the greatest elaboration in Japan. The constructive 
 principle on which they are based has long been used in many countries, but the 
 suggestion seems to have originated in the scales of a fish rather than in the 
 wave. The Tudor architects obtained very light and graceful open-work by 
 setting semi-circular tiles in rows one upon the other. It is not improbable that 
 the wave design in this simple semi-circular form was used in Japan, though I 
 do not remember having come across it. Its most common form is that given in 
 figs. 1 and 4 (pi. 9). The wave curves are segments of circles, one-third of the 
 circumference, the centres being at the inscribed hexagon-points of the series 
 immediately below. The inner circles are concentric. These subsidiary lines 
 constitute another point of difference between the wave and the " fish-scale." I 
 speak with hesitation, but I think I am right in saying that the wave is never 
 found without them ; they are the ripples on its surface, and much of the charm 
 of the design is to be attributed to them. In fig. 1, when worked in wood, these 
 inner lines are incised with the triangular gouge like the diapers, and what I 
 may call the "ripple-length" is equal throughout; in fig 4 the design is treated 
 flat, and the ripple-lengths are unequal. The most elaborate form of the design 
 that I know of is given in fig. 2 (pi. 9). The wave itself is semi-circular, but 
 the ripples are formed from circles which are tangential at the base, which is of 
 course the point of junction of two waves in the series next below. Both figs. 1 
 and 2 are made of plates of hammered bronze. They come from those last court- 
 
33 STUDIES IN THE DECOR ATIYE ART OF JAPAN 
 
 yards in the Shiba temples where the tombs are, and there is no other ornament. 
 It is symbolic of the calm of death, when the soul sets out on its voyage across 
 the quiet sea. The symbolism is complete, the effect on the mind that of 
 absolute repose. 
 
 The small wave design given in fig. 12 (pi. 10), is in red lacquer in low 
 relief. It is less formal in its treatment, and is interesting on account of the 
 foam-swirls at the base of each depression. The Japanese artists have never 
 hesitated to give a definite form to wave-crest and spray in their pictures; and 
 although the western eye gets accustomed to it with some difficulty, its very con- 
 ventionalised form soon "speaks " to the mind and conveys the required impres- 
 sion. 
 
 The wave designs are not all drawn from the quiet sea ; its turbulence sug- 
 gested much beautiful work, which gave free scope to the vigorous curve drawing 
 in which the Japanese excel. Of these the medallion in fig. 14 (pi. 11) is a 
 good example. It comes from a white column in one of the Shiba temples, cut 
 in very deep relief; the background is formed of a conventional design which is 
 one of the cloud-forms. 
 
 The clouds although they have suggested formal designs such as this, are 
 frequently treated in the manner of fig. 13 (pi. 11), in which there is an almost 
 inevitable relation to the treatment of the foam-swirls of the waves, the cirrhus 
 mass being broken up by inturned scalloped curves. The clouds in this figure 
 are painted on a long beam in three shades of pink, blue and green, mixed with 
 body colour, much in the manner of the missal illuminations of the West; the 
 outer edges of the swirls are outlined with black or gold. The cloud-form is 
 treated more rigidly in fig. 9 (pi. 10), which is in incised lines, gilded, on a dark 
 brown beam. 
 
 Turning now to the more formal design, of which fig. 8 (pi. 11), is typical, 
 it is difficult to say whether its origin is wave or cloud, nor is the matter one of 
 very great importance, the principal feature, an inturned scalloped curve, being 
 common to both. The design is cut very deeply, and shows the layers of 
 lacquer, a kind of work with which collectors are familiar in small red lacquer 
 boxes. It appears again in the upright beam in fig. 14, where it is associated 
 with the same design in a more rigid form, serving as background to the wave 
 medallion already noticed. This form is also used in outline in fig. 5 (pi. 9). 
 
 In fig. 3, the same involuted form appears grouped in masses of waves. 
 This is carved in very deep relief and painted in body colour. In fig. 6 of 
 the same plate, the same idea is worked out as an independent design, which is 
 cut shallow, the inner scallop being slightly modelled. In fig. 7, the design is 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 34 
 
 ■drawn more vigorously, and is cut deeper into the wood ; in some cases the 
 panel is cut right through as in the very freely drawn wave with spray in fig. 15 
 {pi. 11). Fig. 10 (pi. 10) is a fanciful adaptation of the cloud-form, painted in 
 the same way as fig. 13. 
 
 The most beautiful design of this class which I have seen is fig. 1 1 (pi. 10), 
 where the crested wave appears in a triple-head figure enclosed in a trefoil 
 border, the interstices being filled with the lines of repose taken from the wave 
 •design in its original form. The carving is in deep relief; the wave lines are 
 gold on a dark blue ground, and the triple crests gold on a crimson ground. 
 
 Note on the construction of jig. 2 (pi. 9). 
 
 The following figure shows the principle on which the complicated wave 
 <lesigu given in fig. 2 (pi. 9) is constructed, one of the waves being given com- 
 plete with all its tangential circles There are six inner circles, the ripples : the 
 centres of each being on the lower half of the vertical diameter, falling on it a 
 ■distance equal to half the breadth of the distance between the tops of the two 
 circles, which I have called in the text the " ripple-length." The lower and 
 shaded parts of the curves represent the side of the triangular grooving which is 
 in shadow. 
 
Kig 1 
 
 Fig. 5 
 
 Fig. 2 
 
 Fig. 4 
 
 
 Fig. 6 
 
 Plate 9 
 
 Waves and Clouds A 
 
& 
 
 Fig 8 
 
 Fig. 11 
 
 X 
 
 Fig. 9 
 
 V 
 
 Fig. 10 
 
 Plate 10 
 
 Fig. 12 
 Waves and Clouds B 
 
Fig. 13 
 
 Fig. 14 
 
 t™ — 7-"-"*"*r 
 
 Fig. 15 
 
 K ■* \ ■■*-' 
 
 Plate 11 
 
 Waves and Clouds C 
 
LATTICE 
 
 WORK 
 
 I come now to a very highly-developed form of decoration, which is used to 
 a large extent for lattice-work in the temples, but of which I have found no trace 
 elsewhere. This may be accounted for by the fact that the domestic architecture 
 •of the Japanese is so rigidly simple that it does not admit of much ornament, 
 and even such public buildings as old Japan needed afforded no scope for its use. 
 It is not symbolic, and has little affinity with anything that has already been 
 considered. Nor is it suggestive, but is, with rare exceptions, pure line for line's 
 ■sake ; its special artistic virtue being its peculiarly effective adaptation to its 
 object. I doubt whether it would, at first sight, be called Japanese ; though in 
 their individuality of conception and ingenuity of construction, the designs possess 
 two at least of the characteristics of Japanese art. 
 
 We must imagine the mind of the decorative artist exercised bv the need for 
 •decorating, not a flat surface, but an open space ; such a space, for example, as 
 the sides of a temple gateway, which in form is not unlike an oid English 
 lych-gate. The result of his cogitation is that he fills it with a lattice or wooden 
 grille, instead of with solid woodwork. The screens which surround the court- 
 yards of the temples are solid only for about half their length ; the upper part 
 being usually composed of latticed panels. I believe this work to be unique. 
 Open-work, in wood and other material, was often used in the architecture of 
 -other countries, of which the " fishscale " pattern, already referred to, is a typical 
 example ; but nothing approaching the Japanese lattices is to be found 
 •elsewhere. 
 
 Occasionally the designs overlap those of the diapers, as was indeed inevi- 
 table ; but as a whole they stand apart, having no relation to any other form of 
 ornament. The new conditions of the work inspired the artists to invent a new 
 from of decoration, the most important feature of which is its dependence for its 
 effect on the sunlight passing through the spaces of the design, the result of which 
 is to emphasise their form, so that in some cases the eye is attracted to them more 
 than to the surrounding design of the woodwork. Occasionally these lattices are 
 placed on a background of red lacquer, which emphasises the space-shapes even 
 more than the sunlight. 
 
40 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 I have considered these lattices on the method of a continuous <>rowth or 
 natural development which seems to me to be clearly traceable. A most 
 complicated design may, of course, be invented without being developed from 
 more simple ones ; but when a natural order is discoverab 1 ^ it is useful to note it r 
 it may be only for the purposes of study ; but it is also at least probable that the 
 gradual development of the more elaborate designs was along these lines, which 
 are those not only for least resistance, but of greatest facility. 
 
 The simplest designs are composed merely of crossed lines, of which fig. 1 
 (pi. V2) is an effective example. It is not made on the crossed-lath principle of 
 our garden trellices; the pieces of wood join in the same plane and do not overlap. 
 But even with such poor materials it will be seen that the commonplace repetition 
 produced by cross lines at regular intervals is avoided. From this we pass at 
 once into the region of ingenious and obscure construction. The design given in 
 fig. '1 of the same plate, made of crossed zig-zag lines, is very prevalent in all 
 branches of Japanese art ; perhaps the lightning suggested it, perhaps it was 
 worked out in playful fancy. The effect is striking though somewhat irritating 
 to the eye ; it cannnot travel smoothly along them, but is perpetually beingjolted,. 
 so to speak, off the rails. It cannot even follow one line continuously, but is 
 perpetually being diverted by some notch in the crossing lines. The design sets 
 up cerebral fidgets and endless reflexions as to how it was drawn. If there 
 be among my readers an architect who is vexed by a lazy pupil, I would 
 suggest that he should be set to work to draw the design in fig. 2, — that 
 in fig. 3 would indeed be better still — worrying out the method of construction 
 for himself. 
 
 But the eye very soon gives up the idea of following the lines, and rests on 
 the spaces which lie between them. The lines indeed are so subordinated in the 
 general effect to the space-figure, that you might imagine the order of things 
 reversed, the spaces filled by solid wood, the lines of the woodwork left blank. 
 This shape (fig.A), is perhaps somewhat bizzarre ; but it is not unpleasing, and 
 it is a recognised device in Japanese heraldry. In the crest of one family it 
 appears in the twin form given in the accompanying figure (B). 
 
 A 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 41 
 
 I believe this zig-zag design to be essentially Japanese ; but the idea of the 
 broken line occurs in this well-known form in Tudor Archi- 
 tecture. It must be set down as one of those occasional and 
 curious resemblances, which are, I think, inevitable coin- 
 cidences, between the decoration of the East and West. 
 
 The method on which the Tudor figure is constructed is 
 identical with that of the Japanese, as is shewn by the thin lines in the 
 drawing. The notch or break bears a certain definite proportion to the breadth 
 of the main line, so that the continuation of the upper edge of this line becomes 
 the lower edge of one of the lower sections : the next or the next but one 
 according to the depth of the break. 
 
 The zig-zag design appears in a triple form in fig. 3 (pi. 12) ; it is most 
 dazzling and bewildering ; it occurs also in outline in fig. 25 (pi. 8), with no less 
 than seven lines. Many variations are also produced by changes in the angle of 
 inclination of the lines and of wood-thickness ; and beautiful effects obtained by 
 the use of gold and colour. In the lattice given in fig. 3, the surface is gold, and 
 the thickness of the wood crimson. 
 
 A glance at the illustrations on plates 12 to 15 will shew how the thickness 
 of the wood comes into play as an elemental feature of the designs, the depth 
 giving, as I have already pointed out, real instead of artificial shadows. The 
 effect is also heightened by the light coming through the other parts of the design,, 
 sharpening the edges of the wood. When very thick wood is used, as in fig. 4 
 (pi. 12), where it is an inch and a half thick, and gilded, the lattice has a very 
 massive appearance. The ornament and the lattice bars are in this instance 
 flush ; but this is somewhat unusual. 
 
 The next step in our artist's upward path of excellence would be the very 
 natural one of designing an ornament for the lozenge spaces of the lattice ; and 
 he also very naturally had recourse first to the standard four-petalled floral 
 forms. The one used in fig. 4, with the scalloped petal seems indeed to have been 
 specially created for the zig-zag lozenge, so beautifully does it fit into its crannies. 
 
 In fig. 5, we have a simple lattice of triple bars, the lozenges being filled 
 with a deeply cut ornament. It will be noticed that the intersection of the 
 lattice bars produces the four little lozenges arranged lozenge-wise, which was 
 noticed in fig. 3 (pi. 4) of the diapers. The ornament in fig. 5 is not flush with 
 the bars, but has the appearance of being imbedded in the design. It is 
 treated in the same way in fig. 6 (pi. 13), which is one of the most graceful 
 of this class, the effect, in spite of the heavy wood used for the bars, being very 
 light. 
 
42 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 The next idea which, as I imagine, came into our artist's brain is given in 
 fig. 7 (pi. 13). Instead of introducing an ornament into the lozenges, he has 
 ornamented the bars themselves by the introduction of thin slips of wood fasten- 
 ed to the inside, which obliterate the lozenge and produce a novel internal shape, 
 less formal and less wearisome to the eve than reiterated undecorated lozenges. 
 This is a simple illustration of the effect alluded to above, produced by 
 different wood-thicknesses. The inner pattern is gold and the bars black. 
 
 The idea of introducing these thin slips of wood as an element of the 
 designs having once been originated, it was rapidly developed ; and we come 
 almost naturally to the very ornamental lattice given in fig. 8 (pi. 13). The 
 point to be specially noticed, in view of the subsequent developments of the 
 idea, is that the internal slips on the four sides of the lozenge, are identical in 
 shape ; and therefore correspond on both sides of the crossing bars. This results 
 from the fact that the lozenge is rectangular and rectilinear; but when lattice- 
 bars are inclined at an acute angle modifications in the shape of internal slips 
 become necessary. 
 
 These modifications will be readily understood if we imagine the lattice, fig. 
 8, to be made of india-rubber, and to be stretched lengthways till the horizontal 
 angles approximate to 30 degrees, which is the angle of grace. The internal 
 ornaments will then become distorted, as in fig. 9. For purposes of construction 
 the following explanation will, I think, be clear. The four internal slips are 
 the same shape, and any two adjacent ones are merely the opposite pair reversed. 
 One interesting feature ot this design is that the space-shape corresponds in 
 its general configuration with the lozenge ornament fig. 6. One other point 
 should be noted ; the reversing process produces the greatest possible want of 
 correspondence in the pairs of slips on the opposite sides of the lattice bars. 
 This lack of symmetry is even more marked in fig. 10 (pi. 13), where the zig-zag 
 lattice is used, with a foliated scroll for the internal slips. These slips are heavier 
 than those in figs. 8 and 9, and the space-shape is much reduced in consequence; 
 but I think it will be clear that in this class the space-shapes hold such an 
 important position in the design that the actual lattice almost becomes the adjunct 
 rather than the principal feature. 
 
 But this lack of symmetry gives rise to another curious effect, which has 
 originated yet another development in these designs. When the eye does 
 concern itself with the woodwork, ignoring the space-shape, it rests on the large 
 design made by four lozenges with their slips at the intersection of the cross 
 bars, which is in the form of a highly ornate St. Andrew's cross. It is however 
 so distorted as to produce little effect. In order to bring this new feature into 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 43 
 
 greater prominence recourse was hud once more to a rectangular arrangement 
 for the cross bars, as in fig. 11 (pi. 14). In this design the principal feature is 
 the foliated Maltese cross which springs from the centre of four of the lattice 
 squares, the crosses themselves being arranged horizontally and diagonally. 
 This lattice is composed of black bars, with blue slips, the thickness of both 
 bars and slips being painted red. 
 
 Fig. 12 (pi. 14), is another very ornamental design of the same class, the 
 bars making rectangular and rectilinear lozenges. The boldness of its curves 
 required no adventitious aid to emphasise them, and the lattice is painted 
 throughout in one colour, a vivid green. The internal slips are nearly flush 
 with the bars, and the shadow lines therefore very slight. 
 
 One minor point in fig. 11 and 12 (pi. 14), must not escape attention. The 
 slips, in addition to the principal cross-figure make also a subordinate cross, 
 which alternates with it in the horizontal arrangement. It is composed of the 
 meeting limbs of four principals, so that the foliation is set inwards. I have not 
 come across any example in which this secondary figure is emphasised by central 
 radiation ; but use is made of this characteristic in the two sets of space-shapes 
 in fig. 14 of the same plate. 
 
 In fig. 13, however, two different cross-forms are in fact introduced. The 
 lattice itself is elaborated by the introduction of double bars alternating with 
 single ones. The bars in this design are of thick wood, gilded ; the slips are 
 thin, giving deep shadow effects, and the two sets of crosses are emphasised 
 by different colours, dark blue and vivid green, with the wood-thicknesses red. 
 
 In fig. 14 (pi. 14), we reach the last stage of development of the lattices. 
 The angle of grace is reverted to, and the space-shape again predominates. It 
 will be noticed that in fig. 12 the method of arranging the internal slips pro- 
 duced two different space-shapes ; by reducing the size of the slips this feature 
 is emphasised, and just as in figs. 11, 12, and 13, there are, as noted above, two 
 sets of cross-figures, so in fig. 14 we get two sets of very prominent space-shapes 
 arranged in horizontal lines. This design differs in its construction from fio-s. 
 9 and 10 (pi. 13), in one important particular. In those the internal slips were 
 all of the same shape, and were merely inverted in order to complete the design 
 on the opposite side of the lozenge ; but this produced a lack of symmetry in 
 the slips on the two sides of the bars. This feature is deliberately emphasised 
 in fig. 14, where the internal slips of the alternate rows of lozenges are of 
 different designs. They have however a " perspective" affinity, produced by the 
 horizontal stretching process which I have imagined to have been applied to fig. 3, 
 which would naturally elongate the initial curve or stem of the design of the slip. 
 
44 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 The natural development is now exhausted. I add, however, a few note- 
 worthy designs, some of which lie outside it. 
 
 In fig. 15 (pi. 14), pure line ornament for the internal slips has been abandon- 
 ed in favour of the quatrefoil, which is the only floral form I have found used in 
 lattice-work. The bars arc reduced to a minimum in order to give a greater promi- 
 nence to the quatrefoils, which are painted alternately bright blue and vivid green. 
 
 In fig. 16 (pi. 15), the principal figure, which has some affinity with Renais- 
 sance scroll ornament at its best, fills four of the lattice squares. In its con- 
 volutions it is based on that cunningly-balanced distribution of light and shade, 
 to be hereafter referred to, which is one of the most distinguishing features of 
 .Japanese decoration. 
 
 In figs. 17 and 11), the circle appears for the first time in these lattices. Its 
 infrequent use in this class of work is worthy of note, in view of the special 
 position which, as we shall presently see, the circle holds in eastern art. It 
 seems probable that these designs owe their origin to foreign sources; though 
 whether they can be said to be so Romanesque in character that this origin 
 would be immediately recognised is doubtful, for Owen Jones indicates geomet- 
 rical mosaic as the salient feature of this style of ornament. It is a fact, 
 however, that both these designs appear among Owen Jones's examples of 
 Romanesque design as it developed from Byzantine; fig. 17 as fig. 20, and fig. 
 19 as fig. :>">, of plate 30 of the "Grammar." In fig, 17, the lattice bars are black 
 and very massive ; the circles are deep blue, the floral ornament light blue, and 
 the wocd-thickness red throughout. In fig. 19, the circles are gold, and the 
 ornament crimson, picked out with pink. 
 
 Fig. 19 is an elaborate chain-work design, which, though clever in concep- 
 tion, is hardly of first rank. The central ornament is, worthy of attention ; it 
 is in blue and red alternating with green and red ; the chain in gold. 
 
 The last design, fig. 20 (pi. 15). comes from the panels of a gateway in 
 Nikko. The ornaments in the hexagons are rather weak; but thev do not force 
 themselves into prominence, the eye being attracted more to the general effect 
 of the massive red lacquer network, which is the setting for the central device, 
 the crest of the Tokugawa Shoguns. 
 
Ik jJf 
 
 Fig. 1 
 
 Fig. 4 
 
 Fig. 2 
 
 
 Fig. 5 
 
 Plate 12 
 
 Fig. 3 
 Lattices A 
 
Fig. 10 
 
 Flute 13 
 
 Lattices B 
 
Fig. 12 
 
 Fig. 13 
 
 *$4£j 
 
 Fig. 14 
 
 Plate 14 
 
 Fig. 15 
 
 Lattices C 
 
Fig 17 
 
 Fig. 16 
 
 Fig. 18 
 
 JU+lrf 
 
 Fig. 19 
 
 Plate 15 
 
 Fig 20 
 Lattices D 
 
THE DECORATION 
 OF THE JAPANESE 
 
 Temple decoration lias been put in the foremost place in this book because 
 it is less known than any other form of Japanese art, and because it is of a 
 higher order than decorative art in its more usual forms, and lies in great 
 measure outside the better-known conventions. As everybody knows who has 
 even the most superficial acquaintance with the country, art abounds through 
 the length and breadth of it ; it is part of the life of the people, manifesting 
 itself in its everv detail. 
 
 Without doubt, much of it with which we are familiar runs in grooves ; 
 tradition seems to have cast its iron hand upon it, convention to have gripped it 
 by the throat. The flying swallow and the pensive crow: the cherry blossom 
 and the pine branch, appear and re-appear, do duty alike in their pictures as in 
 ornament for plates and bowls, for inro and sword-hilt, and for ladies' robes. And 
 yet the charm by which it holds the West is its infinite variety. Of the many 
 curious art-problems which Japan has set before us none so difficult to solve as 
 this: how is it possible for convention and art-vitality so to co-exist? for the 
 mere drudgery which it enforces should destroy all the germs of artistic instinct 
 in its students, leaving few to become scholars, fewer still professors. 
 
 Tbe necessity for more numerous and more rigid conventions than in the 
 West arose in the first place from the overwhelming difficulty of the technique; 
 in the second place from the enormous demand for journeymen artists. Art 
 enters so largely into the conditions of life, even in its commonest forms and 
 daily occurrences, that the supply of workmen could be kept up in no other 
 way. There has been at all times a very holocaust of young men, their art- 
 spirit crushed by the rigour of their apprenticeship, in order to satisfy the 
 ceaseless demand for work of even average merit, their pittance not much more 
 than sufficient to supply the daily rice. The art-students of Kensington know 
 no such toil. Ten hours a day for many days to learn the trick of the flying 
 wing; ten more on other days to know how much water the hinge of the brush 
 must hold to catch the swelling fullness of the swallow's breast; ten more on 
 other days to learn how, with the same brush, to put in the crisp lines for beak 
 
50 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 and eye and claw. And in the end, more often than not, failure to come any- 
 where near to the high standard which the canon of Japanese excellence 
 demands. Yet this stood always to the credit of the system, that men whose 
 unknown names were legion went out into the world sufficiently dexterous to 
 supply decoration for the common things that the people use, whose art-training 
 in the school of convention put their work upon a level which was never so low 
 as that of the journeyman artist-dauber of the West. 
 
 But again, convention had so large a grip of art, so pervaded it, that itself 
 assumed an infinitely various character; and no attempt to summarise even the 
 convention of any particular school could hope to be successful. And yet again, 
 the ait has so many different forms of expression, that the same subjects seem 
 to take a different form, to have a different meaning, arising from their treat- 
 ment and the medium in which they were executed rather than from their 
 intrinsic merits as design. 
 
 This brings me to the most serious difficulty in dealing with Japanese 
 decorative art : the definition of its limitations. In the East pictorial art is 
 used as freely as in the West for purposes of ornament; but the line between 
 representative and decorative art is not so easily drawn, because convention, 
 which is the soul of decoration, enters so largely into the execution of Japanese- 
 pictures: in none more so than in landscape. But when the Japanese use 
 pictorial art for decorative purposes it is in fact their pictorial art, and it con- 
 forms to the canons of that art. It is beside the point to say that those canons 
 are of a low order compared with our own. It is quite possible to admit even 
 that their pictures more nearly approach decorative art by reason of the crude- 
 ness of their conventions ; for that again is beside the point. All the limita- 
 tions, the faults as many consider them, of Japanese pictorial art spring from 
 the imperfect means which the artists have at their command. You cannot do 
 much with water-colours if you have nothing to paint on but blotting-paper; 
 and the rapidly-absorbent paper or prepared silk on which the Japanese are 
 compelled to paint their pictures is not very far removed from it. Hence 
 an initial difficulty, which is responsible for nearly all of the peculiar qualities 
 of their art, good as well as bad. The wonder is that they should have achieved 
 so high a technique. I recommend anyone who is anxious to understand, to get 
 a piece of Japanese paper and try to lay on a wash He will turn from his task 
 a crosser but a wiser man, and be more lenient in his judgments in future. 
 
 The question here involved is an important one and deserves consideration 
 for a short space, for it lies at the foundation of the subject in hand, and not 
 every one holds the same views upon it. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 51 
 
 The greatest purist among decorators would never deny that pictorial art 
 may very properly be applied to the purposes of decoration ; but he would 
 insist, and rightly, that when it is so applied it should conform to the principles 
 of pictorial art, that the work must be judged by its standards alone. There was 
 a time when — perhaps it is so still — the paintings of Angelica Kauffmann made 
 desert plates fetch a great price. To paint pictures on anything out of the 
 •common, on porcelain for example, rather than on canvas or paper, has always 
 had a great attraction for many. Sevres porcelain owes almost as much of its 
 value to the excellent work of the artists who painted the landscapes and the 
 groups upon it, as to the fineness of its paste and the lustre of its colours. But 
 a base form of pictorial work, work without artistic vitality, is not to be tolerated 
 merely because technical difficulties stand in the way of getting good execution 
 on clay. If I paint a landscape on a potter's vessel I shall not be forgiven its 
 ■crudity, its lack of depth and light; nor if it be a human face, its vacuity, 
 simply because of the difficulties which my material sets in my way. Admittedly 
 the work is much more difficult than if I painted my subject on paper or canvas ; 
 but art is power, and if it succeed, I may, perhaps, look for greater praise for 
 my picture, because of the great technical difficulties which I have overcome; 
 but the attempt was of my own seeking and if it fail, I cannot insist that, by 
 reason of them my art finds legitimate expression in a lower range of feeling. 
 Nor, again, for art is long, can the length of time which work would occupy 
 were it well done, be an excuse for doing it badly. If it is essential to my 
 happiness to cover my walls with red roses and ribbons, and my ceilings with 
 chubby cupidsand all the winged hierarchy of artistic space, I may do so, and 
 infringe no real or imaginary law. But the cost, if it be the work of men's 
 hands, or the difficulty of getting good results, if it be processed, will not be 
 an answer to kind friends who tell me that my variegated patches of colour are 
 
 hideous and mere nothingness, or that my cherubs are up aloft in positions of 
 anatomic impossibility. 
 
 The main objection to the use of pictorial art for decorative purposes has 
 always seemed to me to be the deleterious effect which it must inevitably have 
 upon the art of the designer and the decorator. It leads them to ignore the 
 fact that there are canons applicable to their own art; and in the past it 
 induced those monstrosities which the early Victorians delighted in — those 
 paintings of flowers and animals, of human beings even, on plates and teacups, 
 which were the merest daubs. As daubs even they were poor ; for they did not 
 come within measurable distance of the colour-plates of the Botanical Magazines 
 which flourished at the beginning of the last century. The poor ladies who 
 
52 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 worked at threepence a dozen had at least an accurately engraved flower or leaf 
 to d;iul) upon. But the familiar and gruesome bunches of pink roses tied with 
 hlue ribbons are as nothing compared with another horror which still dwells in 
 my memory — ;i wall-paper on which was daubed a trellice with ivy growing all 
 over it. Why anyone should have desired to make-believe that his drawing- 
 room was an arbour I never could understand; but even admitting his aim to 
 be admissible, the similitude was so poor that it would not have deceived the 
 commonest garden spider. Decorative art in its lowest form supplies the 
 means of obviating such terrible productions. 
 
 Let us turn for a moment to the pictorial art of Japan. Much of it depends 
 for its charm on the simplicity of the means by which it produces its effects. It 
 revels in suggestion. The works of one of the greatest of its schools is in great 
 part in monochrome. The treatment of leaves and flowers often approaches very 
 closely to the conventional treatment necessary to ornament. In these simple 
 black and white pictures much of the detail, even of the foreground, is left to 
 the imagination ; the middle distance is veiled in a misty cloud; the distance is 
 suggested by a few delicate, almost disappearing touches. It is all a convention ; 
 but it is a convention which has this much at least to be said in its favour, it is 
 the offspring of necessity. 
 
 But this brings us to another point of much importance in the statement 
 of the case. The nature of the technique conventions is such that the technical 
 difficulties in the way of executing landscapes in materials less easy to mani- 
 pulate than ordinary pigments are obviously much diminished. The canons of 
 this pictorial art can be observed as faithfully by an artist working with lacquer 
 and gold dust as by one who uses water and Chinese ink. The wood-carver and 
 the metal worker, the embroiderer and the dyer, know that the masses of colour 
 which their materials produce may be made to correspond entirely with the 
 masses of full tone in a picture. Again, that wonderful dexterity of workman- 
 ship, which surpasses all we have ever dreamed of in the West, makes the work- 
 men look upon the hammer, the chisel, the needle, or the knife, as no less facile 
 instruments for producing sweeping swelling lines than the brush. And, yet 
 again, whether they work with liquid pigment or stiff enamel, with threads of 
 silk or with metal inlay, all alike possess a complete mastery over the gradation 
 of their tones, even to the vanishing point. And thus the arts of all of these 
 craftsmen are identical both in spirit and in execution with the art of the 
 painter ; the result, monochrome pictures in shades of gold or steel, in patina 
 of varied lustres, in dves or in silk embroidery, which are as effective, and which 
 are endowed with the same charms as the painted picture. Thus it comes about 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 53 
 
 that the lacquer boxes, the porcelain, the silk or cotton raiment, and all the 
 thousand things which add to the charm of life to the Japanese, are embel- 
 lished with landscapes executed in precisely the same way, with the same firm 
 lines and evenly-covered surfaces, with the same gradations of tone, the same 
 dark shadows and fleecy clouds, as those which came from the studios of the 
 Kano masters. 
 
 This class of work is exceedingly interesting from an artistic point of view, 
 and its novelty from our own standpoint makes it fascinating to the collector. 
 It appeals to those learned in such matters on account of the magnificent dex- 
 terity of workmanship which ignores rather than overcomes the difficulties which 
 the most unplastic of materials puts in its way ; but it falls outside the intent of 
 these studies, and I have only alluded to it at this length in order to make tin- 
 scope of them clear. 
 
 I now turn to the purely conventional ornament of the Japanese. There 
 is in it, as in the decoration of all countries, a broad division between conven- 
 tional and naturalistic treatment of its subjects ; but it is, 1 think, more sharply 
 marked here than in any other art. Language in its art department is terribly 
 poor, and writers are driven to invent terms, whieh, as often as not, till their 
 meaning is grasped, seem to be in direct survival from the first confusion of 
 speech. But the two expressions, "natural conventionalism' and "conven- 
 tional naturalism," do, I think, convey a definite meaning. By natural con- 
 ventionalism will be understood the conventional treatment of natural forms. 
 The term serves to mark designs which base their form on nature, and to dis- 
 tinguish them from pure line ornament — " geometrical design ' as it is some- 
 times called : " arabesques," and all which is " line for line's sake," and not 
 convention at all. But in conventional naturalism the form of the flower or 
 leaf, which are its most common subjects, is preserved; the treatment only 
 is conventional. It is "form for form's sake"; and it is in this branch of 
 decorative aft that the Japanese so greatly excel. 
 
 The Flowers of Two Rival Arts. 
 
f,+ STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 Natural conventionalism has already received sufficient illustration in the 
 floral forms used in the diapers. They range from the 4-petalled form, common 
 to all systems, to the pseony and chrysanthemum figured on plate 7. Its limita- 
 tions will clearly appear from the consideration of the methods on which 
 conventional naturalism depends. The more essential line of demarcation which 
 it is the object of the foregoing remarks to emphasise, is that which divides con- 
 ventional ornament in both its forms from naturalistic art in its application to 
 ornament. The field is a very wide one, and I am compelled to adhere to the 
 main idea on which this hook is based — that it is confined to "studies" — and 
 limit the number of illustrations. Those which i have selected are, however, 
 typical of the excellence of this branch of decorative art as practised in Japan. 
 
 The first is a bnnch of pa'onies on plate 17, very free and informal in its 
 inception, and is a fine example. It is taken from the lid of a black lacquer 
 writing-box. The outline of the petals and their arrangement are true to 
 nature: that is how one would draw a pieony. .But the bold outlining is the 
 first step in the process of conventionalising the flower. A comparison between 
 its decorative form given in fig. L6 (pi. 7), with its regularly arranged, and 
 evenly balanced petals, and the flowers in this illustration will shew better than 
 any verbal explanation, the essential difference between the two .styles of dec- 
 oration. In the one the form of the flower is used for the purpose of creating 
 a formal design ; in the latter the flower itself is used as an ornament. And the 
 difference between this decoration and pictorial ait should as (dearly appear. 
 The form and natural structure of the flower are faithfully adhered to, but there 
 any attempt at representing the flower pietorially ceases. Its treatment differs 
 not only in the hard outline, but also in the complete absence of modelling; 
 the bellying curves of the petals are indicated only by the curved lines which 
 stand for the veiuing in nature. And one further point must be noted. This 
 method of treatment distinguishes decorative work from monochrome pictures, in 
 which both outlining and formal veiniug are of course absent. 
 
 One important point should here be noted. A strict adherence to the 
 natural structure of the flower, that is, to its growth, is as important in decora- 
 tive art in both its forms as it is in naturalistic painting. All writers have 
 insisted on this as fundamental to good decoration. But it is worth while 
 repeating the exhortation, and pointing to the Japanese adherence to it, for 
 unless it be obeyed religiously a design is worthless, conveying no meaning or 
 suggestion. The best example of conventional naturalism is the treatment of 
 the lotus flowers, with the gold-lined petals, in Buddhist kakemono, with which 
 many of my readers must be familial-. 
 
STUDIES IN THE DECORATIVE ART O F JAPAN 55 
 
 Reverting to the paeonies in plate 17, the leaves are in the normal shade of 
 dull gold, which is used as a Hat wash, with brighter veining; one only (the 
 yellow one in the illustration) is in very dull gold with darker veins — probably 
 a withered leaf. The petals of the larger flower are of bright gold tinted with 
 crimson ; the smaller flower is a thin lamina of gold foil, veined with dull gold. 
 
 Two other examples of this style will be found at the end of the book. 
 Plate 31, also from the lid of a black lacquer box ; is the Nanten (Nandma 
 domestica), which brightens the Japanese garden with its scarlet clusters in 
 January, and takes the place in English homes of our native holly berries. The 
 sprays of leaves are very simply treated in dull gold, surrounding two crests, 
 with the berries in brilliant red lacquer. 
 
 The freedom of drawing which the direct reference to nature in this work 
 admits of is illustrated by the two iris designs (plates 31 and o2), with which 
 the book ends. They are continuous, and form the two sides of a black lacquer 
 luncheon-box; another similar picture is on the other two sides. The leaves 
 are in dull gold veined with brighter lines; the small spots of white in the draw- 
 ing are tiny patines of bright gold, which always figure largely in gold lacquer 
 work. The ripples of the water are outlined in gold; the buds and the flowers 
 which are black in the illustration are in brighter gold, and those which are 
 left white are treated like the pseony in plate 17, with thin gold foil. 
 
 Two designs are given on plate 18, a pseony and a phoenix with a bough of 
 Pawlownia, the principal feature of which is the conventional foliated scroll, 
 developed from a trailing plant, with primitive leaf-forms. There is in almost 
 every advanced system of decoration a scroll peculiar to it; the most familiar 
 being; the acanthus-scroll of Roman architecture. The one used in these two 
 designs is, I think, purely Japanese, derived from the Chinese. It is an elemen- 
 tary form of one frequently seen on the chased brass-work in the temples, to 
 which I propose to devote a little space presently. In the coloured drawing, 
 plate 19, this scroll becomes more formal, and shews a remarkable analogy in 
 form, in its systematic meandering, and its colour treatment in three shades of 
 body colour, to that used by the early English missal painters, examples of 
 which are given in the " Grammar of Ornament ' (plates 71 to 73), and of 
 which Owen Jones himself made such effective use in his own work — as in the 
 title-page of the " Grammar." This charming panel comes from the pediment 
 to the porch of the shrine to the Abbot Jigen Daishi at Nikko, which is in an 
 unfrequented spot on the hills behind the temple of Iyemitsu. 
 
 The design at the top of plate 20 is a narrow open-woik panel of 
 chrysanthemums and butterflies, in which a very light and graceful effect is 
 
fir. 
 
 studies; in the decorative art of japan 
 
 produced by very simple means. Three shades of gold are used; dull for the 
 leaves, with a brighter tone for the veins, and highly burnished gold foil for the 
 Mowers and wings of the butterflies. 
 
 The principal figure of the lower design on the same plate is a design of 
 great beauty and originality, a conventionalised pine tree. It is one of those 
 knotted gnarled trees which grow on the fast-vanishing moats of Tokyo. In 
 spite of the treatment to which it has been subjected, all the features of its 
 rugged growth have been preserved, and have been bent into the circular form 
 of the design. But this brings us to another peculiar feature of Japanese 
 decorative art, the use of the circle as a basis of design, which requires consid- 
 eration in a special section. 
 
 I conclude this section with a few examples of the adaptation of floral 
 forms to architectural purposes (plate 21). Fig. 4 shows the lotus leaf used as 
 a rail support ; fig 3 is the common form of bronze capping to a newel of red 
 lacquer. In fig. 5 the pointed ball is replaced by an inverted lotus flower, 
 which makes a cap of neat beauty. In fig, 2 the flower is treated more formally, 
 in gilded brass on a newel of black lacquer; and in fig. 1, the same form is 
 used still further conventionalised. 
 
 I'me tree highly conventionalized. 
 
 I (esign for wall-paper. 
 

 < 
 
 os 
 
 D 
 
 ■< 
 Z 
 
 J 
 <! 
 Z 
 
 o 
 
 z 
 w 
 > 
 
 z 
 o 
 
 I- 
 
 I— 1 
 
 CL, 
 
05 
 
 < 
 
 OS 
 
 a 
 
 - 
 55 
 
 o 
 
 z 
 
 a 
 > 
 55 
 
 o 
 
 o 
 
 0> 
 03 
 
Plate 19. 
 
 CONVENTIONAL NATURALISM- C 
 
 with Conventional Sen ill -work. 
 
3 
 
Fig. 1. 
 
 Fig. 2. 
 
 Fig. 
 
 Fig. 5. 
 
 Plate 21. 
 
 FLORAL FORMS USED ARCHITECTURALLY. 
 
THE USE 
 
 OF THE CIRCLE 
 
 IN DECORATION 
 
 The circle for us Westerns is a commonplace figure, possessing little 
 inherent beauty. It has ceased to interest us since we learned its dull proper- 
 ties at school. Even the fact that it has a capacity for holding certain inscribed 
 figures, themselves not without interest, does not excite our artistic imagination. 
 We have never appreciated how adaptable it is to decorative treatment, nor 
 realised that it could serve as a basis for ornament. But in the East the circle 
 holds a charmed position, which I think may be attributed to certain inherent 
 qualities which have passed unnoticed in the West, without having recourse to 
 any theories based on the fact that the Sun's disk is also a circle. Of course 
 the mirror which induced Amaterasu, mother of all curious women, to venture 
 forth from her cave, was round. But although the Sun Goddess, without doubt, 
 presided over the mysterious birth of Japan, and although the mirror is among 
 the sacred treasures, we must I think simply accept the fact that both Sun and 
 mirror are round, , without attempting to derive the artistic properties of the 
 circle from it. 
 
 One of these qualities is a mysterious power of bending everything, no 
 matter how stubborn its nature, to its own shape, so that even that decoration 
 which gives nearest to nature, in spite of its characteristic play of fancy, 
 submits most gracefully to its yoke. Another power it has, that of inspiration ; 
 for it brings out all those fine qualities which are innate to the Japanese artists, 
 invention and composition, developing them to the full. To this belief of mine 
 the pine tree on plate 20 stands as witness. 
 
 A bent and stunted pine tree does not suggest much beauty of line or grace 
 of composition to an English decorator; it may have other qualities which 
 appeal to the artist, but certainly not these. Yet among the many things which 
 live as memories of this land of beautiful things, the gnarled pine trees of Tokyo 
 hold a very foremost place. There is something in the way they droop over 
 the moats, which, for the Japanese, is a silent poem; and it has so acted on 
 their imagination as to inspire some of their most' beautiful decoration. Even 
 the clumsy clusters of needles have suggested as we have seen, a delightful 
 
63 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 formal diaper ; but here the whole tree, with all its ruggedness, has been bent into 
 the charmed circle. And the bough of plum blossom, which the Japanese hold 
 in peculiar tenderness as one of their " Friends in Winter," knotted and ungainly 
 as the pine, has yielded to the same potent influence. On the lid of a gold 
 lacquer box, which is the top figure on plate 22, it forms a design, which for 
 grace of line, and for every other highest quality of perfect decoration, stands 
 as I venture to think, unrivalled. 
 
 Hut before I deal with these and the other circular designs on the coloured 
 plates, 23 and 24, I must devote a litttle space to the circle itself, and its use 
 first in symbolism, and then in art. 
 
 The quality of the eircle with which we are most familiar is that it gives us, 
 by help of its radius, the inscribed hexagon. This is not ignored in the East, 
 as some of the diapers shew ; but its capacity for subdivision in a somewhat 
 subtle fashion, is the feature which has been specially emphasised. 
 
 In this figure, known to the Chinese as "the yavg and the yiv t " a 
 continuous line composed of the upper and lower 
 halves of circles drawn on two opposite radii for 
 diameters, divides the circle into two equal parts, 
 which from their interlocking have caused the whole 
 figure to be taken as the symbol of completeness, 
 and hence perfection. Every correlation of ideas 
 which makes a perfect whole finds expression in this 
 symbol : the male and female elements, light and 
 dirk, positive and negative, and generally all things 
 which are the complement of each other, and are 
 equal to one another. 
 
 An interesting point of contact between the East and the West must here 
 be noticed. The accompanying figure, a well-known 
 feature in Gothic architecture, is based on the same 
 continuous curves, which the customary points and 
 spaces of the Gothic style do not conceal. The sub- 
 sidiary figure made by each half is often used inde- 
 pendently in the tracery of stained glass windows. 
 
 Such coincidences are inevitable, and no conclu- 
 sions of influence or borrowing can be drawn from 
 them; for the primitive elements of line on which 
 decoration is based are so few, that it would be 
 astonishing if the same forms were not evolved independently in different conn- 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 64 
 
 tries. In this case a child playing with a pair of compasses, as those of us who 
 have done it know, would inevitably hit upon this Chinese figure. Although 
 its construction makes the dividing tone a mechanical curve, yet the method is 
 so simple and natural that it may well take rank as a natural curve, and 
 the figure it produces as a natural figure. Such figures have always made a 
 strong a-ppeal to the eastern mind, and were used by them as symbolic of natural 
 things evolved in the same simple and natural manner. The peculiar fascina- 
 tion of the figure in question is, I think, sufficient to account for its symbolic 
 attributes. We are, however, more concerned wifh the artistic use to which it 
 was subsequently put ; and in the designs to be presently studied there is a 
 subtle magic of subdivision of light and shade, which I cannot help thinking 
 owes much of its charm to the influence of this symbol ; for completeness, the 
 perfection which it typifies, is the aim of every Japanese artist, and of the 
 decorators in particular. 
 
 There are other circular forms which must be noticed here, for they are 
 often confused with the yangnnd the yin. They are based on a peculiar form, 
 not often seen independently in art, shaped like a 
 thick comma, formed on a segment of a circle, called 
 the Ihmoe, as to the origin of which there are many 
 theories. The most unlikely of them is that it has 
 been taken from the form of a boxing-glove, which 
 it somewhat resembles. Captain Brinkley, in his 
 Dictionary of the Japanese language, gives what is 
 probably the most orthodox version of its history, 
 as it is based on the derivatives of the word — to?no, 
 a leather shield worn round the wrist by archers, 
 and e, a picture, or figure. Hence tomoe came to be the name given to this 
 comma-figure, because it was formerly drawn on the archers' leather shield. 
 The Chinese character devised for it, (2 , was evidently based on the shape of 
 the figure. This, however, does not give us any clue to its origin ; and I am 
 disposed to think that it was suggested by the form of the eddies of w r ater when 
 it is violently swirled ; the archers taking it as their device arbitrarily, as other 
 devices are taken for mon or emblems. There is no other natural form which at 
 all resembles it, (the comma itself can hardly be called a natural form : at least 
 it is unknown in the East) ; and the suggested derivation has this much in 
 favour of it, that it is in accordance with eastern tradition, which is almost an 
 
 article of artistic faith to the Japanese, to look to nature as the parent of all 
 artistic form. 
 
c>:> 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 This comma-form gives rise to two well-known devices composed of two 
 and three ''commas' respectively, arranged in a circle ; and called the fvlatsu 
 tomoe (the double comma), and the mitsu tomoe (the triple comma). 
 
 The futatsu tomoe is in no sense a " complete 
 design, by reason of the gap between the two commas ; 
 and this distinguishes it from the yang and the yin. 
 Yet, as I have said above, it is often confused with it ; 
 notably, in connexion with the national emblem of 
 Corea, which is the symbol of completeness, four of 
 the " 8-trifframs ' being introduced into the corners 
 of the flag. I have seen the mistake made even in 
 Corea itself, but in places where in all probability 
 the draughtsman, was a foreigner. 
 
 How the " yang and the yin " came to be the 
 emblem of the Hermit Kingdom I have been unable to 
 ascertain. It will not be uninteresting to note another 
 kindred figure, the triple arrangement of curves u«ed on Corean fans, painted 
 green, red, and yellow. This has clearly no affinity with the tomoe, but it is 
 conceivable that it might have been developed from the yaruj and the yin." 
 
 I have also a note of the emblem carved in wood, thus ; but it 
 is probably a degenerate form. 
 
 So the circle being established in emblematic use in the country, and the 
 thing symbolised — perfection — being the attribute of the Sovereign Ruler of 
 Japan, it followed in the natural order of things that the Imperial emblem should 
 be based upon this perfect and complete symbol ; and the national flower, the 
 
STUDIES IN THE DECORATIVE ART OF fAPAN 
 
 60 
 
 chrysanthemum, lent itself readily to the treatment. And for the Shoguns of 
 the Tokngawa dynasty, the leaves of the marsh-mallow did not need much 
 artistic persuasion to fall into a circular design. 
 
 And presently all the nobles followed suit, and the circle became the 
 accepted basis of the mon or family emblem. The number of them in these 
 days is Legion, but among the multitude I select the following characteristic 
 designs. 
 
 5#£ 
 
 
«7 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 The exceptions to this circular law of perfect heraldry 
 are few ; it is however curious to note that the Pavlownia , 
 the second of the Imperial emblems, has never been subjected 
 to it. 
 
 Putting- all questions of "grant" on the side, in select- 
 ing the device for the mon, the family would naturally have 
 been influenced by its favourite flower or plant ; and in this 
 connexion I note two, in which three leaves of the beautiful Salisburia 
 adianlifolia have been woven into an artistic device. 
 
 In designing the mon, an artistic arrangement of the subject was not likely 
 to be forgotten. And so, in the most natural manner possible, came the next 
 step, the adoption of the circle as a basis for a special class of ornament ; which 
 had this special virtue, that while it preserved the rigid outline of the symbol, 
 yet drew from it an inspiration from its internal curves, a suggestion for the 
 development of its own internal decoration, which gave full scope for the play 
 of artistic fancy ; and fancy literally ran riot with it. 
 
 One other development must be noticed before we turn to the designs. In 
 due course the decorated circles themselves became items of ornament, and were 
 used in combination for the decoration of surfaces, which are often, to adopt a 
 term of heraldry, sem£e of circles : semee, of course, with that peculiar and 
 subtle grace of arrangement, that " most admired disorder," which is the key- 
 note of one of the modes of Japanese art. Then, when once these decorated 
 circles were used in this way, they came naturally to be treated in the same 
 manner as other items of ornament — clouds, waves, leaves, flowers, petals; and 
 were massed or broken up, on what, for want of a better name, may be called 
 the " broken-cloud ' principle. I think this term will convey its meaning to 
 all who are in the least familiar with Japanese decoration. The use of these 
 circles in combination, large with small, and in groups of two or three, will not 
 escape attention. 
 
 I come at last to these two plates, 23 and 24, which give a series of designs, 
 in which the circular arrangement of conventionalised flower-forms touches its 
 highest development. They are taken from a very beautiful black lacquer 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 68 
 
 luncheon-box, the property of Captain Brinkley, R.A., of Tokyo. It is a pile of 
 five square trays, with an ultimate lid ; the whole being about 8 inches square, 
 and 15 inches high. The box is a very conservatory of flowers, for it is simply 
 covered with these floral circles of various sizes. There are the preony and the 
 plum, and a wealth of blossom besides — iris, lily, apple, dianthus, lychnis, 
 hydrangea, Cape-gooseberry. Some plants lend themselves more readily than 
 others to this treatment : like the iris, wherein nature's law of growth asserts 
 itself at once to the eye. But with that consnmate art of design, of bending lines 
 to their will, which characterises the Japanese, no plant so stiff or crabbed in its 
 habit, not even the floral monstrosity the " cock's-comb," but is made to conform 
 to this nil-compelling Jaw ; and that, for its obedience, is endowed with a charm 
 which is not its own. The colours in the original are as bright as nature's 
 brightest ; bat the strict canons of conventionalism, on which I have so much 
 insisted, are never once departed from. On the stalks and calyxes are sprinkled 
 those tiny flecks of bright gold leaf, already referred to as being so prominent a 
 feature of the best lacquer work ; but they are so numerous and so minute, that 
 I have found it impossible to reproduce them. 
 
 The great lesson which this series, and every design of the series, teaches, is 
 much needed by western artists. It is not merely the graceful now of line which 
 charms and holds us, but even more the subtle distribution of the masses of 
 colour and space, and consequent repose to the eye ; and this in spite of an 
 inequality which is very perceptible, and a certain touch of the hap-hazard. This, 
 the greatest virtue which designing can possess, is I believe, purely Japanese ; for 
 I cm find but little trace of it even in the work of the Chinese. 
 
 The position which blank space holds in the art of the Japanese is responsi- 
 ble for one of its most subtle and attractive qualities. In their pictures there are 
 great spaces without so much as a line or faint wash of colour. They are not 
 " wasted," as we should think, and sometimes say ; they take their share in the 
 general effect, as any one may prove to his dissatisfaction, if he cut off the blank 
 space from one of his pictures. So it is in these designs. There is in all of them 
 a mysterious sense of proportion of colour to space, never the same, in which 
 the eye delights as much as it does in the delicate tracery of the design itself. 
 
 To unravel the mystery is beyond my power ; but I may, in the lightest and 
 most Japanese fashion, suggest what perchance underlies it. If delight in the 
 beautiful means anything, it must be an intellectual delight, the cause so acting 
 on the brain as to set up those vibrations to which we give the name of pleasure. 
 We use two other terms in our talk of art ; the commonplace, and the eccentric ; 
 they refer to things in which there is no beauty. In the former we find no 
 
»;«> STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 pleasure , the latter we reject as irritating. The commonplace sets up no brain- 
 vibrations at all, leaving it quiescent, but slightly annoyed ; the eccentric stirs 
 up commotion in the brain, which as often as not creates disgust. 
 
 These excluded, we are left with the unknown. For "commonplace' 
 connotes a knowledge of the processes on which such effect as there is depends ; 
 and the " eccentric " forces itself on the attention, induces a restless seeking for 
 knowledge as to how the effect is produced : making for brain-weariness, and the 
 final rejection of so troublesome a thing. In these art appears in its most 
 unbeautiful aspects. Between them lies beauty in its thousand forms, and we find 
 ourselves in the presence of art's mystery, unknown and unknowable, like the 
 greater mystery of life of which it is so large a part. And so between that 
 mental blank, quiescence, and mental commotion, lies calm, contented, undis- 
 turbed repose. 
 
 Yet, although I think that the beauty of Japanese art depends in large 
 measure on the existence of an unknown and unascertainable proportion in its 
 composition between light and shade, between the design and its spaces, the 
 artists do not altogether reject equal proportion, what we call " balance," in some 
 things. There never yet was a proposition in Japanese art that it did not seem 
 itself to refute. In the symbol of completeness, the "yanff&iid the yin" the two 
 parts are not only complementary, they are also equal. It is impossible to 
 imagine that the figure made W a curve composed of the upper and lower halves 
 of two circles on the same diameter, unequal but tangential to the circumference 
 and to each other, thus dividing the circle into two unequal parts, should 
 symbolise completeness. In the " pakwa diapers," which we shall presently 
 study, the space between the emblems holds an equally important position with 
 the emblems the.nsel\es, and in one class design and space are equally balanced. 
 This may of course, be referred to the necessities of the design ; but for its general 
 acceptance in art work of a high order there is another explanation. The same 
 art principle is at work ; the rejection of the commonplace is patent. 
 
 The circle divided by its diameter also gives two equal and complementary 
 figures ; but again, who could imagine hat such a subdivision of it should 
 symbolise "completeness?" It is baldly common-place; it ostentatiously asserts 
 the fact that the two parts are equal. But in the " yan.g and the yin" there is 
 that pleasant little borrowing of each half from each, which arouses our curiosity, 
 quite apart from the substitution of a curved line for a straight. And in the case 
 of the paktva diapers, the necessary equal distribution of design and space which 
 pervades them is hidden in the meandering of the lines holding yet hiding the 
 
STUDIES IN THE DECORATIYE ART OF JAPAN 70 
 
 symbol, the beauty and fascination of them increasing as the secret becomes more 
 and more closely guarded. 
 
 And so I come back once more to these floral designs on my friend's luncheon- 
 box, to the pine tree of the moats of Yeddo, and to the still more beautiful 
 bough of blossoming plum, to note the influence on the designing of the inner 
 curves of the symbol of completeness. They are, I think, directly responsible 
 for the grace of the tangential flow of the inner lines of these designs. There is 
 one among them, borrowed from nature's humblest offspring, the grasses, which 
 tells the story in no unmistakeable fashion. It is one of the most charming and 
 original of the series, with its beautiful flow and play of line, the art of composi- 
 tion so concealed that they seem to be bending to the breezes on the Yumoto 
 
 plain. 
 
 The tangential law of design, on which all writers on the art of decora- 
 tion have insisted, is easy to account for, because it is nature's law of growth. 
 Hut the application of it to these circular designs, and the bold sweeps of 
 their inner lines, are traceable, I think, to the influence of the inner curves of 
 the symbol. It need be put no higher than a reminder of the natural law ; but 
 in some a more direct influence is unmistakeable. 
 
 I have already imagined the circle equally divided by its diameter, and 
 noted the very plain result. Let me now imagine it equally subdivided by any 
 curve passing through the centre, each side of it identical with, but the reverse 
 of the other. Again the circle would be divided into two equal and complemen- 
 tary figures. It would not be commonplace it is true ; but all sense of complete- 
 ness would be gone. There would be no reason apparent for the sub livision ; 
 the resulting figure nothing but the expression of an eccentric thought. Perfec- 
 tion could never take such a thing for its symbol. But the figure which the 
 East has taken to symbolise its notion of perfection, in additfon to the manner of 
 its subdivision, assimilates the natural law of growth, and by the tangential devel- 
 opment of its inner curves, suggests one more charm to the designs based 
 upon it. 
 
 This analysis has perchance gone too far, perchance not far enough, leaving 
 many things unexplained — the ugl\ the influence which individual and educated 
 taste must exert upon the discriminating mind, the thought which shapes itself 
 in the formula of those who refuse the discipline of art — " I know what I like, 
 and I don't like anything else." For the rest, the unexplained, I must rely on 
 that maxim to which even the precision of the law sometimes appeals when a 
 thing is so plain that many words would only confuse its appreciation — res ipsa 
 loquitur, and so leave it. 
 
71 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 There remains but one more design to speak of. On the inside of the lid of 
 the gold lacquer box, the second figure on plate 22, there is a little heap of fallen 
 blossoms ; they have indeed begun to fall on the lid itself. If their arrangement 
 seems at first sight fanciful, the fancy which prompted it had an extraordinary 
 sense of completeness. Not one blossom more, bright with gleaming metal — the 
 newly fallen ones; or of dull gold — those which have lain longer on the ground, 
 could be laid there. It is a beautiful example of that space-value of which I 
 have talked so much. But again, there is something more ; nature once more lies 
 hidden as the explanation of its charm, as the key to the absence of all that we 
 call " method." ( )u a bright spring morning, when the light breeze lifts the 
 flowers from the tree, that is how it lays them, thickest near the parent trunk. 
 
Plate 22. 
 
 CIRCULAR DECORATIVE FORMS— A. 
 
Plate 23. 
 
 CIRCULAR DECORATIVE FORMS — B. 
 
Plate 24 
 
 CIRCULAR DECORATIVE FORMS— C. 
 
SCROLL WORK 
 
 AND 
 
 METAL CHASING 
 
 In the two designs on plate 18 a formally conventionalised scroll is in- 
 troduced as the accompaniment to the pseony and the phoenix. I now give a 
 few examples of the scrollwork used in the temples for the chasing of the metal 
 ornaments and end-caps of the beams and pillars. It is also used on the brass 
 corner-pieces of large lacquer boxes. 
 
 Some apologies are due to my readers for the very crude way in which 
 these designs are presented ; they are merely reproductions of pencil rubbings, 
 taken very hastily many years ago. It is only on the advice of a friend of great 
 experience as an architect, and learned in his art, that they appear in this form. 
 As designs they hold an important place in the decorative art of Japan, and the 
 only other method of including them in this book would have been to have had 
 them redrawn. But the actual lines of the chaser would have been lost, and 
 this in itself seemed so important a point, that I have followed my friend's 
 advice ; 
 
 I start with a very simple design on plate 25. It is the original from 
 which the scrolls on plate 18 spring. That form of it, I think I am correct in 
 saying, is the one commonly used for gold ornament on lacquer, while the one 
 given on plate 25 is more adapted for metal chasing. 
 
 This design seems without doubt to be the Japanese variety of the foliated 
 scroll in Chinese style, which was often copied in Japan, on plate 26. The 
 scalloping of the stem is abandoned, the Japanese designers preferring more 
 flowing lines, that of the foliation reduced, and the volute at the base of the 
 leaves omitted. 
 
 With regard to the Chinese scroll, it seems clear that the scalloped leaf- 
 form was repeated from the form adopted for the inturned edges of the petals in 
 the very conventionalised form of pseou'y of which we have already seen an 
 example in fig. 12 (pi. 6), and which is traditional to Chinese art. Here there- 
 fore, we have the genesis of the Japanese scroll. ■ > 
 
 It appears in a more elaborate form on plate 27, on which are given two 
 brass end-plates of temple beams. The sweep of the stem of the scroll conforms 
 
7G STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 to the best traditions of the classical scroll-work of the West, and is obedient to 
 the law of tangential growth, without which scroll-work is mere nothingness. 
 This law is equally applicable to the foliation; for the scroll must have 
 originated from a trailing or climbing plant; though curiously enough, the 
 acanthus scroll is an artificial creation, for that plant is not a creeper. The 
 distinguishing feature of the scroll at its best, is a complete mastery of free- 
 curve drawing, and it was only natural that the Japanese artists should have 
 produced beautiful specimens of this class of work. They also shewed a re- 
 markable facility for adapting the scroll to the exigencies of the plate to be 
 ornamented. This is particularly noticeable both in the principal curve of the 
 large design on plate 27, and also in the foliage which springs from the terminal 
 volutes of the smaller plate. 
 
 It must not be imagined that I attribute this facility solely to the Japanese 
 decorators; it is the attribute of all masters of the decorative art who have 
 designed scroll-work. There is a wood-cut in Owen Jones' " Grammar," in the 
 chapter on Mediaeval Ornament (p. 102), of an Early English scroll, which 
 adapts itself to the spandrel in which it is placed in a similar manner to that on 
 plate 27. 
 
 In plate 28, the foliation of the scroll is still further elaborated, and much 
 larger. Subject to the absence of modelling, otherwise than by line indications, 
 this scroll has a remarkable affinity to the acanthus scroll of Roman architec- 
 ture, both in the method of conventionalising the foliage, and also in the way 
 in which it almost conceals the stem. In this example there is the same adapta- 
 tion of the curves of the scroll to the form of the metal plate as in plate 27, and 
 the same or even greater vigour in the flow of the lines of the foliation. I do 
 not think it is possible here to make any suggestion of borrowing or even of due 
 influence; it is one more point at which the East has touched the West in its 
 decoration. 
 
 In plate 21) the metal is in the form frequently adopted for the protecting 
 cases at the end of beam.:. A pteony in the conventionally natural form with 
 which we are familiar, surrounded by small leaves, is chased on a background 
 of one of the pakwa diapers. 
 
 I have hesitated to multiply examples of this beautiful form of decoration ; 
 but I cannot refrain from giving one more, an example of the departure of the 
 temple artists from the severe ornament which they brought to so high a pitch. 
 It is a simple little idyll ; the marsh-mallow of the Tokugawas, growing in, and 
 swaying to the running river. 
 
.-'•'- — ■:-.'< ' 
 
 ■:<■-'■■ 
 
 
 
 
 
 ■ h 
 
 
 
 tlt^ if 
 
 
 
 fa 
 ^ t 
 
 * 
 
 « 
 O 
 
 o 
 
 (M 
 

 m 
 
 I 
 
 id 
 as 
 O 
 
 o 
 
 X 
 
 o 
 
 SO 
 
 Oh 
 
o 
 
 I 
 
 id 
 
 at 
 o 
 
 SB 
 
 o 
 
 O 
 v> 
 
 
 
 
 S3 
 
 CM 
 
u 
 
 
 WWk 
 
 /yjif lit 
 
 WmimM 
 
 
 K»> 
 
 * :;■ r . ■■■>;■ ■ 
 
 m 
 
 i 
 
 o 
 
 o 
 « 
 o 
 
 m 
 
 % 
 
 mfr 
 
 \ 
 
 00 
 
 CM 
 

 c 
 
 »— i 
 
 < 
 
 X 
 
 o 
 
 < 
 
 h 
 ■A 
 
 •M 
 
 9 
 
 V. 
 

 c 
 z 
 
 < 
 S 
 
 • r. 
 
 o 
 
 
THE PAKWA DIAPERS 
 
 AND 
 
 KEY-BORDERS OE THE EAST. 
 
 ..." «■■.■■ 
 
 
In the foregoing notes I have repeated the familiar statement that the priest 
 in the East, as elsewhere, by introducing religious symbols into the decoration of 
 the temples, made use of art to remind the casual and the curious of the eternal 
 verities to which those symbols gave formal expression. The method he adopted 
 was to introduce them into the diapers with which he lavishly ornamented the 
 walls, the ceilings, and the doorways ; and he exerted all his ingenuity to attract 
 the eye by using every device for beautifying them, arraying them in brilliant 
 garb of glistening light or splendid colour. 
 
 I propose now to study the gradual evolution of the symbolic diapers and 
 borders. In doing so I am led inevitably to trespass into the regions of a debate 
 of far greater importance than the mere invention of a prettily complicated 
 design. The diapers are known to a few as forming the cloisons to eastern 
 enamels, and they have given them the name of " cloisonne patterns." The 
 borders are known as the " Key "-patterns, or frets, and are generally assumed to 
 have originated in Greece. By those who know of their existence in the East, 
 they are further assumed to have travelled thence through India to China, and so 
 to Japan ; a conclusion, as I think, somewhat hastily drawn, and without due 
 consideration of the premises. 
 
 The original notes from which this more exhaustive study is compiled formed 
 the subject of a paper read before the Japan Society in London, in 1893. During 
 the discussion which followed Professor Anderson, the Chairman, made the 
 following remarks ; — As to the origin of the " Key "-pattern, he should say that 
 it was neither Japanese nor Chinese, but Greek. We found precisely the same 
 " key ' in China as in Japan, and we knew that many branches of art did extend 
 from Greece to Japan, together with the Buddhist religion ; and that the Gncco- 
 Buddhistic arts left in certain parts of India at the time of the conquest of 
 Alexander, were transferred to China, from China to Corea, and from Corea to 
 Japan. He should imagine the Greek " Key "-pattern followed the same course. 
 
 If our old friend, than whom no one was better versed in all learning 
 appertaining to Japan, had no further information to guide us, I think I am 
 justified in saying that the only fact we know of is that this design is of greatest 
 antiquity in the East as it is in Greece. 
 
 The question whether a design, found in two widely-sundered parts of the 
 world, developed spontaneously in both, or was borrowed by one country from 
 the other, and which was the borrower, is in itself interesting. But once it is 
 admitted that a religious symbol enters into the structure and development of the 
 
86 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 design, mere artistic interest gives way to questions of grave importance in the 
 history of the world, in which religion has played a part second only to war. 
 
 These ail studies therefore trench inevitably on the larger question of the 
 early relations of the East with Greece, and the place which Grseco-Buddhism ' 
 takes in the religions of the world. My contribution to that subject must be 
 limited to the statement of certain facts which the design itself in its various 
 forms furnishes. They seem to me to be the premises of which we stand in need; 
 and if I am right in the statement of those premises, I doubt if any other con- 
 clusions than those I have indicated can be drawn from them. 
 
 The designs, both diapers and borders, which form the groundwork of the 
 following discussion, were found in actual use in Nikko and Shiba. They were 
 collected at different times, and without regard to the building up of any theory 
 upon them. But the first important fact that they reveal is that in its most 
 primitive form these diapers were based upon an element of the pakwa symbol ; 
 and secondly, that without much ingenuity, there is discoverable among them a 
 perfect and orderly scheme of development from the primitive form. It is un- 
 necessary to dwell upon the importance of this fact in a question depending 
 almost exclusively upon research; and it is not the only case in which a natural 
 older can be (dearly traced in the art of the temples. 
 
 But there is another and still more important fact which I think must be 
 accepted; that the " Key "-borders were not designs which developed independ- 
 ently of the diapers, but sprang from them. 
 
 A priori, it would, I think, defy human ingenuity to contrive designs of 
 such intricacy except from some definite origin, and on some carefully thought 
 out plan. And it is significant thai there is among the borders a regular order 
 of development which runs parallel with that of the diapers : that the same pakwa 
 element, both in its simple and complicated forms, is as manifestly the basis of 
 all the borders as it is of all the diapers. To this a further fact must be added : 
 that the use of this element in the border is fragmentary, only half of it being 
 taken. Again a priori, it seems to me impossible to conceive that a series of such 
 complicated designs as the borders, should have been evolved by the use of half 
 the (dement independently of another series of equally complicated designs which 
 were being evolved contemporaneously by the use of the whole element. Apart 
 from the fact that the evolution of the diapers was the simpler process of the two, 
 the constructive principle, with its elaborate system of interlocking of the different 
 parts of the (dement, is not likely to have been hit upon for borders independent- 
 ly of the diapers ; for the designer of borders must, as will presently be seen, 
 inevitably have dropped into the diaper during the process of designing, by the 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 87 
 
 more accidental prolongation of bis lines. And lastly, every true " Key "-border 
 '•an in fact, by prolongation of its lines, be developed into a diaper. These state- 
 ments I hope to justify in the following elaborate study of both forms of design. 
 
 Now, supposing them to be justified, what light do they throw on the question 
 whether the " Key "-border developed independently in China and Greece? or 
 whether it travelled with the Sun or from the West eastwards? For myself I 
 believe tins question would be conclusively answered by the answer to another — 
 Did the Greeks ever use diapers ? There is little or no trace of them in Owen 
 Jones' " Grammar of Ornament." But for some people this would not be con- 
 clusive ; and I am therefore disposed to say that the answer depends absolutely on 
 another, and as I believe, crucial question, Was the pakwa at any time a recog- 
 nised religious symbol in Greece? If it was not, then the conclusion seems to be 
 irresistible, that the diapers and their offspring, the borders, originated in the 
 East ; and that the borders, though probably not the diapers, travelled westward 
 into Greece. 
 
 One further, but subsidiary point, is, I think, conclusively shewn, if the 
 symbol was not used in Greece — that the whole of the complicated system of bor- 
 ders travelled together, and not isolated fragments of it, from which the system 
 was reconstructed in Greece. For the process of evolution depends on an accurate 
 and familiar use of the pakwa symbol ; and if that was not known in Greece, it 
 could not have been introduced promiscuously into the different forms of the 
 diaper by unlearned artists. 
 
 The probable development of the soastika, and the designs based upon it, 
 from the pakwa designs, will be noted during the progress of the discussion. 
 
 One word of warning I must here give to the reader. This discussion is in- 
 evitably highly technical, and possibly tedious. It is intended primarily for 
 those who wish to study, and perhaps themselves to construct and use these de- 
 signs. I trust, however, that it may be found sufficiently lucid for the ordinary 
 reader to follow the thread of argument which runs through the discussion, even 
 if he have to accept on trust some of the details of construction on which the de- 
 signs are based, and on which the gradual evolution, of the more complicated ones? 
 both diapers and borders, depends. 
 
 The pakwa is an arrangement of long and two short lines — " divining-rods," as 
 they were called by the initiated — in groups of three, as shewn in fig. I. 
 
 Fig l. 
 
s^ 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 Eight variations in grouping are possible, hence the name the " Eight Tri- 
 grams," by which the complete series is known. Much learning has been be- 
 stowed upon its teaching, but for our purpose it is sufficient to note that each group 
 is supposed to represent some power in nature, active or passive — earth, air, fire, 
 water, thunder, and the rest, 
 
 The lines are found in another form, with the ends of each returned, as shewn 
 in Fn>\ '2. 
 
 Fit 
 
 The symbol does not appear in decoration in its group form, one of the 
 elements of the groups with the returned ends being the basis of the diapers. 
 
 The other well-known symbol which enters into these designs is the 
 svastika, or fylfot (tig. 3). In Japanese heraldry it usually takes the form given 
 in fig. 3a. It might be described as two elements of the pakwa groups with the 
 ends returned in opposite directions, placed across one another at right angles. 
 I hesitate however to adopt this description, as it would appear to suggest that 
 it was in fact its origin, whereas its natural evolution from the pakwa diapers 
 will hereafter become apparent. It probably has an intimate relation with the 
 " Wheel of Life"; for, as the eye travels round from one limb to the other, the 
 mind becomes impressed with a sense of rotatory motion. 
 
 Fig. ::. 
 
 Now, if we take a pair ofthe pakwa elements with the returned 
 set them back to back, as in Fig. 4, and then place two more 
 back-to-back pairs underneath them, so that the returned 
 ends of the top limbs of these second pairs lock with those 
 of the bottom limbs of the first pair, and repeat this regularly 
 over the whole surface, as in Fig. 5, we get an elementary 
 design very much used in the East, which is the pakwa 
 diaper in its simplest form. 
 
 ends, and 
 
 Fig 4. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 89 
 
 Fig. 5. 
 
 In order to obtain a symmetrically locked design, the length of the returned 
 ends must be the same as the thickness of the limb itself, thus making a small 
 square. In the following descriptions of the illustrations this square will be 
 taken as the unit of measurement. The length of the limb in this instance is 5 
 units, and the breadth of the intervening space between all the members of the 
 design one third of a unit. 
 
 The next development consisted of arranging 
 the pairs of limbs " half-back-to-back," as in Fig. 
 6; then, when other half-back-to-back pairs of ele- 
 ments are locked with the first pair, as in the previ- 
 ous case, over the whole surface, we get the design, 
 Fig. 7 ; also a familiar one, which is the second of 
 the pakwa diapers. 
 
 Fiij. fi. 
 
 Fig. 7. 
 
00 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 These two diapers contain the two main constructive 
 whole series is based. We shall presently find the 
 pakwa clement extended and much elaborated, result- 
 ing in most complicated diapers; hut however com- 
 plicated it may he, the principle on which the design 
 is constructed will he found to he invariably, a pair 
 of the elements set either full hack-to-hack, or half- 
 hack-to-hack. All these diapers, therefore, fall into 
 two classes, each being based on one of these princi- 
 ples, and each class will be found to have character- 
 istics peculiar to itself, resulting from its constructive 
 principle. 
 
 I have come across one diaper, given in the Di- 
 aper plates (lig. 8 plate 5), in which the simple 
 pakwa element is arranged in pairs set at right angles 
 to one another, as in Fig. 8, giving the diaper, Fig. ( ->. 
 
 principles on which the 
 
 Fig. 8. 
 
 Fig. !). 
 
STUDIES TN THE DECOR ATTYE ART OF JAPAN 
 
 91 
 
 Two points about the pair of elements in fig. 8 must be noticed. The limb 
 has been lengthened to allow of the interlocking; of the returned ends of the 
 pairs : in this case to 9 units. And the intervening space between the elements 
 lias become an important feature of the design. It is now increased to a full unit ; 
 ami henceforth it always retains this dimension, instead of being one third of a 
 unit as in the first two diapers. 
 
 This diaper (fig. 9) holds an important position in the development of the 
 series. It will be remembered that in one class of lattices the spaces between the 
 lattice-work themselves formed a substantive design, the woodwork appearing to 
 vinish, the eye resting only on the series of pleasantly shaped spaces. Now, in 
 this diaper, the spaces between the limbs give the figure, fig. 10; or taking the 
 half of it, fig. 10a. 
 
 Fig. 10n. 
 
 Fig. 10. 
 
 This figure, 10a, is the pahwa element with a central stem added ; and this 
 stem is retained in all the subsequent variations of it, the figure itself becoming 
 the basis on which all the diapers are developed. Two of these figures are set 
 back-to-back, as in the elementary diapers, and this becomes the unit of design 
 in the future. The reason for taking the half instead of the whole figure (fig. 
 10), will appear as the argument proceeds. 
 
 In order to illustrate what I mean by saying that the spaces between the 
 limbs take their share in the general effect of the design, I have made the 
 following transformation (fig. 11) of figure 9, with the spaces black, and leaving 
 the pakwa elements white. 
 
'.)'_> 
 
 STUDIES IX THE DECORATIVE ART OF JAPAN 
 
 Fig. 11. 
 
 The effect of this, looked at as a mere question of black ;m<l white, or light 
 and shade, it most unsatisfactory ; it throws tire pairs of pakwa elements into 
 high relief, and the predominant black gives no determinate design, but only a 
 confused and meaningless black mass at the meeting points of the ends of the 
 white elements. It has the effect of fret-saw work, 
 and lacks altogether that appreciation of the due 
 proportion between light and shade which is charac- 
 teristic of Japanese art, and of which I have already 
 talked sufficiently. From this stage onwards, the 
 diapers, unlike the lattices, are composed of an equal 
 amount of light and dark ; but this point forms part 
 of the previous discussion. 
 
 We now come to the first of the diapers based 
 on the new form of the element, the limb with the 
 central stem, fig. 10a. It is constructed by pairs 
 arranged back-to-back, with an intervening space of 
 one unit, (fig. L2.) 
 
STUDIES FN THE DECOR ATTYE ART OF JAPAN 
 
 03 
 
 Fig. 13 
 
 In this diaper, fig. K>, the space-shape is identical with the shape of the 
 elemental figure. It is probably the original of what we know as the " Chinese 
 puzzle," as will he seen if pieces of differently coloured cardboard are cut to the 
 shape of fig. 10, and fitted together at right angles to one another. The length 
 of the pakwa element is 7 units, of the returned ends 1 unit, and of the connect- 
 ing limb 7 units, to correspond with the base of the complementary pakwa 
 element (the space-figure) which fits into it. 
 
 This rectangular setting of the pieces of the Chinese puzzle has no further 
 relation with the principle of the diaper, fig. 9 ; for here the space-figures are set 
 at right angles, and not the pairs of elements. The whole design, spaces and 
 elements, fits together. 
 
 In the next development we revert to the interlocking principle on which 
 the earliest diapers were based. The length of the element in fig. 13 does not 
 permit this, and the points of the returned ends only touch. In order, therefore, 
 to allow the elements of each pair which face one another to lock, they are 
 
«u 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 extended by two units, making ( .» units in all, and the returned ends are pro- 
 longed to 2 units, as in fig. 14, which produces the diaper, fig. 15. 
 
 The characteristics of this diaper are, first, that it differs essentially from 
 the Chinese puzzle principle, the pairs of elements standing out prominently as 
 the main feature of the design, and the space no longer produces an identical 
 shape. Secondly, this space runs in and out of the elements in a continuous 
 fillet. In view of what follows hereafter in connexion with the production of 
 the " Key-border," this point must be specially borne in mind. 
 
 In the next development a new idea is introduced 
 into the form of the pakwa element ; the returned end 
 is moved inwards by one unit, as in fig. 10, but the 
 length of the element is as in fig. 13, 7 units. There 
 is no interlocking, the diaper belonging to the 
 " Chinese-puzzle " class. 
 
 The adoption of this new feature in the shape of 
 the element appeare at first sight to be arbitrary. 
 The result of it, however, is very important, for it 
 gives a new resultant figure at the point where the 
 returned ends (though, owing to their new position, 
 they are " ends " no longer) meet, and produces a Fig. 14. 
 
 Fig. 15. 
 
 more dazzling effect than the simpler arrangements. Further, this resultant figure 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 95 
 
 becomes a special feature of the diaper, of great importance in the subsequent 
 developments, as it contains the germ of what will presently appear. 
 
 Fig. 17. 
 
 I pass now to one of the most complicated of the pa/cwa diapers, which is 
 also one of the Chinese-puzzle class, the space-shape being identical with the 
 elemental design, and there being no interlocking of the elements. 
 
 The shape of the element is given in fig. 18 ; its form appears at first sight 
 to be fantastic, no special reason being at first sight discoverable for the introduc- 
 tion of the " crook." But if we look a little deeper we find that if the opposite 
 crooks of the pairs are joined they make the simple pahva element; so that this new 
 unit of design has been produced by laying two single pakwa elements across the 
 back-to-back pair of elements with the central stem, the length of these elements 
 being increased to 11 units, in order to admit of this being done. 
 
 I have given one complete section of this diaper in outline, and it will be 
 seen that after this long process of development we have arrived at the svastika 
 in outline, which I have thickened for emphasis. There are four of them in the 
 section. 
 
9(V 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 It should have been noted before, that although 
 for convenience of demonstration, I have taken the 
 pakwa elements as solid, the diapers are often found 
 in brocades, and sometimes in ordinary decoration, 
 as they are used by the enamellers, in outline. 
 
 From this sudden and unexpected appearance 
 of the svastika some things may he inferred. First, 
 the gradual evolution of the diaper from its primi- 
 tive form to this, one of its most complicated 
 forms, shows that it was not constructed from the 
 svastika. Secondly, that this diaper could not have 
 been devised for the purpose of introducing this 
 second symbol ; for it is clear, not only that its 
 
 Fig. is. 
 
 Fig. li». 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 97 
 
 appearance is fortuitous, but also that no such ornate and complicated design 
 could have been originated at first hand, and this quite apart from the 
 fact that the evidence is conclusive that it did not. We get therefore to 
 this, that there has been a gradual evolution of designs based on the pakwa 
 element, along the lines indicated, and that at this stage the svastika appears. 
 Or, putting this more accurately, at this point in the evolution of the 
 diapers, a new and rather fascinating figure appeared, which having attracted 
 the attention of the devout, was taken out of the diaper, and accorded a 
 special symbolic meaning of its own. Possibly the svastika has been imagined 
 before its spontaneous appearance in the diaper, that this was, so to speak, its 
 second incarnation ; but the suggestion which I have made that this is in fact 
 its origin has somewhat stubborn facts to support it. 
 
 Is there anything in this opinion in conflict with the conclusions come to 
 by the many learned authors who have collected materials on the subject ? 
 Really, I do not not think so. The missionary fathers have called it a croix 
 gammee, but find in it no relation to the true cross. Indeed the teaching of 
 their predecessors, of the time when the challice of Ignatius Loyala brought com- 
 fort to the Christian martyrs of Japan, must have been that the cross was a truer 
 symbol to hold faith by than the svastika or wheel of life. 
 
 Other authors, like Dr. Robert Ferrers, assert its use in all ages and in all 
 lands : connect it with Indian fire-worship : trace it as a mystic evil-dispelling 
 ornament throughout the whole of Europe in its primitive state. Mr. Elkins says 
 that it was the monogram of Vishnu and Shiva, is to be found on the battle-axe 
 of Thor in Scandinavian inscriptions, as an ornament in the crowns of the 
 Bonpa deities of Thibet, and a favourite symbol with the Peruvians. This 
 learned author also says that it was the symbol of the esoteric teaching called 
 " the pure secret of the eye of right doctrine," communicated orally by Buddha 
 before his death to his disciple Maha Kashiapa ; and that it was usually placed 
 on the heart of Buddha, whence it came to be called " heart's seal," as containing 
 within it the whole mind of that divinity. 
 
 But all this leaves the creation of the symbol and the artistic developments 
 based on it untouched, and concerns more its subsequent spreading through 
 the world. It throws its origin into a remote past; but the fact of its 
 appearance, which the evolution of the diapers demonstrates, and which to 
 the devotees of a mystic religion would be looked upon almost as a super- 
 natural creation, merely sets back the origin of the pakwa symbol to a still 
 more remote period. 
 
98 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 I must now revert to the diapers, and have at once to note yet another 
 spontaneous appearance of the svastika in a somewhat curious fashion. The 
 diaper, fig. 20, has already been given on plate ">, fig. 9, in order to shew its 
 artistic effect. 
 
 1 
 
 Fig. 20. 
 
 Its construction must now be studied. The element of which it is com- 
 posed, shewn under the diaper, is the same as fig. 16 broken into mosaics; and 
 the diaper itself is the same as the diaper fig. 17, but with its lines set at an 
 angle, and the whole of it, both element and space-shape* is in mosaic. The 
 result is that a fillet of space is left running round each fragment. Now this 
 fillet produces, as shewn in the outlined portion of the figure, the svastika at 
 regular intervals, in the same way as it is produced in fig. 19. It is in fact the 
 outline of the figure produced in fig. 17 by the contact of the protuberances of 
 the elements, the special feat n re of that diaper to which I have already called 
 attention. The importance of the idea which suggested the moving inwards of 
 the returned end of the pa hint element as shewn in fig. 16, thus becomes 
 manifest. 
 
 I referred in connexion with fig. 19, to a " complete section" of the diaper 
 as having been given in outline. The meaning of that expression must now be 
 explained. Both in fig. 19 and fig. 20, the svastika symbols are produced alternately 
 face-to-face, both horizontally and vertically. The complete set of svastika 
 figures must therefore be taken to be a square of four, those which are diagonally 
 opposite being alike. This four-fold repetition of the symbol becomes itself a 
 feature in subsequent designs, as will be seen hereafter. 
 
STUDIES IN THE DECORATIVE ART OF JAPAtf 
 
 99 
 
 We may now take stock of the different forms of the elements which have 
 come under our notice. They are Nos. 1, 2, 4, and 6 of the series given in fig. 21. 
 But the introduction of the " crook " into the pakwa element (fig. 18) suggests, 
 from the complicated nature of the new element resulting from it, the existence 
 of intermediate stages in the development of the pakwa element as the unit of 
 design for the diapers. The process has been so gradual, that I think it may be 
 taken for granted that there were at least two intermediate stages with which we 
 are not at present familiar. The following seems to have been the probable 
 sequence, Nos. 3. and 5 being new to us. 
 
 !. (Diapers, tigs. 5, 7 and i') 
 
 4. (Diapers, figs. 16 and 20) 
 
 (Diaper, tig. I3,i 
 
 Fig. 21 
 
 (Diaper tig. 19; 
 
 With regard to No. 3, which is composed of the returned end of the element 
 lengthened and itself returned, it must almost certainly have been considered 
 with a view to use. If it is treated in the same way as the other elements, set 
 back-to-back, and the full figure constituted with the whole of the central limb, 
 it will be seen that it produces only a series of isolated figures, as in fig. 22. T 
 have seen this arrangement used in brocade, and also the corresponding one 
 which would be produced if these figures were set half-back-to-back. There is 
 no continuity of design, and therefore they are not diapers, and hold no place in 
 the present discussion. 
 
100 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN. 
 
 is;. 
 
 No. 5 of the series of elements, in whatever manner it may be treated, 
 produces neither continuous nor independent design. 
 
 I am compelled for the moment to leave the diapers, for actual observation 
 supplies me with no other examples. The result of the study of those which I 
 have actually observed, may be stated thus : there is discoverable among them a 
 regularly developed scries, which depends on a regularly arranged series of pakwa 
 elements; and that it begins, as all good series should begin, with a simple, and 
 ends with a complicated form. 
 
 One important point must here be noted; the series of elements, so far as 
 we have studied them, produces diapers which with one exception, are referable 
 to what T have called the " Chinese puzzle " principle, where, to repeat myself, 
 the space-shape is identical with the elemental design-shape. The exception is 
 fig. 15, with its elongated returned ends ; and it seems manifest that there must 
 be a second and more elaborate series of diapers, constructed of elements of which 
 some or all of the parts are prolonged. I did not come across them in Nikko or 
 Shiba, and must therefore seek them aliunde. 
 
 Now, if I take the second of the elementary diapers, fig. 7, where the pairs 
 of pakwa elements are arranged half-back-to-back, and fill in the continuous line 
 which runs round the elements at right angles to the general arrangement of the 
 diaper, I get at once the simple " Key." This is shewn in fig. 23, where the 
 diaper is printed in grey, and the " key " in black, in order to make this re- 
 
STUDIES 7,V THE DECORATIVE ART OF JAPAN. 
 
 ioi 
 
 markable and probably, to my readers, unexpected appearance of what they 
 know as the " Greek key' plainer. In due course I shall explain how it was 
 that this tansverse section of the diaper came to be taken for the production of 
 the " key "-border. 
 
 I next treat the first of the elementary diapers, fig. 5, in similar fashion. 
 The result, shewn in Hg. 24, is not a regular " key "-border ; but it is a well- 
 known variant, sometimes called the "double key." 
 
 Pursuing the idea of taking a transverse section of the diapers, we find that 
 fig. 15 so treated produces the same double "key," (fig. 25), with the addition of 
 
 a central limb. 
 
102 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 Fig. 25. 
 
 From these results a conclusion, apparently obvious, seems in waiting to be 
 drawn — that a transverse section of any diaper will produce a " key "-border 
 corresponding to it.* But a brief study of the other diapers figured in the 
 preceding pages will be sufficient to dispel the illusion that we have arrived at 
 the end of the enquiry, for some of them will not produce a " key " of any sort. 
 We have really got no further than this : that a transverse section of some dia- 
 pers produces a " key," and of some others it produces a variant. Yet this does 
 leave the suggestion a probable one, that all " keys " are produced from diapers, 
 so long as they are constructed on the same principle as fig. 7 : that is to say, 
 diapers in which the component elements are arranged in pairs half-back-to-back. 
 
 A " key " is a continuous fillet ; and this diaper, as well as fig. 5, the other 
 elementary diaper, and also fig. 15, as shewn above, has a continuous line round 
 the elements of which it is composed. But where the " Chinese-puzzle ' princi- 
 ple is the basis of construction, it is clear that there can be no line round the 
 elements, because the elements and the spaces "fit"; and therefore no diaper 
 constructed on this principle will give a " key." 
 
 It has already been noted that the feature which distinguishes fig. If) from 
 the other diapers, the prolongation of the returned end of the pakwa element to 
 2 units, plays an important part in the production of this continuous line. It 
 seems clear, therefore, that if "keys " are always produced from diapers, these 
 diapers must be constructed from elements in which some or all of the parts are 
 prolonged. 
 
 This much said, I propose now to take the first step in the discovery of the 
 series of diapers with which as yet we are not familiar. In doing so I shall 
 assume nothing, but merely avail myself of the suggestion which the foregoing 
 discussion makes inevitable, and turn to the " key "-borders themselves for 
 light. 
 
 Although it is conceded that the eastern and the Greek "keys" are the 
 same, I take the elementary " key " (fig. 26) from the Greek examples given by 
 
 1 For the benefit of those who have studied the paper which I read before the .J;i]>:ui Society, referred to at 
 the beginning of this chapter, 1 ought to confess that this conclusion was in fact drawn in it. Subsequent con- 
 sideration has however shewn that it is only partially true. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 io3 
 
 Owen Jones (plate XV, No. 1). It makes this point clear, that in the Greek 
 " keys," the fillet, or meander, was the same breadth as the elemental " crooks ' 
 which make it. 
 
 Fig. 27. 
 
 Fig. 26. 
 
 This border is composed of a continuous series of units of design in the 
 form of fig. 27 ; and the first thing that strikes one about this unit 
 is its obviously intimate relation with the now familiar palcwa 
 element; that the black " crooks " are apparently half the element, 
 with the returned end prolonged to 2 units. Further, these 
 "crooks " or half elements, interlock in the same way as the elements 
 themselves in the diaper, fig. 15, with a space of 1 unit round them, 
 which makes, as it did in that diaper, the continuous fillet. 
 
 One difference will be noted between fig. 26 and the " key ' produced by 
 the elementary diaper in fig. 23 ; there is a bar both at the top and bottom of 
 the border, from which the " crooks ' spring. It would seem, though for the 
 moment I cannot assume it, as if these bars or borders stand in the same relation 
 to " crooks " or half elements as the central stem (fig. 10a) does to the whole 
 element from which the diapers are constructed. This remains to be 
 proved. 
 
 I have proceeded in my argument so cautiously up to the present, that I 
 take the next step without hesitation. I prolong the lines of fig. 26, and repeat 
 the figure on the other side of the bar, which is shaded to make the procedure 
 clearer. This gives us fig. 28. 
 
 Fig. 28. 
 
104 
 
 STUDIES /.V THE DECOR ATIYE ART OF JAPAN 
 
 This is actually among the " key "-borders given by 
 Owen Jones (pi. XV, No. 17), and is said to have come from 
 the coffers of the ceiling of the Propyhea at Athens. The 
 learned author describes it as two " keys ' placed one above 
 the other, running in opposite directions. There is, I think, 
 another explanation of it, which gives us the clue to the 
 construction of these borders. 
 
 There is running through it a series of regular and re- 
 current figures making together a fish-bone, resembling fig. 
 2i), into which the reversed "crooks" are locked; the result, 
 two " key " fillets running in opposite directions, is inevitable, 
 bearing in mind the fact that the spaces and the "crooks" 
 are the same breadth. It seems obvious that the " fishbone " 
 is the cause and not the consequence. 
 
 At this point yet another conclusion seems to be in 
 waiting, wliich, if it were legitimate to draw it, would bring Fig. '29. 
 
 the argument to a satisfactory termination. For if the process of prolonging the 
 lines be applied to both sides of the border, we should get to a diaper; which 
 again seems to shew the intimate relation between the " keys ' and the 
 diapers. I must, however, proceed by a more circuitous method of reasoning. 
 
 The " fish-bone " seems to have been purposely constructed from a series of 
 entire pakwa elements with the central limb, placed one on top of the other; and 
 if this is so, and if other "keys" can be constructed on this principle, the true 
 
 nirn 
 
 Tinmr: 
 
 Fig. 30. 
 
 Fie. 31. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 105 
 
 inwardness of this figure will become apparent : it establishes the relation between 
 the pakwa and the " key "-borders. 
 
 In order to test this I revert once more to the elementary diaper, fig. 7, and 
 the " key " produced from it, as shewn in fig. 23. The narrow fillet between the 
 elements makes comparison with fig. 26 difficult. I have therefore expanded this 
 diaper so as to give the space its due proportion of one unit ; and for this pur- 
 pose, as explained in connexion with fig. 15, the returned end has been length- 
 ened to 2 units. This gives us fig. 30; and treating fig. 23 in a similar manner, 
 we get fig. 31. The essential difference between figs. 31 and 20, is, as already 
 explained, the bars or borders in fig. 20, from which the half elements spring. 
 
 The next step is easy, and is based on the previous course of the argument. 
 If we take the element with the central limb, fig. 14, and arrange the pairs half- 
 back-to-back instead of back-to-back as in fig. 15, we get to fig. 32, a diaper, the 
 salient feature of which is a series of " fish-bone " figures set alternately up and 
 down, so that the returned ends (2 units long) lock with a space of 1 unit 
 
 Fig. 32 
 
 between them. The ends of the contiguous elements are 3 units apart, in order 
 
 The transverse section of 
 
 to allow the central limbs to come into the design 
 
10(> 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 this diaper gives us fig. 33, which is the "'key," 26 in its entirety : that is, with 
 the top and bottom bars. Fig. 32 is in fact the diaper which 1ms already been 
 
 suggested as resulting from prolonging the lines of the border, fig. 28, both 
 ways. The suggestion made above, that the bars of the " key '-border in its 
 common form hold the same relation to the half elements as the central stem 
 does to the whole element in the diapers is, therefore, borne out. We must 
 now see whether what has been done in this simple case produces the same 
 results in the case of other " keys." 
 
 Fig. 34 is given as a Greek " key " by Owen Jones (pi. XV, No. 2), and 
 he calls it the " raking fret," holding it to be the parent of all the other forms 
 of interlacing ornaments in styles, such as the Arabian, which succeeded the 
 Greek. He attributes to its influence many of the designs found in the 
 Alhambra. 
 
 Fig. .",4. 
 
 I have a treated it in the same way as fig. 26, and as before, the "fish-bone" 
 arrangement of the pakwa, but set at an angle, appears. The returned ends of 
 the elements are lengthened to 3 units. The inclination of the lines merely 
 shews that the Greeks, like the Chinese occasionally sought variety from the 
 normal rectilinear arrangement. We have already had 2111 instance of this prac- 
 tice in the East in the diaper, fig. 22. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 107 
 
 The prolongation of the returned ends may be carried to any extent, ;is in 
 the following example, fig. 35, where the return is 6 units, and the length of the 
 element only 9 units. 
 
 As a principle of construction of " keys ' without the necessity of making 
 the entire diaper, the "fish-bone" arrangement is simple where the pakwa element 
 is used in its elementary form; but it is necessary to see whether it is applicable 
 to more elaborate forms. 
 
 Fig. 36 is another Greek " key " given by Owen Jones (pi. XV, No. 9). 
 The prolonged lines shew that it is based on a " fish-bone ' arrangement of an 
 element, which is a development of No. 3 in the series of pakwa elements given 
 in fig. 21. The returned ends are 4 units, and the second return 2 units. 
 
 j m m m m r- 
 
 jy 
 
 W 
 
 Fis;. 30, 
 
 We may conclude from this that the " fish-bone ' arrangement is as much 
 the clue to the construction of a more advanced order of " key "-borders as it 
 was in the case of simpler examples, and it seems obvious that a great variety of 
 " keys " can be produced by this very simple means. The process of interlock- 
 ing the ends of the elements is of course the point at which the artist's ingenuity 
 comes into play ; but it is precisely the same ingenuity, exercised in precisely 
 similar circumstances, as that which was required for the construction of the 
 diapers out of the same elements. In proof of which I give fig. 37, which is the 
 diaper fully developed from fig. 36. 
 
108 
 
 STUDTES TN THE DECORATIVE ART OF JAP AN 
 
 Fig. 37 
 
 80 far, therefore, as " keys " are concerned, in which the element is not very 
 complicated, their genesis j s clear. They are transverse sections of diapers 
 constructed in the same way as lii»\ 32 : the pakwa element having been selected, 
 and the due proportion of its returns determined with a view to interlocking, 
 they are arranged half-hack-to-back like the simple elements in the elementary 
 diaper fig. 7. One little point may he noted here. Pakwa elements without a 
 central stem give " keys ' without bars at top and bottom ; those with the 
 stem give " keys " with these bars. 
 
 The " fish-bone " arrangement having served its purpose, I have discarded 
 it in adopting this method of construction, which is more scientific ; because if 
 the designer starts with the " fish-bone," he has nothing to guide him in deter- 
 mining the length of the central limb, the important factor on which the inter- 
 locking primarily depends. I have purposely drawn the " fish-hone" in fig. 29, 
 in an arbitrary fashion in order to emphasise this difficulty. In making the 
 diaper the real basis of construction this length is determined, as it has a 
 definite relation to the length of the base of the element. 
 
 So far we have heen concerned with the "' keys " which are produced by the 
 half-back-to-back arrangement of the elements, which have earned the name of 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 109 
 
 "true keys." Owen Jones defines a " key ' as a continuously meandering line, 
 all others as imperfect. The full back-to-back arrangement has somewhat 
 dropped out of sight ; but that also, as we have seen, produces a continuous line, 
 and the section of the diaper a border well worthy of attention. The diaper fig. 15 
 has given us the border fig, 25, and it must be obvious that a variety of designs 
 of great charm may be produced from such diapers, so long as they do not drop 
 into the " Chinese-puzzle " principle which produces no borders. 
 
 The next design, fig. 38, is one of these borders, winch I found at Nikko. I 
 have continued the lines in the same way as in the case of the Greek frets, and it 
 will be seen that the unit of design is a pair of simple pakwa elements, without 
 the central limb, set back-to-back, the returned ends being 4 units, and the second 
 return 1 unit. In order to make this a little clearer, and to enable the com- 
 parison to be made, I have in fig. 38a transposed the black and white parts of the 
 
 0SSM5SSSS 
 
 Fi K . 38 
 
 Fig 38a 
 
 drawing. It gives us the elementary diaper fig. 5, elaborated ; and the fillet is 
 the same as that given in fig. 25, as resulting from the diaper fig. 15. But this 
 is not a continuous meander, for it is broken by tbe vertical lines, which disturb 
 its effect. Indeed, if we turn once more to fig. 24, it will be seen that the black 
 line is not a true meander, for it passes over and omits the vertical white lines. 
 Fillets broken in this way may be taken to be the feature of frets produced from 
 diapers constructed on the back-to-back principle. On the artistic virtue of the 
 uninterrupted meandering line I shall dwell presently. 
 
 But there are undoubtedly many designs which may be described as " imper- 
 fect ' " key "-borders ; and I should define them to be those which are constructed 
 
110 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 upon .sonic principle of their own, and arc not based on a regular diaper. Fig. 39, 
 which is one of Owen Jones' Greek frets, (pi. XV, No. 22) is a good illustration. 
 
 Fig. 39. 
 
 The introduction of an imperfect fret here will give point to the argument. 
 The definition shall be justified at a latter stage when I come to examine the 
 Greek frets more closely ; but it is well that those who are familiar with this 
 class of design should appreciate at once the fact that there is a right and a wrong. 
 The defects in fig. 39 will, I think, he readily appreciated. The repetition of the 
 principal motive at different levels is irritating; although the great essential, a 
 repeated thought is present, the design as a whole lacks continuity of thought. 
 The construction is almost too ingenious ; nothing rests in the mind hut a con- 
 fused maze of lines. There is in fact, if you care to worry it out, a continuous 
 meander, hut it is impossible for the mind to grasp it, to use a figurative expres- 
 sion, it is perpetually being switched oil' the rails. Above all, prolongation of 
 the lines would result in nothing more than a repetition of the maze ; would 
 produce nothing which, by any stretch of the imagination, could he called a 
 diaper. These and other points will he dealt with fully in due course; for the 
 present we must revert to our study of the true "keys." 
 
 I now come to two very complicated borders; the first, fig. 40, from Nikko, 
 the second, fig. 41, based on one of Owen Jones' Greek frets, (pi. XV, No. 12.) 
 
 Fig 4i). 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 I rani rani j? 
 
 ^rni^miHl 
 
 111 
 
 Fig. 41. 
 
 These two designs are examples of the " key "-border in its highest stage of 
 development. It is obvious that although they come from different sources, they 
 belong to the scheme of decoration, could not have been produced independently 
 of one another, and that the second carries on the idea of the first to a higher 
 pitch of perfection. That idea is the emphatic repetition of the svastika, which 
 is produced by the lines of design in the first, and by the space-shape in the 
 second. In order to discover how these two borders came into existence, I have 
 prolonged the lines as in the previous cases, with the following results. 
 
 Fig. 40 developes into fig. 42 ; the black and white having been transposed 
 as in a previous case, to facilitate comparison. 
 
 Fig. 42. 
 
112 
 
 STUDIES TN THE DECORATIVE ART OF JAPAN 
 
 The fi rst point to be noted about this figure is that its formal regularity 
 proclaims it to be a diaper constructed out of pakwa elements, with regularly 
 recurrent features resembling in their nature the diapers we have already studied. 
 The element used is the one with the returned end moved in (No. 4 of fig. 21) ; 
 but the return is 2 units long, and it is moved in 2 units from the end; the 
 length of the element is 13 units, and of the central limb 9 units. The second 
 point is that the " fish-bone " arrangement of this element is as marked as it was 
 in the ease of the simple " keys " when developed. But another feature has been 
 introduced, indicated by shading ; and this new feature has an evident affinity 
 with the pakwa element on which the design is based. It might be described as 
 half the principal element, for it also has a "return," which is moved in from 
 the end by 2 units. 
 
 The border fig. 41, treated in the same way produces the diaper fig. 43, the 
 constructive principles upon which it is based being identical with those of fig. 42. 
 
 Fig. 43. 
 
STUDIES IN THE DECORATIVE ART OF JAPAlSl 113 
 
 The "fish-bone" (printed black throughout) is constructed from the elabo- 
 rate development of the pakiva element, No. 6 of fig. 21, on which the diaper fig. 
 20 is based ; but it is very much extended to permit the parts to interlock. The 
 base of the element is 21 units, the returned ends 2, the "crook" 4, and its 
 return 2 units: the length of the central limb is 13 units. The secondary figure 
 in this diaper has the same affinity with the principal element as it has in fig. 
 42; for it also has a "crook" and a returned end, and is, as before, half of it. To 
 those secondary figures we must now turn our attention. 
 
 When a diaper is constructed on the half-back-to-back principle with an 
 element which has a central limb, a large space is left between the ends of the 
 contiguous elements as already explained in connexion with fig. 32. In that 
 case the space was 3 units; in both of these developed diapers, figs. 42 and 43, 
 it is 7 units. The same thing occurs in the back-to-back diapers, where the 
 central limb also passes between two elements. In the case of the diapers we are 
 now considering, it is hardly necessary to explain that as the space between the 
 halves of the elements which face one another must partake of the character of 
 the elements, so the figure which is used to fill up these spaces must have 
 the same characteristic features as the elements. They also have an affinity 
 with another figure which we have not yet examined. 
 
 The position which these two diapers hold in the enquiry is very important. 
 First, there is an obvious relationship between them and the two earlier diapers, 
 figs. 17 and 19. In them the svaslika was produced in outline only, on account of 
 the closely fitting spaces; in these it is produced in solid form by the space- 
 designs, which, as we know now, are the direct result of the half-back-to-back 
 arrangement of the elements. 
 
 But secondly, it must be obvious to those who have followed the gradual 
 progress of the argument, that not only do these two borders appear at the end of 
 a series of which we have the beginnings, but that the diapers produced by them 
 are also manifestly at the end of a series of diapers which begins with the one so 
 often referred to, fig. 15. We have in fact found the end of the second and more 
 elaborate series, the existence of which has been suggested on a previous page : 
 a series in which some or all of the parts are prolonged. 
 
 And so far as the " keys " are concerned, the same law which connects the 
 early borders with diapers, and conversely the early diapers with borders, con- 
 nects these two elaborate borders with diapers, and conversely, these two elabo- 
 rate diapers with borders. The intimate relationship between " key-borders " 
 and pakwi diapers, which I have suggested throughout as being more than 
 possible, is therefore established. 
 
114 
 
 STUDIES L\ THE DECORATIVE ART OF JAPAN 
 
 The secondary figures which have appeared in these diapers have a close 
 affinity with another kind of border, which has been independently developed 
 from the svastika, which we must now proceed to examine. 
 
 THe SVASTIKA BORDERS. 
 
 The Svastika having, as already explained, appeared during the process of 
 the development of the diapers, gave rise of itself to a series of designs, which 
 
 engendered borders often mistaken for " key "-borders. 
 
 The symbol derives its distinctive character from the ends of its limbs beina 
 returned in opposite directions. By prolonging and again rerturning the ends, 
 and repeating this process, the following figure, fig. 14, was obtained. 
 
 Fie. 44. 
 
 The special characteristic of this developed Svastika is that while each 
 section is identical in the contours of its prolongations, each prolongation is at a 
 different angle, is turned through 90 degrees from the one preceding it. This 
 peculiarity lent itself readily to the artistic conceptions of the Japanese designers; 
 for its appearance was as fanciful as its development was formal ; and when the 
 lines were cut with the triangular gouge, they scattered the light as effectively as 
 the pine design itself. It is not to be wondered at, therefore, that it figures 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 115 
 
 largely, either in whole or in part, in the temple diapers. Occasionally it appears 
 as a fillet, the whole being repeated, as in fig. 4o. 
 
 WJMjMJMj 
 
 Fig. 4"> 
 
 Sometimes one symbol with its developed ends form the basis of an entire 
 design. 
 
 Vet again, isolated sections of the figure were arranged in series to form a 
 border, as in fig. 4<>, which comes from the edge of a water tank in the Temple 
 of Tvevasu at Xikko. 
 
 Fig. 4ti 
 
 This fillet differs in many respects from the " key "-borders. It is an 
 arrangement of independent units of design ; there is therefore no continuity of 
 
 Fij;. 4' 
 
 Fig. 48 
 
 JlE][s|E]|sL 
 
 1 — ' — j | — 
 
 S]|[afEl||7 
 
no 
 
 STUDIES fN THE DECORATIYE ART OF JAPAN 
 
 line, nor any meandering fillet ; and in this instance even the base line is omittted, 
 so that these units spring from nothing. It is, however, exceedingly effective. 
 
 From this sprang the "S "-like figure of which the next two borders are 
 composed, figs. 47 and 48. Its connexion with the developed svaslika is fairly 
 obvious, and it originated in a very simple manner. 
 
 We are now so familiar with the ways of the eastern designer, that it will 
 cause no surprise to find him obtaining a little variety in his borders by arrang- 
 ing the svaslika elements alternately up-side down, as in fig. 49. The " S "-like 
 figure is the intervening space between them, and is of course the black design 
 of figs. 17 and 48. The bars which 1 have introduced are taken from fig. 48. 
 
 Fig 49 
 
 The next development was inevitable. These " S " figures were evidently 
 made to be interlocked ; and so, pruning them of some of their involutions, they 
 were interlocked and produced the borders given in figs. 50 and 50a. 
 
 Fig. 50 is given by Owen Jones among the Greek '■ keys " (No. 7 pi. XV) ; 
 but it does not fulfil his own definition, that a " key ' is a continuous meander; 
 
 Fig. 5(i 
 
 M@MS 
 
 Fig. 50a 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 117 
 
 for the units of- design are " S ' figures, and the space (Joes not meander, but 
 merges in the white hands above and below the design. 
 
 Fig. 50a which I found on the walls of the Arena at Aries, is of course the 
 same design ; but the " S " figure is in white and formed by the spaces, and the 
 second return is 3 units instead of 2. I have given the two drawings of the 
 same design in order to shew the different effects produced according as the 
 principal feature of the design is treated in black or white, as design or space. 
 It also emphasises the point just taken, that this fillet does not produce a genuine 
 meandering line. 
 
 These borders have a superficial resemblance to the '• key "-borders ; but 
 this arises from the affinity between the svastika and the pakiva. Their origin 
 and development shew that they belong to a different order of design, and there- 
 fore cannot be called imperfect " keys." 
 
 One further point in connexion with the svastika must be noticed. Its 
 own complete form, when its returned limbs are prolonged as in fig, 44, suggested 
 larger units of design, composed of four sections, corresponding with the four 
 outlined sections of the diaper, fig. 20 ; that is to say, with the diagonally opposite 
 figures corresponding. Traces of this enlarged figure will be seen in some of the 
 diapers on plates 4 to 8. It also appears in one of the Greek designs given by Owen 
 Jones, (No. If), pi. XV), which I reproduce, fig. 51. 
 
 Jig 
 
 Hour 
 
 Fie. 51 
 
 The top right and lower left corners of this figure linked by the central 
 connecting line, shews it obvious relationship with the svastika. The design is 
 
118 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 elaborated by the interlacing of the lines of other similar figures ; and were it 
 not for tiie small squares in the remaining sections, about which I shall speak 
 presently, it would rank among the best of its class. 
 
 There seems to be little doubt that there is an affinity between the secondary 
 figures in the diapers 42 and 4'A and this " S " figure, for they also are " S " in 
 form. These secondary figures take their shape, as we have seen, from that of 
 the spaces left unfilled by the pakwa elements of which the " fish-bone " is com- 
 posed ; but I think this much may be said, that the existence of the " S " as a 
 recognised form of design gave a certain warrant for its use in these diapers, 
 quite apart from the fact that it was identical with the form of the element itself, 
 and was able to take on its embellishments. The main point is that these con- 
 siderations shew that these secondary figures are not hap-ha/.ard, resulting 
 merely from the desire to fill an awkwardly-shaped space. But this further 
 remark seems also permissible ; that they are a definite feature in the develop- 
 ment of this form of decoration, and point to an all-pervading scheme which 
 links all the parts of the systcM together. 
 
 The results of the enquiry may now he briefly summarised. 
 
 The complexity of the " key " designs puts spontaneous evolution out of the 
 question ; the obvious relation of the " key " with the pakwa symbol leaves no 
 room for gradual evolution from any other figure; and lastly, the dependence of 
 the " key "-borders on the pakwa diapers as the means of construction is so 
 absolute as to leave no possibility of there being any other. 
 
 I must, however, notice a theory which seems to have attracted the attention 
 of students of the subject. During the discussion which followed the reading of 
 my paper before the Japan Society, Mr. Gowland, a member well-known for 
 his erudition, made the following remarks : — 
 
 " The generally accepted explanation of the origin of these patterns -which was Hist pro- 
 pounded by General Pitt-Rivers — is, that fchev were developed from simple concentric circles, 
 which by being subsequently united by a curved line gradually passed into a series of re- 
 entering spirals, and that from these, by the exigencies of ornament in later times, the 
 rectangular forms were evolved. The fact that some of the oldest bronzes id' .Japan — earlier 
 than the 6th century A.C.— are ornamented with these spirals supports General Pitt-Rivers, 
 contention that the circular form preceded the rectangular. There 13 however, no proof 
 that the rectangular forms were developed in this manner in Japan, as no patterns showing 
 the necessary transitional stages have been observed in the country. Possibly they may lie found 
 in China, and, in fact, all the evidence we possess is in favour of that country rather than 
 Japan having been the original home of the pattern. It is indeed not until many centuries 
 after its use in China that we find it on Japanese objects, and then only on those which 
 have been copied from Chinese models." 
 
 Quite apart from the considerations which I have advanced, this theory 
 does not appear to be very satisfactory. Undoubtedly, circular forms may 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 119 
 
 engender rectilinear forms ; but equally they may be engendered by them. There 
 are indeed some well-known curvilinear borders which seem to have an affinity 
 with the " key "-borders ; but, as Mr. Gowland pointed out, the theory of evolu- 
 tion one from the other carries us no distance at all, unless we have the his- 
 torical fact to start from, which of the forms preceded the other in first creation. 
 ] int again, the complicated forms of the " key," looked at as a mere border, and 
 their evident development one from the other in an ordered sequence, do not at 
 all fit in with the theory of evolution from some circular design. Even granted 
 that there is some circular ancestor of the "key' in its simpler forms, it has 
 not, 1 think, ever been advanced that there are circular forms corresponding to, 
 or which could have been the direct parents of, the complicated " keys " ; there- 
 fore we should still have to find the "key' whence these complicated borders 
 came. So that even accepting the circular genesis, we are left at precisely the 
 same point at which we have arrived, that there is a regular and orderly sequence 
 of development traceable from the " key " in its simple form, down to its most 
 elaborate example which has just been studied. 
 
 But I have still to explain and justify the use of the word " border " in its 
 application to the " key " patterns. Of course the shape of them suggests their 
 use as borders; but the point I have now to make is that, although after they 
 had been devised they came to be used independently, in their inception they 
 were used as definite borders to the diapers from which they sprang and that 
 herein lies the secret of their origin. 
 
 We may imagine the artist, designer, his wall covered with a beautiful 
 diaper, vet finding something wanting : something not altogether satisfactory in 
 the abrupt ending to the design at the top of the wall. The idea of a border 
 must have occurred very early in the history of the decorative art, as a much 
 less crude way of finishing off the work than merely cutting the diaper short, 
 with its lines left at a loose end. Of course a border does precisely the 
 same thing, so that apparently there was nothing to be gained by its 
 
 introduction. 
 
 But even such trifles as art-borders, like everything else in the world, have 
 made for themselves a little rule of conduct for those who design them. There 
 is a border which is right, and a border which is wrong; and there can be little 
 doubt that the right border is one which has a definite relation to the design it 
 borders. Otherwise it would be an independent design, its designer rather trou- 
 bled to explain the reason for its being where it is. If the design could speak, 
 and, figuratively, all design should speak, it would confess at once that it was out 
 of place. 
 
120 STUDIES IN THE DECORATIVE ART OF JAPAN. 
 
 The Morris designers discovered this rule very early in their career ; their 
 best friezes always had this definite relation to the designs of the wall-papers. 
 If early memories serve me right, this was one of the first of their art 
 discoveries. 
 
 The reason for the rule is not far to seek, if we remember what designing 
 means, and especially designing which has for its aim the covering of large blank 
 surfaces. It is intended to provide a little recreation for the brain, weary of 
 the flat surfaces of mundane things. If the eye constantly sees the same dulness 
 wherever it turns, it can only impart dull impressions to an already weary brain; 
 but when the blank space is broken up by meandering lines, leading to pleasant 
 spaces, even if they convey no suggestion, of flower or bird, of cloudlet or ripple 
 or stormy sea, still the eye delights in its wanderings, the brain finds pleasure in 
 line for beautiful line's sake. Looking at the diapers merely as an arrangement 
 of lines set in a beautiful maze, we have seen how this effect is enhanced by the 
 play of light and shade, the sun glinting from the high relief, the clouds giving 
 depth to the shadows. But when all this comes to an abrupt termination, when, 
 the glitter is cut off by that most detestable thing a" hard line," dissatisfaction 
 at once supervenes. But if then the eye lights upon a border in the same order 
 of ideas as the decoration of the surface, pleasant surprise takes the place of dis- 
 satisfaction ; the eye follows it in its wandering, the brain is beguiled from its 
 evil humour, and suffering from no sharp shock, continues its train of artistic 
 thought which the design itself has set going. 
 
 But, says the critic, did this pleasant little theory suggest itself to the 
 eastern artists centuries ago ? I answer that it is probable, for they were our 
 masters in design. Moreover, I have a witness ready to give evidence ii: my 
 behalf; a witness, it is true, of quite modern times, for the artist dwelt not more 
 than three hundred years ago. But Japan is a land of tradition ; and we may 
 be sure that he held by his traditions faithfully when he devised that star-like 
 border which I have figured on plate (i , among the diapers. For it is, as I 
 have said, no more than a diagonal section of the pine tree design ; and I have 
 in my mind's eye the very wall on which this diaper glistened in red lacquer, 
 with the star-like border to finish it. 
 
 This question of finishing off a design is really a very troublesome one, 
 needing much discrimination in dealing with it, and many fail most lamentably, 
 as a certain long-forgotten artist did many centuries ago at Aries, when the 
 Roman colonists were building their Arena. On a stone border above a window 
 I found this device, cut as roughly as I have drawn it in fig. 52. 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 121 
 
 Fie. 52 
 
 What a melancholy story of incapacity and indecision these lines tell ; 
 what a dull brain the sculptor must have had to finish off his design in so crude 
 a fashion. The design itself is bad enough, but his instinct told him that an 
 abrupt ending was to be avoided ; and those grotesque figures at the ends were 
 all he could devise. But perhaps it was the work of prentice hands! The 
 master-sculptor must have passed it. And Arlesium was no such mean city in 
 the days of Constantine, that it should have the work of prentices foisted on it. 
 
 And now, what of the Greeks? What of the claim advanced on their 
 behalf that thev were the originators of these famous borders? Are thev truly 
 called " Greek keys"? 
 
 Coming now to the end of the discussion, I find myself sceptical as to the 
 soundness of even my own remarks on the subject, made at the beginning of 
 what has been for me, as well as for those of my readers who have been patient 
 enough to travel it with me, a voyage of discovery. In this one especially I 
 now entirely disbelieve : — that the whole system of borders went from China 
 into Greece, and not some isolated frets from which the eastern system was re- 
 constructed by the Grecian decorators. 
 
 The hypothesis throughout has been that the" key ,: systems of the East 
 and West are identical, in spite of the fact that there seems to be no trace of 
 the eastern diapers in Greece. It is indeed this hypothesis that has given rise 
 to the little calumny, as I think it, that China borrowed its ft key "-borders from 
 the West. There seems to be every reason to believe that isolated fragments 
 only, and not the complete system travelled to Greece ; and that, contrary to 
 what I have supposed the system was never completely reconstructed in its 
 original purity. The internal evidence which the Greek " keys " themselves 
 furnish, seems to indicate that while some notions of the value of the Eastern 
 
122 STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 " key " for decorative purposes were conveyed to the Grecian artists, they knew 
 nothing of its symbolism, and hence there grew out of what they received from the 
 East a very impure series of designs, in which the germ of the original was 
 almost lost. 
 
 I shall admit at once that the case for the Greeks may possibly he put much 
 higher than I am able to put it; for my references to Greek design are based on 
 one author alone. But Owen Jones was a keen admirer of Grecian art, a believer 
 in all its virtues ; and he devoted some attention to the " Greek key." There 
 may be many more examples of it to be found in other books, many more 
 varieties of its development than he selected. But he drew from the best sources ; 
 and, from personal observation, it seems to me probable that his examples are 
 among the best. But on the admissions the case for the Greeks is gone. For the 
 borders actually given by Owen Jones on his plate XV, whether best or worst, 
 did in fact exist, and I am prepared to rest my conclusions on this. 
 
 Here is a summary of the author's appreciation of Greek art. It does not 
 want for enthusiasm. 
 
 Egyptian ornament was derived direct from natural inspiration, but was 
 founded on few types, remaining unchanged, except in execution, during the 
 whole course of the country's civilisation. Assyrian ornament was a borrowed 
 type, possessing none of the characteristics of original inspiration, appearing 
 rather to have been suggested by the ait of Egypt when it had begun to decline, 
 and to have carried the decline still further. But Greek art, though borrowed 
 partly from the Egyptian and partly from the Assyrian, was the development of 
 an old idea in a new 7 direction ; and unrestrained by religious laws, as both the 
 Egyptian and Assyrian arts were, it " rose rapidly to a high state of perfection 
 from which it was able to give forth the elements of future greatness to other 
 styles. It carried the perfection of pure form to a point which has never since 
 been reached, and from the very abundant remains we have of Greek ornament, 
 we must believe the presence of refined taste was almost universal, and that the 
 land was overflowing with artists whose hands and minds were so trained as to 
 enable them to execute these beautiful ornaments with unerring truth." 
 
 I note, in passing, that in similar exuberance of appreciative language have 
 many of us written of the arts and artists of Japan. But the important part 
 of the admission lies in this statement ; and the skilful advocate might legitimate- 
 ly say that no better examples of the " key "-border can be found than those 
 which ligure on Owen Jones' plate XV. But there is more. 
 
 " Greek ornament was wanting, however, in one of the great charms which 
 should always accompany ornament, viz. symbolism. It was meaningless, purely 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 123 
 
 decorative, never representative, and can hardly be said to be constructive ; for 
 the various members of a Greek monument rather present surfaces exquisitely 
 designed to receive ornament, which they did, at first painted, and in later times 
 both carved and painted." 
 
 May I not say that here is an admission which, if I am right in my own 
 premises, destroys all claim on the part of the Greeks to have been the originators 
 of the " key "-borders. They are symbolic; were based on the most ancient 
 symbols of the religions of the East. I perforce omit all reference to Gra^co- 
 Buddhisra ; but so far as the religion of the Greeks went, there does not seem to 
 be a trace of the pakwa, nor of the svastika, in any of their ornament, other than 
 is to be found in these borrowed borders. 
 
 But the strongest evidence comes from the frets themselves. I make bold 
 to say that with the exception of No. 1, the simple " key " : No. 2, the " raking " 
 fret: No. 9, a pure " key " : No. 17, the "fish-bone' (the true inwardness of 
 which Owen Jones does not seem to have realised): and No. 7, the " S " fret, 
 the others are of a very low order, and never touch the purity of design which 
 marks the " key "-borders of the East. 
 
 By a pure design I mean one belonging to a certain class which is 
 based on a definite origin, and which preserves that origin clearly discernible 
 through all its developments. But in these borders, even assuming that Grseco- 
 Buddhism did introduce the pa/civa and the svastika into Greece, there is not 
 one of them which shews their direct influence. Such traces of these svmbols 
 as are discoverable are in designs which (as in No. 5, where a section of the 
 svastika is used) are clearly borrowed, for as I shall shew directly, they were 
 used without intelligence. The artists seem to have been impressed by the idea, 
 by the " crook " and its potentialities for decorative purposes ; they constructed 
 clever devices by combinations of " crooks " ; but their cleverness lost itself in 
 efforts to produce the continuous meander which they seemed to consider all- 
 essential. The results are more often than not merely ingenious mazes ; many 
 
 are meaningless, some of them, I venture to think, unspeakably ugly and 
 irritating. 
 
 Take No. 22, (fig. 39 in the preceding pages), which I have already criticis- 
 ed. It is undoubtedly exceedingly ingenious ; but its very ingenuity bewilders 
 the eye; its structure, judged by Eastern canons, merely fantastic. 
 
 Even No. 11 (the upper part of fig. 53), in which the true principles have 
 been followed, is but a feeble adaptation of these ideas. The svastika appears, it 
 is true, as the regularly recurrent feature; but the effect is that it is produced 
 by a fortuitous assemblage of " crooks," invisible hands guiding them to catch 
 
124 
 
 STUDIES /;V THE DECOR ATIYE ART OF JAPAN 
 
 some miserable fly caught in the meshes of the lines. Yet it is based on the 
 same principle as figs. 38 and 39; the " fish-bone " is the constructive principle, 
 and the secondary "S" figure is true. 
 
 Fig .").•; 
 
 I have copied this No. 11 design in fig. 53, and developed it, as in the other 
 cases, to shew how far below them it stands in the scale of excellence. What is 
 wrong with it is fairly clear. The pakwa element is badly chosen, its extensions 
 are unintelligent ; it is ungraceful in its proportions, suggesting nothing better 
 than a series of square "pothooks and hangers." And the secondary " S "-figure 
 also inevitably lacks both grace and proportion. 
 
 Yet No. 12 (fig. 54) shews that the Greek artists had the best model to 
 work from, the diaper given as fig. 43, which is used in alternate sections. This 
 diaper, as I have said, seems to be the last word in complexity of design, and 
 perfection of result. There is a sense of proportion about the component parts; 
 and even when the diaper is worked out in two colours as I have done in fig. 43, 
 
STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 125 
 
 in order to explain its construction, the pakwa element does not unduly force 
 itself on the attention, but merges into the general effect of the design. 
 
 Yet, so lacking were the Greeks in appreciating its beauty and wonderful 
 complexity, that though they used it, they must needs endeavour to improve on it. 
 
 jiii 
 
 Fig. 54 
 
 In Owen Jones' No. 12 it is combined with No. 11, as fig. 54 shews, the 
 large and small resulting svaslikas alternating. Owing to this combination of 
 two different design, the " S ' figure is not uniform, half being the simple 
 "crook " of fig. 53, and the other half having the return and crook, as in fig. 43. 
 Putting it as a mere question of taste, I venture to think that fig. 54 cannot 
 bear comparison with its original, fig. 40. 
 
 There are further abundant evidences in this page of designs that the Greeks 
 never accepted the eastern " keys " at their best. It would appear as if their 
 chief characteristic, repetition, stood to them for monotony. The last example 
 shews this ; but there are worse signs of their hick of knowledge of the meaning 
 of the borders. Subsidiary and irrelevant ornament came to be introduced ; and 
 the " key," instead of being the leading motive of the design, sank into mere 
 bordering for floral ornament. 
 
 The most prevalent variation on the eastern design was the introduction of 
 a small square containing internal squares and an ultimate dot, very much in 
 the manner of the lozenges in the diapers. This square appears in the large 
 four-fold design, based on the svastika, fig. 51 ; and if we take the two upper 
 sections and imagine the lines prolonged in both directions so as to make a 
 border, we get to the one which seems to have been a general favourite among 
 the Greeks. 
 
 Fig. 5^ shews it as I found it at Nimes on the string-course of the Maison 
 Carree, and also among the relics of the carvings of the Temple of Diana. 
 
126 
 
 STUDIES IN THE DECORATIVE ART OF JAPAN. 
 
 Fig. 55 
 
 Distance from Rome, did not make for deterioration of design or careless 
 work ; for the antiquaries are agreed that the very best traditions of hellenic art 
 were embodied in this beautiful little gem, of which Nimes is justly proud, and 
 now guards in pious memory. If some would say that it is absurd to cavil at 
 one minute detail hidden away in a mass of beautiful decoration, the answer is 
 that so high a tradition would allow no detail to go astray or be wanting in 
 beauty. The fact that this design was used on the MaisonCJarree, shews that it 
 was considered by the Romans as among the best of the Greek border designs ; 
 and that, Owen Jones was right in including other borders in which this feature 
 appears among his examples of the Greek "key '' at its best. 
 
 I must not multiply unduly the examples of the badness of this "best"; 
 but I cannot refrain from giving two instances (figs. 56 and 57) of what, from 
 the stand point of these "keys' ;is the easterns devised them, can only be 
 described as artistic depravity. 
 
 Fig. 56. 
 
 Fi 
 
 g. ■><■ 
 
STUDIES IN THE decorative ART OF JAPAN. \21 
 
 This square ornament must have had a peculiar fascination for the Greeks, 
 for we are perpetually coming across it ; and it was developed in a peculiarly 
 grotesque fashion, two, and sometimes three, of the sides being prolonged into 
 ungainly and meaningless crooks, as in fig. 56 (No. 10, pi. XV. of the "Grammar,") 
 
 That the Greeks did not lack ingenuity, however, is manifest. I select one 
 special instance, which I have not come across in the East, and assume to be of 
 Grecian design. It is a double meandering line, fig. 55, which is not without 
 grace, though the re-iterated "steps" are somewhat irritating. 
 
 Fig 58. 
 
 The structure of this border depends on a locked combination of crooks and 
 " S "-figures, each with a return of two units, which make a svastika at the 
 crossing of the lines. 
 
 With these examples before us of the Greek " key," I take up the threads 
 dropped on a previous page, and discuss once more the meaning of " perfect " and 
 '• imperfect " " key "-borders. 
 
 I think that too much emphasis is laid on the necessity for a continuous 
 meandering fillet. It certainly led the Greeks to devise such abortions as fig. 
 57 ; for even in this there is a continuous line traceable round the inverted "T," 
 which owes its existence entirely to this supposed necessity. (The lines on the 
 right and left of this design are the sides of squares each containing an ornament.) 
 A more accurate definition of a " key "-border is that it is a design in which the 
 returned end of the paktva element, " key "-shaped, or the svastika, where this 
 key-shape is four times repeated, is made the principal motive, which is repeated 
 at regularly recurrent intervals. The discussion shews that this can only be 
 obtained in borders which are directly derived from pakwa diapers. A priori, 
 the regularity which the construction of these diapers demands is the only 
 source from which the same regularity could be obtained in the borders. But 
 the question will bear closer investigation. I assume that a border which 
 belongs to this scheme of decoration must be characterised by a repeated thought ; 
 it is the admitted feature of the Greek as well as of the eastern frets. But a 
 designer is not bound to adopt the " key " ; he may weave his lines as seems to 
 
12£ STUDIES IN THE DECORATIVE ART OF JAPAN 
 
 him best fitted to the object he has in view ; only if he does adopt it, lie must 
 conform to the laws which the use of the " key " has engendered. 
 
 I have laid considerable stress during the argument on the inference which 
 may be drawn from a prolongation of the lines of a border. Now designs 
 such as figs. 06 and 57, when they are treated in similar fashion, also produce 
 something on the other side of their bordering lines ; but it is not a diaper. 
 They do no more than repeat themselves above and below ; the pattern is reversed, 
 that is all. Owen Jones' description of the " fish-bone " design (fig. 29) exactly 
 fits them. But a mere repetition of a design even if it be reversed, does not 
 produce a diaper. I imagine that we are all agreed that a diaper 1 , although 
 it is based on the repetition of a small design, is a decoration which treats the 
 surface as a whole ; the idea of each part reflects the idea of the whole. But a 
 series of borders placed one above the other produces no continuity of surface 
 design, only a layer of bands. And if the bands are themselves ugly and irritat- 
 ing mazes of lines, the effect of repeating them cannot be anything but wearisome. 
 
 By " imperfect " fret I mean, therefore, a fret the design of which, when 
 its lines are prolonged, does not work into a diaper. In other words, I think the 
 diaper controls the true fret : the pakwa diaper the true " key "-border. 
 
 But I am not really much concerned with the definition of the true or per- 
 fect, as distinguished from the untrue or imperfect fret. There must always 
 be an allowance for liking and disliking. The fact that I dislike most of these 
 Greek frets intensely, that they irritate my sense of artistic propriety, will never 
 prevent ray friend from admiring them immensely. The only point really in 
 issue is whence the original simple " key "-border, with its numerous offspring, 
 came. That the system of borders in vogue in Greece travelled thence to China 
 is impossible on the face of the examples we have seen; such things are not to 
 be found in Eastern art, either at its best or worst. On the other hand, the East 
 had a complete and coherent system of diapers and borders, mutually dependent; 
 and so far as Greece is concerned the case can be put no higher than this, that 
 fragments of one part of that system are to be found there buried under a mass 
 of design which had no connexion, and was altogether out of sympathy with 
 that system. The conclusion seems obvious that the East was the originator, 
 Greece the borrower. 
 
 This then T believe to be the conclusion of the whole matter. The meander 
 in which the Eastern delighted to involve the most sacred symbols of his 
 religion became with the Greeks only a maze, in which the eve's bewilderment 
 led to nothing ; and that which was to them nothing led inevitably to meaningless 
 developments. The Eastern artists were not, it is true, fettered in their design- 
 
STUDIES IN THE DECORATIVE ART OF JAPAN' 120 
 
 ing by a religious law ; but the laws of design were to them as a religion, which 
 they could not if they would disobey. The Grecian artists also had laws of 
 design, but in them religion found no part. They developed beautiful forms 
 from nature; the honey-suckle inspired those free brush-strokes with the swell- 
 nv>- curves, which whether as flat ornament or architectural embellishment still 
 delight us ; and herein there is yet another link between the arts of these two 
 nations which knew not one another ; for in swelling brush-strokes the Japanese 
 too delighted, and their art derives much of its beautiful vigour from them. But 
 from these graceful floral forms the Greeks had already devised charming designs 
 for surface decoration. They knew too the virtue of the frieze, and, moreover, 
 acted on that little law which I have imagined to govern all good friezes. There 
 is a page of them (plate XVI) in Owen Jones' " Grammar ' to tell the tale; 
 some graceful, some severe. It is not hard to see how well these borders " went ' 
 with the surface ornamentation of honeysuckle on plates XIX and XX. Almost 
 may we imagine a Grecian household assembled in art conclave to decide the 
 momentous question of the decoration for the new villa, and Owen Jones' plates 
 serving as the decorator's pattern-book. I know which border they would have 
 chosen for each mural design. But as I turn these artistic pages as I turned 
 them many years before I knew the East, the same thought comes over me — 
 What is this plate of " key "-borders doing among the ornament of the Greeks ? 
 Surely it is full of alien ideas ; without coherence among themselves, with no 
 relation to those other designs which so delighted the eyes of Greece. You 
 cannot imagine a wall covered with honeysuckle curves waving in all directions 
 terminating in a key-border. " Keys " in Greece can only have been used for 
 plain surfaces ; there is no trace of surfaces covered with the diapers as in the 
 East. And to this questioning thought I think the answer is this — Whether 
 commerce or religion were the chosen emissary of art matters not at all ; but 
 there came at a certain time to the Greek artists fabrics from a far-eastern coun- 
 try, on which were woven strangely meandering lines, which struck them as a 
 pleasing fancy ; these they borrowed and treated after their own manner. But 
 the meaning of the symbol hidden among the lines did not reveal itself to them ; 
 they knew nothing of the verities of which they spoke to the devotees of the 
 mystic religion which had originated them ; and they used the lines, even the 
 symbols, without intelligence. Their very treatment reveals to us now, as I 
 think, that they were artists unlearned in the mysteries of an alien religion ; and 
 though they were lucky enough sometimes to light upon the best of the models, 
 they bent them to suit their own fancy, which went on its ow r n way leading to 
 mere bewildering mazes of meaningless lines. 
 
1.1<) STUDIES IN THE DECORATIVE ART OF JAPAN. 
 
 It seems to be fairly certain that the triangular grooving which lends so 
 much charm to the Japanese diapers apart from the designs themselves, never 
 reached Greece ; but I am uncertain to what age this form of work must be 
 attributed. One thing strikes me as curious ; the diapers can hardly have failed 
 to come to Greece ; for the enameller's is a very ancient art, and their work 
 would have been brought home from India with the stuffs by the travelling mer- 
 chants of Greece ; but the diapers, which must have been frequent on the enamels 
 never seem to have appealed to the Greeks in such a way as to lead to their 
 adoption. Possibly their formality displeased them ; and the family council, 
 which I have imagined considering the momentous question of design, was loth 
 to part from the waving foliations of their favourite flower. Yet, still relying on 
 Owen Jones, surface decoration composed of small ornaments regularly repeated 
 in spaces formed by intersecting lines, or without them, were common to Egypt, 
 and seem to have reached a certain degree of excellence in Assyrian art. It 
 would seem therefore, as there is no trace of this kind of surface decoration 
 among the Greeks, that the rejection of the eastern diapers followed their prior 
 rejection of those of Egypt and Assyria. 
 
These three concluding plates are referred to on page 55 as illustra- 
 tions of the best form of conventional naturalism : — 
 
 Plate 31, the nanten, in dull gold, with scarlet berries, on black 
 lacquer 
 
 Plates 32 and 33, iris, in dull sold, with Howers, in bright gold 
 foil, on black lacquer. 
 

 late 31. 
 
 THE NANTEN (NANDINA DOMESTieA) CONVENTIONALISED 
 
kites 32 :ui<l :!.'!. 
 
 IKIS, GROWING IN A STREAM, CON VKNTIONALISKD 
 
w 
 m 
 
 7 
 
 \ 
 
 1 
 
 m 
 
 ;n 
 
 
 m 
 
 T 
 
 + 
 
 It 
 0f 
 
 a- * 
 
 9 m 
 
 y Tfi 
 
 r 
 
 Y 
 
 + 
 
 
 0f * 
 
 f 
 
 ^ x-jy ^xxixxixx|> y 
 
 
 ] 
 
 # 
 
 m 
 
 ^/ it? 
 a X 
 
 fli! 
 
 V 
 
 ft 
 
 
 l 
 
 
 aH: 
 
 5/ 
 
 r 
 
 
 >3 
 
 y TfT 
 
 lU 
 
 m 
 
 ft 
 
 7 
 
 ^ A 
 
 ] 
 
 93 
 
 W 
 
 Jft 
 
 te 
 
 ea 
 
 m 
 
 + 
 
 + 
 
 — ^ 
 
 ■^ 
 
 ^ 
 
 ^ 
 
 -b 
 
 A 
 
 ^ 
 
 R 
 
 
 — ■ 
 
 » 
 
 + 
 
 
 X 
 
 H 
 
 B 
 
 « 
 
 SI 
 
 ft 
 
 «i| 
 
r 
 
 CT» 
 
 F^ 
 
 im, 
 
A>>1' 
 
 ^f/M^ 
 
 / /J M STERLING & FRANCINt tLAhi^ ami iwimuil 
 
 ■ NK1484 .P5 s ,ack 
 
 Piqqott, Francis Ta/Studies in the decor 
 
 II 
 
ft 
 
 '"'-•-.<...•" 
 
 
 
 
 p 
 
 ""S-aoT^i 1 * ..' 
 
 ^ ""Kb- ' 
 
 f <3 Vfe> 
 
 
 ^. «i iAii P a 
 
 ftswi"*''^ 
 
 p^s 
 
 ■~.\, . MMM«» 
 
 $£%. 
 
 ^ |F I ^• ! H 1 4#i 
 
 //r^kl 5 '! ■ 
 
 5^ :- " ':■ ff^' JP "■* -J 
 
 si/ 7 ^X;^w 
 
 awl 
 
 ■ 
 
 
 
 
 
 '■'■' 
 
 '■"<..-, ■. -"