Some of the New Zealand patterns (Fig. 43, and Plate VI., Fig. 12) certainly have a superficial resemblance to the more typical scroll patterns from the South-Eastern Archipelago of New Guinea, but there is no ground for comparing them except for this casual resemblance. The bird element is entirely lacking, and there is far less interlocking in the Maori than in the Papuan scrolls; there are also noticeable technical differences. My impression is that the carved designs have been derived mainly from tattooing, and possibly also partly from the dismemberment which so often befalls the conventionalised carvings of their ancestral figures. (Plate VI., Fig. 11.) When one looks at tattooed Maori heads or carvings of human figures one finds that rounded surfaces, such as the wings of the nose, the cheeks, the shoulders and thighs are usually decorated with spiral designs; this is in such places an appropriate device, as it accentuates the features which are ornamented, and personally I am inclined to believe that artistic fitness is the explanation of this employment of the spiral, and that it has been transferred to other objects as being a pleasing design, and that connecting lines have been made to give coherence to the decoration. It is worth noting that in early European art the shoulders and haunches of animals are often decorated with spirals.[17]

THE MATERIAL OF WHICH PATTERNS ARE MADE.

I.—The Decorative Transformation and Transference of Artificial Objects.

Dr. H. Colley March has introduced the term “Skeuomorph”[18] for the forms of ornament demonstrably due to structure. Professor G. Semper[19] “was the first to show that the basket-maker, the weaver, and the potter originated those combinations of line and colour which the ornamentist turned to his own use when he had to decorate walls, cornices, and ceilings.” So write MM. Perrot and Chipiez;[20] but this statement is too sweeping. A considerable amount of ornamentation is doubtless due to technique, but in Europe, Western Asia, and North Africa plant forms have had a great influence in the origin of designs, some of which have been modified by passing through a textile technique.

Given any object, two forces, so to speak, attack it—the utilitarian and the æsthetic. The resultant may be an implement which is solely useful and has little or no beauty to recommend it; or while retaining a full measure of utility, it may be beautified in form or in surface decoration; or, lastly, the object may become so glorified by the artist as to be translated from earthly use into the realm of æsthetics.

1. Transformation of a Solitary Object.

There are numerous examples of the annihilation of the useful by the beautiful. One instance came under my notice at the Murray Islands, in Torres Straits. Formerly when a girl was engaged to be married, in addition to numerous petticoats she wore a number of ornaments suspended from her neck and hanging down her back. The more important of these were white triangular pieces of shell, o, cut out of Conus millepunctatus; turtle-shell (“tortoise-shell”) bodkins (ter), used for shredding the leaves of which their petticoats were made, and for piercing the septum of the nose of infants; turtle-shell fish-hooks, and curious turtle-shell ornaments which are called sabagorar. These latter vary considerably in size, form, and amount of decoration; but by placing a number of them together a sequence can be obtained which illustrates the evolution of the sabagorar from the fish-hook (Fig. 44). Some hook-like objects are slightly ornamented with incised lines, and they might very well serve as fish-hooks; others are clearly totally unfitted for practical use, and may be quite plain or decorated. Fish-hooks (Fig. 44, A) are used in pairs, being fastened at each end of a piece of fine string, which, in its turn, is tied at its middle to the fishing-line proper. When the piece of twine with its hooks was thrown round a girl’s neck, the two hooks would often hang down her back shank to shank. Two sabagorar similarly arranged occur in the British Museum collections. What more natural than that this should be noticed, and to save the trouble of making two sabagorar a double one should be cut out of one piece of turtle-shell. The more remotely from the fish-hook did the sabagorar vary, the larger it became, and in some instances the double form became of considerable size, and the hook portion acquired a slight spiral curvature (Fig. 44, K). In one modified specimen the hooks are actually fused with the shank (Fig. 44, L). It will be also seen that divergent Λ-like processes often occur on the sabagorar, but are never found on the fish-hook.

The betrothal equipment of a girl thus consisted in the main of objects of utility which had reference to her future condition. The turtle-shell objects being easily cut, afforded a convenient field for ornamentation, and most of the ter implements exhibit a little decoration. The comparatively slender fish-hooks provided insufficient surface for ornamentation; the broadening of them for decorative purposes reduced their efficiency, so that in time the latter was sacrificed and a mere ornament resulted.

In the chain of islands which stretch away from the south-eastern end of New Guinea, one finds an interesting metamorphosis of the stone axe. The stone axe was very precious among these people, to whom the art of working in metals is still unknown. A large fine axe would have very considerable value, and the exhibition of it would be a symbol of wealth, and consequently of power. The desire to be recognised as wealthy has resulted in the development of a stone axe of which the stone is very large, often remarkably thin and beautifully polished, and is hafted to an unwieldy handle which may be carved and decorated with shell-money and other ornaments. The value of such an object seems to depend upon the amount of work required to produce it; its inutility enhances the reputation of the wealth of its possessor; thus we appear to arrive at certain primitive conceptions. Work done gives ownership or property. One form of wealth is the possession of unnecessary or useless property; the exhibition of this gives power to the owner.

I have made sketches (Fig. 45) of two axes in my possession. The first (A) is decorated with characteristic ornamentation, consisting in the upper part of birds’ heads and at the handle of the bird and crocodile design; but it is still a useful implement. The second axe (B) has a large thin stone, and is an unwieldy and probably quite useless object.

The late Mr. H. H. Romilly[21] tells us that at Utian (Brooker Island), in the Louisiades, “The stone implements made here are very fine. I got some axes of enormous size, which I am sure could not be intended for use. They seemed rather to be a common possession; perhaps two or three belonged to the village, and were exhibited on state occasions.” The Rev. Dr. W. Wyatt Gill,[22] at South Cape, saw “two axes solemnly carried by the chiefs as a preliminary to peace ... a glance at the slight artistic hafting will convince any one that they are not intended for cleaving timber.” This is all the information we have concerning these axes. It appears that they have come to be recognised as symbols of authority, but it is extremely doubtful whether they are anywhere held as a common possession.

A still more wonderful change has affected certain adzes in the Hervey Islands. (Fig. 46.) The stone blade is a carefully cut and polished piece of basalt, and it has every appearance of being perfectly serviceable; but the elaborately carved handles preclude the idea that in their present state they could be used for practical purposes. In form the handles may be quadrangular, gradually diminishing from the base to the blade, or conical, or polygonal or cylindrical. When short the handles are thick, even to the extent that they can scarcely be grasped by the two hands; these forms too are often perforated by quadrangular holes. One specimen in the Archæological and Ethnological Museum at Cambridge is six feet three inches in length.

Later on (p. 83) I shall describe the ornamentation on these adzes; at present we are merely concerned with the fact that for some reason or another they have become functionless through increase in the size of the handle, and by reason of the weakness caused by deep carving. We have now to trace the meaning of this vagary.

Dr. W. Wyatt Gill, who resided for twenty-two years in the Hervey Islands, and who has been a very careful observer and recorder of Polynesian customs and beliefs, informs us that “The adzes of the Hervey Islanders are frequently hafted with carved ‘pua’ wood. The carving, which is often admirable, was formerly executed with sharks’ teeth, and was primarily intended for the adorning of their gods. The fine-pointed pattern is known as ‘the sharks’ teeth pattern’ (‘nio mango’). Other figures are each supposed, by a stretch of the imagination, to represent a man squatting down (‘tikitiki tangata’). Some patterns are of recent introduction, and being mere imitations of European designs, are destitute of the significations which invariably are attached to ancient Polynesian carving. The large square holes are known as ‘eel-borings’ (‘ai tuna’); the lateral openings are naturally enough called ‘clefts’ (‘kavava’). To carve was the employment of sacred men.” Dr. Hjalmar Stolpe, of Stockholm, who has made a special study[23] of the ornamental art of these people, found in the museum in Chambéry an adze of this kind; according to the account on the label the stone had belonged to a chief, and it was after the owner’s death shafted in this manner that it might be preserved by his family as a remembrance. Dr. Stolpe continues, “The internal probability of the story confirms the truth of the account. Ancestor worship is a characteristic feature of Polynesian religion. The souls of the departed become the guardian spirits of the survivors. Their worship demanded a visible form, under which offerings could be enjoyed by them, and this was found sometimes in the skull itself of the deceased, which was preserved in the house, sometimes in some article of his property. In the latter case scarcely anything could be more suitable than the stone adze, which was the deceased’s most important implement, and which it required so much toil to make. On the Hervey Islands the transition was easier, as there the stone adze itself is considered as a god. Even the fine plait of coco-nut fibre with which the adze is fastened to the shaft was a god, and the method of binding it had, in Mangaia, been taught by the gods. Both during the operation of plaiting and during the decoration of the adze-shaft songs were sung in a low voice to the gods, that they might further the work. The ‘pua’ wood (Fagræa Berteriana) of which the carved adze-shafts are made may also have a religious significance, for Gill speaks of ‘its long branches being regarded as the road by which the spirits of the dead descended to Hades.’”

The following conclusions of Dr. Stolpe’s appear to be warranted:—“From these researches it appears to me to follow that the peculiarly shafted stone adzes of the Hervey Islands have a religious signification, that they are especially connected with ancestor worship, and that they were probably the very symbols under which this worship was performed.”

Dr. H. Colley March[24] has gone a step further, and tries to account for the very remarkable form of the handle of the sacred adze. He says, “It is remarkable that the typical Mangaian axe [adze] was exclusively associated with ‘Tane, the royal-visaged.’ This god was widely venerated over the Pacific; in Mangaia he was especially the drum-god and the axe-god; he presided over the erotic dance as well as over the war dance ... it is evident that the drum was not only associated with a Tane cult in the erotic dance, but was regarded as Tane’s embodiment; when the drum was beaten, it was Tane that was struck, and from the fissure in the drum it was Tane’s voice that issued.” Dr. March quotes a number of extracts from early voyagers, etc., descriptive of various Polynesian drums, and he comes to the conclusion that the upright drums, which were hollowed out of a single piece of wood, were originally derived from bamboo instruments. He figures a drum (Fig. 47) said to have come from Java, which, with the exception of the terminal head, corresponds closely with the drum called naffa which Captain Cook describes at Tonga. He concludes that after the drum “had passed from bamboo to wood, the horizontal instrument assumed the erect form, more appropriate to the god, and was then surmounted, as in the so-called Javan example, by Tane’s head, which subsequently gave place to Tane’s adze. As the cult differentiated, the symbolism differentiated too.” Without going into further detail, in the short thick form of the Mangaian adze, such as Fig. 46, the upper portion of the handle is usually cylindrical. The lower portion is usually quadrangular, or may be polygonal, and looks as if it might be a pedestal for the former. According to Dr. March’s interpretation, the stone implement represents the head of Tane; the upper cylindrical part of the handle is his neck. The lower part of the handle is an artistic analogue of the sacred drum; “the useless transverse closings represent the original bamboo joints, as well as the solid ends of the wooden drum. In spite of the fact that their presence increased the difficulty of hollowing out the shaft, they were reproduced in obedience to a well-recognised law. The square and oblong rectangular openings have an analogous explanation. They indicate the original aperture, whether the slit in the bamboo, or the single or double chink in the wooden drum which was excavated through the drum in order to secure its resonance. The great increase in the number of apertures, helped by rectangular designs on horizontal instruments, took place as an evolution of ornament that largely consists in a multiplication of functionless details.”

It is possible that the adzes from the Hervey Islands, with long, unperforated carved handles, may have a different history from the form illustrated in Fig. 46; they may merely be decorated but useless adze handles. In any case, the above-quoted conclusions of Dr. Stolpe may be accepted.

In the three examples of the metamorphosis of a practical object into an unpractical one just recorded, we have an illustration of the effects of three dominant human forces on these several implements, art, display or wealth, and religion. The result is practically the same in all cases, but the motive leading to it is different. Analogous modifications are everywhere to be met with.

2. Transference of Fastenings.

One of the earliest handicrafts was to fasten two things together. To quote from Dr. H. Colley March,[25] “As soon as man began to make things, to fasten a handle to a stone implement, to construct a wattled roof, to weave a mat, skeuomorphs became an inseparable part of his brain, and ultimately occasioned a mental craving or expectancy.”

In order to securely fasten two objects together, such as splicing wood or fastening a handle to a stone implement, a lashing is necessary, and the nature of the latter varies more or less according to the conditions under which the artificers live. Where mammals are abundant, their sinews afford a readily procured and very strong, fine lashing, but it occurs only in short lengths. The hide of a newly-killed animal is pliant, strong, and can be so cut as to produce long thongs. Owing to the rarity of mammals in New Guinea, and their absence from the Great Ocean, the Papuans, Melanesians, and Polynesians make no use of skins or thongs; sinews may be employed, but the great bulk of all fastening is accomplished by the employment of vegetable fibres. The inner bark of various trees supplies bast and tapa, several vegetables have long fibres which are utilised, but the most widespread and important of all lashings in Oceania is the twisted or plaited string made from the fibres of the husk coco-nut. The latter is known as sinnet, and there are many degrees of excellence in its manufacture; for rough work it is coarsely plaited, but nothing can exceed the delicacy and beauty of the finest sinnet work, such, for example, as occurs on the symbolic adzes of the Hervey Islands, where it was even regarded as a god.

Most of the stone implements of primitive man were fastened in various ways into handles, and an inspection of almost any ethnological collection will demonstrate the diverse methods of lashing employed by even the most backward peoples. For example, we have in Plate I., Fig. 1, an illustration of the fastening of the stone axe of Montezuma II., now in the Ambras Museum at Vienna,[26] but analogous figures will be found in numerous books of travel, or in ethnographical journals and treatises.

The even serving of the lashing gives rise to geometrical figures. One might in some cases describe them as patterns, whose symmetrical disposition gives a pleasing effect.

In process of time the stone spear points of our ancestors were replaced by bronze, and during the evolution of the palstave, or socketed bronze celt (Plate I., Figs. 4, 10, 11), from the flat bronze celt, the method of fastening also changed. But by this time the old style of binding had become so associated in men’s mind with the implement, that it was engraved on the socket of the bronze head as a pattern. Hence most of the ornamentation of bronze implements. (Plate I., Figs. 2-4.) On socketed bronze celts one frequently finds (Plate I., Figs. 10, 11) two, three, or more ridges running from the base to some distance towards the end; three is the most common number of these ridges. They may fade away at their ends, or terminate in slight knobs or annular prominences. The meaning of these characteristic markings is at present obscure, but they appear to be skeuomorphs of lashing.

What are known as “beads” have frequently the same origin; that is, they are reminiscences of fastenings. This is especially evident when the bead is decorated with a twisted design, as occurs in the zonal decoration of a bronze vessel from a Swiss lake-dwelling. (Plate I., Fig. 5.) There is no reason to believe that lashing was actually employed on older forms of Assyrian combs, or prehistoric bone needles or bronze knives, nevertheless the patterns shown in Plate I., Figs. 6, 8, and 9, have doubtless been derived from ligatures; more from the fact that such patterns were familiar, and a feeling for a need of decoration, than for any special appropriateness.

One frequently finds designs in the ornamentation of objects from Oceania which are evidently based upon sinnet lashings. To take a few out of many examples now before me, in Fig. 48 we have a reduced rubbing of a carved cylindrical club, said to come from the Friendly Islands (Tonga); the same kind of club also occurs in Fiji. The decoration of this club irresistibly suggests bands of plaited sinnet irregularly bound round the club.

In these two groups of islands sinnet is often worked into a design that is also copied on the upper part of a carved wooden club. (Fig. 49.) The same kind of lashing is seen in Plate I., Fig. 1. Occasionally, instead of being angular, this pattern is carved in curved lines, and so gives rise to an imbricate pattern, which might be mistaken for a scale pattern.

Other sinnet patterns perhaps occur in the lower part of the decoration of a Tongan club. (Fig. 50.) The design on the upper left-hand corner is evidently copied from matting, and it frequently occurs on these clubs. This figure also illustrates the Tongan peculiarity of inserting little figures into designs, in this case a man and probably a frigate-bird.

I do not wish to suggest that all zigzags included within parallel lines, as in Fig. 48, or such simple designs as those of Fig. 50, are everywhere sinnet derivatives, or otherwise skeuomorphic; some, at least, in the Pacific certainly are. We have seen that birds’ head designs may degenerate into zigzags (Figs. 30, 36), and we shall see that frogs’ legs (Fig. 122, B), snakes (Fig. 103, G, H, K), alligators (Fig. 97, E, F), and even the human form (Fig. 125, A) may pass into zigzags. There are many other possible origins of the zigzag, but in many cases it is probably only a purely decorative motive of no further significance. The simple zigzag can be traced in ancient Egyptian art as far back as 4000 B.C., and, according to Professor Flinders Petrie, it continued popular with a few modifications for about 2000 years, when spots were associated with it, but these were adopted from foreign art. About the eighteenth dynasty the use of the zigzag was discarded in favour of the wavy line and various scroll designs. In all cases it is necessary to study each pattern locally.

3. Skeuomorphs of Textiles.

In Europe a very early form of fabric was wattle-work, formed by the interlacing of flexible boughs and wands. The most ancient huts were doubtless made of wattle-work daubed over with clay. Only very exceptionally are traces of these structures found, as, for example at Ebersberg, where Dr. Keller[27] found, among the débris of a lake-village which had been destroyed by fire, fragments of the clay daubing, “smooth on one side, and marked on the other, with deep depressions of the basket-work.” The pattern thus impressed on the clay is one of repeated straight lines crossed by a contrasted series of curved ones. (Plate II., Fig. 1.) Thus the fire which consumed the house baked its clayey coating, and in this way preserved for us a record of what it destroyed.

I do not know whether the wattle-work has been perpetuated on any object as a skeuomorph, but it is possible that the shape of similarly constructed huts has been continued, as Mr. Charles de Kay suggests,[28] into the round towers of Ireland. He says, “Seeing how the Irish kept heathen ideas in other things, we can perceive how the round wicker house of the Kelt, such as we see it carved on the column of Antoninus at Rome, developed into the wood and wicker outlook tower and beacon, and in skilful hands became the Irish round tower. Christian in usage, they are pagan in design.”

The predatory expeditions of the Scandinavians created a demand for watch-towers and places of temporary refuge; the pattern for these was supplied by the traditional erections of the Gauls, but their translation into “towers more durable, useful, simple yet stately, than anything Ireland had seen before or has seen since,” was due to the skill and experience of “Byzantine craftsmen driven from the East by the bigotry of the image-breaking emperors.”

Mr. de Kay also calls attention to the encircling stone bands, or “string-courses,” as in the round tower at Ardmore, “which repeat, without any useful object in stone, the horizontal bands that strengthened the tall wicker house of the Gauls. Such apparently trivial points weigh heavily in favour of the indigenous character of the round tower of Ireland.”

The interlacing of flexible bands, such as strips of bast, entire leaves as of grass, or shreds of large leaves, is known to almost every people, and is employed in making mats. When the elements employed are all of one size, and when the plaiting is straight, the intersections form regular equilateral rectangles or squares. (Plate II., Fig. 3, and compare the transferred design in Fig. 50.) If the material consists of two colours simple patterns are readily produced, but of necessity the patterns must consist of straight, slanting, or zigzag lines; curves are an impossibility. The same holds good for nearly all forms of matting and basketry which is made of strips of one material, but the constructional surface marking may be rectangles of various shapes and sizes instead of simple squares. (Plate II., Fig. 4.) When one series of the components is twisted, as in Plate II., Fig. 5, there is a kind of flow effect in the intersections.

The making of baskets by laying down the material in a spiral gives rise to different effects, especially when coloured strips are interwoven for decorative purposes—as, for example, in some African baskets and the baskets made by the natives of South Australia, in the neighbourhood of Adelaide. Dr. Keller found in the Lake of Robenhausen a kind of basketry formed by bast, the fibre of the lime-tree, intertwisted among a series of willow rods, the strips “running concentrically in such a way that both together form a structure like that called ‘herring-bone.’”[29] (Plate II., Fig. 2.) It is possible that the pattern in the middle band of Fig. 49, and some of those in Fig. 50, may have been suggested by basketry or plaited fans.

An early type of basket is seen in the Roman corbula (Plate II., Fig. 6), in which the osier rods are placed rectangularly; another, in an ivory plaque from Boulak (Plate II., Fig. 7), in which there is a chevron arrangement. The latter is the more common skeuomorph on European prehistoric pottery, but the rectangular type often occurs, and it may be seen on a Danish food-vessel of the Stone Age. (Plate II., Fig. 8.)

The bottom of a basket, with a cruciform arrangement of the bands, due to the method of weaving, was discovered by Dr. Keller in the Terramara marl-pits of Northern Italy (Plate II., Fig. 9); and a piece of pottery from the same deposit is ornamented with a corresponding skeuomorph (Plate II., Fig. 10).

Dr. Colley March has further developed this subject, and, while I cannot commit myself to several of his conclusions, I do not hesitate to give an exposition of his ingenious views, as they are very suggestive, and even if they are not finally accepted, they will lead to a further examination of the problems:—

“The perpetual concentration of attention, the strain of hand and eye and brain upon the forms of wattle-work and basketry produced an important decorative result. The mind acquired an expectancy of a special mode of curved repetitions. This particular skeuomorph is composed of a band that winds in and out among a row of rods or discs.” (Plate III., Fig. A.)

The “discs” are naturally the cross sections of the vertical elements of the wattle-work—that is the “rods.” “The device underwent a change in opposite directions. The discs grew, or they vanished. In the latter case the band left by itself is the meander, and may be called a curvilinear zigzag. In the former case the discs often became the seat of phyllomorphic invasion, and were transformed into leaves or flowers.

“Examples may be seen on the margin of a bronze shield from Cyprus (Plate III., Fig. 2); on a vessel of terra-cotta from the third sepulchre of Mycenæ (Plate III., Fig. 8); and on an enamelled Roman vase found on Bartlow Hill (Plate III., Fig. 5); whilst a twin-form, which presents both contrast and repetition, occurs on another bronze shield from the Mediterranean (Plate III., Fig. 1) and is the basis of the Assyrian ornament and its Greek variant called the guilloche. (Plate III., Figs. 4, 3.)

“A different skeuomorph is derived from a different method of basketry, in which a single fibre is turned round a row of osier-sticks, so as to produce a wave repetition (Plate III., Fig. B), as may be seen on the pottery of the ancient Pueblos (Plate III., Fig. 6). When these discs disappear, the fibre by itself resembles the Vitruvian scroll, and may be called a curvilinear fret. (Plate III., Fig. B.)

“Whenever the pattern has a stepped form, as on many of the Pueblo vases (Plate III., Fig. 7), it indicates that the methods of textile manufacture had already influenced the eye and mind of the race before the invention or introduction of pottery.”

The scroll-patterns illustrated by Dr. March may at one time and place have had the origin supposed by Dr. March, but it does not appear to me to be probable that they would have arisen in this way both in South Europe and in Mexico. I have shown (p. 51, Fig. 27) how a simple guilloche has arisen from interlocking birds’ heads. The Vitruvian scroll design occurs among the Tugeri head-hunters of New Guinea, and it is most improbable that it owes its origin to basketry. It is probable that the Pueblo pottery with curvilinear patterns, such as Plate III., Fig. 6, is more recent than that with angular designs; but I shall return to this later on. In fact, I would feel inclined to state that Dr. March’s view is possible for the origin of the patterns in question, once and in a restricted locality, but highly improbable for wide application.

There is a great tendency for spirals to degenerate into concentric circles; examples could be given from New Guinea, America, Europe, and elsewhere. In fact, one usually finds the two figures associated together, and the sequence is one of decadence, never the evolution of spirals from circles. The intermediate stage has been aptly termed a “bastard spiral” by Dr. Montelius, “that is to say, concentric circles to which the recurved junction-lines give, to a casual glance, the appearance of true spirals.”[30]

“The strangest skeuomorph of all,” writes Dr. March, “was that common to the early inhabitants of Northern Europe. They were adepts in basketry, and in wattle-work for walls and ramparts. Moreover, the pliant bark of the birch was ever ready to the hand for a thousand purposes of life. The Norwegian still makes hinges for gates and loops for the oar out of the entwisted fibre. The old Norseman spoke of the rudder withy, for the earliest rudder was an oar; and leather thongs were also used to keep the oar against the thole-pin. The skeuomorph consists of a withy wound upon itself. (Plate VII., Fig. 11.) This device, wrongly called a rope-pattern, gained such an ascendency over the northern mind that it was employed sometimes as a symbol (Plate VII., Fig. 12), like the reefing knot on Roman altars. (Plate VII., Fig. 13.) It was used also by the ancient Hittites. (Plate IV., Fig. 1.)

“It is evident that the withy skeuomorph (Plate IV., Figs. 2, 3), the Scandinavian worm-knot, established itself as a necessity of the mind before those men who were dominated by it had discarded a covering of skins for one of cloth; for its type is antagonistic to the regular intersections and the stepped designs of textile fabrics, and no trace of these appears on their early pottery.

“When weaving was at last introduced, so as to be practised by these people, it was probably along with the introduction of metals. But for a while the use of metal only increased the number of twisted things. The words, wire, wicker, and withy are all from the root WI, to plait, and the Teutonic WIRA means filigree, an ornament of twisted filaments of metal; and as the simplest manner of terminating a wire is to coil its end, the earliest filigree is preponderantly spiral. (Plate IV., Figs. 5, 6, 7.) Thus was the way prepared,” concludes Dr. March, “for the advent of the serpent zoomorph, so much affected by Teutons and Scandinavians.”

In early times wooden bands were interwoven to form flat surfaces, as, for example, in the floor of a lake-dwelling at Niederwyl, in Switzerland (Plate IV., Fig. 8), but few traces of the art of “fascining,” as Dr. March points out, remain to us from antiquity, since wood-work rapidly perishes by decay, and is easily destroyed by fire. This art produces a bold decorative effect which appears to have been perpetuated in various ways. Amongst others may be mentioned the interior decoration of an earthen vessel from Ueberlingen See (Plate IV., Fig. 9), a crescent of red sandstone from Ebersberg (Plate IV., Fig. 10), and an incised stone from Hadrian’s Wall, in Northumberland. (Plate IV., Fig. 11.)

So far we have only considered the type of ornamentation which occurs on plaited or woven objects, and these are seen to be conditioned by that particular technique. We have now to see what occurs when a new material is substituted for the old.

There are many varieties of tapa in the Pacific, some of which are coarse and others of extreme fineness and softness. The process of making and decorating tapa has often been described; sometimes the tapa is ribbed, having been beaten with more or less finely corrugated wooden mallets, occasionally it is marked with squares which give it an appearance of having been stamped by a simply plaited mat, but many pieces are quite smooth. There is nothing in the texture or manufacture of tapa to prevent its being ornamented with intricate and involved patterns. As a general rule, all over the Pacific we find that tapa patterns are largely geometrical—that is, they are formed of straight and angled lines; bowed lines, which are grouped into leaf-like designs, are not infrequent, but doubly curved lines and scroll-like designs are extremely rare. The evidence clearly points to a time anterior to the employment of tapa, and when mats and other textiles were the only fabrics; the decoration of these was necessarily angular in style. When tapa became general the older designs were transferred to the new material, and quite irrespective of its capabilities. Only gradually has it been found that the smooth surface of tapa lends itself to a more elaborate decorative treatment. The essential conservatism of the savage precludes rapid emancipation from long existent thralls, especially as the æsthetic mind has, so to speak, become set in angularities.

It is probable that the practice of beating tapa with wooden mallets led to the discovery of printing in colours. The transitions are slight between finding the natural graining of wood impressing itself on the soft tapa, of so cutting the mallets as to produce a regularly grooved surface, and of colouring the blocks, and lastly of making the great printing blocks on which the pattern stands up in relief, which were made in Fiji. Sometimes the lines in relief of printing blocks are made by fastening the mid ribs of palm leaves on to a stout piece of tapa.