With none of these three basic colours was any national feeling a.s.sociated, yet the Persians excelled in the use of blues. The Turkomans of Turkestan and Asia Minor produced better reds than any other colour, and the best yellows, even if generally inferior in positiveness to blues and reds, were those of the Chinese.
Though other primary and secondary colours sometimes result from the application of a single dye, the many thousand different tints can only be produced by the blending of two or more. Moreover, the qualities of the same dye vary greatly, as they depend on the soil where the plant grew, the time of year when it was removed, and the weather and other conditions prevailing during the dyeing.
In nature green is one of the most pleasing colours, but in carpets it is most unsatisfactory, as it has generally a faded appearance, due probably to the fact that one of the dyes of which it is formed by blending is less permanent than the other. The Chinese greens obtained from the buckthorns are generally the best.
Greys and browns are sometimes derived from gall nuts, and reddish brown from henna. For very dark browns and black, iron pyrites has been largely used in both old and modern rugs; but unfortunately the dye has a corrosive effect on the wool, so that the black knots of old rugs are often worn to the warp.
In parts of India flowers of the b.a.s.t.a.r.d teak (_Butea frondosa_) make a favourite dye, from which are produced, by blending with other dyes, a large number of shades ranging from deep yellow to brownish copper tones. Another well known dye is _b.u.t.ti lac_, obtained from an insect, _Coccus lacca_, that lives on the twigs of trees. It is a subst.i.tute for cochineal and produces different shades of red, crimson, terra cotta, and purple, according to the other dyes and the mordants with which it is blended.
Besides these few dyes are innumerable others that are used either singly or in combination. Furthermore, different colour effects are produced by the application of different mordants, which it is necessary to use for the reason that without them many fibrous materials are unable to absorb a large number of the dyes. The most valuable of all mordants is alum; and the sulphate of iron and tin are largely employed in the case of red colours. Of the vegetable mordants, pomegranate rind, which contains some yellow colouring matter, is the best known. Valonia also is sometimes used, as well as limes, lemons, the fruit of the tamarind, and the mango.
In the monograph of Mr. Harris on the "Carpet Weaving Industry of Southern India" are a number of directions from an old ma.n.u.script owned by a dyer who stated that he was the descendant of twenty generations of dyers who originally came from Tabriz, and that he had made his copy from a Persian book of dyes which had belonged to his grandfather. A few of these are given below, because they show not only the dyes and mordants, but also the methods employed.
"Birbuls Blue. Take cinnabar, indigo, and alum, grind and sift lighter than the light dust of the high hills; soak for ten hours; keep stirring it; put in the wool and soak for many hours. Boil for three hours; wash in kurd water, water in which kurds and whey have been well beaten up; leave for three hours, and then wash and beat again in water.
"A Fine Indigo Blue. Take indigo, soak it in water for twelve hours, grind it to a fine paste in a mortar, add some _Terminalia citrina_, pomegranate peel, and alum; and mix thoroughly. Boil; put the water into the hot bath and keep stirring till cold. Now mix in some iron-filings water, and boil steadily for another two and a half to three hours; wash with a beating and dry.
"Ruddy Brown Grey. Take sulphate of iron, _Terminalia citrina_, oak galls, and alum; mix well; dry; then steep for twenty-four hours. Put in the wool; soak it for twenty-four hours, then boil for two or three hours. Dip in a soda-bath, wash, and dry.
"Cinnamon. Take oak galls, acacia bark, cinnabar, and alum, and steep for a night. Put in the wool, and soak for twenty or thirty hours; boil the water for two or three hours and give a soda-bath wash; dip in acidulated water; and wash again with beating.
"Crimson. Take lac colour and cochineal. Steep for from four to six days in the sun, in hot weather for the lesser time, stirring constantly till a rich deep colour comes where some has stood for a few minutes in a thin gla.s.s bottle and settled. Then strain through two cloths, and put in pomegranate rind and good iron-filings water. Add mineral acid; steep wool for thirty-six hours, then boil for three hours, wash well, and dry.
[Ill.u.s.tration: _COLOUR PLATE II-OUSHAK CARPET_
_The colours and pattern of this antique Oushak are similar to those of the best examples that remain of the carpets woven in Asia Minor during the XV and XVI Centuries. The deep blue of the central field, the rich red of the medallions, and the golden yellow of the leaves are entirely unlike the more subdued hues found in Persian rugs. Strongly contrasting with them are the more delicate tones of the tendrils and leaves, which display in their drawing a keen sense of refinement. In the formal pattern of the field are stateliness and elegance; in the narrow borders are simplicity and grace. Such colours and drawing show that the early Asia Minor weavers had an intense appreciation of the enn.o.bling qualities of beauty and harmony._
_Loaned by Mr. James F. Ballard_]
"Pale Greyish Green. Take copper rust, asburg,[5] and alum. Mix well with any hot water, not boiling; soak wool for eighteen hours, then boil for three hours. Give a bath with water acidulated with some limes, and dry in shade.
"Old Gold and Rich Yellow. Take turmeric and asburg, cinnabar and alum.
Soak all night. Steep wool for twenty-four hours, boil for four and a half hours, wash with a beating, and dry in shade.
"Dark Grey. Take of the fruit of _Cupressus sempervirens_, seeds and seed pods of babul (_Acacia arabica_), iron-filings water, and alum.
Steep over night. Now add the water and let it soak for twenty-four hours, then boil for two or three hours, until the colour is right, then wash and dry in the sun.
"Rose Colour. Take ratanjot (_Onosma echioides_), a thought of cochineal, manjit (_Rubia cordifolia_) or lac colour a very little, and cinnabar. Add water, soak them for twelve hours, put in wool, and steep for thirty-six hours; cook it for three hours, then bathe the wool in alum and wash nicely; afterward dry in the shade.
"Persian Scarlet. Take lac colour, and if you choose a little cochineal for richness, and soak from four to six days; strain it in two cloths and add alum and a little turmeric; let it stand for three hours. Put wool in and steep for twenty-four hours, then boil for two hours. Take out the wool and add mineral acid; re-enter wool and boil an hour more.
Wash fifteen minutes when cold, and dry in the shade.
"Saffron Yellow. Take turmeric, cinnabar, and soda, add water and keep for a full day. Then add some alum, make the dip, and soak the wool for thirty hours. Cook it for several hours, and dry in the shade after beating and good washing.
"Rich Yellow. Take asburg and turmeric, soak for a night in water, steep the wool for twenty-four hours, add alum, shake out, and dry in shade."
Identical shades of a number of colours are not produced in all parts of the Orient, not only for the reason that soil, moisture, and climate affect the colour values of dye-stuff, but because each family of dyers preserve inviolable the craft secrets transmitted from their forefathers. Thus it happens that different parts of the rug-producing countries adhere to particular tones that help to identify the locality where the fabrics were woven.
Unfortunately the Western aniline dyes, which were introduced about the year 1860 and quickly adopted because they are cheaper and less complicated in their application, have to such an extent transplanted some of these fine old vegetable dyes that a number of the richest and most delicate colours found in the rugs of a former century are no longer produced. Thus the superb blue of the fine old Ispahans, as well as of l.u.s.tre tiles and illuminated ma.n.u.scripts, belongs to a lost art.
The disadvantages of the aniline dyes are several: they have a tendency to make the fibres of the textile fabric brittle, and when it is wet the colours will frequently run. Some dyes also fade more readily than others, so that if a colour be the product of two or more dyes, the resultant tint may be totally unlike the original. On the other hand, not all vegetable dyes are fast; but as they fade they mellow into more pleasing shades. Efforts have been made to encourage the use of old vegetable dyes; but unless the laws which have been enacted in parts of Asia to restrict the importation of aniline dyes be more stringently enforced than in the past, the cultivation in the garden patch of the dye-producing herbs and plants will soon cease to be the time-honoured occupation it was in days gone by.
Almost as important as the art of preparing the dyes is that of properly applying them to the yarns. It is an art that demands infinite pains in its technique, as well as a lifetime to acquire. It is in itself a separate profession practised by artisans who guard with jealousy the sacred secrets that transmitted from generation to generation occupy their thoughts to the exclusion of almost everything else. The homes of these professional dyers in the larger villages and cities are located on a stream of water which possesses mineral properties that long experience has proven especially suitable as solvents for the different kinds of colouring matter. Ranged about the walls of their low dwellings are jars or vats containing liquid dye of various colours. Suspended above them, from hooks driven into beams, are the yarns from which, after immersion in the proper vats, the liquids are allowed to drain.
After this the yarns are exposed for the proper length of time to the dry air and burning sun. It is, therefore, the suitable mordants, the preparation of the proper dyes for the vats, the immersion of the yarn in correct sequence and for the correct length of time, as well as the exposure to the glare and heat of the sun for a definite period to be gauged to the exact moment, on which the colour results depend. This complicated process by which, for instance, the infinitely different shades of a red, a blue, or a brown may be conveyed to yarn by using the same dyes but by slightly modifying the steps requires the greatest precision, for which no rule but an experience amounting almost to instinct is the guide.
[Ill.u.s.tration: PLATE 6. FERAGHAN RUG]
There was a time when the Oriental had not learned the meaning of _tempus fugit_ or seen the glitter of Western gold, when his dyeing and weaving were proud callings, in which entered his deepest feelings. Then the old vegetable dyes that mellow, grow softer and more l.u.s.trous, were almost exclusively used; but now throughout all weaving countries the dyer has deteriorated so that he can no longer produce some of the rich colours in use half a century ago. Yet remote from the princ.i.p.al lines of travel, on the edges of the desert, in lonely valleys, among rugged mountains, half-tamed tribes are still dyeing their hand-spun yarn as did their fathers" fathers.
CHAPTER V.
WEAVING
NEAR the tents of some nomadic tribes may occasionally be seen crude looms on which are woven some of the most interesting rugs that now reach the Western markets. In all probability they are not dissimilar to what were used thousands of years ago, for it would be impossible to construct a simpler loom. Where two trees suitably branching are found growing a few feet apart, all of the upper branches are removed excepting two, which are so trimmed as to leave a crotch at the same height in each tree. In each crotch is rested the end of a pole or beam, and parallel to it is placed another extending at a short distance above the ground from trunk to trunk. Or, as is more frequently the case, roughly hewn posts are firmly implanted in the ground and horizontal beams are stretched between them. In the upper one is a groove with a rod to which one end of the warp, consisting of strong threads of yarn numbering from ten to thirty to the inch, is attached, while the other end is tightly stretched and firmly secured to the lower horizontal beam. Sometimes the beams to which the warp is attached are placed perpendicularly, so that the weaver may stand and move sideways as the work progresses. But among a very large number of those tribes that are constantly wandering in search of new pastures for their flocks and herds, it is customary to let the loom lie flat on the ground, while the weaver sits on the finished part of the rug.
[Ill.u.s.tration: PLATE A.-AN UPRIGHT LOOM]
Under more favourable circ.u.mstances, when the tribes live in villages or cities, the looms are so made that the weavers are not compelled to bend in order to tie the first row of knots or stand erect to finish the last rows of a long rug. Of the several devices by which the weaver may remain seated while at work, the crudest consists of a plank used as a seat, which rests on the rungs of two ladders placed parallel to each other at the sides of the rug. As the work progresses, the plank is raised and rested upon the higher rungs. More frequently, however, both upper and lower beams of the frame have the shape of cylinders of small diameter, which revolve between the upright posts. The lower ends of the threads of warp are attached to the lower beam, and the other ends may either be wound several times around the upper one or else pa.s.s over it and be kept taut by weights attached to them. Such a loom is generally used for weaving very large rugs, which are rolled up on the lower beam as the work progresses.
In Plate A (Page 45) is represented a loom commonly used in many parts of the Orient. When preparing it for weaving two stakes are driven in the ground at a suitable distance apart, and about them the warp is wound in the way a figure eight is formed. The warp is then carefully transferred to two rods that are attached to the upper and lower beams.
If it has been carefully wound, none of the threads should be slack; but if desired the tension may be further increased by different devices.
Two other rods, known as "Healds," are then attached to the front and back threads of warp; or in the case of a single rod, it is attached to the back threads, as shown in the Plate. A lease rod is next inserted between the threads of warp that cross below the upper beam, and another is placed below it where, if necessary, it is supported in position by loops. When the weaving begins, a short web is generally woven at the lower end to protect the knots from wear. After the first row has been tied, the shuttle carrying the thread of weft is pa.s.sed between the front and back threads of warp; the heald rod attached to these back threads is then pulled forward, so that they are now in front of the others, and the shuttle is pa.s.sed back. If the rug is narrow, only one shuttle is used; but if the rug is wide, or if the weft consists of two threads of unequal thickness, a shuttle is pa.s.sed across from each side.
Every thread of warp is in this way completely encircled by the thread of weft as it pa.s.ses and repa.s.ses. When weaving large rugs, there is an advantage in having two heald rods, as by their use the distance between the front and back threads of warp may be increased. The object of the lease rod is to prevent any slack caused by drawing forward the threads of warp, and is accomplished in a very simple manner, as will be seen by studying the drawing; since when the tension of the back threads is increased by drawing them forward, the tension of the front threads is also increased by displacing the lease rods which thereby stretches them.
[Ill.u.s.tration: PLATE 7. FERAGHAN RUG]
The products of the loom are divided according to their weave into three separate cla.s.ses. The simplest of these are the kilims, which are without pile and consist only of warp and weft to which a few embroidered st.i.tches representing some symbol are occasionally added.
A more elaborately made cla.s.s are the Soumaks. They consist of warp covered by flat st.i.tches of yarn and of a thread of weft which extends across and back between each row of st.i.tches in the old rugs and between each second and third row of st.i.tches in the new rugs. In the narrow, perpendicular lines that define both borders and designs the st.i.tch is made by the yarn encircling two adjacent threads of warp; but in other parts of the rug it is made by the yarn pa.s.sing across two adjacent threads of warp at the front, and after encircling them at the back, recrossing them again at the front. It is then continued across the next pair of adjacent threads of warp. The result is that at the back of these rugs each of the two threads of warp encircled by the yarn appears as a separate cord, while at the front the yarn pa.s.ses diagonally across four threads of warp. As this diagonal movement is reversed in each succeeding row, the surface has an uneven appearance sometimes termed "herring bone" weave.
By far the largest cla.s.s of rugs are those with a pile. When making them, the weaver begins at the bottom and ties to each pair of adjacent threads of warp a knot of yarn so as to form a horizontal row. A thread of weft is then pa.s.sed, as often as desired, between the threads of warp and pressed more or less firmly with a metal or wooden comb upon the knots, when they are trimmed with a knife to the desired length. Another horizontal row of knots is tied to the threads of warp; again the yarn of weft is inserted; and so the process continues until the pile is completed. In tying the knots, work almost invariably proceeds from left to right and from the bottom to the top. It is but rarely that the warp is stretched horizontally and that the knots are tied in rows parallel to the sides. It is still more infrequently that a rug is found in which the knots are tied by working from the centre to the right and left, and to the top and bottom. These interesting exceptions may easily be discovered by rubbing the hand over the pile, when it will be noticed that the knots lie on one another so as to face the same direction, which is the opposite to that in which the work of tying advanced, or as is generally the case, from top to bottom.
The compactness, durability, and value of a rug depend somewhat on the number of knots in any particular area. Yet if the yarn is coa.r.s.e, the rug may be compact even though the number of knots be small; and if the yarn is fine, the rug may be loosely woven, either because the rows of knots have not been firmly pressed down, or because there are several "filling threads" of weft, and still the number of knots be large. A square inch is a convenient size for measurement; but since all parts of a rug are not woven with equal compactness, the measurement should be made in several places if exactness be required. In loosely woven pieces, such as the Oushaks and some of the Genghas, there may be less than twenty knots to the square inch; but among the more closely woven, as the Kirmans and Bokharas, are frequently several hundred.
These knots are of two cla.s.ses, the Ghiordes and the Sehna. The Ghiordes are found in all rugs of Asia Minor and Caucasia, in some of the rugs of India, and in most of the rugs of Persia. They are named after the town of Ghiordes in Asia Minor, where some of the finest Asiatic pieces were made, and which tradition states was once the ancient Gordion, noted even in the days of Alexander. In tying the knot, the two ends of yarn appear together at the surface included between two[6] adjacent threads of warp around which they have been pa.s.sed, so that the tighter the yarn is drawn the more compact the knot becomes. The three different ways of tying this knot are shown in Plate B, Figs. 1, 2, and 3 (Page 49), of which the second is known as a "right hand" and the third as a "left hand" knot. The Sehna knots, which are used in the Turkoman, Chinese, many of the Persian, and in some of the Indian rugs, take their name from the city of Sehna in Persia. In tying them, a piece of yarn encircles a thread of warp and is twisted so that its ends appear at the surface, one at each side of the adjacent thread of warp, as is shown in Plate B, Figs. 4, 5, and 6. According as this thread of warp is to the right or the left of the one they encircle, the knots are known as "right-hand" or "left-hand" knots,[7] but in the appearance of the carpet there is no distinction. If the pile of a rug is carefully parted, the two ends of yarn forming a Sehna knot can be separated; but with the Ghiordes knot this is impossible, as will be understood by studying Plate B, in which Figs. 1, 2, 3, 7, 8, and 9 are Ghiordes knots, and Figs. 4, 5, 6, and 10 are Sehna knots. As a rule, the Sehna knots, which permit of closer weaving and clearer definition of pattern, appear in rugs of shorter nap.
[Ill.u.s.tration: PLATE B.-Nos. 1, 2, 3, 7, 8, 9, Ghiordes knots. Nos. 4, 5, 6, 10, Sehna knots. No. 11, Weft-overcasting. No. 12, Double-overcasting. No. 13, Weft-selvage. No. 14, Double-selvage. Nos.
15, 16, 17, ill.u.s.trate one, two, and three threads of weft pa.s.sing between two rows of knots.]
The nice distinctions in the technique of weaving are rarely understood even by those who are familiar with Oriental rugs. The general pattern, which next to colour is the characteristic that most quickly arrests the attention, is often the sole guide by which novices guess the cla.s.s. The more experienced will observe if the knot be Ghiordes or Sehna, and examine the finish at the sides and ends; but few give the peculiarities of the weave the consideration they deserve. This, perhaps, is because only those who have made a special study would believe the constancy with which members of a tribe or locality have followed the same method of tying the knot and inserting the weft. The different methods of treatment by separate tribes are sometimes only slight, but they afford a most important clue for determining the place of origin of doubtful cla.s.ses. In fact, nearly every cla.s.s has a typical weave differentiating it from all other cla.s.ses. To be sure, there are exceptions to the established type which are inevitable; since, for instance, a man from the Feraghan district might marry a woman from the adjoining Hamadan district, who, to please her husband, might weave a rug with pattern common to his district but follow the style of weaving that she has been familiar with from childhood. Nevertheless, weavers of a particular district adhere more closely to a typical style of weaving than they do to any other characteristic of a rug. Nor is this surprising, since weaving is learned in earliest childhood; and as it contains no elements calculated to stimulate the imagination, it is mechanically followed with stereotyped precision. An innovation in pattern, by copying some strange designs that strike the fancy, is far more likely. These distinctions in weaving may be conveniently divided into those that affect the knot, the warp, and the weft.[8]
THE KNOT.-Not only may a knot be tied as a Ghiordes or a Sehna knot, but it may have other distinguishing peculiarities; as, for instance, it may be of fine or coa.r.s.e wool. This is most conveniently observed at the back, where it will be seen that the knots of rugs such as the Bokhara, Kirman, Jos.h.a.ghan, and Bergamo are tied with fine yarn; while the knots of other rugs, as the Samarkand, Bijar, Gorevan, Kurdistan, Yuruk, and Kazak are tied with coa.r.s.e yarn. Whether the yarn be fine, medium, or coa.r.s.e, all specimens of any cla.s.s will show a remarkable conformity.
Also in some rugs the knots are drawn very tight against the warp, while in others the yarn encircles the warp loosely. Any one who has examined the back of many Sarouks, Kashans, Kirmans, or Daghestans, and rubbed the finger-nail against them, could not possibly mistake them for a Shiraz, Kulah, Yuruk, or Karabagh, which are less tightly woven. Again, as a result of using yarn in which the double thread that forms two or more ply has been very loosely or very tightly twisted together, there is some difference in the direction or slant of the strands forming the yarn, where it shows at the back, though this feature is not p.r.o.nounced.
For example, in most Afghans, Yuruks, Bijars, and others the strands of yarn where it crosses the warp in forming the knot lie for the most part in a direction parallel to the weft; while in other rugs, as Mosuls, Kurdistans, and Kazaks, the strands of yarn slant irregularly.
Furthermore, in some rugs, as the Melez and Yuruks, as a result of the threads of yarn being strung rather far apart, each half of a knot encircling a thread of warp stands out at the back distinctly from the other with clear cut edges; while in many rugs, as the Shiraz or Sehna, each half is very closely pressed together. Also in some rugs, as Sarabends and Afshars, each of these half knots where they show at the back have the same length, measured in a direction parallel to the warp, as width, measured in a direction parallel to the weft; while in such rugs as the Kazaks, since the yarn generally consists of several ply, the length exceeds the width; and in a few rugs the length is less than the width.