SMITHSONIAN INSTITUTION--BUREAU OF ETHNOLOGY. NAVAJO WEAVERS. BY DR. WASHINGTON MATTHEWS, U. S. A. Third Annual Report of the Bureau of Ethnologyto the Secretary of the Smithsonian Institution, 1881-'82, Government Printing Office, Washington, 1884, pages 371-392. ILLUSTRATIONS. PLATE XXXIV. --Navajo woman spinning 376 XXXV. --Weaving of diamond-shaped diagonals 380 XXXVI. --Navajo woman weaving a belt 384 XXXVII. --Zuņi women weaving a belt 388 XXXVIII. --Bringing down the batten 390 FIG. 42. --Ordinary Navajo blanket loom 378 43. --Diagram showing formation of warp 379 44. --Weaving of saddle-girth 382 45. --Diagram showing arrangement of threads of the warp in the healds and on the rod 383 46. --Weaving of saddle-girth 383 47. --Diagram showing arrangement of healds in diagonal weaving 384 48. --Diagonal cloth 384 49. --Navajo blanket of the finest quality 385 50. --Navajo blankets 386 51. --Navajo blanket 386 52. --Navajo blanket 387 53. --Navajo blanket 387 54. --Part of Navajo blanket 388 55. --Part of Navajo blanket 388 56. --Diagram showing formation of warp of sash 388 57. --Section of Navajo belt 389 58. --Wooden heald of the Zuņis 389 59. --Girl weaving (from an Aztec picture) 391 NAVAJO WEAVERS. BY DR. WASHINGTON MATTHEWS. § I. The art of weaving, as it exists among the Navajo Indians of NewMexico and Arizona, possesses points of great interest to the studentof ethnography. It is of aboriginal origin; and while European art hasundoubtedly modified it, the extent and nature of the foreigninfluence is easily traced. It is by no means certain, still there aremany reasons for supposing, that the Navajos learned their craft fromthe Pueblo Indians, and that, too, since the advent of the Spaniards;yet the pupils, if such they be, far excel their masters to-day in thebeauty and quality of their work. It may be safely stated that with nonative tribe in America, north of the Mexican boundary, has the art ofweaving been carried to greater perfection than among the Navajos, while with none in the entire continent is it less Europeanized. As inlanguage, habits, and opinions, so in arts, the Navajos have been lessinfluenced than their sedentary neighbors of the pueblos by thecivilization of the Old World. The superiority of the Navajo to the Pueblo work results not only froma constant advance of the weaver's art among the former, but from aconstant deterioration of it among the latter. The chief cause of thisdeterioration is that the Pueblos find it more remunerative to buy, atleast the finer _serapes_, from the Navajos, and give their time toother pursuits, than to manufacture for themselves; they are nearerthe white settlements and can get better prices for their produce;they give more attention to agriculture; they have within theircountry, mines of turquoise which the Navajos prize, and they have notrouble in procuring whisky, which some of the Navajos prize even morethan gems. Consequently, while the wilder Indian has incentives toimprove his art, the more advanced has many temptations to abandon italtogether. In some pueblos the skill of the loom has been almostforgotten. A growing fondness for European clothing has also had itsinfluence, no doubt. § II. Cotton, which grows well in New Mexico and Arizona, the toughfibers of yucca leaves and the fibers of other plants, the hair ofdifferent quadrupeds, and the down of birds furnished in prehistoricdays the materials of textile fabrics in this country. While some ofthe Pueblos still weave their native cotton to a slight extent, theNavajos grow no cotton and spin nothing but the wool of the domesticsheep, which animal is, of course, of Spanish introduction, and ofwhich the Navajos have vast herds. The wool is not washed until it is sheared. At the present time it iscombed with hand cards purchased from the Americans. In spinning, thesimplest form of the spindle--a slender stick thrust through thecenter of a round wooden disk--is used. The Mexicans on the Rio Grandeuse spinning-wheels, and although the Navajos have often seen thesewheels, have had abundant opportunities for buying and stealing them, and possess, I think, sufficient ingenuity to make them, they havenever abandoned the rude implement of their ancestors. Plate XXXIVillustrates the Navajo method of handling the spindle, a methoddifferent from that of the people of Zuņi. They still employ to a great extent their native dyes: of yellow, reddish, and black. There is good evidence that they formerly had ablue dye; but indigo, originally introduced, I think, by the Mexicans, has superseded this. If they, in former days, had a native blue and anative yellow, they must also, of course, have had a green, and theynow make green of their native yellow and indigo, the latter being theonly imported dye-stuff I have ever seen in use among them. Besidesthe hues above indicated, this people have had, ever since theintroduction of sheep, wool of three different natural colors--white, rusty black, and gray--so they had always a fair range of tints withwhich to execute their artistic designs. The brilliant red figures intheir finer blankets were, a few years ago, made entirely of _bayeta_, and this material is still largely used. Bayeta is a bright scarletcloth with a long nap, much finer in appearance than the scarletstrouding which forms such an important article in the Indian trade ofthe North. It was originally brought to the Navajo country fromMexico, but is now supplied to the trade from our eastern cities. TheIndians ravel it and use the weft. While many handsome blankets arestill made only of the colors and material above described, Americanyarn has lately become very popular among the Navajos, and many fineblankets are now made wholly, or in part, of Germantown wool. The black dye mentioned above is made of the twigs and leaves of thearomatic sumac (_Rhus aromatica_), a native yellow ocher, and the gumof the piņon (_Pinus edulis_). The process of preparing it is asfollows: They put into a pot of water some of the leaves of the sumac, and as many of the branchlets as can be crowded in without muchbreaking or crushing, and the water is allowed to boil for five or sixhours until a strong decoction is made. While the water is boilingthey attend to other parts of the process. The ocher is reduced to afine powder between two stones and then slowly roasted over the firein an earthen or metal vessel until it assumes a light-brown color; itis then taken from the fire and combined with about an equal quantityin size of piņon gum; again the mixture is put on the fire andconstantly stirred. At first the gum melts and the whole mass assumesa mushy consistency; but as the roasting progresses it graduallybecomes drier and darker until it is at last reduced to a fine blackpowder. This is removed from the fire, and when it has cooledsomewhat it is thrown into the decoction of sumac, with which itinstantly forms a rich, blue-black fluid. This dye is essentially anink, the tannic acid of the sumac combining with the sesquioxide ofiron in the roasted ocher, the whole enriched by the carbon of thecalcined gum. [Illustration: PL. XXXIV. --NAVAJO WOMAN SPINNING. ] There are, the Indians tell me, three different processes for dyeingyellow; two of these I have witnessed. The first process is thusconducted: The flowering tops of _Bigelovia graveolens_ are boiled forabout six hours until a decoction of deep yellow color is produced. When the dyer thinks the decoction strong enough, she heats over thefire in a pan or earthen vessel some native almogen (an impure nativealum), until it is reduced to a somewhat pasty consistency; this sheadds gradually to the decoction and then puts the wool in the dye toboil. From time to time a portion of the wool is taken out andinspected until (in about half an hour from the time it is firstimmersed) it is seen to have assumed the proper color. The work isthen done. The tint produced is nearly that of lemon yellow. In thesecond process they use the large, fleshy root of a plant which, as Ihave never yet seen it in fruit or flower, I am unable to determine. The fresh root is crushed to a soft paste on the _metate_, and, for amordant, the almogen is added while the grinding is going on. The coldpaste is then rubbed between the hands into the wool. If the wool doesnot seem to take the color readily a little water is dashed on themixture of wool and paste, and the whole is very slightly warmed. Theentire process does not occupy over an hour and the result is a colormuch like that now known as "old gold. " The reddish dye is made of the bark of _Alnus incana_ var. _virescens_(Watson) and the bark of the root of _Cercocarpus parvifolius_; themordant being fine juniper ashes. On buckskin this makes a brillianttan-color; but applied to wool it produces a much paler tint. § III. Plate XXXVIII and Fig. 42 illustrate ordinary blanket-looms. Two posts, _a a_, are set firmly in the ground; to these are lashedtwo cross-pieces or braces, _b c_, the whole forming the frame of theloom. Sometimes two slender trees, growing at a convenient distancefrom one another, are made to answer for the posts, _d_ is ahorizontal pole, which I call the supplementary yarn-beam, attached tothe upper brace, _b_, by means of a rope, _e e_, spirally applied. _f_is the upper beam of the loom. As it is analogous to the yarn-beam ofour looms, I will call it by this name, although once only have I seenthe warp wound around it. It lies parallel to the pole _d_, about 2 or3 inches below it, and is attached to the latter by a number of loops, _g g_. A spiral cord wound around the yarn-beam holds the upper bordercord _h h_, which, in turn, secures the upper end of the warp _i i_. The lower beam of the loom is shown at _k_. I will call this thecloth-beam, although the finished web is never wound around it; it istied firmly to the lower brace, _c_, of the frame, and to it issecured the lower border cord of the blanket. The original distancebetween the two beams is the length of the blanket. Lying between thethreads of the warp is depicted a broad, thin, oaken stick, _l_, whichI will call the batten. A set of healds attached to a heald-rod, _m_, are shown above the batten. These healds are made of cord or yarn;they include alternate threads of the warp, and serve when drawnforward to open the lower shed. The upper shed is kept patent by astout rod, _n_ (having no healds attached), which I name the shed-rod. Their substitute for the reed of our looms is a wooden fork, whichwill be designated as the reed-fork (Fig. 44, _a_). [Illustration: FIG. 42. --Ordinary Navajo blanket loom. ] For convenience of description, I am obliged to use the word"shuttle, " although, strictly speaking, the Navajo has no shuttle. Ifthe figure to be woven is a long stripe, or one where the weft must bepassed through 6 inches or more of the shed at one time, the yarn iswound on a slender twig or splinter, or shoved through on the end ofsuch a piece of wood; but where the pattern is intricate, and the weftpasses at each turn through only a few inches of the shed, the yarn iswound into small skeins or balls and shoved through with the finger. § IV. The warp is thus constructed: A frame of four sticks is made, not unlike the frame of the loom, but lying on or near the ground, instead of standing erect. The two sticks forming the sides of theframe are rough saplings or rails; the two forming the top and bottomare smooth rounded poles--often the poles which afterwards serve asthe beams of the loom; these are placed parallel to one another, theirdistance apart depending on the length of the projected blanket. On these poles the warp is laid in a continuous string. It is firstfirmly tied to one of the poles, which I will call No. 1 (Fig. 43);then it is passed over the other pole, No. 2, brought back under No. 2and over No. 1, forward again under No. 1 and over No. 2, and so on tothe end. Thus the first, third, fifth, &c. , turns of the cord cross inthe middle the second, fourth, sixth, &c. , forming a series ofelongated figures 8, as shown in the following diagram-- [Illustration: FIG. 43. --Diagram showing formation of warp. ] and making, in the very beginning of the process, the two sheds, whichare kept distinct throughout the whole work. When sufficient stringhas been laid the end is tied to pole No. 2, and a rod is placed ineach shed to keep it open, the rods being afterwards tied together atthe ends to prevent them from falling out. This done, the weaver takes three strings (which are afterwardstwilled into one, as will appear) and ties them together at one end. She now sits outside one of the poles, looking towards the centre ofthe frame, and proceeds thus: (1) She secures the triple cord to thepole immediately to the left of the warp; (2) then she takes one ofthe threads (or strands as they now become) and passes it under thefirst turn of the warp; (3) next she takes a second strand, andtwilling it once or oftener with the other strands, includes with itthe second bend of the warp; (4) this done, she takes the third strandand, twilling it as before, passes it under the third bend of thewarp, and thus she goes on until the entire warp in one place issecured between the strands of the cord; (5) then she pulls the stringto its fullest extent, and in doing so separates the threads of thewarp from one another; (6) a similar three stranded cord is applied tothe other end of the warp, along the outside of the other pole. At this stage of the work these stout cords lie along the outersurfaces of the poles, parallel with the axes of the latter, but whenthe warp is taken off the poles and applied to the beams of the loomby the spiral thread, as above described, and as depicted in PlateXXXVIII and Fig. 42, and all is ready for weaving, the cords appear onthe inner sides of the beams, _i. E. _, one (Pl. XXXVIII and Fig. 42, _hh_) at the lower side of the yarn-beam, the other at the upper side ofthe cloth-beam, and when the blanket is finished they form the stoutend margins of the web. In the coarser grade of blankets the cords areremoved and the ends of the warp tied in pairs and made to form afringe. (See Figs. 54 and 55. ) When the warp is transferred to the loom the rod which was placed inthe upper shed remains there, or another rod, straighter andsmoother, is substituted for it; but with the lower shed, healds areapplied to the anterior threads and the rod is withdrawn. § V. The mode of applying the healds is simple: (1) the weaver sitsfacing the loom in the position for weaving; (2) she lays at the right(her right) side of the loom a ball of string which she knows containsmore than sufficient material to make the healds; (3) she takes theend of this string and passes it to the left through the shed, leavingthe ball in its original position; (4) she ties a loop at the end ofthe string large enough to admit the heald-rod; (5) she holdshorizontally in her left hand a straightish slender rod, which is tobecome the heald-rod--its right extremity touching the left edge ofthe warp--and passes the rod through the loop until the point of thestick is even with the third (second anterior from the left) thread ofthe warp; (6) she puts her finger through the space between the firstand third threads and draws out a fold of the heald-string; (7) shetwists this once around, so as to form a loop, and pushes the point ofthe heald-rod on to the right through this loop; (8) she puts herfinger into the next space and forms another loop; (9) and so on shecontinues to advance her rod and form her loops from left to rightuntil each of the anterior (alternate) warp-threads of the lower shedis included in a loop of the heald; (10) when the last loop is madeshe ties the string firmly to the rod near its right end. When the weaving is nearly done and it becomes necessary to remove thehealds, the rod is drawn out of the loops, a slight pull is made atthe thread, the loops fall in an instant, and the straightened stringis drawn out of the shed. Illustrations of the healds may be seen inPlates XXXV and XXXVIII and Figs. 42, 44, and 46, that in Fig. 46being the most distinct. § VI. In making a blanket the operator sits on the ground with herlegs folded under her. The warp hangs vertically before her, and(excepting in a case to be mentioned) she weaves from below upwards. As she never rises from this squatting posture when at work, it isevident that when she has woven the web to a certain height furtherwork must become inconvenient or impossible unless by some arrangementthe finished web is drawn downwards. Her cloth-beam does not revolveas in our looms, so she brings her work within easy reach by thefollowing method: The spiral rope (Plate XXXVIII and Fig. 42) isloosened, the yarn-beam is lowered to the desired distance, a fold ismade in the loosened web, and the upper edge of the fold is sewed downtightly to the cloth-beam. In all new blankets over two feet long themarks of this sewing are to be seen, and they often remain until theblanket is worn out. Plate XXXV, representing a blanket nearlyfinished, illustrates this procedure. Except in belts, girths, and perhaps occasionally in very narrowblankets, the shuttle is never passed through the whole width of thewarp at once, but only through a space which does not exceed thelength of the batten; for it is by means of the batten, which israrely more than 3 feet long, that the shed is opened. [Illustration: PL. XXXV. --WEAVING OF DIAMOND-SHAPED DIAGONALS. ] Suppose the woman begins by weaving in the lower shed. She drawsapportion of the healds towards her, and with them the anteriorthreads of the shed; by this motion she opens the shed about 1 inch, which is not sufficient for the easy passage of the woof. She insertsher batten edgewise into this opening and then turns it half around onits long axis, so that its broad surfaces lie horizontally; in thisway the shed is opened to the extent of the width of the batten--about3 inches; next the weft is passed through. In fig. 42 the batten isshown lying edgewise (its broad surfaces vertical), as it appears whenjust inserted into the shed, and the weft, which has been passedthrough only a portion of the shed, is seen hanging out with its endon the ground. In Plate XXXV the batten is shown in the secondposition described, with the shed open to the fullest extentnecessary, and the weaver is represented in the act of passing theshuttle through. When the weft is in, it is shoved down into itsproper position by means of the reed-fork, and then the batten, restored to its first position (edgewise), is brought down with firmblows on the weft. It is by the vigorous use of the batten that theNavajo serapes are rendered water-proof. In Plate XXXVIII the weaveris seen bringing down this instrument "in the manner and for thepurpose described, " as the letters patent say. When the lower shed has received its thread of weft the weaver opensthe upper shed. This is done by releasing the healds and shoving theshed-rod down until it comes in contact with the healds; this opensthe upper shed down to the web. Then the weft is inserted and thebatten and reed-fork used as before. Thus she goes on with each shedalternately until the web is finished. It is, of course, desirable, at least in handsome blankets ofintricate pattern, to have both ends uniform even if the figure be alittle faulty in the center. To accomplish this some of the bestweavers depend on a careful estimate of the length of each figurebefore they begin, and weave continuously in one direction; but themajority weave a little portion of the upper end before they finishthe middle. Sometimes this is done by weaving from above downwards; atother times it is done by turning the loom upside down and workingfrom below upwards in the ordinary manner. In Fig. 49, whichrepresents one of the very finest results of Navajo work, by the bestweaver in the tribe, it will be seen that exact uniformity in the endshas not been attained. The figure was of such a nature that theblanket had to be woven in one direction only. I have described how the ends of the blanket are bordered with a stoutthree-ply string applied to the folds of the warp. The lateral edgesof the blanket are similarly protected by stout cords applied to theweft. The way in which these are woven in, next demands our attention. Two stout worsted cords, tied together, are firmly attached at eachend of the cloth-beam just outside of the warp; they are then carriedupwards and loosely tied to the yarn-beam or the supplementaryyarn-beam. Every time the weft is turned at the edge these two stringsare twisted together and the weft is passed through the twist; thusone thread or strand of this border is always on the outside. As it isconstantly twisted in one direction, it is evident that, after awhile, a counter-twist must form which would render the passage of theweft between the cords difficult, if the cords could not be untwistedagain. Here the object of tying these cords loosely to one of theupper beams, as before described, is displayed. From time to time thecords are untied and the unwoven portion straightened as the workprogresses. Fig. 44 and Plate XXXVIII show these cords. The coarseblankets do not have them. (Fig 42. ) Navajo blankets are single-ply, with designs the same on both sides, no matter how elaborate these designs may be. To produce theirvarigated patterns they have a separate skein, shuttle, or thread foreach component of the pattern. Take, for instance, the blanketdepicted in Fig. 49. Across this blanket, between the points _a--b_, we have two serrated borders, two white spaces, a small diamond in thecenter, and twenty-four serrated stripes, making in all twenty-ninecomponent parts of the pattern. Now, when the weaver was working inthis place, twenty-nine different threads of weft might have been seenhanging from the face of the web at one time. In the girth pictured inFig. 44 five different threads of woof are shown depending from theloom. [Illustration: FIG. 44. --Weaving of saddle-girth. ] When the web is so nearly finished that the batten can no longer beinserted in the warp, slender rods are placed in the shed, while theweft is passed with increased difficulty on the end of a delicatesplinter and the reed-fork alone presses the warp home. Later itbecomes necessary to remove even the rod and the shed; then thealternate threads are separated by a slender stick worked in tediouslybetween them, and two threads of woof are inserted--one above and theother below the stick. The very last thread is sometimes put in with adarning needle. The weaving of the last three inches requires morelabor than any foot of the previous work. In Figs. 49, 50, 51, 52, and 53 it will be seen that there are smallfringes or tassels at the corners of the blankets; these are made ofthe redundant ends of the four border-cords (_i. E. _, the portions ofthe cord by which they were tied to the beams), either simply tiedtogether or secured in the web with a few stitches. The above is a description of the simplest mechanism by which theNavajos make their blankets; but in manufacturing diagonals, sashes, garters, and hair-bands the mechanism is much more complicated. § VII. For making diagonals the warp is divided into four sheds; theuppermost one of these is provided with a shed-rod, the others aresupplied with healds. I will number the healds and sheds from belowupwards. The following diagram shows how the threads of the warp arearranged in the healds and on the rod. [Illustration: FIG. 45. --Diagram showing arrangement of threads of thewarp in the healds and on the rod. ] [Illustration: FIG. 46. --Weaving of saddle-girth. ] When the weaver wishes the diagonal ridges to run upwards from rightto left, she opens the sheds in regular order from below upwards thus:First, second, third, fourth, first, second, third, fourth, &c. Whenshe wishes the ridges to trend in the contrary direction she opens thesheds in the inverse order. I found it convenient to take myillustrations of this mode of weaving from a girth. In Figs. 44 and 46the mechanism is plainly shown. The lowest (first) shed is opened andthe first set of healds drawn forward. The rings of the girth take theplace of the beams of the loom. There is a variety of diagonal weaving practiced by the Navajos whichproduces diamond figures; for this the mechanism is the same as thatjust described, except that the healds are arranged differently on thewarp. The following diagram will explain this arrangement. [Illustration: FIG. 47. --Diagram showing arrangement of helds indiagonal weaving. ] To make the most approved series of diamonds the sheds are openedtwice in the direct order (_i. E. _, from below upwards) and twice inthe inverse order, thus: First, second, third, fourth, first, second, third, fourth, third, second, first, fourth, third, second, first, fourth, and so on. If this order is departed from the figures becomeirregular. If the weaver continues more than twice consecutively ineither order, a row of V-shaped figures is formed, thus: VVVV. PlateXXXV represents a woman weaving a blanket of this pattern, and Fig. 48shows a portion of a blanket which is part plain diagonal and partdiamond. [Illustration: FIG. 48. --Diagonal cloth. ] § VIII. I have heretofore spoken of the Navajo weavers always as ofthe feminine gender because the large majority of them are women. There are, however, a few men who practice the textile art, and amongthem are to found the best artisans in the tribe. [Illustration: PL. XXXVI. --NAVAJO WOMAN WEAVING A BELT. ] § IX. Navajo blankets represent a wide range in quality and finish andan endless variety in design, notwithstanding that all their figuresconsist of straight lines and angles, no curves being used. Asillustrating the great fertility of this people in design I have torelate that in the finer blankets of intricate pattern out ofthousands which I have examined, I do not remember to have ever seentwo exactly alike. Among the coarse striped blankets there is greatuniformity. [Illustration: FIG. 49. --Navajo blanket of the finest quality. ] The accompanying pictures of blankets represent some in my privatecollection. Fig. 49 depicts a blanket measuring 6 feet 9 inches by 5feet 6 inches, and weighing nearly 6 pounds. It is made entirely ofGermantown yarn in seven strongly contrasting colors, and is the workof a man who is generally conceded to be the best weaver in the tribe. A month was spent in its manufacture. Its figures are mostly inserrated stripes, which are the most difficult to execute withregularity. I have heard that the man who wove this often draws hisdesigns on sand before he begins to work them on the loom. Fig. 50 _a_shows a blanket of more antique design and material. It is 6 feet 6inches by 5 feet 3 inches, and is made of native yarn and _bayeta_. Its colors are black, white, dark-blue, red (_bayeta_) and--in aportion of the stair-like figures--a pale blue. Fig. 50 _b_ depicts atufted blanket or rug, of a kind not common, having much theappearance of an Oriental rug; it is made of shredded red flannel, with a few simple figures in yellow, dark blue, and green. Fig. 51represents a gaudy blanket of smaller size (5 feet 4 inches by 3 feet7 inches) worn by a woman. Its colors are yellow, green, dark blue, gray, and red, all but the latter color being in native yarn. Figs. 52and 53 illustrate small or half-size blankets made for children'swear. Such articles are often used for saddle blankets (although thesaddle-cloth is usually of coarser material) and are in great demandamong the Americans for rugs. Fig. 53 has a regular border of uniformdevice all the way around--a very rare thing in Navajo blankets. Figs. 54 and 55 show portions of coarse blankets made more for use use thanornament. Fig. 55 is made of loosely-twilled yarn, and is very warmbut not water-proof. Such blankets make excellent bedding for troopsin the field. Fig. 54 is a water-proof _serape_ of well-twilled nativewool. [Illustration: FIG. 50. --Navajo blankets. ] [Illustration: FIG. 51. --Navajo blanket. ] [Illustration: FIG. 52. --Navajo blanket. ] [Illustration: FIG. 53. --Navajo blanket. ] [Illustration: FIG. 54. --Part of Navajo blanket. ] [Illustration: FIG. 55. --Part of Navajo blanket. ] The aboriginal woman's dress is made of two small blankets, equal insize and similar in design, sewed together at the sides, withapertures left for the arms and no sleeves. It is invariably woven inblack or dark-blue native wool with a broad variegated stripe in redimported yarn or red _bayeta_ at each end, the designs being ofcountless variety. Plates XXXIV and XXXV represent women wearing suchdresses. [Illustration: FIG. 56. --Diagram showing formation of warp of sash. ] [Illustration: PL. XXXVII. --ZUŅI WOMAN WEAVING A BELT. ] § X. Their way of weaving long ribbon-like articles, such as sashes orbelts, garters, and hair-bands, which we will next consider, presentsmany interesting variations from, the method pursued in makingblankets. To form, a sash the weaver proceeds as follows: She drivesinto the ground four sticks and on them she winds her warp as acontinuous string (however, as the warp usually consists of threads ofthree different colors it is not always _one_ continuous string) from, below upwards in such a way as to secure two sheds, as shown in thediagram, Fig. 56. [Illustration: FIG. 57. --Section of Navajo belt. ] [Illustration: FIG. 58. --Wooden heald of the Zuņis. ] Every turn of the warp passes over the sticks _a_, and _b_; but it isalternate turns that pass over _c_ and _d_. When the warp is laid sheties a string around the intersection of the sheds at _e_, so as tokeep the sheds separate while she is mounting the warp on the beams. She then places the upper beam of the loom in the place of the stick_b_ and the lower beam in the place of the stick _a_. Sometimes theupper and lower beams are secured to the two side rails forming aframe such as the warp of a blanket is wound on (§ IV), but morecommonly the loom is arranged in the manner shown in Plate XXXVI; thatis, the upper beam is secured to a rafter, post, or tree, while to thelower beam is attached a loop of rope that passes under the thighs ofthe weaver, and the warp is rendered tense by her weight. Next, theupper shed is supplied with a shed-rod, and the lower shed with a setof healds. Then the stick at _f_ (upper stick in Plate XXXVI) is putin; this is simply a round stick, about which one loop of each threadof the warp is thrown. (Although the warp may consist of only onethread I must now speak of each turn as a separate thread. ) Its use isto keep the different threads in place and prevent them from crossingand straggling; for it must be remembered that the warp in this caseis not secured at two points between three stranded cords as is theblanket warp. When this is all ready the insertion of the weft begins. The reed-forkis rarely needed and the batten used is much shorter than thatemployed in making blankets. Fig. 57 represents a section of a belt. It will be seen that the center is ornamented with peculiar raisedfigures; these are made by inserting a slender stick into the warp, soas to hold up certain of the threads while the weft is passed twice oroftener underneath them. It is practically a variety of damask ortwo-ply weaving; the figures on the opposite side of the belt beingdifferent. There is a limited variety of these figures. I think I haveseen about a dozen different kinds. The experienced weaver is so wellacquainted with the "count" or arrangements of the raised threadsappropriate to each pattern that she goes on inserting and withdrawingthe slender stick referred to without a moment's hesitation, makingthe web at the rate of 10 or 12 inches an hour. When the web has grownto the point at which she cannot weave it further without bringing theunfilled warp nearer to her, she is not obliged to resort to theclumsy method used with blankets. She merely seizes the anterior layerof the warp and pulls it down towards her; for the warp is notattached to the beams, but is movable on them; in other words, whilestill on the loom the belt is endless. When all the warp has beenfilled except about one foot, the weaving is completed; for then theunfilled warp is cut in the center and becomes the terminal fringes ofthe now finished belt. The only marked difference that I have observed between the mechanicalappliances of the Navajo weaver and those of her Pueblo neighbor is tobe seen in the belt loom. The Zuņi woman lays out her warp, not as acontinuous thread around two beams, but as several disunited threads. She attaches one end of these to a fixed object, usually a rafter inher dwelling, and the other to the belt she wears around her body. Shehas a set of wooden healds by which she actuates the alternate threadsof the warp. Instead of using the slender stick of the Navajos toelevate the threads of the warp in forming her figures, she liftsthese threads with her fingers. This is an easy matter with herstyle of loom; but it would be a very difficult task with that of theNavajos. Plate XXXVII represents a Zuņi woman weaving a belt. Thewooden healds are shown, and again, enlarged, in Fig. 58. The Zuņiwomen weave all their long, narrow webs according to the same system;but Mr. Bandelier has informed me that the Indians of the Pueblo ofCochiti make the narrow garters and hair-bands after the manner of theZuņis, and the broad belts after the manner of the Navajos. [Illustration: PL. XXXVIII. --BRINGING DOWN THE BATTEN. ] [Illustration: FIG. 59. --Girl weaving (from an Aztec picture). ] § XI. I will close by inviting the reader to compare Plate XXXVI andFig. 59. The former shows a Navajo woman weaving a belt; the latter agirl of ancient Mexico weaving a web of some other description. Theone is from a photograph, taken from life; the other I have copiedfrom Tylor's "Anthropology" (p. 248); but it appears earlier in thecopy of Codex Vaticana in Lord Kingsborough's "Antiquities of Mexico. "The way in which the warp is held down and made tense, by a rope orband secured to the lower beam and sat upon by the weaver, is the samein both cases. And it seems that the artist who drew the original rudesketch, sought to represent the girl, not as working "the cross-threadof the woof in and out on a stick, " but as manipulating the reed-forkwith one hand and grasping the heald-rod and shed-rod in the other. NOTE. --The engravings were prepared while the author was in New Mexico and could not be submitted for his inspection until the paper was ready for the press. Some alterations were made from the original pictures. The following are the most important to be noted: In Plate XXXVIII the batten should appear held horizontally, not obliquely. Fig. 5 is reduced and cannot fairly delineate the gradations in color and regular sharp outlines of the finely-serrated figures. Fig. 53 does not convey the fact that the stripes are of uniform width and all the right-angles accurately made. * * * * * INDEX Blankets, Navajo 380-388 Codex, The Vatican; Illustrating Mexican weaving 391 Colors prepared for Navajo fabrics 376 Cotton woven in Pueblos, Native 375 Dyeing among Navajoes 377 Dyes used by Navajoes 377 Fabrics; Prehistoric textiles of the United States 393-425 Healds of Navajo loom 378 Mode of applying the 380, 384 Zuņi 389 Looms, Navajo 377 Mathews, Dr. W. , Navajo weavers 371-391 Navajo blankets, Varieties of 385-388 Mode of weaving 383 diagonal 383 diamond 384 dyeing 377 dyes 376 healds in loom used 380 looms 377 position in weaving 380 warp of blankets 378-379 sash 388 weavers 371-391 wool 375 Taylor, E. B. , Anthropology cited 391 Warp, Construction of Navajo blanket 378 Warp, Construction of Navajo sash 388 Weavers, Navajo, by Dr. Washington Mathews 371-391 Weaving, Navajo position in 380 wool by Navajoes 375 Zuņi, healds 389 * * * * *