Library- of M CHICAGO Natural History Museum M.-' C-*Q**t1A**N X X X a; JS Oh o > >> bo o "o o. o c o Z E 3 - a: 3 s o z t— < OS o o < z >* ►J bj <: H < O H-t Oh Field Museum of Natural History Founded by Marshall Field, 1893 Anthropology, Memoirs Volume II, No. 3 ARCHAEOLOGICAL EXPLORATIONS IN PERU PART III TEXTILES OF THE EARLY NAZCA PERIOD BY LILA M. O'NEALE ASSOCIATE PROFESSOR OF HOUSEHOLD ART AND ASSOCIATE CURATOR OF TEXTILES IN THE MUSEUM OF ANTHROPOLOGY AT THE UNIVERSITY OF CALIFORNIA WITH INTRODUCTION BY A. L. KROEBER PROFESSOR OF ANTHROPOLOGY IN THE UNIVERSITY OF CALIFORNIA RESEARCH ASSOCIATE IN AMERICAN ARCHAEOLOGY IN FIELD MUSEUM 36 Plates, 2 Colored Plates, 1 Text Figure Second Marshall Field Archaeological Expedition to Peru PAUL S. MARTIN CHIEF CURATOR, DEPARTMENT OF ANTHROPOLOGY EDITOR Chicago, U.S.A. 1937 PRINTED IN THE UNITED STATES OK AMERICA BY FIELD MUSEUM PRESS CONTENTS PAGE List of Illustrations 12i Preface 127 Introduction 131 Garment Types 133 Early Nazca Colors 136 Seven Major Color Groups 146 Red-to-Orange Color Group 146 Orange-to- Yellow Color Group 148 Yellow-to-Green Color Group 148 Green-to-Blue-Green Color Group 149 Blue-Green-to-Blue Color Group 150 Blue-to-Red Color Group 150 Purple-to-Red Color Group 151 Early Nazca Yarns 153 Early Nazca Mantles 154 Mantle Dimensions 155 Mantle Warp-weft Yarn Counts 156 Correlation between Mantle Types and Warp-weft Counts 157 Mantle Yarns 158 Plain Weave Undecorated Mantles 159 Plain Weave Unidentifiable Garment Materials 159 Plain Weave Striped Mantles 160 Striped Cotton Mantles 160 Striped Wool Mantles 161 Striped Mantle Colors 164 Brocaded and Embroidered Mantles 164 Mantles with Needleknitted Edges 172 Unidentifiable Fragments 174 Colors in Needleknitting Specimens 175 Pattern Weave Mantles 178 Gauze Weave Mantles 179 Miscellaneous Mantle Types 180 Interlocking Warps and Wefts 180 Early Nazca Tunics 183 Early Nazca Kerchiefs and Veils 187 Early Nazca Aprons 191 The Miscellaneous Group 192 Bands 192 Woven Bands 193 Plaited Bands and Cords 195 Warp-twined Fabrications 196 Embroidered Bands and Garment Fragments 198 Cords 200 Slings 201 Netting 202 119 120 CONTENTS PAGE Feather Bands 202 Wrappings 202 Pads 203 Stone Wound with String 203 Summary 204 Garments 204 Textures 204 Decorative Features 204 Colors 205 Yarns and Technical Processes 205 Yarns 205 Warp-weft Techniques 205 Single-element Techniques 206 Superstructural Techniques 207 Stitchery 207 Devices to Vary Effect 207 Single-element Manipulation 208 Conclusions 213 Glossary 215 LIST OF ILLUSTRATIONS Specimens designated by numbers containing six figures are in Field Museum of Natural History; those of four figures are in the University of California Museum of Anthropology. The several sites at a locality in the Nazca Valley are distinguished by capital letters; subsites or portions of a cemetery by small letters following; graves by numbers. Thus, "Cahuachi, Ajl0-171218a" means part or fragment a of Field Museum specimen 171218 which was found in grave 10 of portion j of cemetery (site) A at Cahuachi, valley of Nazca. Superior numbers indicate addition to the original list (see explanation, page 131). All Field Museum specimens were excavated or found by A. L. Kroeber in 1925 and 1926; the California specimens were secured by Max Uhle in 1905. These latter (4), marked by an asterisk, are surface finds lacking precise provenience. Arrows by the specimen numbers on the plates denote the direction of the warp. PLATES XXXII. Typical Early Nazca Coloring; 171217 (in color). XXXIII. Typical Early Nazca Brocade and Embroidery; 171218a (in color). XXXIV. Early Nazca Period Garments, a, Cahuachi; Ha-171181; apron, 37}^" x 16lA"; plain weave, needleknitted bindings on string ends, b, Cahuachi; Aj 13-171262; diagram of mantle showing characteristic treatment of edges: side trimmings continued part way across ends, c, Cahuachi; Ajl3-171266b; child's tunic with embroidered bands (Plate LVIa, b, c). d, Cantayo; C-ax cache-171071d ; child's tunic with Kelim slit for neck opening, e, Cahuachi; A15-171308; tunic with checkerboard section in two colors; interlocking warps and wefts (Plate LXVIIgr). XXXV. Pads and a Headband(?). Cahuachi. a, AJ13-171280; girdle(?) or headband(?), 2y2" wide; plain weave, warp stripe (Plate XXXVIId). b, c, Ajll-171236; surface and reverse sides of pad stuffed with cotton, 8^" x 5^"; single-face pattern weave, warp and weft floats (Plate XLVIrf, e, /). d, Ag-171330; pad stuffed with cotton, 10" x 6", approximately. XXXVI. Striped Cotton Materials, a, b, c, Cahuachi; Ha-171183; A14-171305b; A15-171311. d, e, Majoro; A6-170465b, d; schematic representations of colored yarns in warp set-ups for cotton striped materials. XXXVII. Striped Wool Materials, a, Cantayo; C-ax cache-171071a». b, Majoro; A6-170462b. c, d, e, /, Cahuachi; 170211c12; AJ13-171280; Aj'13-171265; Aj 13-171262. g, Cantayo; C-ax 11-171033. h, Cahuachi; A15-171310. i, Ocongalla West; 6-170665; schematic representations of colored yarns in warp set-ups for wool materials, j, Cahuachi; A14-171305c; flat- view diagram of colored weft stripes. XXXVIII. Variations in Tapestry Technique, a, Cahuachi; AjlO-171225; band for kerchief, V/j' wide; Kelim tapestry woven on wefts which cross a space provided for in the warp set-up (Plate LXVIIId). b, Majoro; A6-170465e; band, Yi' wide; Kelim and eccentric tapestry weaving, c, d, Cahuachi; Ajl0-171217a; detail of lower tunic edge showing twining wefts, and figure-8 tapestry weaving on warp loops; fringed band sewn to tunic armscye, 9" overall; Kelim and eccentric tapestry weaving, separately woven and applied fringe. XXXIX. Interlocking in Plain Weaves. Cahuachi. a, c, Ag-171111; diagram of multicolored "patchwork" with stepped fret motives, 5" long, and reconstruction showing method of making (a). 6, d, AjlO-171221; diagram of mantle motive, 15" long; plain weave, interlocking wefts as shown in detail reconstruction (b). 121 122 LIST OF ILLUSTRATIONS XL. Woven Bands and Cords. Cahuachi. a, Ag-171118f2; turban band, %" wide; plain weave with double weft, b, AglO-171227; turban band, 14" wide; plain weave, c, /, Ag- 171109a1; turban band, }4" wide, and detail showing plain weave variation on two warps, d, e, Ha-171180c; tubular cord, %" wide, approximately; single-face pattern weave, underfloat warps. XLI. Gauze Weaves, a, c, Cahuachi; Ag-171110; gauze weave motives within a plain weave cotton fabric, 5J4" long; detail showing manipulation of warps in change from plain to gauze weave with return to original position on the sixth pick of weft, b, Majoro ; A7-170476f1; mantle fragment with attached band, 114" wide; plain weave with gauze motives and applied plaits; band in double-face pattern weave, warp floats. XLII. Plaited Bands and Cords, a, f, g, Majoro; A7-170476d; fragment of double-cloth band, 2J^" wide, with three 16-strand twine plaited ends (g) coming together as a 48-strand flat plait, 3A" wide (/). b, Majoro; A7-170476f; 4-strand flat plait, c, d, Cahuachi; Ea-171124b1, c1; 4-strand round plait, and 7-strand flat plait, e, h, Cahuachi; 170211d6; Ag-171118f2; 8-strand and 9-strand flat plaits. Plaits b, c, d, e, h, y% wide, approximately. XLIII. Pattern-weave Borders, a, c, Cantayo; C-ax 171059; flat-view diagram and design of wool kerchief band, 1" wide; single-face pattern weave, warp underfloats. b, Cahuachi ; Ag-171119; border design, y." wide; double-face pattern weave, weft floats (Plate XLVc). d, Cahuachi; Ag-171118e; design of wool kerchief band, 1}4" wide; single- face pattern weave, warp underfloats (Plate XLVa, b). e, Cahuachi; Aj 10-171226; design of wool kerchief band, V>/% wide; single-face pattern weave, warp underfloats (Plate XLIV6). XLIV. Flat Views of Pattern Weaves. Cahuachi. a, Ajl3-171279a; mantle material, crosses %", approximately; single-face pattern weave, warp and weft underfloats. b, AjlO- 171226; flat-view diagram of wool kerchief band, 1%" wide; single-face pattern weave, warp underfloats (Plate XLIIIe). XLV. Flat Views of Pattern Weaves. Cahuachi. a, b, Ag-171118e; flat-view diagrams of surface and reverse sides of wool kerchief band, 114" wide; single-face pattern weave, warp underfloats (Plate XLIIM). c, Ag-171119; flat-view diagram of mantle border, y;*«>" Reddish- n ' Greenish- Specimen nos. Light Medium Dark wmte Gray grange Blue 170465b 13 E 7 15 C 10 12 B 6 170465d 13 E 7 14 G8 12 B 6 13 A 4 171183 12 D 5 14 B 9 11 B 9 39 E 1 171305b 12 E 6 12 B 6 171311 12 E 6 13 E 7 15 E 11 The wool mantles in this small group are more colorful. Brown wool yarns were used in some garments, for instance, the tunics, but only one of the four mantles is woven with brown yarns. Disregarding the colors in the trimming bands, which will be noted when speaking of the needleknitting specimens, the dyed yarns of the basic webs strongly contrast with those in the cotton webs, probably Natural Brown in some cases. Wool mantles Specimen nos. Red Orange (Brown) Yellow Green Blue Purple Black 170665... 5K10 5 L 11 13 K 8 11 C 3 39 H 3 47 E 5 171262.. 5 L 7 5 L 11 Present 171265.. 5 L 7 5 L 11 Present 171310.. 8H10 14L10 15 A 10 BROCADED AND EMBROIDERED MANTLES The most obvious feature of the small group of mantles coming under this classification is that no two of them are similar in design or in patterning, only in coloring. Among them is Cahuachi mantle 171220, a specimen of such importance as to warrant detailed treatment of its several features.15* '** A detailed discussion of this specimen appears also under the title The Wide-Loom Fabrics of Nazca, Essays in Anthropology in honor of Alfred Louis Kroeber, Berkeley, 1936. EARLY NAZCA MANTLES 165 Mantle 171220 is represented by three large and two small fragments: Ajl0-171218a 38" x 48J^" Ajl0-171219a'-3 80^" X 40", 12" x 18", 3^" x 7" Aj 10-171220 26" x 33" For convenient reference within this section of the paper only, the three fragments of 171219a have been given superior numbers, but in speaking of the garment as a whole it is referred to by the last of its numbers, 171220 (Plate XXXIII). The five fragments, four of which are corners showing loom-string ends and side sel- vages, fit together to form a rectangle at least 6'8" long by 5'5" wide (Plate XLIX). If more evidence than the reconstructed shape were needed to prove that the remnants were originally one web, the similar yarn counts per inch, the identical design motives, and the colors of the decorative yarns could be cited. There is no doubt of the full width measure since there is no suspicion of lengthwise join. The two fragments 171218a and 171220 fit together with no inch of the fabric missing. Straight tears warpwise and weftwise through one of the rectangu- lar design motives leave portions of it on both fragments. In addition, there are the same dark and yellowish stains on both fragments. Any opinion concerning the loom type upon which a 5'5" breadth was woven must be speculative. No one has made a detailed study of the ancient Peruvian artifacts which might be interpreted as loom parts. The direct and indirect evidence to date indicates that the looms were of the type attached to the weaver's belt,16 but a belt loom for a single weaver seems a most dubious explanation for a single breadth 5'5" wide, and a totally inadequate one for two single-breadth Paracas fabrics 7'7" and 8'3" wide in the Museo Nacional collections in Lima.17 The main reasons why the ordinary belt loom type manipulated by a single weaver seem improbable explanations for wide fabrics like the three mentioned concern the limita- tions imposed by the loom itself. A weaver sitting at the center of a loom bar attached to her waist can weave a fabric very little wider than the length of her reach to right and left. She can, it is true, sway to each side, but the distance is limited by her position in the belt contrivance. Before she can enter the weft yarn at either side, the alternate warps of the set-up must be raised by some form of heddle. The kind found in the prehistoric graves is a type known in different parts of the world. It consists of a stick as long as the warp set-up is wide — this length is for fabrics of average width — from which cord loops drop to encircle each of the odd or even warps. A heddle stick 5'5" long with loops encircling the alternate warps would be no easy object to raise, especially since it should be inflexible, therefore of some weight. To envisage raising the whole group of encircled warps against their resistance, at the same time putting through the space so created a sword or batten to keep them separated, these two actions with the body tense, taxes the imagination of a modern craftsman. A sword at least 5'5" long, and of sufficient weight to be effective, could not but be unwieldy. I doubt if one person could manage both heddle and sword if they were as long as the Cahuachi set-up was wide. Perhaps, like the heddles on the looms for weaving pictorial tapestries, the Cahuachi heddles were in sections. And, perhaps, the swords were short and were pushed through the shed as the sections of the warp were raised by the heddles. To put the weft across a wide web is slow work but not difficult. Wound sticks, or spindles, and balls of weft yarns have been found. In either form the weft yarn is not a u Textile Periods, p. 29, footnote 14. 17 E. Yacovleff and J. C. Muelle, Un fardo funerario de Paracas, Revista del Museo Nacional, vol. Ill, nos. 1-2, pp. 77, 79-80, plate le, f, 1934. 166 TEXTILES OF THE EARLY NAZCA PERIOD problem. But battening each line of weft down to the already woven cloth depends again upon the length and type of sword. Maintaining even side selvages, a feature of all Peruvian weaving of whatever period, would also give trouble to the weaver working at a 5'5" breadth. Most modern craftsmen pull through the weft with one hand, at the same time holding with the other the opposite edge to prevent it from drawing in. To stretch across a set-up of 5'5" demands the full reach of a weaver of that height, and to make cloth by repeating the motion 26 times an inch for 6'8" seems pretty improbable. For the 7'7" and 8'3" materials reported from Paracas mummy bundles, a single weaver at a belt loom is an impossible assumption. Since any opinion to date regarding the loom type is speculative, one might imagine a loom at which it was customary for several weavers to work side by side as they did when weaving the Kashmere shawls, or as they do today in weaving the Oriental rugs. Then, too, there is always the possibility that the Early weavers had wide looms similar to the horizontal or the vertical frames in use today among certain African tribes. Possibly, like the ancient Egyptians, the Peruvians used both types and set two weavers at each.is The Lima investi- gators quoted above suggest "some form of large loom whose models have not come down to us for the reason that it was impossible to place the looms in the graves." An analysis of the length measurement of the Cahuachi mantle and the reason behind the assumption that 6'8" is the minimum also involves uncertainty. Specimens 171218a and 171219a1 are opposite corner pieces on the same side, and although the embroidered motives bordering the ends are different, the side borders are identical (Plate LIIc). The same sequence of four colors in the two halves of the design unit — (1) Blue, Green, Yellow- Orange; (2) Blue, Green, Red — is maintained through two complete repetitions, or four half-units, from the corner of specimen 171219a1. The cloth has disintegrated under the last two motives, leaving long loops of the Green yarn still interlocked at the change from Green to Red (Plate LXVIIIe), and the few remaining Red loops are only a fraction of their original length. On the other side of the break the sequence picks up at Blue and continues: Blue, Green, Red — the second half-unit of the repetition — and then unbrokenly to the corner of specimen 171218a where the sequence ends on Blue. In numerals, to represent Blue, Green, Yellow-Orange, Red, the sequence is as follows: 123-124; 123-124; cloth missing— 124; 123-124; 123-1. A determination of the original length of this mantle depends upon two factors in the analysis: the approximate length which seems in reasonable proportion to a mantle width of 5'5", and the clues furnished by the measurements of the border motives. These last are perhaps more tangible and may be considered first. Peruvian color sequences are in most cases consistently maintained. There is no reason to suppose that the destroyed portion of the original mantle between what is now represented by specimens 171218a and 171219a1 was any exception to that rule. There is no reason, either, to suppose that the lengths of the three embroidered elements in each half unit, totaling approximately 8^", should have appreciably varied in the missing portion. The mantle's total length would, then, have been equal to the fragmentary border lengths, 33^" (which includes the length of the fragmentary Red element, 3") and 38" plus the first half-unit of the sequence, the missing Blue, Green, Yellow-Orange motives, approxi- mately S}^", total 80". This is the minimum length. Of course, there is no arbitrary limit to the suggested length, but if the original mantle was longer than 6'8" then the amount must almost certainly have been governed by the combined lengths of the two half-units in the border, approximately 17"-18". In other words, adding IVA" to the lengths of the " H. Ling Roth, Ancient Egyptian and Greek Looms, figs. 1-11, 13, 14, 16, Halifax, 1913. EARLY NAZCA MANTLES 167 two intact borders provides for completion of the fragmentary Red motive and for the missing half-unit, Blue, Green, Yellow-Orange. Any alternative interpretation requires that whole units, each measuring 17"-18", be considered in addition to the 113^" which fills the gap in an otherwise perfect color sequence. From the standpoint of proportions, there is less material upon which to base estimates concerning the length of the original mantle. The Early Nazca collection contains ten mantles furnishing complete dimensions. Their lengths, ranging from 40" through 80" (?), and their widths, from 38" through 65", are tabulated in the section on Mantle Dimensions (Table 22). If we assume the original dimensions of mantle 171220 to have been 80" (333^" + 38"+ missing half -unit, 8%") by 65", the proportion of width to length is 0.812, placing the mantle third in the group of ten. If, however, we assume the original dimensions to have been 97" (33^"+ 38"+ missing half -unit, 8^",+ a complete unit, 17"+) by 65", the pro- portion of width to length is 0.67, first on the list. There is, in reality, no more foundation for one than the other assumption, except that the available mantles show lesser rather than greater differences between their two dimensions. It may be of interest at this point to note that the Cahuachi mantle seems to have conformed to a local style in the matter of proportions. The celebrated embroidered mantles from the Paracas Necropolis repre- senting a period approximately coeval in time19 are not only unlike the Early Nazca mantles in the Field Museum Collection in general appearance and decoration, but also in dimensions. A summary of three basic tables is here presented:20 Range in lengths of 67 Paracas Necropolis mantles: 83"-144". Range in widths of 53 Paracas Necropolis mantles: 31"-64". Range in proportions of widths to lengths of 53 Paracas mantles: 0.242-0.605. Even assuming a length measurement for mantle 171220 which would place its width- to-length proportion first in its group does not alter the fact that the Paracas mantles radi- cally differ in shape from the available Early Nazca examples. Technologically, Cahuachi mantle 171220 presents few unusual details. The basic weave is plain, and the texture has something in common with modern scrim except that the more tightly twisted yarns of the mantle account for a slightly creped surface. The warp-weft counts per inch vary sufficiently to indicate that hand weavers are pretty much alike in all times. The warps were set up 26, 28, and 30 per inch. Wefts count 24 and 26 per inch. The yarn is 2-ply white cotton, spun medium-to-hard twist. This Cahuachi mantle, like so many other textiles in Peruvian collections, brings up the ever-recurring question of whether a certain technique shall be classified as brocade or embroidery. The answer is made no easier by references to older writings in which are described brocades worked with the needle, as well as embroideries accomplished in the loom. Either description confuses by reason of the fact that brocading has been more or less firmly established as a weaving process, and embroidery has for ages been restricted to a method of patterning a fabric already woven. With the best intentions two English writers seek to clarify the situation in the following paragraph: "Many of the Peruvian patterns are woven by the method of brocading. This closely resembles a simple form of embroidery, and it is sometimes difficult to determine the one from the other, brocading being really a form of embroidery applied to weaving. In brocade weaving, the threads forming the pattern are inserted as an addition to the weft threads 19 Textile Periods, p. 25, Table I, Chronological Concordance of Periods. 20 Unpublished MS. 168 TEXTILES OF THE EARLY NAZCA PERIOD and in a line with them during the course of weaving, and this is done with a needle or some form of bobbin, the warp threads being so regulated in their use as to secure the bro- cading threads somewhat slackly at certain defined points in the pattern without them- selves being evident; there may be a special warp for this purpose. The brocading is therefore part of the process of weaving, but it has not the structural element of tapestry. The same process carried out by the needle on the woven fabric is true embroidery."21 The description is accurate and also adequate, if two details be added: first, that the threads forming the weft brocade pattern regularly alternate wherever used with the weft threads of the basic material; and, second, that the "certain defined points" at which the warp threads appear often in themselves make a secondary pattern on the brocading threads as well as "secure" them.22 In the analysis of these Cahuachi mantle fragments for the preliminary report23 the borders and rectangular motives were termed brocades. It is true that the basic fabric is a fairly coarse open-mesh material, and to count its threads in order to plan the geometric elements in the patterns implies no such skill as do other Peruvian techniques of this Early period, the needleknitted bird and flower fringes, for example. But, if one inclined to interpret mantle 171220 as a brocade, one might argue that the straight-line work of the pattern rectangles could be put in more easily with a shuttle weaving over and under stretched bare warps than with an embroidery needle subsequent to the weaving. This reasoning assumes a knowledge of what the Cahuachi weaver considered easy or convenient. And, too, if one of the heavier pattern yarns was never found to cross a basic weft which parallels it on either side, the brocade interpretation could stand against any argument for em- broidery. Perfectly done, embroidery in the brocade manner cannot be distinguished from brocade. The smaller motives which form the lower borders, the finial motives out from the corners of the rectangles, and the side borders with their yarns interlocking at color changes do not bear the stamp of the brocade technique, although it is conceivable that they might have been made by that method. The patterning of Cahuachi mantle 171220 is not the least interesting of its features. There are seemingly illogical differences in the four borders, inconsistencies in the arrange- ment of design forms, and unanswerable questions arising as to the original size of several fragmentary motives. Added to these aspects — and largely because of them — is the re- grettable lack of actual evidence for the patterning of the missing center third of the mantle. Plate XLIX shows schematically the proportions of the reconstructed garment, the placing of its various design areas indicated by capital letters. Each area is a composite of elements more or less similar in appearance, but not identical. The characteristic simple and elaborated elements forming the motives are the following: (1) Lozenge shapes: plain 4-sided diamonds of various sizes, with and without em- broidered center dots (Plate L6); angular figure-8 forms composed of two lozenge shapes, and elaborated lozenge shapes with wide framing lines and hexagonal center dots (Plate La). (2) Triangles: isosceles and equilateral, both with and without embroidered center dots (Plate Lb). (3) Chevrons: small simple forms as well as elaborated ones (Plate Lb). (4) Bird forms: in flight, profile view; in flight, top view (Plates LI, LI I); four heads (?) radiating from horizontal axis, and two heads (?) in profile, bodies forming sides " Mary Symonds and Louisa Preece, Needlework Through the Ages, p. 108. London, 1928. " Textile Periods, plate 23b, shows warp shedding for secondary design. " Idem, footnote 16. EARLY NAZCA MANTLES 169 of chevron (Plate La); bird (?) form with short tail feathers and very long tail feather (?) turned back under body (Plates LI, LI I). (5) Serpent forms: the bodies a composite of lozenges and triangles (Plate Lb); highly conventionalized form, possibly serpent (Plate LIIc, d). These design elements are variously combined in forming ten different border and basic web motives, each of which requires a separate description. The letters correspond to those on Plate XLIX. (A) A border on one long edge (specimens 171219a2 and 171220) composed of 2>£" to 3" units with half -inch spaces between them. Each space is broken by the bill of a bird in profile; alternate birds face in the same direction, a method of countering which seems never to have been practiced to any extent during the known Peruvian periods (Plate LI Id). Colors in repeated sequence, as given above: Blue, Green, Yellow-Orange; Blue, Green, Red. There is little doubt that the rectangular unit as well as the bird form is in embroidery technique. The mantle edge is thickened where several strands of wool yarns have been crowded in between the basic warp threads. The motives are all more or less alike, but what they were meant to represent is not clear. Perhaps they are double-headed serpents, a form to be seen in two of the large rectangles in the center of the web (Plate L6). Wherever the color change takes place, the individual embroidery yarns of the two involved sets interlock underneath the space occupied by the bird form as shown in detail drawing Plate LXVIIIe, /. (B) A border on opposite long edge formed by specimens 171218a and 171219a1. In color sequence and other respects this border is identical to border A except that the bird figures which fill the spaces between the rectangles are represented full top view instead of in profile (Plate LIIc). (C) An end border on fragments 171219a2 and a3. Plates LI6, Llle show the form and arrangement of a design element which may or may not be a bird. Note the apparent lack of system in either the arrangement of the motives in relation to each other, or in their relation to the border as a whole as evidenced by the sizes, spacing, and countering of the design parts. The small birds near the corner are equally illogical. Did the embroiderer begin with these and finish with the more ambitious motives? Or did she weary of doing the larger motives, and end off the row with the small birds as fillers? (D) An end border on fragment 171219a1 (Plate Lla). Profile bird forms in zigzag arrangement, the bill of the upper bird forming the leg of the lower bird, and vice versa. At the corner end of the border, and for no obvious reason based upon spacing, there is a single element identical to the motives in end border C. (E) An end border on fragment 171218a, a second arrangement of the bird motives of border D in a double zigzag which drifts off without reason into a pair of the same motives used alone in border C (Plate Lie). (F) A free standing motive, animal (?) or fish (?) on end of fragment 171219a1 (Plate LI la). In solid color, Blue. (G) Rectangular motives in main web, complete in specimen 171218a and partially complete in 171219a1. The individual elements of this design are lozenges, triangles, and chevrons, so arranged as to form double-headed serpent forms. The stitchery (?) in all the rectangular motives is parallel to the weft yarns, giving the effect of woven brocade (Plate L6). The colors, Red, Green, Yellow-Orange repeated 1-2-3-2, 1-2-3-2, 1-2-3, appear as lengthwise stripes on the Blue ground of the motive. (H) Rectangular motives, portions of which appear in specimens 171218a, 171219a1, and 171220. The rectangles are approximately the same size as those described under 170 TEXTILES OF THE EARLY NAZCA PERIOD G, but the basic elements are differently combined. In addition, each corner is extended by a zigzag line terminating in a bird form (Plate La). The same stripes of Red, Green, and Yellow-Orange yarns run lengthwise through these motives as in the G type. Birds Red and Green on opposite diagonal corners; in Blue and Yellow-Orange on remaining two. (I) A finial in the form of a bird extending out from the corner of a missing design motive. The original shape of the motive cannot be determined. The distance between the finial and the basic web of fragment 171220 is too short to allow a rectangle as large as those lettered G and H. There is no clue to the main design motive except that the bird is embroidered in Red and Green as in H type. (J) A motive of unknown size and shape on fragment 171220 (Plate LII6). Here there is space for a form similar to the rectangles lettered G and H, but the bit of intact embroidery is unlike either, although apparently composed of some of the common elements. The one constant factor in the patterning is the limited range of colors used. If mantle 171220 showed fewer signs of fading, the number of hues might be analyzed with greater certainty. The yarns matched to the printed samples in Maerz and Paul's Dictionary of Color yielded the following results: Bronze Green (16 L 12), two Yellow-Orange hues similar to Raw Sienna (12 J 9 and 13 L 9), two dark Blues (39 H 3 and 40 H 10), a Brown similar to Chestnut (7 H 9), the familiar Brickdust (5 L 11), and perhaps several other Reds which at this time look like faded yarns. Upon analysis it has been found that the Peruvians had dyes to produce a wide range of Reds, but in this particular case it seems wise to stop with the one Red which is certainly present. Many of the yarns in mantle 171220 show strands differing in color, and it is doubtful whether matching an assortment of swatches would mean very much. The impression given by the whole mantle is of embroidered motives in Red, Blue, Green, and Yellow-Orange, with occasional small areas in dark Purplish Brown. The repetition of a series of these colors is nowhere insistent except in the two side borders, A and B, but it does make itself felt as a deliberate repetition. Cahuachi mantle 171220 presents a number of features common to the Early period fabrics as we know them : the cotton basic web with a pattern in wool ; the bold designs, no doubt conventionalized and affected by local tastes, but not rendered meaningless through slovenly handling; the strong rich colors of the dyed yarns, the whole general effect of a garment woven and embroidered by craftsmen who were aware of involved techniques, but who had them under perfect control. This is evident in the evenly spun yarns and the uni- form surface texture of the material. Majoro mantle 170465a is a fragment of incomplete length, but full 37" width. Needle- holes on its side selvage indicate that a second breadth was originally seamed to it. The weave is plain, warp face, and the yarn count varies in the three selected spots 52 x 32, 58 x 36, and 64 x 30. An arbitrary choice of 58 x 32 was made for Table 20. The embroidery is a combination of double running and stem stitch. One side edge, the outer in the original garment, has a band of animal figures (llamas?) with short legs, approxi- mate overall measure of 1" long by ]/^ high (Plate LVIIe). Extending from the corner across the loomstring end for 6" are two rows of birds (?). The yarn of the basic material is 2-ply white cotton. The embroidery is done in 2-ply wools: Purple, of two hues similar to Mauve (46 : G4 and 46 : D6, both colors characteristic of the textiles of this collection), Rose (4 E 10), and Yellow-Green (24 L 1). The colors appear in no regular order except that the two Purples occur together twice, making dark blocks in the border. The question whether the specimen is embroidered or brocaded is not raised by this mantle since the extra yarn is put in parallel to the warps on both side and end. Wherever EARLY NAZCA MANTLES 171 the colors change the embroidery yarns loop about each other as in mantle 171220 (Plate LXVTIIe). The designs are built up by floats over 2, 4, or 6 wefts, and the stitches are so arranged as to give a flat twill effect to the face of the embroidered surface. Cahuachi mantle 171216 is chiefly interesting because of the unusual number of colors brought into its design motives. It is a complete piece, 57" x 473^", in plain weave; warp- weft counts 28 x 34; 2-ply white cotton for the basic web, single-ply and 2-ply dyed wools for the embroidery (Plate LVIII). The surface and reverse sides of the mantle are identical. The stitch is the same double running stitch mentioned above, but this time used alone (Plate hXh). Double running is essentially like weaving in that the needle passes alternately over and under a definite number of warps or wefts — in this mantle three of either — when progressing in a given direction. On the return, progressing in the opposite direction, the same number of basic web yarns are passed over and under, but in reverse order, so that if warps 1, 2, 3 are passed over the first time, on the return they are passed under by the embroidery yarn. The dif- ference in the direction of the stitch, whether made parallel to the warps or to the wefts, varies the surface effect. The work appears to be solid embroidery, although the lines of wool stitchery are separated by a single cotton yarn of the basic weave. Curves and diagonal lines are built up as in weaving by stepping the embroidery stitches to right or left. Each corner of mantle 171216 contains a human figure holding a staff or weapon in either hand. The figures are about 4%" tall, the length of each parallels the weft yarns, and each is divided into approximately the same larger areas of color: face, body, head-dress, tunic, breech clout, staves. The faces are represented with painted markings. At least fifteen colors are involved. Some of the figures and even the details were originally outlined with dark yarn, probably as near Black as was obtainable. All these outlines are either completely disintegrated or remain in the fabric as bits of charred yarn. The mouth and eye details also were probably in blackish yarns since they have fallen out of the material. The greatest number of colors are used in the head-dresses and the staves. One figure has a head-dress embroidered in checks of eight colors; another carries staves divided into ten small color zones. Most of the yarns are still bright, and fifteen were matched to the printed samples in Maerz and Paul's Dictionary of Color with the following results : Reds: 3 J 9, similar to Coral; 4 G 10, similar to Etruscan; 5 L 11, Brickdust. Yellows: 10 J 5, Corn; 10 K 6, Chinese; 13 L 9, similar to Hazel. Greens: 22 J 1, similar to Reseda; 23 E 5, similar to Cedar; 24 A 6, dark Yellow-Green; 24 E 11, similar to Hunter Green. Blue-Greens: 38 J 3, 40 H 9, and 40 J 6, similar to Navy. Purples: 44 B 2, similar to Heliotrope. Purple-Reds: 56 C 5, similar to Burgundy. Cahuachi mantle 171222 is represented by two large fragments 20" x 41}^" (full width) and 28" x 26" (Plate LIXa). Thanks to the maintained color sequence of the embroidered motives, it is possible to fit together the two pieces and to determine the complete size, 50" x 41^". The mantle is plain weave, with warp-weft count 28 x 24 per inch, of 2-ply brown (?) cotton, crepe twist. The allover pattern of open flowers, 5" x 3", is done with fine single- ply wool yarns. The technique employed gives the motives the appearance of having been set into the cloth. The stitch is the well-known "tent stitch," spaced. According to the usual method of making a line of tent stitching the embroidery thread passes over each intersection of warp and weft, missing none. In the Peruvian variety as exemplified by mantle 172 TEXTILES OF THE EARLY NAZCA PERIOD 171222, the embroidery thread passes over alternate intersections of the warps and wefts: over one warp-weft crossing, under one, over one, etc. (Plate LIX6). When, in the next line to be worked, the warp-weft crossings missed are passed over by the embroidery thread, the slanting stitches form diagonals within the motive illustrating the secondary design type of patterning of which the ancient Peruvians were so fond. Plate LIXa shows the shape of the design motive and the sequences of colors in flower shapes and center dots. The yarns matched to the printed color samples24 are the following: Flower 1. Red, 5L9 Purple 2. Purple, 46 D 2 Yellow 3. Light Green, 14 K 1 Red 4. Light Red, 4 G 10 Light Blue 5. Green, 31 C 7 Yellow Center Spot 6. Yellow, 12 L 9 Green 7. Light Blue, 39 H 3 . Light Red 8. Red, 5L9 Blue 9. Blue, 40 H 5 Yellow The Purple, 46 D 2, is slightly brighter than the hue known as Mauve Dust, and forms one of the small group of Bluish-Reds previously mentioned as noticeable in any amount because comparatively rare. In mantle 171222 the areas developed in Mauve are unusually large. An embroidered edge trimming, Cahuachi 171180d, belongs in this group on the basis of its technical features. Its appearance is similar to the needleknitted band with tabs and fringe which is applied to Cahuachi mantle 171309. The embroidered fragment is about 13^" wide plus a 2" fringe (Plate LVIIc). The foundation fabric is woven of 2-ply wool yarns, warp-weft count approximately 30 x 30 per inch. This foundation is almost covered with stem stitch embroidery of fine quality. The motive is the Nazca open flower worked in five or more colors, most of them faded or discolored. One side of the band is edged with close-set rectangular tabs about %" deep. These are built by working rows of blanket stitches (coil without foundation) back and forth across the width (Plate LXVa). The stitches are so small that the fabric has the appearance of having been woven. Tabs made in this fashion are frequently found on fabrics from Early Nazca and Paracas sites. They are described in the section on Mantles with Needleknitted Edges as foundation shapes for needleknitting embroidery. The fringe on the opposite side of the embroidered band was added by weaving between the outer warp of the band and a removable skeleton warp about 2" distant. This type of fringe is characteristic of the period. MANTLES WITH NEEDLEKNITTED EDGES Needleknitting is a term coined to describe the appearance of an embroidery stitch.25 By means of this technique the Early period embroiderers of Nazca and Paracas constructed edge trimmings from simple bindings to complex three-dimensional passementeries. The best-known example of the latter type is the celebrated Larco-Herrera textile from the Necropolis at Paracas.26 An analysis of the five mantle specimens upon which needleknitting is found reveals its range of possibilities. Two of the mantles are colored: 171112 Bluish-Black and 171309 Dark Green (16 E 3). The dimensions of the only two complete garments (171223a, 171224) show them to have been almost square (Table 20). The pair of 25" widths for each are seamed together with tiny stitches taken through each turn of the wefts on the side selvages, " Dictionary of Color. • Textile Periods, p. 32, footnote 25. " J. Levillier, Paracas, A Contribution to the Study of Pre-Incaic Textiles in Ancient Peru. Paris, 1928. EARLY NAZCA MANTLES 173 a process which may or may not have been done at the time the second breadth was in the loom.27 The Bluish-Black mantle, 171112, which looks like a veil or turban, is included in this group for the reason that there are too few similar examples to justify subdivisions. All five mantle specimens are plain weave cottons, generally of finer quality than that of any of the other groups, as has already been noted in the section on the Correlation between Mantle Types and Yarn Counts (Table 24). With the exception of the embroidered edges the mantles within this classification present no distinctive features. Each has a needleknitted edge of wool yarns in a wide range of colors. For the present, it will suffice to say that typical needleknitting was developed in from four to a dozen or more colors. The variety was made possible by the freedom from technical limitations such as were imposed by the loom, and by the fact that short bits of yarn could be employed advantageously. Any generalizations regarding Early Nazca con- ventions or tastes as shown by color choices must be based upon comparisons and occurrence frequencies. These can best be presented at the end of this section in a summary. The motives developed in needleknitting technique by the Early Nazca embroiderers, to judge by the available material, were relatively few: birds of the parrot and humming bird types, full-blown flowers, plant and seed (?) forms, faces, and the stepped fret. The bird and flower motives seem to have been suitable for representation, either flat like plain embroidery, or in three dimensions. Following is a brief resume of the needleknitted details on mantles in the collection. Three of the five have fringes representing birds on a "branch." The fringe sewed to Cahua- chi veil or mantle 171112 may be taken as a characteristic specimen (Plate LXI/). In this fragment, 1}4" wide, plant forms originally alternated with birds' heads. At intervals within the length of the branch there are colored motives representing bodies and three-toed feet. A dozen different colors may be identified, not counting those which are interpreted as faded hues. The fringes on specimens 171223a and 171237 are similar to the above except that the %" width reduces the size of every detail (Plate LXIIa), yet the number of colored yarns employed is 12 and 14. Originally there were flower or plant forms between the birds of the latter specimen as indicated by the remnants below the "branch." Mantle 171224 illustrates one of the many discovered variations of the needleknitting technique. The fragment is 50" long, the complete length. A Red and Purplish Black striped trimming band, approximately 2" wide, remains sewn to one of the side edges (Plate LXIa-c). The set-up allowed for a center space as follows: warps for %", J^" space, warps for %". The wefts crossed the space, thus providing a foundation over which to work the needleknitting. This was done on both sides of the bare wefts, making the band reversible except for the fact that birds' heads rise up from it on the surface side. The realistic effect is heightened by the open bills in Yellow stitchery and the lines on the needleknitted band representing tail feathers. Cahuachi mantle 171309 is a folded bundle of material too charred to open out to measure. The fringed edge trimming illustrates one of the familiar uses of the needleknitting technique. The stitchery veneers a woven foundation tape 8 warps wide. On one side of this veneered band are small tabs in needleknitting made over a foundation of buttonhole stitches, on the opposite side a fringe (Plate LXIrf). The loops forming the 2" fringe were made by weaving yarn from the needleknitting stitches on the lower edge of the tape over and under three warps, two of them close set, the third one the required distance from the 27 Textile Periods, p. 30, footnote 15. 174 TEXTILES OF THE EARLY NAZCA PERIOD other two. This skeleton warp was subsequently removed, thereby allowing the weft loops to twist tightly. Many Peruvian specimens are edged with fringes woven in this manner. Most of them are separately made and joined to the garment. UNIDENTIFIABLE FRAGMENTS Besides those already described, the collection contains a number of fragmentary specimens of needleknitting. Some are similar to the bird and flower fringes, and some are included in this section on mantles only because they have more points in common with the mantle edge trimmings than with those on any other of the garments in the collection. The group, more or less unidentifiable as to original use, falls into four classifications : (1) Bird and flower fringes, presumably the edge trimmings of mantle, veil, or turban types, \yf-Z" wide. The fragments within this first category are very similar. Trancas fragment *9058 {2%" wide) consists of a needleknitted tubular "branch" along which humming birds alter- nate with open flowers (Plate LXII6). The latter are realistic. They have centers, petals, and stems in solid colors and stripes, but no two flowers are alike. The birds' forms are similarly divisioned into color areas of eyes, throats, breasts, wing and tail feathers. There is appar- ently no planned arrangement of the colors in any of these intricate passementeries. The hues, several of each family, range from light to dark, and through various degrees of grayness. The above description will do in a general way for Cahuachi specimens 170211a, 171113, and 171180a. All are fragments ranging in length from 3" to 6". As the summary will show, the number of colors varies, but the bits of each are so small as to make all flower-bird fringes look alike except for size. (2) A stalk with blossoms and leaves (specimen 171114) probably was part of a fringe trimming (Plate LXIIe). It is more elaborate in its detail than the bird and flower motives, but shares the same general characteristics: realistic rendering of design units within the limits of a technique which cannot avoid being bulky, and the division of the units into small color areas. (3) Flat, veneered tapes with fringed and tabbed edges like those of mantle 171309 are represented by five specimens. Cahuachi 171180b1 is a fragment about V/2" wide (Plate LXII0). It consists of a woven tape, and tab foundations covered on all sides with needleknitting. The foundation fabrics have disintegrated but the assumption that they were part of the piece in its original condition is based upon their presence in similar and better-preserved specimens. The design motive in the band might be one of the numerous bird-form abstractions. The second specimen of a veneered tape, 171117, is an unusual example (Plate LXI$r). Geometric motives seem rarely to have been developed in needleknitting. This fragment has pairs of interlocking stepped frets approximately 1^" long, each in a different color. The piece is too small to indicate whether or not a repetition of the colors was carried out. The %" fringe is made as described for mantle 171309 above, by weaving a yarn from the lower edge over and under three warps set close to it, and a fourth skeleton warp the desired depth of the fringe distant from the others. The skeleton warp is withdrawn upon com- pletion of the weaving. The remaining three specimens in this category, fragments 170211b, 171116a, and 171321a are all flat, veneered tapes with tab edges (total width approximately 13^") (Plate LXII/, h). The motive in each is the favorite flower-bird combination. The first specimen still shows two intact humming bird forms, tail feathers spread, bills in an open six-petal flower. The colors of the yarns are dimmed due to the bad condition of the specimen. The EARLY NAZCA MANTLES 175 second piece is in even poorer condition, but remnants of flowers and feathers indicate the familiar arrangement. A dark outline on the tabs of specimen 171321a may be a bird form. (4) Needleknitted ornaments in the form of small squarish human heads (approxi- mately V/i' by V/i') may originally have been part of a decorative edge (Plate LXIfe, i, j). There are seven in the group, three badly charred: Cahuachi specimens 171115, 171180b2, 171321b1-5. The faces of the better-preserved specimens are Red, with features and paint marks in Yellow, Green, and Blue yarns outlined with Black. The rectangular noses are in bold relief, and remnants of "hair" are indicated by long figure-8 stitches across the fore- heads of two specimens. Each head is embroidered in fine needleknitting stitches over a cotton foundation. The surface and reverse sides of each are duplicates except for some deliberate interchange of colors. COLORS IN NEEDLEKNITTING SPECIMENS After amazement at the intricacy of detail in a needleknitted specimen, one's next reaction is to the minute bits of color which appear and disappear in the motives. Then one begins to count. The task does not yield a satisfying scientific result because some of the colors have faded to hues which have counterparts in unfaded yarns: Reds became Golden Browns; Green lost its Blue and became Yellow; and Black most often is Red, Purple, Green, or Brown when matched with samples. Embroidery in Early Nazca times, and this includes needleknitting, makes obvious the fact that the kaleidoscopic effect given by many small areas of color was desired. One type of woven fabric, the multi-colored patchwork, likewise lent itself to the same effect, and by reason of it claims aesthetic as well as technological interest. The colors listed below are those found among the needleknitted specimens. They have already appeared in the various color charts (Tables 3 to 19), and form part of whatever computations were based upon them. The notation follows that of Maerz and Paul's Dictionary of Color in indicating plate, column letter, and row number of the printed samples matched by yarn swatches. The tabulation makes evident that the Early Nazca embroiderers were aware in com- bining colors of the effect of certain values and degrees of purity. They must have been conscious of the fact that the Orange-to- Yellow hues are bright and sparkling in their light values, but appear muddy as they are darkened and grayed. They apparently did not favor the light Reds and Pinks, to judge by the great preponderance of medium and dark hues. And when they chose to use yarns coming within the groups from Yellow around the circle again to Red, they restricted themselves largely to closely related Yellow-Greens (Plate 23), 28 Blue-Greens (Plate 39), 28 and the grayed Purples, previously mentioned as noteworthy. These tendencies argue for as well-established predilections for certain colors as we ourselves entertain. The involved specimens are from only two sites in the same valley, but represent finds from six graves.29 They constitute a fair sample by which to judge the characteristic use of color in a single fabric type. ■ See Table 26, footnote t- "The graves represented are Nazca, Cahuachi III, Ag, AjlO, Ajll, A15, Ha. The Trancas specimen, *9058, is a surface find. 176 TEXTILES OF THE EARLY NAZCA PERIOD TABLE 26 Range of Colors in Needleknitted Specimens! Color Groups Light Hues Medium Hues Dark Hues, BlackJ Red-to-Orange: Plates 1-8 4 J 10 5K3 5K5 5 K 6, Crimson± 5 K 10, Ember 5 L 7, Barberry 5L 9 5 L 11, Brickdust 6 D 11, Gingeri* 8H2 8L 7 8 L 9, Chocolates Orange-to- Yellow : Plates 9-16 10 F 4, Maizei 10 G 2, Buff 10 K 6, Chinese Yellow 11 D 4, Maplei 11 K 8, Gold Leaf 12 F 4 12 Kl 12 K 5 12 L 7, Burnished Gold 12 L 9, Golden Yellowi 13 J 11, Sorrel 13 L 9, Raw Sienai 15 A 9. Dark Beaver 16 L 12, Bronzei Yellow-to-Green: Plates 17-24 20 D 5 20 F 6, Pea Green± 23 E 5 23 H 4 23 J 1 23 J 4, Lincoln Green 23 L 4 24 A 7 Green-to-Blue-Green : Plates 25-32 31 C 7, Poplari 32 J 6, Sprucei Blue-Green-to-Blue : Plates 33-40 38 H 3 39 H 3 39 H 4 39 H 9 39 J 6 Blue-to-Red: Plates 41-48 45 B 1, Zinc 45 I 7, Amethysti 46 A 1, Cement Gray 46 J 6, Old Mauve± 48 J 6 48 L 12, Egg Planti Purple-to-Red: Plates 49-56 t Notation follows that given in Maerz and Paul's Dictionary of Color. J Black, darker than any of the printed samples, appears in practically all of the needleknitted specimens, a true Black, but most nearly matches the darkest Brown, Green, Blue, Purple. * The ± sign indicates an approximate match. It is rarely Due to the probability that some hues were overlooked in matching yarns to the printed samples in the Dictionary of Color, a count is little more than indicative of the minimum number in a needleknitted specimen. Compilation of the number of hues counted in the more familiar types of needleknitting shows, however, the forms which tended to elaboration through the use of many colors: Number of colors, Type of needleknitted specimen at least 3—4 Human head ornaments 7-8 Veneered tapes with tab or fringed edge 8-14 Bird and flower fringes The Needleknitting Technique. — The weavers of the Early period in Nazca seem to have revelled in passementeries of the 3-dimensional type and veneered bands with tabs or fringes EARLY NAZCA MANTLES 177 or both.30 This must mean that the simpler forms in the needleknitting technique, whatever they may have been, were outgrown or had been elaborated. Whatever forms preceded them, the 3-dimensional bird and flower fringes represent needleknitting embroidery at its peak. Nothing as fine as the Larco-Herrera cloth from Paracas has been reported from Nazca Valley, but structurally the specimens from that locality are the same as those from the Necropolis at Paracas. The 3-dimensional types imply that the embroiderer did not depend on the woven material of the garment upon which to work a trimming. She visualized the finished embroidery, and provided the necessary foundations for the specific forms which she meant to make. The needleknitting stitch itself is a variant of the cross-stitch (Plate LXIIIa, b). It may be employed to make a narrow line or a wide band, but it must be constructed over founda- tion material or a core. Specimens of needleknitting in the Early Nazca collection illustrate types of specially constructed foundations such as narrow tapes, shapes made by buttonhole stitchery, and twists of yarn for cores. The examples already described required one, two, or three of these foundation types. Specimen 171309, a veneered tape with tabs, has a woven cotton foundation 8 warps wide. The tabs are constructed of buttonhole stitch (Plate LXIrf). The tape and tab fabric are covered with needleknitting. The squarish human heads also have a cotton foundation made of buttonhole stitches worked row upon row. This foundation is invisible, hidden by fine needleknitting stitches, about 24 per inch. The construction of the bird-flower fringes must have required the greatest technical skill. Veil or mantle 171112 is a characteristic example. The main points in its analysis supplement the description of its motives and colors given above. The foundation for the "branch" element is a woven cotton tape approximately J^" wide. This is completely hidden by the needleknitting stitches which represent the bodies and feet of the birds. The stitches are worked crosswise of the band. The birds' heads which rise from the "branch" are shaped, and therefore a woven foundation was impractical. But a foundation built up of simple embroidery stitches (the buttonhole stitch, or coil without foundation) was feasible. The fabric shown in Plate LXIIIgr was reconstructed stitch by stitch to follow the microscopic analysis of one of the heads in specimen 171112. Shaping is shown to have been accomplished by adding stitches at certain points. The tail feathers were veneered cores formed by tightly twisting three short lengths of cotton yarn, folded end-to-end to make a loop. The three were veneered as a group with several rows of needleknitting stitches. They were then separated so that each represented a single tail feather (Plate LXIII/). It is probable that this same device was resorted to when foundations were required for similar shapes. Whether the object of the embroidery herein called needleknitting was to veneer a flat tape, a tab, a flower form, a bird, or any other foundation shape, the actual appearance of the single stitch units varies but slightly. Reconstructing an ancient technique with modern materials duplicates the effect, but does not necessarily prove that the method followed was that of the original weaver or embroiderer. From that standpoint we cannot be sure until we find a piece of unfinished needleknitting accompanied by the tool or tools used that it was made with an embroidery needle. However, the plain bands of any number of stitch loops, and any of the veneers, in fact any of the details with which I am familiar, may be duplicated according to the following directions: Step 1. Begin with one, or any number of plain cross-stitches made in a horizontal line from left to right. At this point the needle is passed by means of a long stitch on the 30 For an analysis of the technique see Peruvian "Needleknitting," American Anthropologist, n.s., vol. 36, pp. 405-430, 1934. 178 TEXTILES OF THE EARLY NAZCA PERIOD under side of the cloth so as to bring it out into position at the extreme left and just below the row of cross-stitches (Plate LXIIIa, 6). Step 2. Slip the needle — but do not pass it through the cloth — from right to left under each of the crosses in turn. This step completes each needleknitting stitch in the row, and brings the needle out again at the extreme right. Step 3. Pass the needle again through the cloth from the right to the left to bring it into position at the extreme left and just below the preceding rows of needleknitting stitches. Repeat steps 2 and 3. Patterns are made by dropping the yarn of one color to continue with that of another (Plate LXIIIc). PATTERN WEAVE MANTLES Pattern weave differs from brocade. In the latter there are basic warps and wefts which form the fabric independent of the supplementary warps or wefts which form the design motives. These last may be removed without demolition of the cloth itself. But in a true pattern weave, called figure and Jacquard weave by modern weavers, the basic warps and wefts not only make the foundation material, but also float on the surface wherever necessary in developing the design motives. If one of these pattern yarns is pulled out of the cloth some of the basic web is lost. Of the two types, pattern weaving indicates a higher stage of technical development than does brocading. The Early Nazca collection contains one complete and two fragmentary mantles in pattern weave technique. Specimen 171119 is 40" square, the width made by lacing two 20" breadths together. The texture is canvas type, warp face. It has the appearance of being an everyday garment. The weave is plain, except for a narrow double-face pattern- weave border across the ends. An interesting detail shown by this Early piece may also be found in the textiles of the Middle and Late periods in Peru: the crossing and grouping of the basic warps at the point of change from plain weave to the tapestry (Plate LXVIIIc). The method followed is the same for all pieces: pairs are formed by drawing together warps 1 and 3 and warps 2 and 4; warps 5 and 7 and warps 6 and 8, etc. The crossing prevents the weft yarns of the new technique from sliding or crowding the yarns of the finished weaving, and the grouping into 2's doubles the warp strength. In specimen 171119 the plain weave changes, after the warp crossing against slip and grouping, to tapestry weave for a few picks of weft. Tapestry weave is plain weave with the wefts so closely battened together that the warps are covered. It is an effective way to give a boundary line of color to a design band and firmness to an edge. The pattern weave which classifies this specimen is simple, and not too accurately accomplished. The design motive is a diamond with center dot set in a long rectangle below the Red tapestry stripe (Plates XLIII6, XLVc). The colors are dingy, but certain pairs may be distinguished: Orange and Brown, Light Blue and Black (almost a Jet), Pink and Brown, Blue-Green and Brown. Where the colors of two rectangles meet, the individual wefts interlock on the reverse side (Plate LXVIIIe). The remaining two of the trio of pattern weave specimens are alike in technique. They are fragmentary, but their widths, 23" (full width of one breadth, 171279a) and 26^" (partial width of 171218b), indicate that they were mantles. Flat- view diagrams of the designs are shown in Plates XLIVa and XLVII. Both are plain weave fabrics within which the patterns are developed on surface and reverse sides by floats of the warps and wefts. EARLY NAZCA MANTLES I79 Specimen 171279a is the simpler of the two. The set-up is in two colors, White and Brown (natural?) . One unit of the repeat is given : Colors: W BR W BR W BR W BR W BR W No. warps: 36 424252424 36 The wefts cross in exactly the same order, 36 White wefts followed bv 4 Brown wefts 2 White, etc. This weft crossing would result in a simple plain weave plaid if the warps and wefts were not allowed to "float" or "flush" over certain groups. The diagram cited shows the positions and designs developed by these floats. This specimen shows, as many do, the freedom a hand weaver enjoys. A mechanical loom is set in such a manner as to force repetitions of units at certain spaced intervals. In mantle 171279a there are a half dozen different arrangements of spacings and positions of the floats to vary the design units. Mantle fragment 171218b is an elaboration of both the plaid and pattern by contrast with the preceding example. The technique is similar. The set-up is in Brown yarns (8 L 11), Blue (39 E 4), Pink (11 B 6), Jasper Green (32 H 9), and White. The method by which warp and weft yarns are manipulated so that floats of each will develop the design motives is shown in Plate XL VI I. GAUZE WEAVE MANTLES Of the three fragments in this technique only one is large enough to justify placing it in the mantle class. Specimen 171110 is a torn strip 32" long by 9" wide. Its length is the guiding factor to its classification. The background is plain weave, the motives true gauze weave. This latter is made by drawing one active warp of a pair, or two active warps of a group of four over the passive one or two warps, and securing the cross so formed by passing the weft through it. In this specimen warps 1 and 3 are crossed over warps 2 and 4. Peru- vian weavers seemed content only in rare instances to stop with simple crossings of a single or a pair of warps. They usually separated a pair to cross each single with other singles for several passages of the weft. Plate XLIa shows the shifts in position of any one yarn before being returned to its original place in the set-up. The effect is more like lace than is that given by our own standard gauze weaves. The motives in the Peruvian gauze weaves are often elaborate. The style in the Early Paracas period, the Caverns, was for most complicated patterning. This Cahuachi example is white cotton with simple allover pattern in the ubiquitous stepped fret, each motive 5" on the warp by 3" weftwise (Plate XLIc). Fragment 171141 is a plain weave cotton fabric with stepped fret motives in bands, allover effect. The yarn is dyed a darker Blue than any sample shown in Maerz and Paul. Several small Blue wool fragments, Majoro 170476f\ were probably part of a mantle or veil. The weave is plain with cross stripes of gauze forming the pattern. The gauze is made in the usual manner: crossing warps 1 and 3 over 2 and 4 in the same group. The count of the plain weave portion is 34 warps by 36 wefts per inch, a contrast in quality with the two cotton specimens, 30 x 24 and 26 x 26 per inch. The fragments do not contain the complete gauze weave motive. Rows of crossed warps about 2" long are grouped in fours to form a rectangular unit %" wide. The drawing (Plate XLI6) shows the arrangement of these units in a stepped diagonal of gauze weaving. In addition to the decorative weave there are appliqued S-form motives: a 3-strand braid in Gold (12 K 9) wools, and a 4-strand flat braid in Rose (5 K 9) and Dark Green (32 C 8) yarns. Applique" of yarn-made motives is rare in Peruvian textiles. One of the gauze weave fragments of this same mantle has a border tape 1}4" wide still held to it with long whipping stitches. The tape was originally between two webs, since 180 TEXTILES OF THE EARLY NAZCA PERIOD remnants of stitches on its opposite edge show that material has been torn away. The weave is plain, with a pattern in warp floats making lozenge motives through the center of the border. The same three colors appear in the border as in the rest of the specimen, the Blue and Gold in 34" stripes on either side of the Rose center portion. MISCELLANEOUS MANTLE TYPES INTERLOCKING WARPS AND WEFTS A unique method of patterning is illustrated by Cahuachi specimen 171221 (Plate XXXIXrf). The weave is plain throughout, and color changes are effected by interlocking weft yarns at their points of meeting. This is tapestry weaving, but in tapestry the wefts are so closely beaten together that they completely cover the warps, forming a dense fabric. In the Cahuachi specimen the warps and wefts show equally, and the pattern as well as the basic web is as sheer as the quality of the yarn permits. Plate XXXIX6, made after a reconstruction, shows the method of weft interlocking, and the manner in which the triangular projections out from the motive are woven. Those on the long sides are formed by extending the colored weft of the main decorative bands to the desired distance; but each of the triangles edging the outer and inner ends of the figure has its own individual length of colored weft which interlocks at every turn with correspond- ing turns of the white weft of the background. The warp-weft count per inch, 28 x 32, indicates quality of the material. The wool yarns are single-ply wools. None of the mantle corners is complete, but two are in condition to show the color changes. The double-headed monster motives are represented with striped bodies, solid color heads, eyes and smaller details in one or another of the three main colors. Motive No. 1, stripes: head: Blue (39 J 8) Blue Rose (4 H 10) Green (15 J 1) Rose (4 H 10) Blue (39 J 8) Motive No. 2, stripes: head: Gold (12 J 9) Gold Green (15 J 1) Brown (15 A 10) Green (15 J 1) Gold (12 J 9) The yarns are not consistently one color throughout a motive. For instance, both Brick- dust Red (5 L 11) and Rose (4 H 10) are found. The third corner is too fragmentary to show a color sequence, and the fourth corner is missing entirely. Multicolored Patchwork Mantles. — The mass of wool fragments constituting Cahuachi specimen 17021 le gives no indication of its original size, but at least it could not have been a kerchief, nor does its fragile texture suggest a tunic (Plate XXXIXc). It might have been a veil or turban piece, or a mantle. This type of fabric seems to have been well devel- oped even in the Early period. It got its name because of its appearance, although the term "patchwork" usually implies seaming together bits of already-woven fabric.31 In the case of the Peruvian examples, the evidence all points to their having been made wholly on the loom. Successful reconstruction of the technique has been accomplished through the use of supplementary transverse yarns, scaffolding, or skeleton wefts. Between these it was possible to set up warps of different colors, and to so group them that they formed geometric designs. When warps of two colors met on a transverse yarn, they interlocked. This allowed the transverse yarn to be withdrawn upon completion of the pattern; neglected remnants have been the clues in the analysis. Once the warps were placed, wefts of the same color — sometimes the ends of the same lengths of yarn that formed the warp setups — crossed them, 31 Textile Periods, pp. 39, 40, 41, 49-51; figs. 8-10; plates 6a, 19. A.A., n.s., vol. 35, pp. 87-94, 1933. EARLY NAZCA MANTLES 181 and interlocked with wefts from the adjacent motives. Each detail thus woven is mono- chrome, in strong contrast to the other details surrounding it. Some of them, border lines especially, are so small as to have required needles for bobbins in the weaving. This form of woven material permits a range of colors only limited by the number of geometric design units in its pattern. As a matter of fact, were we in possession of a complete piece we might discover one of the elaborately conceived color sequences in which Peruvians of all periods took delight. The design motive in both specimens 170211e and 171111 are lengthwise stripes of interlocked step frets, each outlined with a narrow band of White or Buff. The particular fragment drawn for Plate XXXIXc has ten colors. Matching yarn swatches from the mass of fragments to the color samples in Maerz and Paul's Dictionary revealed more: Red-to-Orange group: Green-to-Blue-Green group: Light: Rose, 1J 9; Red-Orange, 4 G 10, 4 K 9 Dark: 32 J 6 Medium : Barberry, 5 L 7 Blue-Green-to-Blue group : Orange-to- Yellow group: Dark: 40 J 8 Light: Honey, 12 J 6; Gold ±, 12 L 9; Hazel ±, 13 L 9 Blue-to- Red group: Yellow-to-Green group: Dark: Slate, 47 C 6 Medium: Grass ±, 21 J 4 Purple-to- Red group: Dark: 21 E 5 Dark: 56 A 5 Twined and Painted Fragments. — A large irregularly shaped fragment, Cacatilla *8537, is included among the mantle types, although it may have been a veil or turban. The piece was not constructed on a loom or by interlacing warp and weft elements, but by continuously twining pairs of yarns. It is by the separation of these pairs to enclose other pairs, or by the progression of the pairs at angles to right and left that patterns are formed (Plate LIV). The whole fragment is marked off horizontally by three herringbone bands alternating with V/2 squares in the same twining technique. The 2-ply wool yarn seems to have been a Red which in some areas matches Maerz and Paul's Reddish-Purple 47 J 1, in others Havana Rose 6 K 9. Color variations within the same piece are typical of a good many of the wool yarns. Reds often fade to Browns in which Orange, Yellow, or Red may be predominant. Trancas *9056, a cotton fragment about 22" square, is also included within this mantle category for want of any better means of identification than its texture and general appear- ance. Its warp-weft count, 34 x 26, is similar to some among the plain weave, undecorated types. The painted pattern represents birds in rows so closely set together that the effect is an all over. Two typical features of Peruvian design arrangement may be seen in the Trancas fragment. Within a horizontal row of motives, birds with three-feather tails alter- nate with birds with fan-shaped tails. They all fly in the same direction. The birds in the next row are more or less identical with those above them, but they fly in the opposite direc- tion (see Text Fig., p. 135). In a more perfect piece, the diagonal lines of similar motives would be insistent. Schematically, with the two types of birds represented by 3-F (3-feather tails) and F-S (fan -shaped tails), the result shows the following: F-S 3-F F-S 3-F missing The alternation of motive types within a course or row, and the alternating rows of right- face and left-face motives may be recognized as a convention among the prehistoric weavers and embroiderers. This same arrangement is found in both Nazca and Paracas textiles, 3-F F-S 3-F facing right F-S 3-F missing facing left 3-F F-S 3-F facing right F-S 3-F missing facing left 3-F F-S 3-F facing right 182 TEXTILES OF THE EARLY NAZCA PERIOD from periods at the very beginning of the Peruvian time scale as we know it, and apparently it was as aesthetically pleasing to the Inca craftsmen at the end of the prehistoric time scale.32 The White basic web seems to have been washed over with a neutral Green or Gray paint after the pattern was put on. The whole effect of the piece is dull. Colors identified are Golden Brown, Black, Red, and Green. Most of the birds bear seeds, plants, or leaves in their beaks. Certain resemblances to the forms painted on the Early Nazca pottery may be recognized. " Textile Periods, plates 29 and 30. The motive is llamas in these Late textiles, but the Early period birds are arranged in like manner. EARLY NAZCA TUNICS The Early Nazca collection contains four tunics — two children's, two adults' — suffi- ciently complete to allow determination of several characteristic technical and decorative details. Other specimens have been allocated to this group on the basis of their resemblances to the better-preserved pieces. These more or less unidentifiable tunic fragments will be considered separately. The available material is too scanty to justify many inferences. What we should like to know is whether the wool or the cotton tunics were considered the finer, and whether any type of decoration was locally favored. In the following tabulation the full length of the tunic as worn is just half the length of the woven web. Specimen Full length Full width Proportions Warps per inch Wefts per inch number of garment of garment W:L C2 W2 C2 W2 Children's: 171071d 10" 15" 1.5 40 36 171266b 26" 22" 0.85 32 30 Adults': 171217a 33" 46" 1.4 44 20 171308 39" 45" 1.2 42 18 From the standpoint of proportions, the tunics and mantles cannot be compared, since one is woven for wear as a rectangle, and the other, although woven a rectangle, is folded on the crosswise center and seamed up the sides. The textures of the two garment types differ, also. The median warp-weft count for the mantles is 32 x 26, the count for a wool specimen; both cotton tunics are finer than this. The wool mantles are heavy, but less because of the closeness of their basic weaves than because of the weight of their fringed edges; the wool tunics are definitely warp-face, one or two of the fragments in the unidentifiable group being similar to tapestry in quality. Since the tunic is not only a woven web but also a seamed and decorated garment involving a number of different types of workmanship, I shall first present a summary analysis of each of the four specimens, and after that a detailed analysis of each of its technical features. (1) Child's tunic. Cantayo C-ax 171071d (Plate XXXI Vd). Size of complete web, 20" x 15"; side selvages whipped together to form garment; plain weave warp face; 11" Kelim slit left for neck opening. Needleknitted binding around neck opening 7 loops wide, 5 of which show on surface side of the garment; 5 loops wide on edges of armscyes. Fringes, needlemade, 1" deep around bottom; also at armscyes. Yarn counts, 40 warps by 36 wefts per inch of 2-ply brown cotton, hard twist; yarn in bindings, 2-ply wool dyed various colors. The Kelim slit is a familiar Peruvian type of neck opening (Plate LXVIIIgr). It is con- structed by turning wefts proceeding from opposite side edges around adjacent warps at the center of the piece. Since these meeting wefts do not interlock, an opening of the desired size is left in the lengthwise center of the web. The needleknitted edge finish in one or another of its variations is also a familiar technique and occurs in all periods on cloths from central and southern sites.33 The form found on this tunic specimen, a binding, is often patterned, and in this case has small plant-like motives at close intervals (Plate LXIe). These are in a maintained sequence of Light Green, Coral, Dark Green, Yellow, Navy Blue, 33 American Anthropologist, vol. 36, pp. 410-411. 183 184 TEXTILES OF THE EARLY NAZCA PERIOD Orange, and Black on a Rose-Red field. The method of making a patterned band is shown in Plate LXIIIc. The fringes were apparently made after the weaving was completed by drawing two extra strands of yarn down and up through each weft turn at the armscye edges and through each warp turn on the bottom or loomstring ends (Plate LXVIc). Loops of desired length were left, groups which, due to the tightly spun yarns, twisted about each other. (2) Child's tunic. Cahuachi 171266b (Plate XXXI Vc). Size of complete web, 52" (including fringe) x22"; side selvages, whipped together to form garment; plain weave; Kelim slit left for neck opening in both the main web and in the separately woven band which forms the foundation for the embroidery; fringes at armscyes and around bottom of tunic. Embroidery across shoulders of garment in addition to that on the separately woven bands appliqued at neck and at armscye edges. Remnants of feathers in three places between shoulder motives, each feather tied on by its quill. Yarn counts, 32 warps by 30 wefts per inch of 2-ply white cotton; applied fringe yarn, 3-ply white cotton of two different sizes; embroidery yarn, 2-ply wool, various colors. The Kelim slit opening is woven in the customary manner but is later reinforced by two rows of blanket stitching (Plate LXa). The fringes are of two types: that around the bottom of the tunic is made by leaving the desired length of warps unwoven below three heavy loomstring wefts (Plate LXVIa). The 6" depth is unusually long for this type of construction. The armscye fringes might be described as woven and subsequently applied to the edges. Two warps close together and a third 53^" distant provided the foundation for weaving. Upon completion of the desired amount of fringe the third warp was withdrawn, leaving long weft loops to twist together in groups (similar in construction to that shown in Plate hXVld). The embroidery which is done directly on the tunic is in the form of solidly worked rectangles approximately %" by J^r, alternately Red and dark Brown, seven rows of six across one shoulder, and nine rows of six across the other. The embroidered appliqued bands, 1" wide, show flower or fruit forms and monster heads in Rose, Black, two Yellow- Greens, Gold, Ecru, and Brown wool yarns on solid Red fields (Plate LVI). The embroidery is the simple outline or stem stitchery in rows set close together (Plate LXg). For the small rectangles, the usual length of stitch is equivalent to the diameters of four warps. Where the design requires curved lines, the stitches are somewhat shorter. The work progresses from left to right with the needle passing over four warps, back under two on the reverse side, over four from this point, and so on. (3) Man's(?) tunic. Cahuachi 171308 (Plate XXXIVe). Size of complete web 78" x45" (two breadths seam-joined); side selvages whipped together to form garment; plain weave, warp face; interlocking warps and wefts, Kelim and eccentric tapestry variations. Fringe across ends and armscye portion of side edges. Yarn counts, 36, 40, 52 warps by 18, 20 wefts per inch (the differences depend upon the portion of the tunic surface selected for count) of 2-ply natural (?) and dyed(?) wool. The yarn used in the tapestry woven borders is the coarsest. The two webs forming the garment were separately woven. Saddler's seams join the two at center front and back; plain whipping seams close sides (Plate LXVIIc, j). Plate XXXIVe shows the 4J^"-5^" blocks of contrasting color which form the central portion of the tunic front and back. Five skeleton wefts were necessary in order to effect the color changes. The warps for each block were crossed only by wefts of the same color. These interlocked at the side edges. The type of weaving which involves skeleton warps, wefts, or EARLY NAZCA TUNICS 185 both, and interlocking basic warp and weft elements has been called "patchwork" although no stitchery enters in as seems implied by the name. The body of tunic 171308 is Bronze Brown (15 H 11) ; the blocks are Yellow Beige (13 H 7) and Chocolate Brown (8 J 10). Tapestry borders in several other colors are woven about an inch up from the end loomstrings. Upon changing from the plain weave of the garment to the tapestry variation, the warps are grouped in fours, and held in these groups by two rows of twined yarns (Plate LXVIIIa). These outline the two edges of the borders through which a center line of zigzags forms reciprocal triangles. The Orange yarns making the zigzag are put in by a tapestry technique called eccentric wefting.34 Finally the twining yarns and the eccentric wefts are twisted or are plaited together on the reverse of the garment at the side seams (Plate LXVIIIa). The colors of the triangles formed by the zigzags are exchanged at intervals of six inches, approximately, according to the following plan : RRRRRRGGGGGGRRRRRRBBBBBB GGGGGGRRRRRRBBBBBBRRRRRR A fringe of unwoven warp lengths finishes the lower edges of tunic 171308 (Plate LX Via), and separately woven fringed bands are sewn to the armscye edges. These bands, 16" x 1)4!', contain the same zigzag pattern as that of the lower edge borders except that Red yarn always makes the outer triangle. In weaving, this Red yarn is carried beyond the outer edge of the band to a skeleton warp or similar device 7" distant (Plate LXVId). The sub- sequent removal of this warp results in a deep fringe. Both the bottom and armscye fringe yarns cling to each other and twist so tightly that instead of soft loops there seem to be fairly good sized cords pendent from the edges. (4) Man's(?) tunic. Cahuachi 171217a. Size of complete web 66" x 46" (two breadths seam- joined) ; side selvages whipped together to form garment; plain weave, warp face. Embroidery and fringe across ends; tapestry woven band and fringe for armscye portion of side edges. Yarn counts: main web 44 warps by 20 wefts per inch of 2-ply wool, dyed, crepe twist; tapestry bands, 14 warps of 3-ply white cotton, each ply double, by 22 wefts of 2-ply dyed wool per inch (Plate XXXII, Frontispiece). This tunic is seemingly a finer garment than Cahuachi tunic 171308 described above. The basic web is a rich Red and the tapestry band (10J^" x 3") seamed to the armscye has zigzags lengthwise of the piece woven in Kelim and eccentric tapestry techniques (Plate XXXVIIId). The sequence of colors is irregular: Yellow, Red, Blue, Yellow, Purple, Yellow- Green, Red, Blue, Yellow. The separately woven fringes of Blue-Green wool are attached to the outer edges of the bands. These were made over five closely set warps and a sixth placed at a distance of 6". This last was withdrawn upon completion of the weaving. The fringes around the lower edge of the tunic are unwoven lengths of warp loops approximately 1^" deep. A bunch of these loops is separated into two heavy "warps," those coming over the original loomstring forming one, those coming under the loomstring forming the second. A loosely twisted Blue-Green yarn woven over the two "warps" in figure-8 technique pro- duces the effect of a soft flat knob (Plate XXXVII Ic) . Just above the fringe pairs of yarns in two colors (Green-Blue with Blue; Golden Brown with Yellow-Green) are twined to form two narrow borders. The work began at the right, leaving a loose end of the yarn extending beyond the edge, progressed to the extreme left, then back to the right again, leaving a second length of yarn. Repeating with the second yarn of the color pair furnished four ends at the side edge. These were plaited in a flat braid in which was maintained the same arrangement of colors as in the stitchery. 3< M.D.C. Crawford, Amer. Mus. Nat. Hist., Anthr. Papers, vol. 12, p. 91, fig. 19, 1915. 186 TEXTILES OF THE EARLY NAZCA PERIOD Decoration for Kelim Slit Neck Opening. — Cantayo 171071b, a complete specimen, presumably was intended as a trimming for the neck opening of a tunic (Plate LXIIc, d) . The piece is noteworthy because the method of its construction is clearly visible, and the quality of workmanship unusually fine. A plain weave cotton foundation band (14^" x V>/%\ 44 warps x 40 wefts per inch) was woven with a long Kelim slit (Plate LXIIc). A second type of foundation, a small rectangle %" x 3^", was built up row upon row with buttonhole stitches (Plate LXVa). There are sixty-eight of these tiny tabs edging the band. The surface side veneer on band and tabs is needleknitted : a Red ground with bird motives in a variety of colors on the band, a plant-like motive on each tab. Perhaps a regular sequence of colors was planned since three of the motives on one side of the slit are duplicated by three motives on the opposite side. Each bird motive is subdivided into the following color areas: head, eyes, nose, beak, the object held in it, body, wings, wing tips, tail, tail tip. Some of these are given a decorative spot, and every detail is bounded with a Black line one loop wide. The ten or a dozen colors are similar to the Red (5 L 11), Blue (39 J 6), Yellow (10 G 2), Orange (11 K 8, 13 L 9), Yellow-Green (23 J 4, 32 J 6) of the needleknitted bird-flower fringes de- scribed in the mantle section. To develop a design involving a number of colors was not, I believe, especially difficult, but it did mean carrying along all of the yarns required by the motive so that each might be available when needed in a certain position (Plate LXIIIc). The embroidery yarns are usually hard twisted, a feature that would add to the difficulty of manipulating them. Tunic Materials, Identified by Stylistic Features. — Three other Cahuachi fragments are similar in texture and details to tunic 171217a. The most complete is specimen 17021 If, which measures 30" x 15". The fabric is a plain weave, warp-face wool, yarn count 46 warps by 24 wefts per inch, woven with 2-ply yarns. Whipping stitches remain to show the seam- joined breadths down the center of the garment and at the side selvages. Across one edge, corresponding to the bottom of the tunic, there is a short fringe of unwoven warp lengths. Twining weft yarns in Dark Brown and Cream wool yarns head the fringe. This technique is also found on tunic 171217a. Specimens 170211g and 171118c are very small fragments of dark wool fabrics. The first has a warp-weft count of 64 by 24 yarns per inch, and the second 76 by 18 per inch. All the yarns are 2-ply wools. The yarn counts of these fragments together with the two better- preserved garments indicate an appreciable range in quality: 42, 44, 46, 64, 76 warps per inch; 18, 18, 20, 24, 24 wefts per inch. Warp-face materials as closely woven as these have the appearance of fine canvas-like fabrics. One of the main criteria for including specimens 170211f, g and 171118c in the tunic group with specimen 171217a is the similarity of their dyed yarns. As with other specimens, there are too many variations in the color of large areas to be certain of the original hue. Tunic 171217a is Red-Orange, matching Maerz and Paul's 5 J 11 and 5 L 11 (Brickdust); fragment 170211f is slightly grayer and darker, 6 L 11, almost an Indian Red; fragment 171118c is Red, but a darker, Purplish or Haematite Red, 7 J 3; and specimen 170211g is darkest of all, a Purplish Brown, 8 H 6. A fragment of woven fringe, Cahuachi 171118b, is constructed like those found around the tunic armscyes. The foundation is typical: several warps, in this specimen two, close- set, and a skeleton warp at the desired depth of the fringe away from the others. The colors of the wefts are in a regular sequence, to judge from the 3" fragment: Black, Red, Yellow, Green, or Blue as 1-2-1-3, 1-2-1-4, repeat. EARLY NAZCA KERCHIEFS AND VEILS For want of a better term relatively small webs decorated in various ways together with two sets of trimming bands similar to those found on intact specimens are classified as ker- chiefs. They may have been worn as head coverings, or veils, except that the texture of the material does not warrant the latter word in most instances. All but one are cotton pieces. The decorative yarns are invariably wool. The sizes, yarn counts, and methods of decoration by which one sub-group is distin- guished from another appear in the following tabulation : Kerchief Specimen Complete Complete Proportion Warps Wefts type number length width W : L per inch per inch Undecorated 171071a1 14 15 1.07 Weft stripes 171305c Applied bands 170462b Applied bands 171033 16 14 \i 0.906 Applied bands 171059 6M 9 1-38 Applied bands 171226 18H 20 1.08 Embroidered *9120 11H 13H 1.17 Embroidered 171266a 14 15 frag. frag. frag. frag. 16 14 y2 6M 9 18M 20 U-H 13H frag. frag. 30 38 20 26 36 24 42 30 34 22 36 40 34 26 38 32 The main difficulty with the classification as presented is that it does not provide any clue to the differences in texture. These set apart two fragmentary pieces in particular: Cahuachi 171266a and 171305c. Each is further distinguished from the other specimens by having a simple knot tied in one of its corners. Both are in extremely fragile condition, but specimen 171266a was originally 31" wide, to judge by a 153^" breadth and a portion of a second still attached to it by fine seaming stitches. The other specimen is charred beyond possibility of measuring. Both veils are patterned. Specimen 171266a has small rectangular motives (approxi- mately Y% by %") set in one inch from the side edge, and about a half inch apart (Plate LXc). The lines filling the rectangles are made by a double running stitch similar in appearance to twining in basketry or weaving. After one series of running stitches outlined the form of the motive the spaces between the stitches were filled by a return series. The twining effect is gained by inserting the needle first above and then below the original stitch as shown in the drawing. The fragment is not long enough for a repeated sequence, if there ever was one. The single-ply dyed wools used are still bright and fresh looking. Matched to the printed samples in Maerz and Paul's Dictionary, they were found to be the following, in the order given: Tan (approximately 12 L 8), Grass Green (approximately 21J 4), Purple-Red (43 H 5), Shell Pink (1 C 10), Green-Blue (39 H 3), Bronze (14 L 9), Dark Green (32 J 6), Barberry Red (5 L 7), and Yellow-Orange (12 A 8). Cahuachi specimen 171305c is patterned by a very simple method of striping, but one I do not recall having seen before among ancient Peruvian textiles. Weavers of pre-Incaic tapestries usually began their patterned areas and zones with a series of lines and narrow monochrome bands running weft-wise. Other striped fabrics are usually warp face materials, with stripes running warp-wise. This Cahuachi fragment has a yarn count which shows it to be weft face, 20 warps by 26 wefts per inch, but the general effect is that of a gauzy mate- rial, similar to a square count fabric. The unusual feature is the introduction of a series of colored single-ply wool yarns in crosswise stripes of Bronze (approximately 14 L 10 and 16 L 12), Rose (4 1 10), Olive Drab (approximately 15 L 5), Etruscan Red (4 H 11), and Orange- Yellow (11 H 10). These are not battened down tightly, but are slightly spaced to 187 188 TEXTILES OF THE EARLY NAZCA PERIOD the same degree as are the cotton wefts, contrasting strongly with the usual weft-wise stripes of tapestry (Plate XXXVII;'). Majoro 170462b, also veil-like in texture, apparently was used on the head, judging from the quantity of human hair among the fragments (Plate XXXVI 16). It is a plain weave material, made of single-ply creped cotton yarns. A %" band is sewn to the edges of some of the bits. That, too, is plain weave with a set-up for warp stripes: a dull Bluish-Purple (47 E 1) in the center, 3 Brown (15 E 8) and 2 Red (4 J 10) stripes on either side. One other specimen, Cantayo 171033, deserves special treatment (Plate XXXVILj). It is an all-wool striped piece, similar in general effect and decorative features to the Cahuachi striped wool mantles already described. The amount of actual handwork represented by this kerchief is out of all proportion to the effectiveness of the result. First, the weave: a set-up for stripes in three colors, 4 Black yarns, 4 Yellow yarns, 4 Red, 4 Yellow, 4 Black, etc., with a variation in stripe widths due to crowding together the Yellow warps. This makes the Yellow stripe a finer weave than the Black or Red ones. Two complicated edging lengths, each of which is seamed down a long side and part way across the two ends, make this kerchief a miniature mantle. As was noted above, the striped wool mantles of Cahuachi and the Paracas Necropolis embroidered mantles illustrate the same conventional arrangement of trimming: side trimmings — fringes or tabs — and embroid- ered borders which extend around each corner for a distance, leaving a center space at either end (Plate XXXIV6). The 3^" edge trimming of kerchief 171033 consists of three parts, as shown in Plate LXIVc. (1) A brown cotton tape 6 warps wide {}/%), the foundation for stem stitch embroidery which completely covers it. The ground work is Red, and the minute motives (leaves? flowers?) are in six other colors: Rose, Yellow-Orange, Green, Blue-Green, two Blue-Purples. The embroidery is on a smaller scale than is usual through being confined to the space repre- sented by two warps, since the outer two on either side of the tape are used for fringe and "picots." (2) A }4" fringe made of a length of yarn woven figure-8 fashion between one edge warp of the foundation tape and two extra warps, the outer one a skeleton. The latter was withdrawn upon completion of the weaving. (3) Small "picots" set %" or less distance apart on the side of the tape opposite the fringe. In reconstructing these, as the drawing shows, the whipping stitches progressing from the left were elongated to extend between the tape edge and a skeleton yarn. Three of these stitches equalled six "warps" upon which the same length of thread that made them wove the tiny rectangular tabs. The drawing can give no conception of the effect of the fringed tape. It is colorful, but, as has been said, is too small to be truly effective. It com- pares, I think, to the fine needlework of our own times which spells achievement rather than an aesthetic result. The other specimens in the group are similar in texture to some of the mantles and aprons. They are plainly a less elegant type of garment than those already analyzed. Cantayo specimen 171071a1 has no feature technical or otherwise except its size to give it character. Embroidered kerchief *9120 from Nazca Valley and Cahuachi flower mantle 171222 are two of the few allover patterned pieces from this period. The kerchief motives, crosses and rectangles in rows, appear to have been worked with casual regard for sizes, placing, or color sequences (Plate LVa). Several of the crosses are in two colors as if the embroiderer had made use of short lengths of yarn. EARLY NAZCA KERCHIEFS AND VEILS 189 The stitch is a long double running stitch in which the embroidery yarn is left slack enough to cover entirely the basic weft yarns (Plate LXt). The work is mediocre by com- parison with the same type done on the four-warriors mantle 171216. On opposite diagonal corners there are Red and Yellow tassels made by drawing 18" lengths of yarn through the tips and there twisting the ends together. The usual Peruvian colors are present, together with Purple-Red (48 H 2) and dark Purple-Blue (48 E 10), of special interest in the Early Nazca period. Kerchief 171059 is a cotton rectangle measuring less warp-wise than weft-wise (Plate XLIIIc). This is rare. Separately woven pattern bands with warp-loop fringes are sewn to the side edges. Each band is a warp set-up of Red, Black, and Bronze-Green yarns arranged in the following order: stripes of 2 Black yarns and 2 Green yarns on either side of a center portion of 12 yarns : 6 Blacks alternating with 6 Reds. The block pattern in the center portion is single face. Unlike most pattern weaves, this type is a simple variation of the plain weave without floats. All the warps making the side stripes are crossed each time in regular fashion, but upon reaching the 12 Reds and Blacks in the center, the Reds are left down and the Blacks are brought up (Plate XLIIIa). For four picks of weft these only are crossed as in plain weave, over one, under one. Upon completion of the four picks, the Blacks are left down while the Reds are brought up to become warps. The alternation of Reds and Blacks for warps makes loose texture blocks of one color on the surface of the fabric, and long floats of the other color on the reverse side. In contrast to the 34 x 20 cotton yarns per inch in the basic web, the applied all-wool bands have 40 warps and 20 wefts per inch. One of the Cahuachi fragments identifiable only by its style is entered here because of its similarity in design and technique to Cantayo specimen 171059. Cahuachi 171118e is a patterned band 1}4" wide. Its center portion consists of a line of Red rectangles framed and centered with Black (Plate XLIIIrf) . On either side are the usual Red and Black stripes. The set-up for the center alternates 2 Black yarns with 2 Red yarns, total 26. The manipula- tion of the warps is identical to that in specimen 171059; whatever warps are brought up are crossed in plain weave fashion by the weft; the remaining warps, of the contrasting color, are left inactive as floats on the reverse side (Plate XLVa, b). Cahuachi kerchief 171226 (Plate XLIIIe) has a much more ambitious pattern, but is technically the same as Cantayo 171059. The plain cotton web is bordered on each side with separately woven wool bands joined to it by stitchery. The bands have warp-loop fringes. The warp set-up for the stripes on either side of the patterned center is as follows: 4 Red yarns, 4 Yellow-Orange, 4 Red, 4 Dark Green, 4 Red, 4 Purple-Black. Next comes the set-up for the patterned portion : 2 Purple-Black yarns alternating with 2 Yellow-Orange yarns, total 34. The flat-view diagram, made from a reconstruction of the original, shows where the light yarns are brought to the surface to make the pattern, or are left down, inactive, while the dark yarns are brought up to form the background (Plate XLIV6). The interlacing is invariably over one, under one on the surface side. The inactive yarns whether pattern or ground yarns appear as very long floats on the reverse side. The design motive in this pair of borders is the familiar Nazca bird with its bilj open six-petal flower. The arrangement of the birds illustrates the conventional, double inversion involving reversal of the two opposing bird elements on a horfjgnfsfy vertical axis. A pair of woven bands (Cahuachi 171225) are almost certainly kerchief boij measure 213^" by 1^", are fringed at the ends with unwoven warp loops, and are patterned • in a style similar to that of Cahuachi 171226, though not by means of the same technique. 190 TEXTILES OF THE EARLY NAZCA PERIOD As in the case of the other kerchief borders there are outer warp stripes of close-set yarns — Red, Yellow-Orange, Black — on either side of a central portion that is woven in Kelim tapestry technique (Plate XXXVIIIa). In specimen 171225 we have one of the infrequent illustrations of weft-to-warp change (Plate LXVIIIrf). That means that when the warps were set up for the stripes, the two out- side color groups — Red, Yellow-Orange, Black — were spaced far enough apart to allow for the center patterned portion of the band. Each weft crossed the first series of stripes, then the Yi space, and then the second series of stripes, regularly, as in ordinary plain weaving. The wefts crossing the space subsequently became the warps upon which flower motives in a dozen colors were woven in Kelim tapestry. The wefts may have been put in with sewing needles, judging by the fineness of the details. The quality of the band is fine, 44 warps by 20 wefts per inch. The tapestry yarns are single-ply and 2-ply wools. Cantayo 171045, a fragment 83^" by 93/2", is put in the kerchief group because it re- sembles specimens in no other (Plate LV6). One feels that the piece must have been a small rectangle but there is no detail to give certainty. Like the surface find, "Nazca" *9120, the Cantayo specimen is patterned with allover embroidery. The motive is the familiar step-fret done with a double line of stem stitches in dark Brown wool yarn, probably natural color, the motives so arranged as to give prominence to diagonal lines. The scalloped edge finish is of cotton yarn like the basic web except that wherever a dark line of embroidery comes to the edge, the scallop at that point is made with the dark wool yarn. The method of making the scallop detail can best be understood from the drawings in Plate LXVe, /, the former made from a reconstruction. The technique illus- trates the ingenuity of the embroiderer, who could, as she did in this case, combine at will whatever devices in stitchery and weaving suited her purpose. She provided a foundation upon which to work the scallops by making a couple of circular loops of the sewing thread, and between the bottom half of this heavy circle and the edge yarns of the fabric itself she wove in figure-8 fashion. The turns at the top of the circle float across the scallop. EARLY NAZCA APRONS Two complete aprons and portions of two others comprise the material from which to draw generalizations as to this Early Nazca garment. Fortunately, the pieces come from three different graves, which gives some ground for assuming that the characteristic features are local rather than the result of an individual's fancy. Cahuachi specimens 171181 and 171215 are plain weave, undecorated cotton rectangles of medium fineness. Strings or ties of approximately equal length and width extend from each of the four corners (Plate XXXIVa). There is a difference in length of one inch between those extending from the corners on the upper side, say, and those extending from the corners on the lower. Perhaps it is only a coincidence that the measurements differ to that degree on both specimens, or perhaps the method of wearing the garment regulated the lengths. The Red needleknitted edge bindings on all the tie ends in the group seem to indicate a conventional finish. Those on apron 171181 and fragments 171267b are simple 5-loop bindings (Plate LXIIIrf). The ends of tie specimens 171223b and apron 171215 are more elaborate. The first has humming birds alternating with flowers, each tiny motive about %" high, spaced approximately four to the inch (similar to motives in Plate LXIIa). The birds' heads which project from the bindings on apron 171215 are about W high, and each one is developed in four or five colors. Lines of contrasting color on the binding itself repre- sent the tail feathers. Both of these specimens represent very fine needleknitting technique. The usual dozen or more colors are to be identified in the bird-flower bindings. The weaving yarns are 2-ply cotton. All the decorative yarns are 2-ply wools. The tabulation shows the main technical features of this small group. Specimen Complete length Complete width String String Warps Wefts number of web of web lengths widths per inch per inch 171181 37 y2- 15M' 27"; 28"; fr. 13 y2"; fr. 20" 2" 30 30 171215 44' 21' 35M"t 36H"t 4K2' 5H" 40 36 171223b 84H't 7' 30 32 171267b fr. 2" 4' 34 34 t Two strings of this length. 191 THE MISCELLANEOUS GROUP Practically every specimen in this group is in a fragmentary condition. As a conse- quence, the names chosen to distinguish one type from another are often purely descriptive. The particular uses of ancient Peruvian fabrics can be determined from their representations on the pottery, as Montell has so clearly shown us,35 but it is difficult if not impossible to identify the smaller accessories of costume. The variety of details, ornamental or purely utilitarian, which are brought together from the Early Nazca sites represented in the collection for analysis in this section can best be appreciated when they are listed: bands or ribbons, cords, nets, pads (for deforming infants' skulls as at Paracas?), wrappings, slings, a girdle (?), a scrap of feathered material, a scraper (?). In addition to the fact that there are these distinguishable types, there is also a wide range of techniques. BANDS To classify as bands the narrow, individually constructed specimens in the collection, it is necessary to include some fabrications not made by warp-weft interlacing. These are the twined and plaited specimens. The pieces to which the term has been applied are listed in the following table. TABLE 27 Bands: Woven and Fabricated by Other Techniques Length in Width in Techniques Specimen Number Inches Inches Woven Plaited Twined Embroidered 1702nd1-5 frags. Va 5 170211d6 frag. Va 1 170465e frag. 8V2 Vi 1 . . 170476c frags. H i 170476d frag. 8^ Wi 1 1 1 . . 170476e frag. 8 Va . . . . 1 171049 frag. 16 1+ i 171050 frag. \Yi Vi l 171054 frags. ? . . l 171109a1-2 mass Va 2 , . 171109b1-4 frags. Va 4 171109c frags. Va i 171116b frags. Vi i 171118a frag. 17 2H 1 171118f1 frag. Vs i 17111Sf- frag. 14 Va 1 . . 171118^ frag. Vs i , , 171124a frags. Va 1 171124b1-1 frags. Vs 4 171124c1 frags. Vs 1 171124c2 frags. Vs . 1 171140a frags. 1 + l 171140b frags. 1 + l 171140c frags. 1 + l 171227a frag. 82 Va 1 171227b frag. Va 1 , , . , 171238b1-2 frags. 80 Va 2 . . 171258a, b mass Va 2 171280 frag. 14 5 1 35 Gosta Montell, Dress and Ornaments in Ancient Peru, Archaeological and Historical Studies, Goteborg, 1929. 192 THE MISCELLANEOUS GROUP 19.3 WOVEN BANDS Cahuachi 171118a is the simplest of all the specimens in this group. It is a plain weave, cotton fragment, 2J/£" wide, 96 warps by 24 wefts per inch, with only its unusual closeness of weave and its Blue-Black color to give it character. It seems to have been used for binding, to judge by the sharp crease through the lengthwise center. Majoro ribbon 170465e is the most attractive of the plain weave specimens in the group (Plate XXXVIII6). It is in Kelim tapestry technique, of a very fine quality of workmanship, as may be judged by the warp-weft count: 26-28 warps by 176 wefts per inch. The fragment is complete as to width, %'> DUt the count per inch is given in order to better contrast the quality of this Early Nazca tapestry with fine garment pieces of the Epigonal period: 46, 38, 58 2-ply cotton warps by 180, 216, 200 2-ply wool wefts per inch.30 During this latter period tapestry weaving reached its technical peak. The drawing gives a good idea of the design motives and their arrangement. Lozenge shapes with the long axis running cross- wise seem a little unusual. The diamond motive is most often found in the twined pieces, and in them is always set lengthwise. There is a repeated sequence of four colors on the medium Rose (5 H 9) ground: Yellow-Brown (13 K 8), Dark Rose, Rose, Purple-Red (7 C 5). All the outlines around the figures are Brownish Black. One of the noticeable features in this specimen is the disintegration of the Rose yarn, leaving the warps practically bare. I do not recall other specimens in which this color has deteriorated. Usually it is the Blackish or other very dark yarns that fall from the web. Cahuachi specimen 171280, a striped wool fragment, looks as if it were intended for a headband or girdle (Plates XXXVa, XXXVIId). It is 5" full width, folded through the center, whipped together on the edges, and in addition has two heavy twists of yarn sewn to them. The wools used in the stripes, which vary x/i *-%"> appear to be natural colors: "White" (12 F 5), medium Brown (Adobe, 14 D 7), dark Brown (8 C 10 and also 8 E 12). The drawing gives a better idea of the end finish than words can. Majoro band 170476d is the single example of double cloth in the collection (Plate XLIIa). It is double cloth of the standard type constructed of two independent sets of warps and wefts. The latter, in addition to crossing their respective warps, whether they be inter- changed with the opposite set or not, turn about each other at the side edges, thereby locking the webs at this point. The simple design motives are rectangles the width of the band on a Rose (5 K 9) ground. Each motive is woven in three colors provided for in the warp set-up: 4 warps of Burgundy (56 H 9, which looks Black in the piece), 4 warps of Green (32 C 8), 2 warps of Gold (12 L 9). A single Rose yarn alternates with each yarn of any other color. When the Rose set is crossed separately, the web is monochrome; when the tri-color set is crossed, there are warp stripes. All the yarns are 2-ply wools. Three short fragments in a combination interlacing-twining technique extend from the end of the 8" band. In reconstructing the technique it was feasible to begin with the inter- lacing which furnishes the foundation of the strand. It is not real plaiting, not a technique which makes a uniform strand if used alone, and yet it is not weaving either, since it requires but one set of elements (Plate XLIIflf). Following is the series of movements for strands in positions 1, 2, 3, 4 (the strands shift from one side to the other, but each is identified by its position at the moment, not its original one) : 3 over 2, under 1 ; 4 under 3, over 2. Those two movements provide a center for the twining strands, 6 to twine around strands 3 and 4, and 6 to twine around strands 1 and 2. Plate XLIIgr shows how the lower member of a twining pair always comes up to the right of the upper member. It also shows the small 36 Textile Periods, p. 45. 194 TEXTILES OF THE EARLY NAZCA PERIOD centers formed by the four interlaced foundation strands which are independent of the twining elements. The three short lengths made by interlacing-twining, 16 strands in each, are brought together into one 48-strand flat plait. The technique is simple: single outside strands alternately the right, then the left, come to the center crossing over 4, under 4, over 3, with the left active strand always crossing over the right active strand. What gives the plait a complicated appearance is the cross striping. This, too, is simple, being dependent upon the initial placing of the strands coming from the three 16-strand interlaced-twined lengths. The colored yarns in each were arranged in the following order: 11 22 33 44 11 22 33 44 11 22 33 44 44 33 22 11 44 33 22 11 44 33 22 11 Majoro 170476d is only a fragment of its original length, but the remnant suggests that the band, whatever it was used for, must have been in the quality class of accessories. Even in its present state, it illustrates the fineness of Early Nazca double cloths, 16-strand interlaced-twined braids, and a 48-strand plait of unusual color arrangement. Cahuachi specimens 171109a1-2 represent a type of narrow weaving to which has been given the name "turban band" (Plate XLc, /). What the original length of these specimens was, and on what form or foundation they were shaped for wearing can only be conjectured. Specimen 171109a1 is literally a formless mass of material, measuring approxi- mately 25 yards. Specimen 171109a2 consists of three bunches, each in form resembling a festoon. The best-preserved is 9" across approximately, composed of a group of 34 separate wool tapes wound around at intervals of 12", 16", and 20" from one end with wool strands to hold the group together. The construction of these turban bands is by plain weaving, but done so ingeniously as to create a unique result. There are only two warps and six wefts, 2 Red, 2 White, 2 Blue (171109a1). The interlacing over the two warps is regular, but it is done by each of the six wefts in turn: Red, Red, White, White, Blue, Blue. This brings all to the same side of the warps. On the return to the opposite side, the last weft across becomes the first, and the order is Blue, Blue, White, White, Red, Red. The drawing shows the scalloped edges which result. Imagine this a quarter of an inch wide and correspondingly thick. Specimen 171109a2 is identical except for the colored wefts: Green, Brown, Red. Striped Turban Bands. — The seven remaining specimens listed in the column of woven bands are alike in technique : narrow set-ups of wool warps, the odd-numbered of one color, the even-numbered of a contrasting color. Plain weaving brings up all of one color on the first pick, all of the contrasting color on the second pick, and so on alternately. Crosswise stripes are the result. The weft elements are of three types: (1) A single weft element consisting of a 2-ply wool yarn or a doubled 2-ply yarn crosses from side to side regularly (Plate XL6), as in specimen 171227. Such wefting produces a thin turban band. (2) A pair of weft elements which enter the shed separately, one from each side, passing each other and exchanging positions. Specimen 171124a has a warp set-up of Yellow and Red yarns, but the two Green-Blue wefts form visible color lines on the side edges. (3) Two pairs of wefts of different colors work in this manner: when the odd warps are raised, the pair of one color enter separately from opposite sides of the web, passing each other in the shed. On the next pick, when the even warps are raised, the pair of the second color enter separately from opposite sides of the web, passing each other in the shed (Plate XLa). Besides involving three more weft yarns than the first method does, and making THE MISCELLANEOUS GROUP 195 the turban band thicker, the edge is given added strength and decoration. Specimen 171118f2 has Red and Brown warps, and pairs of Red and Brown wefts. The turban bands with crosswise stripes are of fairly consistent quality. The warp-weft counts are approximately 10 or 11 by 2 or 3 for the quarter inch width, equivalent to 40-44 by 12-14 for the inch. The colors, also, are within a narrow range. Following the plan adopted for other subgroups in the collection, the tabulation shows the combinations which seem to have been favored by the Early Nazca weavers of turban bands. None of the colors is unusual. The Reds tend toward Purple or Orange, the Oranges are similar to Gold or Tan, the Greens are very dark, or Blue-Greens, and the Blues deep and rich. There are no Purples, and no pale colors of any group. TABLE 28 Turban Band Colors Specimen Number Warps Red-to- Orange Orange-to- Yellow Yellow-to- Green Green-to- Blue-Green Blue-Green- to-Blue Wefts 170211d> 170211d2 7L 2 6K9 7L 2 7L 2 6K9 7L 2 7L 2 6K11 6K11 6K11 • 6K9 6K11 6K11 7L 2 8H10 24 J 11 24 J 11 31 H 6 31 H 6 31 H 2 31 H 2 31E5f 39 E 8 39 E 8 39 E 8 24 J 11 7L 2 170211d3 6K9 170211d4 170211d6 171109b1 171109b2 12 L 11 12 L 8, 8 H 10 8H10 7L 2 8H10 8H10 31 H 6 171109b3 6K9 171109b4 171118f2 171124a 171227a 12 L 11 8H10 12 L 7 7L 2 7 L 2, 8 H 10 31 H 6 31 H 2 171227b 171238b1 12 L 8 12 L 8 6K9 171238b2 171258a1 12 L 8 12 L 8 39 E 8(?) 171258a2 7L2 t Warps originally 39 E 8(7) faded to 31 E 5. PLAITED BANDS AND CORDS The ancient Peruvians knew many different types of plaits, flat, round, and square. They knew, also, how to use few and many strands in their making, but until more study is made of the slings, especially, we shall have to conjecture that the main object in multi- plying the number of strands was to strengthen the plait, and that the addition of colored strands was incidental. That is, in comparison with the variety of plaited fabrications in the collections, the ones that evidence interest in combining colors are definitely in the minority. One-color plaits are most common, two colors are much less often found, although typical of the 4-strand round braids. Three or more colored yarns in the same plait make it unusual. For that reason the two types of interlacing and plaiting already described for Majoro 170476d are important as showing this form of technical achievement in the Early Nazca period (Plate XLII/, g). The turban bands illustrating plaited work as listed in the tabulation of band types (Table 29) total eleven. There are no square plaits in the collection. These seem to have been the particular contribution of the Late and Inca technicians. But there are fiat and round plaits with the 3-strand flat, our own most common variety, represented by only three examples.37 One of them is, however, the single example of a 5-color braid (171116b), 37 Textile Periods, p. 33, footnote 20. 196 TEXTILES OF THE EARLY NAZCA PERIOD probably a fragment from some decorative detail. The colors are the familiar ones: Orange, Blue, Cream, Red, Green. The second 3-strand flat is the result of grouping 9 strands into 3's (171118^), and the third example is suspect. The present 3-strand plait on specimen 170476P was originally a 4-strand, in all probability, but lost its dark strand of wool through deterioration. It is represented in Plate XLI6 as a 4-strand, to match the companion plait in the S-shaped appliqued motive. TABLE 29 Plaited Turban Bands: Number and Color of Strands Specimen number Flat braids; number of strands Round braids; number of strands Strand colors 170211d6 8 Red 171109c 8 Red 171116b 3 Orange, Blue, Cream, Red, Green 171118!' 9,3 Red 171118P 4 Yellow, Brownt 171124b1 4 Red, Green 171124b4 4 Red 171124b3 4 Yellow, Brown 171124b4 4 Blue, Brown 171124c1 7 Yellow 171124c2 4 Yellow t Yellow and Brown probably natural color of wools. The 4-strand round plaits are almost invariably in two colors. One method of making is to place the two pairs of colors in order: Red, Red, Green, Green, in positions 4, 3, 2, 1 (Plate XLIIc). First step: Pass the active element in position 1 under strands in positions 2 and 3, and back over strand 3. The active element is now in position 2. Second step: Pass the active element in position 4 under strands in positions 3 and 2, and back over strand 2. The active element is now in position 3. Repeat steps 1 and 2. The 4-strand flat plait represents very simple interlacing : the outside right strand crosses its neighbor to the left; the outside left strand crosses its two neighbors to the right (Plates XLII6, LXVIIIa; the latter is an unusual use of plaiting as a closing of a seam). The 7-strand flat plait is also simple interlacing: the outside right strand always passes to the left under 2 strands, over 1, under 1, over 2 (Plate XLIIrf). One type of 8-strand flat plait (specimen 170211d6) has a single active element, the out- side right strand. Each strand as it succeeds to this outside position at the right, passes consistently over 2 strands, under 2, over 1, under 2, to become the left strand (Plate XLIIe). The 9-strand flat plait (specimen 171118^) has two active elements, the outside right and the outside left strands. Each passes to the center, first the right one, under 2 strands, over 2 strands; and then the left active element, under 2 strands, over 2 strands, one of these last being the former right active strand (Plate XLlIh). WARP-TWINED FABRICATIONS Small all-wool fragments, Majoro 170476c, e, represent a class of materials of which we have comparatively few examples. Twining is familiar as a basketry technique, but in basketry the weft elements are active. Weft twining has already been described in con- nection with the decorative yarn finishes on tunics 171217a and 171308 (Plate LX6). The loom with its warp yam held in position does not lend itself to the construction of fabrics in which it is necessary to twist one warp yarn completely around another. Twining THE MISCELLANEOUS GROUP 197 is not to be confused with gauze weaving, which it resembles superficially. In the latter the active warp of the pair shifts from side to side, makes a half turn, but never encircles the passive warp. But if the present Early Nazca collection contains only small fragments of bands in warp twining, it is probable that more and varied pieces will be made available for study at a later date. The technique was surely well known to the Early period weavers. A small group of fabrics from three Paracas Caverns graves furnishes two good examples.38 Complete measurements can be given for Paracas Caverns specimen 8430. Held without stretching, it is about 52" long by 22" wide. The description as given in the Revista is as follows: "There are two very heavy loomstrings at either end. This suggested that the two-ply wool yarns might have been stretched in some sort of frame. It was found possible to duplicate the effect although as in all reconstructing of old techniques it is impossible to state with certainty that the method used duplicates the ancient method of fabrication. "In the twining technique, pairs of warps work together, separating to form units of a new pair, and, in this particular type of twining, coming back to their original position after a set number of twists (fig. 33). Any twist of one yarn about another at one end of a group of yarns held in a definite relation to each other, is duplicated in reverse at the opposite end. Since this is so, two meshes in the fabric are built up with each set of twining motions, and the actual work need progress only to the mid-point of the piece. At this point the problem is to keep the whole web from loosening and untwisting to its original state of plain stretched warps. In both specimens, when the twining turns from each end had approached to within an inch of each other, the fastening of the twists was made secure by "chaining" the yarns from one side edge to the other. With the series of yarns still held taut, pair number one was extended over pair number two and number two was pulled through the loop formed; pair number two was extended over pair number three and those yarns pulled up through the loop, and so on. The effect is a ridge of chain loops through the center with the pattern and ground meshes diverging in opposite directions from the ridge." To return to the Nazca fragments, specimen 170476e is the simpler of the two (Plate LUIa). It consists of pairs of dark Green (32 C 8) and Yellow (12 K 9) yarns in order, Yellow (1 pair), Green (4 pairs). The actual number of units, I should judge from a re- construction of the technique, is not important except that sufficient must be provided for the desired width. The arrangement of units at the center is reversed as follows: YY GG GG GG GG YY YY GG GG GG GG YY The right-hand Yellow element always turns over the left-hand Yellow element, but in manipulating the two pairs of Green, the effect of counterpairing is given by twining the inside elements (2 and 3) away from the center and over the outside elements (1 and 4). Between each two turns of any single active pair anywhere in the piece a passive pair is enclosed. The chevron pattern is formed by the appearance and disappearance of the pairs of Yellow twining elements. How closely these twined yarns were set may be appreci- ated by the width of the bands, %" complete. Majoro specimen 170476c (Plate LI 1 16) requires a set-up of 52 pairs of Green, Yellow, and Red yarns in the following order: GYGYGYGYGRRRRGYGYGYGYGRRR R— center This brings us to the center of the %" band, and to a reversal of the order of the yarns. There is nothing difficult about the actual manipulation of the strands in twining. Working 38 Tejidos del Periodo Primitivo de Paracas, Revista del Museo Nacional, vol. 1, pp. 60-80, Lima, Peru, 1932. 198 TEXTILES OF THE EARLY NAZCA PERIOD with long strands of twining yarns probably necessitated winding the ends into small balls to keep them from tangling. Or perhaps the twining at one end of a band of stretched warps duplicated itself at the other end. The placing of the colors results automatically in the production of lozenge shapes if the twining itself is correctly done. That is, it is possible to form diamonds of Yellow and Green, one within the other, only if the active Green pair, for instance, is active only so far as is necessary to form the side of its own diamond. When a corner is reached, the active twining pair must become passive, must be enclosed within a pair of yarns of the contrasting color. The drawing shows the changes each pair of strands makes in producing lozenge shapes. EMBROIDERED BANDS AND GARMENT FRAGMENTS Cantayo 171049, a fragment 16" long by V/% wide, was originally completely covered with rows of stem stitchery (Plate LVIIb). The stitch is one of the simplest in the em- broiderer's repertory, and is usually made without variations. If the thread is always kept either above the needle or below it, the surface of the embroidery takes on a twilled effect (Plate LXgr). Specimen 171049 has stitches of uniform length, taken over 4 warps or wefts, or their equivalent, depending upon the direction of the line. Since the warp-weft count of the cotton foundation band is 38 x 24 per inch, the wool embroidery surfacing is of fine quality. The ground is an Orange-Red (6 H 10), the motives similar to catfish (?) in Khaki Yellow (13 J 7), Yellow-Green (15 H 4), and dark Green-Blue (40 E 3). Cantayo fragment 171050 is also embroidered in stem stitchery which completely covers the cotton foundation (Plate LVIId). The work is even finer in this specimen than in the last due to the finer warp- weft count, 42 x 34 per inch. Lengths of stitches vary, floating as they do over 2, 3, and 4 foundation yarns. The ground is Rose Red (619). The design motives are block-S forms, about a half-inch high, and slanting at an angle of 45°. One complete repeat and a remnant of a second indicate a sequence of six colors: dark Green, Yellow, pale Green, Purple-Red, Yellow-Green, light Rose. Part of the second sequence has been lost in the deterioration of the wool yarns, but colors in the 3rd, 4th, 5th, and 6th places remain. Just beyond the solid embroidery on one long edge there is a line of small plant-like motives which give the effect of tabs. Each detail on these is made by a single stitch. Cantayo specimen 171054 consists of very small and fragile bits of embroidered bands, probably neck or armscye trimmings. What is left of them shows that they are similar to the better-preserved pieces described above: a fine weave cotton foundation material, which also gauges the quality of the embroidery, since it was customary to count the yarns for the stitches by twos and fours; small design motives made up of spirals and circles (possibly a flower), and at least four colors counting the Black outlines to the motives. This specimen illustrates one of the stem stitch variations: the stitches on one row have been worked with the thread consistently thrown above the needle; in the next row, the thread was kept below the needle (Plate LX/). The counterpairing produces a very different effect from the usual twilled surface of solidly worked stem stitches. The four specimen numbers, Cahuachi 171140a, b1-2, c, cover from 90-100 small bits and shreds of embroidery, the largest of which measures 5J4" by 2%", but they represent the finest stitchery in this Early Nazca collection. It is inevitable that they should remind one of the Paracas Necropolis embroideries both in their technical and design features, and an analysis of the fragments indicates the stylistic similarities between the embroideries from the two sites. Although these embroidered bits are listed under 171140a, b1-2, c, the four subdivisions will be treated as one since each contains a similar assortment of fragments. That is, frag- THE MISCELLANEOUS GROUP 199 merits in each of the four are wool embroidery on cotton foundation webs, and wool em- broidery on wool foundation webs. The Paracas garments are also very often combinations of separately woven cotton and wool materials embroidered in the same designs. But these Cahuachi all-wool remnants represent a style different from any Paracas style with which I am familiar. Dividing the fragments on the basis of foundation materials, we have evidence of two different garments, one Brownish-Black cotton with an embroidered border seamed to its edge, and the other White cotton with allover embroidery. There is a possibility, of course, that these two were originally parts of the same garment, but that seems unlikely. The Brownish-Black cotton fragment, about 23^" by 2", is a plain weave, warp-face material with yarn count 88 x 30 per inch. To its side selvage is whipped a shred of a Red wool embroidered border. This border is composed of three narrow bands, each separately woven, embroidered, then seamed together (Plate LVIIa). There is no doubt that the three parts were planned for this specific purpose, but why they should have been made separately is inexplicable. The two %' bands are warp-face set-ups of Red and Black warps (count 48 x 30 per inch), the Black yarns making narrow stripes near the outer edges. The em- broidered motives on these are germinating seeds, a familiar design element on Nazca pots and on Paracas Necropolis garments. The 1J4" center band is plain weave Red wool (count 40 x 40 per inch) embroidered with a series of birds and monsters with protruding tongues. The stitchery is the very simple stem stitch with thread consistently kept below the needle. The surface has the usual twilled effect. The group of fragments which seem to me to belong to a second garment or detail are equally well executed although a trifle coarser. This is due to the fact that the cotton foundation material, by contrast with the wool, has a yarn count of 32 x 32, and that stitches which float over four such yarns are of necessity longer than those that float over four smaller yarns in a cloth of higher count. The fragments suggest that the cotton garment was edged with tabs of at least three sizes and two types. The two larger sizes were 2" deep by 1XA!' wide, and 1}4" deep by 2" wide. The colors and motives seem to be the same. Tabs like these may have been made on the side edges of a web by using a skeleton warp (Plate LXV6). The regular wefts might then have extended out and around the skeleton. They would in turn become warps upon which to weave the tabs, a method suggested in the drawing. Or the tabs may have been made on the regular warps. The Paracas weavers made tabs on the ends of webs by adapting the Kelim tapestry technique to this use (Plate LXV6). Neither of these suggestions has been entered in Basic Table 30. Dissection proves that the Cahuachi tabs have three side selvages. They are bound with 3-stitch needle- knitting (Plate LXIIIe). The small tabs vary in size, but measure a little less than 1* by %". They are made of rows of buttonhole stitchery, 28 stitches by 28 courses per inch (Plate LXVa). The colors are the familiar embroidery and needleknitting colors: Reds (1 J 9 and 5 L 11), Brown (15 C 10), Greens (16 L 12, 23 E 5, 32 J 6), Blue (39 H 3). The Blacks are dark Greenish and Brownish-Blacks. By contrast with the quantities of sumptuous embroideries from the Paracas Necropolis these scanty bits from Cahuachi make a poor showing. It is not reasonable to suppose, however, that they represent the only fine work from the region. No doubt the original garments from which specimens 171140a, b1-2, c, came were unusual, but they need not be considered unique. Except for the joining together of the separately woven embroidered bands, which may indicate a weaver's personal preference, the basic materials, embroidery 200 TEXTILES OF THE EARLY NAZCA PERIOD techniques, quality of workmanship, coloring, and design motives approximate the elaborate style of the Paracas Necropolis garments. CORDS This name is arbitrarily given to three specimens: an example of 4-stitch tubular needleknitting, a 2-strand fragment of natural color maguey(?), and a flattened tubular piece made by an ingenious variation of a simple weaving technique. The 2-strand cord (Cahuachi 171124d) is just such a fragment as might be found among the useful binding materials of many a modern Indian group. In making the Peruvian specimen, approximately 16 single plies were twisted together, then doubled. Doubling preserved the twist for each ply, and gave twice the strength to the cord. The needleknitting fragment (Majoro 170476f2) was made in the same manner as has been described in the section on plain weave striped mantles with needleknitted edges (Plate LXVc). It consists of a core veneered with rows of stitchery which are worked round on round. In offering an analysis of the method of making the tail feathers and plant stalks of Cahuachi bird-flower fringes, the core was described as a twisted foundation strand of cotton. This extra element seemed necessary to furnish a little more body than a wool core of the same size could furnish. But, in the Majoro specimen, wool yarns of four colors are used alone. Each strand is active in its turn in forming the striped pattern, and while it is active the passive three serve as core. The Majoro fragment is only about four inches long, but that length is sufficient to illustrate one of the most interesting features of Peruvian color use: the combination of an unbroken repetition of three colors — Green alternating with Red and Yellow — and a simple repetition of two unit sizes — Yi repeated twice and %" repeated twelve times. In the following, the repeated letters standing for the colors give an approximate effect of the result: GGGG RRRR GYGRGYGRGYGR GGGG YYYY GRGYGRGY G R G Y— repeat The flattened tubular piece, Cahuachi 171180c, is a wool fragment 8%" long by about Y% wide. It is made of fine yarns, approximately 40 x 22 per inch. The weave is, for the most part, plain, although there are places in some of the design repeats where warp yarns float over two weft picks. The small rectangular motives suggest bird forms (Plate XLe). They are all alike, about an inch or more long, and appear on an Orange ground. The design areas in each rectangle are approximately the same, but the same colored yarns do not appear in adjacent motives nor do they appear in any two motives developing simul- taneously on opposite flat sides of the tubular cord. The reconstruction from which Plate XLd was drawn does not pretend to dupli- cate the actual area shapes of any of the partly deteriorated motives in the specimen. It demonstrates the technique by which such motives could have been woven. The warp set-up included 8 Orange yarns for each rounded edge, and 12 for each flat side, 40 in all. Then, to judge from appearances, there were 24 of each of the colored yarns used in the motives: Cream, Rose, Blue, Black. These 136 warps must all have been held taut and kept crowded together. A continuous Orange weft yarn wove over one, under one, spirally, enclosing the whole warp group. In reconstructing the piece I brought to the surface all the Orange warps and passed the weft under and over them alternately as in plain weave. Two such rounds of the weft made the solid color Orange bar between each two motive rectangles. In the Majoro specimen, each color area is outlined with Black. The drawing of the recon- struction does not show any outlines, but simply how by leaving the Orange yarns down, THE MISCELLANEOUS GROUP 201 any other desired warp color could be brought up from the encircled group to appear on the surface for as few or as many of the picks of the weft as was desired. Red seems to be part of each design motive, exchanging places from one flat side of the tube to the other with Blue and Cream. There is nothing complicated or difficult about the technique itself, but as Crawford has said, "of all weaving tricks tubular weaving seems the most unlikely for the primitive craftsman to stumble upon. Today the most common example of this class is the pillow slip. Yet in the Peruvian collection ... we find a narrow tubular ribbon in which the warps of great variety of color produce design. In order to make the texture solid the colors on one side have been drawn through to make designs on the other. This is an application of the principle of double cloth weaving."39 SLINGS Cahuachi specimens 171289a-g are slings and fragments of slings combining maguey and cotton. There are an intact specimen (Plate XLVIIIa), 2 more or less complete speci- mens bound together with their own cords, 2 intact maguey centers and 2 fragmentary centers, a finger loop and cord. Total: 3 slings; 4 identifiable center fragments. The rem- nants evidence that the slings were all practically alike in appearance. The 7" centers con- sist of two flat plaits of maguey fibers brought to a point at each end by a wrapping of 2-ply cotton cord. Stout 3-ply cotton cords about ^s" m diameter were also included in the wrappings. One of these 2- or 3-ply cords ends in a plaited finger loop of Blue, Yellow, and White wool yarns. The other cord ends in a flat plait "handle" of cotton. The three speci- mens on which these parts are intact are wound around with their own cords. Technically speaking, the slings are simple. The central, maguey portion is plaited by a method which can use any even number of strands (Plate XLVIIIc). Whatever the number chosen the strands are divided into two equal groups. The outside right strand crosses all of its group to the center; the outside left strand crosses all of its group, and also the former outside right strand. These two movements alternate without change. The result is a smooth plait over a core which is thin or thick according to the number and size of strands chosen. The center plaits of all the sling fragments are about y? wide by Vs"-M" thick. The finger loop by which the sling was held upon release of the stone is a flat braid %" wide. The loops on the two intact slings are alike in being bound to the ends of the hard 3-ply cords by wrapping, perhaps with the single plies of the cords themselves. The color effect of the loops — dissection of the specimens did not seem justified — can be gained by plaiting 32 strands with colors in the following order: 1-1-2-2-1-1-3-3-1-1-2-2-1-1-3-3, and reverse. The series of crosswise herringbone stripes is made by working alternately first with the outside right, then with the outside left strands: right over four strands, under three; left over four strands, under four, including the former right strand. The effect is similar to that shown in Plate XLII/. The "handle" on one of the bound slings (171289a) seems to be of considerable length. It is formed of a thick group of long cotton yarns held by its center in the twist made by doubling the end cord (Plate XLVIII6). The cotton yarns are then plaited into a 4-strand round plait as described in the section on turban bands and shown in Plate XLIIc.40 Cahuachi 171314 is a sling entirely different in type from the maguey-cotton slings (Plate XLVIIId). It is incomplete, without a hint of the lengths or forms of end finishes, " M. D. C. Crawford, Peruvian Textiles, Amer. Mus. Nat. Hist., Anthr. Papers, vol. 12, p. 95, 1915. 40 Compare with maguey sling from the Early period at Paracas, described and pictured by E. Yacovleff and J. C. Muelle, Revista del Museo Nacional, vol. I, p. 48. 202 TEXTILES OF THE EARLY NAZCA PERIOD but a change in the color of the yarns at one end indicates that there was a finish. The length, originally, was over 60", but how much over depends upon whether the two smaller fragments are parts of one end or two. Specimen 171314 is made of all-wool yarns, natural light Cream and dark Brown, probably llama wool, and dyed Yellow and Red. The center is in Kelim tapestry technique too fine in quality for the rope-like ends. The design is geometric, a series of hollow squares with stepped frames set diamond-wise, the spaces between them spanned by bars of the same Yellow. Where all the warps are constricted near the ends, the weave changes to plain striped tapestry. These same center portion warps are merged with the active elements in forming the rope-like ends. The technique used is called wrapped weaving, a term taken over from basketry.41 In the sling each warp element makes a single turn around a spiralling weft element by going down on the far side of the weft, and coming up on the near side. The same technique with change in the number of yarns or strands passed over is the stem stitch of embroidery already referred to many times. The semi-realistic serpent design on this sling is given added interest by the manner in which the dark and light blocks spiral around the rope. The smaller ends of the ropes are done in the same technique but the motive is simpli- fied to zigzag lines. Remnants of Yellow and Red yarns on one fragment end suggest tassels or deep fringe. NETTING Two netted specimens, Cahuachi 171118g and 171118h, illustrate the two simplest methods of constructing this type of fabric : knotless netting, or half -hitching, the coil without foundation of basketry; and the netting with simple or finger knot (Plate LXV-2; 171140a, b2; 171180c, d; 171217a; 171225; 171227a, b; 171238b1-2; 171258a'2, b; 171262; 171265; 171280; 171308; 171310; 171314 Cotton and wool 37 54 *9058; *9120; 170211a, b; 170462b; 170465a, e; 171045; 171049; 171050; 171054; 171059; 171071b, d; 171111; 171112; 171113; 171114; 171115; 171116a; 171117; 171119; 171125; 171140b1, c; 171180a, b'-2; 171181; 171215; 171216; 171218a; 171219a; 171220; 171221; 171222; 171223a, b; 171224; 171226; 171237; 171266a, b; 171267a; 171289f, g;J 171305c; 171309; 171321a, b1-5 Warp-weft techniques Plain weave, number of specimens*; 95 114 1. 1 warp, 1 weft 91 110 *9056; *9120; 170211b, c'-2, d'-*, f, g, h, i; 170413a, b; 170462a, b; 170465a, b, c, d; 170476a, b, f1; 170665; 170677a, b; 171033; 171045; 171049; 171050; 171054; 171055; 171059; 171071a1-2, b, c, d; 171109b1-'; 171110; 171112; 171118a, c, d,e, P; 171119; 171124a; 171140a, b'-2, c; 171141; 171180d; 171181; 171182a, b; 171183; 171213; 171214; 171215; 171216; 171217a, b'-2; 171218a, b; 171219a, b; 171220; 171221; 171222; 171223a, b; 171224; 171225; 171226; 171227a, b; 171236; 171237; 171238a, b'-=; 171258a1-2, b; 171262; 171265; 171266a, b; 171267a, b; 171279a, b; 171280; 171305a, b, c; 171306; 171308; 171310; 171311; 171312; 171330 la. 2 warps, 1 weft tt lb. 1 warp, 2 wefts 1+. Interlocking warps, wefts 4 4 170211e; 171111; 171221;** 171308 Basket type t The difference in the two totals is mainly due to giving individual numbers to specimens technically identical, but different in quality or color. They may be recognized by the letters and superior figures. t Specimen contains maguey fiber. 1 Frequencies of subvarieties of a process may total higher than the figure given for the process, because of co-occur- rences in one specimen of several subvarieties. tt The blanks indicate the non-occurrence of techniques found in textiles of the Middle and Late periods. •• Interlocking wefts only. TABLE 30 (continued) Tapestry, number of specimens^ 12 12 Monochrome 4 4 171117; 171119; 171217a; 171265 Kelim 5 5 170465e; 171217a; 171225; 171308; 171314 Eccentric 3 3 170465e; 171217a; 171308 Interlocking weft 1? 1 171125 Underfloat weft Single warps wound 1 1 171225 Figure 8 3 3 171045; 171180b2; 171217a Twill types Double cloth 1 1 2 warps, 2 wefts; variants 1 1 170476d 2 warps, 1 weft Pattern weave, number of specimens 12 10 Single-face, underfloat warps 4 4 171059; 171118e; 171180c; 171226 Single-face warp and weft floats 4 171218b; 171236; 171279a, b Double-face, 1 warp, 2 wefts 5 1 171119 Double-face, 2 warps, 1 weft 3 1 170476f* Wrapped weave 1 1 Single weft 1 1 171314 Multiple weft Gauze weave 3 3 170476fl; 171110; 171141 Single-element techniques Half-hitching, coil without foundation 2 2 171071e; 171118g Netting, with knots 1 1 171118h Knitting 25 || Plaiting: braids, number of specimens^ 9 15 Round 3 6 171118f3; 171124b1-4; 171289a Flat 11 12 170211d«; 170476d, f1 (2); 171109c; 171116b; 171118P (2); 171124c1-2; 171217a; 171289a Square Plaited finish of warps, basketry type Twine-plaiting, "lace" 6 4 *8537; 170476c, d, e Weave-plaiting, cords 1 ... If Frequencies of subvarieties of a process may total higher than the figure given for the process, because of co-occur- rences in one specimen of several subvarieties. || Reclassified under needleknitting embroidery. 209 TABLE 30 (continued) Superstructural techniques Brocade, number of specimens 3 3 Single-face Double-face (or embroidery^ 3 3 171218a; 171219a; 171220 Edge finishes, number of specimens'1 16 45 Fringes Applied, needlemade 7 6 171033; 171071d; 171117; 171180d; 171265; 171309 Applied, extra length weft to skeleton warp 4 5 171118b; 171217a; 171262; 171266b; 171308 Warps left unwoven 6 6 170211f; 171217a; 171225; 171226; 171266b; 171308 Tassels 3 2 *9120; 171314 Needleknitted cords, tabbed and fringed 4 171033 ;Jt 171125; 171262; 171265 Needleknitting, 3-dimensional edge trims, etc 11 11 *9058; 170211a; 171112; 171113; 171114; 171180a, b1-2; 171215; 171223a, b; 171224; 171237 Tabs 5 7 Woven 5 2 171140b1-2 170211b; 171071b; 171116a; 171140c; 171180d; 171309 Scallops: needlemade; woven 3 171045; 171109a1-2 Stitchery Seaming, number of specimens^} 27 Whipping; zigzag; "sham" hem stitch 26 22 170211f; 170462b; 170476f1; 170677a, b; 171033; 171059; 171125; 171140c; 171214; 171217a; 171224; 171226; 171237; 171262; 171265; 171266a, b; 171280; 171305a; 171308; 171330 Saddler's, lacing 2 5 171119; 171217a; 171262; 171308; 171310 Running 4 2 170465c; 171306 Hemming 2 171215; 171266b Wrapping of core "Warps" added to edge 2 2 171033; 171045 Embroidery, number of specimens*! 25 48 Blanket stitch 1 171266b Stem, outline 12 15 170465a; 171033; 171045; 171049; 171050; 171054; 171140a, b1-2, c; 171180d; 171218a; 171219a; 171220; 171266b Needleknitting 3 36 *9058; 170211a, b; 170476^; 171033; 171071b, d; 171112; 171113; 171114; 171115; 171116a; 171117; 171125; 171140b1-2, c; 171180a, b1-2; 171181; 171215; 171223a, b; 171224; 171237; 171262; 171265; 171267b; 171309; 171321a, b'-5 Couching Chain 3 ... f Frequencies of subvarieties of a process may total higher than the figure given for the process, because of co-occur- rences in one specimen of several subvarieties. it Similar to others in the group in effect; center an embroidered band. 210 TABLE 30 (continued) Figure 8 Seaming stitches for embroidery 7 4 Double running 3 *9120; 170465a; 171216 Tent hemming 1 171222 Twined stitches 1 171266a Devices to vary effect Yarn spinning Slack twist Crepe twist 9 10 170413a; 170476a, b; 170677a, b; 171182a; 171237; 171262; 171265; 171305c Two tone 2 170211c1-2 Structural, set-up of loom Warp face, plain weave 25 78 *9056; *9120; 170211b, d1-5, f, g; 170413a, b; 170462a, b; 170465a, b, c, d; 170476a, b; 170665; 170677a, b; 171033; 171049; 171050; 171055; 171059; 171071a2, b, d; 171109b1-4; 171110; 171112; 171118a, c, e, f2; 171124a; 171140b2, c; 171180d; 171183; 171213; 171215; 171217a, b2; 171218a; 171219b; 171220; 171222; 171223a, b; 171225; 171227a, b; 171236; 171237; 171238b1-2; 171258a1-2, b; 171262; 171266a, b; 171267a; 171279b; 171280; 171305a, b; 171306; 171308; 171310; 171311 Drawing in for stripes, patterns 34 49 170211c1-2, d1-*, e; 170462b; 170465b, d; 170476d, f1; 170665; 171033; 171059; 171071a1; 171109b1-4; 171111; 171118e, f2; 171124a; 171140b2; 171180c; 171183; 171218b; 171224; 171225; 171226; 171227a, b; 171236; 171238b1-2; 171258a1-2, b; 171262; 171265; 171279a, b; 171280; 171305b; 171308; 171310; 171311 Scaffolding weft for interlocking plain weave 2 3 170211e; 171111; 171308 Scaffolding weft for shaping web Warp locking, end-to-end 3 3 170211e; 171111; 171308 Tab formation • 5 2 171140b1-2 Loom joining of widths 2 2? 171224(?); 171262(?) Spaced warps 3 2 171224; 171225 Tubular construction 1 1 171180c Warp element manipulation Crossing against slip 1 171119 Grouping for tapestry 3 171119; 171217a; 171308 Weft element manipulation Color changes for cross stripes 8 11 170211c1-2, e; 171071a1; 171111; 171218b; 171236; 171279a, b; 171305c; 171311 Weft grouping, for size 1 1 171314 Weft lock, as in tapestry 2 2 170476d; 171221 Warp lock, as in tapestry 2U TABLE 30 (continued) Warp-weft lock, as in tapestry 1 1 171125 Counterpairing of wefts 2 170211f; 171217a "Facing" one color with another Double set of weft, plain weave 3 2 171118^; 171124a Weaving techniques Loose beating up 5 6 *9056; 170211c1-2, h; 170476a, b Weft-to-warp change 1 1 171225 Padding yarns introduced Kelim slot for neck opening 2 3 171071b, d; 171266b Single element manipulation Three-dimensional knitting 11 Counterpairing in embroidery, twining 1 3 170476c, e; 171054 Interlocking of embroidery yarns 4 170465a; 171218a; 171219a; 171220 Plaiting element manipulation Color variations in braids 11 170476d, P; 171116b; 171118f3; 171124b1-4; 171217a; 171289f, g Surface decoration Painting 1 1 *9056 Tie-dyeing Feathers, applied 1 2 171071c; 171266b || Reclassified under needleknitting embroidery. 212 CONCLUSIONS We may conclude from the 163 specimens in the Early Nazca collection that the weavers of the period not only had and knew how to handle fine cotton and wool yarns, dyes, weaving, and stitchery techniques, but that they were familiar with the variety of possibilities offered by combinations. Although the examples are few in number, there are garments illustrating from one to fourteen techniques. Most of them are basic, yet many are adaptations to designs or uses. Very little of the workmanship is slovenly, and none of it gives the impression of being experimental. Even by today's standards of fineness and uniformity, yarns, weaves, and embroidery meet the test. Table 30 weights a simple whipped seam equally with a complicated edge finish: each counts one. On that basis, many specimens are entered once, twice, or more times in the table. The fragmentary condition of a good part of the collection has been mentioned. This condition is undoubtedly responsible for the numbers of specimens falling in the groups of the fewest techniques, but it is not responsible for the great preponderance of plain weaves. As in all other Peruvian periods, the plain weave with its obvious color variations was a fundamental technique. Under single technique classification, plain weave is not first in frequency, because, in all likelihood, a rectangle with no decorative features was rare, and the fragments lack edges or corners showing the combinations made in the original garment. Table 31 summarizes by numbers and percentages the material given in Table 30. 213 TABLE 31 Range of Technological Processes Found in the Early Nazca Specimens Number of techniques Number of Per cent of Number of techniques represented Number of represented in specimens collection by all specimens in group specimens each specimen 1 33 20 9: needleknitting plain weave plaiting various (6) 14 5 5 9 2 31 19 17: plain weave + 1 other technique plaiting + color variation various combinations of 2 techniques 20 5 6 3 47 29 22: plain weave + warp face + stripes plain weave-j-warp face+stitchery various combinations of 3 techniques 21 9 17 4 ■ 18 11 23: plain weave + warp face +2 techniques various combinations of 4 techniques 15 3 5 20 12 27: plain weave +4 other techniques various combinations of 5 techniques 17 3 6 5 171119 171140b1- 171140c 171224 3 15: plain weave +5 other techniques 5 7 2 170476d 170476^ 1 11: double cloth +6 other techniques plain weave +6 other techniques 2 8 2 171225 171265 1 14: plain weave +7 other techniques 2 9 1 171033 0.6 9: plain weave +8 other techniques 1 10 2 171262 171266b 1 15: plain weave +9 other techniques 2 13 1 171308 0.6 13 1 14 1 171217a 0.6 14 1 Totals 163 163 214 GLOSSARY OF TERMS WITH DIRECT APPLICATION TO THE EARLY NAZCA COLLECTION Apron: a rectangular web with strings or ties from the corners (Plate XXXIVa). Batten (also called sword) : the shaped piece of wood by means of which the shed is kept open for the weft yarn, and each inserted weft is driven down to the partially woven web. The word is both noun and verb. Blanket stitch (also called coil without foundation, half-hitching, and buttonhole stitch) : an embroid- ery stitch used for protecting or reinforcing edges (Plate LXa). Used in Early Nazca textiles for founda- tions under needleknitting veneer, and for tabs (Plates LXIIIa, LXVa). Brocade: a form of superstructural patterning in which supplementary yarns develop design motives by means of floats. The extra yarns usually alternate with the basic yarns of the fabric. In double-face brocading the pattern floats are as important for the reverse side as for the surface side. Brocaded mate- rials are often indistinguishable from embroideries (Plates L-LII). Buttonhole stitch: see Blanket stitch. Count: the number of warp and weft yarns per unit of measurement, the inch. For closely woven materials, count indicates the quality of the fabric. Counterpairing : in twining or weaving, the turning of the two elements of adjacent pairs toward or away from each other, thereby forming a plaited effect (Plates LI I la, LX6) ; in stem stitch embroidery, the throwing of the thread first above and then below the needle on alternate stitches (Plate LX/). Crepe twist: an extra amount of twist given in yarn spinning, which results in a pebbly surface of the woven fabric upon release from the loom. Crossing against slip: the exchanging of the regular positions of the warp yarns with neighboring warps to insure keeping in place the last wefts put across the web; usually done between plain weaving and a tapestry border (Plate LXVIIIc). Double cloth: a reversible fabric requiring two sets of warps arranged one above the other, each with its own weft. Ordinarily, the sets are of different colors. To make the pattern, certain reverse-side warps are raised to replace surface-side warps which are lowered. Colors are exchanged, and ties are formed between otherwise separate portions of the fabric. Since each set of warps, no matter what its position, is crossed only by its own weft of the same color, strongly contrasted design areas are produced (Plate XLIIa). Double running stitch (also called punto scrito) : an embroidery stitch in which the material passed over by the needle on the first line of running stitches is covered by a second line (Plate LXc, h, i). Double set of wefts: two wefts enter the shed from opposite sides, and cross each other in the shed (Plate XLa). Drawing in for stripes, patterns: setting up the loom with colored warp yarns (Plates XXXVI, XXXVII). Finger knot: the simplest knot which may be tied with a single element; it gets its name from the method of turning the element around the left forefinger in order to make a loop (Plate LXVA). Float: a warp or weft yarn free for a distance upon the surface of the fabric. Patterns are built up by means of floats. Fringes: extra lengths of wefts. Most fringes have plain-weave tape-like headings from 2-5 warps wide. The fringe proper is the extra length of weft which turns about a skeleton warp set the desired distance from the other warps. Upon completion of the weaving, the skeleton warp is withdrawn, leaving tightly twisted weft loops. A similar fringe is made by the extension of the regular weft to a scaffold yarn (Plate LXVId). Needlemade. The Nazca and Paracas types are similar: loops made by drawing an extra yarn through whipping stitches on the edge of the garment are twisted tightly into a fringe (Plate LXVIc). Unwoven warp lengths. See Plate LXVIo. 215 216 TEXTILES OF THE EARLY NAZCA PERIOD Gauze weave: manipulation of certain warps by drawing 1 and 3 over 2 and 4, and securing the cross with a passage of the regular weft to produce the effect of openwork. The Peruvian gauze is used in combination with plain weave (Plate XLIo-c). Half-hitching: see Blanket stitch. Plate LXVgi shows use in netting. Heddle: a device for separating the warps into what are called sheds for the insertion of weft. The primitive heddle, also called heald, is a rod from which string loops depend to encircle alternate warps. When the rod is drawn up, the odd or even warps are separated from the other half of the set-up. Hemming stitch: a seaming stitch used for fastening down edges (Plate LXVIIA). Interlocking: embroidery yarns. A method by which yarns of two colors, instead of building up their respective motives independent of each other, loop about each other at the common boundary line (Plate LXVIIIe). Warps and wefts. Multicolored patchwork constructed by means of skeleton warps and wefts around which the set-up yarns turn. The weaving is so accomplished that the scaffolding yarns are unnecessary upon completion of the fabric, since both basic sets interlock (Plates XXXIXa, c, LXVIIgr). Warp-weft type. A technique normally found in tapestry. Two wefts interlock with each other, at the same time enclosing a warp (Plate LXIVe). Weft yarns. A method by which color change is effected at the boundary of a design motive; normally found in tapestry (Plate XXXIX6). Kelim slot: an opening of the desired length left between two adjacent warps by turning around them the weft yarns which come from opposite sides of the web (Plate LXVIIIjr). Lacing stitch: see Saddler's stitch. Loomjoin: done with an extra length of yarn independent of the weaving elements, which draws together a breadth of cloth already woven and a breadth on the loom by engaging corresponding weft turns on the adjacent edges (Plate LXVII/). Loom strings: the first two, three, or more heavy wefts put across the warp yarns (Plate LXVIo). Mantle: the largest rectangular garment found among the Early Nazca pieces; usually formed by joining two breadths of material. Patterned in the loom, embroidered, and edge-trimmed (Plate XXXIV6). Needleknitting : a name given to the modern plaited cross stitch which is known to embroiderers in various parts of the world as Ceylon stitch, Portuguese stitch, etc. The Peruvian variety is identical to the modern stitch in method of making (Plates LXI-LXIII). Needleknitted cords, tabbed and fringed : cords with tiny buttonhole stitch tabs on one side, fringes on the other (Plate LXIV). Needleknitting, three-dimensional : flower and bird forms developed in the needleknitting technique over woven or blanket-stitch foundations (Plates LXI-LXIII). Pattern weave: double-face, 1 warp, 2 wefts: design motives made by weft floats independent of the warp set-up. In these motives the bobbin carrying the decorative weft floats over and under groups of warps. The weave resembles brocading except for the fact that there are no alternating basic weft yarns in the pattern (Plate XLIII6). Double-face, 2 warps, 1 weft: design developed by warp floats, a method of patterning dependent upon the colors set up for warps (Plate XLI6). Single-face, underfloat warps: a fabric with design developed by the appearance of certain warps on the surface of the web; when these warps are not necessary to the pattern, they float on the reverse side (Plates XLrf, e, XLIIIa, c-e, XLIV6). Single-face, warp and weft floats: a fabric with design dependent upon arrangement of colors in the set-up, and in their subsequent shedding. The latter accounts for the warp floats, and the weft floats are made by carrying the bobbin yarn over and under groups of warps (Plates XLIVa, XLVIo-/, XLVII). Pick: one length of weft from side to side of a design motive (in tapestry), or of a breadth of fabric in the loom. The term is used both as a noun and as a verb meaning to insert the weft in the shed. Plain weave: the interlacing of a single weft yarn over and under single warps. Plaiting: synonymous with braiding. GLOSSARY 217 Running stitch: the simplest of all seaming and embroidery stitches; identical to darning. Rarely found on Peruvian garments of any period (Plate LXVIIi). Saddler's stitch: a seaming stitch which laces the edges of two webs together without overlapping (Plate LXVIIc). Scaffolding yarns: warp or weft yarns forming temporary foundation elements for multicolored patchwork fabrics, end-to-end warp locking, and separately woven fringes and tabs (Plates XXXIXa, c, LXIVa, c-e, LXV6, LXVIa, d). Scallops: needlemade: an edge finish formed by weaving over and under with the needle between the selvage and a loop of yarn made as required (Plate LXVe, /). Woven: a unique form resulting from weaving narrow bands on two warps, the number of wefts indeterminate. In the Nazca example the wefts are carried over and under the two warps in the following order: 1-2-3-4-5-6*6-5-4-3-2-1 (Plate XLc, /). Set-up: see Warps. "Sham" hem stitch: a seaming stitch (Plate LXVIId). Shed: the triangular space produced by raising alternate warps (in plain weaving), or groups of warps (in pattern weaving), by means of a heddle device. The weft is carried through this space from side to side of the web. Skeleton yarns: see Scaffolding yams. Sling: various types both simple and elaborate are found in sites from the different Peruvian periods. Technologically, slings are alike in having slotted or net-like center portions, long plaited or wrapped- weave cords from each end, finger loops and tassels as finials for the cords (Plate XLVIII). Square count: an equal number of warps and wefts per unit of measurement, the inch. Stem stitch (also called crewel and outline stitch) : a simple line embroidery stitch made similar to back stitch in seaming, wrapped weave in basketry, and Soumak stitch in rug work. Usually the work proceeds from left to right by entering the needle between the 4th and 5th yarns on the surface side, and bringing it up from the reverse side between the 2nd and 3rd yarns (Plate LXd-g). The quality of work, whether fine or coarse, is dependent upon the warp-weft count of the fabric. Most of the Peruvian stem stitch embroidery appears as solid filling for pattern motives. Stock-dyed fibers: raw stock, usually wool, dyed before spinning it into yarn. Superstructural techniques: those not functional to the making of the basic web. In brocade and embroidery the decorative yarns do not constitute an essential part of the structure. Tabs: needlemade: blanket or buttonhole stitches worked row on row (Plate LXVa). Woven : Plate LXV6 suggests two types dependent upon scaffolding yarns. One type is made by an extension of the regular wefts which become warps; a second type is made by weaving upon warp loops leaving Kelim slots between adjacent tabs. Tapestry: eccentric: a type in which warps and wefts interlace at angles other than right angles. Generally used to develop irregularly shaped design motives. Figure-8: a type in which the weft interlaces as in plain weave with two single warps or two groups of warps (Plate LXIVa). Kelim : a type distinguished by slots at the sides of pattern areas. The wefts of one color turn on an edge warp ; those of the adjoining color turn on the adjacent warp. The length of the slot is governed by the size of the color area (Plates XXXVIIIa, d, LXVIIId). Monochrome: a plain weave in which wefts of one color are battened together so closely as to com- pletely cover the warps. Single warps wound: a method by which a very narrow line of color is produced. A single warp or a single warp group is wrapped closely with weft yarn (Plates XXXVIIIa, LXVIIId). Tent stitch: a slanting stitch used by the embroiderer to veneer single yarns of the basic material by winding variously colored threads about them (Plate LIX6). Tunic: a sleeveless shirt formed of one or two breadths of material. An opening for the head is - '. - TEXTILES OF THE EARLY XAZCA PERIOD left in the center seam, or provided for by a Ketim sloe The side edges are seamed to allow for medium size annscyes (Plate XXXIYc-e). Taiban bamds: very narrow colored ribbons in weaving or plaiting techniques Plate XL). Twine plaiting: lace" constructed of pairs of yarns which twine about each other for distances dependent upon the design. Where two pairs meet to cross, each separates and plaits as in modern bobbin bee (Plates LIII, UV). Twined stitches: made with two weft yarns, or with yarns in needles as in true embroidery. The introduced strands twist a half turn about each other, enclosing groups of warps between each two inter- sections. This is the simplest two-strand twine (Plates LX6, LXYIIIa). Warp; the yarns stretched on the loom preparatory to weaving. The completed warp series is called the set-up. Warp face: the appearance given to a fabric by a preponderance of warp yarns over the number of weft yarns per inch. Most of the striped materials in the Early Xazca collection are warp face. Warps added to edge: devices by which dements were provided for weaving very small edge details (Plates LXTVe, LXVe). Warp element manipulation: management of the warp yarns so as to farm patterns by means of warp floats (Plates XLIII-XLYII). We*: a textile fabric, especially one under construction on the loom. Weft (also called woof, filling, pick): the yarn carried by a bobbin or shuttle from one side to the other across the warps in weaving. Weft element manipulation: management of the weft yarns so as to form patterns by means of floats, as in Naming The majority of brocaded fabrics are <*Ta"»pL>g of weft manipulation • Plates L-LI I . Weft face: the appearance given to a fabric by a preponderance of weft yarns over the number of warp yarns per inch. Tapestry is the most extreme example of a weft-face fabric. Weft-to-warp change: presupposes a space in the warp set-up across winch the regular weft was carried. Upon completion of the weaving, these bare wefts were woven upon (probably with needles as in other fine tapestry work) as if they were warps Plates XXXYIIIa, LXYIILf . Whipping stitch: a very shallow wrapping stitch taken at right angles to two fabric edges. A whipped seam opens out flat. The stitch was most frequently used by the ancient weavers to fasten together separately woven or constructed parts of garments. In several specimens possibly loom- joined I it is found taken between the first and second edge warps and through each of the corresponding weft ; on the two breadths. Probably, in those instances, the stitch drew the two edges together as in (Plate LXY1I/). Wrapped weate (also called Soumak stitch): farmed by carrying the weft element forward and once around a single warp at a time. Similar to stem stitebery which is done on fabric, in contrast to wrapping on bare warps (Plate XLVIIM, detail l. Zigzag stitch: a decorative aiming stitch with the advantage of strength. Plate LXYU6 shows the method of making: the needle is brought out at A, is inserted at B, is brought out a second time at A; the needle is then inserted at C, is brought out at D, is again inserted at C, is brought out a second time at D, and repeat. Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XXXIV rfrzsz^ '"Mii ilii I lull I I EARLY NAZCA PERIOD GARMENTS > X! X X * E O > o ! o — c .2 2 3 55 o E 3 s Q 55 « Q , ■n+fr 1 3& si I i i T w m m 171305 6 I7M83 I7I5II 17046 5 -•? Iii; III k Ii ii|! I .Mi ♦i.i' lni j'ti I'1 I'ii V'l ■l" I 1:1 I'l'l ill! !'l I'll] I'll '!' !i Ii, i ! ! Ii ," I1 1 till! !i!!l r'l" s III!] ' I I "II I!1" In, [llll I'M • I, 'M : I'M IH ::; !« i^i pin '■ l,i' in' jn| !! i 1 I 'ill 'III i I i'i'Ii 1 hi Ml, 'i1' V'l III ITO4€5-B| STRIPED COTTON MATERIALS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XXXVII STRIPED WOOL MATERIALS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XXXVIII d VARIATIONS IN TAPESTRY TECHNIQUE Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XXXIX INTERLOCKING IN PLAIN WEAVES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XL I7lll8f* 171180c WOVEN BANDS AND CORDS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XLI I70476f' 111110 — * GAUZE WEAVES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XLII musf2 PLAITED BANDS AND CORDS X Cu, t— 1 "3 > g '5 § ■ : ($ 2 -, ^'' < c _ ;.-.:-.s:. »■:.«. B w 05 a Q es o ca u > w is z E- - o o £ S H Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XLIV 17I27S-A | mm an: - m «< fcE*i ^ »"£« 'X' "l:m-j"'.l."" m am **t. vwm /tr. V** ' rr/i «W Z&te '<" "" ,' ' met ' .•*• . <<*.<• ]~*i.\ I x«r •"S\ _ via ^f.% **\\<— a- FLAT VIEWS OF PATTERN WEAVES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XLV a a a ima a aiaia a aiaia • ■ a aiaia a a a aiaiaiaia ■ a a a a aiaum! a a a aiaia a aiaiaiaia a aiaia a aiai mill mtWm\ n\ ■■■■■■■■•lalilaiitili ■■■iiilillllllill iIiiiiiiiiiIi III aula a a alliaiiilil III! • aaaiaiaima aiaia a aiaia la i a a a • aiaiaiaiaiaia a a a aiaiaiaia ■ imaiaia a ■ • amiiaiaiaia a a a iiaia a laiaia lal :: ■ a a ■ ■ in a aiaia a aiaia a la a aian |a|alallillllliala|a|ala|a|aiaiaiaiilalala|aiaia a a a Jami a aiaia la laiaia a i^^i:=ilSi:i5iiHliS;pa|iL ■la la.aia a a a aiaiaiaiaiaia a la a • L FLAT VIEWS OF PATTERN WEAVES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XLVI l"l f*l f"T f*1 r'1 ■"• IP* *'* »"* *'^ "* ,.}J ,.!.. ,.J.a ,.j.l ,.jj T.;.j r + J| ,.lj r.jJ r..l r.j •J r+J »•:-« -■»"• ■■■M -<•' r-b-» «■■«- r-J-Ji r-t-* r + - riy '■{-» H-« •■*-* '■{-* •■<•' *<•* r*"" '+* r+* \ r-j"J h-j n** H-* »■<•* '•;-* »-s-« r-s-* r-s-j r-s-* h ... ,-,.i ,JJ *■{•* p-5-J »■!•' r-S-« r-jl-* •"!"" H-* r1*J r-5-' r-<-' r<— '■!••" !■+■• ,■-{-* «■»!" ft-* »•!-* rS"* » r-H y-H '<■' r+ rX-J r-ir* t"^J r^*J 171236 I e, surface »J r-».i ,.;. ■' --!-* r-t-» «•*' ;-:-» -- r A-* r^J»* r^l-* f>"* rv rv' f, reverse FLAT VIEWS OF PATTERN WEAVES > -1 X •S s o > o E 01 o o o .a a - z o £ 3 2 "3 :_. I — 1 w — > «*! W •s i l I 55 Ed H H <: Oh i I S < fa O H l-H > ■in 1 H in j 1 3 fa in' I !ljl l i — : — _,>*: — _ Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate XLVIII SLINGS AND SLING DETAILS X X S o > S2 o "o a o o .a z o £ 3 3 2 E £ a > i '5 £ a p c H fr- iz I— I O 2 Q W Q < o OS 05 3 E £ o £ a a s I 1 z "8 £ 3 n w <: a t» » E O a a Q OS o n Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LII 171Z18A , 1TUI9 A' IJlilgA1 1 1" 1 BORDER MOTIVES IN MANTLE 171220 Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LIII 1704 76 e 1704 lb C TWINE-PLAITED BANDS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LIV TWINED "LACE" 4-fl»5J7 T Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LV Oil 09 2MB 3D H u&e 9li0 TBI T,l^) EMBROIDERED KERCHIEFS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LVI I7IZ66 b i7iztfcb EMBROIDERED TUNIC BANDS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LVII l7IKOabc 171180 d EMBROIDERED BANDS Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LVIII ams I7IZI6 171216 1 71216 EMBROIDERED MOTIVES ON MANTLE 171216 Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LIX U'—M C/i-rvJ Luu irnifn irnir ' Luu LC—fJ ut/O L/l jt, LWU LJ^iU LJ tJu (ajOuljOu-jvK i r*~*i n^v) rwcj |Lq4l UJLD j ! 1/2.. a.' Ui..s\.i L'7..:vJ !7liZ2 I'll <■. : '■>:•■! •■> <■> «■>•■.■> '■>!■■ 3R . i»m rggrr»sn' aEpanii I) 1PB IVSJ T*«* WW I7UZZ EMBROIDERED MANTLE 171222 X E o > o S ii o & o c <3 >> »-< o £ 3 u O s 3 0) ca 3 "3 E in h ■— z w o H « H Q o « pa 3 H Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LXI ni224 — profile b ni224 — reverse C ■ajgpf-JH»!-4fr. frfeJWT Wit J". *~4 ruu b h i , NEEDLEKNITTED BANDS AND FRINGES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LXII iruaob' 171071 b iiiiiiii I70ZH b NEEDLEKNITTED BANDS AND FRINGES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LXIII f 9 NEEDLEKNITTING TECHNIQUES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LXIV EDGE CORDS, TABBED AND FRINGED Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LXV 71140 b'-' 35at3BU H ■ tioaail ■ (IttaaaO ■ 11 tjanatfll ivaaanUaadc: :J3r2^jJSB3B^33zi3«B33anc]SBBnajaaBBBBnaciaac nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnr llllb! lTO+7«r gggaa DOC3D3 DjjaDDDDaaDaaa DDDDaDooDaaQDD-- D0O0aQODDa0a0QDD3DDQDnOa0 QuaQOuoaooDaoapQODaoooQDp miieg 1 71118 K MISCELLANEOUS TECHNIQUES Field Museum of Natural History mitt to ^SgBSsri 1 fc Anthropology, Memoirs, Vol. II, Plate LXVI saansn: a3=^s=s^* 17'OTio 3D"" 10. FRINGE TECHNIQUES Field Museum of Natural History- Anthropology, Memoirs, Vol. II, Plate LXVII 1 170*77 b B^B Mr* 1712(4 170465 c I'OWlb SEAMING STITCHERY TECHNIQUES Field Museum of Natural History Anthropology, Memoirs, Vol. II, Plate LXVIII iptttiq i»mi» Sxjudbuuupuuuuuu JODaDUUCJLiDUlJUtjQU janaauaucuuuDUULi jnnaDDuucLiuuuci-' ]o^c3uuuul:uuu' SannouuuLCLHJu )rrlnnffT-»»jYi.. - joaBuatJH joaai — 5oaai__ Satmauu Spuuniiiiuii... iuuuouuuuuu UUUr- juuuuuuuuuuuiiutjnij: H>2t«A I1I2I9A* UULJULJUi.'L'ULKJlltjIJUUUUlJIIUIIlJUlJUl yp.y^ULiyi'UliUUy'JUUflUULI'JULJJLJUL ^ffsllff i»ion