The join is a technical feature that indicates standards of craftsmanship. It is customary in weaving materials with end as well as side selvages to give more or less attention to the closing of the s.p.a.ce between the weaving proper and the heading strip. When the warps in the form of a skein had been spread out evenly and bound in place to the end bars, the ancient weavers on two-bar looms first wove a shallow heading strip to secure the warps in their positions and to establish the ultimate width of the fabric, a practice followed by some modern weavers today. Then the weaver reversed the loom end for end to begin what became the weaving proper, and continued until the length was complete. Difficulties or indifference to appearance very often resulted in a general looseness of texture where standard-size tools had to be removed and the interlacing done more or less by the fingers. Plate 2,a, b shows heading strips of different depths, fairly wide join areas in which the wefts are more widely s.p.a.ced, and above these, the compact texture of the weaving proper.

Three finely woven cloths, one of them shown in plate 5,d, exemplify warp locking, end-to-end. This technique is known from the earliest periods on the coast in the so-called patchworks from Nazca Valley graves. It occurs also in Middle- and Late-period textiles.[5] The methods of lengthening the warps by the addition of new ones vary, but one feature is common to all those examined: the supplementary transverse yarns are in effect scaffold or skeleton wefts.[6] In the Chincha cloths, the two warps interlock as shown in the reconstruction in plate 5,a. In two Chincha plain-weave cloths, as in the Nazca patchworks, the warps of two colors meet on the skeleton weft.

Two specimens in lot 4- (3890a and 4056) are poor in quality of craftsmanship. Careless weaving resulted in the breaking of several warps, uneven shedding, and puckering in the center of the web. A three-inch difference in the length between the two side edges of specimen 4-4056 was probably due to slanting of the warping stakes (fig. 6). There is also a difference between the widths of the ends of each cloth, in one of them as much as three inches. Different weights of yarn are used, their twists ranging from soft-to-medium to crepe.

[Ill.u.s.tration: =4-4056= Fig. 6. Diagram of a web showing an irregular shape which may have resulted from careless warping. Occurrences of plain-weave variations are indicated by symbols for units: +, one-by-one; =

, one-by-two; ++, two-by-one; and #, two-by-two (27 in. 26 in.

over all).]

In specimen 4-3890a the warps were grouped in pairs throughout the breadth of the cloth. In the first eight and one-half inches of the length, the weft is single and for the remaining fifteen and one-half inches the wefts are paired. This results in plain-weave variations of two-by-one, or semibasket weave, and two-by-two, or basket weave.

In setting up the loom for specimen 4-4056, twelve inches of the breadth were warped with units of single 2-ply warp yarns (fig. 6, right) and the remaining fourteen inches were set up for units of twin warps (fig. 6, left). Several plain-weave variations were found. The weaver introduced single and twin wefts at irregular intervals throughout the length of the cloth. Therefore, in the portion where the single warp unit interlaces with single wefts, a simple one-by-one, or plain weave results; where the single warp unit interlaces with twin wefts, a one-by-two, or semibasket weave occurs.

In the portion of the breadth where warps are paired, interlacings of two-by-one, or semibasket weave, and two-by-two, or basket weave, occur.

Owing to the difference in length between the two side edges of specimen 4-4056, the weaver started making adjustments before she had woven half the length of her cloth. In order to restore a working edge at right angles to her warps, she introduced incomplete or fill-in wefts; that is, weft yarns entered on the long side and carried a distance across the web and then turned back in the next shed (pl.

8,a). The largest number of fill-in wefts occurs roughly at a point about a third up from the end. Here, seven wefts were introduced, one after the other, all entering from the same side of the web (pl.

5,e). The distance across the web that these various wefts were carried ranges from ten to twenty inches. At each turning point of the weft there is a kelim slit.

FOOTNOTES:

[5] L. M. O'Neale and A. L. Kroeber, Textile Periods in Ancient Peru: I, basic tables at end of plates.

[6] L. M. O'Neale, Textiles of the Early Nazca Period, Field Mus. Nat.

Hist., Anthrop. Mem., 2:180, 1937.

St.i.tCHERY

Three very familiar needle techniques occur on the Chincha plain-weave cloths. Breadths are seamed together with whipping st.i.tches or running st.i.tches, or are laced together with the antique seam, often called the baseball st.i.tch (pl. 8,d). The effects vary with the depth and tightness of the seaming. Some of the whipping st.i.tches are left loose so the two breadths lie flat, their selvages barely touching; other st.i.tches are drawn so tightly that the selvages form a ridge (pl.

5,b). The smallest st.i.tches are taken under two or three warps less than one-eighth inch deep and about one-eighth inch apart. Deeper st.i.tches found on the coa.r.s.e wrappings and one bag (4-3889) range from a quarter-inch to three-eighths of an inch in depth and the same distances apart.

The baseball st.i.tch, if well done, can bring the selvages of two breadths together in a flat seam (pl. 5,d). The Chincha types range from very loosely drawn to tightly drawn threads.

Running and double running st.i.tches (pl. 8,b, c), never very carefully executed on the plain-weave specimens, fasten down all the patches, hems, and occasionally the edges of lapped seams in which one breadth is extended conspicuously over another. Specimen 16-1229 has such a seam with a six-inch overlap. When running st.i.tches are small, they range from one-eighth to one-quarter inch in length with approximately the same distance between them. Many more are from one-quarter to one-half inch long, especially on the numerous patches (pl. 3,d), and the distances between the st.i.tches may be even longer. When running st.i.tches are used for the hems, the cloth edges, including selvages, are turned under twice, just as is our customary procedure. Double running st.i.tches on a bag (4-3889c) are about a quarter-inch long.

Because of the variety of uses to which running st.i.tches are put, they outnumber the other types two to one in the 16- lot, being often combined with the whipping and baseball techniques.

Needleknitting, a decorative st.i.tch which occurs frequently on Cahuachi (Early Nazca) textiles[7] is the edge finish on four of the Chincha plain-weave cloths (pl. 5,c). From the side, the st.i.tch resembles a whipping st.i.tch except for its compactness and the fact that the lines of thread are upright, not slanting; from the edge, the st.i.tch resembles a chain (pl. 8,f, g). The Chincha variety differs slightly from that on the Early Nazca textiles: st.i.tches taken straight over the edge alternate with those linked together with the chain effect.

Patching and Mending

Any form of repair technique in Peruvian textiles is rare. Many of the materials show wear and occasionally coa.r.s.e st.i.tches are put through the cloth to draw the edges of a tear together; otherwise there is little to suggest concern with prolonging the life of a garment.

In a series of Chincha domestic cloths there are eleven patched specimens but not one trace of reweaving as in darning techniques.

Apparently the unusual number of mended cloths interested the collectors in the field for, although a half dozen fragments appear to have been reduced to their present size, the patched portions have been carefully preserved. Fragmentary lengths of these textiles ranged from 14 to 28 inches; widths ranged from 13 to 30 inches.

Certain generalizations are pertinent to all the mended fragments.

There is no evidence of the use of a cutting tool; the edges of the patching pieces were torn or snagged along a thread. If the selvages were somewhat worn, the seamstress did not remove them but made a deep turn to fold the worn part to the underside. Much of the patching material was perceptibly worn to begin with; three brown specimens were badly disintegrated. On specimen 16-1259 there are four overlapping layers of patching material in one spot. Generally, the worn spot covered by a patch or several patches is an area in which the weaving was poorly done.

The patched fabrics are in the medium- and coa.r.s.e-texture groups with the exception of one fine cloth (16-1224). The repair materials fall within all three texture groups. A third of the patches (11 out of 31) were of striped materials, most of which are of better quality than the base fabric. Patches too small to cover the entire worn area are pieced out by overlapping them with a second piece of material.

More than a third of the patches were taken from the edges of the breadths, as the stripes indicate.

Techniques used to fasten down the edges are hemming and whipping. The workmanship is fairly coa.r.s.e, the lengths of the individual st.i.tches approximately a quarter-inch long. Standards were much below those held by the weaver, but this difference is not surprising.

Threads employed for the patching suggest that the seamstress used odds and ends of weaving yarns. Two or more kinds ordinarily appear on a single patch, one of them usually a coa.r.s.e white cotton thread of fairly loose twist. Some threads are used single in the needle, some double. Colors are browns, blues, orange, yellow, the last happening to be short lengths of wool. There is no evidence that the seamstress attempted to match the yarn to either the ground or the patch materials. Where we start a new length of thread with a knot made at the end, these Chincha threads begin with a half-hitch around the first st.i.tch taken through the cloth (16-1238) or with a st.i.tch through the cloth and a knot tied with the short and long sections of the thread (16-1261).

Plate 3,d shows one of the typical patched cloths. Four fabrics are represented: the base material, medium fine; and the three patching fabrics, the lower patch very fine and the upper right and left patches coa.r.s.e. The most complicated arrangement of patches is found on a specimen (16-1240) composed of two breadths seamed together. The overall measurements of the torn rectangular fragment are 17 inches warpwise by 30 inches weftwise. Within this area are nine different pieces of cloth, seven of them covering worn spots or poorly woven areas.

FOOTNOTE:

[7] Ibid., pl. 53, a-c, p. 210.

PATTERN

The only colored decoration on the Chincha domestic cloths is in the form of stripes. This section presents an a.n.a.lysis of the types found on thirty-odd specimens.

Stripes in this sample group either border the edge of the cloth or make an allover pattern. With the exception of four cloths, the stripes are warpwise of the materials; these four have stripes both warpwise and weftwise, and thus may be cla.s.sified as plaids. Edge stripes occur in combination with an allover strip pattern in specimen 16-1287 and in combination with plaid in specimen 4-3973d (pl. 6,f).

There are no cloths crossbanded only with colored wefts.

Apparently there was no preference as to the texture most appropriate for patterning by stripes; both fine and coa.r.s.e cloths are thus decorated. For example, specimen 16-1225 is very fine (thread count, 102 warps by 42 wefts per in.) and specimen 16-1234 is medium coa.r.s.e (count, 36 warps by 28 wefts per in.). Both cloths are allover striped. Edge stripes occur on a relatively fine cloth, specimen 16-1255a (count, 62 warps by 40 wefts per in.), and also on a coa.r.s.e cloth (count, 28 warps by 24 wefts per in.).

Five cloths in the Chincha lot are allover striped. One (16-1252) has solid blue and brown stripes at irregular intervals. The arrangement contrasts with the regularity of the other allover-striped materials and of the symmetrical plaids. Other allover stripes (fig. 8,a; pl.

7,c) have units a quarter-inch wide, brown on a neutral ground.

There is both color and texture interest in these specimens. The brown warp units are in pairs, the neutral-color warp units between each two brown units are alternately all single warps and all pairs of warps.

As a result, every other neutral-color stripe is appreciably thinner than its neighbor stripes (pl. 7,c). The third allover striped specimen (16-1224) is alternately blue and neutral color, each stripe unit approximately one-sixteenth inch wide (fig. 8,c). Specimen 16-1225 has striping in the same colors and to it is seamed a piece with blue on a reddish-orange ground. The blues appear to have been the same, but the cloth, otherwise in good condition, is so badly faded that the photograph does not reveal the stripes in the blue-orange section (pl. 5,d). The fourth allover-stripe pattern is common to two specimens, one of them shown in figure 8,b. The colors blue and tan stand out from a neutral ground. The sequence is blue-blue-tan, blue-blue-tan, and repeat. The stripes measure one-sixteenth inch in width and are about the same distance apart.

[Ill.u.s.tration: Fig. 7. Schematic representations of stripings with color changes indicated by symbols. Units consisting of pairs of warps represented by pairs of triangles. Chevron marks center of bilateral groupings of colors.]

The four fragments symmetrically plaided with an identical arrangement of warp and weft stripes (16-1279; 16-1303) probably came from the same cloth despite the different numbers.

Edge stripes, the most numerous group, vary in width from three-sixteenths inch to one and three-eighths inch. They are simple in construction, eight of the thirteen being symmetrical both in arrangement and count of colored warps. The semblance of balance is marked, also, in those stripes which are not symmetrical.

The edge stripes with two exceptions (16-1260, a kerchief, and full breadth 16-1287) border only one of the selvages on the complete widths a.n.a.lyzed for this section. The opposite selvages have hanging threads, remnants of the st.i.tchery which originally seamed two breadths together. The stripes decorated the outside edges of this seamed rectangle.

No specimen in the Chincha plain-weave group has stripes showing more than three colors, exclusive of the color of the ground material. The ground color is usually neutral and may originally have been white or brown cotton. The most frequently occurring color in the stripes is brown, followed by blue. Red and rose occur only twice.

In five specimens we found the warps used in pairs. In specimens 16-1224 (fig. 7,a) and 16-1280 (fig. 7,k) the colored warps are paired, the ground is set up with single warps; in 16-1240 (fig.

7,j), the stripe warps and certain sections of the ground warps are paired, the greater portion is set up with single warps. In several specimens the otherwise uniform setup of single colored warps is broken by a warp unit comprising a pair (fig. 7,f), and in two specimens (cf. fig. 7,d) the series of single warps is broken by two pairs of warps in one of the stripes. These units may have been deliberately planned by the weaver, since they are maintained for the entire length of the preserved stripe.

All of the Chincha striped cloths examined for this study were woven either in the over-one-under-one interlacing or its variation, twin warps crossed by single weft, a technique sometimes designated as the semibasket weave. What textural differences there are between the colored stripes and the ground material are the results of combining the single-warp plain weave with its twin-warp variation. The following tabulation shows the occurrences of these two techniques among the thirteen striped pieces in figure 7: