1 The Metropolitan Museum of Art's Department of the Arts of Africa, Oceania, and the Americas (AAOA) is a significant repository of approximately 100 works of Precolumbian Andean feather art (illus 1—feather panel installation in MMA Gallery). For this essay, I will use examples from AAOA’s collection to discuss both material and technical aspects of the pre-conquest feather traditions of South America as a corollary to the Mesoamerican feather art traditions that are the heart of this seminar. This collection, while not encyclopedic in temporal range, contains multiple examples from single caches as well as individual specimens, whose technical features and iconography associate them with cultural centers on the North, Central and South coasts.
2The Department also keeps a small collection of Amazonian 20th century ethnographic featherwork; which utilize similar materials and construction techniques. In addition to their appealing artistic qualities; they are useful because their generally more sturdy nature allows them to be closely examined. The ethnographic featherwork will not be discussed here, but these 20th century artifacts demonstrate that Precolumbian feather traditions have, in part, a trans-Andes origin, and suggest that Amazonian featherwork has likely remained a living tradition from prehistoric times up to the present.
3As in Mesoamerica, brightly colored and intricately patterned featherworks were highly regarded by the ancient peoples of the central Andes, which includes the distinct cultural traditions of the highlands and the coast. Many examples of feathered ornaments and ritual objects, as well as garments and hangings patterned by feather mosaic, have been recovered from the arid coastal region, where they were preserved for centuries and, often, for millennia (illus 2—photos of coast/highlands).
4Although some conclusions may be drawn with regard to the great antiquity and wide cultural distribution of Precolumbian feathered artifacts, there are several interrelated issues that hinder a greater understanding of these objects. Chief among these is the fact that, in the past, many featherworks, if not the majority, have entered into art museums and private collections through dealers, usually without provenience or with only anecdotal information concerning their excavation. More often than not, the featherworks have been separated from related burial materials. It has been reported that the early generation of archaeologists sometimes supplemented excavated materials with purchased additions, which were not always clearly labeled. Unlike the situation in Mesoamerica, there are no Andean Precolumbian manuscripts and the ethno-historical materials from the post-conquest era, which can be very useful, generally only describe the rich feathered textiles and accouterments of the Inka, and cannot be used to interpret the functions and meaning of earlier feathered works. As a result, we have an incomplete picture of methods used to hunt birds and/or harvest feathers; of the trade routes and trading patterns; and the craft habits and geographic locations of the feather workers. Finally, complicating all of the above is the fact that feathered surfaces lend themselves more easily than woven textiles to reconstructions and restorations that are not readily detected.
5The Department of the Arts of Africa, Oceania, and the Americas is fortunate to have examples of feather mosaics that have not been altered or over-restored. With, however, the notable exception of the group of blue and yellow panels, some of which were radiocarbon dated, and the artifacts classified as Chimú (Rowe 1984:155ff), the remaining featherworks are either unattributed or tentatively attributed to either Wari or Inka culture. In light of this uncertainty, the physical characteristics of the objects themselves are the best evidence we have for an understanding of Andean featherwork traditions. Additional insights can be gained from ornithological, archaeological, ethno-historical and art historical works regarding Andean feather creations. The information presented here is based upon my examination of particular pieces and the work of many scholars in these fields, most notably those listed in the bibliography.
6Before discussing the AAOA artifacts, a few general comments about feathers are in order. Feathers, like scales, fur, hair, as well as horn, hooves and nails, are formed from the structural protein, keratin. The different physical characteristics found in each of these manifestations are due primarily to the thickness of the protein layer. Sulfur cross-linking between the amino acid chains makes keratin a hard and quite insoluble protein. Feathers are adapted to be light, yet flexible and tough. The shapes of feathers are related to their individual functions on the bird’s body, but it is beyond the scope of this paper to discuss feather anatomy. Published examples of recovered artifacts, however, suggest that Precolumbian Peruvian feather artists employed contour feathers exclusively in their creations. A contour feather is a vaned feather. Contour feathers include both the smaller feathers that create the outer covering of the body of the bird and the larger flight feather of the wing and tail. The majority of artifacts in collections are embellished with the typically small body feathers, which have symmetrical vanes to create feather mosaics. (illus 3—feathers with parts labeled)
7The larger, generally asymmetrically vaned tail and wing flight feathers are infrequently preserved in archaeological contexts. However, Guaman Poma, the 17th century chronicler who prepared a history of the Andean region for King Phillip III, is, among others, an important source for illustrating feathers, as he believed the Inka used them. These drawings can now be viewed in facsimile at the Guaman Poma website. http://www.kb.dk/elib/mss/poma/ For example, Guaman Poma shows Mayta Capac (mss page 98), the fourth Inka king, wearing a headdress crowned with what could be interpreted as flight feathers. Guaman Poma also provides evidence of colorful forest birds being kept as pets by Inca royalty. An illustration of the twelfth Inka queen, Chuqui Llanto, (mss page 142) mourning her murdered husband shows a woman attendant holding a large and possibly feathered, umbrella. At her feet are a parrot and macaw. These “pets” would provide a ready source of colorful feathers, but additional sources would have also been necessary, given the large numbers of small feathers needed to create feather mosaics.
8To date, no comprehensive investigative study of Andean Precolumbian feather creations has yet been published. However, two important scholarly publications have begun to explore this topic—Ann P. Rowe’s chapter on featherwork in Costumes and Featherwork of the Lords of Chimor (1984) and Kay Candler’s article “Ancient Plumage: Featherworking in Precolumbian Peru” in The Gift of Birds (1991). Rowe, curator of Western Hemisphere collections at the Textile Museum in Washington DC, applies an analysis of the structural details embedded in woven fabrics of secure provenience in order to classify comparable feathered textiles as originating from the Chimú culture. She suggests that while local iconography may be appropriated by a dominant culture, the intrinsic qualities of craft-habits are seldom transferred from one culture to another. Candler, an anthropologist working as the assistant curator for the eponymous exhibition at the University Museum of the University of Pennsylvania, is interested in broad cultural questions such as the relationship between birds and featherwork and the cultural significance of feathered dress in Precolumbian societies, but she also understands the need to use structural details as an analytical tool.
9Ornithologists and historians have observed that out of hundreds of available bird species, the peoples of Precolumbian Peru generally seemed to prefer feathers from a select few species. For the Gift of Birds, the 1991 exhibition and catalog of featherworks of native South Americans peoples mentioned above, ornithologist Mark Robbins, a specialist in South American birds, surveyed both archaeological and ethnographic featherworks and determined that the Precolumbian feathers from the collection featured in the exhibition seemed to be from six species: macaw, parrot, toucan, cotinga, tanager, trogon (Reina &Pressman 1991:114). Several years earlier, ornithologist John P. O’Neill, also an ornithologist of South American birds, identified feathers for textile curator Ann Rowe’s 1984 exhibition and catalog Costumes and Featherwork of the Lords of Chimor at the Textile Museum in Washington, DC. He also identified feathers for independent scholar Ann Paul’s 1990 publication Paracas Ritual Attire. For Rowe’s publication, O’Neill often suggested more than one species when it was difficult to identify the bird. However, he concluded that most of the artifacts he examined used feathers from very few species. He cites Amazona parrots, macaws, tanagers and Muscovy ducks as the main species. (Rowe 1984:147)
10Two further points concerning feather identification should be made. O’Neill observes that small white body feathers are generally undistinguished and could come from many different species. Related to this fact is his further finding that some artifacts contain similarly colored feathers from different species (assuming they are not later restorations) (Rowe 1984:146). These two observations lead to his conclusion that for the creators of some featherworks, the overall color effect seems to have been more important than the particular bird species. This raises interesting questions concerning the source of power inherent in this type of prestige ornamentation because his conclusion suggests that for some Precolumbian peoples, power and prestige may derive from a general embodiment of “birdness” inherent in a feathered garment. However, Garcilaso’s well-known description of the use of two corequenque feathers by the Inka king (Gheerbrant 1961:190, Candler 1991:9) suggests that other Precolumbian peoples were more selective, finding power and prestige inherent in particular types or colors of feathers.
11However, it is wise to remain cautious, yet open-minded, when it comes to feather identification. Even when a study collection of positively identified contour feathers of tropical birds can be consulted, feather identification remains difficult because, as in the case with white feathers, similarly colored feathers from different species can be difficult to distinguish, especially to non-specialists; and it is difficult, if not impossible, to identify species for partial and/or degraded feathers or feathers that cannot be examined on both sides. However, it is clear that the majority of brightly colored feathers used in the Precolumbian featherworks found in many museum collections, including the MMA, are primarily from the bird species identified above; and these species are, or were once, native to the regions east of the Andes. But the question of why the feathers of so few bird species were selected does remain another unanswered line of inquiry concerning artifacts.
12 The majority of the MMA feathered textiles consists of feather mosaics, which have long been considered one of the most spectacular manifestations of Precolumbian textile arts. In general, the MMA feather mosaics can be divided into three groups: full-size garments, miniature garments, and large panels of uncertain use. In addition, there are two small rectangular bands and a small square cloth, all of uncertain function. Finally, there are also two feather mosaic headdresses.
13Loose feathers as well as feathered artifacts have been found in archaeological sites from the Preceramic period onwards. There is evidence that the Chavin culture held feathers in high regard. Caches of loose feathers as well as those turned into “featherwork” have been recovered at El Paraiso (Burger 1992:41) and at Aspero on the Peruvian coast (Feldman 1986:35). A small feathered object composed “from about a dozen parallel rows of red and yellow feathers” held together by strings and without a textile backing has been described among the buried caches of offerings found in the summits of the platform mounds at the Huaca de los Idolos at the Aspero site. This late preceramic site dates to approximately 2750 BC. Some of the colorful loose feathers recovered from La Galgada (occupation dates approximately 3000BC to 1500BC, in the highlands have been tentatively identified as coming from species living in the rainforests of the eastern side of the Andes (Grieder 1988:73-75; Burger 1992:54).
14In the south, at the well-known Paracas cemeteries, archaeologists have uncovered a substantial body of small featherworks enclosed within the mummy bundles. Eugenio Yacovleff, an archaeologist working at Paracas presented a compilation of ethno-historical references for feathers. He identifies the birds and describes feathered fans, tassels, batons, slings and headdresses recovered from a group of opened bundles (Yacovleff 1933:145ff). Ann Paul also describes some of the objects contained within the bundles, adding turbans and small cloths to the list. She notes that featherworks were placed either close to the body or in the penultimate layer and that those items placed in the layer closest to the body are items usually associated with the head (Paul 1990:42-43). Both Yacovleff and Paul report that the feathers in question are from local birds and from the tropical forest, which points to a relationship between the peoples of the coast and the highlands to the peoples of the tropical forest. Neither Yacovleff nor Paul reports finding an example of polychrome-feathered mosaic in the Paracas bundles.
15The creation of multicolored patterns and images by the technique of feather mosaic seems to be a later development, although it has not been possible to establish a locus (or loci) for this achievement. As with much of Precolumbian textile arts, most feather mosaics have been found at central and south coast sites, although some of these are thought to be imports from other regions (Candler 199l:2). Interestingly, published examples of feather mosaic garments, headdresses and accessories, whether they have been recovered from properly excavated archaeological sites or are of unknown provenience, are always very sophisticated creations. They appear to come from traditions where the formats for their design and the manufacturing methods have been established for a long time. To date, it has not been possible to find examples of scientifically excavated feathered artifacts that suggest a cultural link between these earlier finds and the feather mosaics found in many collections. All the feathered mosaics held by the MMA, are thought to be products of Wari or later cultures, and they are thought to date from the 7th century AD onwards.
16The feathering techniques used for the MMA feather mosaics can be grouped into two broad categories according to how the feathers have been attached to the support. Most commonly the feathers are attached mechanically, by stitching, and less commonly, by adhesive. Stitched feather mosaics rely on differently colored feather strings, which are fastened to a cloth by stitching the strings into place according to a predetermined design. Feathers attached to the substrate with adhesive were placed individually on the foundation material, also according to a predetermined design.
17There are a number of differences in the precise way the feather strings are created and attached to the foundation fabric; some of these methods have been diagramed and described (Mead 1908, Yacovleff 1933, Green 1991). Invariably, the feathers are first attached to a string. Usually, the shaft (rachis) is bent over and a string is knotted around the fold. When the feathers are held with a single knotted string they tend to turn in all directions, so that, in most cases, a second string is added. This second string can be tied around the feather in several different ways. The feathered strings will then lie flat with all the feathers facing in the same direction. Two examples of feather strings are shown here (illus 4—Precolumbian feather strings). The feathers secured with two strings were, no doubt, easier and quicker to position onto the substrate. Once the differently colored strings were prepared, rows of similarly sized but differently colored feathered strings were shingled over the surface of a woven textile according to the requirements of the design. When the feather artist needed to change the feather color according to the design plan, the feather string would be cut and a new, differently colored string begun. The feathered mosaics I have examined in the MMA collection have also used this same method to change color, although it has not generally been possible to move all the fragile feathers at color change interfaces in order to observe if there is ever a color change without a cut string. The feather strings are held in place by stitching them to the foundation fabric with another yarn. Again the stitching techniques are varied, some of these have been described, and they will not be further discussed here.
18In the MMA collections, the feather strings documented to date, are always attached to woven cotton fabrics. The strings are most commonly cotton, but occasionally an unidentified hard plant fiber is used for the feather strings in these works. Based upon the examination of the MMA collection, the use of cotton or plant fiber strings in a feathered object appears to have been a random association, perhaps based upon what was available to the feather artist. However, feather mosaics using strings of hard plant fibers have also been reported elsewhere and it has been suggested that in some cases there may be a relationship between a particular type of feather and the fiber used for the string (Greene 1991:25). When feather mosaics are created by these techniques, the tips of the feathers always point down and the feathers are viewed in this orientation when they are worn or displayed. The cotton foundation fabrics are quite varied, both in yarn make-up and structural details; some of these characteristics will be highlighted as necessary. Collectively, these details—the types of fibers used to make the strings; variations in the method of attaching the feathers to the strings and of stitching the feathered strings to the woven textiles; differences in yarn make-up and in the woven structure of the base fabric, as well as the orientation and seaming of the cloths suggests both centuries long traditions and multiple centers of feather working. So the apparent gap of approximately seven centuries in the archaeological record between the small Paracas featherworks and the feather mosaics thought to belong to Wari material culture remains unexplained.
19Less typically, feathers are attached to a substrate with an adhesive. The MMA possesses one such object, a feathered crown. Ann Rowe (1984) attributes this crown to the Chimú culture although it was said to have been part of a cache found in the Chancay Valley on the Central Coast. This crown is a complex object created from metal, cotton, skin, reeds and feathers. The methods used to fabricate this construction are too detailed to be discussed here, but it is interesting to note that in order to create the multi-colored checkerboard pattern in feather mosaic (illus 5—detail of MMA1987.394.655 crown) tiny feathers were adhered with an adhesive to a thin light brown cotton fiber “felted” support. The adhesive, which is now dark and hard, gives body to these two materials and together they form a stiff construction. Where the feathers are not too degraded or embedded in the adhesive, it is possible to detect that they do not seem to be chopped and they are applied in overlapping rows from the top down, so that all the feathers are pointing up. These bright feathers have been tentatively identified as being from the Paradise Tanager (Rowe 1984:156) and they seem to have been particularly chosen for their tiny size; the sharply defined colored blocks are created with only minimal trimming of some of the larger feather tips.
20Aside from this example, the remaining feather mosaics in the collection are all of the stitched variety. The earliest example is a tabard (29.146.25), which is defined as an over-the-head garment, left unseamed at the sides (Rowe 1984:176) (illus 6—detail of 29.146.25 yoked tabard). Both the feathers and the foundation fabric are generally intact. There are three darned “native” repairs along the narrow (bottom) edges, whether these are pre- or post excavation repairs is unknown. This tabard has generic features found in several of the other feathered garments and panels, thus is discussed here to illustrate spinning and weaving methods and materials used in a particular type of ground cloth that is commonly associated with cultural groups centered in the southern highlands. It has been cataloged as Wari and is thought to date from the 7th to 10th centuries.
21The foundation fabric is composed of three cotton panels, both woven in a plain weave structure, where the per centimeter warp count is slightly higher than that of the weft. Three colors of cotton fibers—ecru, tan and brown have been used. It was not possible to determine if the mid-range color is a unique shade or a mixture of the darker and lighter colors. The panels have a random grouping of differently colored bands running in the warp direction. The make-up of the warp and weft yarns are the same; all are Z-spun yarns, two of which are plied S. At the headings, multi-plied cords are used. The thickness of the yarns and the degree of spin varies. Some of the warps are bi-colored—that is, composed of two differently colored yarns. The weft yarn is doubled and in some areas one yarn is ecru, the other brown. Both the bi-colored warps and differently colored weft yarns are inconsistently used throughout the web.
22The two larger panels are seamed together along side selvages, leaving an opening in the center for the head to pass through. Along one narrow end, the side selvage of a smaller complete panel has been stitched to the heading cords of the larger panels. When the tabard is worn, the warps of the larger panels run vertically, while the warps of the smaller panel run horizontally. The proportion of the smaller panel to the larger seamed panels is approximately 1:7. Collectively these technical details of yarn-make up, woven structure and construction details give the overall impression that the quality and type of woven cloth used for this garment was not important; it would not, after all, be seen. Low quality, slightly warp face cloth is a feature that appears in other examples from the collection, although each textile varies somewhat from what has been described here.
23The simple, yet bold, decorative pattern of this feather tabard, with a distinct front and back, is very different from the intricately patterned and technically sophisticated Wari camelid tunics where both front and back sides are the same. One side of the tabard is patterned with a yellow U-shaped yoke centered on a ground of turquoise feathers and this side has generally been considered the front of the tunic. At the top center of the yellow yoke there is a small turquoise square of feathers. The other side is composed of four horizontal bands of turquoise feathers each patterned with four connected stepped triangles worked in red-orange feathers.
24This tabard employs the general techniques of feathering described above. Each side is worked separately. Straight rows of colored feathers are stitched into place starting at the bottom of each side of the tunic and worked toward the shoulder. Each subsequent row overlaps the previous row by about 1/3 to ½ the length of the feather. When a new color feather is introduced for reasons of the design, the feather string is snipped and secured. A new string is set-in near to the cut, often slightly above it. Where the first row of yellow feathers begins the curved part of the yoke, it breaks the process of using feather strings in straight rows. This first row is slightly curved just along the bottom. The remaining feathers continue in straight rows. This process is continued up to approximately where the tabard would begin to turn around the shoulder. This area is typically left blank, because the feathers would stick out as they round the curve, rather than lie flat. A plush velvety surface seems to have been the desired effect.
25The feathers used in this tunic are all small body feathers with symmetrical vanes. It is likely they come from macaws. The flat red-orange and yellow feathers have a spiked hair-like appearance that is found on the breast feathers of both the blue and yellow macaw and the scarlet macaw, whereas the blue feathers tinged with green iridescence seem to be the back feathers of the blue and yellow macaw. It is interesting that the blue feathers, which have a sturdier shaft and closed vanes with rounded tips, were chosen for the background, while the chest feathers, which have thinner shafts, flatter tips and a more open structure were chosen for the designs. The feathers are generally used without modification, although some of the red and yellow feathers may have been clipped across the tops to sharpen the edge of the design.
26Perhaps the most well-known feather mosaics are the extraordinary group of large rectangular feather panels said to have been found on the far south cost of Peru near the Ocoña River valley in 1942-43 (illus 7—MMA1979.206.467 blue/yellow hanging). They are believed to have been from a single cache, estimated to have totaled 96 works. Most of the panels were said to be yellow and blue but some panels of single colors, red or yellow, were also reported. Several small groups of these panels came into the collection of the Museum of Primitive Art during the late 1950’s and eventually became part of the MMA collection. Details of this find and some details of the subsequent disbursement of the pieces were discussed in a 1958 exhibition checklist written by Junius Bird, the archaeologist and curator emeritus of the Department of Anthropology, American Museum of Natural History in New York City (Bird 1958:unpaged).
27The circumstances of the find led to questions about the date and cultural affiliation of these showy artifacts. Originally cataloged as Inka by Bird, samples from three panels were radiocarbon dated in the mid-1960’s. The results of the C-14 tests were published by the long-time professor of anthropology at the University of California, Berkeley, John Rowe (Rowe and Menzel 1967:27-28). The radiocarbon dating placed the samples within the early Middle Horizon, approximately the 8th century AD, and their provenience was changed to Wari. Although the panels look similar, this similarity is somewhat misleading. An examination of the back reveals differences in the foundation fabrics. Only a few of the panels have been thoroughly examined and only a few general remarks will be made here. The quality of both the yarn make-up and the weaving is very skilled in all panels. When compared to the feather mosaic garment just described, the cotton yarns of the blue and yellow panels are more finely spun and it seems the cotton fibers were once quite white, although darker colored yarns appear randomly in some panels. Only a portion of the differing structural details will be mentioned here. The foundation fabrics have different warp orientations and structural details. Warps can run vertically or horizontally. A foundation fabric can be a single piece of fabric or it can be constructed of two pieces. On one panel, one of the short sides has been folded toward the front of the fabric. Whether this was done to achieve a specific final dimension or for some other purpose, is not known. Folding a section of the ground fabric outward has also been described for some feathered Chimú tabards and is used systematically in the miniatures described below. The feather strings are mostly cotton, but there is one yellow and blue panel with all strings composed of vegetable fibers. The blue and yellow feathers are similar, if not identical, to the feathers described for the tabard and have been identified as blue and yellow macaw. Perhaps due to the simplicity of the design, the feathers are not altered in any way. The panels have a woven heading tape and braided ties (or the remains of ties) of camelid fiber, at either side. These ties suggest the panels were secured to some type of structure; however, neither the purpose of the panels, nor the intended arrangement has been determined, and they remain frustratingly illusive.
28Caches of objects can add a further dimension to our knowledge of Precolumbian cultures. The Museum of Primitive Art had, in the late 1950’s, the good fortune to acquire a group of miniature feathered textiles, all of which came to the MMA. Like the tabard and the blue and yellow feather panels, they are believed to date from the 7th—10th centuries. These miniatures are thought to have been part of a large cache of feathered textiles, which were discovered in the lower Ica valley. A memo from Junius Bird in the AAOA department files indicated that he thought they dated to the Inka period and he reconstructed the original cache as numbering approximately 117. He noted that the cotton base fabrics were not standardized, that the cloths had been folded to size and that the workmanship was sloppy. From these features, Bird concluded that they were unlikely to have served any “practical function” and they “may have been made hurriedly as grave offerings.” He further noted these may have been meant to represent garments, but that further substantiation would be necessary before coming to a final conclusion.
29These small objects languished in this limbo for at least three decades. In 1996, Mary Frame reexamined these textiles and concluded that they were created in order to conceptualize gender differences (MMA catalog cards and Boone, ed. 1996:421). The miniatures with the yoke appear to all be representative of men’s garments. The determining features indicating maleness include a vertically oriented warp with excess fabric folded toward the outside of the garment. The garments are seamed along both sides and the neck opening is either snipped-in or is a dark feather placed where the opening should be (illus 8—MMA1979.206.683 male miniature).
30The second and larger group with checkerboard patterns is differently fashioned in order to represent female dress. In particular it is a garment type that was once little understood but is now generally accepted (Kajitani 1982:72-73; Desrosiers 1988:44). Here the warps are horizontal; excess fabric is folded toward the inside of the finished work; the fold runs along one side and the other side is seamed to make a tubular shape. The upper edge is seamed, but in such a way so as to suggest openings for the arms and head. (illus 9—MMA1979.206.658 female miniature)
31For both groups, the foundation fabrics are generally single pieces of cloth woven from cotton yarns. The structures are all plain-weave, but warp and weft counts vary as do the make-up and quality of the yarns and the weaving, as Bird noted. Some cloths have paired warps, suggesting the influence of Chimú weaving techniques. Other fabrics have damages that suggest they could have been recycled for this use. But no matter what the shape of the original fabric, they all have a final shape that, when folded, is square or rectangular. Interestingly, the overall dimensions vary from less than 4” square to more than 12”. Several features of the ground cloth, such as the uneven quality of the yarns and woven structures, make a persuasive case for associating some of them, culturally, with the full-sized garments already described. Like the full size garments, the designs, with a few exceptions, are different from front to back on both the male and female garments. The feather strings are constructed and attached in the manner described above. The strings of feathers are cut when the color changes and although the feathers are generally not altered dramatically, occasionally the tip of a feather is trimmed, most noticeably in the checkerboard design dresses, to clarify the pattern. A variety of feathers have been employed in the creation of these miniatures and not all the feathers have been identified.
32The final group of feathered garments in the collection consists of works thought to have been produced by the Chimú or somewhere within their circle of influence. In addition to establishing the date and provenance of these feathered creations, Ann Rowe has analyzed the yarn make-up and some details of the woven structures (Rowe 1984:176-178). All use S-spun single yarns; both warps and wefts are paired for the tabards. All use a version of the knotted and stitched feather mosaic technique for patterning. Two of the most distinctive are worthy of further discussion.
33These two garments have a quadrant-based design, which is a typical format of Chimú feathered tabards. One tabard had been cut into two halves and each came into the Metropolitan Museum as separate accessions: 63.163 and 69.62.2. The sides are identical; both feature four right-facing crested birds. The remaining tabard, 59.135.8, features four multicolored winged creatures on one side; unfortunately the feathers have been removed from the other side, so the design is forever lost. While acknowledging the Chimú features of the ground fabric, neither piece easily fits into the corpus of Chimú feathered textiles.
34The foundation fabric of the crested bird tabard is very degraded (illus 10—MMA 63.163 half tabard), but seems to have been woven from beige cotton yarns. It is composed of two long, narrow panels seamed together along a side selvage leaving a center opening for the head. Excess fabric along the heading and finish selvages of both halves of the crested bird panels have been folded toward the outside of the garment and the seam covered with the first row of feathers. The outward folding suggests both a “male” garment and a particular (perhaps standard) finished dimension. The warp orientation, as worn, was vertical. The shoulder area was left unfeathered.
35For some design elements, such as the feet and beak details, the feathers have been dramatically shaped; this is the first time that severe feather trimming appears in the MMA feather collection. In addition, feathers are sometimes attached to the strings in groups of two or more and in the details of the eyes and beaks a brightly colored feather has been glued to the tied feathers with a now dark and hard adhesive. Both these features seem to be time-saving techniques. The profile-crested bird is part of the Chimú iconographic corpus; but the design makes sparing use of brightly colored tropical feathers. Rather, a somber brown and white palette of Muscovy duck feathers predominates, which suggested to Rowe a post-conquest date, after the trans-Andes feather trade had been disrupted (Rowe 1984:176-178).
36The foundation fabric of the second tabard, 59.135.8, exhibits Chimú structural details in yarn make-up and in the paired warps and wefts (illus 11—MMA59.135.8 tabard). Other features seem anomalous to Chimú-type feathered garments. The hovering, armless and legless, winged front-facing figures do not seem to reflect Chimú iconographic preferences; although the staring eyes and sharp teeth do appear on figures in the “plain crescent headdress” style. The feathering of this garment also features several interesting anomalies. The first is the diversity of feathers found on one single garment. There are at least eight different types of bird feathers appearing on this one object and with, perhaps, the exception of the black feathers, all are brightly colored, thus likely from the tropical zone. No other feathered garments in the MMA collection feature such a diverse collection on a single item. The iridescent green feathers that have been tentatively identified as trogon (Rowe 1984:176 photo caption) are rarely found on Andean Precolumbian featherworks. Furthermore, the size of the feathers found on this tabard has a range also not seen in the other MMA feathered artifacts. The smallest feathers average about 1.5 cm in length (after folding), while the largest originally averaged about 12-13 cm in length.
37The high relief at which the smallest of these feathers stand out from the background is remarkable and upon close examination these feathers were discovered to have been grouped into small feather tufts of four or more, held together with an adhesive, which is now dark and hard. It is difficult to examine many of these tufts closely, due to their fragility and the thick cotton strings used to fasten them together. It is not been possible to see precisely how the shafts are folded over and tied to the heading string. This feature, and the variation where two or three feathers are grouped and tied without the use of adhesive, have subsequently been observed on other Precolumbian feathered mosaics in the collection, but these observations are preliminary and further examination is necessary before any conclusions can be put forth.
38In conclusion, the collection of the Department of the Arts of Africa, Oceania and the Americas quite effectively captures the diverse picture of Precolumbian featherworks from the 7th century to the Colonial period. It is my hope that these feathered mosaics, largely unrestored, will prove to be an invaluable primary source for comparative purposes. A great deal of information can be recovered from a close examination of the woven structures and construction details of these works. The next step will be to begin detailed comparative studies with featherworks in other collections; one goal of which would be to sort out repairs and reconstructions on featherworks from original materials. This information alone would provide a valuable resource to those attempting to create a history of Precolumbian featherworks.