Virginia S. Wimberley, Asst. Professor
Department of Clothing, Textiles, and Interior
Design University of Alabama, Tuscaloosa, AL
with Commentary by
Christopher Carr, Professor
Department of Anthropology
Arizona State University
Tempe, AZ 85287-2402
The following questions were asked in order to address these three topics.
(1) How commonly do textiles occur on copper items of all types, such as breastplates, headplates and celts?
(2) Do sites vary in the commonality with which textiles occur on copper items of all kinds?
(3) Do basic artifact types differ in the commonality with which textiles occur on them?
(4) Do textiles tend to cover 80% or more of an item-side or less than 80% of an item-side, considering all kinds of copper artifacts for all sites studied?
(5) Does the areal coverage by textiles on items differ by site?
(6) Does the areal coverage by textiles on items differ by type of artifact?
(7) When textiles occur on a copper artifact, do they tend to occur on one side or
(8) Do copper artifacts of different types differ in whether textiles occur on one
side or both sides?
(9) How many cloths of identifiably different technical construction tend to occur
on copper items, regardless of their specific type?
(10) Does the numbers of different cloth constructions per artifact type differ
among the sites studied?
(11) Does the number of different cloth constructions occurring per item differ
among the artifact types studied?
(12) Do any items having multiple textile fragments from the same or different
textile construction have their textile fragments oriented differently?
(13) How do fabric structures, fiber diameters, yarn diameters, and yarn spacing
per centimeter (thread count) vary by site?
(14) How do fabric structures, fiber diameters, yarn diameters, and yarn spacing per centimeter (thread count) vary by the kind of artifact?
Questions 4, 7, 9, and 12 were asked specifically to determine whether textiles on headplates, breastplates, and/or celts derived from these objects having been placed in textile bags or wrapped in textile cloth just prior to burial, or whether the textiles had different origins. One of these alternatives is that textiles were cut out into culturally- meaningful shapes and applied to the artifacts, perhaps in the process of developing copper corrosion patinas (see Parts I and II).
In this report, the primary observations were made by V. Wimberley and D.A. Wymer. The technical, textile analysis and written report were the work of Wimberley. C. Carr was responsible for making the Canon color prints on which some observations were made, sample selection, calculating some summary statistics reported here, and the commentary at the end of the report.
Items were prioritized for examination according to the following criteria: (1) the presence of substantial pieces of textiles or their pseudomorphs; and (2) the potential of the item bearing artistic imagery, perhaps produced in part with textiles, as assessed by Carr.
For fiber and yarn characteristics, five measurements were made for diameters and averaged. A hand held protractor was used to determine the angle of twist for single and ply yarns. A protractor disc in an eyepiece would have been preferable but was not available. For thread counts of vertical and horizontal elements, five measurements taken by sampling over the entire sample were made, using an ocular micrometer disc and then averaged. In the case of some fragments, five measurements were not possible due to the small size of the textile; therefore, as many measurements as possible over the entire area of the sample were made and averaged.
Table 5.2 reports the assessment by Wimberley and Wymer of a somewhat different sample of 66 items (132 item-sides) by visual inspection of the actual items, rather than Canon color prints of them. In a few instances, the assessment of textile presence was made by Wymer, alone. Due to time limitations, Wimberley surveyed in detail only those artifacts judged to possibly have textile remains, as evaluated by Wymer’s inspection, and those assessed as important for imagery by Carr, rather than every artifact. Of the 132 item-sides surveyed, 68 (51.5%) had textile remains. Of the 66 items, 23 (34.8%) had fabric on two sides, 22 (33.3%) had fabric on one side, and 21 (31.8%) had fabric on neither side. Only 3 celts of 14 (21.4%) had textile on both sides; 3 (21.4%) had textiles on one side and 8 (57.1%) had no fabrics.
Table 5.3 provides a breakdown of these results by the kind of item (breastplates, celts, headplates) and by site. The sample is definitely skewed toward Seip and Hopewell. The Hopewell Mound Group had 5 out of 10 breastplate sides exhibiting fabric, but all with areal coverage of less than 80. Hopewell also had 3 of 5 celt sides with fabric present at 5 percent areal coverage. This would appear to indicate that the textile was not necessarily bagging or wrapping material, but without provenience information it is difficult to say more. These artifacts have been handled extensively and textiles may have broken off or been reduced to power. The preservation of the appearance of these items in there present state by Carr’s Ektachrome and digital photography will help track what happens to the textiles over time as they remain in storage and are used for research. The other sites of Rockhold, Liberty, Ater, and Ft. Ancient have few artifacts with textiles adhering to copper. These sites were represented by one to two items possessing any areal coverage of fabric, all considerably less than 80 percent, which was chosen as likely indication of wrapping or bagging. There are textiles from these sites analyzed by the author in 1997 that are not attached to copper. Breastplates from these sites have the most textile evidence, with some celts possessing textiles or pseudomorphs. Headplates, originally thought to have remains reminiscent of a textile, universally proved to have no attributes of textiles in this Oho Historical Center sample. However, one cannot assume that headplates do not generally retain textile evidence; there is one headplate specimen at the Ross County Historical Society, Chillicothe, OH, which had much cordage on both sides.
To consider whether textiles covered one or both sides of an artifact, and the amount of areal coverage by textiles, please refer to Tables 5.2 and 5.4. Eighty percent or greater coverage of an artifact side by textile is a rare occurrence. For the 66 items examined, of the 68 item-sides having textiles, 48 (70.6%) have textiles on less than 50% of their area, 10 (14.7%) have textiles on about 50% of their area (21.2%), and only 10 (14.7%) have textiles covering more than 50% of their area.
Few artifacts had more than one cloth construction on one side (Table 5.5). Of the 84 copper items with textiles or their pseudomorphs, only two items (2.4%) had two different cloths, and only one item (1.2%) had three. There was some mixture of one textile construction with feather and/or hair or fur fiber.
Table 5.6 lists the different textile constructions found on the surveyed artifacts. The number of different textile constructions was limited to four types: oblique interlacing, spaced 2-strand twining, alternate pair twining and spaced alternate pair twining. Only two items had two identifiably different textile constructions on the same side of the artifact; these items were both breastplates, from the Hopewell site (See Table 5.5). There is a possible example of featherwork, where feathers were attached to a textile substrate. Only a few places have the feather shafts bent in the traditional form for attachment; yet there is a definite textile grid over the entire plate side. This artifact requires further analysis.
Table 5.6 shows that oblique interlacing occurred only on breastplates for the sites of Seip, Hopewell and Ater. Alternate pair twining was found on one breastplate from Seip. Spaced alternate pair occurred on ten breastplates from Seip, one breastplate and one celt of Hopewell site origin, and one Rockhold breastplate. Spaced-2-strand twining was characteristic of only one breastplate from the Hopewell site. Five breastplates from the sites of Ater, Hopewell, Liberty and Seip had unidentifiable textile constructions. Because the sample is so biased in numbers toward Seip and Hopewell, it is not possible to generate any conclusions for between site and artifact type occurrence.
Information of more detailed technological attributes, including thread and ply diameters, and yarn spacings, are shown in Tables 5.7. The yarns were universally Z- spun singles and then plied S for all sites.. It would further the analysis of textile structure occurrence if the artifacts were put in archaeological context. There may be differences in the attributes of textiles associated with bone. There is one incidence of thicker elements or yarn diameters having been used for a textile adhering to a plate with bone, compared to the same construction adhering to plates without human remains.
Wimberley would like to compare the fiber and textile constructions that occur on copper to those textiles which were not attached to copper [studied during the summer of 1997]. This may answer the question of whether there are differences between textile structures appearing on copper and those not associated directly with copper from the same Hopewellian sites.
Finally, Wimberley noted higher or lesser twisted yarns within the textile structures on several breastplates. Table 5.7 reports yarn diameters and thread counts. This anomaly of twist per centimeter is being investigated for several possible reasons. Once a spinner is experienced, the twist per centimeter remains fairly consistent for the yarn type being produced for a particular use. In the case of the breastplates in Table 5.7, the bast-like yarns all visually appear to be of the same fiber type and were used for the same twined construction, which would normally mean that all yarns used would have the same twist per centimeter for the entire cloth when made by one spinner. The presence of these varying yarns may indicate that more than one person provided yarn for the textile or that yarns, which had been prepared for different purposes, were incorporated into the structure, perhaps due to time constraints for quick preparation.
It is possible, and actually thought probable, that each of these formation processes had roles in the occurrence of textiles on the copper breastplates and celts. Occasional items (e.g., the very large celt from Seip) have fabric remnants and/or pseudomorphs in a regular, grid-like pattern over most of their area and both sides. This pattern suggests the wrapping of the item in cloth or its placement in a cloth bag. At the same time, of the 66 items studied, only 23 (34.8%) have textiles on both sides. Thus, for the majority of items, the explanation of a cloth wrapping or placement in a bag is not likely.
The idea that the textiles were the fortuitous result of having lain against clothing, with subsequent corrosion and incorporation of the textiles in the artifact-corrosion mass, finds some support in that 22 of the 66 items studied (33.3%) have textiles on one side, alone. However, few of these items have textiles covering any large area of the side on which they occur. Of the 22 items with textiles on one side, 17 (77.3%) have textiles covering less than 50 of their area, while only 2 (9.1%) have textiles on about 50% of their area and only 3 (13.6%) have textiles on more than 50% of their area. More broadly, of 68 item-sides with fabrics present on them, 48 (70.6%) have fabrics covering less than 50% of a side, while only 10 (14.7%) have fabric covering about 50% of a side and only 10 (14.7%) have fabric covering greater than 50% of a side. These situations tend to support the alternative explanations, that textile forms that were cut out and applied to the artifacts as decorations or to create images through differential patination. The last two options cannot be separated by the kinds of data tabulated here.