Chapter 4

The mat, a very excellent piece of work, is 6 feet 6 inches by 3 feet 4 inches. By reference to plate II it will be seen that it is neatly and artistically made and quite well preserved. The strands are from one-third to three-sixteenths of an inch in width and are even on the edges and smoothly dressed on the back. The hard, glistening outer surface of the cane is light in color and the dressed surface is dark naturally or artificially, and the weaving is so managed that a tasteful border and a checkered effect are produced by alternately exposing the light and dark sides. This piece probably very fairly represents the split-cane work of the whole cane-producing region. A similar piece of work from the gulf coast is illustrated in figure 12.

Inclosed with the mat were three pieces of fabric of especial interest, all pertaining, no doubt, to the costume of the person buried. The piece of cloth shown in plate III probably served as a mantle or skirt and is 46 inches long by 24 wide. It is of coarse, pliable, yellowish-gray stuff, woven in the twined style so common all over America. The fiber was doubtless derived from the native hemp, and the strands are neatly twisted and about the size of average wrapping cord. The warp strands, 24 inches in length, extend across the piece; and on the left margin, as seen in the illustration, they are looped for the passage of a gathering string, while on the left they have been cut to form a short fringe. The opposing series (the woof strands) have been passed through with the length of the cloth in pairs, which are twisted half around at each intersection, inclosing the web strands in alternating pairs as shown in detail in figure 5. These twined strands are placed three-eights of an inch apart, the web being so close that the fabric is but slightly open. The twined strands are carried back and forth in groups of four as shown at the ends in the plate, and are knotted as illustrated in the figure.

A piece of fabric of much interest is presented in plate IV. It may be an unfinished garment of the class shown in the preceding illustration, but it is more likely a complete skirt, the narrow woven band with its gathering string serving as a belt and the long fringe being the skirt. The length at the gathered edge is 34 inches, and the pendant length is 20 inches. The material and the weaving are the same as in the piece of cloth already described, although the work is somewhat coarser. A detailed study of the border is given in figure 6, the vertical series of threads being pulled apart to show more distinctly the manner of combination.

The two pieces just described would seem to correspond pretty closely with the garments formerly worn by women and girls of the lower Mississippi country, as illustrated by Du Pratz in a plate facing page 310, volume II, of his Histoire de la Louisiane. His plate is reproduced in figure 7. The following are translations of his descriptions of the garments delineated:

The women in warm weather have only a half ell of limbourg, with which they are covered; they fold this cloth around the body and are well clothed from the waist to the knees; when they have no limbourg they use in the same way a deer skin.[54].

When the girls reach the age of eight or nine years they are clothed from the waist to the ankles with a fringe of threads of mulberry bark, fastened to a band

which is attached below the abdomen; there is also another band above the abdomen which meets the first at the back; between the two the body is covered in front by a network which is held there by the bands, and at the back there are merely two large cords, each having a tassel.[55]

Of equal interest to the preceding is the badly frayed bag shown in plate V. It is 20 inches in length and 13 inches in depth. The style of weaving is the same as that of the two preceding examples; a peculiar open effect is produced by the rotting out of certain strands of dark color, which were arranged in pairs alternating with eight lighter threads. The construction of the border or rim of this bag is quite remarkable. As shown in figure 8, the upper ends of the vertical

strands are gathered in slightly twisted groups of four and carried up free for about two inches, when they are brought together and plaited with remarkable neatness into a string border. As if to convey to the curious investigator of modern times a complete knowledge of their weavers' art, the friends of the dead deposited with the body not only the fabrics worn during life but a number of skeins of the fiber from which the fabrics were probably made. This fiber has been identified as that of the _Cannabis sativa_, or wild hemp. Two of the skeins are shown in plate V.

The presence of these unworked materials makes it probable that the individual burned was a female, for the distaff and the loom have been and are universal emblems of the practical enslavement of that sex.

A small but very instructive group of burial fabrics is preserved in the National Museum. These specimens were found with a desiccated body in 1877 in a cave 8 miles from Mammoth cave, Kentucky. They consist of a number of bags and other articles woven in the usual styles of bast and hemp. Nearly all of the articles are worn or fragmentary, but the fiber is wonderfully preserved and the original colors are as fresh as if the burial had taken place but yesterday. There are three wide-mouthed, shallow bags, resembling the one from Tennessee illustrated in plate V. The largest is 34 inches long when closed, and 15 inches deep. Both web and woof are of bast. There is a border of open work bound by a plaited band as seen in figure 8, and the manner of weaving is identical with that shown in that figure. The second bag is 22 inches long and 16 deep. The web is of bast, the woof of hemp. The smaller specimen is 14 by 9 inches and is made exclusively of hemp, and is thus much more pliable than the others. The small remnant of a larger bag shows a web of heavy, plaited bast strands resembling the specimen impressed on pottery and shown in _a_, plate IX. Besides these pieces there is a bit of heavy, compactly woven stuff, resembling the broad part of a sling, which shows traces of a geometric pattern, and a piece of flattish rope 12 feet long and 12 inches broad plaited very neatly of hempen twine.

Among a number of cave relics from Kentucky donated to the Museum by Mr. Francis Klett, are some textile articles. Among these is a sandal or moccasin woven or plaited very neatly of bast. It is shown in figure 9. Prof. F. W. Putnam and other explorers of these caves have obtained numerous textile articles of interest.

CHARRED REMAINS OF FABRICS FROM MOUNDS.

That the well-preserved fabrics just illustrated represent fairly the textile work of the mound-builders is practically demonstrated by the evidence furnished by the mounds themselves. From hundreds of sources come the same story; and it is not necessary here to enter into any elaborate discussion of the subject or to multiply illustrations. I present in plates VI and VII specimens of mound fabrics which, since they were burned with the dead, undoubtedly formed part of the clothing of the living or were wrappings of articles deposited with the bodies. These coarse cloths may be considered as fairly representing the weaving of the mound-builders. There are among them some finer examples of weaving than those obtained from the caves and shelters of Tennessee and Kentucky, but there is nothing specifically different in material or methods of combination, and there is nothing whatever to suggest a higher stage of culture than that of the historic Indian.

The fiber is quite fine and is more probably of hemp than of the bark of trees. The strands are generally well twisted and even, the twist being in most cases to the right, or as if twisted on the thigh with a downward movement of the right-hand, the thread being held in the left. As in the case of cave fabrics as well as the work of the modern peoples of the region, the weaving is nearly all in the twined style, of which there are two varieties; one in which each strand of the web is in turn inclosed simply by the woof twisted in pairs, and the other in which alternate pairs of the web strands are inclosed by the twined pairs of the woof. Cloths woven in the first method are often quite close, as the woof threads are readily pressed or pounded down on one another entirely hiding the web strands, giving a fabric of much compactness and strength. The second variety is usually somewhat open and net-like, and very often the pairs of twined woof strands are placed far apart, as shown in several of the illustrations given in this paper. The finest mesh observed is in the first of these styles, and includes about twenty intersections to the inch.

From the Ohio mounds also there are examples of plain as well as of diagonal interlacing. In appearance the cloth is much the same as that done in the twined style. In a few cases a border or selvage of very simple construction is seen. A looped margin for the passage of a gathering cord is common.

In plate VI a number of bits of charred cloth are shown; being quite black the camera fails to give them with clearness, but the drawings presented in plate VII serve to make clear all details of the strands and their combination. The charring has taken place in cremating the dead, in the burning of offerings or through accidental subjection to heat. In some cases very considerable portions of the cloth are found, but it is usually in a very fragile state and little has been preserved.

Specimens preserved in this way are obtained from a large area, including the Ohio and a large portion of the Mississippi valleys.

FABRICS PRESERVED BY CONTACT WITH COPPER.

The preservation of woven textures through association in burials with implements or other articles of copper is of common occurrence. Our museums contain many examples of copper celts retaining on their surfaces portions of cloth so well preserved that the fibers retain much of their original strength as well as color. In plate VIII three examples are shown from a mound near Davenport, Iowa, and a fourth from a mound near Savannah, Georgia. The fabrics on _a_ and _b_ are of the twined style and, although occurring 800 miles apart, are identical in every respect. The cloth on _c_ is very closely woven and has the appearance of simple interlacing. The finest piece of work that has come to my notice is a bit of cloth from a mound in Pike county, Ohio. It has from thirty-five to forty strands to the inch, and looks much like coarse twilled goods. It is woven in the twined style, however, and is therefore of native origin. It was preserved by contact with a large number of copper beads, four of which are shown in the cut, figure 10.

Traces of basketry are rarely preserved either by charring or by contact with copper. Matting is occasionally preserved in these ways. Figure 11 illustrates a piece of rush matting found fixed to the surface of a bit of copper in a mound near Augusta, Georgia.

The weaving of the hair of many species of quadrupeds, the buffalo, the opossum, the rabbit, etc., is noted by a number of authors, and a few specimens of haircloth have been recovered from mounds. Mr. Henry R. Howland found in a mound near Alton, Illinois, two varieties of cloth preserved by contact with a copper ornament representing a turtle-shell; they are described as follows:

Closely fitting over the outer surface of the copper shell is, first, a woven cloth of a vegetable fibre, similar in its general character to the outer matting above described, but of a stronger and better preserved fibre, apparently more like that which forms the woven coating of the Davenport axes. This is covered in turn with a softer, finer fabric, now of a dark-brown color, formed of twisted strands, laid or matted closely together, though apparently not woven. The material of which these strands are formed proves, under microscopic examination, to be animal hair.[56]

An illustration of ancient split cane matting is presented in figure 12. The specimen was obtained from Petite Anse island, near Vermilion bay, southern coast of Louisiana, and a photograph was presented to the Smithsonian Institution in 1866, by J. F. Cleu. The following description, as given by Prof. Joseph Henry, appears on the label attached to the specimen:

This fragment of matting was found near the surface of the salt, and about 2 feet above it were remains of tusks and bones of a fossil elephant. The peculiar interest in regard to the specimen is in its occurrence in situ 2 feet below the elephant remains, and about 14 feet below the surface of the soil, thus showing the existence of mart on the island prior to the deposit in the soil of the fossil elephant. The material consists of the outer bark of the common southern cane (_Arundinaria macrosperma_), and has been preserved for so long a period both by its silicious character and the strongly saline condition of the soil.

FABRICS IMPRESSED ON POTTERY.

It was a common practice among the aborigines to employ woven fabrics in the construction and ornamentation of earthenware. Impressions were thus left on the clay, and by baking these were rendered as lasting as if engraved on stone.

From no other source do we obtain so wide a range of fabrics. The fabric-marked vases and sherds are obtained from mounds, graves, and village sites all over the country. There is not a state within the Mississippi or Atlantic drainage that does not furnish some example of the preservation of native fabric impressions on earthenware. The perfection with which every character of these textures is preserved is well shown in a number of the figures here introduced.

A somewhat extended study of this subject was published in the Third Annual Report of the Bureau of Ethnology, and illustrations of nearly all the styles of weaving were given. As indicated by subsequent investigations, a number of slight inaccuracies of analysis and drawing occur in that paper, but they are of such minor importance that detailed correction is unnecessary.

It would seem that imprints of cloth woven in the plain interlaced style appear to be quite rare, although it is difficult, from the impressions on clay, to distinguish this from other forms when the threads are closely impacted. In somewhat rare cases the interlacing is so arranged and alternated as to give diagonal effects as in a specimen shown in figure 13. These effects are peculiar to the interlaced fabrics, not being produced in twined or netted work.

It has been supposed that vessels of clay were often modeled in baskets, and that the native earthenware preserved numerous impressions of baskets. On closer analysis these impressions turn out to be the application of pliable cloths, or of cords singly or in groups, or of stamps covered with textiles or having geometric textile-like patterns engraved on them. I can not recall a single example from eastern United States in which it is entirely clear that the clay vessel was modeled in a basket. The impressions of basket work occasionally seen are only partial, having been applied after the vessel was practically finished.

I present in figure 13, a small earthen vessel from a mound in North Carolina, the entire exterior surface of which is marked with a fabric, a pliable cloth or bag woven in the twined styled. The impressions are not the result of a single application of the texture, but consist of several disconnected imprintings as if the hand or a paddle covered with cloth had been used in handling the vessel or in imparting a desired finish to the surface.

Specimens of diagonal fabrics, restored from potsherds, are given in figures 14 and 15. The first is a very neatly woven diagonal from the ancient pottery of Polk county, Tennessee. Two series of cords have been interwoven at right angles to each other, but so arranged as to produce the diagonal effect. One series of the cords is fine and well twisted, the other coarser and very slightly twisted. The second is a piece of matting restored from the impression on a small piece of pottery collected in Alabama. It was probably made of rushes or heavy blades of grass.

Twined weaving prevails in the fabrics impressed on pottery as in those from all other aboriginal sources. An example of the simplest form, obtained from a small fragment of pottery found in Polk county, Tennessee, is shown in figure 16. Two series of threads are interwoven at right angles, the warp being arranged in pairs and the woof singly.

At each intersection the pairs of warp threads are twisted half around upon themselves, inclosing the woof threads and holding them quite firmly, so that the open net-like effect is well preserved even under strain or in long continued use. There are many varieties of this form of fabric resulting from differences in size and spacing of the threads. These differences are well brought out in the succeeding figures.

In figure 17 we have a characteristic example of this fabric, obtained from a fragment of pottery from a mound at Sevierville, Tennessee.

The impression is quite perfect. The cords are somewhat uneven, and seem to have been only moderately well twisted. They were probably made of hemp fiber. It will be observed that the threads of the web are placed at regular intervals, while those of the woof are irregularly placed. It may be noticed that in one case the woof has not been doubled, the single thread having, as a consequence, exactly the same relation to the opposing series as corresponding threads in simple interlacing. The impression, of which this is only a part, indicates that the cloth used in shaping the vessel was considerably distorted when applied to the soft clay.

Nowhere else are found so many fine impressions of fabrics on clay vessels as in the ancient salt-making localities of the Mississippi valley. The huge bowls or vats used by the primitive salt-maker have generally been modeled in coarse, open fabrics, or have had cloths impressed upon them for ornament. In figures 18 and 19 fine examples of these impressions are given. The latter engraving illustrates a specimen in which every detail is perfectly preserved. Only a small portion of the original is shown in the cut. It is noticeable that the cords are quite heavy and well twisted, although the spacing is somewhat irregular.

The example given in figure 20, impressed on a fragment of clay from Arkansas, has an ornamental border produced by looping the cords of the web, which seem to have been five in number, each one passing over four others before recrossing the frame. A specimen showing a somewhat different border is given in figure 21.

The interesting specimen illustrated in figure 22 was obtained from a small fragment of pottery found in Ripley county, Missouri. The combination of the two series of strands clearly indicates the type of fabric, the twisted cords of the woof being placed very far apart. The warp is of braid formed by plaiting strands of untwisted fiber, probably bast. All the details are shown in the most satisfactory manner in the clay cast.

In figure 23 we have a similar fabric closely woven or impacted. I have made the drawing to show fillets of fiber appearing at the ends; these do not appear in the impression. It is highly probable, however, that these fillets are plaited bands, as in the preceding example. They are wide and flat, giving somewhat the effect of basket-work of splints or rushes.

Another variety of the twined fabrics, distinguished by peculiarities in the combinations of the threads, is illustrated in figures 24 and 25. The threads of the warp are arranged in pairs as in the specimens already described, but are twisted in such a way as to inclose two of the opposing series instead of one, each succeeding pair of warp threads taking up alternate pairs of the woof threads. Figure 25 is from a small piece of pottery exhumed from a mound on Fain island, Jefferson county, Tennessee. The threads of the woof are quite close together, those of the web being far apart.

That the native love of decoration had a marked influence on the weavers' art in its simplest and rudest as well as higher forms is well evinced even in the meager vestiges brought to light by researches in the mounds. Decorative borders and fanciful combinations of strands are shown in some of the preceding cuts, and figure 26, copied from a pottery fragment obtained in the Ohio valley, indicates a more ambitious attempt at embellishment. The fabric was evidently of ornate design and the execution excellent.

Plate IX is intended to convey a clear notion of the nature and appearance of fabric-marked pottery and of the manner of securing positive impressions in clay. Three bits of pottery from Illinois are placed at the left, and the three casts appear at the right. All illustrate open fabrics of comparatively simple pattern done in the characteristic twined style.