University of Virginia Library

Section F

ARTIFACTS OF PERISHABLE MATERIALS[204]

By Harry Tschopik, Jr.

The number of perishable objects recovered from the excavations
was relatively large, considering their provenience and the open nature
of the site. In all, 26 specimens were studied and classified according
to the following categories, arranged in the order of their abundance:
matting, fabrics, basketry, wood, and leather. In practically all cases
the specimens were extremely fragmentary and in an advanced state of
decay. A very large proportion was associated with the burials in
rooms 2 and 5 (cf. table 3).

Matting.—Three varieties of matting, comprising 16 specimens,
were represented in the collection: plaited, twined, and threaded. Of
these, all except one were associated with burials.

Of the 12 plaited mats, 10 examples were twilled in the under-three-over-three
technique,[205] and 1 in the under two-over-two. The
woven elements probably were of Yucca glauca Nutt. and ranged from
3 to 5 millimeters in width in given specimens, with an average width
of 4 millimeters.[206] In not a single specimen was a portion of the selvage
preserved. The twelfth specimen was in such a fragmentary condition
that, although the technique was clearly twilled, the details could
not be observed with accuracy.

Two examples of twined mats were encountered. One is made of
reeds[207] placed parallel and twined together at intervals of 8 centimeters
along their lengths.[208] The nature of the twining material is uncertain,
but an adobe cast which accompanied the specimen indicates that it
was a twisted multi-fiber cord, 1 millimeter in diameter.[209] The twining
is of the simple type in which the 2 elements are half-twisted about one
another after twining each reed.

The second twined specimen is made also of reeds (Sporobolus
wrightii Munro), but in a finer and slightly different technique. This
technique of twining differs in that the elements are half-twisted on
one another after twining each pair of reeds, and are spaced at intervals
of 6 centimeters along their lengths. The twining material is a 2-ply
fiber cord with a right spiral, two twists per centimeter, and 1
millimeter in diameter.[210]

A third type of mat, represented by a single specimen, is made in
the threaded technique.[211] The reeds (Sporobolus wrightii Munro) are
placed side by side and are held together by cords which pass through
them at right angles to the long axis. Although the intervals of
threading were not determined, each cord is 2-ply, with a right spiral,
2½ twists per centimeter, and 1 millimeter in diameter.

Fabrics.—Specimens in this category are 5 in number, and are of
two types: twined woven and simple under-and-over weaving. In all
cases, the fabrics were found in connection with burials.

Fabrics of cordage made in the twined woven technique are represented
by 4 very fragmentary specimens of which 2 are probably a
variety of feather cord cloth. In the remaining 2 examples, the presence
of feathers could not be detected. As regards the details of the
weaving technique, the twined fabrics seem to be identical technologically
with the fur cloth example illustrated by Guernsey and Kidder,
except for the fact that the warp selvages are not represented in the
Bc 51 specimens.[212]

In all examples of feather cord cloth of which the writer is aware,
the foundation of the weft element consists of some twined cord
around which whole or split quills were then wrapped. In respect,
however, to the feather cord cloth examples from Bc 51, at least one
feature seems quite unusual. Microscopic analysis revealed no indications
whatsoever of a vegetable fiber cordage foundation for the
weft element.[213] A short length of weft, which seems to meet this requirement,

was produced experimentally in the following manner. Turkey
tail feathers were split down the quill, and the coarser body of the
quill was then scraped away with a knife. Following this, the two
strips of quill with barbs and down attached were loosely twisted upon
one another to produce a loose, 2-ply cord with the barbs of the
feathers serving as the pile. It has been noted elsewhere in the case of
fur cloth specimens that "strips of tough skin with the hair on were
sometimes twisted upon themselves instead of being wound about a
cord."[214]

The cordage of the Bc 51 specimens, including the warps of the
feather cord cloth, is of the 2-ply variety, with a right spiral. In one
example, the weft selvage is a heavy cord measuring 4 millimeters in
diameter, while the other elements average 1½ millimeters in thickness,
with 1½ twists per centimeter. The wefts are somewhat finer and
more tightly twisted, averaging 1 millimeter in diameter, with 2
twists per centimeter. In the other specimens, the twining elements
and wefts are of equal weight, the cordage being 1 millimeter in
diameter, with 2 twists per centimeter. In these examples, it should be
pointed out, none of the selvage was preserved.

But a single example of plain weave cotton cloth (under-one-over-one)
was recovered. Since the specimen is small, neither pattern nor
selvage is represented. It seems probable that the wefts are the coarser
threads, while the more tightly twisted ones are the warps. This being
the case, there would be 14 warps and 11 wefts per centimeter. Both
elements seem to be single-ply.

Basketry.—Of the three examples, two are made in the close coiled
technique.[215] Both are made of some dicotyledonous wood which may be
Rhus trilobata, though conclusive identification is lacking. One specimen,
represented by the bottom fragment of a basket, is made on the
two-rod-and-bundle triangular foundation of the coil in progress as
well as the bundle of the coil below. The basket is worked on the concave
surface, to the left of the worker, with a counter-clockwise spiral.
There are 16 stitches per inch, but the number of coils per inch could
not be determined.

The second close coiled specimen is made on a single-rod foundation
with non-interlocking stitches which split the rod of the coil below.
The stitches are split, quite frequently, on the non-work surface. The
basket is worked on the concave surface, to the left of the worker, with
a counter-clockwise spiral. There are ten coils and ten stitches per
inch.[216]

The third basketry specimen is very interesting and deserves a
more detailed treatment. It seems possible that it is a fragment of a
"bird's-nest" coiled storage basket of the general type reported from
among the Pima, Yuma, and most southern California groups. The
specimen measures 11 by 30 centimeters, and is composed of roots
(Rhus, species probably trilobata) which vary from 2 millimeters to 2
centimeters in diameter. The arc of the fragment is such as to suggest
a basket some 2½″ in diameter. Fragments of a variety of twilled
mat were found on the concave surface of the specimen in such a
position as to suggest a lining.

The fragment was encountered on the floor in the southeastern
corner of room 18, at a depth of 40″, in association with a Red Mesa
Black on White bowl. There is a possibility that it may also have been
associated with burial Bc 60/25.

Wood.—Two specimens of wood are reported in the field catalog
cards. One, which was not seen by the writer, is described as a weaving
stick "of somewhat roughened wood, thicker at one end, with a
cleft in it." The measurements given are: length, 8½″ (216 mm.);
width ¾″ (19 mm.); thickness, ¼″ (6 mm.) to ½″ (12 mm.).

The second specimen is curious, and it is impossible to say whether
it has been worked or not. It is a small, flat, charred strip of bark
identified as Pinus ponderosa or Pinus brachyptera Engelm. The strip
measures 3½ centimeters long by 23 millimeters wide, and is beveled
at either edge to a thickness of 2 millimeters. The central thickness is
1 millimeter.

Leather.—The one fragment of leather recovered seems to have
been a part of a strap. One straight edge apparently shows evidence
of having been cut with a sharp instrument. The specimen measures 3
centimeters in length, 1 centimeter in width, and 1 millimeter in
thickness.

Comparisons and Distributions.—Since the perishable objects
from mound Bc 51 have now been described, it remains to compare
them with similar objects from Bc 50[217] and Łeyit Kin,[218] since it might
be expected that these collections would exhibit great uniformity.

In respect to twilled matting it is noted that frequent examples
were encountered at Bc 50, as at Bc 51, in connection with burials; but,
whereas all specimens but one from Bc 51 were of the under-three-over-three
variety, several examples from Bc 50 were twilled in an
under-two-over-two technique. Again, though Bc 50 produced an
example of checkerwork matting (under-one-over-one), this technique
was not encountered in the collection from Bc 51. In regard to twined
mats, the 2 specimens from Bc 51 seem to have been made of Sporo-

bolus, although they are technically comparable to those from Bc 50,
where either Equisetum or Sporobolus was employed.[219] The threaded
matting technique seems not to have been represented at Bc 50. The
specimens of matting from Łeyit Kin are not fully described.

Feather cloth, which was represented by several examples in the
Bc 51 collection, was not encountered at Bc 50, although from the
latter site fragments of what may have been a twined bag are
described. Plain woven cotton cloth was not found at Bc 50. Textiles,
apparently, are not represented at Łeyit Kin.

As regards basketry, no coiled specimens were recovered at Bc 50,
although some sherds from the superstructure level at this site retained
impressions of coiled baskets on their exteriors. These coiled
baskets seem to have been made on a single-rod foundation, and to have
been considerably finer in texture than the single-rod specimen from
Bc 51, since the impressions indicate 15 coils to the inch. The remains
of a twilled ring basket were discovered at Bc 50, but were lacking in
the Bc 51 collection. On the other hand, "bird's nest" coiling was not
represented at the former site. With the exception of a few charred
fragments of coiled basketry which seem to have been of the "rod and
bundle" type, no actual specimens of basketry were recovered at Łeyit
Kin.[220] Adobe casts, however, reveal the presence of twilled baskets, two
of which seem to have been made in a fancy diamond weave, and a
third which probably represents the remains of a twilled ring-basket
manufactured in the under-two-over-two technique.[221] Judd has reported
a two-rod and splint basket from his Chaco Canyon pithouse
No. 2.[222] In general, this technique seems to be a well-recognized variation
of the more common two-rod and bundle variety.

To sum up the foregoing, therefore, we may state that, whereas
the collections of perishable artifacts from Bc 51 and Bc 50 resemble
and to a large extent complement one another, they are far from
being identical in regard to their total assemblages. The collection
from Łeyit Kin seems to have little in common with these. This is
probably to be interpreted as being due to the arbitrary selective factors
of environment which govern the preservation of perishable objects,
and is of no archaeological significance. A composite list of the
artifacts of this category from the three sites under consideration gives
us a rather adequate picture. The objects are categorized according
to their probable use as follows:

Containers:

• 1. Coiled basketry

  • a. Two-rod-and-bundle

  ---

  [99

  
  
  

  • b. Single-rod

  • c. "Bird's nest" storage baskets (?)

• 2. Twilled ring-baskets

• 3. Twined bags (?)

Clothing:

• 1. Cotton textiles

• 2. Twined-woven feather robes

Matting (for household, burial, and roofing usages):

• 1. Plaited mats

  • a. Checkerwork

  • b. Twilled under-two-over-two

  • c. Twilled under-three-over-three

• 2. Twined rush mats

• 3. Threaded rush mats

Since the culture elements listed above have, with the exception of
coiled basketry, received exceedingly slight attention from the archaeologist,
it might be profitable to trace the distribution of each within
the Southwest. But since coiled basketry is still manufactured by
numerous Southwestern groups, and because it has received a somewhat
detailed treatment in the literature, this trait has assumed a
comparative significance which seems to warrant a more thoroughgoing
examination. It will, therefore, be treated separately later.

With regard to the other elements of material culture described
above, the proposition is very different indeed. Either it has been
tacitly assumed that these traits have no comparative value, or investigators
have been disinterested in them. Additional factors, such as
those of preservation and the technical difficulties involved in the study
of the specimens, have served to relegate these artifacts to obscurity.
It seems probable, however, that these objects, if properly studied,
could shed their own peculiar light on the problems of Southwestern
archaeology.

No exhaustive examination of the literature in regard to these culture
elements has been made, and conclusions at this time seem inadvisable;
indeed, it seems possible at present only to indicate certain
problems.

Weaving in Cotton.—Textile weaving in cotton seems to have
been introduced into the Southwest at about the beginning of Pueblo I
times in the San Juan region, and to have persisted among the Pueblos
into historic times.[223] That this trait was not present in Basket Maker
II times seems conclusive and, if it occurs in Basket Maker III sites, it
seems to have been rare. The extent to which this trait occurred to the

north and west is as yet problematical, but in respect to the southern
periphery of the Southwest, we are on firmer ground. Cotton textiles
were discovered in the Sacaton Phase levels at Snaketown,[224] and persisted
in that area into Classical Hohokam times.[225] Indeed, Beals has
shown that the weaving of cotton cloth, as practiced by living ethnic
groups before 1750, extended in a continuous belt along the western
coast of Mexico from the Southwest to the Valley of Mexico.[226] The
earliest appearance, however, of cotton textiles in northern Mexico is
as yet unknown. Dry cave material which seems to represent a BasketMaker-like
horizon in northern Mexico did not reveal the presence of
this culture element.[227] The eastern boundary of cotton growing and
cotton weaving is equally vague. These techniques seem not to have
been known to the cultures of the Lower Pecos River and the Big Bend
region of Texas.

Feather Cord Cloth.—At precisely what period feather cord cloth
appears in the central Pueblo area is uncertain. It seems probable
that the trait is first found in this region in late Basket Maker III
times.[228] In respect to northeastern Arizona, Guernsey states that
feather cord cloth appears in Pueblo I times in the Marsh Pass
region and continues in use into the Pueblo IV period.[229] On the
Pueblo III horizon this trait occurs widely: at Mesa Verde,[230] Pueblo
Bonito in Chaco Canyon,[231] the Mancos-La Plata region of southern
Colorado,[232] at Jemez Cave, New Mexico,[233] and in the Sierra Ancha
Mountains of central Arizona, where it is mentioned that feather
cord cloth is rare.[234] Mr. Paul Reiter informs me that feather cord
cloth was also found at Chetro Ketl in Chaco Canyon. Fewkes also
found the trait present at Casa Grande in southern Arizona.[235] At
Pecos[236] and at Puyé,[237] this trait persisted into Pueblo IV times. In
all these cases, it is to be noted, the technique of manufacture was not
that represented in the collections from Bc 51; instead, fiber cords
were wrapped with strips of quill or with strips of bird skin with the
feathers attached.

The precise boundaries of the trait of feather cord cloth are extremely
vague. But a single example, and this doubtful, has been
discovered in the Big Bend area.[238] For northern Mexico, where very
little work has been done, only one occurrence seems to have been
reported, this being found in the Basket Maker-like culture of the
Upper Rio Fuerte region of southern Chihuahua.[239]

In view of this fragmentary evidence, one wonders how the data
on the distribution of feather cloth are to be interpreted. Technically
some feather cloth is identical with fur cloth, although the latter trait
seems to have a far wider distribution. It may be possible that the
making of feather cloth is to be correlated with the keeping of the
domesticated turkey, although the data seem to indicate that feather
cloth appears in the Southwest somewhat before the appearance of the
latter. Again, feather cloth was made in regions where the domesticated
turkey is not known to have been kept.

Plaiting.—In regard to the technique of plaiting[240] the problems
are complicated at the outset by the simplicity of the mechanical principles
involved and the great distributional ranges of these techniques
in North America.

Turning our attention to the distribution of plaiting, we are confronted
immediately with a basic problem: Is there any justification
for studying the distribution of a technique irrespective of the cultural
form in which it manifests itself? May we, for example, equate
the presence of a twilled mat in one region with the presence of a
twilled ring-basket in another? Underlying such an equation is the
assumption that the techniques are historically related and that it is
the technique which is the fact of greater significance. The problem
of independent invention versus diffusion need not here concern us,
since our problems are of a descriptive rather than of an interpretative
nature. The issue in question is the comparability of techniques.

That such an assumption is not considered entirely unwarranted
is indicated by the frequency with which it has been employed. Wissler
shows the distribution of stone sculpture to have been continuous from
the Valley of Mexico to Central America irrespective of whether this
sculpture adorned a temple or an isolated monolith.[241] Again, Beals
traces the distribution of metallurgy in Mexico, disregarding the articles
manufactured.[242] Although it is true that he refers in his tables

to the individual articles made, it is evident that it is the total distribution
which he seeks to demonstrate.

In the Southwest, the plaiting technique was used to manufacture
several types of articles: mats, sandals, baskets, and occasionally
cradles. Since in this era a reasonably accurate chronology affords a
check on the data, it may be well to test this procedure in question by
applying it to the technique of plaiting.

In the classical Basket Maker area of the Pueblo plateau, twilled
mats seem to be absent in both the Basket Maker II and III periods.
Again Guernsey, speaking of twilled ring-baskets, states: "This type
of basket does not occur, so far as we know, in the Basket Maker II
or Basket Maker III period, but is the commonest type found in the
Pueblo III period."[243] In fact, the only twilled specimens from Basket
Maker times in this region seem to be a bag-shaped basket made in the
under-two-over-two technique,[244] and occasional twilled yucca leaf
sandals.[245]

But as we approach the Big Bend region of Texas, twilled objects
become more common. In the Guadalupe Mountain country of southeastern
New Mexico, several types of plaited artifacts are found on a
general Basket Maker-like horizon. These take the form of checker
mats,[246] checker weave baskets, twilled mats under-two-over-two, and
twilled sandals.[247] In the Big Bend region proper, Setzler states that
checker and twill under-two-over-two mats occur in quantities on a
similar Basket Maker-like horizon.[248] In addition baskets are made in
both twill and checker techniques.[249] It seems significant that the
under-three-over-three twill technique does not appear to have been
used in the Big Bend either for mats or for baskets.

To the west and north of the Pueblo plateau, sites so far excavated
seem not to have yielded artifacts in the twill plaited technique;
none were encountered in either Gypsum or Lovelock Caves.[250] Steward
reports checker mats from caves in the Great Salt Lake region of
Utah, although the culture represented here is almost certainly of a
later date.[251] To the south, checker and twill are represented in dry
cave material from Coahuila, Mexico.[252] Again, Zingg reports that
under-three-over-three twilling is characteristic of the Basket Makerlike

culture of the Upper Rio Fuerte region in the State of Chihuahua,
Mexico.[253]

In Pueblo I and II times, twilled ring-baskets (under-three-over-three)
are found in northeastern Arizona,[254] and in the Hohokam area
of southern Arizona.[255] During these periods twilled ring-baskets and
mats (under-two-over-two; under-three-over-three; checker) are
found in Chaco Canyon.

By Pueblo III times, the plaiting technique becomes widespread
throughout the Pueblo area, and is utilized in several ways. Sandals
made in checker or under-two-over-two technique occur widely; at
Jemez Cave,[256] in the Mancos-La Plata region of Colorado,[257] in northeastern
Arizona,[258] and in the Mimbres Valley.[259] Ring-baskets in under-two-over-two,
under-three-over-three, or in a combination of these techniques
occur so frequently in Pueblo III sites as to be the rule rather
than the exception. The same may be said of plaited mats, which are
made in the three techniques: under-three-over-three, under-two-over-two,
and checker, though the latter technique seems to be the most
unusual of the three. At this time the twill technique is employed
occasionally in making cradles.[260] In the Hohokam area, under-two-over-two
twilled mats seem to have been used.[261]

In Pueblo IV times, although the data are less complete,[262] it seems
that the general distribution of articles manufactured in plaited techniques
resembled that of the Pueblo III period.[263]

The extent to which plaited articles survived into historic times
was not investigated. Suffice it to say that twilled ring-baskets are
still manufactured by several Pueblo groups, notably the Hopi and
Zuñi.[264] Twilled mats in the under-two-over-two and under-three-over-three
techniques were formerly made by the Pima.[265]

From the foregoing, therefore, it would seem unwise to attempt
to trace the distribution of twill plaiting in the Southwest irrespective
of the object manufactured. For the very nature of the distributions
and manifestations of this technique should make us suspicious of our

historical identities. Holmes observed that "no class of articles of
textile nature were more universally employed by the aborigines than
mats of split cane, rushes, and reeds . . .".[266] The distribution of
twilled and checker basketry is equally widespread, since baskets in
these techniques are found in such widely separated regions of North
America as the Northwest Coast, the Southwest, and the Southeast.

It seems significant, therefore, that checker and under-two-over-two
twilled mats are found on a Basket Maker-like horizon in the Big
Bend area, whereas they seem not to have appeared in the San Juan
region until Pueblo I times or later. It also seems significant that the
under-three-over-three twill technique of the Pueblo area has not yet
been reported from either the Hohokam or the Big Bend regions. In
America north of Mexico, there seem to have been at least two regions
in which twill plaiting was centered: the Southeast and the Northwest
Coast. Mexico proper may have comprised a third. It is, therefore,
felt that the factors involved in the study of plaiting in the
Southwest are far from being simple, and that each technical variety
and each type of artifact must be treated for present purposes as a
separate historical entity.

Twined Mats.—The difficulties pertaining to plaiting techniques
apply also to the question of twined mats. In the case of the latter,
however, twined mats made of rushes or reeds have such a widespread
distribution in North America, both in regard to time and space, that
their comparative significance is at present very slight. In addition,
they have received such slight recognition in the literature that we are
able here only to note their presence and absence. It is highly probable,
however, that all twined mats in all places in North America
are not historically related, and it seems very likely that a careful
study would reveal significant sub-types.

Twined rush mats, according to Guernsey, do not occur in Basket
Maker II or III sites in northeastern Arizona, but appear in Pueblo I
times and persist in this region into the Pueblo IV period.[267] Interestingly
enough, however, numerous and varied examples of twined
mats were recovered by Nusbaum from a Basket Maker cave in Kane
County, Utah.[268] Twined mats were found in caves in the Guadalupe
Mountain region of southeastern New Mexico, where they are said to
have been rare.[269] In respect to this trait, northern Mexico is an unknown,
but for the northern and western periphery of the Southwest,
we have a certain amount of data. In Nevada, twined mats were found
to the exclusion of plaited types.[270] Again, in the Great Salt Lake

region of Utah, twined mats were common, whereas plaited examples
were rare.[271] In the Fremont River region of Utah, twined mats are
found in Pueblo sites.[272] By Pueblo III times, twined reed and rush
mats are found widely throughout the Southwest, and the same is probably
true for the Pueblo IV period.

In view of the difficulties stated at the outset, interpretation at this
time is difficult. The data, however, seem to indicate that there is a
greater possibility that the technique of twined matting spread into
the Southwest from the Great Basin, or from the north, than that
it entered this area from the east.

Threaded Matting.—This technique has been mentioned so rarely
in the literature that it is quite impossible to trace its distribution in
either time or space. Pearce and Jackson describe it from a Big Bend
cave dweller site,[273] while Fewkes found it to be represented at Mesa
Verde.[274] It has also been reported by Morris from Aztec[275] and by
Reiter from Chetro Ketl.[276]

Coiled Basketry Techniques in the Southwest:
Distributions and Continuities

During the six years which have elapsed since the appearance of
Dr. Weltfish's excellent paper on Southwestern basketry techniques,
a rather extensive body of pertinent data has been accumulated in the
course of various excavations.[277] Particularly is this true of the areas
of southern New Mexico and Arizona and of the Big Bend region of
Texas. In view of this, it is felt that a summary and re-examination
of the data relating to the distribution of Southwestern coiled basketry
techniques in both time and space will not be inappropriate. In this
way we will be enabled to evaluate with greater precision the appearance
of given techniques, such as those represented in the collections
from Bc 51, in a particular locality during a particular period.

The method, therefore, will be to present the distributions of these
techniques on maps and in tables, and finally to draw such conclusions
regarding the continuities and historical implications of this craft
as seem tenable in view of our present knowledge. Before we begin to
examine the data, however, it seems advisable to make some observations
with reference to the distribution of coiled basketry and to inquire
into some problems relating to this distributional study.

It is continually being brought to our attention that the development
of the Indian cultures of the Southwest cannot be properly understood
unless we conceive this territory in the broadest possible sense.[278]
Especially is this true in the case of coiled basketry, and constant reference
must be made to regions geographically remote. In view of this
it seems important to begin with a brief résumé of the distribution of
this culture element in the New World.

Wissler traced the distribution of coiled basketry throughout the
Americas and concluded that this craft was principally confined to a
belt which paralleled the Pacific coast of North America from Siberia
through Alaska to the Mexican border. He noted that it occurred
sporadically among the northern Algonkians, but that "in Mexico the
technique disappears and does not come to notice again until we reach
Patagonia."[279] More recent investigations, however, have changed the
picture to a considerable extent. Coiled basketry fragments were
found in the Cenote at Chichén Itzá, in Yucatan, and more recent evidence
of the presence of this technique has been discovered in the
Province of Coclé in Panama.[280] Finally, Nordenskiöld has shown that
in South America this craft is distributed along the west coast, in
what appears to be a continuous belt from Panama to Tierra del
Fuego.[281] In addition to its wide distribution in the New World, coiled
basketry is found under conditions which suggest a certain antiquity.[282]

In North America, Weltfish finds that coiled basketry occurs in
six of the present-day basketry making areas which she delineates.[283]
Each of these coiled basketry areas is characterized by certain technical
traits in common, although there is a certain amount of overlapping,
as is to be expected, in adjacent areas. To mention these
areas briefly, there are the following: Salish, Mackenzie, Basin, Central
California, Southern California, Southwest. In addition, she
finds that coiled basketry occurs sporadically in the Plains (Pawnee
and Arikara) and in the Northeast (Ojibwa, Menomini, and Chippewa).[284]

The value of coiled basketry in comparative studies must not be
underestimated. The fact that it is usually encountered in archaeological
sites in a fragmentary condition is compensated for by the
fact that the techniques of construction are not obliterated in the finished

product as is often the case with pottery. The perishable nature
of basketry is, of course, a handicap; however, specimens of relatively
great antiquity have been recovered from dry caves and from such
desiccated regions as the coast of Peru, and some are found preserved
in open sites where moisture has not destroyed them. As in the case
of pottery, many of the items composing the trait of coiled basketry
are mechanically independent of one another and tend to objectify
themselves in the process of manufacture.[285] This is particularly true
of four such items, which have widespread comparative value:[286] (1)
the nature of the foundation of the coil; (2) the surface from which
the basket was worked: (3) the direction in which the work proceeds;
(4) the character of the stitches.

As to foundation, Weltfish sees three principal types and their
variations represented in North America:[287] (1) multiple (a bundle of
grass or splints, etc.); (2) triangular (two-rod-and-bundle; three-rodetc.);
(3) vertical (including single rod). These terms are descriptive,
and apply specifically to characteristic coils of particular baskets when
viewed in cross-section.

Since a woman engaged in making a basket may work either on
the concave or convex surface, with the coil in progress extending
either to her left or to her right, these criteria have important comparative
value. But although these two features, work surface and
direction of the work, are mechanically independent of one another,
the latter feature cannot be determined until the former has been
established.[288] In this connection, the direction of the spiral coil, i.e.,
"left spiral" or "right spiral," may be disregarded as a comparative
feature, since it is to be stated as a product of the direction of the
sewing and the surface worked.

The purpose of the binding stitch is to hold the foundation of the
coil in progress together and to attach it to the completed coil immediately
below. The sewing may proceed in three different ways; the
stitches of the coil in progress may: (1) split the stitches of the completed
coil below; (2) interlock with them; (3) not interlock with
them.[289]

Other features in all probability have comparative value, but are
more difficult to control than those mentioned above. In ceramics,
shape and design have been exploited to the fullest extent in a comparative
sense, but in the case of basketry, certain difficulties present
themselves. In basketry, the possible choice of designs is restricted
to a certain extent by the technological processes involved in the pursuit
of the craft. Furthermore, as regards the products of contemporary

Indian basket makers, a lack of systematic and standardized
description, as well as, in many cases, of adequate photographs
tends to nullify the comparative value of design and shape. Coupled
with this is the fact that these features usually may not be observed
in fragmentary archaeological specimens. Texture seems quite important
as a comparative feature, and although individual skill is
almost certainly a factor, this feature may be treated objectively by
counting the number of coils and stitches per inch or centimeter. The
former unit would appear to be more useful since basketry texture,
when described in the literature, is usually given in terms of inches.
Again, it is conceivable that the plant materials used in coiled basketry
might be of value in comparative studies, although the choice of these
is naturally limited by geographical consideration. Aberrant coiled
basketry techniques such as "sifter coiling" have not been considered,
because of the scanty treatment which they have received in the
literature.[290]

It would now seem appropriate to examine the limitations of the
comparative features worked out by Weltfish, and to point out certain
difficulties involved in their application. Certain of these difficulties,
most of which have been touched upon by Weltfish, are inherent in
the data themselves, while others result from the treatment of the
data in the literature.[291] Still other difficulties, as is inevitably the
case, arise out of the classification of the data and result from the
problems of chronological position and precise allocation of the specimens.

The fragmentary nature of the archaeological record is an initial
difficulty. For large areas and relatively great spans of time the data
are completely lacking, while for other regions and periods they seem
to be adequate. Arising out of this fact is the question of sampling.
In most cases we are aware at the outset that our samples are insufficient,
and for the main purposes of the present study it has been
assumed that the mere presence of a particular technique at a stated
time and place is in itself a significant fact.

The ethnological record is far more complete; but even here, as
is also true for the archaeological data, descriptions in the literature
are frequently so scanty as to be of little comparative value. This
is a situation, however, which can and should be remedied.

Turning now to the matter of classification, we are up against one
of the most general and basic problems of anthropology. It seems
reasonably clear that no classification serves for all times and for
all purposes. All classificatory systems are arbitrary, and each is
made for its own particular purpose; for without a specific purpose a
classification is meaningless. Accordingly, the merit of each must be

judged in relation to the degree to which it approximates the specific
aim in view. It is further desirable that a classification be as objective
as is possible, for out of classification grows typology. One who
sets about, therefore, to create a classification usually attempts on the
one hand to achieve a particular purpose, and on the other to be as
objective as possible. But in striving toward an aim and toward
objectivity, care must be taken not to do violence to known facts.

Weltfish has evolved a classification of coiled basketry techniques
in order to study the distribution of the types which have arisen out
of her classification.[292] She has also studied these distributions in a
temporal sense in that she has noted parallels between ancient and
modern basketry making areas.[293] Since we have already had occasion
to refer to the features classified, let us inquire into certain difficulties
inherent in the application of this classification. Since, in the manufacture
of a coiled basket, the surface worked and the direction in
which the work proceeds are mechanically unrelated to one another,
these comparative features objectify themselves in a given basket and
are, in a sense, directly comparable in all coiled baskets.[294] In regard,
however, to the other two comparative features under consideration,
the nature of the foundation and the character of the stitches, the
subjective element looms large indeed. Here we are confronted with
the comparability of these features, particularly on the archaeological
level, whenever we compare two baskets. What, precisely, is comprised
by a "multiple" foundation? Are we to include bundles of
grass, reeds, rods and splints under this single term? Are we even
justified in classifying two baskets in which the bundle foundations
consist of two different species of grass in the same category? To do
so is to equate these forms technologically. If this were done, and if
these forms were in fact technologically incomparable, any conclusions
which would be based on the distributions of these forms would be
misleading. Again, Weltfish has classified the manner of stitching
coiled baskets according to the following types: split, interlocking, and
non-interlocking. She further divides the "split" category into three
subdivisions: split on the work surface, split on the non-work surface,

split on both surfaces.[295] Of these forms, we are informed further that
the "interlocking" type is relatively unusual, which means that the
majority of baskets are made with non-interlocking, non-split stitches,
or with stitches of one of the split varieties. The question is, therefore,
do these forms constitute clear-cut, empirical types? Observations
made by the writer on specimens from Coahuila, Mexico, have
served to convince him that such is not always the case. In a single
example, some stitches were split on both surfaces, some on the work
surface or on the non-work surface alone, while others interlocked.[296]
Again, while investigating basket making among a contemporary
Navaho group, it was observed that stitches were quite frequently split
on the non-work surface.[297] When questioned about this, the women
invariably maintained that this was accidental, that the stitches should
not have been split. It would seem, therefore, that the nature of the
stitching is not always clear-cut, and that sufficient leeway should be
given to account for individual variability and accidents. Whereas,
it is obvious that experience in handling many baskets is invaluable
in making decisions, it would seem desirable to attempt to discover
some means of objectifying these judgments by determining the central
tendency in each case. This might be accomplished, for example,
by counting the stitches in the various descriptive categories and
calculating percentages.

The foregoing is not to be construed as destructive criticism; no
superior ways of dealing with the problems in question are offered in
return. But it seems necessary to keep these classificatory limitations
in mind in order to safeguard conclusions.

The final difficulty with which we have to deal is that of chronology.
For certain regions of the Southwest, notably the plateau of
New Mexico and Arizona, and adjacent portions of Utah and Colorado,
we have at our disposal both a relative and an absolute chronology.
But for other regions we are less fortunate. Although in
several of these, especially in the region of southern Arizona, we
possess a relative chronology which seems to be accurate in its essentials,
no great detail has yet been achieved in the problem of correlations
with known areas.[298] For other regions, such as sections of
Utah, southern California, western Texas, and northern Mexico, our
time estimates are based on little more than pure speculation. For

this reason, the data relating to coiled basketry have been grouped
on three distribution maps. The first map presents present-day conditions;
the second, the distribution during Pueblo III and Pueblo IV
times; and the third, the distribution prior to the Pueblo III period
of the central Anasazi region. This method, it is realized, is far from
being satisfactory; but, since we cannot attempt to be more accurate
than the data permit, this scheme seems to afford some of the leeway
necessary in allowing for time lags in peripheral regions and other
chronological discrepancies.

As thorough an attempt was made as time permitted to canvass
the literature for data concerning Southwestern coiled basketry.
Many descriptions were found to be so vague as to be of little value,
and hence have not been utilized. No systematic attempt, however,
was made to re-examine the sources used by Weltfish, since the baskets
described in these have, in many instances, been again studied by her.

Keys to Maps and Tables.—As stated above, the method adopted
here has been to plot the distribution of Southwestern coiled basketry
on three maps. In so doing, only one of the four technological units
under consideration, the nature of the foundation, has been represented
on the maps; the remaining units have been listed in three tables
which accompany the maps. It may be thought that, because foundation
has been represented diagramatically on the maps, this feature has
been given more emphasis than the others. Emphatically, however,
this is not the case. In the historical speculations based on the distributions,
all of the four features here considered have been taken
into account, and no one has been emphasized at the expense of the
others. Foundation was selected for representation on the maps
solely because it seemed the feature best suited for this purpose.

The symbols plotted on the maps attempt to represent in a diagramatical
manner the arrangement of the foundation elements, as
these would appear in cross section, within the coil of the basket.
These arrangements are: triangular (represented by a triangle);
vertical (represented by a rectangle); circular (represented by a
circle). Irregular foundation types will be explained by footnotes
in the tables. Within the foundation symbols, the nature of the
foundation elements is represented as follows: (1) stippling for a
bundle of pliable fibers or grass; (2) black dot for rod; (3) half
circle for split-rod; (4) horizontal bar for slat; (5) small open circle
for reed; (6) hatched circle for single-rod foundation; (7) two concentric
circles for "bird's nest" coiling. Hence, map 7, 30 represents
two-rod-and-bundle triangular foundation; map 7, 35A, two-rod-and-bundle
vertical foundation; map 7, 15, "bird's nest" coiling;
map 7, 10, bundle foundation; map 10, 19A, rod surrounded by bundle
foundation. The numbers on the map are to be referred to the accompanying
tables.

The tables list the character of the stitches, the surface worked,
and the direction in which the work proceeds. A blank space indicates
a lack of data; a question mark, the uncertainty of the data.

Conclusions.—Some data on coiled basketry distributions in the
Southwest have now been presented, and there remains but to draw
such conclusions from the resulting distributions as seem defensible
in view of our present knowledge. Although the data on which our
classifications have been based are not completely above question, the
data seem to permit no alternate method of treatment. Granted,
provisionally, therefore, that the techniques which we have described
and classified are comparable, we may proceed with our conclusions.

Let us first examine the special aspects of the distributions. Weltfish
has concluded that coiled basketry areas exist today, and are
observable also for the past. She has summarized these and given

the characteristics of each.[299] In map 7, an attempt has been made
to bring out this point by means of lines drawn around the modern
areas as we now conceive them. These lines are intended simply as
guides in interpretation. They are not to be considered boundaries in
the strict sense, and obviously do not exist in literal reality.

TABLE 6—KEY TO MAP 7

                                                                                                         

Number	Group	References	Stitches	Surface	Direction
1	Maidu	Weltfish, G., 1930a, p. 465	Non-interlocking; split on concave	Convex	Left
2A	Washo	Weltfish, G., 1930a, p. 465	Non-interlocking; split on concave	Convex	Left
2B	Washo	Weltfish, G., 1930a, p. 465	Non-interlocking; split on concave	Convex	Left
3A	Miwok	Weltfish, G., 1930a, p. 465	Interlocking	Convex	Left
3B	Miwok	Weltfish, G., 1930a, p. 465	Interlocking	Convex	Left
4	Owens Valley Paiute	Steward, J., 1933b, pp. 270-272	Non-interlocking	Convex	Left(?)
5A	Mono	Gifford E. W. 1932a, p. 27; Weltfish, G., communication	Non-interlocking	Either	Right
5B	Mono	Gifford E. W. 1932a, p. 27; Weltfish, G., communication	Non-interlocking	Either	Right
6	Yokuts	Weltfish, G., 1930a, p. 467	Non-interlocking	Either	Right
7	Chemehuevi[300]	Weltfish, G., 1930a, pp. 466-468	Non-interlocking	Either	Either
8	Kawaiisu	Weltfish, G., 1930a, p. 468	Non-interlocking	Either	Right
9A	Chumash	Kroeber, A. L., 1925, p. 560
9B	Chumash	Weltfish, G., 1930a, pp. 465-466	Non-interlocking	Either	Right
10	Luiseño	Sparkman, P. S., 1908, pp. 204-205		Concave
11A	Cahuilla	Kroeber, A. L., 1908, pp. 41-51
11B	Cahuilla	Kroeber, A. L., 1908, pp. 41-51
12	Southern Diegueño	Spier, L., 1923, p. 347
13	Kamia	Gifford, E. W., 1931, p. 39
14A	Yuma	Forde, C. D., 1931, pp. 124-125
14B	Yuma	Forde, C. D., 1931, pp. 124-125
14C	Yuma	Forde, C. D., 1931, pp. 124-125
15	Cocopa	Gifford, E. W., 1933b, p. 270
16	Seri	Kroeber, 1931, pp. 59-60		Convex	Left
		McGee, W. J., 1898, p. 208
17A	Maricopa	Spier, L., 1933, pp. 122-125
17B	Maricopa	Spier, L., 1933, pp. 122-125		Concave	Right
18A	Pima	Kissel, M. L., 1916
18B	Pima	Kissel, M. L., 1916	Non-interlocking	Concave	Left
19A	Papago	Kissel, M. L., 1916
19B	Papago	Kissel, M. L., 1916	Non-interlocking	Concave	Left
20	San Carlos Apache[301]	Roberts, H. H., 1929, pp. 153-163	Non-interlocking	Either	Left
21	Yavapai	Gifford, E. W., 1936, pp. 282-283	Non-interlocking	Concave	Left
22	Walapai	Kroeber, A. L., 1935, p. 79
23	Havasupai	Spier, L., 1928, pp. 133-138	Non-interlocking	Either	Left
24A	Hopi[302]	Jeancon, J. A., and Douglas, F. H., 1931	Interlocking	Either	Left
		Colton, M. R. F., 1931, p. 6
24B	Hopi	Weltfish, G., 1932a, p. 34	Non-interlocking	Convex	Left
25	Kaibab Paiute	Weltfish, G., 1930a, p. 470	Split on concave	Convex	Left
26	Moapa Paiute	Weltfish, G., 1930a, pp. 469-470	Interlocking	Convex	Left
27A	Northern Paiute[303]	Weltfish, G., 1930a, p. 469	Interlocking or non-interlocking	Convex	Right & left
27B	Northern Paiute	Weltfish, G., 1930a, p. 469	Interlocking or non-interlocking	Convex	Right & left
28	Ute	Weltfish, G., 1930a, p. 470	Split on concave	Convex	Left
29	Navaho	Franciscan Fathers, 1929, pp. 291-297	Non-interlocking	Concave	Left
30	Zuñi	Weltfish, G., 1932a, pp. 36-37	Non-interlocking	Convex	Left
31	Sia	Weltfish, G., 1932a, pp. 35-36	Non-interlocking	Convex	Left
32	Santa Ana	Weltfish, G., 1932a, p. 35	Non-interlocking	Convex	Left
33	Pecos	Kidder, A. V., 1932, pp. 296-297	Non-interlocking
34A	Jicarilla Apache	Weltfish, G., 1930a, p. 470	Non-interlocking	Either	Left
34B	Jicarilla Apache[304]	Douglas, F. H., 1934, p. 55
35A	Mescalero Apache[305]	Weltfish, G., 1932b, p. 115, fn. 25	Non-interlocking	Either	Left
35B	Mescalero Apache	Weltfish, G., 1932b, p. 115, fn. 25	Non-interlocking	Either	Left
35C	Mescalero Apache	Weltfish, G., 1932b, p. 115, fn. 25	Non-interlocking	Either	Left

In order to discuss the interrelations between these areas in both
time and space, it is necessary to summarize briefly the salient characteristics
of each. Turning our attention first to map 7, which represents
present-day conditions, the area of southern California may be
characterized as follows: bundle foundation; non-interlocking stitches;
concave work surface (trays) or convex work surface (ollas); either
work surface (burden baskets); direction of work to the right. An
exception is seen among the Chumash, who also used the three-rod
triangular foundation of reeds.

A second area is comprised by central California, the Basin, and
central Arizona. Throughout, this area is characterized by a three-rod
triangular foundation (with the additional use of a single rod in
the northern portion of the area) and a left direction of the work.
To the north, the stitches tend to be interlocking or split, while to the
south non-interlocking stitches seem characteristic. Again, the convex
work surface is used to the north, while either surface is used to
the south, although in the latter area there is apparently a preference
for the concave work surface. In general, the Basin seems intermediate
and variable, as indeed might be expected, since in this region
the techniques of three areas come together.[306] The Northern Paiute
use both right and left direction of work, although the former may
possibly be due to Salish influence to the north.[307]

Northeastern Arizona and northwestern New Mexico comprise
a third area, the characteristics of which are: two-rod-and-bundle
foundation; non-interlocking stitches; concave surface worked on
trays, convex on ollas; direction of work to the left. It may be objected
that this ware is no longer manufactured by Pueblo groups, but
that it was manufactured in the nineteenth century seems certain.
Specimens were collected in the late nineteenth century by such reliable
men as J. W. Powell and J. Stevenson.[308] In addition, Weltfish
has pointed out that the baskets which were collected among the
Pueblos cannot be attributed to the Navaho, the only living group now
manufacturing this ware, since the designs are non-Navaho.[309] She
also points out the probability that this ware persisted among the

Pueblos into the Pueblo V period, since twilled ring-baskets, an ancient
Pueblo trait, are still manufactured at several villages.[310] The notable
exception in this coiled basketry area is found among the Hopi. Here
a ware is made with a bundle foundation and interlocking stitches.

A fourth, although less well defined area, seems to exist to the
north of the Pueblo area and to the east of the central Basin region.
This area includes the basketry of the Ute, Kaibab Paiute, Shoshoni,
and Bannock.[311] The ware manufactured in this area is characterized
by two-rod or three-rod vertical foundation; stitches split on the nonwork
surface; convex work surface; left direction of work. The
Cheyenne, Arapaho, Kiowa, and Comanche manufactured gambling
baskets which seem to be more or less typical of this area. Since
they are tray-shaped, however, these baskets were worked on the
concave surface.

A fifth area may be found about the Lower Colorado and in
Arizona to the south of the Gila River. In this region, the problem is
rather complex and difficult to understand. The Maricopa and, possibly,
the Yuma manufactured a ware which seems to be affiliated
with that of the southern California coiled basketry area. The Pima,
Papago, and Seri, on the other hand, made a bundle-foundation type

basketry which seems to differ in a technological sense chiefly in the
left direction of the work. Generally speaking, however, this area
seems to be characterized by a conspicuous lack of coiled basketry
among the Yuman-speaking peoples of both southern California and
the Colorado-Gila region in general. A positive characteristic of the
area is the manufacture of "bird's nest" coiled ware. This latter
ware is also made to a certain extent in the southern California basketry
area, where it is found, for example, among the Cahuilla.

Turning our attention now to map 8, we find that similar areas
existed in Pueblo III and IV times, although the data are far less complete.
Whereas the Pueblo area is considerably larger it is again
characterized by the two-rod-and-bundle foundation ware in which the
same aggregate of items occurs. Vertical foundation types, single-rod,
and three-rod triangular forms are to be found along the northern
periphery. Central Arizona is characterized by a three-rod triangular
ware, although a bundle foundation type with a left direction of work
is also found. In southern California a bundle foundation ware
occurs which is said to be typical of that made in the same region
today.[312] In western Texas, a bundle foundation ware occurs which is
typical of this area at an earlier time.

The distributions presented on map 9 represent conditions from
Basket Maker II through Pueblo II times. In the Pueblo plateau
area, the two-rod-and-bundle triangular aggregate is again characteristic.
On the northern periphery, there is considerable variation in
foundation type, and vertical forms occur sporadically. During this
interval, a single rod foundation with interlocking stitches is found
to be very widespread, occuring sporadically from Nevada and northern
Utah to southern New Mexico.[313]

For southern California during this early period, no data are
available. They are, however, rather plentiful from the area comprised
by northern Coahuila and the Big Bend region of Texas. The
consensus of opinion seems to indicate that for the western portion of
the Big Bend region, the characteristic coiled basketry type has a
bundle foundation; stitches split on the non-work surface; concave
work surface; left direction of work. This is also a secondary type
in the Lower Pecos River region, where the characteristic form has
a bundle foundation; interlocking stitches; concave work surface;
left direction of work.[314]

TABLE 7—KEY TO MAP 8

                                                               

Number	Site	Period	Reference	Stitches	Surface	Direction
1	Culberson County, Tex.	1300-1600	Jackson, A. T., 1937, p. 157	Non-interlocking; split on convex	Concave	Left
2	Swarts Ruin, N. M.	Pueblo III	Cosgrove, H. S. and C. B., 1932, pp. 67-68
3	Montezuma Cave, N. M.[315]	Pueblo III?	Weltfish, G., 1932a, p. 28	Non-interlocking	Either	Left
4	Village of the Great Kivas, N. M.	Pueblo III	Roberts, F. H. H., Jr., 1932, p. 134
5	Pecos, N. M.	Pueblo III	Kidder, A. V., 1932, pp. 288-289
6A	Jemez Cave, N. M.	Pueblo III	Alexander, H. G., and Reiter, P., 1935, pp. 49-50
6B	Jemez Cave, N. M.	Pueblo III	Alexander, H. G., and Reiter, P., 1935, pp. 49-50	Non-interlocking
6C	Jemez Cave, N. M.	Pueblo III	Alexander, H. G., and Reiter, P., 1935, pp. 49-50
7A	Mesa Verde, Colo.[316]	Pueblo III	Weltfish, G., 1932a, pp. 16-17	Non-interlocking	Either	Left
7B	Mesa Verde, Colo.[317]	Pueblo III	Weltfish, G., 1932a, pp. 16-17	Non-interlocking	Either	Left
7C	Mesa Verde, Colo.[318]	Pueblo III	Weltfish, G., 1932a, pp. 16-17	Non-interlocking	Either	Left
8	Sandal Cliff House, Colo.	Pueblo III?	Weltfish, G., 1932a, p. 19	Non-interlocking	Concave	Left
9	Moki Canyon, Utah	Pueblo III?	Weltfish, G., 1932a, pp. 7-8	Non-interlocking	Concave	Left
10	Battle Canyon, Utah	Pueblo III?	Weltfish, G., 1932a, p. 15	Non-interlocking	Concave	Left
11	Allan Canyon, Utah	Pueblo III?	Weltfish, G., 1932a, p. 16	Non-interlocking	Concave	Left
12	Betatakin, Ariz.	Pueblo III	Weltfish, G., 1932a, pp. 6-7	Non-interlocking	Concave	Left
13	Canyon del Muerto, Ariz.	Pueblo III	Weltfish, G., 1932a, pp. 9-10	Non-interlocking		Left
14A	Pueblo Bonito, N. M.	Pueblo III	Weltfish, G., 1932a, pp. 21-22
14B	Pueblo Bonito, N. M.	Pueblo III	Weltfish, G., 1932a, pp. 21-22
15	Sikyatki, Ariz.	Pueblo IV	Weltfish, G., 1932a, p. 34	Interlocking	Either	Left
16	Chevlon Ruin, Ariz.	Pueblo IV?	Weltfish, G., 1932a, p. 27	Non-interlocking	Concave	Left
			Fewkes, J. W., 1898b, p. 527
17	Palatki, Ariz.	Pueblo III?	Weltfish, 1932a, p. 25	Non-interlocking
18A	Sierra Ancha, Ariz.	Pueblo IV	Haury, E. W., 1934, pp. 73-76	Non-interlocking
18B	Sierra Ancha, Ariz.	Pueblo IV	Haury, E. W., 1934, pp. 73-76
19A	Casa Grande, Ariz.[319]	Pueblo IV	Weltfish, G., 1932a, pp. 27-28	Non-interlocking	Concave	Left
19B	Casa Grande, Ariz.	Pueblo IV	Fewkes, J. W., 1912, p. 147
20	Las Acequins Ruin, Ariz.[320]	Pueblo III	Haury, E. W., 1935, Mss.	Non-interlocking
21	29 Palms, Calif.	Pueblo IV?	Campbell, E. W. C., 1931, p. 63	Non-interlocking
22A	Mesa House, Nev.[321]	Pueblo III?	Hayden, I., 1930, pp. 59-60	Interlocking
22B	Mesa House, Nev.	Pueblo III?	Hayden, I., 1930, pp. 59-60	Interlocking

The data from Coahuila are too meager, at present, to form a
basis for definite conclusions, but a general affiliation with the Lower
Pecos-Big Bend area may be postulated. Throughout this region, a
bundle foundation type with non-interlocking and non-split stitches
occurs. There is some evidence for the presence of a single-rod foundation
type, but the stitches here do not seem to be of the interlocking
variety. For northern Chihuahua but two examples of coiled basketry
seem as yet to have been described. One specimen differs from the
wares of the Lower Pecos-Big Bend region, chiefly in having a right
direction of work. The other is typical San Juan Basket Maker coiled
ware.

In the western and southern regions of the Great Basin, two
wares are represented at this time: a three-rod triangular type and a
single-rod form with interlocking stitches.

To the north of the Pueblo plateau, we find, at this time, vertical
foundation forms, the single-rod type with interlocking stitches, and
bundle foundation forms.

It will be noted that in the case of all three distributions plotted,
intermediate types appear at the peripheries. Such a condition is to
be expected, and could, indeed, be predicted, although the precise
aggregate of items could not be foreseen.

Map. 9. Coiled Basketry: Distribution During Basket Maker II—Pueblo II

TABLE 8—KEY TO MAP 9

                                                                                                                             

Number	Site	Period	Reference	Stitches	Surface	Direction
1	Upper Rio Fuerte, Mex.[322]	BM?	Zingg, R. M., Mss.	Non-interlocking; non-split	Concave	Right
2A	Western Coahuila, Mex.[323]	BM?	Peabody Museum Collection No. 22774	Stitches split on both surfaces
2B	Western Coahuila, Mex.	BM?	Peabody Museum Collection No. 22841	Stitches split on both surfaces	Concave	Left
2C	Western Coahuila, Mex.	BM?	Peabody Museum Collection No. 22774	Interlocking; split on work surface	Convex	Left
3	Western Coahuila, Mex.	BM?	Peabody Museum Collection No. 22687	Non-interlocking; non-split	Convex	Left
4A	Val Verde County, Tex.	BM?	Pearce, J. E., and Jackson, A. T., 1933, pp. 106-114	Interlocking
			Martin, G. C., 1933, pp. 55-59
4B	Val Verde County, Tex.	BM?	Pearce, J. E., and Jackson, A. T., 1933, pp. 106-114	Non-interlocking; non-split
			Martin, G. C., 1933, pp. 55-59
4C	Val Verde County, Tex.	BM?	Pearce, J. E., and Jackson, A. T., 1933, pp. 106-114	Split
			Martin, G. C., 1933, pp. 55-59
5	Culberson County, Tex.	BM?	Jackson, A. T., 1937, p. 185	Non-interlocking; split on concave surface
6	Brewster County, Tex.	BM?	Weltfish, G., 1932a, p. 33	Non-interlocking; split on non-work surface	Concave	Left
			Setzler, F. M., 1932, p. 136
7A	Brewster County, Tex.	BM III-Pueblo I	Setzler, F. M., 1933, p. 56	Non-interlocking; split on non-work surface	Concave	Left
7B	Brewster County, Tex.	BM III-Pueblo I	Setzler, F. M., 1933, p. 56	Interlocking	Concave	Left
8A	Brewster County, Tex.	BM?	Coffin, E. F., 1932, p. 38	Non-interlocking; non-split?
8B	Brewster County, Tex.	BM?	Coffin, E. R., 1932, p. 38	Split on non-work surface?
9A	Guadalupe, Mts., N. M.	BM II-Pueblo II?	Mera, H. P., 1938a, pp. 50-52	Interlocking
9B	Guadalupe Mts., N. M.	BM II-Pueblo II?	Mera, H. P., 1938a, pp. 50-52	Some interlocking; some split
9C	Guadalupe Mts., N. M.	BM II-Pueblo II?	Mera, H. P., 1938a, pp. 50-52	Some non-interlocking; some split
10A	Guadalupe Mts., N. M.	BM?	Howard, E. B., 1930, p. 197	Non-interlocking?
10B	Guadalupe Mts., N. M.	BM?	Howard, E. B., 1930, p. 197	Interlocking
10C	Guadalupe Mts., N. M.	BM?	Howard, E. B., 1930, p. 197
11	Dona Ana County, N. M.	BM?	Weltfish, G., 1932a, pp. 31-33	Non-interlocking	Concave?	Left
12	Harris Village, N. M.	Pueblo II	Haury, E. W., 1936a, p. 78	Interlocking; nonsplit
13A	Chaco: Bc 51, N. M.	Pueblo II?		Non-interlocking	Concave	Left
13B	Chaco: Bc 51	Pueblo II?		Non-interlocking; surface split on non-work	Concave	Left
14	Piedra District, Colo.	Pueblo I	Weltfish, G., 1932a, p. 21	Non-interlocking	Concave	Left
15A	San Juan County, Utah[324]	BM III	Weltfish, G., 1932a, p. 14	Non-interlocking	Concave	Left
15B	San Juan County, Utah	BM III	Weltfish, G., 1932a, p. 14	Non-interlocking	Concave	Left
15C	San Juan County, Utah	BM III	Weltfish, G., 1932a, p. 14	Non-interlocking	Concave	Left
16A	San Juan County, Utah	BM II	Weltfish, G., 1932a, p. 16	Non-interlocking	Concave	Left
16B	San Juan County, Utah	BM II	Weltfish, G., 1932a, p. 16	Non-interlocking	Concave	Left
17	Moki Canyon, Utah	BM	Burgh, R., 1937, p. 10	Interlocking
18A	Fremont River, Utah	Pueblo II	Morss, N., 1931, pp. 72-74
18B	Fremont River, Utah	Pueblo II	Morss, N., 1931, pp. 72-74
19A	Kane County, Utah	BM	Weltfish, G., 1932a, p. 14	Non-interlocking	Concave	Left
19B	Kane County, Utah	BM	Weltfish, G., 1932a, p. 14	Non-interlocking	Concave	Left
20A	Grand Gulch, Utah	BM	Weltfish, G., 1932a, p. 12	Non-interlocking	Concave	Left
20B	Grand Gulch, Utah	BM	Weltfish, G., 1932a, p. 12	Interlocking
21A	Ashley Creek, Utah[325]	Pueblo II?	Peabody Museum Collection No. A-7942	Non-interlocking; split on non-work surface	Concave	Left
21B	Ashley Creek, Utah	Pueblo II?	Peabody Museum Collection No. A-7949	Interlocking; rod not split
21C	Ashley Creek, Utah	Pueblo II?	Peabody Museum Collection No. A-7943	Interlocking; split on non-working surface	Convex	Left
22A	Promontory Point, Utah	BM	Steward, J. H., 1937, pp. 33-35	Interlocking
22B	Promontory Point, Utah	BM?	Steward, J. H., 1937, pp. 33-35	Split
22C	Promontory Point, Utah	BM?	Steward, J. H., 1937, pp. 33-35	Non-interlocking
23A	Western Utah	Pueblo I?	Steward, J. H., 1936, p. 55	Non-interlocking
23B	Western Utah	Pueblo I?	Steward, J. H., 1936, p. 55
24A	Lovelock Cave, Nev.	?	Weltfish, G., 1932b, pp. 110-111	Non-interlocking or split on non-work surface	Convex	Left
24B	Lovelock Cave, Nev.	?	Weltfish, G., 1932b, pp. 110-111	Interlocking; nonsplit	Convex	Left
24C	Lovelock Cave, Nev.	?	Weltfish, G., 1932b, pp. 110-111	Interlocking; rod split	Convex	Left
24D	Lovelock Cave, Nev.	?	Weltfish, G., 1932b, pp. 110-111	Interlocking; rod not split	Convex	Left
25	Paiute Cave, Nev.	?	Harrington, M. R., 1930, p. 118
26	Chinlee Valley, N. M.	BM II	Guernsey, S. J., 1931, p. 70	Non-interlocking	Concave?	Left
27	Segi Canyon, Ariz.	BM III	Guernsey, S. J., 1931, p. 78	Non-interlocking	Concave?	Left
28	Segi Canyon, Ariz.	Pueblo I	Lockett, H. C., 1934, p. 13	Non-interlocking	Concave?	Left
			Guernsey, S. J., 1931, p. 95
29A	Canyon de Chelley, Ariz.	BM	Weltfish, G. 1932a, p. 10	Interlocking
29B	Canyon de Chelley, Ariz.	BM	Weltfish, G. 1932a, p. 10	Non-interlocking	Concave?	Left
30	Canyon del Muerto, Ariz.	BM III	Weltfish, G., 1932a, pp. 8-9	Non-interlocking	Either	Left
31	San Francisco Mts., Ariz.	Pueblo II	Bartlett, K., 1934, pp. 44-45	Non-interlocking	Concave	Left

To point out but a few examples, the Chemehuevi (map 7, 7), who
are located midway between an area characterized by right direction
of work on the one hand and left direction on the other employ either
direction. The Western Mono (map 7, 5), similarly located between
two areas, employ foundations typical of both. Again, it seems possible
that two-rod-and-bundle triangular ware manufactured in the Guadalupe
Mountains (map 9, 9A) in early times had interlocking stitches
due to the proximity to the interlocking wares of the Lower Pecos
River region.

Turning our attention now to the temporal aspects of the distributions,
there seems to be a remarkable continuity of coiled basketry
techniques throughout the development of the Southwestern cultures.
Weltfish has treated this problem at some length, and has shown that
"the parallels for the prehistoric technical types in all cases are to be
found in modern areas closely contiguous to the ancient sites."[326] In
this paper she has shown that the prototype of the northern vertical
foundation wares of the Shoshoni, Ute, and Plains tribes is to be found
in the work of the Ozark Bluff-Dwellers. She also concludes that
certain technological forms of coiled ware from Lovelock Cave are
represented today among the Maidu, Washo, and Pomo in California.
Perhaps the continuity of two-rod-and-bundle triangular ware in the
Pueblo area is most striking of all, since it has persisted with little
change from Basket Maker II times to the present day.

We are now confronted with a problem most vital to all who
attempt to reconstruct the history of human culture: the question of
cultural stability. Weltfish correctly states that the question of "the
stability . . . of technical traits" in an important problem in an historical
study of coiled basketry.[327] The precise problem, however, remains
undefined, and the word "stability" is conspicuously lacking in her
conclusions. Instead, she implies, as previously stated, that there is a
marked continuity in basketry techniques which accounts for the
similarities between ancient and modern wares. And stability, it would
seem, is a necessary prerequisite to the continuity, whether the latter
concept be expressed or implicit. What, therefore, can be meant in
this instance by "stability"? Does it suffice for all purposes simply to
say that "stability" is the state or quality of being stable? It is in
this general sense that Boas has applied this concept to culture when
he contrasts "stability" with "change."[328]

In particular instances, however, we are interested in the question
of precisely what in a culture makes for stability or instability.
The issue, it seems, is that of the definition of a problem, rather than
of a concept. Mead has treated at some length the question of the

stability of "elements" within "complexes" in Polynesia.[329] But this
does not help us materially with the problem at hand, since we are
concerned with but a single trait—coiled basketry—and not with trait
complexes.[330]

What problems, therefore, are involved in the question of the
stability or instability of Southwestern coiled basketry techniques?
Weltfish's conclusions suggest that one problem is the question of the
stability of a trait within an area. That this problem may be illuminated
by the Southwestern data she has demonstrated with adequacy.
The very application of the Southwestern basketry data to this particular
problem, however, involves yet another problem: the question
of the stability of a given trait within a specified culture. That these
two problems are intimately related seems obvious. The evidence furnished
by Southwestern archaeology indicates that from time to time
new increments of population, as well as new traits and trait complexes,
entered this area from the outside. In order for the stability of
a given trait to be maintained within a specified area in a case where
the population or culture is wholly or partially altered, it is necessary
either that the trait be taken over by the new population or culture, or
that it be perpetuated by the older inhabitants of the area. Today, for
example, the Navaho represent the only Southwestern group to manufacture
two-rod-and-bundle coiled basketry.[331] There seems to be little
room for doubt that the Navaho took over this trait from the earlier
inhabitants of the Pueblo plateau after they had arrived in approximately
their present position. The trait seems not to have been reported
outside of the Southwest, and the probability that the Navaho
invented this basketry type in an area in which it had been manufactured
since early times seems unlikely. The stability, therefore, of
two-rod-and-bundle basketry within the Southwestern area has been
maintained in this single instance (e. g., Navaho-Pueblo) by the fact
that the Navaho adopted it; the instability of the trait in the nineteenth
century Pueblo culture is evident, since the manufacture of
this basketry type among the Pueblos has ceased in recent times
altogether.

Again, basic to both of the problems touched upon above, is the
question of the stability of items within a trait. It seems evident that,
unless the aggregate of items composing a trait are stable, the trait
itself will be in a constant state of flux. Setzler has put forward this
question in connection with differences in foundation in otherwise similar

baskets.[332] He points out that a bundle foundation basket with
interlocking stitches is characteristic of the Lower Pecos River region
of Texas. These specimens resemble a form found as intrusive among
typical Basket Maker examples. This intrusive ware, however, is characterized
by a single-rod foundation. How, then, may we regard the
instability or stability of the aggregate of items which compose a
trait? This appears to be an issue which underlies the problems of
independent invention and diffusion. To this most basic question there
is, at present, no wholly adequate answer. We may simply weigh the
evidence in each case with the pious hope that our judgment will later
prove correct. Since, however, cultural change normally progresses at
a pace which is perceptible only over larger time intervals, it seems
probable that much light could be shed on the problem of the stability
of culture as a whole by examining the stability of items within traits.

Let us turn now to certain historical problems suggested by the
distributions. Since the data are of a fragmentary nature, and because
we are dealing here with but a single trait, we must resort to speculation
pure and simple. But speculation is useful, for granting that we
are speculating, certain intriguing possibilities present themselves.

Let us first examine the possibilities suggested by the distribution
of bundle foundation wares. On a general Basket Maker-like horizon we
find bundle foundation baskets; non-interlocking, non-split stitches;
usually with a concave work surface; worked to the left in the northern
Coahuila-Big Bend region. In Pueblo IV times, these are found
in southern Arizona at Casa Grande and in the Sierra Anchas. Today,
these are made by the Pima and Papago. A similar basket, but with a
right direction of work, has been found on a Basket Maker-like horizon
in northern Chihuahua. Again, in Pueblo IV times, there are indications
that this ware was manufactured in southern California. Present
day distributions reveal that it is made by the Seri, the Maricopa, and
the southern California groups. Does the Basket Maker-like ware
from northern Coahuila and the Big Bend represent the prototype of
modern Pima-Papago ware? Are the beginnings of modern southern
California basketry to be seen in northern Chihuahua on a Basket
Maker-like horizon? One wonders also at the relationship between these
two styles, technically similar except for the direction of work.

In the case of two-rod-and-bundle triangular ware, Weltfish has
pointed out that the Lower Rio Grande forms the southeastern boundary
of this style.[333] It is, indeed, interesting that, to date, with the
exception of this extension along the Lower Rio Grande and the example
from Chihuahua, no two-rod-and-bundle triangular basketry
has been reported to the south of the plateau lands occupied by the
Anasazi peoples. Does basketry also reflect the cultural hiatus which

seems to exist between the Anasazi and the Hohokam? Will later
researches indicate that a bundle foundation basketry style was typical
of the latter culture?[334] It is unfortunate that no basketry remains
on a Basket Maker horizon have been produced from southern Arizona.

The single-rod basketry style with interlocking stitches presents
a puzzling problem. Weltfish considers it to be intrusive in the Anasazi
area, where it occurs sporadically in all periods with the exception of
present times.[335] Today it is found in the Basin and in central California.
For all periods, as is also true in the case of the vertical foundation
styles, this ware has essentially a northern distribution. Is it
an ancient type? With what culture or cultures may it be identified?

"Bird's nest" coiled storage baskets seem first to appear—as far
as present knowledge goes—on a Pueblo IV horizon in the Sierra Anchas,
and may possibly also have been utilized at Casa Grande. As has
been stated, the specimen from Bc 51 is so doubtful that it seems unwise
to consider it at this time. In modern times this storage ware has
been manufactured by the Pima and Papago, the Yuman speaking
peoples of the Lower Colorado and Gila Rivers and of southern California,
and some of the southern California Shoshonean groups. This
ware, today, seems to be centralized among the Yuman speaking
peoples. All this suggests several interesting possibilities. There is
some evidence furnished, both by the bundle coiled basketry and the
"bird's nest" storage ware, that may possibly indicate some relationship
between the Pueblo IV inhabitants of the Sierra Anchas and the
modern Pima and Papago. Is there some relationship between the
Hohokam peoples and the southern Yumans?[336] Will later excavations
in Hohokam sites reveal the presence of "bird's nest" coiling?

Three-rod triangular ware seems to have been concentrated in
the Great Basin since early times. Although we have no evidence for
Basket Maker times in central Arizona, this basketry style is represented
in this region in Pueblo IV times. Today it is manufactured in
central and northwestern Arizona by the Western Apache and the
Plateau Yumans respectively.[337] Did the Western Apache enter Arizona
in Pueblo IV times by way of the Great Basin, bringing this style
of basketry with them? Gladwin believes that they appeared in the
region inhabited by the Salado peoples early in the fourteenth century.[338]

Weltfish has pointed out that this style is independent of the
typical Anasazi ware; that its distribution is more southern; that it
tends to be identified with a later horizon.[339]

In summing up, it may be stated that coiled basketry seems to
have earned its place as a trait of some comparative significance. It
reflects cultural relationships and differences, and is, therefore, a
useful tool in reconstructing the prehistory of aboriginal America.
Finally, it serves to shed its own light on some interesting theoretical
problems which are applicable to anthropology as a whole.