116 BUREAU OF AMERICAN’ ETHNOLOGY [BULL. 65 
red and natural color, each four threads wide; the latter (A—1709, 
Ruin 9) three-thread stripes of red, brown, and natural color. The 
red and brown dyes seem to have rotted the cotton to some extent, 
as the colored strands are considerably more fragile than the others. 
All varieties of cotton cloth were evidently made on the hand 
loom, an apparatus probably similar to, though lighter than, that in 
use by the Navaho to-day. The selvage or edge-binding of these 
ancient fabrics is like that of the better class of Navaho blankets; 
that is to say, the threads are held in place by a pair of heavy, three- 
strand strings intertwined as shown in figure 44. For the warp this 
fastening was probably done at the time of the “ set up ” outside the 
‘oom, according to the Navaho practice. The only difference be- 
tween the Cliff-house method and that of the Navaho is that the 
Navaho warp is held by three strings braided together, rather than 
by two twisted ones, as in the majority of the ancient cotton fabrics. 
There is one specimen in our collection, however, in which the warp 
cords are held, as in Navaho blankets, by a 
braid; a further elaboration is introduced 
in the form of a row of twined weaving 
in red threads between the selvage and 
the body of the cloth (A-1426, Ruin 6). 
The weft, of course, was worked into its 
“edge cord” during the weaving, .prob- 
ably again in the same way as is done by 
Hike: ic of cotton the Navaho, namely, by attaching two 
strings to the lower beam of the loom on 
each side just outside the warps and twisting them together as the 
web grew, including one turn of the weft in each twist (fig. 44). The 
selvages produced are neat, firm, and considerably heavier than the 
body of the fabric; they are so much stronger than it that strips of 
the selvage were usually left after the rest of the cloth was worn out, 
and these were often reused for sandal ties, pot harness, and other 
secondary purposes. When cloth was cut across for any reason, it 
was kept from raveling by rolling up the edge and sewing it over 
and over with stout thread. 
As was stated above, the rags recovered do not give us much idea 
of the original size of the pieces of cloth produced in the looms. The 
largest fragment in our collection is 19 inches long. The series of 
loom attachments observed in the kivas, where presumably these 
fabrics were produced, average about 4 feet in length. 
Other fabrics—Twined woven textile is represented in the col- 
lection by two small specimens. One (A-1398, Ruin 5) is perhaps 
Matthews, 1884. 
