ba BUREAU OF AMERICAN ETHNOLOGY [ BULL. 65 
toward the edge and locked under the elements following them in 
the series (fig. 42). These are the common methods. We have two 
specimens showing variants of the same styles: one, A-1702, Ruin 
7, with a double border made by bringing the elements back from 
the edge, past the junction of body and selvage to a second edge, 
and then back again to the junction where they end as usual. The 
other (A-1515, Ruin 7): the odd elements in this case, instead of 
being merely cut off, are carried, before ending, a few turns to the 
right, thus supplementing the twined binders and giving greater 
strength to the fabric (pl. 44, d). 
These rush mats all appear to have been square or rectangular 
with rounded corners. Their size was fairly uniform, having been 
Kore 
oii 6 
pt kh ed 
a ee 
ee Semel 
Fig. 42.—Rush-mat selvage. 
limited by the length of the component rushes. Our nearly com- 
plete specimen (Ruin 2) measures 3 feet 2 inches in one direction 
and seems to have been nearly square. A similar mat from Spruce- 
tree House, Mesa Verde, figured by Nordenskidld? is square and 
measures 1.2 by 1.2 meters. Cummings? gives a description of one 
that is 3 feet by 2 feet 6 inches. 
Twistep ELemMEeNts, STRINGS AND THREAD 
Cotton.—Cotton string was used for practically nothing but the 
weaving of cloth and the sewing of garments. The finest of it is 
single strand, about one-fiftieth inch in diameter. This was employed 
in the better grades of cloth, but was not strong enough for sewing, 
11893, pl. xlviii, 4. 21910, p. 35. 
