576 AKCHEOLOGICAL RESEARCHES IN ARGENTINA AND BOLIVIA. 



having been provided with handles. I made the discovery, how- 

 ever, very soon, that the implement had had a handle, for on several 

 of the specimens I obtained there were plain marks of friction against 

 some hard object on one of the surfaces of the projecting part, 

 which itself resembled a handle. On some of the axes is j^lairdy 

 to be seen that the Avooden handle has produced a shallow, level 

 furrow^ in the stone (pi. viii, fig. 4). Subsecpiently I found in 

 another grave a handle bent at an acute angle and shaped at one 

 end for fastening the ax more securely to it. I placed one of the 

 axes before described in contiguity to this handle, and thereby con- 

 vinced myself that the worn surface on the ax very closely corre- 

 sponded to the level surface of the handle; the two par-ts were prob- 

 ably united together by means of something wound around them 

 or b}^ being incased in leather. It is possible that resin or some 

 other adhesive substance may have been employed to prevent the 

 ax gaping; the fact of its having been insecurely attached is, how- 

 ever, amply proved by the considerable amount of wear shown on 

 the handle. 



The fragile material of which the ax is made renders it an impos- 

 sibility that it can have been used for dealing with any hard sub- 

 stance; on the other hand, this ax would be an admirable implement 

 for peeling oif the soft exterior of the pillar cactuses, Avhen the hard 

 internal stem of the plant is to be got at. With the trifling exception 

 of small bushes and the Polylepis racemosa^ which resembles a shrub, 

 and, moreover, occurs veiy sparsely, there do not exist any trees but 

 the cactuses in the whole of the Puna that yield wood for manufac- 

 turing purposes or for fuel. These cactuses (jdI. ii, fig. 1) attain a 

 respectable growth. The adaptability of the wood of the cactus for 

 manufactures is shown both by the loom (pi. iii, fig. 1) made of cac- 

 tus wood and by the door (pi. ii, fig. 2) of the same material. 



About 8 kilometers northwest of Casabindo, in a narrow valley, 

 there are numbers of remains of round stone huts. They are, indeed, 

 so numerous that we may well assume that a population of several 

 thousands dwelt here. Stone huts of a similar appearance are still 

 used in the Puna, though only sparsely. The illustration (pi. in, 

 fig. 2) shows a modern hut of the kind. The roof timbers are of 

 cactus w^ood, covere(,l with sticks and straw. The valley leading from 

 Casabindo to the ruined city is hemmed in by fairly steep slopes, 

 covered almost throughout with stone terraces of from 1 to 2 meters 

 in height antl of similar a])[)earance to the irrigation terraces of the 

 Inca period, which are so common in Peru. Plate iv, figure 1, repre- 

 sents a terraced mountain slope at Casabindo. Plate iv, figure 2, shows 

 a small gi-otto, the entrance to which has been blocked up with a 

 stone wall. Grottoes of that kind are of very general occurrence in 

 the vicinity of the irrigation terraces. At the foot of the wall there 



