Page:Isaiah Bowman - Desert Trails of Atacama (1924).pdf/343

 of the stone work is distinctive. However, to reach the con- clusion that the climate has changed it is necessary to deter- mine by accurate means the actual water resources today, and this has not yet been done.

More circumstantial is the argument in favor of a change of rainfall in the case of the Arroyo de Sayate in the Puna de Jujuy. The andenes here appear to be arranged in a manner to facilitate irrigation by canals supplied by the Sayate, but no trace of a feeding canal has been found, and it is believed that a canal was never employed. Yet the rainfall today is not sufficient to permit culture without arti- ficial irrigation. Watering by hand does not seem possible, for it would require a population much larger than could be con- tained in the little ravine in prehistoric times or than is indi- cated by the number of skeletons in the graves. The traces of water conduits encountered on the terraces of Sayate are interpreted as indicating the practical beginnings of a system of conserving the natural rainfall from terrace to terrace with- out permitting it to run off violently as it would do if left to take its natural course to the floor of the ravine.

On the other hand, cultivation, plus the greater water- holding capacity of the flat and walled terraces, is enough to account for the conditions described without invoking a change of climate. From the large number of remains of maize in the graves of Sayate it is concluded that maize was the principal food of the valley and that without doubt it was the principal plant cultivated on the terraces, for to find its remains in such great abundance would seem to argue against its being carried by the pre-Spanish Indians from the lower ground in the valleys of Salta and of Jujuy or from the Desert of Atacama. The other food plants that were grown in the Punain pre-Hispanic times, namely, potatoes, coca, quinoa, and the like, had a secondary importance, and their cultivation was carried on without andenes. Periods of rest of three, four, five, or even seven years between periods of culture, which in turn endure for three or four years at a time, are known to every student of Central Andean culture; and this may well account