tree, so that the forest is shrinking rapidly, rotting stumps attesting its former limits. Passing through the forest, we emerged suddenly on open country, and, shortly after, passed a ruined tower, which, if I understood the Tartar rightly, marks the point where the Suanetian sheep and cattle raiders used, in the old days, to be held in check; presumably, therefore, it marks the point above which sheep and cattle were not, in those old days, ever pastured. Below this point one may seek in vain, not merely for a tree, but even for the smallest bush. As I walked down the valley, I could not resist the conclusion that the presence or absence of forest in the Bashil valley had been determined by the presence or absence of sheep and goats. And though I am doubtless generalising on very insufficient data, I am much inclined to attribute the extraordinary contrast between the treelessness of some of the northern valleys and the dense forests of the southern, less to climatic differences than to the form in which the wealth of their respective inhabitants exists. In the one case oxen, horses, sheep, and goats; in the other, well-tilled and neatly-fenced fields and orchards. Though at first sight it appears difficult to believe that sheep and goats can destroy the forest over great stretches of country, a careful examination of the upper Bashil Su shows that the cause is sufficient to produce a continuous
22
