The question whether central Asia is becoming drier has been discussed for many years. Kropotkin held that there is a gradual desiccation irrespective of periods of greater or less rainfall. Ellsworth Huntington has written repeatedly on this problem and in The Pulse of Asia (19°7); Palestine and its Transformation (1911) and other writings has accumulated evidence to show that the fluctuations in the prosperity of ancient civilizations are ulti mately due to climatic conditions. Thus in the Nearer East there was progress and peace, save for dynastic wars, when the rain fall was sufficient, but when it began to diminish the peoples of the poor lands and deserts were set in motion. For example, after 1700 B.C. we hear of invasions from the desert and elsewhere. Kassites invaded Media, Elam and Babylonia, and Egypt was swamped by the Hyksos, who brought their families, flocks and herds, and introduced the horse to Egypt, and we know that famine was the reason for the migration of Hebrews into Egypt. For zoo years after the expulsion of the Hyksos Egypt again prospered. About 1400 B.C. the Aramaeans invaded Syria from the south and the Hittites from the north. The second millennium was a period of depression in the ancient East; from Arabia, Central Asia and Europe new nations immigrated into the lands of ancient civilization. This universality was probably due to widespread adverse economic conditions which were beyond human control. This sort of thing has happened time after time, and analogous incidents have doubtless occurred elsewhere among peoples beyond the ken of history. The theory that the increasing aridity of parts of Central Asia is the direct result of climatic change has been disputed by several authorities (cf. A. Stein, Geog. Journ. lxv., pp. 1925; R. C. F. Schomberg, G.J., lxxii., p. 357, 1928).
Elaborate hypotheses have been promulgated about sunken continents. There is a considerable literature about "Atlantis" and Macmillan Brown adduces evidence satisfactory to himself in favour of a sunken continent or, at all events, of great land areas in the Pacific. Naturally the existence of great land masses in these oceans in human times would have had a considerable effect on migrations if, as is claimed, they belong to human times, but, on the other hand, existing distributions can, on the whole, be satisfactorily explained by present conditions.
We do know, however, that great tracts of land have been alternately submerged and elevated within relatively recent times, for example, men hunted game over what is now the North sea, and coastal areas are constantly liable to fluctuations of level. There is thus ample evidence for compulsory migrations or shift ings of populations due to causes entirely beyond the control of man.
The direction in which all migrations take place is conditioned by geographical factors. We may take it as axiomatic that peo ple will usually seek conditions similar to those to which they have become accustomed. This is, on the whole, true for modern emigrations and colonizations. There is a distinct tendency for
people to emigrate to similar latitudes or to countries with a climate which resembles that of the home country. Northern Europeans seem to experience greater difficulty in acclimatiza tion in tropical countries than do south Europeans. The subject is a difficult one as it is complicated by habits of life and by tropical diseases, as well as by the effects of heat and moisture. North ern Europeans by taking suitable precautions can live long and healthy lives in the tropics, but it has yet to be shown that their de scendants can live the whole of their lives there and indefinitely continue the race. The only definition of a "white man's coun try" is one which includes the latter proviso. A transplanted population which has to be continually replenished by fresh im migrants and has to be diminished by the return of the children to the home country cannot be described as a true migration, it is merely an occupation of the country. In primitive times, that is, before the period' of the means of efficient overseas transport, migratory movements of a people were perforce usually very slow and thus it was possible for acclimatization to take place to some extent en route. But this may not have been always opera tive, as there seem to be indications that the Amerinds of Guiana, for example, are not yet thoroughly acclimatized to their tropical habitat.
We may take it as a general rule that voluntary migratory movement in early times took place only along lines of least re sistance. Definite barriers which could not be crossed, or only to a small degree, were ice and snow, oceans, inland seas, lofty, inhospitable mountain ranges, dense forests, jungle-clad, tropical mountain regions, swamps, very poor steppes and deserts. Un der exceptional circumstance or through dire necessity these might be partially negotiated by small bands, but normally they would prove to be definite obstacles to the migration of whole communities. On the other hand, grassy plains and plateaux, savanna or bush veldt, loess lands, etc., afforded highways for migration. On the whole, mountain ranges running east and west do form appreciable barriers, those running from north to south may do so for people wanting to travel east or west, but for those travelling from north to south, or vice versa, they may afford toilsome but practicable bridges along which people accustomed to a temperate climate might cross unhealthy tropical regions, all the time keeping to healthy conditions. It is, however, unwise to suggest actual routes for prehistoric migrations until geographers and others have provided maps which give the orography, climatic conditions and flora; for these determine the lines of least resistance or, at all events, the routes along which difficulties could be readily overcome, and at the same time the maps would supply information concerning routes which would be quite impracticable.