[20] I say "_large_ areas" for the sake of argument; but the same correlation between distribution and affinity extends likewise to _small_ areas where only _small_ differences of affinity are concerned. Thus, for instance, speaking of smaller areas, Moritz Wagner says:--"The broader and more rapid the river, the higher and more regular the mountain-chain, the calmer and more extensive the sea, the more considerable, as a general rule, will be the taxonomic separation between the populations"; and he shows that, in correlation with such differences in the _degrees_ of separation, are the _degrees_ of diversification--i. e., the _numbers_ of species, and even of varieties, which these topographical barriers determine. I cannot well conceive a _reductio ad absurdum_ more complete than this. But, having now presented these most general facts of geographical distribution in their relation to the issue before us, we may next proceed to consider a few illustrations of them in detail, for in this way I think that their overwhelming weight may become yet more abundantly apparent. * * * * * It will assist us in dealing with these detailed illustrations if we begin by considering the means of dispersal of organisms from one place to another. Of course the most ordinary means is that of continuous wandering, or emigration; but where geographical barriers of any kind have to be surmounted, organisms may only be able to pass them by more exceptional and accidental means. The principal barriers of a geographical kind are oceans, rivers, mountain-chains, and desert-tracts, in the case of terrestrial organisms; and, in the case of aquatic organisms, the presence of land. But it is to be observed that, as regards marine organisms, any considerable difference in the temperature of the water may constitute a barrier as effectual as the presence of land; and also that, in the case of all shallow-water faunas, a tract of deep ocean constitutes almost as complete a barrier as it does to terrestrial faunas.