At about the same period when icebergs were transporting boulders in N. America as far as 36 deg. south, where the cotton tree now grows in South America, in latitude 42 deg. (where the land is now clothed with forests having an almost tropical aspect with the trees bearing epiphytes and intertwined with canes), the same ice action was going on; is it not then in some degree probable that at this period the whole tropical parts of the two Americas possessed{370} (as Falconer asserts that India did) a more temperate climate? In this case the Alpine plants of the long chain of the Cordillera would have descended much lower and there would have been a broad high-road{371} connecting those parts of North and South America which were then frigid. As the present climate supervened, the plants occupying the districts which now are become in both hemispheres temperate and even semi-tropical must have been driven to the Arctic and Antarctic{372} regions; and only a few of the loftiest points of the Cordillera can have retained their former connecting flora. The transverse chain of Chiquitos might perhaps in a similar manner during the ice-action period have served as a connecting road (though a broken one) for Alpine plants to become dispersed from the Cordillera to the highlands of Brazil. It may be observed that some (though not strong) reasons can be assigned for believing that at about this same period the two Americas were not so thoroughly divided as they now are by the West Indies and tableland of Mexico. I will only further remark that the present most singularly close similarity in the vegetation of the lowlands of Kerguelen's Land{373} and of Tierra del Fuego (Hooker), though so far apart, may perhaps be explained by the dissemination of seeds during this same cold period, by means of icebergs, as before alluded to{374}. {370} Opposite to this passage, in the margin, the author has written:--"too hypothetical." {371} The Cordillera is described as supplying a great line of invasion in the _Origin_, Ed. i. p. 378. {372} This is an approximation to the author's views on trans-tropical migration (_Origin_, Ed. i. pp. 376-8). See Thiselton-Dyer's interesting discussion in _Darwin and Modern Science_, p. 304. {373} See Hooker's _Lecture on Insular Floras_ in the _Gardeners' Chronicle_, Jan. 1867. {374} <Note by the author.> Similarity of flora of coral islands easily explained. Finally, I think we may safely grant from the foregoing facts and reasoning that the anomalous similarity in the vegetation of certain very distant mountain-summits is not in truth opposed to the conclusion of the intimate relation subsisting between proximity in space (in accordance with the means of transport in each class) and the degree of affinity of the inhabitants of any two countries. In the case of several quite isolated mountains, we have seen that the general law holds good.