258 PROCEEDINGS OF THE AMERICAN ACADEMY. 



ants of the second seed would at once be subjected to the same modifying 

 influences that had affected those of the first. 



It is thus clear' that the insular form, having once started its diver- 

 gence from the continental type, would be likely to differentiate itself 

 more and more and not be much affected by the occasional arrival of 

 scattered and isolated seeds from the continent. Furthermore, if seeds 

 of the Chatham form were by any chance carried to other islands of the 

 archipelago, farther specialized races would be almost certain to arise 

 bearing much the same relation to that of Chatham that it bore to the 

 continental form, and quite as little affected by subsequent seedings. 

 Thus unchecked, the races would have every opportunity to develop into 

 more and more highly differentiated forms, varieties, and ultimately into 

 well-marked species characteristic of particular islands. 



This may all seem purely hypothetical, but it will be seen that the 

 conclusions rest upon only two very natural premises, namely, that seed- 

 transference between the mainland and the islands or between the islands 

 themselves, does not occur in the case of particular plants oftener, on the 

 average, than once in several years, and in the second place that plants 

 have multiplied on the islands as rapidly as they have frequently been 

 observed to multiply elsewhere. When these two not unreasonable 

 postulates are admitted, it is clearly no harder to account for the exist- 

 ence of a " harmonic" flora on islands of emergence than of subsidence. 

 Indeed, it is what is to be expected in an archipelago of either kind 

 where seed-transference is relatively rare. That this condition obtains 

 in the case of the Galapagos is clearly shown by the existing diversity in 

 the floras of the different islands, — a condition which could not continue 

 if seed-transference were very common between the islands. The fact 

 that it is not more frequent is perhaps sufficiently explained by the arid 

 and sterile shores, which would certainly offer to most seeds washed 

 thither by oceanic currents an exceedingly poor chance of surviving. 



The great existing difference between the Galapageian flora as a 

 whole and that of tropical America is doubtless due not only to the 

 differentiation of the insular forms, but also and perhaps in a consider- 

 able degree to changes which have been simultaneously going on in the 

 continental flora itself. Thus the ancestral forms of many Galapageian 

 2~>lants, for instance of the above mentioned Euphorbia viminea (which 

 might reasonably be sought in the deserts of Peru), have probably 

 failed to persist at all upon the mainland. The fact that the florulae of 

 the several islands are on the whole much more like each other than 

 any one of them is like any part of the continental flora is perhaps 



