No. 615] SHORTER ARTICLES AND DISCUSSION 271 



of fertilization. Many of the genera which are themselves not 

 endemic on any island are nevertheless represented there now 

 only by endemic species. In such cases it seems clear that the 

 first representative of the genus to invade the island has since 

 disappeared there entirely and been replaced by local species. 

 Table II gives the percentage of such genera (1101 endemic but 

 represented only by endemic species) for each of the three plant 

 types which we have discussed and for all the islands. 



It is evident that genera in which the "original species"' has 



among vascular cryptogams, thus suggesting that hybridization 

 has resulted in the "swamping out" of the early forms. If local 

 adaptation and natural selection alone were at work, it is hard 

 to see why extinction should not be equally common in all these 

 groups. The facts point to the importance of hybridization in 

 completely altering specific type when a group of individuals 

 have been isolated from the main body of the species. 



Against the soundness of these conclusions several points may 

 be urged. Vascular cryptogams are perhaps inherently less 

 variable and quick to produce new species than flowering plants. 

 It may be, too, that cross-fertilization is much more common 

 among them than is generally believed. Whether the recognized 

 "species" among these plants is the equivalent of the "species" 

 among angiosperms, or is a much more inclusive group, is also 

 a matter of doubt. These points can not well be brought against 

 the glumaceous monocotyledons, however, as contrasted with the 

 petaliferous types. Whatever its interpretation, the fact seems 



