DERIVATION OF THE FLORA OF HAWAII 21 



ably post-glacial, long after the time at which the ancestors of the exist- 

 ing birds are supposed to have reached Hawaii. 



The strand flora of Hawaii is notably poor in species when com- 

 pared with most tropical regions. The principal ocean drift at present is 

 from the northeast, but the likelihood of many species from northwest 

 America, which might have reached Hawaii as drift, becoming estab- 

 lished in such a different climate, is hardly probable. Just how such wide- 

 spread species as Ipomoea pes-caprae, Scaevola, Vigna, Hibiscus and 

 Terminal reached the islands is doubtful, but very probably by drift, 

 as many strand plants have special adaptations for carriage by water. 



As we have already intimated, it is likely that a considerable number 

 of species have reached the islands through the agency of wind. This is 

 probably the case with some of the ferns, including the two American 

 species ; and we are inclined to regard this as probable also for most of 

 the Compositae, and perhaps for other forms with very small and light 

 seeds. 



That any considerable number of species, except the littoral ones, 

 have reached the islands as drift, is very improbable. Guppy suggests 

 that certain genera, e. g., Erythrina, Canavalia, Sophora, which possess 

 littoral species elsewhere but which in Hawaii are represented by inland 

 species only, are descended from littoral species which came as drift, but 

 which have become extinct. While this may be true, it seems rather im- 

 probable. 



While recognizing the difficulties involved in any attempt to explain 

 many perplexing facts connected with the origin of the Hawaiian flora, 

 it seems to the writer that the evidence offered by the advocates of the 

 theory of the strictly oceanic character of the archipelago is by no means 

 adequate, as it involves assumptions, such as the extraordinary role of 

 fruit-eating birds in seed distribution, which are, to say the least, quite 

 incapable of satisfactory proof. 



We believe that the evidence for a more or less direct connection 

 with some continental area to the southwest is very strong, and the ad- 

 mission of such a connection will readily explain most of the difficulties 

 that are involved in the assumption that the archipelago has always been 

 completely isolated. 



A study of the ocean depths would indicate that the whole Poly- 

 nesian area might represent the remains of a former continental area, or 

 series of large continental islands, connected at some time, perhaps in 

 the early Tertiary, with Australasia and the Malayan Archipelago. Some 

 such connection seems necessary to explain the great preponderance of 

 Australasian and Malayan types in the flora of Hawaii, especially such 

 forms as offer no ready means of transport by any known agency. 



