Hawaiian Flora — FOSBERG 
shoots that still survive, and none has pro- 
duced a large number. Furthermore, all these 
genera are among the more doubtfully Ameri- 
can of the lot. 
There are only two strand species in the 
American element — Lycium carolinianum and 
Jacquemontia sandwicensis. Chenopodium oahu- 
ense grows at sea level as well as in dry up- 
lands. Of course, some of the pantropic 
strand plants may be of American origin. 
Of the total, only Sapindus and Psychotria 
have seeds too large to be easily distributed 
by wind. 
There are about eight cases of probable 
bird dispersal — plants whose seeds logically 
might have stuck to birds’ feet or feathers. 
Prominent among these is Fragaria chiloensis, 
which grows on sea beaches from Chile to 
Alaska and in the uplands of the island of 
Hawaii. Alaska is the summer home and 
Hawaii the winter home of the Pacific plovers 
and curlews. In addition to these, there are 
about eight other plants with fleshy fruits 
which may have been brought in birds’ in- 
testines, though this is less likely over such 
distances. 
Human agencies cannot be absolutely ex- 
cluded in about 10 cases, though the possi- 
bilities have been carefully weighed, and only 
about 2 of these 10 are regarded as at all 
likely. Those that seem really to have entered 
Hawaii by human introduction have been ex- 
cluded as non-indigenous. One cannot posi- 
tively exclude very early historical introduc- 
tion for a few plants, such as Hesperocnide and 
Daucus, or prehistoric human transport for 
such as Argemone; but it is unlikely. The 
Hesperocnide is considered an endemic species 
and the Argemone an endemic variety of 
Argemone alba. 
Gossyptum tomentosum is a special case. Cyto- 
logical investigations by Hutchinson, Steph- 
ens, and Silow have led them to the conclu- 
sion that this species and the two widespread 
cultivated American cottons form a closely 
related group derived by hybridization be- 
tween an Asiatic cotton and a wild diploid 
205 
American cotton. They think that this hy- 
bridization followed prehistoric human intro- 
duction of an Asiatic cotton into America 
(where it does not now persist), and that 
Gossypium tomentostm was then carried back 
to Polynesia and to Hawaii by Polynesian 
travelers. 
There are several weaknesses inherent in 
this theory. Even supposing that the Poly- 
nesians had made such voyages, it seems 
scarcely likely that they would have selected 
for taking back the one perfectly useless 
cotton of the three, or that it would not 
have persisted elsewhere along the route in 
Polynesia. The fiber of Gossypium tomentosum 
is only a few millimeters long. The greatest 
cause for doubt, however, lies in Dr. Show’s 
statement (in conversation, 1949) that Gos- 
sypium tomentosum is closely related to the 
cultivated cottons. Morphologically, at least, 
this does not seem to be true. T, H. Kearney, 
long an authority on cottons, has told me 
that he regards it as closest to a wild species 
of the Galapagos Islands. I am well ac- 
quainted with Gossypium tomentosum and with 
both cultivated American cottons and find 
little similarity. 
I suggest that Gossypium tomentosum be re- 
examined cytologically, using material about 
whose origin and identity there can be no 
doubt. It may be that there has been a con- 
fusion with the forms of Gossypium harhadense 
that have long been introduced and estab- 
lished in Hawaii. 
To return to general considerations, it 
seems fairly safe to assume that identity or 
close relationship with American species in- 
dicates that isolation from them has not been 
of very long standing. The lack of extensive 
evolutionary differentiation suggests the same 
thing. It is realized, of course, that there may 
well be exceptions to these generalizations. 
But when almost 60 per cent of the presumed 
American stocks in the flora are identical 
with or very close to their American relatives, 
and when over 94 per cent have not given 
rise to any number of evolutionary progeny. 
