Dioecism in Hawaiian Flora — Gilmartin 
ent areas may be the result of saltational specia- 
tion. That short term dioecism can occur very 
quickly is undeniable. This type of dioecism is 
present according to D. Lewis (1942) in those 
families that have only a few dioecious mem- 
bers. He feels that the majority of dioecious 
species show the short term type of dioecism 
that is controlled by few genes and is reversible, 
and he postulates that it is only within families 
like the Salicaceae that dioecism is irreversible. 
Yampolsky and Yampolsky (1922) estimated 
that about 75 per cent of all the families of the 
phanerogams have some dioecious species. 
The percentage of families with dioecious 
species does not differ appreciably in the two 
study areas, Hawaii and southern California, 
being approximately 15 and 14 per cent respec- 
tively. In comparing these percentages with the 
estimate of 75 per cent by Yampolsky and 
Yampolsky (1922) caution is indicated in 
drawing conclusions. In neither of these two 
floras does the distribution of dioecious species 
per family seem to reflect the distribution for 
families of the world. While this could simply 
indicate that the sample in regard to species per 
family is inadequate here, it could also indicate 
that factors are at play in both these floras which 
are not generally found throughout the world. 
If it is true that dioecism has largely, if not 
totally, evolved in situ in the indigenous dioe- 
cious species of the Hawaiian Islands, and the 
evidence is strong that it has, why has selection 
acted favorably toward dioecism? The answer 
would be that dioecism is a very effective mech- 
anism for insuring outcrossing. Outcrossing 
produces heterozygosity, and heterozygosity pro- 
vides more potential recombination types. In a 
small population, such as would exist shortly 
after the arrival and establishment of long-dis- 
tantly dispersed species, the gene pool would 
be relatively small. Additional genotypes com- 
ing in from other populations would be rare 
indeed. A factor which favored outcrossing and 
the concomitant greater variability would prob- 
ably be selected for. Lewis (1942) and Grant 
(1958) both have emphasized that the genetic 
system and its controlling factors are themselves 
subject to evolution. The factors making up a 
genetic system include mutation rate, chiasma 
frequency, meiosis, fertility, and, of course, 
the breeding system. 
289 
Whitehouse (1950), in emphasizing the im- 
portance of outbreeding to adaptation, has also 
suggested that when species are once adapted 
to a given situation the outbreeding mechanism 
may then be lost. Some current workers, study- 
ing small populations, seem to feel that self- 
pollination systems will be selected for and 
retained. Moore and Lewis (1965:113) write, 
"In a very small population genes promoting 
self-pollination would be at an advantage. Once 
established, an inbreeding race would become 
self-perpetuating." Apparently, however, if the 
selection pressure is great enough, the inbreed- 
ing system need not be self-perpetuating. Such 
a situation with strong selection pressure for an 
open recombination system would exist in the 
Hawaiian flora. The Hawaiian Islands include 
in a small area a large number of very different 
habitats. These habitats may change quickly 
because of the high rainfall producing landslides 
and because of volcanic effects. Different eco- 
logical niches are suddenly made available in 
such a situation. Baker (1953) has emphasized 
the greater advantage that an outbreeder has 
over an inbreeder in occupying a new niche. 
DISCUSSION 
The present author maintains, therefore, that 
selection pressure for an outcrossing breeding 
system has been very strong in Hawaii and has 
resulted in the survival of a higher percentage 
of dioecious species and survival of a lower 
percentage of hermaphrodite species, than are 
found in the North American continent. In the 
tropical continental flora there may be almost as 
high a percentage of dioecism as in Hawaii. 
The evidence is too flimsy to attempt to draw 
any conclusive comparisons between a conti- 
nental tropical flora and an insular tropical 
flora. However, it would not be too surpising 
to discover that a flora such as that in Ecuador 
actually did show a percentage of dioecism 
somewhere between that of an insular tropical 
flora such as Hawaii and that of a continental 
temperate flora such as California. One would 
expect that in a continental tropical flora the 
selection pressure for dioecism, an outcrossing 
breeding system, might be less than in an 
isolated flora such as in Hawaii with a tropical 
climate and yet somewhat more than in a 
