Dioecism in Hawaiian Flora — Gilmartin 
I obtained for the circumscribed area in south- 
ern California is more complete than that for 
the Hawaiian Islands since a considerable num- 
ber of species have been added to the indigenous 
Hawaiian flora as new discoveries of species 
were made since 1888. A much smaller number 
of additional indigenous species have been 
added to the flora of California since 1935. 
Nevertheless, Hillebrand’s Flora is relatively 
complete. He included slightly more than 700 
species among the indigenous phanerogams of 
Hawaii. St. John (1946) estimated that the 
indigenous phanerogams in Hawaii included 
1,795 species and infraspecific taxa. If the 
species only are considered, it is very likely that 
there would be fewer than 1,200. The number of 
dioecious species in Hawaii (based on Hille- 
brand, 1888) and that of southern California 
(Munz, 1935) are compared in Table 1. Also 
compared is a bare representation of the flora 
of Ecuador, a tropical continental flora, the data 
for which are based on Diels’ report (1938) 
of several months’ collecting and include a total 
of only 658 species. This must represent a very 
small portion of the entire flora. Furthermore, 
Diels probably included a disproportionate 
number of species from the alpine region of 
the Andes where he spent more of his collecting 
time. Therefore, some reservations should be 
held in comparing the data on the flora from 
Ecuador with those of the other two floras. 
However, this information is included because 
it represents, perhaps poorly, both a tropical 
flora (as does the Hawaiian) and a continental 
flora (like that of southern California). 
Table 1 indicates that the percentage of 
TABLE 1 
Comparison of the Total Number of 
Indigenous Angiosperm and Gymnosperm 
Species in Three Areas as 
Related to Dioecism 1 
AREA 
TOTAL 
SPECIES 
DIOECIOUS 
SPECIES 2 
PER CENT 
DIOECIOUS 
Hawaii 
730 
39 
5% 
So. Calif. 
2,335 
67 
3% 
Ecuador 
658 
20 
3% 
1 Data compiled from Hillebrand (1888) for Hawaii, Munz 
(1935) for southern California, and Diels (1938) for 
Ecuador. 
2 Only those species which are listed as dioecious are 
included; polygamous or gynodioecious forms are not in- 
cluded. 
287 
dioecism for the southern California flora is 
about 3 per cent and that of the Hawaiian flora 
about 5 per cent. This can be compared with 
figures of Yampolsky and Yampolsky (1922) 
which show 5 per cent complete dioecism for 
flowering plants of the entire world. If the 
number of dioecious species among the Hawaiian 
flora is incorrect it is very likely that the error 
was in underestimating dioecism. Hillebrand 
often was forced to describe species on the 
basis of inadequate specimens collected by 
someone else, and perhaps accompanied by poor 
notes. Thus, he was not always able to state 
whether a species was dioecious, monoecious, 
or hermaphroditic. 
In compiling the data for Table 1, only those 
species clearly described as "dioecious” were in- 
cluded in that category. The percent of dioecism 
might have been higher if gynodioecism had 
been included. It is therefore safe to assume 
that the proportion of dioecious species in the 
Hawaiian flora is at least 5 per cent and very 
likely higher. The estimate of dioecism for 
the continental, temperate flora of southern 
California, as obtained from Munz’s Manual , 
is probably much more accurate, and indeed it 
compares well with the estimate of Baker (1966, 
personal communication) for all of California — 
2.4 per cent. A two-by-two contingency table, 
chi square test for homogeneity was made for a 
comparison between dioecism in the two study 
areas, southern California and Hawaii. The 
results show that the difference in ratios is 
significant (P < .005). It is therefore extremely 
likely that a real difference exists in the amounts 
of dioecism in the two floras. 
DIOECISM AND WOODY HABIT 
A comparison was then made to determine 
whether the difference in percentage of dioecism 
in the two areas might merely reflect the higher 
proportion of woody species in all tropical 
floras. Woody species are never annuals, and 
annuals do not usually have any sort of out- 
crossing mechanism. Baker (1959^) emphasized 
the point that there is a close association between 
outcrossing and perennial habit. Trees generally 
show outcrossing and heterozygosity (Baker, 
1959^). 
Table 2 presents the results of the compari- 
