Polynesian Species of My op orum— W ebster 
This last character is undoubtedly valuable 
in delimiting species, although it may oc- 
casionally be inconstant (on the islands of 
Hawaii and Raivavae). The abrupt basal 
curvature of the style observed in some of 
the specimens from southern Polynesia is an 
inconstant character and its importance is 
hard to assess on the basis of the small 
amount of material studied. 
The number of stamens is one of the 
classical key characters for separating M. 
sandwicense from the other Polynesian species, 
but this distinction breaks down when a 
large number of Hawaiian specimens is ex- 
amined. The length and height of insertion 
of the stamens were found to be quite 
inconstant, and the degree of divergence of 
the anther cells is dependent on age and is 
of no taxonomic value at all. 
The drupes of most of the Hawaiian 
varieties differ from the other Polynesian 
species in having a greenish- white exocarp. 
Some plants on the island of Hawaii (var. 
Fauriei), however, may have either whitish, 
pink, or purplish drupes. The drupes of M. 
Stokesii and M. rapense are pink or red and 
brick-red or purplish-red, respectively, while 
those of M. laetum are said to be purple. 
Judging from the variability of color in M. 
sandwicense var. Fauriei, however, it would be 
unwise to consider differences in shade of 
color as especially meaningful. The presence 
of color in the drupes of var. Fauriei and 
its absence in the other Hawaiian varieties 
do seem to be significant, however. 
The bony endocarp is usually subglobose 
or top-shaped and often ridged at the top or 
along the sides. Its shape, although extremely 
variable, is a distinguishing character in some 
varieties and species. The number of cells in 
the drupe is inconstant, but a knowledge 
of the range of variation will assist in placing 
a specimen in the proper category. Counting 
the number of cells is not always easy as 
some of the abortive ones may be quite small. 
There is often a considerable degree of ster- 
ility, for many drupes will be found to 
55 
contain only one or two cells with well- 
developed seeds. 
TAXONOMIC CATEGORIES 
The species is here considered to be a 
population whose individuals have in com- 
mon an ensemble of fundamental morpho- 
logical characters different from that present 
in other species. It is reasonable to suppose 
that this morphological distinctiveness is 
due to various types of barriers (such as 
geographical, physiological, and genetical) 
which prevent a free genetic exchange with 
other species. 
The categories subspecies and variety, as 
here used, apply to groups which occupy a 
discrete region within the total area of the 
species and which are characterized by a lack 
of sharp morphological distinctness from 
other groups within the species. A sub- 
species differs from a variety in its greater 
morphological differentiation, but does not 
necessarily occupy a greater area. Both are 
considered to lack genetic isolation — that is, 
varieties of the same species would be ex- 
pected to interbreed freely when brought 
together — and no doubt often owe their 
existence to geographical barriers. 
In recent years there has been considerable 
discussion among taxonomists as to whether 
the geographically delimited groups within a 
species should be given the rank of sub- 
species or of variety. However, in this work 
it has seemed better to use both categories 
in order better to accommodate the great 
sub -specific range of differentiation. 
The form, in the present intrepretation, is a 
group of plants which occur within a variety 
and differ by very minor or inconstant char- 
acters ; they may or may not be geographically 
localized. Specimens may also be referred to 
forms when they are inadequate to give an 
accurate concept of the population. Inasmuch 
as the forms listed in this study are not con- 
veniently separated morphologically and to 
some extent are designations based on in- 
complete knowledge, it seems unwise to 
