54 
PACIFIC SCIENCE, VoL V, January, 1951 
by observation of a great many specimens— 
is essential in defining the taxonomic charac- 
ters of the various groups. 
Growth habit, as has been indicated, is so 
greatly modified by environment that it is 
probably of no taxonomic significance. The 
stem is significant only in distinguishing var. 
stellatum from var. Degeneri, the former hav- 
ing pubescent branchlets. 
There are many leaf shapes but few of 
them are constant in any one population. 
Serrate margins are found sporadically among 
nearly all the Hawaiian varieties and must be 
used with extreme caution as a taxonomic 
character. On the other hand, this serrature is 
constant in the southern Polynesian forms 
and may be used dependably to separate M. 
Stokesii from M. rapense. Conspicuous dif- 
ferences in leaf texture occur among the 
Hawaiian plants, but in the main these appear 
to reflect environmental differences. The 
relative conspicuousness of the nervation 
depends in turn on leaf texture and is thus 
obviously unreliable. 
The leaves of all the Polynesian species 
have subepidermal pellucid-punctate tissue, 
and the distribution and type of these pel- 
lucid spots appear to be good specific cri- 
teria. M. Stokesii and M. rapense are difficult to 
separate from the New Zealand M. laetum 
except for the fact that the glandular spots 
on the leaves of the latter are far more 
conspicuous. In M. laetum the pellucid spots 
are visible by reflected light (on the upper 
leaf surface) and give a distinct "polka dot” 
effect in transmitted light. The spots of the 
two Polynesian species are smaller, less con- 
spicuous in transmitted light, and almost 
invisible in reflected light. 
A readily recognizable and often valuable 
leaf character in Hawaiian Myoporum is the 
presence of pubescence. Unfortunately, while 
pubescent leaves are constant features of 
varieties stellatum and Degeneri they may 
appear sporadically in populations of other 
varieties which ordinarily have glabrous 
leaves. These sporadic occurrences seem 
often to be a response to wounding caused 
by grazing animals, although in some cases 
the lower "sucker” shoots from an uninjured 
plant may bear pubescent leaves. In fact, 
even if the leaves on basal or wound shoots 
are not pubescent, they are often serrate and 
larger than normal leaves. Furthermore, there 
is evidence that at least in some groups of 
Myoporum sandwkense the "juvenile” leaves 
of a young plant are of the abnormal serrate- 
pubescent type, while those formed later are 
entire and glabrous. Further investigations 
on this subject are much to be desired. 
The number of flowers in the axillary 
clusters varies not only between plants but 
even on the same branch and thus does not 
appear to be of any great taxonomic value. 
The same can be said of the flower pedicels, 
which vary widely, but inconsistently, in 
length and thickness. 
The number and length of the calyx lobes, 
within certain ranges of variation, are sig- 
nificant, while their shape (on account of its 
plasticity) is usually less so. The size and 
shape of the corolla have not been used for 
distinguishing varieties; although there are 
readily apparent differences, they are over- 
lapping and inconstant. The amount of 
pubescence on the corolla throat is usually 
quite variable and may or may not have 
significance. 
The ovary is generally uniform in shape and 
constant in size, within any one variety. It is 
glabrous and has a somewhat expanded, dif- 
ferently colored basal region which seems to 
be nectariferous. At anthesis the flowers may 
be noticeably fragrant, although I have 
personally found the flowers of M. sand- 
wkense to be almost odorless. In marked 
contrast to the Hawaiian species in this 
regard are M. Stokesii and M. rapense, whose 
flowers are credited in collectors’ notes as 
ranging in scent from "strong sweet” to 
"fetid.” 
The style furnishes some of the best tax- 
onomic criteria, especially in its length and 
in the presence or absence of pubescence. 
