8 
PACIFIC SCIENCE, Vol. VII, January, 1953 
lands Indies. In Hawaii this type of species ra- 
diation has been observed in groups as diverse 
as the birds, the drosophilids, and the land 
snails. However, possibly because the marine 
forms have not been as completely isolated, 
or possibly because the marine animals have 
not been as thoroughly studied, the phenom- 
enon ha,s not been especially noted in the 
animals inhabiting the reefs and shores. 
The third possibility is that the variation is 
not unusual or unique to Hawaii and occurs 
in all areas but has been overlooked to a large 
extent by other workers. They may not have 
had a large enough series of specimens of one 
species to give a picture of the variation, or, 
having the specimens, they may have been too 
imbued with the type-specimen concept to 
discuss the variations in their papers. It is true 
that in many cases the earlier workers did dis- 
cuss these variations; de Man, for example, 
presented long numerical tables on the varia- 
tions in some of his species. Yet in his keys, 
de Man (1911: 192) used characteristics that 
could easily overlap, e.g., 
i. Rostrum usually 3-times (2.75-3.25) as long as 
wide. . . . Dactylus of third legs 3-times as long 
as thick at the base. . . . 
ii. Rostrum 2 2-times as long as wide. . . . Dactylus of 
third legs 2.51 times as long as thick at its base. . . . 
Coutiere has erected species upon differ- 
ences with even less allowance for variation. 
De Man states in his introduction (1911: 133) : 
Coutiere also first called attention to the great impor- 
tance of relative measurements . . . and it was just by 
means of these new characters that often specimens of 
small size of Alpheus and Synalpheus proved to be species 
that were still unknown, while formerly such specimens 
usually would be regarded as juvenile forms or at most 
varieties. 
Coutiere stated (1909: 36) that some of these 
species were even based largely on "geo- 
graphic characteristics." For example, "5. 
paulsoni and the other Indo-Pacific forms are 
not represented [in America] by identical 
forms, but the differences are at times so 
slight that, without indication of locality, the 
identification would be very difficult." 
The lack of a sufficient number of speci- 
mens of a given species to show the extent of 
variation often renders it necessary to erect 
new species for what may later be found to be 
divergent members of the same species. This 
has been done in the past, and I fear that in 
this paper it may have been done again. For 
example, Crangon pseudopugnax Banner is sep- 
arated from C percyi (Coutiere) chiefly by the 
differences in the base of the rostrum, and the 
character seems to be quite clear-cut and de- 
cisive. However, of the two species only nine 
specimens are known from Hawaii, and per- 
haps, if a large number of specimens were col- 
lected, it might be found that the two species 
described are merely the extremes of a vari- 
able single species, with specimens of inter- 
mediate form of the rostrum more plentiful. 
Only by extensive studies can it be deter- 
mined which of these postulated causes for 
the observed variation is the correct one. To 
arrive at a solution of the problem it will be 
necessary to study not only a large series of 
specimens from Hawaii but also many speci- 
mens of these species from other portions of 
their ranges. Specimens from the southwest 
Pacific especially should be studied, as that 
area appears to be the center of distribution 
for the Indo-Pacific members of the family. 
Through such a study some of the criteria 
based on small differences in the relative pro- 
portions, like those of the antennular or carpal 
articles, will perhaps have to be discarded. 
With any such change in the criteria, some of 
the species now recognized will fall into sy- 
nonymy. It may also be necessary to find new 
and more stable characteristics upon which to 
base specific descriptions. 
In any case, the work done in this present 
study is sufficient to cast doubts on the valid- 
ity of some of the species that have been de- 
scribed on the basis of slight differences in 
what appear to be variable characteristics. 
Family CRANGONIDAE Weber 
Carapace smooth, provided with cardiac 
grooves; rostrum reduced, antennal and bran- 
chiostegal spines always absent; in known Ha- 
waiian species carapace more or less project- 
