240 
PACIFIC SCIENCE, Vol. IX, April, 1955 
variety of artificial containers. At least two 
types are represented. One, from a paint pail, 
resembles Espiritu Santo larvae from bamboo 
and has a very short siphon, long hair 4-X 
and a few short comb scales. The other form, 
from an old tire, contains a variety of larval 
forms, some of which have a much shorter 
hair 4-X, less than 1.5 the length of the anal 
saddle, and the stellate tufts with longer and 
more numerous branches. 
Collections are available from Aneituym and 
Futuna in the southern New Hebrides. The 
populations on Aneituym are quite distinct 
from those on Efate and have been found 
breeding in coconut shells, a tree hole, a tire, 
and a tin can with brackish water. All the 
larvae, except those in brackish water, have 
a long dorsal anal gill and all possess rela- 
tively short comb scales. The tree hole larvae 
have slightly longer, sharply pointed comb 
scales instead of blunt, and a greater number 
of branches in the stellate tufts. The popula- 
tions on Futuna are quite distinct from those 
on Aneituym and have been found breeding 
in a rock hole in coral, a tree hole, a canoe 
and a small ground pool. All the larvae have 
the anal gills shorter, the comb scales more 
pointed, hair 4-X shorter and, except for those 
from the tree hole, a longer siphon. Larvae 
from the different habitats appear to be more 
similar than on the more northern islands and 
surprisingly the tree hole larvae have a shorter 
siphon than any of the others. 
A single collection from a hole in a trunk 
of Araucaria cookii has been examined from 
E. Lifu in the Loyalties. These larvae resemble 
most closely those from New Caledonia and 
have a long siphon and about 30 long comb 
scales (Fig. 5 h). 
From New Caledonia I have seen five larval 
collections, three of them without habitat 
data, the others from tin cans and from flower 
bracts of palms. One of the larval collections 
without habitat data (La Foa) is probably 
from tree holes since the larvae have the 
characteristic pigmentation and development 
of stellate hairs associated with this habitat. 
In addition adults reared from larvae col- 
lected in bamboo stubble were examined. All 
the larvae fall within the atypical melanesiensis 
complex and exhibit much less variation in 
the length of the siphon and the number and 
character of the comb scales while retaining 
the usual variation in the branching of the 
stellate tufts. It is possible to segregate the 
New Caledonia populations, along with those 
from E. Lifu and Art Is., from the remaining 
atypical melanesiensis into a southern group 
on the basis of the characters indicated in the 
diagnosis. As will be noted below, the adults 
are also distinguishable. On the other hand 
there appears to be some intergradation of 
characters in the southern New Hebrides. The 
larvae from flower bracts of palms (Fig. 5 g) 
appear to be distinct from the others espe- 
cially in the extreme development of the 
dorsal anal gill and may represent a good 
species comparable to the folicola of Espiritu 
Santo. Since I have seen but a single cc flec- 
tion of these larvae and the other material 
from New Caledonia is very scanty, addi- 
tional material is necessary before a decision 
as to the status of this form can be made. 
Finally, I have seen a single larval skin from 
a "rock pool into which assorted debris had 
been flushed by heavy rain" (Laird) on Art 
Island, Belep group (N. W. of New Cale- 
donia). This larva resembles the New Cale- 
donia and Loyalty atypical melanesiensis . 
The adults of melanesiensis are quite variable 
in minor details of coloration. I have not been 
able to correlate any of these differences in 
the adults with the different larval ecological 
forms. Such correlation is particularly difficult 
because of the total lack of individual rear- 
ings. It is not even known whether or not 
some of the individual differences in the larvae 
are sexual. On the other hand geographical 
differences in adult coloration are quite evi- 
dent. The forms in the New Hebrides are 
much darker than in New Caledonia and E. 
Lifu, as indicated in the diagnosis. The ma- 
terial from the southern New Hebrides is too 
scanty to determine whether or not this char- 
