426 
PACIFIC SCIENCE, Vol. V, October, 1951 
ubiquitous puddles, collections from "fresh 
surface water" had a more scattered distri- 
bution. This was directly correlated with the 
availability of these habitats. 
4. Ninety-two per cent of all collections were 
from fresh water, only 8 per cent from brack- 
ish. All brackish habitats were temporary or 
semi-permanent and were sharply localized. 
This was due largely to the scarcity and total 
non-productivity of brackish marshes and 
swamps. In several areas (especially Gama- 
dodo and Morobe) extensive mangrove 
swamps were present but mosquito larvae 
were entirely absent, probably due to a sig- 
nificant daily tidal fluctuation. 
5. Seventy-five per cent of the collections were 
from temporary water, 17 per cent from semi- 
permanent, and only 8 per cent were from 
permanent water. Since there was a remark- 
ably high percentage of collections from 
temporary water, I believe that a significant 
degree of control could have been achieved 
by dealing solely with these habitats. 
6. Approximately 50 per cent of all collec- 
tions were from puddles and artificial con- 
tainers. This figure emphasizes the over- 
whelming importance of man-made habitats. 
Table 3 gives the habitat indices for each 
fresh-water species represented by more than 
10 collections. In each case only habitats with 
indices greater than 10 are included. It is felt 
that lower figures represent habitats too 
rarely utilized to be significant. 
The habitats are arranged in a sequence 
from clear, natural surface water to polluted 
water in artificial containers. Within this 
sequence they fall into three major categories, 
each with its characteristic group of species. 
1. In clear, natural surface water, character- 
istic species were Bironella spp., U. argyro- 
tarsis^ and C. fraudatrix. 
2. Artificial surface water was almost en- 
tirely puddles. Characteristic species were: 
An. farauti, C. pullus, C. halifaxi, An. punctu- 
latus, A. vexans, and C. annulirostris. 
3. In water in containers, characteristic 
species were: (a) A. kochi, A. keefei, and Arm. 
hreinli, which were found primarily in natural 
containers; and (b) U. papua, U. nigerrima, 
A. albolineatus, A. notoscriptus, Arm. milnensis, 
Toxorhynchites spp., T. himaculipes, A. scuteU 
laris, and C fragilis, which were found in both 
natural and artificial containers. 
Very few species bridge the gap between 
surface and container habitats to a significant 
degree. Only C. papuensis, A. novalhitarsis, and 
A. papuensis were frequent in both surface 
water and in water in containers. 
Of the disease vectors, An. farauti and An. 
punctulatus occurred primarily in artificial sur- 
face water habitats whereas A. scutellaris was 
found primarily in containers. 
Habitat Factors 
Exposure: In most cases shade or exposure 
was a natural characteristic of the habitat. | 
For example, tree holes and mountain creeks, 
by their nature, are shaded, whereas coral 
pools are exposed. Species with high indices 
for such habitats showed correspondingly 
high values for shade or for exposure. In these 
cases the actual role of this factor could not 
be determined. However, for several species, 
exposure seemed significant. An. punctulatus 
and A. vexans were common in exposed 
puddles, whereas C. annulirostris was more 
frequent in shaded puddles. A. albolineatus 
and C. fraudatrix were found in a wide habi- 
tat range but typically in shaded situations. 
Superficial character of water: Based on 
organoleptic methods, the habitats were 
divided into four categories: clear, muddy, 
stagnant, polluted. For the most part, a par- 
ticular species was not found solely in any 
one of these categories. This would be ex- 
pected because of the wide habitat range for 
nearly every species. However, the following 
were frequently associated with these water 
types: Cledx— Bironella spp., H. genurostris, 
A. keefei, and A. dasyorrhus; muddy — An. 
punctulatus, C. halifaxi, and C. pullus\ polluted 
— Arm. breinli. Arm. milnensis, and C. fragilis. 
