396 
PACIFIC SCIENCE, VoL V, October, 1951 
more specific microhabitats, which are listed 
in Table 2. 
The term "habitat index" is used as a 
convenient way of referring to the ratio be- 
tween the number of collections of a species 
in a particular habitat and the total number of 
collections of that species. 
Habitat Factors 
An extensive literature has accumulated in 
an attempt to determine the effect of chemical 
and physical factors on the breeding habits of 
various species. Many factors have been in- 
vestigated, but specific conclusions concern- 
ing the role of any one factor as an absolute 
indicator of potential breeding are almost 
entirely wanting. 
A comprehensive quantitative study of the 
physio-chemical factors such as light inten- 
sity, dissolved oxygen, etc., was impossible 
under our field conditions. Our aim was to 
describe the general character of the habitat 
by careful observations using quantitative 
determinations when possible. Therefore only 
major factors such as exposure, superficial 
character of the water, vegetation, pH, were 
considered. In this report the habitat factor is 
given as the percentage of our collections in 
which that factor was recorded. 
It should be noted that stagnant water was 
always static, with varying amounts of or- 
ganic debris in the early stages of decompo- 
sition. In polluted water, decomposition was 
always well advanced and produced a foul 
smell. pH was determined in the field for 
many but not for all collections, using 
Squibb’s nitrazene paper (range 4. 5-8.0). 
Salinity was measured with a hydrometer in 
all cases of suspected brackish water and was 
considered experimentally for An. punctulatus 
and for An. farauti. 
Species Association 
The most common method of presenting 
numerical data indicating the degree of 
association is in the form of a table showing 
the actual number of times any two species 
were found together in individual collections. 
However, Hacker (1921) proposed the "Asso- 
ciation Unit" and Dice (1945) the "Associa- 
tion Index" as measures of the degree of 
association between species. A combination 
of the methods of Hacker and Dice was used j 
in determining my association indices. ! 
Collections of any one species were either 
pure or in various combinations with one or 
more additional species of mosquito larvae, j 
The value (P) of pure collections for any one 
species (A) is the ratio between the number ; 
of times the species was collected alone and j 
the total collections of species A. Then a [ 
measure of the value of mixed collections ! 
(M) can be taken as 100-P. Each cohabitant j! 
is assigned a relative portion of this total as || 
follows: hM/t, where h equals the number of |i 
times species A was found with any other 
species and t equals the number of times j 
species A was found with all cohabitants. In j 
this paper, species having an association of i' 
less than 4 per cent have been grouped and 
given as "others." 
ll 
Bironella I 
Because of the confusing status of various ! 
species in the genus Bironella, it was impossi- | 
ble accurately to determine species of this 
genus during the collecting period. Labora- 
tory rearing was so difficult that only five 
adults were available for checking: two speci- 
mens from Milne Bay were identified by Dr. 
Stone as B. travestitus and three from the same | 
locality as B. soesiloi sensu Swellengrebel and j 
Rodenwaldt. Because of the rather limited j 
literature on the biology of Bironella it is con- | 
sidered worth while to present these data 
even though more than one species is repre- | 
sented. 
distribution: Bironella species were pres- ! 
ent in collections from Milne Bay, Morobe, 
Finschhafen, and Saidor. Collections were 
nowhere numerous and the number of indi- 
viduals in each collection was always rather 
limited. It seems that Bironella species would | 
have been represented more abundantly if 
extensive collecting had been possible from 
