environment, whereas Cx. halifaxii occurred frequently
in shaded areas and bamboo groves and showed a
significant positive coefficient for environment (Table
3). The coefficient for season was not significant for
Cx. halifaxii, Notonectidae, Anisoptera and Chaoborus
but significantly positive for Veliidae (Table 3).
Spearman’s rank correlation coefficients (r
s
)
between frequencies of occurrence of mosquitoes and
predators in each habitat type category are shown in
Table 4. An. sinensis and Cx. tritaeniorhynchus showed
positive correlation with all groups of predators. Of
them, correlation was significant with Notonectidae,
Veliidae and the combined all-predator groups for both
An. sinensis and Cx. tritaeniorhynchus. Correlation
with Anisoptera was also significant in Cx.
tritaeniorhynchus. Ae. japonicus and Cx. kyotoensis
showed no significant correlations with predators,
except for negative correlations with Notonectidae.
They showed a weak but non-significant negative
correlation with the combined all predator groups. Ae.
(Stegomyia) spp. and T. bambusa showed negative
correlations with all the predator groups. The
correlation was significant in most cases, except for
the relationship between T. bambusa and Cx. halifaxii.
Both Ae. (Stegomyia) and T. bambusa showed
significant negative correlation with the combined all
predator groups.
The similarity in occurrence patterns of
mosquitoes and predators corresponded to the actual
encounter with predators, expressed in the proportion
of habitats with predators (Table 5). Predators were
present in high proportions of the habitats of An.
sinensis and Cx. tritaeniorhynchus both in summer and
autumn (Table 5). Cx. kyotoensis co-occurred with
predators in more than one-third of its habitats in
summer (Table 5), which is largely due to co-occurrence
with Cx. halifaxii (9 sites). Proportion of habitats with
predators was lower for Ae. japonicus (Table 5). Ae
(Stegomyia) spp. rarely co-occurred with predators, and
all of the habitats of T. bambusa were predator-free
(Table 5).
Table 6 shows Hurlbert’s (1969) C8 association
index of mosquitoes and predators for containers of
0.01 – 1 m2 in shaded sites and bamboo groves surveyed
in summer. A significant positive association was
detected between Cx. kyotoensis and Cx. halifaxii. No
significant correlation was detected between Cx.
halifaxii and other
non-predaceous mosquitoes.
In large containers in open sites, Notonectidae
showed significant negative associations with Ae.
japonicus, and combined all non-predaceous
mosquitoes in summer (Table 7). From 15 containers
of D0.1 m2 in open sites where Notonectidae was found
at the summer survey, only one mosquito larva (An.
sinensis) was collected. In addition, a significant
negative association was detected between Ae.
japonicus and Chaoborus sp., and a significant positive
association was found between combined all nonpredaceous mosquitoes and Cx. halifaxii (Table 7).
However, in the autumn survey, no significant
correlation was detected between mosquitoes and
Notonectidae or other predators, either when the
abundant species were examined separately or when all
non-predaceous species and all predators were
combined (Table 8).