Characterizing larval breeding sites: competition and predation
The density of Ae. aegypti was similar in forest and peridomestic larval breeding sites in La Lopé (Figure 2c, W = 33.5, p = 0.961). Other mosquito species were recorded only in the La Lopé forest, with a similar density between Ae. aegypti present and absent site (Figure S3a, W = 121.5, p = 0.723). The frequency of finding predatoryToxorhynchites larvae (Table S3) did not differ among larval breeding site groups (ꭓ2=3.67, df=3, p=0.30), between habitats (ꭓ2=1.14, df=1, p=0.29), or between Ae. aegypti present and absent sites (ꭓ2=0.41, df=1, p=0.52).
In Rabai, Ae. aegypti density was significantly lower in domestic larval breeding sites (Figure 2d, forest vs. domestic: W=19, p<0.001; peridomestic vs. domestic: W=140, p=0.013). The density difference between domestic and forest containers was mainly driven by the larger size of the former. In contrast, the difference between domestic and peridomestic sites was due to the higher number of mosquitoes found in the latter (negative-binomial model, rate ratio=2.89, p<0.001). Peridomestic and forest sites had similar Ae. aegypti density (W=69, p=0.583). When including other mosquito species, domestic breeding sites also had lower overall mosquito density than the other two habitats (Figure S3b, Table S5). Within the forest, Ae. aegypti present and absent sites had similar mosquito density.