Figure legends
Figure 1. Sampling locations in (a) La Lopé, Gabon, and (b) Rabai, Kenya. The inset in each graph shows the location of the field site in continental Africa. In (a), each point represents a sampling location where one to multiple larval breeding sites were found. In (b), each point represents a single larval breeding site. The color of the point indicates the habitat category: red - domestic (village indoor), yellow - peridomestic (village outdoor), green - forest. The satellite images were from (a) Google Satellite and (b) Bing Satellite in QGIS.
Figure 2. Environmental conditions of larval breeding sites in La Lopé (a, c, e, g) and Rabai (b, d, f, h). (a-b) Principal component analysis (PCA) of all physical variables. The first two PCs are shown, and the variance explained by each PC was indicated in the axis label. Each point represents a single larval breeding site. Colors and point shapes indicate habitat and whether Ae. aegypti were found in the sites, respectively. An eclipse was drawn for each larval breeding site group with a 75% confidence level. The colors of the eclipses represent habitat types and match the colors of the points. The solid and dashed eclipses correspond to Ae. aegypti present and absent sites. The original variables were overlaid on the PC1-PC2 plate with major variables labeled. (c-d) Comparison of Ae. aegypti density between habitats. “+” and “-” denote Ae. aegypti presence and absence, respectively. The boxplots show the minimum, 25% quartile, median, 75% quartile, and maximum. Asterisks indicate statistical significance (p<0.05). (e-f) Comparison of microbial density across larval breeding site groups. (g-h) NMDS analysis of bacterial community compositions. The color and shape of points and eclipses are the same as in the PCA (panel a-b).
Figure 3. Chemical profile of volatile samples collected from Rabai larval breeding sites. Each row represents a compound, and each column represents a larval breeding site. The five columns of points between the compound names and the heatmap summarize whether the compounds were present in each of the five larval breeding site groups. The color and shape of the points are the same as in Figure 2. The color of each cell in the heatmap quantifies the concentration on a log scale. Gray cells indicate that the compound was not found in the larval breeding sites, according to the GC-MS results. The inset Venn diagram shows the total numbers of compounds unique in each habitat or shared between different habitats.
Figure 4. Two-choice laboratory oviposition assays testing preference for field-collected waters, pH, shading, a combination of water pH, salinity, and shading,Ae. aegypti larval density, and bacterial culture. Colony-wise results are shown in Figure S11 in the Appendix. The two choices in each assay are described in detail in Table S2 in the Appendix. Higher OIA implies a preference for the forest condition. Each point represents the OAI of one cage with five gravid females. The mean and 95% confidence interval (CI) were estimated by beta-binomial models. The asterisks and ‘ns’ above each colony indicates whether the 95% CI excludes zero. No significant differences were found between habitats or between colonies in any experiments.
Figure 5. Five-choice laboratory oviposition assays testing preference for bacterial density. Five cups were provided in each cage with increasing bacterial density at 0, 2x105, 1x106, 5x106, 2.5x107 cells/mL (details in Table S2), which correspond to the five columns (left to right) in each panel. Each line connects the five egg counts in one cage. Colors represent the habitats of the colonies. Multiple colonies from the same habitat in Rabai were combined in this figure, and colony-wise results are shown in Figure S12 in the Appendix. A negative binomial model was used to fit the results of each oviposition assay. The model estimates the mean number of eggs in each bacterial density and a 95% confidence interval, shown by the open circles and the error bars, respectively.