Experiment Two
Given that the top-down effects of crayfish were so strong, essentially eliminating juvenile snails, we were concerned that the top-down effects of crayfish were masking any positive effects of herbicides on snail production and schistosomiasis risk. Consequently, we conducted a follow-up experiment where we crossed the same herbicide exposures with the presence and absence of crayfish. However, using five crayfish led to complete depredation of snails in the predator tanks, reducing our number of tanks with live snails in half (n = 32). Subsequently, we found that little variance in Bi. glabrata biomass was explained by treatment (Table S4), and likelihood ratio tests did not detect significant herbicide effects in our nested models (Table S5). An interaction term between crayfish presence and herbicide type for infected, uninfected, and total Bi. glabrata biomass, after model selection (Tables S11-S13), showed no significant interaction between crayfish presence/absence and herbicides (Table 2). Focusing on tanks without crayfish predators, the herbicides atrazine and acetochlor were associated with increases in the average biomass of infected snails (55 %) and total snails (131 %), respectively (Table S13; Fig. 3a,3c). In contrast, the herbicide alachlor was associated with a decrease in uninfected snail biomass (126%) in tanks without predators (Table S14; Fig. 3b). A Pearson’s correlation between snail biomass and infected snail biomass suggested a significant positive association (rho = 0.67, t30 =4.96, p < 0.001; Fig. 1c). Moreover, infected snail biomass was a significant positive predictor of total cercariae in tanks (Fig. 1d).