BackgroundThe host species composition in a household and their relative availability affect the host-feeding choices of blood-sucking insects and parasite transmission risks. We investigated four hypotheses regarding factors that affect blood-feeding rates, proportion of human-fed bugs (human blood index), and daily human-feeding rates of Triatoma infestans, the main vector of Chagas disease.MethodsA cross-sectional survey collected triatomines in human sleeping quarters (domiciles) of 49 of 270 rural houses in northwestern Argentina. We developed an improved way of estimating the human-feeding rate of domestic T. infestans populations. We fitted generalized linear mixed-effects models to a global model with six explanatory variables (chicken blood index, dog blood index, bug stage, numbers of human residents, bug abundance, and maximum temperature during the night preceding bug catch) and three response variables (daily blood-feeding rate, human blood index, and daily human-feeding rate). Coefficients were estimated via multimodel inference with model averaging.FindingsMedian blood-feeding intervals per late-stage bug were 4.1 days, with large variations among households. The main bloodmeal sources were humans (68%), chickens (22%), and dogs (9%). Blood-feeding rates decreased with increases in the chicken blood index. Both the human blood index and daily human-feeding rate decreased substantially with increasing proportions of chicken- or dog-fed bugs, or the presence of chickens indoors. Improved calculations estimated the mean daily human-feeding rate per late-stage bug at 0.231 (95% confidence interval, 0.157–0.305).Conclusions and SignificanceBased on the changing availability of chickens in domiciles during spring-summer and the much larger infectivity of dogs compared with humans, we infer that the net effects of chickens in the presence of transmission-competent hosts may be more adequately described by zoopotentiation than by zooprophylaxis. Domestic animals in domiciles profoundly affect the host-feeding choices, human-vector contact rates and parasite transmission predicted by a model based on these estimates.