This thesis describes a study of the dog and fox reservoir populations of Leishmamainfantum in an endemic focus in Amazon Brazil Following a brief review of the relevant literature on canine visceral leishmaniasis in Chapter 1, the aims of the study are described Broadly they were to (I) quantify the courses (and their inter-relationships) of infection, disease and infectiousness in dogs, (2) investigate the implications for culling strategies, (3) assess the relative role of dogs and foxes in transmission. Chapter 2 describes the overall study design. A cohort of 126 previously unexposed native dogs was established in 24 endemic study communities and monitored at bimonthly intervals over a period of 24 months. A total of 756 sera were tested by ELISA, and 514 bone marrow samples were examined either as smears, or following inoculation into hamsters or culture medium. Clinical examinations were performed on 116 dogs on 562 independent occasions, and 50 dogs were experimentally exposed to colony bred Lu. longipalpis (the vector) in 173 xenodiagnostic trials. Longitudinal demographic and serological data were also collected on the resident dogs of 15 communities and on a sympatric free-ranging population of crab-eating foxes, Cerdocyon thous. These data were added to those for the same populations collected during a previous study (by the author) giving 5 years of data, all of which is described in this thesis. Using comparable methods as for dogs, 37 foxes were clinically, serologically and parasitologically sampled on 74 occasions, 26 of which were also examined by xenodiagnosis in 44 trials. The principal results of this thesis are presented in Chapters 3- 6. Chapter 3 describes the demographic parameters relevant for transmission dynamics. These showed that the resident dog population had a high turnover rate (0 42 per year) characterised by a mortality rate of 0.40 per year. Dog abundance was sustained by immigration (owner-mediated) rather than by birth. The mortality rate was positively associated with seroprevalence, incidence, and sandfly abundance in houses and animal sheds. The seroconversion rate between villages was positively associated with sandfly abundance in houses and animal sheds, and negatively associated with the mean number of dogs per household The fox population replacement and mortality rates were similar to dogs, but with no evidence of mortality due to Leishmania. Chapter 4 describes the courses of infection and disease in the sentinel population. By the end of the study 80 dogs were identified as infected, representing 93% of the 86 dogs which remained in the study for > 3 sample rounds The seroconversion and serological recovery rates were 0 269 and 0 006 per month, respectively, the average time to infection was 115 days, and time from infection to seroconversion was 94 days All dogs developed one or more clinical signs of CVL by the end of 24 months, with an average time from infection to clinical onset of 66 days. Only 9% of dogs classified as symptomatic by their longitudinal clinical profile fully recovered from the disease. Clinical severity outcome was positively associated with antibody titre, parasite isolation success, and mortality. Forty-nine percent of dogs with confirmed Leishmania infections died by the end of the study. The risk of mortality was positively associated with the severity of infection and disease. In Chapter 5, the results from serial xenodiagnosis of the sentinel population revealed that the onset of infectiousness occurred a median 128 days after seroconversion with a latent period (time from infection to infectiousness) of 222 days. The heterogeneity of infectiousness was extensive both between and within dogs: 45% of seropositive dogs were observed to become infectious; 20% of the infectious dogs were responsible for 80% of all sandfly infections. Infectiousness was positively associated with high antibody titres, and severe clinical signs. However, neither antibody titre, clinical signs, biochemical parameters, nor any combinations of these, proved to be reliable surrogate markers of infectiousness. Chapter 6 provides comparable longitudinal serological and parasitological data for the fox population. Fox infection rates were similar to those for dogs, though there was an absence of clinical signs. No foxes were observed to be infectious to Lu. longipalpis by xenodiagnosis. Chapter 7 concludes that (1) the dynamics of canine L. infantum infection and disease in Marajo is similar to that in Europe, (2) the detection of potential clinically severe cases may be possible in early infection (e g for treatment by the veterinarians), however (3) selection of infectious dogs for targeted control (treatment, elimination or other) is not possible using the immunological and clinical parameters as surrogate markers, (4) foxes are not important for human transmission in the presence of infected (infectious) dogs.