This thesis presents an investigation into chemotherapeutic use on commercial, land-based trout farms. Fourteen farmers, who produced approximately one third of the trout grown in Ontario, Canada, participated in a 21 month prospective epidemiological study from October 1992 to June 1994. Participants recorded details of chemotherapeutic treatments, as they initiated them, on special forms. During monthly farm visits, further details about treatments were gathered using structured interviews. Since most farmers did not keep tank-level production records of sufficient quality for use in research, a semi-structured interview was used during farm visits to gather other tank-level health, production and management data. The risk of chemotherapeutic contamination of fish sold for human consumption by these Ontario land-based trout farms was judged to be extremely small. Chloramine-T or farmalin were the chemotherapeutics used in nearly 90% of treatments. Antimicrobials were only used on one farm and never within 60 days of slaughter. Of the 730 treatments initiated, only 8% occurred in fish larger than 200 g. Forty-two percent of treatments were therapeutic, 42% were preventive, and 16% were repeat treatments due to treatment failures. Most therapeutic treatments (65%) were initiated for "gill diseases", as diagnosed by the farmers. The proportion of therapeutic treatments that required re-treatment(s) varied significantly by farm. Over 92% of prophylactic treatments occurred on just four of the farms. In multivariable models, the number and probability of treatments of tanks of fish were significantly associated with fish size (negative) and their time-at-risk (positive). This confirmed the need to control for fish size, and resolved the question of whether treatment frequency measures in aquaculture must incorporate time-at-risk. Fixed farm effects were also statistically significant. Unconfounded by fish size, morbidity varied significantly with time-of-year (lowest in September though November). Morbidity or treatment incidence rates have not previously been reported for land-based fish farming. An incidence density measure was developed, to quantify chemotherapeutic treatment incidence at the tank level. The risk of treatment of fish was reasonably constant over time within each of four fish size categories. Therefore, treatment incidence was summarized within each of those size categories.