Studies were carried out in Manitoba, Canada, to evaluate the efficacy of three repellent products for protection of human subjects against mosquito bites. All test substances contained the active ingredient N,N-diethyl-3-methylbenzamide (deet); two were polymer-based creams (3M Ultrathon Insect Repellent and Morflex DEET Insect Repellent 30) and the third (Muskol Insect Repellent) was an alcohol-based pump spray formulation. Application of repellent was to the forearm and lower legs of subjects at 0.75 or 0.83 mg deet/cm2. Exposure to mosquito attack was continuous, and efficacy was determined by measuring complete protection time (CPT). Regardless of whether delivered as a polymer cream or in alcohol, mean CPT was similar for the tested repellents at 623 +/- 107 to 644 +/- 163 min. By contrast, mean CPT for the different test subjects showed significant variation, ranging from 531 +/- 42 to 756 +/- 54 min. Mosquito collections from untreated human test subjects, Centers for Disease Control and Prevention (CDC) light traps and sweep-netting consisted primarily of Ochlerotatus sticticus (Meigen) and Aedes vexans (Meigen). Relative catch of these two species was similar for different sampling methods through much of the day, but not in the evening, when CDC light traps oversampled Ae. vexans relative to untreated human subjects. Results are used to highlight the need to account for intersubject variation when designing repellent studies, and also are used as a basis to discuss limitations associated with using relatively few subjects when testing repellents.