Reproductive endocrine patterns were characterized in female ocelots (Leopardus pardalis; n = 3), tigrinas (Leopardus tigrinus; n = 2), and margays (Leopardus wiedii; n = 2) housed in captivity in southern Brazil. Females were maintained as singletons and exposed to natural fluctuations in photoperiod. Cyclic changes in ovarian steroids were monitored by analyzing estrogen and progestogen metabolites in fecal samples collected five times weekly for 14 to 18 months. Based on intervals between fecal estrogen peaks, mean (± SEM) duration of the estrous cycle was 18.4 ± 1.6 days for the ocelots (range, 7–31 days; n = 75 cycles), 16.7 ± 1.3 days for the tigrinas (range, 11–27 days; n = 23 cycles), and 17.6 ± 1.5 days for the margays (range, 11–25 days; n = 32 cycles). Fecal progestogen analyses combined with two laparoscopic observations of the ovaries confirmed that ocelots and tigrinas did not ovulate spontaneously. In contrast, non-mating–induced luteal phases of 40.1 ± 6.3 days in duration (range, 30–60 days) were observed frequently in both margays. There was no evidence of gonadal seasonality in margays in either follicular or luteal activity. In ocelots, cyclic changes in estrogen excretion were observed during each month of the year; however, only one female cycled continuously. In the other two ocelots, periods of acyclicity of several months’ duration were observed. It was not possible to conclude whether tigrinas were aseasonal because estrous cyclicity was observed in only one of two individuals. In the female that cycled, a 3-month period of acyclicity was observed in the late fall/early winter. These data demonstrate similarities among three felid species of the genus Leopardus, including evidence they are polyestrous but experience unexplained periods of ovarian inactivity. Only the margays differed by exhibiting occasional spontaneous, non-mating–induced ovulations. Historically, these species have not bred well in captivity. However, it is hoped that understanding the biological similarities and differences among them could lead to improved management strategies that may one day result in increased reproductive success. Zoo Biol 20:103–116, 2001. © 2001 Wiley-Liss, Inc.