Estradiol (E2), testosterone (T) and progesterone (P4) concentrations were determined by enzyme-immunoassay in aqueous extracts of fecal samples obtained during anestrus, proestrus, estrus and metestrus of 11 nonpregnant and 11 pregnant bitches. Fecal hormone concentrations (ng/g) changed in relation to stage of cycle. Mean fecal steroid concentrations in 22 anestrous bitches and 3 ovariectomized bitches were low and similar for E2 (53 +/- 5 and 27 +/- 2), T (60 +/- 7 and 36 +/- 6), and P4 (62 +/- 6 and 86 +/- 15). Within 0 to 3 d of the ovulatory LH surge fecal E2 reached peak concentrations (301 +/- 38). The T peaks (281 +/- 41) were coincident or 1 to 3 d later. Fecal P4 was then elevated for approximately 2 m.o. Between Days 26 and 45 after ovulation, mean fecal P4 concentrations were higher (P < 0.05) in pregnant (401 +/- 60) than in nonpregnant bitches (164 +/- 23) and peak fecal P4 concentrations in individual animals were higher (P < 0.01) in pregnant (812 +/- 121) than in nonpregnant bitches (425 +/- 97). In the same period mean concentrations of E2 (117 +/- 13 vs 61 +/- 5) and T (102 +/- 10 vs 70 +/- 6) were also higher (P < or = 0.05) in pregnant than in nonpregnant bitches. Serum E2, T and P4 concentration were positively correlated (P = 0.1) with concentration in fecal samples obtained one day after serum collection. Although serial fecal ovarian steroid concentrations demonstrate the time course of ovulatory cycles, the diagnostic value of individual fecal samples appears limited. The ratios of peak to basal values were approximately 6, 5 and 7 for E2, T and P4, respectively, and were considerably lower than ratios of 12 to 50 previously reported for serum or plasma concentrations. The results demonstrate that there are pregnancy-specific increases in P4, E2 and T production reflected in fecal concentrations. While such increases are reflected in fecal samples, they are generally not evident in serum or plasma concentrations because of increased hemodilution, metabolism and clearance in pregnant bitches. The physiological stimulus for these increases, presumably ovarian in origin, or the potential role of prolactin is not known.