Many techniques commonly used by herpetologists for monitoring amphibian populations and communities yield censuses of the total adult population size (N). However, for many studies, e.g., of reproductive output, development of populations and potential for evolutionary changes, the effective population size (N_e) must be known. While modern molecular techniques make it possible to measure N_e, they are expensive, work-intensive and may not be possible or necessary for many questions on amphibian reproduction. For females of two species of water frog (Rana lessonae and R. esculenta), we investigate the effectiveness of several techniques to determine the presence or absence of eggs. The direct methods are (1) dissection of dead frogs and (2) a small skin incision into the abdominal side of live females. The indirect methods, all applied to live frogs, include (3) visual inspection of body shape, (4) tactile inspection of the epidermis, (5) transillumination with a strong cold light source, (6) ultrasound, (7) electromagnetic measurement of total body electrical conductivity (TOBEC), (8) calculation of body condition index and (9) analysis of blood plasma testosterone titers. Only two indirect methods were somewhat successful at predicting whether females were gravid. Testosterone titers (9) yielded the best results (ca. 80–90% effective). Body condition (8) was also significantly related to egg presence or absence, but predicted gravidity only weakly. We suggest that a combination of skin incision and hormone analysis provides a fairly good estimate of gravidity. When complemented by mark-recapture techniques and performed on the same individuals at different times of the season, this combination yields estimates not only of the reproductive output of the study population but also of the relative contribution of different females.