Three generations of mallard ducks (Anas platyrhynchos) were fed either a control diet or a diet containing 0.5 ppm mercury in the form of methylmercury. The levels of mercury in adult tissues and eggs remained about the same over 3 generations. The methylmercury diet had no effect on adult weights or weight changes during the reproductive season. Females fed a diet containing 0.5 ppm mercury laid a greater percentage of their eggs outside their nestboxes than did controls, and also laid fewer eggs and produced fewer ducklings. Methylmercury in the diet appeared to result in a small amount of eggshell thinning. Ducklings from parents fed methylmercury were less responsive than controls to tape-recorded maternal calls, but were hyper-responsive to a frightening stimulus in avoidance tests; there were no significant differences in locomotor activity in an open-field test. J. WILDL. MANAGE. 43(2):394-401 The purpose of this paper is to describe reproductive and behavioral effects of methylmercury on 3 generations of mallard ducks. Findings for the 1st and 2nd generations were reported earlier (Heinz 1974, 1975, 1976b, 1976c, Heinz and Locke 1976). Here, I have combined the data for all 3 generations to determine changes from generation to generation and overall effects that might not have been statistically significant in the analysis of single generations. To my knowledge, there have been no reported multiple generation studies on the effects of mercury or any other environmental pollutant on waterfowl. Such studies have been reported for Japanese quail (Coturnix coturnix japonica) (Carnio and McQueen 1973), ring-necked pheasants (Phasianus colchicus) (Dahlgren and Linder 1974), and ring doves (Streptopelia risoria) (Peakall et al. 1972). I thank the following people for their help in the study: C. S. Cruger, R. B. Frederick, D. C. Gray, R. G. Heath, P. A. Heinz, G. Hensler, and N. C. Miller. J. B. Elder provided the Morsodren used in this study and others. METHODS An abbreviated description of the care of adults, care of eggs and young, and behavioral tests is given below. Additional details were reported previously (Heinz 1974, 1976b, 1976c). A 1-way analysis of variance was used to check for changes in mercury levels in tissues from generation to generation. Reproductive and behavioral comparisons involved 2 factors (treatment and generation); these comparisons were made with a 2-way analysis of variance. A significance level of 0.05 was used. Generation effects are not discussed because significant differences from 1 year to another, if any, would not be related to mercury treatment but to other factors such as weather.