Northern Prairie Wildlife Research Center
Canvasbacks, unlike mallards, are almost exclusively overwater nesters and thereby suffer little predation by red foxes, which are reluctant to enter water (A. B. Sargeant, unpublished data). Canvasbacks are, on the other hand, exposed to marsh-dwelling predators, particularly the mink. Eberhardt (1974) summarized the few available data on mink predation on adult ducks on prairie marshes and found that mink commonly take diving ducks, mostly females (78%-- Table 18). Although data are inadequate to ascertain how substantial such losses are, the mink is widely distributed and generally common to abundant throughout the canvasback breeding range. We do not believe that predation by mink on canvasbacks is of the magnitude exerted by red foxes on mallards, but it is undoubtedly widespread and likely selective for females. Other predators are probably less effective than mink and all together have nearly equal impact on males and females. Canvasbacks, like mallards, appear to suffer higher predation losses among females than among males.
Of particular importance is that the sex balance of canvasback populations does not tend to be restored by hunting. Indeed, the reverse appears to be true. Gels (1959) reported that hunting mortality on females was more severe than on males. His conclusion was based on an analysis of recovery rates of birds banded during winter in four States. Olson (1965) gathered experimental and observational evidence that suggested canvasback females and juveniles are decidedly more vulnerable to hunting over decoys and on passes than are adult males.
The consequence to canvasback populations of females suffering more predation and hunting mortality than males can be seen by referring to Fig. 4, which shows how the sex ratio varies as a function of the ratio of male to female survival rates and recruitment. In mallards, for which hunting mortality tends to compensate for sex-related differences in predation, the ratio of the survival rates was calculated in Part One as P = 1.1. In light of the information that predation and hunting are both female-selective, the ratio in canvasbacks is likely to be much higher. Suppose as an example that p= 1.3, i.e., males survive at a rate 1.3 times that of females or the female survival rate is only 77% of the male rate. Under such an assumption, the asymptotic sex ratio climbs (Fig. 4) and, moreover, becomes strongly influenced by the productivity of the population. At P= 1.3, the asymptotic sex ratio is 186:100 if recruitment to the spring population is 1.0 young of each sex per adult female, a reasonable level in a stable population. If recruitment is only 0.6 young of each sex per female, however, the asymptotic sex ratio is 260:100, while recruitment of 1.4 leads to a sex ratio of 165:100. Thus, on the basis of the above assumption we would anticipate the sex ratio in canvasbacks to be much more distorted than in mallards, particularly if recruitment is low.
Such indeed is true. Canvasback populations are thought to contain about 65-70% males (e.g., Trauger 1974), which implies sex ratios of 186-233 males per 100 females. Olson ( 1965: 128) suggested, "Differences in vulnerability to hunting may be the cause of an imbalanced sex ratio in canvasback ducks." While the statement is true, a highly distorted sex ratio could not result solely from differential hunting mortality of the magnitude suggested for this species, but it could result from female-selective hunting mortality, combined with female-selective mortality due to predation, particularly in years of low productivity.
It is not clear how far these results can be generalized. Certainly duck species in biotic areas other than the Prairie Pothole Region may be subjected to a completely different array of mortality factors. To illustrate how dissimilar the situation can be, we refer to a recent paper (Belopol'skii et al. 1975) on the sex ratio in the eider (presumably Somateria mollissima) in the Soviet Union. Of 77 embryos examined in late stages of development, the authors concluded that 62.4% were males, a statistically significant departure from equality. In contrast, the sex ratio in mature eiders was very nearly balanced. The authors attributed the change between the secondary sex ratio and the quaternary ratio to heightened mortality of males resulting from greater loss to predators, hunters, and perhaps parasites. The situation in this species thus seems to be the complete opposite of that in the mallard and canvasback, in which the sex ratio changes from nearly even at hatching toward a preponderance of males among adults. We cite the example only to caution against careless overgeneralization of our results and interpretations.