Northern Prairie Wildlife Research Center
Foraging canvasbacks were collected on semi-permanent and permanent wetlands from late April to early July, 1977-81. Most birds were observed foraging for at least 10 minutes before collection. Esophageal contents were removed immediately and preserved in 75% alcohol (Swanson and Bartonek 1970). The volume of each food was determined by water displacement and results were expressed as percent occurrence and mean percent volume (Swanson et al. 1974). Fragments of snail shells were excluded from calculations.
Diets of females were evaluated relative to their stage of reproduction. Reproductive periods were categorized by the female's ovarian condition (Hohman 1985, Barzen and Serie 1990): (1) prenesting from arrival until the beginning of rapid follicle growth (largest follicle diameter less than 7.5 mm); (2) rapid follicle growth (RFG) during ovarian follicle growth, largest follicle at least 7.5 mm and no postovulatory follicles present; (2) laying postovulatory follicles present and at least one developing follicle in the ovary of oviduct; (3) incubation brood patch present and follicles regressed; and (4) renesting brood patch and postovulatory enlarged follicles evident and follicles developing (largest follicle at least 7.5 mm). Diets of breeding males were pooled across the breeding season. Because the number of birds collected from each reproductive period was limited, we pooled data from all years.
We used multiple analysis of variance to compare mean percent volumes of plant and animal foods of male and female canvasbacks and, for females, among reproductive periods. Percentages were transformed (arcsine squareroot) before analysis to improve normality and equality of variances.