Response of Breeding Canvasbacks to Nest-fencing and a Reduction
in Mammalian Egg Predation
MICHAEL G. ANDERSON AND ROBERT B. EMERY
Duck Unlimited Canada, 1190 Waverley Street, Winnipeg, MB R3T 2E2,
Canada; Delta Waterfowl and Wetlands Research Station, Route 1,
Portage la Prairie, MB RIN 3A1, Canada
Breeding canvasbacks in the Minnedosa pothole country of southwestern Manitoba have been surveyed in a similar manner most springs since the early 1950s. These counts revealed marked fluctuations in canvasback numbers among years, but no long-term trends. Three times the population approached but never exceeded 4.2 pairs per km². Because of strong female homing, and relatively constant apparent annual survival, we suspected that these population changes resulted from year-to-year variation in nesting success, and hence, recruitment of locally reared young to the breeding population. Nest success varied greatly among years, correlated with changes in water levels. To test the idea that local nesting success limits the size of the Minnedosa canvasback population (and to address a number of related hypotheses), we selected two study blocks, each about 16 km² in area, with similar wetlands and equivalent numbers of breeding canvasbacks. Beginning in 1983, one block was treated as an experimental area where predation was reduced by erecting temporary fences around canvasback nests in flooded emergent vegetation. The other block was monitored as a control where nothing was done to enhance nesting success. Nest fences were approximately 20 m long and 1.2 m high, made of 5-cm mesh stucco wire and anchored in position with six or seven pieces of 1.2 cm conduit pipe. Fences were set out in a "U" shape to screen nests from the land side yet allow the hens swimming access to their nests from the water side. In addition to fencing nests, some canvasback clutches nearing hatch with less than five canvasback eggs (mostly due to intrusions by parasitic redheads) were supplemented with eggs collected off the study areas.
Canvasback production on the experimental block improved significantly and remained high until declining during the severe drought of 1988. On average, 221 canvasback ducklings were hatched on the experimental area each summer versus 112 on the control area. Apparent nesting success for fenced and unfenced nests, for all years combined, were 77% and 50%, respectively. In 1984-85, this difference was 88% versus 23%; in later years there was little difference between fenced and unfenced nests. We speculate that fences reduced the likelihood of predators locating nests by deflecting searching animals away from nest sites. Nest survival was greater when several meters of thick cover was present between the fence and the nest. We suspect that fences did little to deter predation if predators detected a nest or a hen. By 1988, the experimental population had grown to greater than 9.6 pairs per km², up 163% from 1983, and more than doubled the highest density of breeding canvasbacks recorded at Minnedosa in 33 years. Populations in the control area fluctuated, as expected, following variation in nesting success. Clearly, recruitment of young, as influenced by nest success and redhead parasitism, is a major factor affecting numbers of canvasbacks at Minnedosa. If duckling production can be improved it appears that young female canvasbacks will home and local population densities can be increased.
Previous Section -- Predator Management in Western Minnesota
Return to Contents
Next Section -- Effects of Skunk Control on Duck Nest Success in Western Montana
