USGS Home
Contact USGS
Search USGS

Recaptures and resightings provide an indication of homing (returning to a breeding ground used the previous year). The statistic generally computed from such data is often called "return rate," the proportion of marked birds captured and released on an area in year t that are recaptured or released on that same area in year t+1. The return rate incorporates three probabilities: 1) the annual probability of survival; 2) the conditional probability of returning to the area of banding, given that the bird is alive the year following banding (homing); and 3) the conditional probability that the bird is recaptured or resighted, given that it returned to the area (resighting). Return rates can be used to draw tentative inferences about homing, but survival and resighting are also incorporated into this statistic.

For hunted species in North America, it should be possible to directly estimate a homing probability unconfounded by survival and resighting probabilities. Survival rate estimates based on band recovery data should reflect "true" survival. Capture-recapture data and models, on the other hand, yield survival estimates (Jolly 1965; Seber 1965, 1982) that reflect a product of both homing and "true" survival probabilities. If both types of survival estimates (band recovery and capture-recapture) can be obtained for annual banded samples in a particular location, then an ad hoc estimate of homing can be obtained. Based on this approach the probability of Pintails returning to a molting marsh was estimated at about 0.8 (Anderson and Sterling 1974), and the homing rate of female Wood Ducks to a breeding area was estimated to be nearly 1.0 (Hepp, Hoppe, and Kennamer 1987).

1. Species

Hochbaum (1946) noted that dabbling ducks (exemplified by Mallard and Pintail) pioneered new breeding areas rapidly, whereas diving ducks (such as Canvasback and Redhead) did so much more slowly. Johnson and Grier (1988) tabulated return rates estimated for several species of ducks, but comparison of such rates is questionable, because they also reflect recapture or resighting probabilities and the size of the area, which likely varied substantially from study to study. We note three studies that provided interspecific comparisons.

Lokemoen, Duebbert, and Sharp (1990) estimated homing rates (computed by dividing return rate by estimated survival rate) of female Mallards, Gadwalls, and Blue-winged Teal on two North Dakota study areas. Their weighted average homing rates were relatively high for Gadwalls (0.62) and Mallards (0.49), but low for Blue-winged Teal (0.07). Sowls (1955) found that return rates of Gadwalls (0.37), Northern Pintails (0.39), and Northern Shovelers (0.42) were relatively high, whereas those of Mallards (0.13) and Blue-winged Teal (0.14) were low. Coulter and Miller (1968) marked adult female ducks in Maine and Vermont and reported return rates of 0.42 for Mallards, 0.25 for Black Ducks, and 0.30 for Ring-necked Ducks. Savard and Eadie (1989) reported similar return rates to a study area by Barrow Goldeneyes (0.59) and Common Goldeneyes (0.56).

Johnson and Grier (1988) used duck and pond count data and found substantial variation in the degree to which 10 different species of ducks exhibited opportunistic responses to variable wetland conditions. A high degree of opportunism and pioneering was found for Mallards, Pintails, Blue-winged Teal, and Northern Shovelers. Lower pioneering was found for American Wigeon, Gadwalls, Green-winged Teal, Canvasbacks, Redheads, and Lesser Scaup.

2. Age and Breeding Experience

The few studies of return rate in male ducks show higher rates among older individuals (Blohm 1978 for Gadwalls, Poston 1974 for Northern Shovelers, and Anderson 1985 for Canvasbacks).

Previous breeding and nesting experience may be an important determinant of the probability that a bird will home to a breeding area. The success of the previous breeding effort appears to influence homing in females. Doty and Lee (1974) observed a return rate of 0.52 for female Mallards that nested successfully in the year of marking versus 0.16 for those that were unsuccessful. Mallard and Gadwall (but not Blue-winged Teal) hens that were successful the previous year showed higher return rates, and estimated homing rates than unsuccessful hens (Lokemoen, Duebbert, and Sharp 1990). Mihelsons, Mednis, and Blums (1986) found that adult female Northern Shovelers and Tufted Ducks in Latvia that lost their clutch or brood in the year of capture were less likely to return than successful females. Bellrose, Johnson, and Meyers (1964) noted that adult female Wood Ducks usually returned to areas where they last nested successfully. They also suggested that yearling Wood Duck hens and adult females that were unsuccessful in the previous season were most likely to pioneer new habitat.

3. Sex and Pair Status

Does homing, resighting, or survival contribute most to the above differences between males and females? Any sex-specific differences in sighting probabilities would favor the more conspicuous males. Survival rates are either similar for the two sexes (in young ducks) or higher for males (in adult ducks), so the higher return rates for females certainly can not be attributed to higher survival. The cited differences in return rates likely reflect a greater probability of homing by females than males.

Anderson and Henny (1972) examined recoveries of Mallards banded in eastern or western breeding areas but recovered in central prairie and north-central breeding areas. Consistently larger proportions of males than females were recovered in breeding areas other than those in which they were banded, again suggesting less homing by male than female ducks.

In contrast to ducks, family groups of most North American geese generally remain together for nearly a year (Mayr 1942, Bellrose 1980). Pair bonds are thought to be maintained for life (Bellrose 1980), with both members of the pair homing to the natal area of the female to breed (Cooke, MacInnes, and Prevett 1975; Cooke and Sulzbach 1978). A male Snow Goose that loses his mate will likely pair anew on the wintering grounds or during spring migration, possibly with a female from another breeding colony. Yearling Canada Geese often return to their natal area with the family group and then disperse after arrival on the breeding grounds (Sherwood 1967, Bellrose 1980). Sherwood (1967) reported that both one- and two-year-old females tended to remain near their natal area, whereas yearling and older unpaired males tended to disperse. Surrendi (1970) transplanted juvenile Canada Geese and found that a higher proportion of females than males returned to the release area.

Rohwer and Anderson (1988) related the differences among species in philopatry between males and females to the nature of pair bonds and the time of formation. Species with long pair bonds tend to have more similar homing rates between sexes than those that pair seasonally.

4. Wetland Habitat and Population Density

Johnson and Grier (1988) provided a conceptual framework for considering homing, pioneering, and their relation to wetland conditions. They defined three strategies: 1) a pure homing strategy in which birds simply return to the natal area to breed each year; 2) a purely opportunistic strategy in which birds breed at the first suitable (in terms of water conditions and other habitat requirements) location encountered during spring migration; and 3) a mixed adaptive strategy in which birds home to the breeding area used the previous year but move to other areas if current-year habitat conditions are unfavorable. Johnson and Grier (1988) used aerial survey data on both ducks and ponds and concluded that homing was most prevalent among Redheads, Canvasbacks, Mallards, Gadwalls, Northern Shovelers, and Lesser Scaup. Although all species exhibited some features of an opportunistic strategy, such features were especially notable in Mallards, Pintails, Blue-winged Teal, and Northern Shovelers, all of which tend to use temporary wetlands to some extent. Canvasbacks, Redheads, and scaup use more stable wetlands and exhibited less covariation with pond numbers. The adaptive strategy, as indicated by overflights in dry years, was exhibited by most species.

The role of density in influencing settling patterns of breeding waterfowl is not yet clear. Density is presumed to exert an important influence on animal location in most hypotheses dealing with habitat selection (e.g., Fretwell 1972). Dzubin (1969) observed that Mallard pairs arriving on the breeding grounds at ponds occupied by another pair of Mallards often dispersed to other nearby ponds and noted that high Mallard densities can result in some pairs making long-distance emigrations to new habitats. Pospahala, Anderson, and Henny (1974) believed that such density-dependent movement of pairs, together with differences in reproductive output associated with favorable versus unfavorable breeding habitats, played an important role in the regulation of North American Mallard populations. Nonetheless, extremely high breeding populations have been reported, usually in areas with high nest success.