Northern Prairie Wildlife Research Center
We found that reproductive performance of canvasbacks varied widely among years. Age related differences were evident between SY and ASY females but few could be detected between TY and ATY females due to the small sample sizes. Low water conditions and poor nesting success in 1977 resulted in a missing age cohort, which affected our ability to detect age-related differences in later years. Older female canvasbacks tended to arrive earlier than SY females, particularly during the warmest spring, which agrees with previous findings that cool spring temperatures delay arrival times in other prairie-nesting ducks (Langford and Driver 1979, Hammond and Johnson 1984, Afton 1984, Cowardin et al. 1985). Additionally, our finding that nest initiation date was influenced by April temperatures and female age, and further by drought in 1977, also agrees with previous studies (Rogers 1964, Stoudt 1982, Krapu et al. 1983, Afton 1984). Delayed nesting by SY females has been documented for other waterfowl (Krapu and Doty 1979, Afton 1984, Cowardin et al. 1985, Lokemoen et al. 1990) and may be a function of experience and/or physiological conditioning for nesting (Barzen and Serie 1990). Annual variation in nest initiation may be partially related to variation in arrival dates (Afton 1984).
Clutch size in our study was similar to, or slightly higher than, other estimates (Olson 1964; Smith 1971; Stoudt 1971, 1982). Variation in clutch size may result from differences in nesting chronology or from variable effort in finding eggs outside of the nest bowl. In our study, clutch size did not vary among years despite marked changes in water conditions. However, the seasonal decline in clutch size for canvasbacks was more rapid over a 10-day period than that reported for mallards, gadwalls (Anas strepera), or blue-winged teal (A. discors) (Cowardin et al. 1985, Lokemoen et al. 1990).
Interspecific nest parasitism rates in our study did not seem to affect nesting success, although they were similar to other estimates for canvasbacks (Olson 1964, Sugden 1978, Stoudt 1982). Other studies found that parasitism reduced the host's clutch size, hatching success, and nest success (Erickson 1948, Weller 1959, Joyner 1976, Sugden 1980, Bouffard 1983). Weller (1959) suggested that 60-70% of the nests must be parasitized with 4-6 eggs/nest before egg and nest losses become measurable. We found that nest parasitism by redheads approached these levels, but most nests had < 4 parasitic eggs, and nest success was independent of the number of parasitic eggs in the host nest. We found that mean hatching success of canvasback eggs varied by 10% between parasitized and unparasitized nests, but because of the extreme variability in our samples, we could not detect any significant effects.
Although we excluded 4 nests with evidence of intraspecific parasitism from our analyses, nest parasitism by other canvasbacks was largely undetected and may have inflated clutch size data in some cases. Sorenson (1990) found that at least 36% of canvasback nests on our study area during 1986-88 were parasitized by other canvasbacks. He found that higher rates of intraspecific parasitism were associated with the drought conditions in 1988, and a high proportion of yearlings in the population during 1987 and 1988. Thus, intraspecific parasitism may be more common when numbers of young females are high relative to available nesting habitat.
Nest success seems to be influenced primarily by predation and environmental factors each year. Although differences among ages were not significant, nest success tended to be highest in ATY females. This was most apparent in 1978 and 1979 when habitat conditions may have been least limiting due to improved water conditions (success rate of ATY females was 13-44% higher than that of younger birds in the same year). Most waterfowl studies have found no age-related differences in nest success (Johnson 1978, Baillie and Milne 1982, Cowardin et al. 1985), although nest success in lesser scaup (Aythya affinis) (Afton 1984) and wood ducks (Aix sponsa) (Hansen 1971) was lowest in young or inexperienced females. Nest success in waterfowl has been related primarily to water or habitat conditions and predation (Rogers 1964; Smith 1971; Stoudt 1971, 1982; Cowardin et al. 1985). We found that predation rates were higher during dry years, when access to over-water nests was greatest. Hammond and Johnson (1984), reported that weather effects, such as low temperatures or heavy precipitation, likely confound variable predation rates. While we found no direct effects of May pond numbers or seasonal decline in wetland numbers on nest success, the high number of nest failures we attributed to predation leads us to believe that it is probably the primary extrinsic factor influencing nest success. Mammalian predation is the primary factor affecting success of waterfowl nests in the pothole region (Stoudt 1971, Klett et al. 1988, this study), and Stoudt (1982) implicated raccoons (Procyon lotor) as the major mammalian predator of canvasback nests.
Hen success is a function of both nest success and renesting rate and, thus, is the principal variable governing waterfowl reproductive performance (Cowardin et al. 1985). We found higher hen success among older females than SY females, which was likely a function of renesting effort since no difference was found in nest success among age groups. However, few studies have reported a measure of hen success or examined individual reproductive differences among females relative to age. In our study, hen success provided a better measure of reproductive success among age groups than any other measure because it was determined using marked females with known nest histories or from brood observations. Nest success was probably over-estimated because it was determined only from nests of marked birds and did not include nests found abandoned, destroyed, or hatched. Thus, hen success better reflected an individual's contribution to annual productivity by accounting for renesting, particularly in years when habitat conditions were favorable.
Observations in our study indicated that in years of moderate to good wetland conditions, female canvasbacks of all ages nested. However, capabilities to renest may be the critical factor affecting annual productivity in dry or moderately wet years and seems to be strongly age-related. During years of poor wetland conditions, as in 1977 and 1980, nesting was reduced and no SY females nested. Most known renesting efforts were by ATY females, and renests by SY females were rarely found in any year. Other studies of age-related productivity in waterfowl have reported similar age-related differences (Mihelsons et al. 1970, Afton 1984, Cowardin et al. 1985, Lokemoen et al. 1990), which may be a function of experience or the physiological capability of young females to renest quickly following the loss of a nest (Barzen and Serie 1990).
Observations of our returning marked females suggest that they often use the same feeding and nesting ponds each year. Spring wetland conditions did not affect return rates of any age class, although in dry years, fewer birds overall settled in their respective breeding areas. In wet to moderately wet years, we found similar return rates for TY and ATY females. Low return rates for TY females from the 1974 age class (40%) were probably due to the small number of marked birds available; whereas low returns of TY females from the 1975 cohort were likely influenced by the drought in 1977.
As in gadwalls and blue-winged teal, but unlike mallards (Lokemoen et al. 1990), return rates of canvasbacks were not affected by the success of their first year's nesting effort. The effect of previous nest success on return rates of older birds was marginal and less than that reported for mallards or gadwalls (Doty and Lee 1974, Lokemoen et al. 1990). Thus, the dominant factor influencing return rates of canvasback females was age. Lower return rates for young females also were found in other studies of canvasbacks (Anderson 1985), redheads (Johnson 1978), lesser scaup (A. D. Afton, in Johnson and Grier 1988), mallards (Doty and Lee 1974, Bishop et al. 1978, Lokemoen et al. 1990), and gadwalls (Blohm 1979, Lokemoen et al. 1990) and may reflect lower survival rates for birds in their first year of life.
Return rates from our study, when compared with previous estimates of survival rates, seem to agree closely and give further support to the idea of strong breeding philopatry by canvasbacks. The highest survival rate estimate for adult females marked on the Chesapeake Bay wintering grounds (Nichols and Haramis 1980) is similar to the average return rates of adult females in this study (69% vs. 64% ), indicating that most, if not all, surviving adult females return to their natal area. The average return rates of juveniles as SY females is similar to an early survival estimate of immatures from the prairie region of Canada (Geis 1959). A small proportion of SY females may breed elsewhere in their first year or two, or remain on other areas until returning to their natal area for their first breeding effort, as indicated by the few indirect or delayed returns. Although Johnson and Grier (1988) found that canvasbacks seem to occupy habitat as it is encountered, little direct evidence shows pioneering by canvasbacks.