Reproductive success of exotic mute swans in Connecticut

Mute Swans (Cygnus olor) are native to Eurasia and were introduced to North America, where free-ranging populations have existed since the early 1900s (Allin et al. 1987). Originally observed in New York, breeding populations of Mute Swans have expanded into all of the Atlantic coastal states from Maine to Maryland and also have become established in Michigan, Minnesota, Wisconsin, British Columbia, and Ontario (Willey and Halla 1972; Reese 1975, 1980; Allin et al. 1987; Knapton 1993). Allin et al. (1987) synthesized 33 years of population data on Mute Swans from the Atlantic Flyway and estimated a 1987 population of 5,300, with most of the birds in New York, Connecticut, Rhode Island, and New Jersey. In the southern New England states of Massachusetts, Rhode Island, and Connecticut, Mute Swans constitute a single population with considerable interchange of individuals among states (Willey and Halla 1972).

During the 1960s and 1970s, the Chesapeake Bay population nested mostly on estuaries and tidal rivers (Reese 1975, 1980), as did the southern New England population (Willey and Halla 1972, Kania and Smith 1986). For instance, 90% of the Mute Swans in Rhode Island in 1967 nested on estuaries, and no nests were more than 5 km inland (Willey and Halla 1972). Since then, large numbers of Mute Swans in southern New England have started breeding at inland sites on lakes and ponds.

Expanding Mute Swan populations may have detrimental effects on native biota. One concern is the influence of swan herbivory on aquatic vegetation (Willey and Halla 1972, Reese 1975, Allin et al. 1987, Conover and Kania 1994) and associated macro-invertebrates (Krull 1970); another is that Mute Swans are aggressive toward other waterfowl species (Willey and Halla 1972, Kania and Smith 1986, Allin et al. 1987, Conover and Kania 1994). Because many native species of waterfowl nest at inland sites, Mute Swans may pose more of a threat to native avifauna at inland sites than at estuaries and tidal rivers (Willey and Halla 1972, Kania and Smith 1986, Conover and Kania 1994).

The present study was conducted to measure reproductive success of the portion of the Mute Swan population that nests in Connecticut and to determine whether Mute Swans nesting on lakes and ponds are as successful as those nesting on estuaries and tidal rivers. Despite the increase in Mute Swan numbers in North America, little is known about their reproductive biology. Reese (1975, 1980) recorded population levels and reproductive success of Mute Swans on Chesapeake Bay from 1968 to 1978. To our knowledge, the only subsequent study of Mute Swans in North America was conducted in southern Ontario by Knapton (1993).

Methods.-Mute Swan nesting data were collected in Connecticut from 1982 to 1990. The study area extended from the Housatonic River to the Connecticut River and included most of the nesting Mute Swans in Connecticut. Population trends were monitored using data from Audubon Christmas Bird Counts in Connecticut and from midwinter waterfowl counts conducted by the Connecticut Department of Environmental Protection using observers in fixed-wing aircraft.

Nests were located during coastal flight surveys conducted each spring by the Connecticut Department of Environmental Protection. Mute Swans nests are readily identifiable from the air owing to the large size and bright white plumage of adult swans. Additional nests were located by ground and boat surveys and by reports from local citizens. We believe that coverage of the study area was complete and that nearly all nests were located.

For each nest, we counted the number of eggs or cygnets present and recorded the presence or absence of parents. Most nests were visited every two weeks. Each nest was classified into one of two habitat types: lakes and ponds (hereafter "lakes") or tidal rivers and estuaries (hereafter "estuaries"). We defined clutch size as the maximum number of eggs observed in a nest and brood size as the maximum number of cygnets observed per swan pair. The maximum number of cygnets was always observed on the first visit after all cygnets hatched (i.e. we observed no adoptions, "gang" brooding, or creching). Cygnets that survived to 1 September were considered to have fledged.

Nests were incubated constantly; hence, the absence of an adult at the nest indicated that a nest had terminated. To determine a nest's fate, we located the parents in their territory as soon as incubation ceased and checked for the presence of cygnets. We confirmed nest failures by checking the nest to see if any eggs had hatched. We distinguished between hatched and depredated eggs by the presence of detached shell membranes (Klett et al. 1986).

We estimated nesting success, defined as the probability of a nest surviving the incubation period to produce at least one cygnet (Johnson et al. 1992), using both the "apparent" (proportion of nesting attempts that were successful) and the Mayfield (Mayfield 1961, 1975) methods. The Mayfield method is based on estimating daily survival rate (DSR) during the time nests are exposed to risk and ascertaining the outcome of the attempt during the observation period. In most studies, this observation interval (exposure days) for successful nests is the number of days between discovery of the nest and the estimated hatching date. The latter is determined by estimating the age of eggs when found and subtracting this value from the mean number of days in the laying and incubation periods for the species in question (Klett et al. 1986).