Incubation constancy and mass loss in the Common Eider Somateria mollissima

Birds that raise young without the help of their mate must balance their own daily energy needs against energy expenditure on clutch or brood protection and incubation (Carey 1996). While leaving the nest to feed increases the probability of egg predation, incurs extra energy costs for locomotion and rewarming the eggs, and lengthens the incubation period, an increase in incubation constancy requires a greater use of endogenous reserves for daily energy expenditure. As a response to these conflicting needs, large waterfowl species have apparently developed fasting endurance, as they possess the capacity to accumulate significant fat reserves before nesting (Afton & Paulus 1992). In contrast, small-bodied waterfowl are apparently unable to store sufficient fat to sustain themselves through the incubation period, and therefore exhibit lower nest attentiveness than larger species because they must spend more time feeding (Afton & Paulus 1992). Female Common Eiders Somateria mollissima lose 23–36% of their body mass (1773–2300 g) during incubation (Korschgen 1977, Franzmann 1980, Parker & Holm 1990, Gabrielsen et al. 1991, Harðardóttir et al. 1997). They are believed to fast during incubation, which provides a plausible explanation for this substantial loss in body mass (Milne 1976, Korschgen 1977, Parker & Holm 1990). However, complete food deprivation has never been documented, and there is anecdotal evidence that eiders feed occasionally during incubation (Swennen et al. 1993). The extent of such feeding is unknown. Fasting endurance in the Common Eider is usually interpreted as an adaptation to minimize the risk of egg predation (Milne 1976, Korschgen 1977). Indeed, nesting success varies greatly in this species (25–90%), probably because eider nests sometimes suffer from heavy predation by diurnal predators such as large corvids and larids (Bolduc 1998, Bolduc & Guillemette 2003). Thus, the number and timing of incubation 'recesses' for drinking or feeding (day or night) may be of paramount importance to reduce the risk of egg predation in this species. Using electronic nest monitors, we investigated the incubation behaviour and body mass variation of Common Eider nesting on Saltholm Island, Denmark. The study was conducted on Saltholm Island (55°39′N, 12°45′E), Denmark, during spring 1996. The island is an EU Special Bird Protection Area of about 40 km2 where human access is restricted during the birds' breeding season. Grasses cover most of the island, and large mudflats and marshes surround the island. We recorded nest attentiveness using electronic nest monitors similar to those used by Gloutney et al. (1999). Nest (and contents) mass was recorded every minute and stored in a data logger. Departures of the incubating female corresponded to large decreases in weight of the nest. We used six nest monitors and installed them beneath the first eider nests found between 8 and 18 April. One of these nests was predated after 10 days, and another was destroyed when a cow stepped on the side of the device after 26 days of recording. The other four monitors recorded continuously until the eggs hatched. Results include data from all devices except the one predated. When monitors were installed, the nests contained one to three eggs; in two cases, nests were found unattended, without down, and covered by grass. The clutch-size was five for all five nests monitored, as verified by counting hatchlings or egg membranes in nests at the end of the study. At the removal of nest monitors, hatchlings with dried down were present in all nests except for one where hatchlings were still wet. Common Eiders lay one egg per day (Swennen et al. 1993). We therefore estimated the beginning of incubation (clutch-completion to hatching, Drent 1975) by adding the difference between the clutch-size and the number of eggs in nests when found, to the day the nest was located. Similarly, the laying-date of the first egg was calculated by subtracting the number of eggs present when found from the date the nest was located. Female body mass was determined by subtracting the nest-content mass during female departures (averaged over 30 min after departure) from that before departure (averaged over 30 min). Prior to installation, we calibrated monitors with a known mass. Units used by nest monitors were generally close to grams (one unit = 0.907–1.017 g). We corrected the data using the appropriate factor to obtain mass in grams for data analysis. We compared female mass before and after a recess to evaluate body mass variation during recesses. Because of potential bias caused by female movements and nest maintenance, we double-checked each mass gain and mass loss by also subtracting the mass before a recess from the mass before the next recess. Then, only those results leading to the same conclusion (both gains or both losses) were used to compute the extent of mass gain and mass loss during recesses. All statistics were computed using SAS v.6.12 (SAS Institute 1996), and results were declared significant at a P-value < 0.05. We used standard errors (se) to express sample variation around the mean, unless specified otherwise. Mean recess duration differed significantly between the laying and incubation periods (paired t-test, mean diff. = 342.7 min, df = 4, t-ratio = 2.99, P = 0.04) (Fig. 1). Recess duration during laying was 421.2 ± 52.1 min on average (range 4–1613 min), and tended to decrease as laying progressed (Fig. 1). During incubation, recesses lasted 14.2 ± 6.1 min on average (range 3–42 min), and their duration was unrelated to number of days into incubation (regression, R2 = 0.007, P = 0.69). Each incubation was generally characterized by two long periods without recess, which occurred about 10 and 20 days after the first egg was laid (Fig. 1). Each of these periods lasted about 4 days (range = 2.85–5.98 days) and 5.5 days (range = 4.01–7.95 days), respectively. Duration and timing of incubation recesses with occurrence of observed mass gain for five female Common Eiders monitored on Saltholm Island, Denmark, during the spring of 1996. Day 5 corresponds to the ending of laying and beginning of incubation. Note that day 5 finishes with its label along the x-axis graduation (midnight). Female order follows Table 1 order. Low-frequency, short-duration recesses yielded to high daily incubation constancy. During the last days of laying, female Common Eiders stayed on nests for 60–70% of the time on average. After laying was completed, incubation constancy increased rapidly to reach 95% and 98% on the first and second day of incubation, respectively. Thereafter, eiders spent 99.5% (± 0.2) of their time sitting on the nest (Table 1). Our results on Common Eider incubation constancy are among the highest known for waterfowl (Afton & Paulus 1992). For example, incubation constancy ranges between 80.3 and 84.5% in three species of diving ducks (Mergini), and between 79.0 and 89.2% in dabbling ducks (Anatini) (Afton & Paulus 1992). Only geese (Anserini) have an incubation constancy (81.7–99.5%) approaching those that we recorded for eiders. Daytime recesses occurred in similar proportion during laying (4/15, Fig. 1) as during incubation (8/61) (χ2 = 1.663, P = 0.20). However, recesses occurred in greater proportion during night-time than daytime (64/76) (G = 39.1, P < 0.01). Most of the daytime recesses that occurred during incubation related to one female only (6/8) (Fig. 1), that took no recess after 06:20 h. Sunrise occurred between 03:00 h and 04:20 h, and sunset between 18:00 h and 19:15 h during our study. It appeared that this 'daytime' female timed its recesses a little later after sunrise than the other four, which took their recesses between 20:00 h and 01:00 h (Fig. 1). Swennen et al. (1993) also observed that most recesses occurred at night, and so we interpret the night-time occurrence of incubation recesses as a behavioural adaptation to minimize egg predation by diurnal birds. Of 76 recesses recorded by nest monitors, 51 were associated with mass loss. In 13 cases it could not be determined whether eiders lost or gained mass (see Methods), and in 12 other cases, females gained mass (117 ± 75 g). Only two mass gains were recorded after day 10 following the onset of laying. Four females gained mass by day 10, which were always followed by several days sitting on the nest without leaving. Other researchers have noticed incubating Common Eiders feeding occasionally, but this has been considered rare (Campbell 1975, Swennen et al. 1993, Criscuolo et al. 2000). On another Danish colony (M. Guillemette pers. obs.), it was observed that females usually took flight from the colony and landed at sea where they drank immediately upon landing. Some females made unusually long dives and showed a distended neck on their way back to their nests suggesting that they had ingested food. Therefore, we suggest that the 12 mass gains (of 117 g on average) that we recorded were caused by the ingestion of food, and not only water. Nevertheless, the low frequency of mass gains observed in our study and the daily diving time usually required to maintain energy balance in this species (about 140 min, Guillemette 1998) indicate that ingestion of food is negligible during incubation in this species. Common Eider body mass decreased significantly with the number of days into incubation for all five females (regressions, all P < 0.001, R2 = 0.913–0.944). Rates of body mass loss averaged 37.5 g/day (range: 27.8–41.9 g/day, Table 1). An ancova (body mass = days in incubation + female + days * female) indicated that rate of mass loss did not differ among females (day-female term, df = 4, F = 2.26, P = 0.075), even though they differed in average body mass (female term, df = 4, F = 13.82, P < 0.001). This is the highest rate of body mass loss recorded in incubating Common Eiders (Fig. 2) or among waterfowl generally (see table 3.13 of Afton & Paulus 1992). For comparison, mass loss in Snow Geese Anser caerulescens and Canada Geese Branta canadensis ranges between 11.2 and 41.7 g/day during incubation. Rate of mass loss of incubating Common Eiders predicted by body mass at start of incubation for different populations. Source of data with subspecies: a, Gabrielsen et al. (1991), S. m. borealis; b, Parker and Holm (1990), S. m. borealis; c, Laurila and Hario (1988), S. m. mollissima; d, Erikstad and Tveraa (1995), S. m. maollissima; e, Harðardóttir et al. (1997), S. m. islandica; f, Korschgen (1977), S. m. dresseri; g, Franzmann (1980), S. m. mollissima; h, this study, S. m. mollissima. We found no relationship between the rate of mass loss and body mass at the start of incubation (regression, R2 = 0.219, P = 0.43). Using data from the literature, however, we found that the rate of mass loss during incubation was related linearly to body mass at the start of incubation (Fig. 2), when averaged over a colony or a local population. Afton and Paulus (1992) found a similar relationship among waterfowl species. We found no relationship either between incubation constancy and the rate of mass loss (regression, R2 = 0.021, P = 0.82), or between incubation constancy and body mass at the start of incubation (regression, R2 = 0.015, P = 0.85). This may be because incubation constancy varied only a small amount (CV = 0.2%), whereas body mass at the start of incubation and the rate of mass loss varied much more (CV = 8.5% and 16.3%), and sample size was low. In their review, Afton and Paulus (1992) observed that species showing a high incubation constancy are characterized by a large body mass and high relative mass loss. This study was funded jointly by the National Environmental Research Institute (NERI), Department of Coastal Zone Ecology, Denmark and by the Department of Natural Resources Sciences of Macdonald College, McGill University, Canada. NERI also kindly authorized our work in the national nature area at Saltholm Island. Electronic nest monitors were provided by the Canadian Parks Service, Québec Region, Mingan Archipelago National Park Reserve. We are grateful to Scotland Talley for proofreading the manuscript.