Behav Ecol Sociobiol (2002) 52:282–288 DOI 10.1007/s00265-002-0523-x
O R I G I N A L A RT I C L E
Sveinn Are Hanssen · Halvor Engebretsen Kjell Einar Erikstad
Incubation start and egg size in relation to body reserves in the common eider Received: 15 June 2001 / Revised: 20 June 2002 / Accepted: 22 June 2002 / Published online: 17 July 2002 © Springer-Verlag 2002
Abstract Avian incubation is often initiated before all eggs are laid. In altricial birds this has been proposed to facilitate brood reduction through asynchronous hatching. However, in precocial birds eggs normally hatch synchronously even if incubation has started before all eggs are laid. Patterns of incubation start may be the adaptive trait selected for both in altricial and precocial species. Several hypotheses have been proposed to explain the timing of incubation start in birds. Decreasing egg-size after incubation start may be adaptively related to an early incubation start, either to ensure synchronous hatching or to decrease fitness cost of late hatched eggs. We have measured individual body condition, egg size and start of incubation in common eider Somateria mollissima, a precocial sea-duck which does not feed during the incubation period. Females in poor body condition start to incubate earlier in the laying sequence than those in good body condition. Furthermore females in poor body condition lay smaller final eggs than females in good body condition. The laying of smaller eggs late in the sequence is therefore probably related to energetic or nutritional state. We propose that females in poor body condition start to incubate early to shorten the nest period in order to reduce their mass loss, but at the cost of reduced size and growth of the ducklings from the eggs laid after incubation start. Females in good body condition on the other hand postpone incubation start at the cost of a longer incubation period and a higher mass loss to the benefit of synchronized hatching and a higher survival of ducklings. Communicated by C.Brown S.A. Hanssen (✉) · H. Engebretsen · K.E. Erikstad Norwegian Institute for Nature Research (NINA), Department of Arctic Ecology, The Polar Environmental Centre, N-9296 Tromsø, Norway e-mail:
[email protected] Tel.: +47-77-750416, Fax: +47-77-50401 S.A. Hanssen · H. Engebretsen · K.E. Erikstad Biology Department, Faculty of Science, University of Tromsø, N-9037 Tromsø, Norway
Keywords Common eider · Egg size · Female body mass · Incubation start · Somateria mollissima
Introduction Breeding is an energy-demanding process and many reproductive parameters such as offspring size, -number and laying date may be constrained by energy demands (Ricklefs 1974; Drent and Daan 1980; Clutton-Brock 1991). However these same traits could also be affected by selective pressures which would induce the appearance of adaptive patterns (e.g. Stearns 1992). One of these traits may be the start of incubation, which in many birds takes place before all eggs are laid. In altricial birds this leads to asynchronous hatching and size hierarchies of siblings. Such size hierarchies of young have been viewed as an adaptation to facilitate brood reduction where small chicks are placed at a disadvantage in relation to its larger siblings if the parents fail to supply enough food for all the young (Lack 1954). However, support for this and other hypotheses treating asynchronous hatching as an adaptive trait have not been conclusive (e.g. Amundsen and Stokland 1988; Stoleson and Beissinger 1995; 1997; Stenning 1996; Vinuela 2000). This has led to hypotheses stating that the pattern of incubation initiation per se could be selectively favored, rather than the consequent size hierarchy (Clark and Wilson 1981; Magrath 1990; Amundsen and Slagsvold 1991; Stoleson and Beissinger 1995, 1997). Many precocial birds also start incubation before the laying of the last egg (Norton 1972; Caldwell and Cornwell 1975; Afton and Paulus 1992; Flint et al. 1994; Persson and Göransson 1999; Hübner et al. 2002), nevertheless, all chicks usually hatch within 24 h and leave the nest soon after (Bjärvall 1968). In precocial birds like ducks, geese and many waders, the young feed themselves and parental costs are mainly related to anti-predator behavior and leading young to food (Boyd 1953; Black and Owen 1989; Sedinger and Raveling 1990; Bustnes and Erikstad 1991). Therefore, parental effort is largely independent
283
of brood size and it is unlikely that precocial birds should apply a brood reduction strategy (e.g. Robertson and Cooke 1993, but see Friedl 1993). Studying the start of incubation in species with synchronous hatching of eggs may enhance our understanding of any possible adaptive explanations for an early incubation start, because we need not take into account possible adaptive values of asynchronous hatching patterns. Many of the hypotheses predicting a facultative timing of incubation start are developed with altricial species in mind, and mostly as alternatives to hypotheses explaining hatching asynchrony (reviews in Magrath 1990; Stenning 1996; Stoleson and Beissinger 1995, 1997). An early incubation start is predicted if it reduces length of incubation and increases egg, chick or adult survival (Dunlop 1910; Clark and Wilson 1981; Hussel 1985; Bollinger et al. 1990). Furthermore, early incubation start and increased female nest attentiveness during laying may reduce intraspecific brood parasitism (Kendra et al. 1988) or prevent the nest site from being occupied by another individual (Beissinger and Waltman 1991). An early incubation start and early hatching may also lead to the matching of the peak of some resource or favorable condition (Hussel 1972; Clark and Wilson 1981; Slagsvold 1986a). In this case the timing of incubation start should vary depending on when in the season the clutch is initiated. The need for synchronous hatching in precocial birds is a factor that might conflict with an early incubation start. However, several mechanisms to facilitate synchronized hatching in the presence of an early incubation start have been proposed. Embryos in late eggs can be stimulated to accelerate development and hatch in a shorter time than earlier laid eggs, or they hatch not fully matured. This mechanism has been documented in lesser snow goose Anser caerulescens (Davies and Cooke 1983), pheasant Phasianus colchicus and mallard Anas platyrhynchos (Persson and Andersson 1999) where the synchronization of hatching is facilitated by communication between embryos in the last days before hatching (Vince 1964, 1968; Freeman and Vince 1974; Lauch et al. 1988). Another hypothesis (not necessarily mutually exclusive) is that females lay smaller eggs after they start to incubate. This hypothesis was originally proposed by Parsons (1972) and is based on the assumption that smaller eggs require less incubation (Worth 1940; Rahn and Ahr 1974; Arnold 1993). Several studies of both ducks and geese have documented a decrease in egg size with laying sequence (Leblanc 1987; Arnold 1991; Swennen and Van der Meer 1992; Flint and Sedinger 1992; Robertson and Cooke 1993; Williams et al. 1993; Erikstad et al. 1998; Hübner et al. 2002). Earlier hatching of the late eggs means that they may hatch less developed than their siblings, and young from the smaller late laid eggs also have slower growth rates (Erikstad et al. 1998; Anderson and Alisauskas 2002). In many precocial waders, ducks and geese the female may leave behind the partly hatched or unhatched chick from the last laid egg because they need to depart to feeding areas
with the rest of the brood (e.g. Williams et al. 1993), suggesting that an early incubation start in precocial birds may reduce the fitness of late hatched young. Smaller late eggs in clutches, where incubation has started early, may also be an adaptation to decrease the cost of asynchronous hatching, because the female thereby invests less resources in the late eggs that hatch less developed and with reduced survival chances (Quinn and Morris 1986; Williams et al. 1993). The aim of the present study was to examine if the start of incubation in female common eiders Somateria mollissima is initiated before all eggs are laid and, if so, to examine if incubation start could be related to laying date, clutch size, female body condition or egg size gradient. We removed the first laid egg from the nest and replaced it with an artificial egg with a temperature logger. The timing of incubation start (as indicated by a rise in egg temperature), egg volume, laying date and female body condition were measured. We especially focussed on the timing of incubation in relation to female body condition. Body reserves are crucial in capital breeders and individuals with low body mass are expected to save energy, possibly at the cost of reduced current reproduction. To our knowledge this is the first study to explore such a relationship in any detail in a precocial bird.
Methods This study was conducted in the 1995 and 1996 breeding seasons in an eider colony on Grindøya near Tromsø, northern Norway (69°49′N, 18°15′E). Grindøya is an island of 0.65 km2 where ca. 500 pairs of eider breed. The common eider is a long-lived seaduck and, as in many precocial birds, the female accumulates endogenous reserves prior to laying which she then uses during both the egg laying and the incubation periods (Alisauskas and Ankney 1992; Afton and Paulus 1992). In the common eider, anorexia during breeding results in a mass loss of 40% during egg laying and incubation (Korschgen 1977; Parker and Holm 1990; Erikstad and Tveraa 1995). The female eider is incubating and caring for ducklings without assistance from the male. The time from the laying of the last egg until hatching is 22–26 days (Erikstad and Tveraa 1995). The eider is an open nesting species and after egg laying is completed as much as 53% of the nests may not hatch any eggs because of egg predation by crows and gulls (Erikstad and Tveraa 1995). The colony was visited daily from the start of egg-laying in mid-May to find first laid eggs. In nests containing only one egg we replaced this egg with a plastic coated silicone egg with an electronic temperature sensor inside. This sensor was connected to an electronic data logger (tinytalk type temp. 75) via a cable from the egg. The logger registered egg-temperature (±0.02°C) every hour. The loggers were removed from the nests 2–3 days after clutch completion and moved to another nest to increase sample size. Females were caught 1–5 days after the laying of their final egg and weighted using a pesola spring balance (±2.5 g.). We also measured wing length (±0.5 mm.) as an index of female structural size. An earlier analysis showed that wing length explained 93% of the variance in the first principal component (n=325), in an analysis based upon wing length, bill and head measurements (Erikstad and Tveraa 1995). To obtain an estimate of body mass reserves at the start of incubation we entered the time elapsed from start of incubation until weighing and structural size (wing length) in a regression analysis with body mass as a the dependent variable. The mass of females increased with wing length (partial r2=0.13, df=20, P=0.02) and
284
Fig. 1 An example of the measured egg temperature for the first 10 days of egg laying and incubation for one female Common Eider Somateria mollissima. The solid line shows the fitted logistic growth curve decreased with the time since incubation had started (partial r2=0.18, df=20, P=0.01). The residuals from this regression model were used as an index of body mass reserves. The term body condition will hereafter refer to this index. Female common eiders in the study area lay from 3 to 6 eggs, but 4- and 5-egg clutches are the most common (Erikstad et al. 1993). To ensure sufficient samples in each clutch size category we have used only 4- and 5-egg clutches in the analyses. All eggs were individually numbered in the laying order, and egg-length and breadth were measured (±0.1 mm) using a vernier caliper. An index of egg volume was calculated using the formula, (volume = length × breadth2 × π × 0.000164) (Guild 1974) which describes most of the variation in fresh egg mass in the study area (Erikstad et al. 1998). In all analyses the mean temperatures every 2 h were used as single observations. To estimate incubation start and incubation temperature we analyzed each nest separately and fitted a logistic growth curve to each data series, estimating the parameter values for each nest. The expression for the logistic growth curve is
This model has three parameters: k, n0 and r. The parameter n0 is the intercept and the parameter k is the asymptote, i.e. the asymptotic temperature during incubation. The parameter r is a measure of the rate of increase. To estimate the parameters k, n0 and r, we used nonlinear curve fitting (PROC NLIN, SAS 1999). Several studies indicate that there is no embryonic development below 25–27°C (White and Kinney 1974; Drent 1975; Haftorn 1978, 1981, 1988). We defined start of incubation as the time the fitted temperature curve exceeded 27°C to ensure a conservative estimate of incubation start. Figure 1 shows an example from one of the clutches in 1996. Table 1 Egg volume (mean, maximum and minimum) in cm3, egg laying date in May, start of incubation (number of eggs laid), the difference in volume (%) between the largest and last laid egg, body mass (g) and body condition index among Common Eider Somateria mollissima females producing 4- and 5-egg clutches. All values are given as mean ± SE, sample sizes (number of nests) are given in parentheses and all df=1,20. Data from two different years (1995 and 1996)
Watson et al. (1993) found that the eider lays 1 egg per 28 h, and we therefore assumed that the second egg was laid on average 14 h after the discovery of the first egg. The following eggs were assumed to be laid with 28-h intervals. Using this information we calculated the number of eggs that had been laid when the incubation temperature was reached for each nest. The laying intervals for different eggs may not be constant; laying interval may decrease with increasing laying sequence (Eisenhauer and Kirkpatric 1977). Others have reported a longer interval between the penultimate and ultimate egg than among other eggs in the sequence (Schubert and Cooke 1993). Our estimate of the start of incubation in relation to laying sequence presented here should therefore be considered conservative, i.e. birds may have started incubation even earlier in the laying sequence. We analyzed the patterns of egg volume in relation to position in the laying sequence, controlling for the effect of the individual female, in a nested analysis with egg volume as dependent variable and female number and egg number nested within female number as independent variables according to Jover et al. (1993). We analyzed the relationship between the estimated number of eggs laid after start of incubation in relation to year, laying date, clutch size, relative egg volume differences, female body condition and all second order interactions using a covariance analysis. We used a standard backwards procedure starting with a full model and sequentially deleting the variable contributing least to the model. The relative difference between the largest and the last laid egg in the clutch was analyzed using the same procedure as above. We entered the same independent variables in the initial full model. All values are presented as means ±SE and all tests are twotailed according to SAS (SAS 1999). P-values less than 0.05 were considered statistically significant.
Results Egg volume and egg-laying date did not differ between 4- and 5-egg clutches, but the start of incubation in relation to laying sequence was on average 1 day earlier for 4-egg clutches than for 5-egg clutches (Table 1). Hence both females laying 4-egg- and 5-egg clutches laid the same average number of eggs after incubation had started. Egg volume differences (between largest and last egg in the clutch) seemed to be larger within 5- than 4-egg clutches; this was, however, not significant (Table 1). Egg volume differences showed a pattern in relation to laying sequence where egg size generally declined with position in the laying sequence (Fig. 2). For 4-egg clutches: female effect, F8,34=1.6, P=0.19; position in laying sequence nested within female, F9,34=3.6, P=0.01. And for 5-egg clutches: female effect, F11,56=0.6, 4-egg
5-egg
ANCOVA Year
Mean egg volume (cm3) Min. egg volume (cm3) Max. egg volume (cm3) Egg laying date Start of incubation Eggs laid after incubation start Relative egg size differences Body mass (g) Body condition
103.5±1.3 (9) 96.5±1.7 (9) 108.3±1.3 (9) 27.0±1.0(9) 2.6±0.2 (9) 1.4±0.2 (9) 11±1 (9) 1,870±41(9) 19.7±30.0(9)
102.3±1.3 (12) 94.0±1.5 (12) 108.2±1.5 (12) 27.4±1.0 (12) 3.5±0.2 (12) 1.5±0.2 (12) 13±1 (12) 1,885±88 (12) –14.8±29.1(12)
Clutch
F
P
F
P
0.6 1.2 0.6 6.0 1.2 1.2 0.5 0.5 0.6
0.45 0.28 0.43 0.03 0.29 0.29 0.48 0.50 0.45
0.4 1.4 0.01 0.03 10.1 0.1 3.3 0.09 0.7
0.52 0.25 0.92 0.86 0.005 0.71 0.09 0.77 0.41
285
Fig. 2 Mean egg volume (±SE) in relation to laying sequence for Common Eider clutches of 4 (n=9) (circles) and 5 (n=12) (squares) eggs. See text for statistical analyses Fig. 4 Egg volume differences (relative differences between the largest and last egg in the clutch) in relation to female body condition (residual body mass controlling for body size and number of days the female had been incubating when she was captured and weighed) in Common Eider clutches of 4 (circles) and 5 (squares) eggs. See text for statistical analyses
There was no significant relationship between the timing of incubation start and relative egg size difference (F1,20=0.97, P=0.34).
Discussion
Fig. 3 The estimated number of eggs laid after the start of incubation (females lay 1 egg per 28 h) in relation to female body condition (residual body mass controlling for body size and number of days the female had been incubating when she was captured and weighed) in Common Eider clutches of 4 (circles) and 5 (squares) eggs. See text for statistical analyses
P=0.85; position in laying sequence nested within female, F12,56=1.8, P=0.09). The results from a backwards ANCOVA on incubation start in relation to clutch termination showed that the estimated number of eggs laid after start of incubation were related to body condition (F1,20=15.1, P
