Flexibility in the parental effort of an Arctic-breeding seabird

Authors


*Correspondence author. E-mail: aharding@usgs.gov

Summary

  • 1Parental investment strategies are considered to represent a trade-off between the benefits of investment in current offspring and costs to future reproduction. Due to their high residual reproductive value, long-lived organisms are predicted to be more reluctant to increase parental effort.
  • 2We tested the hypothesis that breeding little auks (Alle alle) have a fixed level of reproductive investment, and thus reduce parental effort when costs associated with reproduction increase.
  • 3To test this hypothesis we experimentally increased the flight costs of breeding little auks via feather clipping. In 2005 we examined changes in the condition of manipulated parents, of the mates of manipulated parents, and of their chick as direct measures of change in parental resource allocation between self-maintenance and current reproduction. In 2007 we increased sample sizes to determine whether there was a physiological cost (elevated corticosterone, CORT) associated with the manipulation.
  • 4We found that: (i) clipped birds and their mates lost more body mass than controls, but there was no difference in mass loss between members of a pair; (ii) clipped birds had higher CORT levels than control birds; (iii) there were no inter-annual differences in body mass and CORT levels between clipped individuals and their mates at recapture, and (iv) chicks with a clipped parent had lower peak and fledging mass, and higher CORT levels than control chicks in both years.
  • 5Contrary to our hypothesis, the reduction in body mass of partners to clipped birds suggests that little auks can increase parental effort to some extent. Nonetheless, the lower fledging mass and higher CORT of chicks with a clipped parent indicates provisioning rates may not have been fully maintained.
  • 6As predicted by life-history theory, there may be a threshold to the additional reproductive costs breeders will accept, with parents prioritizing self-maintenance over increased provisioning effort when foraging costs become too high.

Introduction

Iteroparous individuals should balance their investment in the current reproductive attempt against future opportunities to breed (e.g. Roff 1992; Stearns 1992). The safeguarding of self-maintenance is particularly relevant to long-lived species, because a small reduction in adult survival can have a large negative impact on lifetime reproductive success (Charlesworth 1980). There has been much interest in understanding how long-lived seabirds optimize the balance between current and future reproduction (e.g. Erikstad et al. 1998; Hanssen et al. 2005). Adhering to a fixed level of investment in the current reproductive event is one strategy by which birds can maximize their lifetime reproductive success (Sæther, Andersen & Pedersen 1993; Navarro & González-Solís 2007). Whilst there is considerable evidence to support this hypothesis (e.g. Ricklefs 1987, 1992; Hamer & Hill 1994; Mauck & Grubb 1995), other studies have shown that provisioning effort can be flexible, and adjusted according to the offspring's demand for energy (e.g. Tveraa et al. 1998; Granadeiro et al. 2000; Velando & Alonso-Alvarez 2003), even, in some cases, at a cost to adult survival (Jacobsen, Erikstad & Sæther 1995; Golet et al. 2004).

Strategies of parental investment may be condition-dependent. Seabirds live in a stochastic environment, where foraging conditions can vary widely among years (e.g. Barrett & Rikardsen 1992). Such variation may favour a breeding strategy where parental effort is flexible, and adjusted according to the conditions of the breeding season, the parents’ prospects of survival, and the fitness of the offspring (Erikstad et al. 1997, 1998; Velando & Alonso-Alvarez 2003). The level of endogenous energy reserves may govern parental investment decisions, with individuals unwilling to increase effort in raising young when their reserves drop below a critical threshold (Johnson, Erikstad & Sæther 1994; Chaurand & Weimerskirch 1994; Tombre & Erikstad 1996; Tveraa et al. 1998). Individuals with high endogenous reserves may have the flexibility to increase parental effort, whereas individuals with low reserves may be unable or unwilling to do so, that is, they will exhibit a fixed level of investment. In addition to body condition, food availability and the survival prospects of the offspring, parents should also optimize their parental care decisions in relation to the effort of their partner (Houston & Davies 1985). Studies of bi-parental care systems have shown that the reduction in the provisioning effort of one parent often results in increased effort by its partner (e.g. Paredes, Jones & Boness 2005), although there may not be full compensation (Houston, Szekely & McNamara 2005).

Studies examining life-history trade-offs in long-lived species have used a variety of experimental methods to test the hypothesis that an increased cost of reproduction leads to a reduction in parental effort (e.g. Sæther et al. 1993; Mauck & Grubb 1995; Navarro & González-Solís 2007). Although many such studies have assumed that mass loss in response to increased reproductive costs represents an increase in parental effort, and a corresponding decline in condition (e.g. Velando & Alonso-Alvarez 2003), there are difficulties in interpreting changes in body mass (Holt et al. 2002; Nisbet et al. 2004). Mass loss during the breeding season has been reported in a number of birds, and has often been attributed to a loss of reserves associated with increased reproductive demands: the reproductive stress hypothesis (Ricklefs 1983; Nur 1984; Golet & Irons 1999). This hypothesis is supported by studies that have compared mass change in un-supplemented and food-supplemented individuals, or mass change in years of good vs. poor natural food supply (Merila & Wiggins 1997; Holt et al. 2002; Nagy, Stanculescu & Holmes 2007). However, other studies suggest that body mass loss may be an adaptive adjustment for reducing flight costs and increasing foraging efficiency during an energetically demanding period of the breeding season: the adaptive mass loss hypothesis (Nordberg 1981; Gaston & Jones 1989; Moreno 1989; Croll, Gaston & Noble 1991; Slagsvold & Johansen 1998). These two alternative explanations for mass loss in breeding birds may operate differently in each sex (e.g. Ritz 2007), and at different stages of breeding (e.g. Quillfeldt, Masello & Lubjuhn 2006). Moreover, the two explanations are not mutually exclusive because stress induced mass loss automatically results in a more efficient weight for flight (Ritz 2007).

One way to interpret changes in body mass and determine the role of endogenous reserves in investment decisions is to make a second, independent measure of energy balance. The steroid hormone corticosterone (CORT) plays an important role in an individual's adaptive response to environmental stress (Wingfield et al. 1998; Romero, Reed & Wingfield 2000). Individuals have been shown to respond to a wide variety of stressors such as adverse environmental conditions, predators, and parasites by increasing secretion of CORT (e.g. Wingfield & Romero 2001; Raouf et al. 2006). Levels have also been shown to be elevated during food shortages (Kitaysky et al. 1999a; Pravosudov et al. 2001), and in individuals in poor body condition (Kitaysky, Wingfield & Piatt 1999b; Romero & Wikelski 2001). However, although a short-term increase in CORT may help facilitate metabolism of endogenous energy stores or shift an individual's behaviour to foraging during stressful periods (Wingfield, O’Reilly & Astheimer 1995; Sapolsky, Romero & Munck 2000), high levels of CORT over an extended period of time can also have deleterious consequences. For example, there is evidence that increased CORT can reduce cognitive abilities (Kitaysky et al. 2003) and immunity (Saino et al. 2003), and a growing number of studies have also shown that high levels of CORT can impact survival (Romero & Wikelski 2001; Brown et al. 2005; Kitaysky, Piatt & Wingfield 2007). Therefore, measuring changes in body mass and the CORT levels of parents can demonstrate the extent of physiological stress associated with a loss of body mass, such that mass change can be interpreted in relation to reproductive costs. In addition to chick growth, circulating levels of chick CORT can also be used as an index of the amount of food chicks receive; food shortages in nest-bound chicks result in an elevation of CORT (Kitaysky, Wingfield & Piatt 2001; Kitaysky et al. 2003), which may facilitate begging and result in increased parental provisioning (e.g. Kilner, Noble & Davies 1999).

In this study we used this two-pronged approach to examine the flexibility of parental effort in the little auk (Alle alle). The little auk is a small, planktivorous seabird that breeds in the high Arctic and lays a single egg annually (Stempniewicz 2001). Like many diving seabirds, little auks have short wings as a morphological adaptation to underwater propulsion during prey capture (Taylor 1994). As a result of their small wing area and subsequent high wing loading, little auk flight is very costly (Gabrielsen et al. 1991). Although individuals may benefit from carrying extra body reserves as a buffer against adverse foraging conditions, additional body mass also greatly increases flight costs (Witter & Cuthill 1993; Veasey, Metcalfe & Houston 1998). Optimal body reserves should therefore be minimal, and this may limit the flexibility of parental effort in the little auk.

Given these physiological traits, and the expectation that long-lived, iteroparous species with low annual fecundity take less of a mortality risk during a given breeding attempt than more r-selected species (Erikstad et al. 1997, 1998; Golet & Irons 1999; Golet et al. 2004), we hypothesized that little auk parents have a fixed level of reproductive investment, and will reduce effort if the costs associated with reproduction increase. We predicted that parents should (i) maintain their own condition, and (ii) allocate fewer resources towards rearing their offspring, if costs associated with providing food for their chick increase. To test these predictions we experimentally increased the flight costs of breeding little auks via feather clipping, and examined changes in body mass and circulating levels of CORT of manipulated individuals, their mates, and their chicks, as direct measures of a change in parental resource allocation between self-maintenance and current reproduction.

We took ethical considerations into account when designing the experimental manipulation. It is possible to experimentally increase flight costs by adding a weight (e.g. Wright & Cuthill 1989; Sæther et al.1993), or clipping feathers to decrease wingspan or area (e.g. Mauck & Grubb 1995; Weimerskirch, Chastel & Ackermann 1995; Navarro & González-Solís 2007). Weights add a permanent handicap if the bird is not recaptured, whereas feather clipping has the advantage that these feathers will grow back during post-breeding moult (Stempniewicz 2001). Thus, we decided to manipulate wing area by removing two primaries (4 and 5) on each wing, thus reducing wing area and increasing flight costs (Winkler & Allen 1995; Moreno et al. 1999; Sanz, Kranenvarg & Tinbergen 2000). As flight has high energy requirements in little auks (Gabrielsen et al. 1991), we expected that a small decrease in wing area would significantly increase provisioning costs.

Methods

study system

Fieldwork was conducted from 11 June to 20 August 2005 and from 22 June to 17 August 2007 at Kap Höegh (70°43′ N, 22°38′ W) on the east coast of Greenland. Little auks nest in rock crevices on the scree slopes of Kap Höegh. They are largely planktivorous, with calanoid copepods (Calanus species) dominating diet across their breeding range (Pedersen & Falk 2001). Little auks catch their prey underwater during wing-propelled dives, and parents carry fresh, intact zooplankton back to their single chick at the colony in a throat pouch (Stempniewicz 2001). Little auks have minor sexual dimorphism, with males slightly larger on average than females (Jakubas & Wojczulanis 2007), and are socially monogamous (Lifjeld et al. 2005). After hatching, the chick is brooded constantly for the first 3–4 days, beyond which time it is attended briefly during provisioning bouts (Stempniewicz 2001). Both parents share incubation and chick-feeding duties, although there is evidence that the male does the majority of provisioning towards the end of the chick-rearing period, before escorting the fledgling to sea (Harding et al. 2004). Chicks are fed approximately five meals per day (Stempniewicz 2001), and leave the colony when they reach, on average, 25-days-old (Harding et al. 2004).

experimental design

We conducted two experiments, one in 2005 and one in 2007, to examine the flexibility of parental effort in little auks. Indices of breeding success suggest that conditions in the two years were fairly similar (Harding et al. unpubl. data): median hatching date was 15 July (n = 46 nests) in 2005, and 18 July (n = 62) in 2007; hatching success (number of chicks hatched/number of eggs monitored) was 90·7% (n = 43 nests) in 2005, and 88% (n = 109 nests) in 2007; fledging success (number of chicks fledged/number hatched) was 92·3% (n = 43) in 2005, and 93% (n = 61) in 2007, and chicks fledged at a similar age in both years (2005, median = 25·5 days, n = 29; 2007, median = 26·0 days, n = 23). Moreover, the lack of any significant difference in the baseline CORT levels of adult breeding little auks sampled during the mid chick-rearing period in each year (t = 0·76, df = 59, P = 0·45: 2007 mean = 2·18 ng mL−1 ± 0·31, n = 39; 2005 mean = 1·90 ng mL−1 ± 0·18, n = 35) suggests that on average parents were in similar physiological condition in both years.

We experimentally increased the flight costs of breeding little auks in both years via feather clipping (removal of primaries 4 and 5 on each wing), and used changes in the condition (as reflected in body mass and CORT levels) of manipulated parents, the mates of manipulated parents, and their chick as direct measures of change in parental resource allocation between self-maintenance and current reproduction. Determining the level of parental flexibility requires knowledge of both members of the breeding pair, because reduction in the effort of one parent may trigger compensatory care by its partner (Paredes et al. 2005). In 2005, we therefore compared changes in condition within manipulated individuals and their mates to changes within unmanipulated pairs. The 2005 protocol required the recapturing of both members of experimental and control pairs, and this constraint limited our sample size. With the aim of increasing sample sizes, we changed the protocol in 2007 to use non-clipped breeding birds caught at random at the colony as controls. Experimental birds were compared with birds caught at a similar time as the initial capture (incubation colony controls) and recapture (chick-rearing colony controls). Chick growth and baseline CORT levels were measured in all experimental and control nests from both years to determine whether parents altered their provisioning behaviour in response to the increased flight costs.

nest selection

Little auks are sensitive to disturbance during the incubation phase of their breeding cycle, and may abandon breeding in response to disturbance. However, they are less prone to desertion towards the end of incubation. Most researchers therefore estimate lay dates from known chick hatch dates, and base the decision on when to start nest checks at the end of the incubation period from knowledge of previous breeding chronology. In this study, nests were not disturbed until the estimated time of late incubation, when an area was searched for active nest-sites with accessible nest-chambers. A total of 32 nest-sites were marked in 2005, and randomly split into 14 experimental and 18 control nests. Egg abandonment was a problem for some birds caught on the nest, and 10 of the 32 nests failed prior to hatch (6 control and 4 experimental). Although it is generally agreed that little auks can be caught on the nest during late incubation, the high failure rate in 2005 suggests some nests may have been disturbed too early. In 2007 we marked 30 experimental nests, 7 (23%) of these failed to hatch. An additional 76 nests were monitored for colony-level values of productivity and chick growth, and 4 (5%) of these failed to hatch.

method details

2005

We caught both members of control and experimental pairs during late incubation (‘Initial Capture’). Each parent was caught in the nest. Birds were weighed using a Pesola balance accurate to ±2·0 g. A blood sample was taken from which circulating baseline levels of CORT were later measured. All birds were bled according to a standardized technique (Kitaysky et al. 1999b), with an initial blood sample (< 100 µL) collected from the brachial vein within 3 min of capture. In birds, it takes at least 3 min for levels of CORT to begin to rise in the blood in response to a stressor (Romero & Reed 2005). In this study we also didn't find a significant relationship between time after capture and CORT within 0–3 min after capture (linear regression analyses: R2 < 0·01, F = 0·008, df = 133, P = 0·929; mean capture-bleed time = 2·31 ± 0·19 min, n = 134 birds). Thus this first sample provided a baseline measure of circulating CORT and did not reflect the stress induced by capture in little auks. One member of each experimental pair was randomly chosen and the 4th and 5th primary feather from each wing clipped at the base. All birds were marked with an individual colour ring combination, and released back into their crevice immediately after measurements were taken. After collection of blood, samples were centrifuged, and the plasma preserved in 70% ethanol. Spun red blood cells were stored for subsequent molecular sexing of individuals.

We attempted to re-catch all experimental and control parents at 7 days (median = 8 days, range 6–11 days) after initial capture (‘Recapture’). Although the number of days between initial capture and recapture was relatively short, the chances of catching parents at a later date diminishes rapidly as parents spend progressively less time with their growing chick. Starting on day 6 after initial capture, nests were checked approximately every 12 h. All birds were weighed on recapture, and blood samples taken for CORT analysis (see above). Birds were then released back into their crevice.

2007

We caught both adults from 30 experimental nests during late incubation. Birds were weighed, and a blood sample was taken for analysis of baseline CORT (see Methods above). Using exactly the same methods as in 2005, one member of each pair was randomly chosen and the 4th and 5th primary feather from each wing clipped at the base. Effort was made to re-catch birds after 7 days (median = 11 days, range 7–14 days), and all recaught birds were weighed, had a blood sample taken for CORT analysis, and were then released back into their crevice.

We compared baseline levels of circulating CORT in experimental birds at initial capture and recapture with birds caught at the colony during a similar time of the season. Experimental adults at initial capture (11–14 July) were compared with birds caught between 7 and 8 July (n = 32; incubation colony controls). Recaptured experimental adults (22–25 July) were compared with birds caught between 19 and 20 July (n = 28; chick-rearing controls). Both groups of control birds were captured using noose carpets (Pedersen & Falk 2001), and blood was sampled from all birds for baseline CORT levels (following same methods as above). Only breeding birds were sampled, and breeding status determined by the presence of a brood patch or full gular pouch (Stempniewicz 2001).

Chicks

We monitored growth and fledging condition of all experimental and control chicks that could be easily reached (2005, n = 10 experimental and 12 control chicks: 2007, n = 17 experimental and 32 control chicks). We focussed analysis of chick growth on indices of peak mass and fledging condition, because the effects of the manipulation were more likely to be seen in older chicks, due to their high energy demands and greater time spent under the experimental regime. All nests were routinely visited every third day until hatching, although the frequency of these nest-checks increased to daily during the period of adult capture (see above). During each visit the nest chambers were checked with a flashlight, and the presence of adult, egg or chick was recorded. Where an adult blocked our sight of an egg or chick, the adult's brooding-posture and the presence of eggshell fragments were used as evidence of hatching.

Chicks were first handled when they were older than 5 days, and the parents had finished brooding. Chicks were then visited every 3 days, and body mass (±2·0 g) was measured using a Pesola balance. Like many species of bird (e.g. Ricklefs 1968; Gray & Hamer 2001), little auk nestlings attain a peak mass then undergo a period of mass recession before they fledge (Konarzewski & Taylor 1989). Peak mass was defined as the highest mass measured for an individual chick, and fledging mass as the last measurement prior to chick departure from the colony. Baseline levels of circulating CORT in chicks were also sampled as an additional measure of physiological condition. Blood samples (< 50 µL) were taken from chicks at c. 15-days-old for CORT analysis (see Methods above). We continued to check nests every 3 days until chick departure. Chicks that had reached 20-days-old when they disappeared from the nest were considered fledged (Harding et al. 2004).

laboratory analysis

Corticosterone (CORT)

Previous studies have measured CORT levels in plasma frozen on collection in the field (e.g. Kitaysky et al. 1999a). However, due to logistical limitations we were unable to freeze samples and had to rely on another method of preservation. A trial was conducted in Hornsund, Spitsbergen, in order to validate the use of ethanol as an alternative preservative. Plasma samples from 25 individual little auks were split in two, with half stored in ethanol and half frozen. There was a strong significant positive relationship between frozen CORT levels and CORT levels in plasma preserved in ethanol (R2 = 0·99, t = 45·64, df = 23, P < 0·001, Fig. 1). Serial dilution of little auk plasma samples preserved in ethanol (or being frozen) yielded displacement curves (between sample concentration and % bound to steroid antiserum) parallel to standard hormone. These validations confirmed the effectiveness of ethanol as a field preservation method for little auk plasma samples.

Figure 1.

The relationship between concentrations of CORT in the same little auk plasma samples (n = 25) split into two subsamples and subsequently preserved either in ethanol or by freezing (see Methods for details).

Total CORT was measured using a radioimmunoassay. Each whole sample (containing equal known volumes of blood plasma and ethanol) was equilibrated with 2000 cpm of tritiated CORT prior to extraction with 4·5 mL distilled dichloromethane. After extraction, percent tritiated hormone recovered from each individual sample (average hormone recovery was 96·7%, SD = 3·88) was used to correct final values. Samples were reconstituted in PBSG-buffer and combined with antibody and radiolabel in a radioimmunoassay. Dextran-coated charcoal was used to separate antibody-bound hormone from unbound hormone. Inter- and intra-assay variations were < 8% and 2%, respectively; assay sensitivity was 7·8 pg tube−1.

Molecular sexing

Genomic DNA was extracted from all blood samples following the salt-extraction protocol described in Medrano, Aasen & Sharrow (1990) and modified as in Sonsthagen, Talbot & White (2004). The DNA was suspended in 100 µL TE, quantified using fluorometry, diluted to 50 ng µL−1, and stored at −20 °C until analysis. We amplified DNA by polymerase chain reaction (PCR) using the P8/P2 primer set to determine gender of each bird, based on the chromo-helicase-binding domain (CHD) gene (Griffiths et al. 1998). PCR reactions and visualization techniques followed Handel et al. (2006). The reaction yields a 379-base-pair (bp) product from the Z-chromosome (for both males and females), and a 403-bp product from the W-chromosome (females only). Sex was assigned based on the absence (male: ZZ) or presence (female: ZW) of the band for the W chromosome.

statistical analysis

Mass

In 2005 we compared body mass at initial capture between experimental and control groups using an anova. Only nests where chicks hatched and both members of the pair were caught during both the initial capture and recapture period were included in the analysis of differences in parental response between experimental and control nests (n = 6 clipped nests, n = 5 non-clipped). The effect of clipping on the change in adult body mass (g day−1) between initial and recapture period was tested in an ancova, treating mates within a nest as a repeated-measure factor to avoid pseudo replication. Parental sex was included as a factor and chick age at time of recapture as a covariate. To examine the energetic repercussions of changes in parental body mass, we modelled the relationship between body mass and flight costs with the software program Flight V1·18 (Pennycuick 2006), based on a wing area of 0·0217 m2, a wingspan of 0·424 m (Welcker unpubl. data), an estimated flight speed of 20 m s−1 (Pennycuick 1987; Elliott et al. 2004), and a basal metabolic rate of 2·059 W (Gabrielsen et al. 1991).

Body mass of birds at initial capture was compared between 2005 and 2007 using a mixed-effects anova, with year and group (clipped bird vs. partner) and their interaction as treatment effects. To account for the fact that measurements of partners of a pair are not independent, ‘nest’ was included as a random factor. Similarly, we compared body mass of experimental birds at recapture in 2005 and 2007 using a mixed-effects ancova. To control for differences in the time between initial capture and recapture ‘number of days from capture’ was included as a covariate.

Corticosterone (CORT)

A model comparing CORT in 2005 with control nests was not possible due to low sample sizes and high inter-individual variability in baseline CORT levels (e.g. Cockrem & Silverin 2002). Instead, we compared circulating levels of baseline CORT in 2007 clipped birds and their partners at initial capture with late incubation colony controls using a mixed-effects anova with ‘nest’ included as a random factor. In a separate mixed-effects anova, we compared levels of baseline CORT in recaptured clipped birds and their partners with chick-rearing colony controls.

Baseline CORT levels of experimental adults during the initial capture session in 2005 were compared to those at initial capture in 2007 using a mixed-effects anova. We then compared baseline CORT of birds in both years at recapture, using a mixed-effects ancova, including the number of days from initial capture to recapture as a covariate. Changes in CORT between capture and recapture were also compared using a mixed-effects ancova, including the same covariate.

Chicks

Differences in baseline CORT were examined with an anova, with year and treatment (control vs. clipped) as treatment effects. We tested for a difference in peak mass with an ancova, including age at peak mass as a covariate. Fledging mass was examined in the same way, in an ancova, with fledging age as a covariate.

All statistical tests were performed in statistica and r 2·2·1. CORT data were log-transformed prior to analysis to attain normal distribution. Statistical significance was assumed at P < 0·05. Unless otherwise indicated, values reported are means ± 1 SE.

Results

adult mass

There was no significant difference in body mass at initial capture between treatment groups in 2005 (F1,49 = 1·21, P = 0·277; experimental group mean = 159 g ± 1·98, n = 26, control group mean = 156 g ± 2·33, n = 25). There was an effect of experimental group (F1,5 = 8·45, P = 0·034) and no effect of sex (F1,5 = 0·003, P = 0·959) on changes in body mass between time of initial capture and recapture. The amount of body mass loss did not differ between mates within the same nest (F1,6 = 0·05, P = 0·82), and there was no interaction between experimental group and repeated factor (F1,6 = 0·01, P = 0·94). Both clipped birds and their partners lost more mass on average than control birds and their partners (Fig. 2). Loss of body mass in experimental pairs resulted in an 11% decrease in flight costs, calculated after Pennycuick (2006). Mean flight costs for clipped birds and their partners during the initial capture period was 14·8W (mean mass = 161·7 g), whereas flight costs for these same birds during recapture was 13·2W (mean mass = 144·9 g).

Figure 2.

Changes in body mass (mean ± 1 SE) of 2005 experimental and control adults.

There was no difference in experimental adult body mass during the initial capture period between years (F1,30 < 0·01, P = 0·999), groups (clipped birds and their partner; F1,27 = 0·79, P = 0·382), or the interaction between year and group (F1,27 = 0·03, P = 0·865; Fig. 3a). Similarly, there was no significant difference in experimental adult body mass at recapture between years (F1,26 = 3·61, P = 0·069), groups (clipped birds and their partner; F1,16 = 3·49, P = 0·080), or the interaction between year and group (F1,16 = 0·08, P = 0·776; Fig. 3a).

Figure 3.

Values of (a) body mass in experimental birds (clipped birds and partners to clipped birds) and (b) circulating baseline CORT levels (mean ± 1 SE; log-transformed) during the initial capture and recapture session in both years (2005 and 2007). Sample sizes during initial capture: 2005, clipped n = 10, partners = 9; 2007, clipped n = 22, partners = 19. Sample sizes at recapture: 2005, clipped n = 7, partners = 8; 2007, clipped n = 17, partners = 14.

corticosterone (cort)

There was no difference in circulating levels of baseline CORT between 2007 experimental birds (clipped birds and partners) at initial capture and incubation colony controls (F2,17 = 0·92, P = 0·416; incubation colony control mean (log-transformed) CORT = 0·25 ng mL−1 ± 0·04, n = 32; clipped birds = 0·20 ng mL−1 ± 0·06, n = 18; partner to clipped birds = 0·18 ng mL−1 ± 0·04, n = 22; Fig. 4). In contrast, there was a difference in baseline CORT between experimental birds at recapture and the chick-rearing colony controls (F2,10 = 7·13, P = 0·012). Clipped birds had significantly higher baseline CORT levels than the colony controls (Post-hoc contrast with Holm's correction for multiple comparisons, P = 0·018; chick-rearing colony control mean (log-transformed) CORT = 0·24 ng mL−1 ± 0·04, n = 28; clipped birds = 0·46 ng mL−1 ± 0·06, n = 17; partner to clipped birds = 0·32 ng mL−1 ± 0·06, n = 14; Fig. 4).

Figure 4.

2007 CORT levels (mean ± 1 SE; log-transformed) in two capture groups: (i) experimental pairs at initial capture (clipped birds, n = 22; partners to clipped birds, n = 19), and incubation colony control birds (n = 32); and (ii) experimental pairs at recapture (clipped birds, n = 17; partners = 14) and chick-rearing colony birds (n = 28). There was no significant difference in CORT levels in the three groups during the first capture. CORT levels in clipped birds were significantly higher than controls during the second capture session (P < 0·05).

Baseline CORT levels during the initial capture period did not differ between years (F1,30 = 0·78, P = 0·383), or groups (F1,26 = 2·04, P = 0·165), but there was a significant interaction between year and group (F1,26 = 6·27, P = 0·019; Fig. 3b). Clipped birds had higher circulating CORT levels at initial capture than their partners in 2005, but there was no difference between clipped birds and their partners in 2007 (Fig. 3b). Change in baseline CORT between initial capture and recapture did not differ between years (F1,16 < 0·01, P = 0·986) or group (clipped birds and their partner; F1,16 = 0·16, P = 0·696). There was however a significant interaction between year and group (F1,16 = 8·88, P = 0·009). Partners to clipped birds increased CORT levels more than clipped birds in 2005 (clipped bird mean change (log-transformed) =+ 0·22 ng mL−1 ± 0·73, n = 7; partner mean change = + 1·83 ng mL−1 ± 0·53, n = 8), whereas clipped birds increased CORT more their partners in 2007 (clipped bird mean change (log-transformed) = + 1·58 ng mL−1 ± 0·44, n = 17; partner mean change = + 0·52 ng mL−1 ± 0·46, n = 14). Despite differences during initial capture, baseline CORT levels of experimental birds at recapture did not differ between years (F1,26 = 0·33, P = 0·573) or groups (F1,16 = 1·14, P = 0·302), and there was no significant interaction between year and group (F1,16 = 3·43, P = 0·083; Fig. 3b).

chicks

There was little difference in fledging success between experimental and control chicks: all control and experimental chicks in both years fledged except for one chick in the 2005 control group, which died of predator-inflicted injuries. There was an effect of experimental treatment on levels of chick baseline CORT (anova: F1,66 = 21·92, P < 0·001), but no effect of year (anova: F1,66 = 0·48, P = 0·49) or the interaction between year and treatment (anova: F1,66 < 0·01, P = 0·962). Chicks with a clipped parent in both years had higher levels of baseline CORT than control chicks (Table 1, Fig. 5). There was an effect of treatment (anova: F1,57 = 9·77, P = 0·003) and year (anova: F1,57 = 6·54, P = 0·013) on fledging mass, but no interaction between treatment and year (anova: F1,57 = 1·74, P = 0·193). Chicks with a clipped parent fledged at a lower mass than control chicks on average, and chicks in 2005 fledged at a lower mass than chicks in 2007 (Table 1, Fig. 5). Peak mass differed between treatment groups (anova: F1,57 = 7·84, P = 0·007), but not between years (anova: F1,57 = 1·38, P = 0·245), and the interaction between year and treatment was not significant (anova: F1,57 = 2·61, P = 0·112). Experimental chicks had lower peak mass than controls (Table 1).

Table 1.  Indices of the condition of chicks from clipped and control nests in 2005 and 2007. Pairs of values that differ significantly are indicated with matching superscripts. Fledging mass differed between years, with chicks lighter in 2005 on average (P < 0·05). Peak mass, fledging mass and baseline CORT differed between chicks in clipped and control groups (P < 0·05). There was no significant interaction between year and treatment group in any of the three indices
  20052007ClipControl
ClipControlAllClipControlAllAllAll
Peak mass (g)n1011211732492743
Mean116·70127·27122·24124·18125·19124·84121·41d125·72d
SE2·572·632·152·101·751·341·751·46
Fledging mass (g)n1011211732492743
Mean105·50116·82111·43c113·35118·16116·49c110·44b117·81b
SE3·642·322·422·251·721·402·051·40
Log-CORT (ng mL−1)n1012221632482644
Mean0·540·240·380·500·190·300·51a0·21a
SE0·090·040·060·090·030·040·060·03
Figure 5.

Chicks in 2005 and 2007 with a clipped parent had higher baseline CORT levels (mean ± 1 SE ng mL−1; log-transformed) and fledged at a significantly lower body mass (g) than chicks in the control group (n = 27 experimental and 44 control chicks).

Discussion

In this study we tested the hypothesis that little auk parents reduce parental effort if costs associated with reproduction increase. We predicted that parents should (i) maintain their own condition, and (ii) allocate fewer resources towards rearing their offspring, if costs associated with providing food for their chick increase. To test these predictions we experimentally increased the flight costs of breeding little auks via feather clipping. In 2005 we examined changes in the condition of manipulated parents, of the mates of manipulated parents, and of their chick as direct measures of change in parental resource allocation between self-maintenance and current reproduction. In 2007, we increased sample sizes to determine whether there was a physiological cost (elevated CORT) associated with the manipulation.

Our main results can be summarised as follows: (i) In 2005, both clipped birds and their partners lost more body mass than control parents, but there was no difference in mass loss between members of a pair; (ii) in 2007, clipped individuals had higher CORT levels than unmanipulated controls; (iii) there were no inter-annual differences in body mass and CORT levels between clipped individuals and their mates at recapture; (iv) chicks with a clipped parent had higher CORT levels, and lower peak and fledging mass than control chicks in both years. Below, we discuss the implications of these results under three broad categories: flexible parental effort, mass loss, and compensatory parental care.

flexible parental effort

The observed changes in chick and adult physiological condition are most likely explained by changes in parental behaviour and energy expenditure, with the handicapping of one of the parents triggering a behavioural/physiological cascade: (i) the clipped birds either reduced their provisioning effort in response to increased flight costs, or had impaired foraging ability; (ii) chicks increased CORT secretion in response to nutritional stress (Kitaysky et al. 2001); (iii) CORT increased chick begging behaviour (Kitaysky et al. 2003); and (iv) mates of clipped birds responded to begging chicks by an increase in provisioning effort (Burford, Friedrich & Yasukawa 1998; Kilner et al. 1999; Kitaysky et al. 2001). Behavioural observations of both parents are needed to qualify our interpretation of these physiological results.

Contrary to our hypothesis that little auk parents will not increase effort in rearing their current offspring, the mass loss by the partners of the clipped bids suggests that little auks do have the capacity to increase parental effort if required. Nonetheless, the lower fledging mass of chicks with a clipped parent suggests that provisioning rates were not fully maintained, and that partners of clipped birds were unable or unwilling to fully compensate for their partner's presumed reduction in food delivery. Thus, as predicted by life-history theory, there may be a threshold to the additional reproductive costs breeders will accept, with parents prioritizing self-maintenance over increased provisioning effort when foraging costs become too high (e.g. Velando & Alonso-Alvarez 2003). The similarity in body mass and CORT levels between clipped birds and their partners at recapture in both years is also indicative of a threshold to investment in the current reproductive attempt. A threshold to depletion of endogenous reserves linked to increased CORT secretion may be serving as an important mechanism by which parents monitor their own mortality risk. For example, Romero and Wikelski (2001) found a threshold relationship between the body condition of Galápagos marine iguanas (Amblyrhynchus cristatus) and CORT levels, with CORT levels rising disproportionally in individuals in extremely poor condition. They also showed a strong negative relationship between CORT levels and survival. An experimental manipulation of CORT in little auks is needed to determine if elevated baseline CORT signals critical depletion of endogenous energy reserves, and whether parents that have reached that threshold would abandon their breeding attempt. It is also essential to establish the relationship between circulating CORT levels and survival to fully interpret our results.

Although behavioural data are required to fully interpret the link between parental behaviour and chick CORT, our study suggests that little auk parents have some ability to modulate their provisioning effort, and that chicks increase CORT secretion when food is limited. The short high Arctic breeding season, and the existence of post-fledging parental care of little auk chicks may select for a fairly fixed developmental period and a more flexible parental investment strategy. These results contribute to accumulating information on the adrenocortical responses of chicks to nutritional stress (Sims & Holberton 2000; Wingfield & Kitaysky 2002; Kitaysky et al. 2003), and may help disentangle the various constraints that shape patterns of parental investment.

mass loss

Seasonal loss of body mass has been interpreted both as a short-term reproductive cost (e.g. Nur 1984), and an adaptation to increase flight economy (e.g. Freed 1981). It has been suggested that mass loss is an important adaptation in seabird species within the family Alcidae because of their exceptionally high wing loading and flight costs (Gaston & Jones 1989; Jones 1994). The adaptive mass loss hypothesis suggests that body mass loss is caused by intrinsic factors, independent of environmental processes. Although there was no significant difference in the amount of mass lost between clipped birds and their partners, the mechanism behind this loss differed between the two groups and clipped birds did lose slightly more mass on average. The drop in body mass of clipped parents most likely reflects the experimentally increased costs of flight, whereas the drop in body mass of the mates of clipped birds probably indicates an increase in their provisioning effort. These results provide strong evidence that changes in little auk body mass are not just caused by intrinsic or programmed processes, and that a reduction in body mass can reflect a change in parental effort. Although the increase in circulating CORT suggests that mass loss in both clipped birds and their mates coincided with nutritional limitations, the distinction between adaptive mass loss and reproductive stress is blurred because stress induced mass loss automatically results in a more efficient weight for flight (Ritz 2007). The loss of body mass in clipped birds and their partners in 2005 resulted in an 11% decrease in flight costs. Although this percentage is overestimated for the clipped birds because of their experimentally reduced wing area, there is likely to be an effect of the reduction in body mass given that even a 5% experimental increase in flight costs has been shown to affect parental effort (e.g. Velando & Alonso-Alvarez 2003). Mass loss could therefore be stress-induced, but the actual amount lost adjusted in response to the availability and predictability of prey, and the survival chances of the parent and their offspring. This threshold of body mass may therefore differ among individuals and populations.

compensatory parental care

In species where both parents care for their offspring there is often a conflict of interest between males and females over their respective parental contributions. Compensatory parental care should be especially important in long-lived birds that rear a single chick in a given breeding attempt, because lack of compensation could lead to breeding failure (e.g. Paredes et al. 2005). According to evolutionary stable strategy models, bi-parental care should only be stable when parents partially compensate for a reduction in care by the partner (e.g. Houston & Davies 1985; McNamara, Gasson & Houston 1999). Our results support this prediction.

There was no difference in breeding success between clipped and non-clipped birds, despite the differences in chick fledging condition. There is little evidence for a positive relationship between body mass at fledging and the post-fledging survival of alcid chicks (Harris & Rothery 1985; but see Gaston 1997). Moreover, fledgling size may not be as closely linked to post-fledging survival in little auks, which fledge at only 67–82% of adult body mass (Stempniewicz 2001) and receive some parental provisioning out at sea (Harding et al. 2004), compared to species that complete their growth at the nest-site. It is therefore possible that parents were compensating for their partner's reduction in care with the minimum effort required to ensure a successful breeding attempt. Conversely, there may be long-term costs associated with elevated CORT secretion during development (Love, Bird & Shutt 2003; Wada, Hahn & Breuner 2007), with even a small increase in CORT during the chick's development affecting quality (low growth efficiency and compromised cognitive abilities) later in life (Kitaysky et al. 2003, 2006). Further experimental study is required to examine parental investment decisions in relation to elevated CORT levels during chick development and effects on post-fledging offspring survival.

Conclusions

This experiment manipulated reproductive costs in a high Arctic seabird in the field, and measured three parameters indicative of a change in parental resource allocation between self-maintenance and current reproduction: change in the mass and CORT levels of (i) manipulated and unmanipulated parents, (ii) their partners, and (iii) chicks reared by these parents. This approach has highlighted both the importance of experimental design and the complexity of interpreting strategies of parental effort. For example, without measurement of the clipped partner's response we would have concluded that little auks have a fixed level of effort.

Our results best fit the flexible effort hypothesis, and suggest that little auks have some limited ability to adjust their reproductive effort in a given breeding attempt. This flexibility will presumably depend on the conditions of the breeding season and the parent's prospects of survival, and may therefore differ among individuals and populations. Breeding success in both our study years was high and suggests that local foraging conditions were fairly good; results may have been different under poorer foraging conditions, where parents are more constrained. Little auks breeding in poor environmental conditions may have fewer reserves, and therefore less ability or willingness to increase parental effort without a risk to their own survival. Conversely, birds breeding in consistently poor conditions may carry more reserves as an ‘insurance policy’. Further long-term manipulative studies of the fitness of individuals breeding under different environmental conditions are needed to more fully understand how parents optimize lifetime reproductive success in a stochastic environment.

Acknowledgements

Many thanks to H. Routti, J. Fort, R. Orben, E. Weston, M. Anne Pella-Donnelly, and K. Holser for their incredible energy and assistance in the field, and M. Munck, NANU Travel, C. Egevang, F. Delbart, T. and T. Fischbach, and R. and J. Harding for their essential logistical support. We thank Z. Kitaiskaia for her expertly performed hormonal assays, and J. Schmutz, and two anonymous reviewers for insightful comments on the manuscript. This project was funded by the French Polar Institute Paul-Emile Victor (Grant 388) to DG, and the National Science Foundation (grant 0612504 to AH and NJK), and was supported by USGS-Alaska Science Center. Additional funding was provided by NPRB grant #RO320 and NSF EPSCoR NSF # 0346770 to ASK. All field work in East Greenland was conducted with the permission of the Greenland Home Rule Government, Ministry of Environment and Nature (Danish Polar Center Scientific Expedition Permit 512-240), and under permits of the Ethics Committee of the French Polar Institute (MP/12/24/05/05). Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Ancillary