Do females trade copulations for food? An experimental study on kittiwakes (Rissa tridactyla)

Behavioral Ecology doi:10.1093/beheco/arl090 Advance Access publication 13 December 2006

Do females trade copulations for food? An experimental study on kittiwakes (Rissa tridactyla) Bart Kempenaers,a Richard B. Lanctot,b,c Verena A. Gill,b,c Scott A. Hatch,b and Mihai Valcua Max Planck Institute for Ornithology, Behavioural Ecology and Evolutionary Genetics Group, Postfach 1564, D-82305 Starnberg, Germany, bUS Geological Survey, 1011 East Tudor Road, Anchorage, AK 99503, USA, and cUS Fish and Wildlife Service, 1011 East Tudor Road, Anchorage, AK 99503, USA

a

ndividuals of many bird species copulate more frequently than necessary for fertilization (Birkhead and Møller 1992), despite the potential costs of this behavior (Daly 1978; Sheldon 1993). Thus, adaptive explanations have been sought for why males and females copulate with multiple partners or repeatedly with the same partner, during a single reproductive cycle (Westneat et al. 1990; Birkhead and Møller 1992; Hunter et al. 1993; Keller and Reeve 1995; Reynolds 1996; Jennions and Petrie 2000). For males, the issue is clear. Males increase their reproductive success by mating with multiple females because they can gain extrapair fertilizations and by mating multiply with their own mate because it reduces the risk of paternity loss (Birkhead and Møller 1992; Møller and Birkhead 1992). Why females copulate more than necessary for fertilization remains puzzling. One of the hypotheses discussed in the reviews cited above is that females obtain immediate material benefits by trading copulations for nutrients or other resources. In many animals, males provide nutrients to the female during the mating season. Courtship or nuptial feeding has been well studied, particularly in insects, and it has been suggested that female fitness is enhanced by nutrients derived from the male’s gift (Thornhill and Alcock 1983; Vahed 1998). Similarly, courtship feeding occurs in many bird species, and females may depend on the food brought by the male for the production of their eggs (Royama 1966; Tasker and Mills 1981; Lifjeld and Slagsvold 1986).

I

Address correspondence to B. Kempenaers. E-mail: b.kempenaers@ orn.mpg.de. Received 15 March 2006; revised 10 November 2006; accepted 11 November 2006.

Even though some observations suggest that females trade copulations for food (e.g., Mills 1994; Velando 2004; Tryjanowski and Hromada 2005) or for nesting material (Hunter and Davis 1998), the hypothesis that females copulate repeatedly with their own mate to obtain immediate material benefits (Hunter et al. 1993) has never been adequately tested. Although the observation that the frequency of copulation correlates with the frequency of courtship feeding (e.g., Mills 1994; Velando 2004) is predicted by the hypothesis (Hunter et al. 1993), it does not exclude alternative hypotheses. The correlation can also arise if females use the frequency of courtship feeding to assess the genetic or parental quality of their mate. Obtaining extra food must be costly for the male in terms of time and energy expenditure, and the frequency of courtship feeding could thus be an honest indicator of his quality or of his ability or willingness to invest in the current brood (Helfenstein et al. 2003). Thus, females might respond by copulating more frequently with males that provide more courtship feeding 1) because they trade copulations for food (immediate material benefits; Hunter et al. 1993), or 2) because they obtain genetic benefits for their offspring (indirect benefits; Jennions and Petrie 2000), or 3) because they increase the probability that the male will invest in the offspring (paternity assurance; Birkhead and Møller 1992). It is generally difficult to distinguish between direct and indirect benefits of female choice for males that provide food because, if the ability to provide courtship feeding is related to heritable variation in condition, high-quality males would provide better direct benefits and good genes (Kokko et al. 2003). To experimentally test the ‘‘immediate material benefits’’ hypothesis, Hunter et al. (1993) proposed to ‘‘manipulate the copulation frequency and record whether an


2006 The Authors This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

Females of many species copulate more frequently than necessary to fertilize their eggs despite the potential costs. Several studies, particularly on socially monogamous birds, have suggested that females obtain immediate material benefits by trading copulations for nutrients or other resources. We experimentally tested this hypothesis by manipulating the food resources available to prelaying female black-legged kittiwakes (Rissa tridactyla). If female kittiwakes trade copulations for courtship feeding because they need the extra resources, well-fed females (experimental group) should be less willing to copulate compared with females that are more food limited (control group). Contrary to our predictions, we found that close to the start of laying experimental females copulated more frequently with their mate than control females. We also observed that males from the experimental group fed their mate at least as often as males from the control group. In experimental pairs, we still observed a positive correlation between the rate of copulation and the rate of courtship feeding. Our results thus refute the immediate material benefits hypothesis. Currently available data are consistent with the hypothesis that prelaying courtship feeding is a form of mating effort. We suggest that the rate of courtship feeding might be a sexually selected trait, on which females base decisions about timing and frequency of copulations, but this remains to be tested. Key words: begging, courtship feeding, frequent copulation, material benefit hypothesis, mating effort, sexual conflict. [Behav Ecol 18:345–353 (2007)]

Behavioral Ecology

346

METHODS Study species and study area Black-legged kittiwakes are pelagic, colonial cliff-nesting seabirds (Cullen 1957; Danchin and Nelson 1991) that feed on small fish and marine invertebrates (Suryan et al. 2000). The species breeds along the coasts of the North Pacific and North Atlantic oceans and is particularly common in arctic and subarctic regions. Food availability has been shown to play an important role for breeding success (Roberts and Hatch 1993). We studied a population of breeding kittiwakes in 1998 on an abandoned US Air Force radar tower on Middleton Island (5828#N, 14619#W), Alaska. This 1295-hectare island is located in the north-central Gulf of Alaska about 80 km south of the entrance to the Prince William Sound and 18 km from the edge of the continental shelf (Gill et al. 2004). In 1994 and 1995, plywood paneling and ledges were added to the walls of the tower about 13–15 m above ground level. In total, about 1000 pairs bred on the tower in 1998, with almost every artificial ledge occupied by a breeding pair.

The tower contained a total of 144 artificial nest sites. Each nest site (ledge) was 35 cm wide by 24 cm deep and separated from neighboring sites by vertical pieces of plywood (20 cm high). Sliding panes of 1-way mirror glass (30 cm high by 26 cm wide) and plastic feeding tubes (20 cm long by 5 cm wide) were installed at each nest site. The feeding tubes, made from plastic pipe cut lengthwise to form a tray, slide back and forth through the wall, allowing food to be presented unobtrusively from inside the tower. The 1-way glass windows allowed us to observe each pair from close range from inside the tower, and a groove below each window could be used to capture individuals with wire hooks. Soon after arrival and long before nest completion, all adults were captured on the nest site and banded with a unique combination of color bands and a US Fish and Wildlife Service metal band. This allowed us to identify males and females resident at each nest site and identify the occasional intruder. Nest sites were checked each morning and each evening, and the amount of nesting material and the appearance of new eggs was noted. Each egg was measured and numbered with a felt-tip marker pen. Experimental procedure For this study, we chose 4 groups of 10 adjacent nest sites (2 horizontal rows of 5 windows). Within each group, nest sites were randomly assigned to the experimental or control treatment (i.e., fed or nonfed). All nest sites were occupied by a pair, but not all pairs ended up breeding, so the sample sizes were reduced to 19 experimental nests and 16 control nests. Food was supplied twice a day from 8 May onward (prelaying) until clutch completion. Birds were fed around 9:00 AM and 5:00 PM local time. Each feeding tube was carefully cleaned with water and bleach, rinsed in pure water, and filled with pieces of freshly thawed herring (Clupea harengus) or capelin (Mallotus villosus). Both fish species are a natural prey of breeding kittiwakes in the northern Gulf of Alaska and the Prince William Sound (Roby et al. 1999; Ainley et al. 2003). The diet was supplemented with vitamin B1 (thiamine) to offset losses of that nutrient associated with freezing (Crissey 1998). Each tube contained on average 400 g of fish and was placed into its slot in the wall so that it sat at one side of the nest site. Food from the tube was readily taken by both pair members. An earlier study (1996–1997) showed that supplemented kittiwake pairs consumed on average 292–324 g of fish prior to laying and on average 258–282 g during laying (Gill 1999). Thus, the amount of fish we offered was rarely completely consumed during the prelaying or laying period. In other words, the experimental pairs were fed ad libitum. We never observed males directly feeding the supplied fish to the female. To safeguard against food stealing by birds other than the experimental pair, we constructed vertical dividers to separate each nest site. An analysis of regurgitated food by control and experimental birds (in 1996 and 1997) showed that food stealing by unfed birds was absent or unimportant (Gill 1999). Behavioral observations Observations started on 11 May and finished on 12 June. From 11 May onward until they finished egg laying, each pair was observed every day for 2 h, 1 h in the morning (between 7:00 and 10:00 AM) and 1 h in the afternoon (between 3:00 and 6:00 PM). The low frequency of begging, courtship, and copulation behavior and the close proximity of the nests allowed each observer to watch a group of 10 pairs (windows) simultaneously. In total, we conducted 1843 hours of observations, but we could use only 1337 observation hours because

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

increase/decrease in copulation frequency results in an increase/decrease in material benefits for the female.’’ We suggest that a better (and perhaps also a more practicable) test of the hypothesis is to experimentally vary the amount of resources available to females. We performed an experiment on black-legged kittiwakes (Rissa tridactyla), manipulating the food availability of prelaying females. In this species, males frequently feed their mate before egg laying, and copulations often occur immediately after courtship feeding (Chardine 1987; Neuman et al. 1998). When the male arrives at the nest site, females are often observed begging for food. Observational and experimental studies (food supplementation) have shown that the amount of food available to the female during the prelaying period influences the timing of laying, clutch size, egg mass, the mean duration of incubation shifts, and hatching and fledging success (Gill and Hatch 2002; Gill et al. 2002). This suggests that the food delivered by the male might form an important proportion of the female’s diet. The kittiwake is a long-lived, socially monogamous species, with low frequencies of extrapair paternity and divorce (Hatch et al. 1993; Helfenstein et al. 2004). Nevertheless, there is an intersexual conflict over the frequency of copulation (females refuse cloacal contact in about half of the male copulation attempts, see below) and over sperm ejection (Helfenstein et al. 2003). Even in the absence of extrapair paternity, it is likely that the risk of losing paternity selected for frequent copulation as a paternity assurance mechanism in this and other seabirds with a similar life history (Hunter et al. 1992; Helfenstein et al. 2004). Males that copulate frequently with their female have a higher probability to fertilize the eggs, because it is more likely 1) that they obtained most copulations and hence transferred most sperm and 2) that they achieved the last copulations before egg laying (see Hunter et al. 1992). We provided a group of kittiwake pairs with food during the prelaying and laying period (experimental group), whereas another group of pairs did not receive any food (control group). If female kittiwakes trade copulations for courtship feeding because they benefit from the extra resources, wellfed experimental females should be less willing to copulate compared with control females that are more food limited, assuming that copulations are costly and that the costs are similar for the 2 groups. The immediate material benefits hypothesis predicts 1) that control females copulate at a higher frequency compared with fed females and 2) a positive relationship between the frequency of courtship feeding and copulation for control pairs but not for experimental pairs.

Kempenaers et al.

•

Courtship feeding and copulation

Data analysis Each of the relevant behaviors (begging, courtship feeding, and copulation) can only occur when both pair members are present at the nest. Moreover, each of the behaviors rarely occurred more than once within a 5-min period. Therefore, we used generalized linear mixed models (GLMM) with a binomial error structure (as described in Venables and Ripley 2002) with the frequency of the behavior as the dependent variable and the number of 5-min periods in which both pair members were present at the nest site on a given day as the binomial denominator. When comparing experimental and control pairs, we can exclude a bias due to the time of day because control and experimental pairs were always observed simultaneously, both during the morning and during the afternoon session. Furthermore, the effect of time of day on the frequency of the behaviors did not differ between the treatments (interaction between time of day as a fixed factor with 2 levels and treatment: all P . 0.10). Thus, for further analyses, we combined the data from the morning and afternoon sessions. The first observations of a pair were made between 34 and 15 days prior to the start of laying (i.e., day –34 and –15, where day 0 is the day on which the female laid the first egg). We used all observations until day 0, because the frequency of the behaviors dropped dramatically after this day (e.g., no copulations were observed after day 0). To account for changes in behavior over the prelaying period, we used day relative to the start of egg laying (egg-date) as a covariate in all models. Because the experiment was set up as a block design (4 blocks of 10 adjacent nest sites, with 5 experimental and 5 control pairs each), initial models contained ‘‘pair nested within block’’ as a random factor. We used Akaike’s Information Criterion (AIC) to choose between this and a simplified model containing only pair identity as a random factor (Buckland et al. 1997; Johnson and Omland 2004). AIC measures the lack of model fit to the data while controlling for the number of parameters in the model. It can be calculated as AIC ¼ deviance 12(number of parameters), whereby the second term provides a penalty for models with more parameters. The 2 models (pair nested within block vs. pair) did not

Figure 1 Correlation between the courtship feeding rate and the copulation rate for 35 pairs of black-legged kittiwakes (Spearman rank correlation, rs ¼ 0.44, P ¼ 0.008; added for comparative purposes, see Table 2). Larger circles indicate 2 overlapping data points.

differ significantly (likelihood ratio tests; time pair present: v2 ¼ 0.86, P ¼ 0.35; begging rate: v2 ¼ 0.07, P ¼ 0.80; courtship feeding rate: v2 ¼ 0.00, P ¼ 0.99; number of copulation attempts: v2 ¼ 1.95, P ¼ 0.35; copulation success: v2 ¼ 0.01, P ¼ 0.91; df ¼ 1 in all cases). Thus, we preferred the model with the lower AIC value, which was always the model with pair identity as a random factor. The models described below thus included pair as a random factor, egg-date as a covariate, and treatment (experiment/control) as a fixed factor. To check for a treatment effect, we started by testing for an interaction between egg-date and treatment. Only if the interaction was not significant (P . 0.10), did we test the main effects of treatment and egg-date (Crawley 2002). All statistical analyses were performed with the free software R (R Development Core Team 2005). GLMM models were fitted using the glmmPQL function of the add-on package MASS (Venables and Ripley 2002). Likelihood ratio tests were performed with the add-on package lme4. Data shown are mean 6 standard error. RESULTS In total, we observed 150 copulation attempts, of which 74 (49.3%) were successful, that is, led to cloacal contact. We also observed a total of 555 begging events and 146 courtship feeds. Female kittiwakes copulated on average 41.0 6 4.9 times per clutch (N ¼ 35; estimated by extrapolation assuming 14 h of daylight), and each copulation consisted of 3.9 6 0.2 cloacal contacts. Only 18.6% of all copulation attempts (N ¼ 150) and 13.3% of successful copulations (N ¼ 74) were closely linked in time to a courtship feed (i.e., observed within the same 5-min period). Overall, the copulation rate correlated positively with the rate of courtship feeding (Figure 1; GLMM with copulation frequency as the dependent variable: F1,602 ¼ 24.79, P ¼ 0.0001), but this relationship did not differ between experimental and

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

some pairs disappeared before laying or did not lay eggs (see above). During every observation period, we confirmed the identity of each of the pair members by their unique color bands. At the start of each observation, and at 5-min intervals thereafter, we noted whether the male or the female (or both) were present at the nest site. Throughout the 1-h focal, we recorded each of the following behaviors every time it occurred: 1) begging by the female (moving the head up and down and pecking at the bill of the male), 2) courtship feeding by the male, 3) copulation attempt (male mounts female but does not achieve cloacal contact), 4) copulation (with cloacal contact). Because we could watch the birds on the ledge from a distance of less than 50 cm, we could accurately score whether cloacal contact occurred. A copulation often consisted of bouts of cloacal contacts and we also counted the number of cloacal contacts in each bout. No male was ever observed trying to forcefully attempt a copulation, and no extrapair copulations were observed. Some female cooperation was required even for a copulation attempt because the female had to sit down to allow the male to mount. The banded birds from the experimental and control group were never observed courtship feeding or copulating away from their nest site, despite observations of individuals on the beach and on the roofs of nearby buildings (see also Mills 1994).

347

Behavioral Ecology

348

of the number of cloacal contacts per successful copulation (Figure 3d; Table 1). DISCUSSION Courtship feeding and copulation: evidence for the material benefits hypothesis

control pairs (treatment: F1,33 ¼ 1.79, P ¼ 0.19; courtship feeding rate 3 treatment: F1,601 ¼ 1.14, P ¼ 0.29). Courtship feeding within the same 5-min period did not increase the probability that a copulation would be successful, either in experimental birds (F1,66 ¼ 0.00001, P ¼ 0.99) or in control birds (F1,40 ¼ 0.10, P ¼ 0.76). The time both members of the pair spent together at the nest site decreased as laying approached but did not differ between the experimental and control pairs (Figure 2; Table 1). Shortly after the start of the experiment, females from control pairs begged more often for food than females from experimental pairs (Figure 3a), as expected because the latter received food ad libitum. However, this difference disappeared closer to egg laying (significant egg-date 3 treatment interaction; Table 1). Overall, the begging frequency strongly increased toward the start of laying. Surprisingly, fed females begged as much as unfed females closer to egg laying during the presumed fertile period (Figure 3a). The frequency of courtship feeding also changed significantly with egg-date: males fed their females more often closer to egg laying. Despite higher courtship feeding by males in the experimental group just prior to the start of laying, there was no overall significant difference between the control and experimental pairs (Figure 3b; Table 1). The frequency of copulation attempts, the proportion of copulation attempts that were successful, and the number of cloacal contacts all increased toward the start of laying. When all copulation attempts were considered (Figure 3c), fed females copulated more often than control females closer to laying but not earlier (significant egg-date 3 treatment interaction; Table 1). Males from experimental and control pairs were equally successful, both in terms of the proportion of copulation attempts that led to cloacal contact and in terms

Potential shortcomings of the experimental design Instead of outright rejecting the immediate material benefits hypothesis, we consider potential reasons why our experiment

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

Figure 2 Proportion of the total observation time that both pair members were present at the nest site together (fed pairs, N ¼ 19; control pairs, N ¼ 16) in relation to egg-date (day 0 ¼ start of egg laying). Shown are mean 6 standard error.

Hunter et al. (1993) proposed that females might obtain direct benefits by trading copulations for food or other resources. The ‘‘immediate material benefits’’ hypothesis predicts that females in pairs with a high copulation frequency gain more provisions (Hunter et al. 1993). The hypothesis gained support based on 5 types of observations. We summarized the support for 4 of these observations, based on a comprehensive literature review, in Table 2. 1) A positive correlation has been found between the frequency of courtship feeding and the frequency of within-pair copulation. 2) Copulations are often directly preceded by courtship feeding. In 2 studies, significantly more extrapair than within-pair copulations were preceded by courtship feeding (Table 2). Although the number of observed extrapair copulations is small, this suggests that females might trade extrapair copulations for resources (see also Hunter and Davis 1998). 3) In some studies, copulations preceded by courtship feeding were more likely to be successful than copulations without courtship feeding. 4) Copulations preceded by courtship feeding often resulted in a higher number of cloacal contacts per copulation or lasted longer, presumably, leading to increased sperm transfer. 5) Copulations preceded by courtship feeding with high-quality prey items were more likely to be successful. This was shown in 2 studies: one on the great grey shrike Lanius excubitor (Tryjanowski and Hromada 2005) and one on the osprey Pandion haliaetus (Mougeot et al. 2002). In the latter, both the fact that food was brought and its quality (fish size) influenced copulation success. In contrast, in the common tern Sterna hirundo, there was no relationship between the size of the fish delivered and the likelihood of copulation or copulation success (number of cloacal contacts; Wiggins and Morris 1986). Overall, these observations do not contradict the hypothesis that females trade copulations for food because they benefit from the nutrients, but they do not critically test it either. We presented the results of the first experimental test of the hypothesis. Contrary to the predictions from the hypothesis, we found 1) no reduced copulation frequency in experimental (fed) compared with control (nonfed) pairs and 2) a similar positive relationship between courtship feeding frequency and copulation frequency in experimental and control pairs (Figure 1). Thus, our results clearly refute the immediate material benefits hypothesis. We showed that even though fed females initially begged at a lower rate, their begging rate increased and became similar to that of control females toward the start of egg laying (Figure 3a). Moreover, in the 10-day period before egg laying, experimentally fed females received more courtship feeds (Figure 3b) and allowed their mate to mount more often than control females (Figure 3c). Because there was no difference in copulation success or in the number of cloacal contacts between control and experimental females, fed females copulated more frequently with their social mate, particularly during the fertile period (close to laying). We now discuss the potential weaknesses of our experimental design and some alternative interpretations of our results and of those reported previously.

Kempenaers et al.

•

Courtship feeding and copulation

349

Table 1 Description of GLMM describing the effects of the feeding treatment and date relative to the first egg (egg-date) on the presence at the nest site and on begging, courtship feeding, and copulation behavior of black-legged kittiwakes Explanatory variable

Parameter estimate

Standard error

Time pair presenta

Intercept Treatmentb Egg-date Egg-date 3 treatment Intercept Treatmentb Egg-date Egg-date 3 treatment Intercept Treatmentb Egg-date Egg-date 3 treatment Intercept Treatmentb Egg-date Egg-date 3 treatment Intercept Treatmentb Egg-date Egg-date 3 treatment Intercept Treatmentb Egg-date Egg-date 3 treatment

0.25 0.03
0.011 0.004
1.77 0.09 0.05 0.05
2.63 0.36 0.14 0.005
4.04 1.08 0.07 0.07
0.09
0.40 0.06
0.04 1.53 0.02 0.02 0.002

0.05 0.07 0.002 0.004 0.18 0.25 0.01 0.02 0.35 0.46 0.03 0.045 0.30 0.37 0.03 0.03 0.26 0.33 0.02 0.02 0.13 0.17 0.01 0.013

Begging

Courtship feeding

Copulation attempts

Copulation successd

Number of cloacal contacts

a b c d

F

P

F1,62 ¼ 0.50 F1,64 ¼ 25.56 F1,62 ¼ 0.97

0.48 ,0.0001 0.33

Not testedc Not testedc F1,625 ¼ 10.88

0.001

F1,33 ¼ 1.09 F1,626 ¼ 43.42 F1,625 ¼ 0.012

0.30 ,0.0001 0.91

Not testedc Not testedc F1,625 ¼ 6.92

0.009

F1,31 ¼ 0.0003 F1,98 ¼ 8.89 F1,97 ¼ 2.30

0.99 ,0.0001 0.13

F1,26 ¼ 0.0004 F1,44 ¼ 12.50 F1,44 ¼ 0.03

0.98 0.001 0.86

Time both pair members were present at the nest site together. Estimate relative to control group (¼ 0). Main effects were not tested when the interaction was significant. The proportion of copulation attempts that led to cloacal contact. GLMM with the number of successful copulations as the dependent variable, and the number of copulation attempts as the binomial denominator.

might have failed to detect a reduction in the copulation frequency in experimentally fed females. The experimental manipulation might not have worked as expected because of the way in which an animal perceives the manipulation and uses the information. We assumed that females assess their nutritional needs based on the amount of food available, even if that food is provided in a tube at the nest and that they respond accordingly. This assumption seems reasonable, but it is hard to verify. Food obtained via courtship feeding might have been of higher quality or might have had unknown qualities that were missing from the experimentally provided food. We attempted to circumvent this problem by 1) feeding females with frozen fish of species they also obtain via courtship feeding (see Methods) and 2) supplementing freezing-induced loss of nutrients (vitamin B1). However, we cannot exclude that females obtained a more diverse diet including other essential nutrients through courtship feeding. The observation that fed females continued begging at a high rate (Figure 3a) seems to support the above explanations, but an equally plausible alternative interpretation for this observation is presented below. Alternatively, the food availability for kittiwakes foraging in the Prince William Sound in 1998 might have been so good that experimental feeding had little effect on the nutritional needs of females. Although we have no detailed information on food availability or on the time spent foraging by control and experimental birds in 1998, this seems unlikely for 2 reasons. 1) Pacific kittiwakes show a lower productivity than their Atlantic conspecifics, and there is circumstantial evidence that this is due to a deficient food supply (Gill and Hatch 2002 and references therein). 2) Experimentally provided food has a positive effect on reproduction. In 1998, fed females laid

their eggs significantly earlier (larger data set, experiment: 4 June 6 0.6 days, N ¼ 121, control: 9 June 6 0.7 days, N ¼ 79, t ¼
5.56, P , 0.0001), and they produced a significantly larger clutch (experimental: 1.81 6 0.04, N ¼ 121, control: 1.67 6 0.05, N ¼ 81, t ¼ 2.20, P ¼ 0.029) with heavier eggs (experimental: 49.5 6 0.4 g, N ¼ 121, control: 48.5 6 0.4 g, N ¼ 81, t ¼ 2.02, P ¼ 0.045; also see Gill and Hatch 2002; Gill et al. 2002). Given the differences in clutch size and egg mass between the 2 treatments, one could argue that experimental females copulated more frequently to ensure that their eggs were fertilized either because they needed to fertilize more eggs or because their eggs were more valuable. However, clutch size and mean egg mass did not explain a significant amount of the variation in copulation frequency either on their own (clutch size: F1,30 ¼ 0.14, P ¼ 0.71; mean egg mass: F1,30 ¼ 0.02, P ¼ 0.89) or in interaction with egg-date or treatment (all P . 0.5). Frequent copulation and the relationship with courtship feeding We now consider alternative explanations for why females copulate frequently with their partner and why this is related to the level of courtship feeding. Both our experimental data and the observational data from the literature (Table 2) clearly suggest that there is a positive association between the rate or quality of courtship feeding and the frequency of copulation. The relationship between the frequency of courtship feeding and copulation might not be causal. Both behaviors occur at the nest site, so a positive correlation could simply arise

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

Behavior

350

Behavioral Ecology

because of variation in the amount of time males and females spend at the nest. This would also explain why courtship feeding and copulation were not tightly temporally associated. In general, this is possible, but it cannot explain our results because we controlled for the amount of time the pair spent together at the nest site in our analyses (see Methods; note that the results remain similar when we did not control for time spent together). Furthermore, the time pair members spent together at the nest site did not differ between the control and experimental group (Figure 2). A positive relationship between the frequency of courtship feeding and copulations might simply be a consequence of food availability. When food is abundant, providing courtship feeding would be less energetically costly for males, and copulating would be less energetically costly for females. This could explain why in the experimental pairs courtship feeding and copulations occur at a higher frequency (at least close to laying) than in control pairs. This hypothesis can be tested with a modified experimental design where only females get access to the food. The hypothesis then predicts that experimental and control pairs do not differ in the frequency of courtship feeding, but experimental females should still copulate at a higher frequency. Our data are also in accordance with the hypothesis that females use the amount of courtship feeding as a signal of

their mate’s genetic or parental quality. If true, it would explain why experimental females copulated more frequently than females from control pairs close to laying (Figure 3b,c). If males that provide more courtship feeding are of higher quality, then females might copulate more frequently with them for several reasons. 1) Females might use frequent copulation to prevent their mate from copulating with other females (female mate guarding hypothesis; Petrie 1992). 2) Females might want to ensure that their high-quality mate fertilizes all the eggs (indirect benefits hypothesis; Jennions and Petrie 2000). This might be relevant if females are forced to copulate with other males or if the cost of refusing to copulate is high (e.g., because of male aggression). Although we did not observe any extrapair copulations, extrapair copulations have been documented in another population of the black-legged kittiwake (Helfenstein et al. 2004). 3) Females might want to assure their mate that he fathered all her offspring, which might positively influence his subsequent paternal investment (paternity insurance hypothesis; Birkhead and Møller 1992). This would be particularly relevant if the rate of courtship feeding signals parental quality or ability or willingness to invest in the brood. Our data do not allow us to differentiate between these hypotheses, but the observation that the copulation rate and the number of cloacal contacts peaked during the fertile period (close to the start of laying)

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

Figure 3 Behavior of experimental (fed, N ¼ 19) and control (nonfed, N ¼ 16) pairs of black-legged kittiwakes in relation to eggdate (day 0 ¼ start of egg laying). Shown are mean 6 standard error. (a) Female begging rate, (b) courtship feeding rate, (c) copulation attempt rate, and (d) number of cloacal contacts per successful copulation.

Kempenaers et al.

•

Courtship feeding and copulation

351

Table 2 Review of avian studies on the relationship between CF and WP or EP copulations

Type

Common tern Sterna hirundo

WP WP WP EP WP WP WP EP WP WP WP WP WP EP WP WP WP

Red kite Milvus milvus Montagu’s harrier Circus pygargus Hen harrier Circus cyaneus African marsh harrier Circus ranivorus Lesser kestrel Falco naumanni American kestrel Falco sparverius Feral fowl Gallus gallus domesticus Great grey shrike Lanius excubitor

Yellow-legged gull Larus cachinnans Red-billed gull Larus novaehollandiae scopulinus Lesser black-backed gull Larus fuscus Kittiwake Rissa tridactyla Great skua Catharacta skua Osprey Pandion haliaetus

a b c d

Copulation and CF frequency positively correlated

Copulation attempts preceded by CF more likely successful

More cloacal contacts when copulations preceded by CF

No — —

No — —

1 2 3

— Yesb Yes No No Yesd No No

Yes — —

4 5 6

— No No No

7 8 9 10

%

N 82 28 74 2 93 352 292 73 138b 107 1247 74 82 3 168 385 210

—

—

11

— No (r ¼ 0.02, N ¼ 22) Yes

35 50 3 100a 42 81 31 0 35 40 21 13 26 100a 47 16 33

No No Yes

12 13 14

WP WP WP WP

Yes Yes — No (r ¼ 0.15, N ¼ 10)

53 28 73 50

86 98 59 125

— — — Yes

WP WP — WP EP

Yes No Yes

56 — — 73 100

387 — — 22 22

— — —

— — Yes (copulation duration) — — — No (copulation duration) — — —

19 20 21

Yes (food quality)

—

22

— Yes (rs ¼ 0.51, N ¼ 15) Yes (rs ¼ 0.51, N ¼ 41) Yes (rs¼ 0.71, N ¼ 21) — Yesc (rs ¼ 0.45, N ¼ 10) — Yes No Yes (rs ¼ 0.44, N ¼ 35) —

—

Ref.

15 16 17 18

CF, courtship feeding; WP, within-pair copulation; EP, extrapair copulation. Significant difference between within- and extrapair copulations. Statistical analyses suffer from pseudoreplication. Not significant (P ¼ 0.19). Combined with data from herring gull Larus argentatus. 1, Wiggins and Morris (1986); 2, Blanchard and Morris (1998); 3, Gonza´lez-Solı´s et al. (2001); 4, Velando (2004); 5, Tasker and Mills (1981); 6, Mills (1994); 7, Brown (1967); 8, Neuman et al. (1998); 9, Helfenstein et al. (2003); 10, This study; 11, Catry and Furness (1997); 12, Birkhead and Lessells (1988); 13, Green and Krebs (1995); 14, Mougeot et al. (2002); 15, Mougeot (2000); 16, Arroyo (1999); 17, Picozzi (1984); 18, Simmons (1990); 19, Dona´zar et al. (1992); 20, Villarroel et al. (1998); 21, Pizarri (2003); 22, Tryjanowski and Hromada (2005).

supports the second and the third hypotheses. Note that the observations summarized in Table 2, although often used as evidence for the immediate material benefit hypothesis, also support these alternative hypotheses. The hypothesis that courtship feeding signals male quality is also supported by other observational studies on kittiwakes (Helfenstein et al. 2003) and on other birds (see below). Helfenstein et al. (2003) showed that the rate of courtship feeding was individually repeatable between years and negatively related to male arrival date, suggesting that the rate of courtship feeding is a reliable index of male quality. Furthermore, males with a higher courtship-feeding rate were paired to females that laid larger clutches and repaired faster in the next year, suggesting that these males are more attractive to females (Helfenstein et al. 2003). Similarly, in the red-billed gull Larus novaehollandiae, females that received a high number of courtship feedings were less likely to engage in extrapair copulations and to divorce the next season (Mills 1994), and in the feral fowl (Gallus gallus domesticus), courtship feeding rate was positively correlated with male social status (Pizarri 2003). Interestingly, in the kittiwake, male courtship feeding rate does not always correlate with male chick-feeding

rate, suggesting that courtship feeding is not necessarily a signal of male parental quality or ability (Helfenstein et al. 2003; but see Niebuhr 1981; Wiggins and Morris 1986; Green and Krebs 1995). The puzzling observation that experimentally fed females begged for food as often as control females, particularly close to egg laying (Figure 3a), is consistent with the hypothesis that females beg to assess the quality of their mate via courtship feeding because this should be independent of their own nutritional status. Courtship feeding and copulation: the male perspective Our experimental results can also shed some light on the question why males invest in courtship feeding. Two not mutually exclusive hypotheses have been considered to explain why males provide food to females during the mating period (reviewed in Vahed 1998). 1) Courtship feeding could function to increase the fitness of the male’s offspring via investment in the female. Feeding the female might increase the number and/or quality of the offspring she can produce and could thus be seen as a form of paternal investment (paternal

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

Species

Copulations preceded by CF

Behavioral Ecology

352

CONCLUSION Our experimental results contradict observational studies that concluded that females manipulate males to obtain immediate material benefits (e.g., Neuman et al. 1998; Gonza´lez-Solı´s et al. 2001; Velando 2004; Tryjanowski and Hromada 2005). Although females obviously obtain material benefits via courtship feeding, we found no support for the hypothesis that reducing the female’s nutritional need results in a lower copulation frequency. Our results do not contradict the hypothesis that the rate of courtship feeding during the (pre) fertile period is a sexually selected trait and an honest indicator of male quality, but this needs to be further tested. In socially monogamous species, the rate of courtship feeding might influence female decisions regarding the timing of pairing, divorce, and within- and extrapair copulations (Mills 1994; Helfenstein et al. 2003; Pizarri 2003).

We thank Cheryl Bishop and Shiloh Shulte for conducting behavioral observations, and Naomi Bargmann, Cheryl Bishop, Elke Claus, Laurent LeBourg, Anoma Patirana, Shiloh Shulte, and Brigid Toner for help in the field. We are very grateful to the Federal Aviation Administration for allowing us to use their shower and laundry facilities and for letting us ‘‘hitch a flight’’ off the island. The constructive comments from an anonymous reviewer greatly improved the manuscript. R.B.L. was supported by a postdoctoral fellowship of the Belgian National Fund for Scientific Research (held at the University of Antwerp, Belgium). This study was supported by the Konrad Lorenz Institute for Comparative Ethology in Vienna (B.K.) and by the US Geological Survey’s Alaska Science Center (R.B.L., V.A.G., and S.A.H.). Funding to pay the Open Access publication charges for this article was provided by the Max Planck Society.

REFERENCES Ainley DG, Ford RG, Brown ED, Suryan RM, Irons DB. 2003. Prey resources, competition, and geographic structure of kittiwake colonies in Prince William Sound. Ecology. 84:709–723. Arroyo BE. 1999. Copulatory behavior of semi-colonial Montagu’s harriers. Condor. 101:340–346. Birkhead TR, Lessells CM. 1988. Copulation behaviour of the osprey Pandion haliaetus. Anim Behav. 36:1672–1682. Birkhead TR, Møller AP. 1992. Sperm competition in birds. Evolutionary causes and consequences. London: Academic Press. Blanchard L, Morris RD. 1998. Another look at courtship feeding and copulation behavior in the common tern. Colon Waterbirds. 21:251–255. Brown RGB. 1967. Courtship behaviour in the lesser black-backed gull, Larus fuscus. Behaviour. 29:122–153. Buckland ST, Burnham KP, Augustin NH. 1997. Model selection: an integral part of inference. Biometrics. 53:603–618. Catry P, Furness RW. 1997. Territorial intrusions and copulation behaviour in the great skua, Catharacta skua. Anim Behav. 54:1265– 1272. Chardine JW. 1987. The influence of pair-status on the breeding behaviour of the kittiwake Rissa tridactyla before egg-laying. Ibis. 129:515–526. Crawley MJ. 2002. Statistical computing: an introduction to data analysis using S-plus. West Sussex (UK): Wiley. p. 761. Crissey SD. 1998. Handling fish fed to fish-eating animals: a manual of standard operating procedures. Washington (DC): U.S. Department of Agriculture, Agriculture Research Service, National Agriculture Library. Cullen E. 1957. Adaptations in the kittiwake to cliff-nesting. Ibis. 99:275–302. Daly M. 1978. The cost of mating. Am Nat. 112:771–774. Danchin E, Nelson JB. 1991. Behavioral adaptations to cliff nesting in the kittiwake (Rissa tridactyla): convergence with the gannet (Sula bassana) and the black noddy (Anous tenuirostris). Colon Waterbirds. 14:103–107. Dona´zar JA, Negro JJ, Hiraldo F. 1992. Functional analysis of matefeeding in the lesser kestrel Falco naumanni. Ornis Scand. 23: 190–194. Gill VA. 1999. Breeding performance of black-legged kittiwakes (Rissa tridactyla) in relation to food availability: a controlled feeding experiment [Master’s thesis]. Anchorage (AK): University of Alaska. Gill VA, Hatch SA. 2002. Components of productivity in black-legged kittiwakes Rissa tridactyla: response to supplemental feeding. J Avian Biol. 33:113–126. Gill VA, Hatch SA, Lanctot RB. 2002. Sensitivity of breeding parameters to food supply in black-legged kittiwakes Rissa tridactyla. Ibis. 144:268–283. Gill VA, Hatch SA, Lanctot RB. 2004. Colonization, population growth and nesting success of black oystercatchers following a seismic uplift. Condor. 106:791–800. Gonza´lez-Solı´s J, Sokolov E, Becker PH. 2001. Courtship feedings, copulations and paternity in common terns, Sterna hirundo. Anim Behav. 61:1125–1132. Green DJ, Krebs EA. 1995. Courtship feeding in ospreys Pandion haliaetus: a criterion for mate assessment? Ibis. 137:35–43. Hatch SA, Roberts BD, Fadely BS. 1993. Adult survival of black-legged kittiwakes Rissa tridactyla in a Pacific colony. Ibis. 135:247–254.

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

investment hypothesis; Ligon 1999). In our experiment, we would then predict that male kittiwakes do not take the supplemental food or simply give it to the female. In contrast, we observed that males from experimental pairs ate some of the food we offered, but they never directly fed it to the female. We would also predict a lower courtship-feeding rate for females that have ad libitum access to food than for control pairs because the former require less paternal investment. We rather found the opposite pattern in our study (Figure 3b). Similarly, male ospreys did not significantly reduce their own courtship-feeding rate in response to food supplementation of the female (Green and Krebs 1995). 2) Courtship feeding could function to attract females, to obtain copulatory access to females, or to maximize ejaculate transfer. Thus, courtship feeding could be a form of mating effort (mating effort hypothesis; Vahed 1998 and references therein). If so, we would expect a higher frequency of courtship feeding in males of experimental pairs because these males had easy access to food (assuming that courtship feeding is costly). As predicted, in the period closest to egg laying, experimental males fed their female more often than control males (Figure 3b), although this effect was not significant (Table 1). Thus, our data suggest that courtship feeding in kittiwakes might be a form of mating effort not paternal effort. The distinction between these hypotheses is not trivial for understanding the evolution of mating behavior. Under the paternal investment hypothesis, courtship feeding is a form of cooperation between the parents, whereas under the mating effort hypothesis it reflects a potential conflict of interest between the male and the female. A recent review of nuptial feeding in insects (Vahed 1998) suggested that there is little support for the paternal investment hypothesis in this taxonomic group. Rather, the amount of nuptial feeding seems to be selected in response to the risk of sperm competition because females provided with more food (e.g., larger spermatophores) allow males to copulate longer and transfer more sperm. Even though it has been demonstrated that females receiving more or larger gifts produce more or heavier eggs, this could be manipulation of female fecundity by the male via hormonal substances in the ejaculate, which has rarely been controlled for (Vahed 1998). Because most birds are socially monogamous and offspring fitness strongly depends on parental care, males and females might have more similar interests (Ligon 1999). Thus, Ligon wrote that courtship feeding is ‘‘One of the most common forms of parental investment by males, apart from care of offspring.’’ However, mating decisions such as engaging in extrapair copulations and divorce may considerably increase the conflict between the social partners and can be related to the level of courtship feeding (Mills 1994; Helfenstein et al. 2003).

Kempenaers et al.

•

Courtship feeding and copulation

Picozzi N. 1984. Breeding biology of polygynous hen harriers Circus c. cyaneus in Orkney. Ornis Scand. 15:1–10. Pizarri T. 2003. Food, vigilance, and sperm: the role of male direct benefits in the evolution of female preference in a polygamous bird. Behav Ecol. 14:593–601. Reynolds JD. 1996. Animal breeding systems. Trends Ecol Evol. 11: 68–72. Roberts BD, Hatch SA. 1993. Behavioral ecology of black-legged kittiwakes during chick rearing in a failing colony. Condor. 95:330–342. Roby DD, Jodice PGR, Turco KR. 1999. Diet composition, reproductive energetics, and productivity of seabirds damaged by the Exxon Valdez oil spill, 1998. Exxon Valdez oil spill restoration project annual report (restoration project 98163G). Corvallis (OR): Oregon State University, Oregon Cooperative Fish and Wildlife Research Unit, Department of Fisheries and Wildlife. Royama T. 1966. A re-interpretation of courtship feeding. Bird Study. 13:116–129. Sheldon BC. 1993. Sexually transmitted disease in birds: occurrence and evolutionary significance. Philos Trans R Soc Lond B Biol Sci. 339:491–497. Simmons RE. 1990. Copulation patterns of African marsh harriers: evaluating the paternity assurance hypothesis. Anim Behav. 40:1151–1157. Suryan RM, Irons DB, Benson J. 2000. Prey switching and variable foraging strategies of black-legged kittiwakes and the effect on reproductive success. Condor. 102:374–384. Tasker CR, Mills JA. 1981. A functional analysis of courtship feeding in the red-billed gull, Larus novaehollandiae scopulinus. Behaviour. 77:221–241. Thornhill R, Alcock J. 1983. The evolution of insect mating systems. Cambridge (MA): Harvard University Press. Tryjanowski P, Hromada M. 2005. Do males of the great grey shrike, Lanius excubitor, trade food for extrapair copulations? Anim Behav. 69:529–533. Vahed K. 1998. The function of nuptial feeding in insects—review of empirical studies. Biol Rev. 73:43–78. Velando A. 2004. Female control in yellow-legged gulls: trading paternity assurance for food. Anim Behav. 67:899–907. Venables WN, Ripley BD. 2002. Modern applied statistics with S. 4th ed. New York: Springer. Villarroel M, Bird DM, Kuhnlein U. 1998. Copulatory behaviour and paternity in the American kestrel: the adaptive significance of frequent copulations. Anim Behav. 56:289–299. Westneat DF, Sherman P, Morton ML. 1990. The ecology and evolution of extra-pair copulations in birds. Curr Ornithol. 7:331–369. Wiggins DA, Morris RD. 1986. Courtship feeding and copulatory behaviour in the common tern Sterna hirundo. Ornis Scand. 19: 163–165.

Downloaded from http://beheco.oxfordjournals.org/ by guest on August 6, 2015

Helfenstein F, Tirard C, Danchin E, Wagner RH. 2004. Low frequency of extra-pair paternity and high frequency of adoption in blacklegged kittiwakes. Condor. 106:149–155. Helfenstein F, Wagner RH, Danchin E, Rossi J-M. 2003. Functions of courtship feeding in black-legged kittiwakes: natural and sexual selection. Anim Behav. 65:1027–1033. Hunter FM, Burke T, Watts SE. 1992. Frequent copulation as a method of paternity assurance in the northern fulmar. Anim Behav. 44: 149–156. Hunter FM, Davis LS. 1998. Female Adelie penguins acquire nest material from extrapair males after engaging in extrapair copulations. Auk. 115:526–528. Hunter FM, Petrie M, Otronen M, Birkhead T, Moller AP. 1993. Why do females copulate repeatedly with one male? Trends Ecol Evol. 8:21–26. Jennions M, Petrie M. 2000. Why do females mate multiply? A review of the genetic benefits. Biol Rev. 75:21–64. Johnson JB, Omland KS. 2004. Model selection in ecology and evolution. Trends Ecol Evol. 19:101–108. Keller L, Reeve HK. 1995. Why do females mate with multiple males? The sexually selected sperm hypothesis. Adv Study Behav. 24: 291–315. Kokko H, Brooks R, Jennions MD, Morley J. 2003. The evolution of mate choice and mating biases. Proc R Soc Lond B Biol Sci. 270:653–664. Lifjeld JT, Slagsvold T. 1986. The function of courtship feeding during incubation in the pied flycatcher Ficedula hypoleuca. Anim Behav. 34:1441–1453. Ligon JD. 1999. The evolution of avian breeding systems. New York: Oxford University Press. Mills JA. 1994. Extra-pair copulations in the red-billed gull—females with high-quality, attentive males resist. Behaviour. 128:41–64. Møller AP, Birkhead TR. 1992. Frequent copulation and mate guarding as alternative paternity guards in birds: a comparative study. Behaviour. 118:170–186. Mougeot F. 2000. Territorial intrusions and copulation patterns in red kites, Milvus milvus, in relation to breeding density. Anim Behav. 59:633–642. Mougeot F, Thibault J-C, Bretagnolle V. 2002. Effects of territorial intrusions, courtship feedings and mate fidelity on the copulation behaviour of the osprey. Anim Behav. 64:759–769. Neuman J, Chardine JW, Porter JM. 1998. Courtship feeding and reproductive success in black-legged kittiwakes. Colon Waterbirds. 21:73–80. Niebuhr V. 1981. An investigation of courtship feeding in herring gull Larus argentatus. Ibis. 123:218–222. Petrie M. 1992. Copulation frequency in birds: why do females copulate more than once with the same male? Anim Behav. 44:790–792.

353