Sex-biased preferential care in the cooperatively breeding Arabian babbler

Authors


A. R. Ridley, DST/NRF Centre of Excellence, Percy Fitzpatrick Institute of African Ornithology, University of Cape Town, Rondebosch 7701, South Africa.
Tel.: +27 216503291 ext. 3634; fax: +27 216503295;
e-mail: aridley@botzoo.uct.ac.za

Abstract

In many social birds there are sex differences in dispersal patterns, with males commonly remaining in their natal group whereas females typically disperse at adolescence. Group members may therefore increase their fitness by preferentially caring for offspring of a particular sex according to social circumstances. Although previous studies have focussed on intragroup social factors that may affect preferential care, we propose that the relative size of neighbouring groups is of primary importance. Here we show that in the cooperatively breeding Arabian babbler (Turdoides squamiceps), parents preferentially feed male offspring when relative group size is small, and female offspring when group size is large. Unlike parents, helpers consistently favour young of the opposite sex to themselves, suggesting the risk of competition with members of the same sex for future breeding opportunities may override other considerations. These results emphasize the complexity of investment strategies in relation to social circumstances and the variable benefits of raising males vs. females in a species with sex-biased philopatry.

Introduction

When young are cared for by parents, the benefits of investment often differ between individual young, generating variation in the optimal level of parental investment. This can cause parents to care for some categories of offspring more than others. For example, in passerines with bi-parental care, parents may preferentially care for chicks of a particular size, sex or phenotype (Harper, 1985; Stamps et al., 1987; Lessells, 1998) as a strategy to maximize the benefits received from investment in the brood (Hasselquist & Kempenaers, 2002; Lessells, 2002a). In many cooperative breeders, where fitness increases with group size, theory suggests that parents should preferentially care for the philopatric sex (Griffin & West, 2003; Griffin et al., 2005), and preferential care should be most pronounced in small groups that stand to gain the greatest benefit from an increase in group size (Courchamp et al., 1999; Kokko et al., 2001). Currently, evidence to support this comes most convincingly from the cooperatively breeding meerkat (Suricata suricatta, Desmarest). In this species, females breed in their natal group more often than males, are more likely to feed young than males, and are more likely to feed female pups (Clutton-Brock et al., 2002). It is possible that similar patterns of care occur among avian cooperative species, but this remains to be determined.

The concept of sex differences in offspring ‘value’ was first applied to the manipulation of offspring sex ratios by Hamilton (1967) who argued that parents should invest less in the sex that is most likely to compete with current group members or each other for resources. This theory, extended by Clark (1978), generated the local resource competition (LRC) and local resource enhancement (LRE) hypotheses, and has helped interpret patterns of sex bias in some cooperatively breeding species. For example, in the Seychelles warbler (Acrocephalus sechellensis, Oustalet), parents on low quality territories bias the prelaying sex ratio towards the dispersing sex to avoid competition for limited local resources (Komdeur et al., 1997). However, there is increasing evidence that many species are unable to bias the sex ratio of young prelaying (Komdeur & Pen, 2002). In such cases, investing preferentially in the most valuable young post-hatching may be an effective strategy to ensure the survival of young that will contribute most to parental fitness.

In cooperatively breeding species, parents and helpers stand to gain different fitness benefits from raising young (Emlen, 1997) and this may affect their investment strategies. For example, in contrast to parents, helpers may prefer to raise young that will aid them in the future as cohort partners in dispersal attempts (Ligon & Ligon, 1983), or as breeding partners (Craig & Jamieson, 1988). Despite these different benefits, little attention has been paid to the different investment strategies that may arise as a consequence. In this paper, we investigate whether adults bias care toward particular young, and whether patterns of care vary between parents and helpers in the cooperatively breeding Arabian babbler (Turdoides squamiceps, Cretzschmar), using field observations. Arabian babblers are sexually monomorphic and obligately cooperative. Reproduction is commonly dominated by a single breeding pair, with remaining group members helping to raise the brood (Zahavi, 1990). Helpers of both sexes assist in rearing young, but females commonly disperse at 2–3 years of age whereas males are philopatric (Zahavi, 1990). This analysis focuses on two issues. First, we look at what factors affect adult preferences for feeding particular young and secondly, we investigate the benefits of preferential care on fledgling weight gain and survival.

Methods

Data collection

Our study was based on habituated groups of Arabian babblers located in the Shezaf Nature Reserve, Negev desert, southern Israel (30°48′N, 35°13′E). All birds in the study population were individually recognizable by a unique combination of colour rings. Each nest was checked daily to confirm the exact day of both hatching and fledging. Within broods, all nestlings fledged on the same day. Preferential care was investigated during the post-fledging stage when young are completely dependent on adult group members for food (from day of fledging until eight weeks post-fledging), which is the period of highest mortality in this species (Ridley, 2003). Each group was observed three times per week for 3–4 h per observation session (n = 469 observation hours, 164 observation sessions). As it is not possible to determine sex from external characteristics in this species, small blood samples (50 μL) were collected from nestlings via brachial venipuncture. Nuclear DNA was extracted using a standard phenol/chloroform/isoamyl alcohol protocol and PCR-based molecular sex determinations were conducted using the method described by Fridolfsson & Ellegren (1999).

The size and number of all food items delivered by all adults to all dependent fledglings was recorded using a handheld data logger. All food items were divided into four size classes (see Radford & Ridley, 2006 for size classification). To determine the average biomass of each size class, 50 prey items representative of each size class were weighed. The total biomass each adult fed to each fledgling was calculated as the sum of the number of prey items in each size class multiplied by the average weight of that size class. Provisioning rates per adult were calculated by dividing the total biomass fed by the number of hours in each observation session.

Body mass was measured by weighing each group member at the start of each observation session. All observation sessions began within 15 min of first light, before any provisioning activity began. Individuals were enticed to stand on a top-pan balance (400 ± 0.1 g) for a small food reward. For fledglings, the change in body mass over the period between fledging and independence was taken as the difference in body mass at fledging and body mass in the first week following nutritional independence for each individual. Because of the drought conditions during the study, most fledglings died before reaching adulthood, making it impossible to analyse differences in survival to maturity. Instead, survival was measured as the number of days post-fledging that each fledgling survived.

Analysis

Analyses were conducted on 14 mixed-sex broods (average brood size 2.6 ± 0.4 fledglings, range 2–4) from 10 groups. Data were analysed in Genstat 8.1 (8th Edition 2005, Lawes Agricultural Trust, Rothamsted, UK). Linear Mixed Models (LMMs) were used so that random terms (which allow the analysis to take account of repeated measures on the distribution of the data) could be defined in addition to fixed terms. The method of fitting LMMs to data followed Crawley (2002). Model simplification using backward-elimination was adopted. Terms were systematically removed from the model and only put back if their removal resulted in a significant loss of variance explained. All two-way interactions were tested, but only those that were significant are presented. To facilitate interpretation in the case of three-way interactions, the data were split into separate categories for subsequent LMM analysis. Where two terms were significantly associated, both terms were added separately to the model. Of the two associated terms, the one which contributed most to model explanatory power was retained.

To identify the factors affecting adult provisioning patterns, the average biomass fed to each fledgling by each adult per hour for each observation session was entered as the response variable in a LMM with an identity link function. The response variable was log-transformed to achieve the correct ‘scale’ for the data. Only data collected when two or more brood members were alive were included in the analysis. The potential explanatory terms fitted were: (i) fledgling characteristics [sex, brood size, age (number of days post-fledging) and body mass], (ii) adult characteristics (sex, breeding status, body mass) and (iii) social factors (relative group size (focal group size/average size of neighbouring groups), actual group size (number of adults) and adult sex ratio). Breeding status was categorized as: ‘breeders’ (the dominant male and female in each group) and ‘helpers’ (all other adult group members), as reproduction by subordinates is rare in this species (Lundy et al., 1998). Group, adult, brood and fledging identity were included as random terms in the model. An additional LMM was conducted to determine the effect of preferential care on the change in body mass between fledging and independence. The response term was the difference in body mass between fledging and independence for each fledgling. The potential explanatory terms fitted were: fledgling sex, brood size, and the average biomass received per hour from all adult group members combined. Group and brood identity were included as random terms in the model.

Survival data were analysed using a Cox's proportional hazards regression. Observations ended 200 days after fledging and individuals still alive after this point were considered as censored data. The response variable was the number of days post-fledging that an individual survived, and the potential explanatory terms tested were fledgling sex, body mass at fledging, brood size at time of fledging and preferential care. The term preferential care was divided into two categories: ‘preferentially fed’ (the fledgling from each brood that received the most food per hour from adults) and ‘nonpreferred’ (incorporating all other fledglings). Analysis was conducted using a backward elimination process until only those terms that explained a significant amount of variation in survival were retained.

Results

We found significant variation in the amount of food that adults fed to male vs. female young, and the feeding preferences of breeders differed from those of helpers (Table 1, Fig. 1). In groups that were smaller than their neighbours, breeders fed male young more than females (back-transformed log-values for male fledglings: 0.29 ± 0.02 g h−1, females: 0.15 ± 0.03 g h−1), whereas in groups that were larger than their neighbours, breeders fed female young more than males (male fledglings: 0.07 ± 0.01 g h−1, females: 0.19 ± 0.02 g h−1Table 1). Relative group size, rather than group size per se, was the primary influence on the provisioning patterns of breeders (LMM: fledgling sex × relative group size χ2 = 11.66, P ≤ 0.001, fledgling sex × group size χ2 = 4.16, P = 0.041).

Table 1.   The minimal models generated from Linear Mixed Model analyses of the terms associated with the average biomass fed to each fledgling per hour by each adult during the post-fledging dependent stage.
Model termWald statistic (χ2)d.f.PAverage effectSE
  1. Analysis was conducted on 164 observation sessions (average feeding rate/hr/adult) of 58 adults (22 breeders and 36 helpers) provisioning 38 fledglings from 14 broods in 10 different groups. The response term was the log-value of the biomass fed to each fledgling per hour by each adult. Group, adult, brood and fledgling identity were included as random terms in each model.

All adults
 Constant   0.420.05
 Fledgling sex × relative   group size × breeding   status14.331< 0.001− 0.620.16
 Brood size13.941< 0.001−0.170.05
 Relative group size0.7710.38− 0.370.14
 Breeding status0.3710.54− 0.020.03
 Fledgling sex0.0510.820.020.06
Breeders
 Constant   0.460.06
 Relative group   size × fledgling sex11.661< 0.0010.870.25
 Relative group size0.6710.41− 0.460.26
 Fledgling sex0.0910.76− 0.040.08
Helpers
 Constant   0.320.04
 Fledgling sex12.941< 0.0010.250.06
 Brood size10.4310.001− 0.170.04
 Fledgling sex × adult sex4.4110.04− 0.180.09
 Adult sex0.1710.680.060.08
Figure 1.

 The average biomass that male vs. female fledglings received per hour from breeding adults in relation to relative group size. Raw data values are displayed. Each point represents the average biomass fed to each fledgling by each breeding adult per hour during the post-fledging dependent period. Fitted lines are generated from the predicted values of the Linear Mixed Model presented in Table 1.

In contrast to breeders, there was no effect of relative group size on the provisioning patterns of helpers (Table 1). Among helpers, there was a difference in provisioning patterns according to helper sex. This trend was driven primarily by male helpers, who fed less food to male young than female young (back-transformed log-values for male fledglings: 0.24 ± 0.04 g h−1, females: 0.49 ± 0.06 g h−1), with no difference in the amount of food that female helpers fed to male vs. female young (male fledglings: 0.36 ± 0.07 g h−1, females: 0.41 ± 0.09 g h−1).

Young received considerable benefits from preferential care, with a strong effect of provisioning rates on both fledgling growth and survival (Table 2). All fledglings gained body mass over the dependent period (Fig. 2), but preferentially fed fledglings (i.e. those individuals in each brood that received the most food) gained more body mass than their brood mates during the two-month period between fledging and independence (Body mass change for preferentially fed young: 24.6 ± 0.4 g, all other fledglings: 20.1 ± 0.8 g). Fledglings that were preferentially fed also survived longer (Cox Proportional Hazards Regression: χ2 = 4.211,34P = 0.04, Fig. 3). From 14 broods, 10 preferentially fed fledglings survived to the end of the observation period (200 days) compared with only seven other brood members (30.4% of fledglings that did not receive preferential care).

Table 2.   Linear Mixed Model analysis of the terms associated with fledgling change in body mass in the period between fledging and independence.
Model termWald statistic (χ2)d.f.PAverage effectSE
  1. Analysis was conducted on body mass changes for 38 fledglings from 14 broods in 10 different groups. Group and brood identity were included as random terms in the model.

Minimal Model    
 Constant   22.040.76
 Biomass   received    hour−14.9710.032.941.28
Additional   terms tested    
 Brood size0.4410.51  
 Sex0.0910.77  
Figure 2.

 Fledgling change in body mass in the period between fledging and independence in relation to total biomass received from all adult group members combined. Raw data values are displayed. Each point represents absolute body mass change for each fledgling over the post-fledging dependent period. The fitted line is generated from the predicted values of the Linear Mixed Model presented in Table 2.

Figure 3.

 Survival probability in relation to preferential care. ‘Preferentially fed’ refers to the fledgling from each brood that received the most food from adult group members, ‘nonpreferred’ refers to all other fledglings.

Discussion

In the Arabian babbler, breeders vary the level of care given to male vs. female young according to relative group size in a pattern that provides support for the LRC and LRE hypotheses (Clark, 1978). Breeders in small groups biased care towards male young, who are most likely to contribute to the defence of local resources (LRE), whereas breeders in large groups biased care toward females, who are least likely to compete with existing group members for resources (LRC). In groups that are smaller than their neighbours, the production of males could provide considerable benefits, as groups containing many males are more effective at territory defence and less vulnerable to invasion by conspecifics (Sherman, 1995; Ridley, 2003). Conversely, in large groups the production of females could provide considerable benefits, as females are more likely to successfully disperse and establish a breeding position elsewhere (Yaber & Rabenold, 2002; Ridley, 2003), whereas males join the end of the social queue and may eventually compete for a breeding position, as well as potentially depleting territory resources (Pen & Weissing, 2000; Kokko & Ekman, 2002). The provisioning behaviour displayed by breeders provides considerable support for the idea that parental investment per offspring is free to vary (Lessells, 2002b, compared to the assumption of fixed parental investment sensu Fisher, 1930), allowing parents to maximize the benefits obtained from investment in young. However, we did not measure the begging behaviour of fledglings, nor the effect of begging on adult behaviour. Thus, we are unable to rule out the possibility that provisioning behaviour could have been affected by offspring that differ in their competitive begging abilities (e.g. Godfray, 1995).

If the production of male vs. female young brings variable benefits to breeders according to relative group size, then the most efficient strategy to gain fitness benefits from investment would be to produce only the sex that brings parents the greatest benefits under current conditions. Although the ability to manipulate offspring sex prelaying has been demonstrated for several species (e.g. Seychelles warbler, Komdeur et al., 1997), conclusive evidence is lacking for many other avian species (Lessells, 1998; Komdeur & Pen, 2002). Arabian babbler females may be unable to control the sex of offspring prelaying, and in such cases, preferential investment in particular young post-hatching appears an efficient strategy of manipulating the survival of young that provide breeders with the greatest fitness benefits. This is confirmed by the significant effect of preferential care on the survival and development of young in this species.

Unlike breeders, helpers provide less care to fledglings of the same sex as themselves, and this trend is driven primarily by males. It appears unlikely that this pattern of care occurs as a result of males preferring to invest in offspring that may become future breeding partners (e.g. Craig & Jamieson, 1988), as there is evidence for inbreeding avoidance in this species (Zahavi, 1990; Lundy et al., 1998). It also seems unlikely that males favour young that are potential future dispersal partners because (i) successful lone dispersal is relatively common and (ii) when social dispersal does occur, it usually involves members of the same sex (Ridley, 2003). Rather, it appears that by providing less care to male fledglings, male helpers reduce the number of competitors for future breeding opportunities. This pattern of care may be prevalent in male helpers only because they are philopatric and therefore more likely to be affected by intragroup same-sex competitors than female helpers.

The preferential care displayed by male helpers in the Arabian babbler contrasts with patterns of care displayed by helpers in other cooperative species. In the meerkat, where helpers increase reproductive success (Clutton-Brock et al., 2001), preferentially caring for females (the philopatric sex) appears the best strategy to enhance the fitness benefits gained from group-living (Clutton-Brock et al., 2002). In contrast, the negligible effect of group size on reproductive success in the Arabian babbler (Zahavi, 1990; Ridley, 2003) suggests that male helpers stand to gain more fitness benefits by reducing the number of future breeding competitors than the potential benefits gained from group augmentation.

Acknowledgments

We are grateful to Professors Amotz and Avishag Zahavi for allowing us to work at their field site and for useful discussions and advice and to Professor T. H. Clutton-Brock for supervising this study. We thank Matthew Bell, Kat Munro, Sarah Ross-Viles, Anat Shapira and Kate Smith for valuable help in the field. For advice, assistance and comments we are grateful to A. Cockburn, S.J. Hodge, J. Juritz, C.M. Lessells, A.N. Radford, A.F. Russell and A.J. Young and two anonymous reviewers. We thank Professor P. Parker and K. Halbert for help with the PCR work and use of lab resources. This research was funded by grants from the Cambridge Commonwealth Trust, New Zealand Federation of University Women and the Wingate Scholars Program to A.R. Ridley and fellowship support from the University of Missouri St-Louis to K.P. Huyvaert.

Ancillary