Climate predicts which sex acts as helpers among cooperatively breeding bird species

Abstract

Among avian cooperative breeders, help in raising offspring is usually provided by males or by both sexes. Sex bias in helping should evolve in response to sex-specific ecological constraints on independent reproduction, with mate shortage for males and breeding vacancy shortage for each sex. Given that male-biased adult sex ratios are prevalent among birds, we predict that male-only helping mainly occurs in temperate species where fast population turnovers deriving from low adult annual survival allow all adult females to hold breeding vacancies, whereas some males overflow as helpers, and both-sex helping in tropical species where saturated habitats prevent not only males, but also females from breeding themselves. As expected, we found that across species, adult survival increased towards tropical zones and warmer climates, and higher adult survival tended to be associated with both-sex helping. Furthermore, sex bias in helping was predicted by latitude and ambient temperature. Our findings of demographic response of species to climate as a potential determinant of bias in helper sex uncover how ecological constraints operate to limit independent reproduction in sex-specific ways.

1. Introduction

Among the nearly 10 000 bird species in the world, 9% are cooperative breeders in which individuals help to rear the young of others [1]. An interesting phenomenon in association with this social system is sex-specific helping, by males only in some species or by both sexes in others, with female-only helping being very rare [2]. Interspecific variation in sex bias of helpers also exists among eusocial species, with the cause being attributed to ecological factors as well as haplodiploidy [3]. However, this question has never formally been addressed in birds [4], which have cooperative breeding systems different from those found in eusocial taxa.

There is general consensus that helping is a ‘best-of-a-bad-job’ strategy owing to environmentally imposed constraints [5], under the assumption that inclusive fitness benefits that a helper gains from helping kin, or direct benefits from helping non-kin, are smaller than those derived from its own reproduction [6]. Therefore, a key step towards understanding interspecific variation in sex-biased helping is to identify which and how ecological constraints act in a sex-specific way. Known constraints may typically be classified into two distinct categories: availability of mates and breeding vacancies (a single measure of territory and nest site, akin to breeding opportunity turnover [7]). Obviously, the former must act on males only, and the latter on both sexes, because a breeding vacancy is necessary for successful reproduction in each sex. The fact that male-biased adult sex ratios are most widespread in birds, including cooperative breeders [8,9], inspired us to suggest the following logic for how either constraint may operate. If species have lower adult survival rates in temperate than tropical zones [10], population turnover will increase at higher latitudes [11], given that no latitudinal gradient in annual production of offspring is evident [12]. As a consequence, breeding vacancies will be readily available for temperate species but limited for tropical species. Following the ‘pigeonhole principle’ in combinatorial mathematics [13], we predict that female mate shortage and male-only helping should occur in species inhabiting temperate regions where readily available breeding vacancies may allow all adult females in a population to breed independently, with male-biased adult sex ratios leaving some males unmated as helpers. By contrast, both-sex helping should occur in species inhabiting the tropics, where slow breeding vacancy turnover would limit reproductive opportunities for both adult males and females.

Using a dataset consisting of phylogenetically diverse cooperatively breeding birds, we tested the predictions that (i) there is a latitudinal decrease in adult annual survival rate across species, (ii) species with male-only helping are more likely to be associated with a lower adult annual survival rate and species with both-sex helping with a higher adult annual survival rate, and (iii) the probability of male-only helping increases when species live further away from the equator. As climates vary systemically across latitudinal gradients, we also investigated interspecific variation in adult survivorship and helping sex in relation to several climate variables to explore the potential climatic mechanism underlying the patterns described here.

2. Material and methods

(a). Data collection

A literature search was conducted to determine adult annual survival rate and sex bias in helping for cooperatively breeding bird species (electronic supplementary material, table S1). Since sex-specific data on survival rate were unavailable for many species, we used the average of both sexes for some species where the data were provided. Sex bias in helping was classified as two levels, male-only or both-sex helping, given that few species have female-only helpers. We considered a species to be male-only helping if more than 95% of helpers in a population are males, and both-sex helping if helpers of either sex account for more than 5% of all helpers. The tolerance of 5% was set considering the possibility that a few adult individuals miss breeding opportunities for stochastic reasons rather than mate or breeding vacancy limitation. We included species with ‘helpers-at-the-nest’ (a breeding pair plus delayed dispersing offspring), ‘cooperative polygynandry’ (alloparental caretakers sharing brood paternity) and ‘joint nesting’ (more than one breeding pair uses a single nest, with helpers being alloparents). Species in which all helpers are juveniles (dusky moorhens Gallinula tenebrosa) or failed breeders (long-tailed tits Aegithalos caudatus, silver-throated tits A. glaucogularis) were excluded, because such helpers do not suffer mate or breeding vacancy shortages. Of the 142 species constituting our sample of species, four (2.8%) had female helpers only, showing the rarity of this sex-biased pattern. Our final dataset contained 138 species, including 88 species with information on adult survival rates. These 138 species distributed between 0° and 58.8° in absolute latitude represent 17 orders and 58 families, with 62 having male helpers only and 76 having helpers of both sexes (electronic supplementary material, table S1).

We extracted six climate parameters at the sites where social organization of the species in the dataset were studied. These included annual mean temperature, intrayear temperature variation, interyear temperature variation, annual total precipitation, intrayear precipitation variation and interyear precipitation variation from the environmental geographic information system (GIS) layers (see electronic supplementary material, methods).

(b). Statistical analyses

All analyses were conducted with Bayesian phylogenetic-mixed models with Markov chain Monte Carlo, using the ‘mulTree’ package [14] (see electronic supplementary material, methods) in R [15]. Ten randomly selected phylogenetic trees from [16] were used to control for phylogenetic uncertainty. The raw data were z-transformed to allow a comparison of effect sizes among different predictors. When investigating whether latitude or climate affects adult annual survival rate, we included body mass as a covariate, because larger species usually survive better [17]. In the analyses with more than one predictor variable, we used a stepwise forward regression method to find out the significant predictors (for details see electronic supplementary material, methods).

3. Results

Bayesian phylogenetic-mixed models revealed a significantly negative correlation between adult annual survival rate and latitude among 88 cooperatively breeding bird species, after correcting for body mass (table 1 and figure 1). The analyses further showed that annual mean temperature and annual total precipitation significantly predicted interspecific variation in adult survivorship, which increased as ambient temperatures and precipitation increased. Standardized regression coefficients suggested that the effect of latitude on survival rate was similar to that of each of the two climate variables. Four other climate variables were non-significant predictors of adult annual survival rate.

Table 1.

A summary of Bayesian phylogenetic-mixed models shows the significant predictors of adult annual survival rate and sex bias in helping (male-only or both-sex) across cooperatively breeding bird species. Predictor variables were selected following a stepwise forward regression method.

models		standardized effect size	95% CI
dependent variable	significant predictor variables
adult annual survival rate	latitude	−0.21	−0.41, −0.01
	body mass	0.34	0.09, 0.58
helping sex	adult annual survival rate	1.52	0.04, 3.10
helping sex	latitude	−1.24	−2.38, −0.15
adult annual survival rate	annual mean temperature	0.29	0.09, 0.48
	body mass	0.34	0.10, 0.58
adult annual survival rate	annual total precipitation	0.26	0.07, 0.46
	body mass	0.30	0.06, 0.54
helping sex	annual mean temperature	1.46	0.35, 2.64
helping sex	intrayear temperature variation	−1.28	−2.45, −0.17

Figure 1.

(a) Latitude and ambient temperature as the significant predictors of adult annual survival rate, (b) adult annual survival rate as the significant predictor of sex bias in helping (male-only, 0; both-sex helpers, 1) and (c) latitude and ambient temperature as the significant predictors of sex bias in helping across cooperatively breeding bird species. The graphs are generated by plotting z-transformed raw data along with the regression lines predicted by the Bayesian phylogenetic-mixed models.

No significant correlation between sex bias in helping and adult annual survival rate was found when using data from the 88 species (standardized effect size: 0.96, 95% CI: −0.47 to 2.48). However, the correlation became significant when excluding four male-only helping species whose annual survival rates were extraordinarily high (greater than or equal to 0.95), with male-only helping more likely being associated with lower and both-sex helping with higher adult survivorship (table 1 and figure 1).

Latitude, annual mean temperature and intrayear temperature variation remained significant predictors of the sex composition of helpers across 138 cooperatively breeding birds, with a similar effect size (table 1). Species with male-only helping were more likely to inhabit higher latitudes and colder, seasonal environments, whereas species with both-sex helping occurred more often at lower latitudes with warmer and less-seasonal environments (figure 1). Four other climate variables had no significant effect on sex bias in helping. When we restricted the data to a subset of the 88 species having information on adult survival rates, our conclusions remained unchanged (electronic supplementary material, table S2).

4. Discussion

We showed that adult annual survival rate of cooperatively breeding bird species can be predicted by latitude, ambient temperature and precipitation. This is consistent with previous findings showing that low temperature is the major factor determining individual mortality in birds, most likely owing to increased thermoregulatory challenges and decreased food availability in cold environments [18,19]. Cold climates at high latitudes, along with low food availability owing to poor rainfall, may further pose a great survival challenge [18,19]. Importantly, helper sex was significantly correlated with adult survivorship, when excluding the four species with survival rates greater than or equal to 0.95. This exclusion is warranted, because the values are more likely to be an overestimate for any bird species especially high-latitude ones in which a reduction in survivorship has been detected using the current dataset even including the four species. Moreover, we provide evidence for the expected patterns of sex bias in helping in relation to latitude and climate. These results together suggest that the mechanism underlying interspecific variation in which sex in helping can be attributed to demographic responses of species to climate.

The idea proposed by this study, the prevalence of male-biased adult sex ratios along with the availability of breeding vacancies for all females at high latitudes, may also explain the rarity of female-only helping among cooperatively breeding birds. A non-exclusive explanation is the high probability of an excess of females leading to polygamy rather than cooperative breeding.

A major implication of this study is to bring insights into the ecological constraints hypothesis, which, as a widely accepted explanation for the evolution of cooperative breeding in birds, does not consider where and how ecological constraints act in a sex-specific way [5]. Here, the key step towards addressing this question is to recognize mate and breeding vacancy as two distinct types of ecological constraints, treating the latter as the sum of a range of factors critical for successful reproduction in different species.

Alternatively, male-only philopatry and helping might be attributed to female-biased natal dispersal. If so, we should expect that natal dispersal in species with both-sex helpers would be unbiased with respect to sex. However, that was not the case as these species also perform female-biased dispersal, as done by most birds [20]. One might argue that a male-biased adult sex ratio, the precondition of the current hypothesis, is a consequence, rather than a cause, of delayed natal dispersal by males. Indeed, natal dispersal may cause high mortality in female juveniles. However, our hypothesis emphasizes adult sex ratios, which more likely derive from high mortality of adult females owing to greater reproductive costs [8,9].

We acknowledge that more evidence is needed for testing our hypothesis that demographic response of species to climate determines which sex helps, including the information on adult sex ratios of cooperatively breeding species for comparative approaches. Moreover, manipulating adult sex ratios and breeding vacancies in a population to monitor changes in sex composition of helpers may also provide direct tests of this hypothesis.

Ethics

This study does not require ethical permit, because its data are based on published sources.

Data accessibility

Data are available as supplementary material.

Authors' contributions

X.L. designed the research; G.Y., Q.Z. and X.L. collected the data; G.Z. performed statistical analyses; X.L. wrote the original manuscript; all authors contributed to the writing of the original manuscript and revisions. All authors gave final approval for publication and agreed to be held accountable for the content of this article.

Competing interests

The authors have no competing interests.

Funding

X.L. was supported by a grant (no. 30830019) from the National Natural Science Foundation of China.