SEARCH

SEARCH BY CITATION

Keywords:

  • cooperation;
  • Hamilton’s rule;
  • helping;
  • indirect fitness;
  • kin selection;
  • meta-analysis;
  • relatedness

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

Hamilton demonstrated that the evolution of cooperative behaviour is favoured by high relatedness, which can arise through kin discrimination or limited dispersal (population viscosity). These two processes are likely to operate with limited overlap: kin discrimination is beneficial when variation in relatedness is higher, whereas limited dispersal results in less variable and higher average relatedness, reducing selection for kin discrimination. However, most empirical work on eukaryotes has focused on kin discrimination. To address this bias, we analysed how kin discrimination and limited dispersal interact to shape helping behaviour across cooperatively breeding vertebrates. We show that kin discrimination is greater in species where the: (i) average relatedness in groups is lower and more variable; (ii) effect of helpers on breeders reproductive success is greater; and (iii) probability of helping was measured, rather than the amount of help provided. There was also an interaction between these effects with the correlation between the benefits of helping and kin discrimination being stronger in species with higher variance in relatedness. Overall, our results suggest that kin discrimination provides a route to indirect benefits when relatedness is too variable within groups to favour indiscriminate cooperation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

A major problem for evolutionary biology is explaining how selection favours cooperative behaviours that benefit other individuals (reviewed by Sachs et al., 2004; Lehmann & Keller, 2006; West et al., 2007a,b). Hamilton’s (1964a,b, 1970) theory of inclusive fitness provides a potential solution to this problem by showing that individuals can increase their indirect fitness by helping relatives. In his original papers, Hamilton (1964a,b, 1971, 1972, 1975) pointed out that the degree of relatedness required to generate indirect benefits could arise via two routes: individuals preferentially interacting with closer relatives (kin discrimination), or through limited dispersal (population viscosity) which increases the probability that individuals will interact with relatives. However, there has been little overlap between the theoretical and empirical research on these processes, with the theoretical literature focusing on limited dispersal and the empirical literature focusing on kin discrimination (West et al., 2002; see discussion for microbial exceptions).

In cooperatively breeding vertebrates, a dominant pair usually produces the majority of the offspring, while the cost of caring for offspring is shared with nonbreeding subordinate helpers (Jennions & Macdonald, 1994; Cockburn, 1998; Hatchwell & Komdeur, 2000; Clutton-Brock, 2002; Griffin & West, 2003; Koenig & Dickinson, 2004). Empirical research on the importance of indirect fitness benefits in explaining such helping behaviour has focused on kin discrimination. However, if relatedness between interacting individuals within groups is high, then it is still possible that indirect fitness benefits will be important, even with indiscriminate helping. Griffin & West (2003) and Boomsma (2007) have argued that one way to test this hypothesis is to examine whether kin discrimination is weaker in species where within group relatedness is higher and/or shows less variation.

We test this prediction with a meta-analysis across cooperative breeding birds and mammals. Griffin & West (2003) have previously shown that the extent of kin discrimination is positively correlated with the benefits of helping behaviour. This is predicted by inclusive fitness theory: when helping provides greater benefits, indirect fitness from preferentially helping closer relatives will be greater. Consequently, our major aim here is to test the prediction that the level of kin discrimination across species should be correlated with within group relatedness as well as the benefits of helping. Furthermore, we test the prediction that kin discrimination will be stronger in studies measuring the probability of, rather than, the amount of help provided. This is because the amount of help may depend on other factors, such as the helper’s physical condition, which can influence the cost of helping (Emlen & Wrege, 1988; Griffin & West, 2003). We test this prediction by examining differences in the strength of kin discrimination between species where the probability or amount of help provided has been measured. A comparative test of the relative importance of these different possible explanatory variables has only just become possible, thanks to: (i) developments in meta-analysis methodology that allow multivariate and formal phylogentic analyses (J.D. Hadfield & S. Nakagawa, unpublished; Nakagawa et al., 2007; Adams, 2008; unpublished et al., 2009; Knowles et al., 2009; Lajeunese, 2009) and (ii) recently published data that expands the number of species studied sufficiently to allow meaningful multivariate analyses.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

Data collection

We have previously presented data collected on the strength of kin discrimination, rKin (the effect of relatedness on the probability of becoming a helper either in natal group or as immigrant, or amount of help provided by helpers), and the effect of helpers, rHelp (the effect of help on fledgling success or, where possible, survival of offspring to the following year) in Griffin & West (2003) and Griffin et al. (2005). All studies used in these previous analyses were included in the present analysis. The dataset was updated to include all studies published since 2005 presenting data that could be used to obtain values for rKin and rHelp. We located relevant papers through keyword searches on Web of Science and forward and backward citation searches on key papers. We also contacted researchers to obtain unpublished data. Full, updated datasets used in analyses are given in Appendix Tables A1–A3. We have been able to add five species to the dataset used to calculate the correlation between rKin and rHelp presented in Griffin & West (2003): Aegithalos caudatus (long-tailed tit), Picoides borealis (red-cockaded woodpecker), Corvus c. corone (carrion crow), Nesomimus parvula (Galapagos mockingbird) and the Manorina melanophrys (Australian bell miner). The studies used to obtain values of rhelp vary in whether they controlled for potential confounding factors such as territory and breeder quality. However, we do not expect this to drive relationships between variables, but instead create variation in the data.

In addition to the relationship between rKin and rHelp we measured the relationship between the strength of kin discrimination (rKin) and the mean and variance in relatedness between helpers and offspring that could potentially be helped, and whether helpers in a species were typically natal to the group in which they helped or immigrants (Appendix, Tables A1–A3). In species where helpers are typically retained natals, it is predicted that helping will be indiscriminate because average relatedness is high and that discrimination will be stronger in species with nonnatal helpers because of increased variation in relatedness. Species were categorized as either ‘mainly natal’ or ‘nonnatal’ from descriptions of their breeding system from the literature (T.H. Clutton-Brock & Sharp, personal communication). For example, Suricata suricatta (meerkats) were categorized as ‘mainly natal’ as helpers are mainly offspring from previous litters that have not dispersed from the natal territory (even though there are also immigrant helpers present) (Clutton-Brock et al., 2001). Ceryle rudis (Pied kingfisher) was categorized as ‘nonnatal’ as helpers are not the offspring of the breeding pair, and may breed with the breeding female (Reyer, 1984).

When measuring the mean and variance in relatedness we aimed to capture relatedness between offspring and potential helpers, but it is often not specified whether studies included nonhelpers in their analyses. This potential bias is likely to result in variance in relatedness being underestimated and mean relatedness over-estimated. However, any bias is expected to be equally likely in species with limited dispersal as those with high dispersal, and reduce the ability to detect an effect of mean and variance in relatedness on kin discrimination rather than drive relationships. Furthermore, most of the studies we used to extract this data were examining the effect of kinship on helping (the strength of kin discrimination) and so would be expected to include individuals that did not help as well as those that contributed to help. The methods used to measure relatedness also varied across studies (genealogical vs. molecular genetic, see appendix), but we found no evidence that this significantly affected relatedness estimates (mean relatedness: GLM with binomial error distribution, F1, 13 = 0.18, P = 0.68. Variance in relatedness: GLM with normal error distribution, relatedness method F1, 13 = 0.97, P = 0.34) or kin discrimination when entered into the model outlined in Table 1 (F1, 5 = 0.25, P = 0.64).

Table 1.   Linear mixed model of predictors of kin discrimination (Zr-kin) across cooperatively breeding vertebrates.
Fixed termsParameter estimate (β)SE95% LCL95% UCLDFFP
  1. Effect sizes of rkin and rhelp were Z transformed prior to analysis and the parameters estimates are presented on the Z scale. The response variable Zr-kin was weighted by the inverse variance. Significant values are shown in boldface type. LRT, log-likelihood ratio test, LCL, lower confidence limit, UCL, upper confidence limit.

  2. Denotes terms included in final model. Nspecies = 14, Ngenera = 14, Nfamilies = 14, Norders = 4, Nclasses = 2.

Zr-help†     1, 103.660.09
Mean relatedness† −2.620.94−4.68−0.551, 107.800.02
Variance in relatedness†     1, 91.270.30
Probability vs. amount of help†Amount0.100.09−0.110.311, 1010.240.01
 Probability0.450.090.260.64
Natal helpers     1, 33.500.17
Zr-help*Mean relatedness     1, 62.240.18
Zr-help*Variance in relatedness† 56.1525.080.75119.931, 95.010.05
Zr-help*retained natals     1, 40.310.61
Zr-help*probability vs. amount of help     1, 81.430.27
Random termsVariance ComponentSE95% LCL95% UCLDFLRTP
Class†0.000.000.000.0010.001.00
Order (class)†0.000.000.000.0010.001.00
Species (order class)†0.050.030.020.30113.480.0002

Meta-analysis

We conducted a meta-analysis on studies examining kin discrimination (rKin) across vertebrate species using a multivariate linear mixed effects model with restricted maximum likelihood estimation (REML) conducted in sas version 9.2 (Littell et al., 2006). Prior to the analysis, effect sizes were Z-transformed

  • image

ZrKin was weighted by the inverse variance to account for variation in sample sizes between studies. The variance was calculated by the reciprocal of the sum of the conditional variance,

  • image

where n is the sample size of the study (Raudenbush, 1994; Nakagawa et al., 2007). We analysed variation in ZrKin in relation to the following fixed effects: (1) the benefits of helping behaviour (ZrHelp, covariate); (2) the probability of helping (nspecies = 13) vs. the amount of help provided (nspecies = 13. For three species, measures of both amount and probability were available) (two level factor), (3) mean relatedness within groups (covariate), (4) variance in relatedness within groups (covariate), and (5) whether helpers were mainly natal (nspecies = 4) to the group or nonnatal (nspecies = 17) (two level factor; Clutton-Brock & Sharp, personal communication). We checked whether the sample size of the studies used to calculate ZrHelp had an effect on our results by entering the inverse variance of ZrHelp as a covariate in our analyses and in all cases this was nonsignificant (P > 0.30).

The published studies on kin discrimination represent a diverse range of bird and mammal species and for some species there were multiple studies that examined both the probability and amount of help directed towards related and unrelated individuals. Nonindependence of data has been dealt with in the past by taking species averages and calculating independent contrasts across phylogenies (Felsenstein, 1985; Harvey & Pagel, 1991; West & Sheldon, 2002; West et al., 2005; Nakagawa et al., 2007; Adams, 2008). However, mixed model meta-analysis can deal with the nonindependence of data through random effects that account for intra-group correlations, avoiding data averaging and allowing the full dataset to be utilized (Hadfield & Nakagawa, submitted; Nakagawa et al., 2007; Adams, 2008; Chapman et al., 2009; Knowles et al., 2009; Lajeunese, 2009). We therefore took into account the nonindependence of data arising from multiple studies on the same species, and from the phylogenetic relationships between species by defining a nested random effects structure whereby species were nested within order and order was nested within class. Only order and class were entered into the model because in our dataset genera and families only contained single species and therefore there were only multiple species at the taxonomic levels of order and class. The significance of fixed effects (factors and covariates) were examined using Wald type adjusted F statistics and the effect with the highest P value was sequentially dropped until only significant terms (P < 0.05) remained (Crawley, 2002). The Kenward & Roger (1997) method for estimating standard errors for parameter estimates and denominator degrees of freedom was used as it is specifically designed for models with multiple random effects and unbalanced data, increasing the accuracy of significance tests (Kenward & Roger, 1997; Littell et al., 2006; Bolker et al., 2009). The significance of random effects was assessed using log-likelihood ratio tests (LRTs) (Self & Liang, 1987). Details of all analyses are provided as electronic supplementary material (Tables S1–S4).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

We found that variation in kin discrimination was explained by three different effects. First, kin discrimination was weaker when mean relatedness between individuals was higher (mean relatedness: F1, 10 = 7.80, P = 0.02; Table 1; Fig. 1a). This suggests that individuals are less likely to discriminate between kin and nonkin in species that live in groups with closer relatives. Second, kin discrimination was stronger when the probability rather than amount of helping was measured (amount vs. probability of help: F1, 10 = 10.24, P = 0.01; Table 1; Fig. 1b). Third, consistent with Griffin & West (2003), we found a positive relationship between kin discrimination and the benefits of helping (Table 1; Table S1; Fig. 1c). However, in the present study there was additional complexity with the relationship between kin discrimination and the benefits of helping being dependent upon variance in relatedness (ZrHelp*variance in relatedness: F1, 8 = 5.01, P = 0.05; Table 1, Fig. 2). As predicted by inclusive fitness theory, the relationship between kin discrimination and the benefits of helping was stronger when variance in relatedness between group members was higher (Table 1; Fig. 2). Finally, after taking into account mean and variance in relatedness between individuals, kin discrimination did not significantly differ between species with and without natal helpers (natal helpers: F1, 3 = 3.50, P = 0.17).

image

Figure 1.  Kin discrimination (ZrKin) across cooperatively breeding vertebrates. (a) Kin discrimination in relation to average relatedness between individuals. Solid line represents predicted relationship from the linear mixed model (Table 1) with 95% confidence intervals (dotted lines). (b) The difference in kin discrimination when the probability of helping and the amount of help provided were measured. Bars represent means ± SE. (c) The relationship between the benefits to offspring of helping (ZrHelp), and kin discrimination. Solid line represents predicted relationship from the linear mixed model (Table S1) with 95% confidence intervals (dotted lines).

Download figure to PowerPoint

image

Figure 2.  The effect of the interaction between variance in relatedness and the benefits to offspring of helping (ZrHelp) on kin discrimination (ZrKin). Surface of relationship is visualized using a loess smoothing procedure. The grey circles indicate data points below the line whereas black data points are above the line.

Download figure to PowerPoint

We tested the robustness of our results in three ways. First, data on mean and variance in relatedness was only available for 14 species (Ngenera = 14, Nfamilies = 14, Norders = 4, Nclasses = 2). We therefore re-ran our analysis removing measures of relatedness from the analysis to use all data on the other explanatory variables (18 species, Ngenera = 18, Nfamilies = 16, Norders = 4, Nclasses = 2). Variation in kin discrimination was once again explained by the benefits of helping, and whether the probability or amount of helping was measured (ZrHelp: F1, 11 = 5.69, P = 0.04. Amount vs. probability of helping: F1, 16 = 6.63, P = 0.02; Table S1). However, we found that species where helpers are mainly natals had significantly lower kin discrimination (mean ± SE ZrKin = 0.18 ± 0.07) than species with nonnatals (mean ± SE ZrKin = 0.57 ± 0.15. F1, 11 = 5.68, P = 0.04; Table S1). This is perhaps unsurprising given that the presence of natal helpers is likely to be a crude indicator of mean relatedness between individuals when direct measures of relatedness were not entered into this analysis. Second, the number of species in our dataset is small relative to the number of explanatory variables and this may give spurious results. We therefore analysed the effect of each explanatory variable on kin discrimination separately, which also allowed us to utilize all data available for each explanatory variable. Once again the significance of results did not change (ZrHelp: F1, 17 = 5.13, P = 0.04; mean relatedness: F1, 13 = 5.42, P = 0.04; amount vs. probability of helping: F1, 22 = 6.77, P = 0.02; natal helpers: F1, 20 = 9.25, P = 0.007; Table S2). Finally, we re-ran our analysis after removing Dacelo novaeguineae (kookaburra) and Phoeniculus purpureus (green woodhoopoe) in turn. We removed D. novaeguineae because relatedness was given as band-sharing co-efficients from DNA fingerprinting (Legge & Cockburn, 2000), which relates less directly to the r in Hamilton’s rule. We removed P. purpureus because it could be argued that the experimental approach used to assess helping behaviour by Du Plessis (1993) does not measure the selected response to differences in relatedness. However, removing D. novaeguineae and P. purpureus from our analyses did not change the significance of any main effects (Tables S3 and S4).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

As predicted by inclusive fitness theory, we found that kin discrimination was: (i) weaker in species where the average relatedness within a group was higher and less variable (Fig. 1a; Fig. 2), and (ii) stronger in species where the benefit of helping was greater (Fig. 1b). Indiscriminate helping can lead to substantial indirect fitness benefits when within-group relatedness is high and shows little variation, reducing selection for kin discrimination (Griffin & West, 2003; Boomsma, 2007). This can occur, either through dispersal patterns (specifically, strong philopatry) or low extra-pair mating by the breeding pair and/or low breeder turnover. We also found that the interaction between variation in relatedness and benefit of helping was key – the benefits of helping were correlated more strongly with kin discrimination in species with higher variance in relatedness between individuals. This demonstrates that selection for kin discrimination is greatest when high indirect benefits from helping combine with high variance in relatedness, which renders indiscriminate helping an unreliable way of directing help towards kin. Finally, we found that kin discrimination is greater when the probability rather than the amount of help provided is measured. The amount of help given may be influenced by a greater number of factors, such as the condition of helpers, that vary the costs to individuals of helping (Emlen & Wrege, 1988; Griffin & West, 2003) and so is a potentially less reliable indicator of kin discrimination.

These results build on previous work by Griffin & West (2003), demonstrating that variation in the importance of indirect fitness benefits can be explained across cooperatively breeding species by the reproductive benefits helpers provide and by population structure: relatedness between offspring and potential helpers. We tried to capture population structure both by categorizing species according to their dispersal patterns (natal vs. nonnatal helpers) and by measuring mean and variance in relatedness directly. There are pros and cons of these different approaches. Categorizing species as natal and nonnatal is likely to capture cues that animals use to assess likely relatedness to offspring in their group. That said much information is lost in this broad-brush approach. For example, in white-fronted bee-eaters and western bluebirds helpers that may be classified as nonnatals are often failed breeders that return to their natal nest after dispersal to help their parents (Emlen & Wrege, 1988; Dickinson et al., 1996). Furthermore, helpers that remain in their natal territory may make assessments of their relatedness to offspring based on other cues such as breeder turnover. However, what is key to assessing the role of indirect fitness benefits in the evolution of cooperation is the mean and variance in relatedness. In many cases this can measured directly, thanks to detailed long-term studies (Koenig & Dickinson, 2004). As we have shown this provides greater resolution in explaining kin discrimination across species in comparison to categorizing species according to their dispersal patterns.

Hamilton (1964a,b, 1971, 1972, 1975) demonstrated that limited dispersal leads to high within group relatedness, and hence favours cooperation. However, a potential problem with this idea is that limited dispersal can also lead to increased competition between relatives, which can reduce or even completely remove any effect of relatedness on selection for cooperation (Hamilton, 1971, 1975; Queller, 1992; Taylor, 1992a,b; West et al., 2002; Griffin et al., 2004). One way around this problem is if individuals disperse in groups of relatives (budding dispersal), which maintains relatedness within groups, but reduces competition between social partners (Gardner & West, 2006; Lehmann et al., 2006; Kummerli et al., 2009). This pattern of dispersal has been observed in several cooperative breeding vertebrates and may be an important factor in maintaining the indirect fitness benefits that individuals gain (Haldane, 1932; Clutton-Brock, 2002; Sharp et al., 2008; Williams & Rabenold, 2005; Bradley et al., 2007; Metheny et al., 2008).

Care should be taken to not over-interpret our results. We have investigated the two different routes by which cooperative breeders can gain indirect benefits from helping – population viscosity and kin discrimination. In contrast, we have not investigated the importance of direct fitness benefits, and so our results do not measure the relative importance of indirect and direct fitness in favouring helping behaviours (Griffin & West, 2003). Related to this, previous work has argued the importance of direct fitness benefits by demonstrating that levels of helping are adjusted in response to the cost of helping (e.g. Clutton-Brock et al., 1999, 2000). However, adjustments in cooperative behaviour in response to changes in the cost of helping, correspond to the c term of Hamilton’s rule (1963, 1964a,b), and so are predicted if the benefits of cooperation are either direct or indirect (e.g. Cant et al., 1996; Field et al., 2006).

To conclude, Hamilton (1964a,b, 1971, 1972, 1975) originally suggested that high relatedness could arise as a result of kin discrimination or limited dispersal. While the role of limited dispersal has gained much attention in microbes, where it has been shown to influence both cooperation and parasite virulence (Griffin et al., 2004; Kerr et al., 2006; MacLean & Gudelj, 2006; West et al., 2006; Boots & Mealor, 2007; Diggle et al., 2007; Gilbert et al., 2007; Ross-Gillespie et al., 2007; Kummerli et al., 2009; Wild et al., 2009), it has attracted less attention with respect to other taxa (West et al., 2002). Our results suggest that both kin discrimination and limited dispersal are important in cooperatively breeding vertebrates, and that they interact (see Rousset & Roze, 2007 for a general overview of theoretical work on kin discrimination). Specifically, kin discrimination increases indirect benefits when relatedness is too variable within groups to favour indiscriminate cooperation. A major future task is to link determinants of within group relatedness, such as the mating system, to the extent and form of cooperation (Boomsma, 2007).

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

We thank the Royal Society, ERC and an EGI fellowship to CKC for funding; Tim Clutton-Brock, Shinichi Nakagwa, Ben Sheldon and Stuart Sharp for useful discussion; the field workers that collected the data we have utilized.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

Appendices

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

Appendix

Table A1 List of studies providing data that were used to measure the effect of kinship on the amount or probability of help (rKin). Full details of how r-values were converted from test statistics can be found in Table S1 of Griffin & West (2003) unless provided below.

Common nameSpeciesReferencen Probability or amount of help measured?rKinNotes
Arabian babblerTurdoides squamicepsWright et al. (1999)92Amount−0.0471
Australian bell minerManorina melanophyrusClarke (1984); Wright et al. (in press)7Amount0.376 
Australian magpieGymnorhina tibicenFinn & Hughes (2001)72Probability0.045 
Brown hyaenaHyaena brunneaOwens & Owens (1984)159Amount0.185 
Carrion crowCorvus c. coroneCanestrari et al. (2005)28Amount0.2892
Dwarf mongooseHelogale parvulaCreel et al. (1991)181Probability0.283 
Florida scrub jayAphelocoma c. coerulescensMumme (1992)49Probability0.4063
Galapagos mockingbirdNesomimus parvulusCurry (1988)292Probability0.124 
Green woodhoopoePhoeniculus purpureusDu Plessis (1993)4Amount0.245 
Grey-capped social weaverPseudonigrita arnaudiBennun (1989)8Probability0.66 
Grey-capped social weaverPseudonigrita arnaudiBennun (1994)19Amount0.2794
KookaburraDacelo novaeguineaeLegge (2000)94Amount−0.156 
LionPanthero leoGrinnel et al. (1995)23Probability0.219 
Long-tailed titAegithalos caudatusRussell & Hatchwell (2001)17Probability0.882 
MeerkatSuricata suricattaClutton-Brock et al. (2001)43Amount0.2275
Pied kingfisherCeryle rudisReyer (1984)17Amount0.7216
Red-cockaded woodpeckerPicoides borealisKhan and Walters (2000)1184Probability0.062 
Seychelles warblerAcrocephalus sechellensisKomdeur (1994)112Probability0.633 
Seychelles warblerAcrocephalus sechellensisKomdeur (1994)6Amount0.901 
Spotted hyaenaCrocuta crocutaMills (1985)262Probability0.173 
Stripe-backed wrenCampylorhynchus nuchalisRabenold (1985)97Amount−0.208 
Superb fairy-wrenMalurus cyaneusDunn et al. (1995)23Amount−0.288 
Western bluebirdSialia mexicanaDickinson et al. (1996)321Probability0.326 
White-browed scrubwrenSericornis frontalisMagrath & Whittingham (1997)68Probability−0.0697
White-fronted bee-eaterMerops bullockoidesEmlen & Wrege (1988)59Amount0.200 
White-fronted bee-eaterMerops bullockoidesEmlen & Wrege (1988)203Probability0.5908
  • 1)
     There is a typo in the supplementary information table of Griffin & West (2003): the sample size associated with the P-value 0.128 is 75 and not 92 as stated. (The authors thank Stuart Sharp and Tim Clutton-Brock for bringing this error to our attention.)
  • 2)
     Conversion to rKin using χ2 = 8.08, rather than P = 0.02 (as used by Griffin & West (2003)) and revised sample size of 36. (The authors thank Stuart Sharp and Tim Clutton-Brock for bringing appropriate sample size to our attention.)
  • 3)
     Average rKin value for study was obtained from statistics measuring the effect of relatedness on the relative contribution to feeding (measured as feeding visits per hour) between ‘nonbreeders and nestlings’, F(1, 18) = 1.62, giving effect size of relatedness on feeding, rKin = 0.287, and between ‘nonbreeders and failed breeders’, F(1,26) = 2.38, giving effect size of relatedness on feeding, rKin = 0.290.
  • 4)
     Average rKin value (probability and amount combined) revised to 0.428 from 0.386 as published in Griffin & West (2003). Revised value does not alter any conclusions of previous analyses.
  • 5)
    r-Value used in previous analyses (r = 0.346, Griffin & West, 2003) was converted from P-value (P = 0.33) assuming one-tailed test. Corrected r-value assumes two-tailed test. (The authors thank Stuart Sharp and Tim Clutton-Brock for bringing this error to our attention.)
  • 6)
     Average rKin value (probability and amount combined) revised to 0.721 from 0.756 as published in Griffin & West (2003). Revised value does not alter any conclusions of previous analyses.
  • 7)
     The following chi-sqaure values were converted to give rKin measurements for the white-browed scrubwren: χ2 = 7, n = 68 gives r = −0.321; χ2 = 1.5, n = 63 gives r = 0.15; χ2 = 0.8, n = 63 gives r = 0.11; χ2 = 6, n = 63 gives r = −0.309; χ2 = 0.9, n = 63 gives r = 0.069. These were averaged to give rKin for study as a whole.
  • 8)
     Average rKin value (probability and amount combined) revised to 0.563 from 0.545 as published in Griffin & West (2003). Revised value does not alter any conclusions of previous analyses.

Table A2 List of studies providing data that were used to measure the effect of helpers on raising offspring to independence (rHelp). Full details of how r-values were converted from test statistics can be found in Table S2 of Griffin & West (2003) unless provided below.

Common NameSpeciesReferencenrHelpNotes
Arabian babblerTurdoides squamicepsWright (1998)270.490 
Australian bell minerManorina melanophyrusClarke (1989)120.6351
Australian magpieGymnorina tibicenP. Finn (Pers. Comm.)80.241 
Carrion crowCorvus c. coroneCanestrari et al. (2008)4530.1212
Dwarf mongooseHelogale parvulaCreel et al. (1991)190.656 
Florida scrub jayAphelocoma c. coerulescensMumme (1992)370.396 
Galapagos mockingbirdNesomimus parvulusCurry & Grant (1989)4500.1103
Green woodhoopePhoeniculus purpureusDu Plessis (1993)1440.102 
KookaburraDacelo novaeguineaeLegge (2000)24−0.187 
Long-tailed titAegithalos caudatusHatchwell et al. (2004)870.3144
MeerkatSuricata suricattaA. Russell (Pers. comm.); Clutton-Brock et al. (2001)1390.323 
Pied kingfisherCeryle rudisReyer (1984)250.822 
Red-cockaded woodpeckerPicoides borealisLennartz (1987)930.3141
Seychelles warblerAcrocephalus sechellensisKomdeur (1994)150.662 
Sociable weaverPhiletairus sociusDoutrelant et al. (2004)770.2681
Stripe-backed wrenCampylorhynchus nuchalisRabenold (1984)1040.584 
Superb fairy-wrenMalurus cyaneusDunn et al. (1995)92−0.0355
Western bluebirdSialia mexicanaDickenson et al. (1996)6130.108 
White-fronted bee-eaterMerops bullockoidesEmlen & Wrege (1988)1040.592 
  • 1)
     Details of how r-values were converted from test statistics can be found in Table 1 of Griffin et al. (2005).
  • 2)
     Wald statistic = 6.45 (n = 453) on effect of helpers on the probability of producing a fledgling gives effect size of r = 0.119; Wald = 6.85 (n = 453) on effect of helpers on number of fledglings produced gives effect size of 0.123. R-values were averaged to give overall rHelp value for study.
  • 3)
     R-help calculated from effect of helpers on fledgling success: F = 5.4, P = 0.02. Sample size was not given in text but was assumed to be 450 from Table 2 of Curry & Grant (1989).
  • 4)
     R-help calculated from effect of helpers on recruitment: F = 9.56, n = 87.
  • 5)
     R-help was given as r = −0.05 in Griffin & West (2003), has been corrected, treating P-value as two-tailed. The authors thank Stuart Sharp and Tim Clutton-Brock for bringing this error to our attention.

Table A3 List of studies providing data on mean and variance in relatedness between helpers/potential helpers and beneficiaries. Relatedness measures are based on molecular genetic data, unless otherwise stated.

Common nameSpeciesReferencenMean relatednessVar. relatednessNotes on data used in relatedness calculations
Arabian babblerTurdoides squamicepsWright et al. (1999)960.4290.121Relatedness between adults and offspring in ‘family’ and ‘nonfamily’ groups merged, using pedigree data confirmed by DNA fingerprinting; see text and Table 1 from ref.
Australian bell minerManorina melanophrysWright et al. (in press)2010.1960.020From Fig. 1 of unpublished manuscript
Dwarf mongooseHelogale parvulaCreel & Waser (1994)3600.3220.038Relatedness measured between helpers and breeders (Fig. 1 in ref.)
Florida scrub jayAphelocoma c. coerulescensMumme (1992)490.3880.031Relatedness measured between helper and nonhelping adults and offspring (Fig. 7a in ref.). Genealogical data used to assess relatedness, assuming monogamy as confirmed by genetic analysis.
Galapagos mockingbirdNesomimus parvulusCurry (1988)2900.2370.053Calculated from Table 3 in ref., including only those categories where identity of both parents confirmed.
KookaburraDacelo novaguineaeLegge & Cockburn (2000)2680.6230.149Calculated from Fig. 2b of ref. Relatedness value is band-sharing co-efficient and therefore not comparable with measures of relatedness derived from other studies.
Long-tailed titAegeithalos caudatusHatchwell et al. (2002)2610.0760.023Relatedness between males (potential helpers) and potential beneficiaries (males in nests within 900m radius) calculated from Fig. 2b from ref.
MeerkatSuricata suricattaGriffin (1998)2640.2420.025See Tables 5.2, 6.1 and 6.2 from ref.
Red-cockaded woodpeckerPicoides borealisKhan & Walters (2000)11840.3890.011See Table 3 from ref. Based on genealogical data.
Seychelles warblerAcrocephalus seychellensisKomdeur (1994)1020.3360.043Relatedness measure based on genealogical data assuming that breeding pair are parents of offspring at nests with only one egg laid, Table 1 from ref.
Stripe-backed wrenCampylorynchus nuchalisRabenold (1985)1510.3460.033See Table 1a from ref. Based on genealogical data.
Superb fairy-wrenMalarus cyaenusDunn (1995)2710.2530.031See Fig. 1 and Table 1 from ref.
Western bluebirdSialia mexicanaDickinson (1996)3210.2260.037See Fig. 2 and Table 4 from ref. Mainly genealogical, subset of relationships confirmed with DNA fingerprinting.
White-fronted bee eaterMerops bulockoidesEmlen (1988)3020.2250.040See Fig. 1 from ref. Based on genealogical data.

Appendix: References

Bennun, L. 1989. Communal breeding in grey-capped social weavers (Pseudonigrita arnaudi). D. Phil. thesis, Oxford University, Oxford.

Bennun, L. 1994. The contribution of helpers to feeding nestlings in grey-capped social weavers. Pseudonigrita arnaudi. Anim. Behav.47: 1047–1056.

Canestrari, D., Chiarati, E., Marcos, J. M., Ekman, J. & Baglione, V. 2008. Helpers but not breeders adjust provisioning effort to year-round territory resource availability in carrion crows. Anim. Behav.76: 943–949.

Canestrari, D., Marcos, J. M. & Baglione, V. 2005. Effect of parentage and relatedness on the individual contribution to cooperative chick care in carrion crows Corvus corone corone. Behav. Ecol. Sociobiol.57: 422–428.

Clarke, M.F. 1984. Co-operative Breeding by the Australian Bell Miner Manorina-Melanophrys Latham – a Test of Kin Selection Theory. Behav. Ecol. Sociobiol.14: 137–146.

Clarke, M.F. 1989. The pattern of helping in the bell miner (Manorina melanophrys). Ethology80: 292–306.

Clutton-Brock, T.H., Brotherton, P.N.M., Russell, A.F., O’Riain, M.J., Gaynor, D., Kansky, R., Griffin, A., Manser, M., Sharpe, L., McIlrath, G. M., Small, T., Moss, A. & Monfort, S. 2001. Cooperation, control, and concession in meerkat groups. Science291: 478–481.

Creel, S.R., Monfort, S.L., Wildt, D.E. & Waser, P.M. 1991. Spontaneous lactation is an adaptive result of pseudopregnancy. Nature351: 660–662.

Creel, S.R. & Waser, P.M. 1994. Inclusive fitness and reproductive strategies in dwarf mongooses. Behav. Ecol.5: 339–348.

Curry, R.L. 1988. Influence of kinship on helping-behaviour in galapagos mockingbirds. Behav. Ecol. Sociobiol.22: 141–152.

Curry, R.L. & Grant, P.R. 1989. Demography of the cooperatively breeding galapagos mockingbird, Nesomimus parvulus, in a climatically variable environment. J. Anim. Ecol.58: 441–463.

Dickinson, J.L., Koenig, W.D. & Pitelka, F.A. 1996. Fitness consequences of helping behavior in the western bluebird. Behav. Ecol.7: 168–177.

Doutrelant, C., Covas, R., Caizergues, A. & du Plessis, M. 2004. Unexpected sex ratio adjustment in a colonial cooperative bird: pairs with helpers produce more of the helping sex whereas pairs without helpers do not. Behav. Ecol. Sociobiol.56: 149–154.

Du Plessis, M.A. 1993. Helping behaviour in cooperatively-breeding green woodhoopoes: selected or unselected trait? Behaviour127: 49–65.

Dunn, P.O., Cockburn, A. & Mulder, R.A. 1995. Fairy-wren helpers often care for young to which they are unrelated. Proc. R. Soc. Lond. B259: 339–343.

Emlen, S.T. & Wrege, P.H. 1988. The role of kinship in helping decisions among white-fronted bee-eaters. Behav. Ecol. Socobiol.23: 305–315.

Finn, P.G. & Hughes, J.M. 2001. Helping behaviour in Australian magpies, Gymnorhina tibicen. Emu101: 57–63.

Griffin, A.S. 1998. A Genetic Analysis of Cooperative Breeding in Meerkats. University of Edinburgh, Edinburgh.

Grinnell, J., Packer, C. & Pusey, A.E. 1995. Cooperation in male lions – kinship, reciprocity or mutualism. Anim. Behav.49: 95–105.

Hatchwell, B.J., Ross, D.J., Chaline, N., Fowlie, M.K. & Burke, T. 2002. Parentage in the cooperative breeding system of long-tailed tits, Aegithalos caudatus. Anim. Behav.64: 55–63.

Hatchwell, B.J., Russell, A.F., MacColl, A.D.C., Ross, D.J., Fowlie, M.K. & McGowan, A. 2004. Helpers increase long-term but not short-term productivity in cooperatively breeding long-tailed tits. Behav. Ecol.15: 1–10.

Khan, M.Z. & Walters, J.R. 2000. An analysis of reciprocal exchange of helping behavior in the red-cockaded woodpecker (Picoides borealis). Behav. Ecol. Socobiol.47: 376–381.

Komdeur, J. 1994a. Experimental-evidence for helping and hindering by previous offspring in the cooperative-breeding Seychelles Warbler Acrocephalus-Sechellensis. Behav. Ecol. Socobiol.34: 175–186.

Komdeur, J. 1994b. The effect of kinship on helping in the cooperative breeding Seychelles Warbler (Acrocephalus-Sechellensis). Proc. R. Soc. Lond. B256: 47–52.

Legge, S. 2000. Helper contributions in the cooperatively breeding laughing kookaburra: feeding young is no laughing matter. Anim. Behav.59: 1009–1018.

Legge, S. & Cockburn, A. 2000. Social and mating system of cooperatively breeding laughing kookaburras (Dacelo novaeguineae). Behav. Ecol. Socobiol.47: 220–229.

Lennartz, M.R., Hooper, R.G. & Harlow, R.F. 1987. Sociality and cooperative breeding of red-cockaded woodpeckers, Picoides borealis. Behav. Ecol. Socobiol.20: 77–88.

Magrath, R.D. & Whittingham, L.A. 1997. Subordinate males are more likely to help if unrelated to the breeding female in cooperatively breeding white-browed scrubwrens. Behav. Ecol. Socobiol.41: 185–192.

Mills, M.G.L. 1985. Related spotted hyaenas forage together but do not cooperate in rearing young. Nature316: 61–62.

Mumme, D.L. 1992. Do helpers increase reproductive success? An experimental analysis in the Florida scrub jay. Behav. Ecol. Socobiol.31: 319–328.

Owens, D.D. & Owens, M.J. 1984. Helping-behaviour in brown hyaenas. Nature308: 843–845.

Rabenold, K.N. 1984. Cooperative enhancement of reproductive success in tropical wren societies. Ecology65: 871–885.

Rabenold, K.N. 1985. Cooperation in breeding by nonreproductive wrens – kinship, reciprocity, and demography. Behav. Ecol. Socobiol.17: 1–17.

Reyer, H.-U. 1984. Investment and relatedness: a cost/benefit analysis of breeding and helping in the pied kingfisher (Ceryle rudis). Anim. Behav.4: 1163–1178.

Russell, A.F. & Hatchwell, B.J. 2001. Experimental evidence for kin-biased helping in a cooperatively breeding vertebrate. Proc. R. Soc. Lond. B268: 2169–2174.

Wright, J. 1998. Helping-at-the-nest and group size in the Arabian Babbler Turdoides squamiceps. J. Avian Biol.29: 105–112.

Wright, J., Parker, P.G. & Lundy, K.J. 1999. Relatedness and chick-feeding effort in the cooperatively breeding Arabian babbler. Anim. Behav.58: 779–785.

Wright, J., McDonald, P.G., te Marvelde, L., Kazem, A.J.N. & Bishop, C.M. In press. Helping effort increases with relatedness in bell miners, but ‘unrelated’ helpers of both sexes still provide substantial care. Proc. R. Soc. Lond. B.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  10. Supporting Information

Table S1 Linear mixed model of predictors of kin discrimination (Zr-kin) with the mean and variance in relatedness excluded from the model to utilize data from all species

Table S2 Linear mixed model of kin discrimination (Zr-kin) entering each explanatory variables on their own

Table S3 Linear mixed model of predictors of kin discrimination (Zr-kin) removing data on the kookaburra, Dacelo novaeguineae

Table S4 Linear mixed model of predictors of kin discrimination (Zr-kin) removing data on the green woodhoopoe, Phoeniculus purpureus

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
JEB_1853_sm_si_Supplementary Material.doc109KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.