Behavioral Ecology doi:10.1093/beheco/arn097 Advance Access publication 8 August 2008
Task partitioning increases reproductive output in a cooperative bird Amanda R. Ridleya and Nichola J. Raihanib Department of Science and Technology/National Research Foundation Centre of Excellence, Percy Fitzpatrick Institute, University of Cape Town, Rondebosch 7701, Western Cape, South Africa and bLarge Animal Research Group, Department of Zoology, Downing Street, Cambridge, CB2 3EJ, UK
a
hereas environmental and phenotypic influences have been found to strongly influence reproductive success (Clutton-Brock 1988), the number of young that breeding individuals can produce per season is often limited by the amount of time it takes for young to become independent (Burley 1980; Verhulst et al. 1997). In multiple breeders (species that invest in more than 1 brood per breeding season), parents may trade-off investment in first-brood young against investment in subsequent broods. However, a decline in parental investment can be detrimental to young (Lycett et al. 1998), with Smith and Fretwell (1974) suggesting that offspring fitness is directly related to parental investment. This concept of the life-history trade-off between the quality and quantity of young produced has received considerable support, strengthened by recent evidence of long-term monitoring of offspring fitness (Gillespie et al. 2008). Among bird populations, there is evidence for highly variable periods of parental investment (Vega 2005; Ridley and Raihani 2007a). For example, in the Seychelles fody (Foudia sechellarum), fledglings received care for up to 4 months if they were the only brood of the season but only 2 months if parents invested in subsequent broods (Vega 2005). Such a large decline in the length of care received can have negative consequences for young, both in the short and long-term (Verhulst et al. 1997; Ridley and Raihani 2007a, Russell et al. 2007). Consequently, the optimal interbrood interval should be determined primarily by the trade-off between the benefits of continued investment in first broods and the initiation of new breeding attempts. Perhaps due to the high cost of simultaneously provisioning multiple broods, brood overlap (where dependent young from different-aged broods are raised simultaneously) is relatively rare among birds (Møller 2007). Although there are several cases where provisioning of the first brood occurs simulta-
W
Address correspondence to A.R. Ridley. E-mail: amanda.ridley@uct. ac.za. Received 24 January 2008; revised 4 July 2008; accepted 4 July 2008. The Author 2008. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail:
[email protected]
neously with incubation of a second brood (e.g., Burley 1980; Weatherhead and McRae 1990; Verhulst and Hut 1996), there are very few cases of simultaneous provisioning to multiple broods laid by the same breeding pair. However, in cooperatively breeding species, where multiple adults help to raise young, the cost of investing in overlapping broods is likely to be much lower than for species with biparental care. This is because helpers can lighten the load for breeders (Crick 1992), allowing them to invest in second broods sooner than they would otherwise be able. For example, Langen and Vehrencamp (1999) found that helpers in the white-throated magpie-jay (Calocitta formosa) became the primary caregivers of the first brood, allowing the breeding pair to reduce their provisioning effort and initiate a second brood before the first brood were nutritionally independent. Similar patterns have been observed in other cooperatively breeding bird species (e.g., splendid fairy-wrens, Malurus splendens, Russell and Rowley 1988; apostlebirds, Struthidea cinerea, Woxvold and Magrath 2005) and may be considered a form of interbrood division of labor among adult group members. Because the duration of the breeding season often determines the number and quality of clutches that can be produced (Burley 1980; Smith et al. 1989; Shutler et al. 2006; Møller 2007), interbrood division may be the best strategy to maximize reproductive output without affecting offspring survival and fitness in cooperatively breeding species. With multiple adults available to care for young, dividing care of different broods between subsets of adults could allow for shorter interclutch intervals and greater annual reproductive output. Task partitioning, where individuals within a social group perform different tasks, is considered to be of paramount importance in explaining the reproductive success of eusocial insects (Wilson 1971). In these species, the capacity for concurrent operations (through specialized worker roles) facilitates greater productivity (reviewed in Bourke and Franks 1995). We suggest that similar interbrood task partitioning may occur in cooperatively breeding vertebrates, allowing greater reproductive output without compromising the level of care offspring receive (cf., biparental species).
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
Parents often face a trade-off between the quality and quantity of young produced because terminating investment in current young could result in lower survival and future reproductive success, whereas initiating new breeding attempts could result in greater production of young. In cooperatively breeding species, helpers may alleviate this trade-off by assuming the role of primary caregivers to first broods, liberating breeders to initiate subsequent breeding attempts without compromising the level of care offspring receive. Here, we investigate the occurrence and consequences of brood overlap in the cooperatively breeding pied babbler (Turdoides bicolor). Brood overlap occurred only in groups and resulted in breeders primarily investing in second broods while helpers continued to provide care to first broods, resulting in dependent young from overlapping broods being raised simultaneously. Interbrood partitioning of care during brood overlap resulted in a greater production of young per season in groups (cf., pairs) without any effect on offspring survival, thus representing a reproductive benefit of task partitioning in cooperatively breeding species. Key words: brood overlap, interbrood division, pied babblers, provisioning behavior, reproductive success. [Behav Ecol 19:1136–1142 (2008)]
Ridley and Raihani • Task partitioning in a cooperative bird
In this paper, we investigate the occurrence of brood overlap and the consequences for offspring survival in the cooperatively breeding pied babbler (Turdoides bicolor), a medium-sized (75–95 g) sexually monomorphic passerine. Young receive an extended period of postfledging care in this species (up to 4 months postfledging, Ridley and Raihani 2007a) and second broods are commonly initiated while first broods are still completely nutritionally dependent. The purpose of this study was 3-fold. First, we asked whether the presence of helpers shortened the interclutch interval. Second, we asked whether the occurrence of overlapping broods affected patterns of provisioning behavior among adults. Finally, we asked whether the occurrence of overlapping broods affected offspring survival.
Study site and description Data were collected from October 2003 to December 2007 at the Kuruman River Reserve in the southern Kalahari desert, South Africa (2658#S, 2149#E). The study area is semiarid grassland and acacia savanna, with an average annual rainfall of 217 mm (for a detailed site description, see Raihani and Ridley 2007a). All pied babblers at the study site were individually recognizable using a unique combination of colored rings and were habituated to observation from a distance of approximately 2–3 m (for details of the habituation process, see Ridley and Raihani 2007b). The number of fully habituated groups present in the study population varied, averaging 11.2 6 2.1 (standard error, range 7–14) groups per year. Pied babbler groups typically comprise a dominant breeding pair (identified by aggressive, courtship, and mating behaviors) and several nonbreeding adults (individuals greater than 12 months old) that help to raise the young of the dominant pair, with a mean group size of 4.2 6 0.3 (range 3–12) adults. Subordinates are primarily retained young of the dominant pair, but in some cases, adults immigrate into groups as subordinate helpers (on average less than 5% of helpers per year are nonnatal). Helpers can be both male and female (average annual helper male:female sex ratio 1:2.63). Pairs attempting to breed without helpers comprise a very small proportion of the population (mean: 7.3 6 1.1% of adults per year, accounting for 10/181 breeding events observed over the duration of the study period). Breeding activity (defined as nest building, incubation, or provisioning young) occurs during the hot, rainy season (typically September–March), although breeding is occasionally observed until May (,5% of all breeding attempts). During each breeding attempt, a clutch of 2–5 eggs (mean 2.7 6 0.3) are laid and incubated for 13–15 days. Posthatching, young remain in the nest for 16.4 6 0.1 days (range 13–19, Raihani and Ridley 2007b). Postfledging, young remain nutritionally dependent on adult group members for an extended period (58.9 6 1.9 days, range 41–99, Ridley and Raihani 2007a). Groups can have up to 3 (mean: 1.25 6 0.15) successful breeding attempts (where at least 1 fledgling per brood reaches nutritional independence) per breeding season. Data collection Behavioral data Each group was visited 3 times per week during the breeding season for at least 3 h per observation session (n ¼ 2870.5 observation hours), and all behavioral activities were recorded ad libitum (Altmann 1974). Nests were checked daily to determine exact incubation, hatching, and fledging dates. During each observation session, the size and number of all food items delivered by all adults to all offspring were recorded
using a handheld data logger. All food items were divided into 4 size classes (for size classification, see Raihani and Ridley 2007a). Size classes were converted into biomass values (grams) by weighing 50 prey items representative of each class. The average biomass that each adult provisioned to each brood per hour was calculated as the sum of the number of prey items in each size class multiplied by the average weight of that class, divided by the number of observation hours. The body mass (grams) of each group member was measured by enticing individuals to stand on a top-pan balance for a small food reward at first light (before foraging had begun). Sexing Because it is not possible to determine sex from external characteristics in this species, small blood samples (50 lL) were collected from adults via brachial venipuncture. Nuclear DNA was extracted and polymerase chain reaction–based molecular sex determinations were conducted using the method described in Radford and Ridley (2006). Rainfall data Rainfall data were collected daily from the weather station located in the Kuruman River Reserve. Rainfall (millimeters) was summed to give the total amount of rainfall that fell in the 2 months prior to the event of interest. We considered 2 months as the appropriate time period to indicate food availability because there is commonly a protracted period of time between rainfall and increased insect abundance (Cumming and Bernard 1997). Analysis Interclutch interval The interclutch interval was measured as the number of days between fledging of the first brood and incubation of the second brood. Analysis was confined to first broods that hatched within the first 2 months of the breeding season (mid-September to mid-November) to avoid seasonal effects influencing the likelihood of renesting. Data were restricted to instances where young from the first brood survived to nutritional independence (defined as the point when young are no longer provisioned by adults and are able to self-forage, Ridley and Raihani 2007a) to avoid the influence of nest failure on time to renesting. The effects of group size (number of adults), brood size (number of fledglings in the first brood), breeding female body mass (average body mass at dawn in the 2 weeks prior to egg laying), and rainfall in the 2 months prior to incubation (millimeters) were included as potential explanatory terms in a linear mixed model (LMM) with a normal distribution of errors and an identity link function. Group identity was included as a random term. Data come from 37 second-brood attempts by 12 groups (range 1–6 second-brood attempts per group over the study period, average 2.7 6 0.4). The best-fit model was determined using the statistical concept of model selection using maximum likelihood estimation in the program MLwiN (v 2.02, Centre for Multi-Level Modelling, University of Bristol, UK). Using the Akaike’s information criteria for small data sets (AICc) approach (Burnham and Anderson 2002), a series of models were tested, with each model representing a biological hypothesis. First, we tested a basic model including only the constant, the random terms, and the residual variance. In subsequent models, we used the conservative approach of adding one explanatory term at a time to the basic model. If more than one term received considerable support, another model investigating the terms together was generated. Akaike weights were calculated for each model. These indicate the relative support for each model in relation to the other models, with support increasing with the Akaike
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
MATERIALS AND METHODS
1137
Behavioral Ecology
1138
Table 1 Model selection for the terms associated with the number of days until incubation of the second brood of the season after a first brood successfully fledged Deviance
K
AICc
DAICc
Akaike weight
Basic Group size 1 first-brood size
312.71 294.23
3 4
315.37 299.29
15.48 0
0.001 0.88
Group size Group size 3 breeding female body mass First-brood size Group size 3 rainfall Total rainfall in prior 2 months Breeding female body mass
300.47 299.04 306.86 307.21 310.88 311.99
4 6 4 6 4 4
304.19 305.35 310.58 313.71 314.60 315.70
4.89 6.24 11.28 14.42 15.30 16.40
0.08 0.04 0.003 0.001 0.004 0.002
Effect 6 standard error GS: 216.38 6 BS: 229.35 6 213.47 6 20.23 6 219.66 6 0.06 6 20.40 6 3.58 6
4.10 11.10 4.76 0.06 12.12 0.11 0.29 4.16
Analysis was conducted on 37 second-brood attempts by 12 groups. The table lists all candidate models tested (see Materials and methods for description of predictor terms). The basic model included the constant, the random term (group identity), and residual variance (r2). Deviance portrays the 22 log-likelihood output of each model. K is the number of parameters estimated in each model. AICc is for a small data set, which assesses the parsimony of each model. DAICc is the difference between that particular model and the one defined as the best model (in bold) for this set of models. Akaike weight indicates the relative support for each model in relation to the other models. GS, group size; BS, brood size.
weight value (Burnham and Anderson 2002). Overall, the model with the highest Akaike weight was considered the best model, but only significantly so if it differed from other candidate models by at least 2 AICc units. AIC and akaike weights for each candidate model are presented in Table 1; Wald statistics and P values for the best-fit models are presented in text. We used this method to determine the best-fit model for all subsequent analyses. Brood overlap We investigated the effect of brood overlap on adult provisioning behavior to first-brood young using data from 10 broods provisioned by 36 adults from 5 different groups. Data were restricted to instances where second broods hatched while first-brood fledglings were still entirely dependent on adults for food to eliminate the effects of fledgling age (and foraging independence) on provisioning behavior. The biomass delivered (grams per hour) to first-brood fledglings (whole brood combined) was calculated for each adult for 1) the 2-week period prior to overlap with a second brood and 2) the 2-week period after the second brood hatched and used as the response term. Observation times when adults were not available to provision fledglings (i.e., when they were nest building or incubating) were subtracted from the observation time used to calculate provisioning rates per hour for each adult. Broods that suffered losses due to predation or disease were excluded from analyses. Adult age (months posthatching), rank (breeder or helper), sex, weight (average body mass at dawn during the provisioning period), the presence of a next brood (present/ absent), brood size, group size, adult:fledgling ratio, and rainfall were included as predictors in the analysis. Group and adult identities were included as random terms. We then investigated the factors determining adult investment in incubation of second broods while first-brood young were still nutritionally dependent. For each adult, the amount of time spent incubating per observation hour was calculated for the entire incubation period and used as the response term. Adult rank, sex, and provisioning rate (to first-brood fledglings), together with rainfall, group size, first-brood size, and adult:fledgling ratio, were included as predictor terms. Group and adult identity were included as random terms. To determine whether changes in provisioning behavior to fledglings were a consequence of interbrood division of labor (where distinct subsets of adults provide care to different broods), rather than load lightening (where breeders lower their investment in the presence of helpers), we investigated
the factors determining adult provisioning rates to secondbrood young. Data were restricted to second broods that hatched while the first brood was still nutritionally dependent and were restricted to the first 2 weeks posthatching. First broods that began to forage independently during this period were eliminated from analyses. For each adult, the amount of food provisioned to nestlings per hour was calculated for the entire 2-week period and used as the response term. Rainfall, group size, nestling brood size, adult rank, sex, adult:fledgling ratio, and amount fed per hour to fledglings (averaged for each adult for the 2 weeks posthatching of the second brood) were included as predictor terms. Group and adult identity were included as random terms. Offspring survival effects To determine whether groups were able to raise more young per season than pairs, we looked at the effect of helper presence on reproductive output. The total number of broods produced by each group or pair per breeding season in which at least 1 young reached nutritional independence was specified as the response term. Group type (pair or group) and rainfall (averaged over the entire breeding season) were included as predictor terms, and group identity was included as a random term. Analysis was conducted on 40 broods that hatched from 17 groups or pairs over 4 breeding seasons. To determine whether brood overlap affected offspring survival, we used a Cox’s proportional hazards regression. We investigated survival over the period from fledging until individuals reached adulthood (1 year postfledging). Individuals still alive after this point were considered censored data. The response variable was the number of days postfledging that an individual survived, and the potential explanatory terms tested were group size, brood size (at time of fledgling), and brood overlap during dependent stage (yes/no). Analysis was conducted using a backward elimination process until only those terms that explained a significant amount of variation in survival were retained. RESULTS After successfully fledging a first brood, pairs never attempted to renest again that breeding season, whereas, groups always attempted to renest if the first brood successfully fledged (pairs ¼ 0/5 renests, groups ¼ 25/25 renests, Fisher’s Exact test, P , 0.001). Interclutch interval was significantly shorter in larger groups (Figure 1, Table 1) and in groups with a smaller number of
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
Model
Ridley and Raihani • Task partitioning in a cooperative bird
1139
fledglings surviving from the first brood (LMM: v2 ¼ 6.34, P ¼ 0.020). Second broods were initiated while first broods were still nutritionally dependent relatively frequently (65% of all second broods initiated). However, owing to the high levels of nest predation experienced by this species (37% of all nests predated at the nestling stage, Raihani and Ridley 2007b), the occurrence of simultaneous provisioning to multiple broods comprised only 13.5% of all breeding attempts. First-brood young received less food per hour from each adult in groups with a high adult:fledgling ratio (Table 2; LMM: v2 ¼ 41.13, P , 0.001). There was no difference between breeders and helpers in the amount of food delivered to first-brood young when there was no brood overlap. However, breeders were the primary investors in the initiation of a second brood while a first brood was still nutritionally dependent and fed first-brood young at a significantly lower rate during this time period (Table 2, Figure 2; LMM: v2 ¼ 9.71, P ¼ 0.002). Nest building took an average 6.6 6 0.7 days from
Figure 2 The biomass fed (grams per hour) by breeders versus helpers to firstbrood fledglings in relation to brood overlap. Mean 6 standard errors are generated from the predictions of the model with the highest Akaike weight presented in Table 2.
initiation to completion (range 3–13 days). Of 1728 observations of nest-building behavior, only 0.8 6 0.2% involved helpers. In addition, breeders were the primary incubators of the second brood (Table 3; LMM: v2 ¼ 46.89, P , 0.001), undertaking more than 85.1 6 1.2% of all incubation events observed. The more time spent incubating the second brood the less time breeders spent provisioning first-brood young (Table 3, Figure 3). While nest building and incubating, individuals were unable to provision fledglings (these behaviors were mutually exclusive because the foraging group was usually separated from the nest area by a considerable distance). Instead, helpers became the primary provisioners of first-brood fledglings while a second brood was incubated (Table 3, Figure 3; breeders: 0.22 g/h, helpers: 0.91 g/h; LMM: v2 ¼ 18.47, P , 0.001). The age of first-brood fledglings also influenced individual contributions to second-brood incubation, with incubation occurring less often when first-brood fledglings were very young (Table 3; LMM: v2 ¼ 17.26, P , 0.001). After a second brood hatched, there was a large decline in the amount of food delivered to the first brood by breeders
Table 2 Model selection for adult provisioning rates to first-brood fledglings Model
Deviance
K
AICc
DAICc
Akaike weight
Basic Brood overlap 3 rank 1 adult:fledgling ratio
103.02 43.91
4 8
106.43 51.06
55.37 0
9.1 3 10213 0.96
Brood overlap 3 rank Brood overlap 1 adult:fledgling ratio
50.52 58.08
7 6
57.67 63.94
6.609 12.88
0.03 0.006
Adult:fledgling ratio Brood size Brood overlap (yes/no)
79.07 80.52 86.73
5 5 5
83.69 85.14 91.35
32.63 34.07 40.29
7.9 3 1028 3.8 3 1028 1.7 3 1029
Group size Rank
93.07 101.23
5 5
97.67 105.84
46.62 54.78
7.2 3 10211 1.2 3 10212
Rainfall Adult sex
102.58 102.71
5 5
107.2 107.33
56.14 56.26
6.2 3 10213 5.8 3 10213
Effect 6 standard error
Brood overlap 3 rank: 0.41 adult: fledgling (A:F) ratio: 20.09 0.40 Brood overlap: 0.29 A:F ratio: 20.08 20.08 0.18 No: 0.0 Yes: 20.31 20.07 Dominant: 0.0 Subordinate: 0.12 0.01 Male: 0.0 Female: 20.06
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6
0.13 0.01 0.08 0.07 0.01 0.01 0.04 0.0 0.07 0.02 0.0 0.09 0.02 0.0 0.08
The table lists all candidate models tested (see Materials and methods for description of predictor terms). Analysis was conducted on 580.8 observation hours of provisioning by 36 adults to 10 broods from 5 groups. Group and adult identity were included as random terms. The model selected as the best predictor is highlighted in bold.
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
Figure 1 Interval (days) between fledging of the first brood and incubation of the second brood of the season in relation to group size. Raw data values are displayed. The line of best fit is generated from the predictions of the model with the highest Akaike weight presented in Table 1.
Behavioral Ecology
1140
Table 3 Model selection for terms associated with the amount of time spent incubating a second brood while a first brood was still nutritionally dependent Deviance
K
AICc
DAICc
Akaike weight
Basic Rank 3 biomass fed to fledglings per hour Rank
25.30 256.76 251.02
4 7 5
21.19 247.65 245.35
46.46 0 2.30
4.4 3 10211 0.54 0.17
Rank 1 fledgling age
252.55
6
245.21
2.44
0.16
Rank 3 biomass fed to fledglings per hour 1 fledgling age Rank 3 group size Fledgling age Group size Biomass fed to fledglings/hour Adult:fledgling ratio Adult sex
255.80
8
244.80
2.85
0.13
247.21 220.41 29.16 28.79 26.69 26.31
7 5 5 5 5 5
238.01 214.74 23.50 23.13 21.02 20.64
9.55 32.91 44.15 44.52 46.63 47.00
0.005 3.8 3 1.4 3 1.2 3 4.0 3 3.3 3
25.70
5
20.03
47.62
2.5 3 10211
Brood size
1028 10210 10210 10211 10211
Effect 6 standard error
0.53 Dom: 0.0 Sub: 20.40 Rank: Sub: 20.37 Age: 0.002 Rank 3 biomass: 0.51 Age: 0.006 20.05 0.008 20.03 0.17 20.02 Male: 0.0 Female: 0.08 0.02
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6
0.10 0.0 0.04 0.05 0.001 0.09 0.002 0.005 0.002 0.02 0.09 0.01 0.0 0.07 0.04
The table lists all candidate models tested (see Materials and methods for description of predictor terms). Analysis was conducted on 72.2 observation hours of incubation of 10 second broods by 36 adults in 5 groups. Group and adult identity were included as random terms. The model selected as the best predictor is highlighted in bold.
(Figure 2). This decline does not reflect a decline in effort by breeders because they instead became the primary provisioners of the second brood (Table 4; LMM: v2 ¼ 4.47, P ¼ 0.039) feeding second-brood nestlings at almost twice the rate of helpers (breeders: 0.32 6 0.5 g/h, helpers: 0.19 6 0.06 g/h). Unlike breeders, helpers continued to provision first-brood young at a similar rate as prior the second brood hatching (Figure 2). This resulted in significant brood overlap, with 2 broods being fed simultaneously by different subsets of adults within the same group. Interbrood division among adults resulted in reproductive benefits: groups were not only able to raise more young to independence than pairs per season (groups: 2.59 6 0.56, pairs: 0.67 6 0.22) but also able to successfully raise more broods (average number of broods with at least 1 young surviving to independence per season: groups ¼ 1.27 6 0.14; pairs ¼
Figure 3 Time spent incubating a second brood (minutes per hour) in relation to the biomass provisioned to first-brood young (grams per hour) during the same period for breeders (open circles, solid line) versus helpers (closed circles, dotted line). Raw data values are displayed. The lines of best fit are generated from the predictions of the model with the highest Akaike weight presented in Table 3.
0.56 6 0.18; v2 ¼ 11.19, P , 0.001). Young did not appear to suffer increased mortality from the presence of overlapping broods: there was no difference in the survival of young from independent broods compared with those from overlapping broods (Cox’s proportional hazards regression: v2 ¼ 0.721,114, P ¼ 0.40; Figure 4). DISCUSSION Brood overlap, where 2 nutritionally dependent broods are being provisioned simultaneously, may resolve the trade-off between terminating investment in first broods and the initiation of subsequent broods. In pied babblers, brood overlap results in the production of more broods per season by groups than pairs. Groups always initiated a second brood if the first brood survived. In contrast, pairs never initiated a second brood if the first survived. This is presumably because either 1) pairs may be birds of lower quality or inhabit territories of lower quality than breeders in cooperative groups and therefore do not have sufficient resources to raise more young or 2) pairs cannot terminate investment in the first brood while they are still dependent, as without helpers to feed them offspring would likely starve. The latter is recognized as one of the major constraints on reproductive output in multiple breeding species with biparental care (Smith and Fretwell 1974; Kluyver et al. 1977; Burley 1980). Brood overlap may be a better solution to raising more young than females simply laying more eggs in a single clutch because 1) multiple clutches may spread the predation risk, 2) there is a limit on the number of eggs that can be successfully incubated (Monaghan and Nager 1997), and 3) in species where the number of helpers present is unpredictable due to dispersal dynamics, large clutches may be a risky strategy. Successful brood overlap in pied babblers is not achieved by more adults feeding young indiscriminately but by the occurrence of task partitioning, with different subsets of adults providing care for different broods. In this species, breeders divert investment to second-brood young, whereas helpers continue to provide care for the first brood, similar to the behavior observed in the cooperatively breeding white-throated magpie-jay (Langen and Vehrencamp 1999). This behavior does not appear to be costly to first-brood young, with no effect of brood overlap on the probability of offspring survival to adulthood.
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
Model
Ridley and Raihani • Task partitioning in a cooperative bird
1141
Table 4 Model selection for adult provisioning rates to second-brood nestlings hatched while a first brood was still nutritionally dependent Deviance
K
AICc
DAICc
Akaike weight
Basic Rank
233.27 242.11
4 5
229.16 236.44
7.28 0
0.02 0.66
Rank 3 biomass fed to fledglings per hour Biomass fed to fledglings per hour Group size Adult:fledgling ratio Adult sex
242.95 235.88 235.42 233.30 233.29
7 5 5 5 5
233.84 230.21 229.76 229.19 229.18
2.61 6.23 6.69 7.26 7.26
0.19 0.03 0.02 0.02 0.02
Rainfall Rank 3 group size Group size 3 biomass fled to fledglings per hour Brood size Group size 3 sex
234.04 237.36 236.86 233.28 235.68
5 7 7 5 7
228.37 228.25 227.75 227.61 226.56
8.07 8.19 8.69 8.84 9.88
0.01 0.01 0.01 0.01 0.01
Effect 6 standard error
Dom: 0.0 Sub: 20.15 20.10 20.09 0.02 0.001 Male: 0.0 Female: 20.01 20.001 20.03 20.01 0.001 0.003
6 6 6 6 6 6 6 6 6 6 6 6 6
0.0 0.05 0.05 0.06 0.01 0.009 0.0 0.05 0.001 0.02 0.01 0.024 0.007
The table lists all candidate models tested (see Materials and methods for description of predictor terms). Group and adult identity were included as random terms. Analysis was conducted on 527 observation hours of provisioning by 36 adults to 10 broods from 5 groups. The model selected as the best predictor is highlighted in bold.
Thus interbrood division may allow first-brood young to still receive the benefits of extended postfledging care (Heinsohn 1991; Ridley and Raihani 2007a) even after a second brood has hatched. Interbrood division may not only allow groups to raise more young per season but could also provide other benefits as well. By becoming the primary caregivers of the first brood, helpers may ‘‘lighten the load’’ for breeders (Crick 1992). Previous research in the pied babbler (Ridley and Raihani 2007a) and other species (Hanssen et al. 2005; reviewed in Linden and Møller 1989; Møller 2007) have found that the cost of current reproduction and breeder condition is an important determinant of investment in subsequent broods. Additionally, breeder condition has been shown to have important influences on longevity, immunocompetence, and future reproductive potential (Dijkstra et al. 1990; Deerenberg et al. 1997; Hanssen et al. 2005). Thus, by transferring the care of first broods to helpers, task partitioning could allow breeders to invest in subsequent broods without a decline in breeder survival or condition. Further work is required to determine the long-term effects of group size and interbrood division on breeder longevity and fitness in the pied babbler.
Although there is considerable evidence that group size can provide reproductive benefits, most benefits have been measured in terms of provisioning rates and predator vigilance (reviewed in Dickinson and Hatchwell 2004). However, the results presented here suggest that the occurrence of task partitioning may influence reproductive output. Task partitioning, where individuals within a social group perform different tasks, is considered to be of paramount importance in explaining the reproductive success of eusocial insects (Wilson 1971). In these species, the presence of distinct worker castes allows overlapping generations of young to be cared for simultaneously (reviewed in Bourke and Franks 1995). In pied babblers, by dividing care of overlapping broods between subsets of adults, groups have shorter interclutch intervals and are able to successfully raise more broods per season than pairs and this appears analogous to the efficiency of worker castes in eusocial insects. By contributing a significant proportion of the care provided to first-brood young, the presence of helpers was an important factor influencing the initiation of second clutches. What remains unclear is why this arrangement is apparently stable in this system. It is possible that 1) breeders may be better than helpers at caring for younger broods or 2) breeders limit helper access to the nest area to prevent egg dumping or infanticide, with several observations of egg eating by helpers (Ridley AR, Nelson-Flower MJ, unpublished data) supporting this possibility. It is also possible that helping activities (e.g., incubation vs. provisioning fledglings) vary in terms of costs. For example, in meerkats (Suricata suricatta), babysitting is more costly than social digging and is the activity that the dominant pair invests in least often (Clutton-Brock et al. 2004). Whether or not the different roles taken on by breeders and helpers vary in costs (in terms of effort and vulnerability to predation) remain to be determined for pied babblers. However, several previous studies have found large variations in the cost (Russell et al. 2003) and participation in different helping activities among group members (CluttonBrock et al. 2004; Arnold et al. 2005), suggesting that this is a possible function of task partitioning that merits investigation. FUNDING
Figure 4 Kaplan–Meier estimates of the probability of fledglings surviving to adulthood (1 year postfledging) in relation to the occurrence of overlapping broods. Data were based on 64 young raised in nonoverlapping broods and 51 young raised in overlapping broods.
Newnham College, Cambridge; Association for the Study of Animal Behaviour; Department of Science and Technology/ National Research Foundation Centre of Excellence to A.R.R.; Natural Environment Research Council to N.J.R.
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
Model
1142
We thank the Northern Cape Conservation Authority for research permits and the Kalahari Research Trust for supporting our establishment of a study population at the Kuruman River Reserve. We thank the Kotze’s and de Bruin’s for land access and all Meerkat Project members for encouragement and support. Lucy Browning, Krystyna Golabek, Sarah Knowles, and Martha Nelson provided valuable contributions to data, life-history records, and habituation. Thanks also to 2 anonymous reviewers for useful comments and advice. This research was approved by the Ethics Committee, Department of Zoology, University of Cape Town.
REFERENCES
Langen TA, Vehrencamp SL. 1999. How white-throated magpie-jay helpers contribute during breeding. Auk. 116:131–140. Linden M, Møller AP. 1989. Clutch size, cost of reproduction, and covariation of life history traits in birds. Trends Ecol Evol. 4: 367–371. Lycett JE, Henzi SP, Barrett L. 1998. Maternal investment in mountain baboons and the hypothesis of reduced care. Behav Ecol Sociobiol. 42:49–56. Møller AP. 2007. Interval between clutches, fitness and climate change. Behav Ecol. 18:62–70. Monaghan P, Nager RG. 1997. Why don’t birds lay more eggs? Trends Ecol Evol. 12:270–274. Radford AN, Ridley AR. 2006. Recruitment calling: a novel form of extended parental care in an altricial species. Curr Biol. 16: 1700–1704. Raihani NJ, Ridley AR. 2007a. Adult vocalizations during provisioning: offspring response and post-fledging benefits in wild pied babblers. Anim Behav. 74:1303–1309. Raihani NJ, Ridley AR. 2007b. Variable fledging age according to group size: tradeoffs in a cooperative bird. Biol Lett. 3:624–627. Ridley AR, Raihani NJ. 2007a. Facultative response to a kleptoparasite by the cooperatively breeding pied babbler. Behav Ecol. 18: 324–330. Ridley AR, Raihani NJ. 2007b. Variable post-fledging care in a cooperative bird: causes and consequences. Behav Ecol. 18: 994–1000. Russell AF, Sharpe LL, Brotherton PNM, Clutton-Brock TH. 2003. Cost minimization by helpers in cooperative vertebrates. Proc Natl Acad Sci USA. 100:3333–3338. Russell AF, Young AJ, Spong G, Jordan NR, Clutton-Brock TH. 2007. Helpers increase the reproductive potential of offspring in cooperative meerkats. Proc R Soc Lond B Biol Sci. 274:513–520. Russell EM, Rowley I. 1988. Helper contributions to reproductive success in the splendid fairy-wren (Malurus splendens). Behav Ecol Sociobiol. 22:131–140. Shutler D, Clark RG, Fehr C, Diamond AW. 2006. Time and recruitment costs as currencies in manipulation studies on the costs of reproduction. Ecology. 87:2938–2946. Smith CC, Fretwell SD. 1974. Optimal balance between size and number of offspring. Am Nat. 108:499–506. Smith HG, Kallander H, Nilsson J-A. 1989. The significance of clutch overlap in Great Tits (Parus major). Ibis. 131:589–600. Vega LB. 2005. The ecology and conservation of the Seychelles fody, Foudia sechellarum [PhD thesis]. Reading (UK): University of Reading. Verhulst S, Hut RA. 1996. Post-fledging care, multiple breeding and the costs of reproduction in the great tit. Anim Behav. 51: 957–966. Verhulst S, Tinbergen JM, Daan S. 1997. Multiple breeding in the great tit. A trade-off between successive reproductive attempts? Funct Ecol. 11:714–722. Weatherhead PJ, McRae SB. 1990. Brood care in American robins: implications for mixed reproductive strategies by females. Anim Behav. 39:1179–1188. Woxvold IA, Magrath MJL. 2005. Helping enhances multiple components of reproductive success in the cooperatively breeding apostlebird. J Anim Ecol. 74:1039–1050. Wilson EO. 1971. The insect societies. Cambridge (MA): Belknap Press.
Downloaded from http://beheco.oxfordjournals.org at Macquarie Univ Library on August 9, 2010
Altmann J. 1974. Observational study of behaviour: sampling methods. Behaviour. 49:227–267. Arnold KE, Owens IPF, Goldizen AW. 2005. Division of labour within cooperatively breeding groups. Behaviour. 142:1577–1590. Bourke AFG, Franks NR. 1995. Social evolution in ants. Princeton: Princeton University Press. Burley N. 1980. Clutch overlap and clutch size: alternative and complementary reproductive tactics. Am Nat. 115:223–246. Burnham KP, Anderson DR. 2002. Model selection and multimodel inference: a practical information-theoretic approach. New York: Springer. Clutton-Brock TH. 1988. Reproductive success: studies of individual variation in contrasting breeding systems. Chicago: University of Chicago Press. Clutton-Brock TH, Russell AF, Sharpe LL. 2004. Behavioural tactics of breeders in cooperative meerkats. Anim Behav. 68:1029–1040. Crick HQP. 1992. Load-lightening in cooperatively breeding birds and the cost of reproduction. Ibis. 134:56–61. Cumming GS, Bernard RTF. 1997. Rainfall, food abundance and timing of parturition in African bats. Oecologia. 111:309–317. Deerenberg C, Arpanius V, Daan S, Bos N. 1997. Reproductive effort decreases antibody responsiveness. Proc R Soc Lond B Biol Sci. 264:1021–1029. Dickinson JL, Hatchwell BJ. 2004. Fitness consequences of helping. In: Koenig WD, Dickinson JL, editors. Ecology and evolution of cooperative breeding in birds. Cambridge: Cambridge University Press. p. 48–66. Dijkstra C, Bult A, Bijlsma S, Daan S, Meijer T, Zijlstra M. 1990. Brood size manipulations in the kestrel (Falco tinnunculus): effects on offspring and breeder survival. J Anim Ecol. 59:269–285. Gillespie DOS, Russell AF, Lummaa V. 2008. When fecundity does not equal fitness: evidence of an offspring quantity versus quality tradeoff in pre-industrial humans. Proc R Soc Lond B Biol Sci. doi: 10.1098/rspb.2007.1000. Hanssen SA, Hasselquist D, Folstad I, Erikstad KE. 2005. Cost of reproduction in a long-lived bird: incubation effort reduces immune function and future reproduction. Proc R Soc Lond B Biol Sci. 272:1039–1046. Heinsohn RG. 1991. Slow learning of foraging skills and extended parental care in cooperatively breeding white-winged choughs. Am Nat. 137:864–881. Kluyver HN, van Balen JH, Cave` AJ. 1977. The occurrence of timesaving mechanisms in the breeding biology of the great tit, Parus major. In: Stonehouse B, Perrins C, editors. Evolutionary ecology. Baltimore (MD): University Park Press. p. 153–169.
Behavioral Ecology
