Abstract

In species with biparental care, there is sexual conflict over parental investment because each parent benefits when their partner bears more of the reproductive costs. Such conflict can be costly for offspring, but recent theoretical work predicts that parents can resolve sexual conflict through behavioral negotiation, specifically by alternating their trips to provision nestlings. However, this idea has received almost no empirical attention. In this study, we test the hypothesis that parents alternate their delivery of food to offspring in long-tailed tits ( Aegithalos caudatus ) and investigate whether this coordination of parental care is associated with greater reproductive success. We show that parents alternate provisioning trips more than would be expected by chance and that parental alternation is repeatable across multiple observation periods at a nest. More alternation is associated with increased visit synchrony and increased food delivery to nestlings. Moreover, we found that nests with more alternation were less likely to be predated, probably resulting from reduced activity around the nest when parents coordinate their provisioning behavior. Our results support the hypothesis that alternation of offspring provisioning is a behavioral adaptation to reduce the costs of sexual conflict.

INTRODUCTION

In most vertebrates, especially birds and mammals, parents must provide some form of parental care for the successful production of offspring ( Clutton-Brock 1991 ). However, parental care is costly and has been linked to reduced lifespan and future reproductive output ( Stearns 1992 ). Therefore, a parent should invest according to the optimal trade-off between the benefits of caring for current offspring and the costs of that investment for future reproduction ( Williams 1966 ). In species with biparental care, an individual’s optimal parental investment also depends on the amount that its partner invests in the joint offspring. The shared benefits of offspring fitness means that each parent should prefer to invest less in the current offspring while its partner invests more. Thus, there is conflict between the parents over investment ( Trivers 1972 ), which may in turn be costly to offspring ( Parker 1985 ; McNamara et al. 2003 ).

Various theoretical models have sought to explain how stable systems of biparental care can evolve despite parental conflict over care. Early evolutionary models like that of Houston and Davies (1985) considered parental investment as a fixed trait that could change over evolutionary time. These models have gradually been succeeded by more biologically realistic “negotiation” models ( McNamara et al. 1999 , 2003 ; Johnstone and Hinde 2006 ; Johnstone 2011 ; Lessells and McNamara 2012 ) that accommodate the possibility of parents behaviorally negotiating their relative investment. Negotiation models predict that partial compensation, where 1 parent partially increases effort in response to a decrease by its partner, facilitates stable biparental care and prevents exploitation by either parent.

The predictions of negotiation models have been empirically tested many times, especially in birds where biparental care is the norm ( Cockburn 2006 ). A meta-analysis of experimental studies shows that, in general, parents do compensate incompletely for changes in care by their partners (reviewed in Harrison et al. 2009 ), as predicted by negotiation models. However, this effect is not universal across species: Some studies find complete compensation for a change in partner effort ( Mrowka 1982 ; Paredes et al. 2005 ), whereas others show no response ( Slagsvold and Lifjeld 1988 ; Schwagmeyer et al. 2002 ) or even a matching of effort between parents ( Hinde 2006 ; Meade et al. 2011 ). Hinde and Kilner (2007) suggested that this variation across species could be a function of the mechanisms through which negotiation operates. If parents are to behaviorally respond to each other’s effort, they must somehow integrate information about the investment of their partner. This may be achieved indirectly (e.g., from nestling condition or begging signals [ Lessells and McNamara 2012 ]) or directly from their partner’s behavior ( Dall et al. 2004 ). In the latter case, observation of a partner’s nestling provisioning frequency may provide a relatively simple way to monitor the investment of that individual. Despite its potential importance for testing negotiation models, the mechanisms of how negotiation would operate have received surprisingly little empirical attention.

Negotiation models also predict that each parent is forced to reduce its own investment below the level at which offspring fitness would be maximized ( McNamara et al. 1999 , 2003 ; Houston et al. 2005 ; Lessells and McNamara 2012 ). If so, offspring will suffer a fitness cost from parental conflict, a prediction borne out by empirical studies ( Royle et al. 2002 , 2010 ). However, Johnstone et al. (2014) have recently argued that this conflict and its potential cost may be reduced through a process referred to as “conditional cooperation” ( Keser and van Winden 2000 ; Gächter 2007 ), a tit-for-tat style alternation of provisioning where individuals withhold provisioning until the partner has provisioned. Johnstone et al.’s (2014) model predicts that alternation of provisioning trips results in greater total parental investment and an optimal provisioning rate that maximizes both parents’ fitness. The authors tested their model using data on great tits Parus major and found empirical evidence for rates of alternation of provisioning trips in that species that were greater than expected by chance ( Johnstone et al. 2014 ). However, whether alternation actually mitigates sexual conflict and improves reproductive success in wild systems, thus functioning as an evolutionary mechanism for maintaining biparental care, was not tested and remains unclear.

In practice, alternation of parental provisioning may also require that male and female nest visits are synchronized so that each parent can directly monitor their partner’s investment. This type of nest visit coordination has received far more attention than alternation, with studies focusing on 3 principal functions for synchronized provisioning behavior. First, if provisioning has a signaling function ( Kokko et al. 2002 ), then it may pay to synchronize visits with other carers at the same nest or nearby nests (e.g., Doutrelant and Covas 2007 ; but see McDonald, Kazem, et al. 2008 ; McDonald, Marvelde, et al. 2008 ). Second, synchronized provisioning may function to reduce predation risk for carers or for broods by reducing nest conspicuousness (e.g., Martin et al. 2000 ; Raihani et al. 2010 ). Finally, synchronous nest visits may serve a similar function to that proposed for alternation, by facilitating efficient provisioning of broods ( Shen et al. 2010 ) and thereby enhancing reproductive success ( Mariette and Griffith 2012 , 2015 ).

In this study, we investigate the alternation of provisioning visits by parents in socially monogamous long-tailed tit Aegithalos caudatus pairs and subsequently test the relationship between alternation, synchrony, and reproductive success. We first pool provisioning data from across breeding pairs to establish whether, across the population as a whole, alternation of provisioning trips occurs more frequently than expected by chance. Subsequently, we use between-pair variation in observed alternation to explore the predictors and correlates of this behavior. Previous experimental investigation of parental investment in the long-tailed tit has shown that parents match experimental changes in their partner’s provisioning rate ( Meade et al. 2011 ), suggesting that individuals monitor and coordinate their partner’s efforts with their own. Because long-tailed tit adults may forage 100 m or more from the nest ( Gaston 1973 ; Hatchwell BJ, unpublished data), active alternation of provisioning trips is only possible if parents are able to observe each other entering the nest—possibly by ensuring that nest visits occur synchronously or close together. We test 3 hypotheses: 1) parents alternate provisioning trips with those of their partner; 2) alternation confers fitness benefits for parents and their offspring; and 3) alternation is achieved through synchronous arrival at the nest, allowing each parent to observe the provisioning behavior of its partner.

METHODS

Study system

This study was based on a long-term dataset from a population of between 25 and 72 pairs of long-tailed tits in the Rivelin Valley, Sheffield, UK (53°23′N, 1°34′W). Each year, at least 95% of all adults in the population are individually recognizable from unique color ring combinations. Nestlings are ringed when 11 days old, and any unringed adult birds are caught in mist nets and color ringed (under British Trust for Ornithology license). Breeding pairs are identified in early spring; nests are found during nest building and are then monitored until fledging or failure. Typical clutch size is 10 eggs, which are incubated for around 15 days; hatching is synchronous and nestlings usually fledge aged 16–17 days old. Nests with nestlings are typically observed for 1h (mean ± standard deviation = 52±0.25min; range = 30–120min) on alternate days from day 2 of the nestling period (hatch day = day 0) until failure or fledging; the identity of each provisioning parent and the time of each visit are recorded in minutes (for further details on provisioning observations, see MacColl and Hatchwell 2003a ).

In this article, we use provisioning observations of breeding pairs recorded in 2000–2007 and 2010–2011. Although approximately half of breeding pairs with nestlings have helpers at the nest who provision the offspring ( Hatchwell et al. 2004 ), we restrict our analyses to nests without helpers. We also excluded any observation periods where adult provisioning rates were experimentally manipulated for other behavioral studies (e.g., Meade et al. 2011 ). Brooding of young nestlings by females reduces female nestling provisioning rates ( MacColl and Hatchwell 2003b ), so we only included observations that took place after females had ceased brooding (i.e., after day 5 of the nestling stage). Therefore, although nests are typically watched every 2 days from hatching to fledgling, the number of watches per nest used in our analyses is reduced by these constraints, as well as by nest failure and poor weather conditions. The final dataset after these exclusions included 248 nest watches at 98 nests, with an average of 2.5 watches per nest.

Calculating alternation of provisioning visits

For each nest watch, we calculated observed alternation, A , as A = F /( t − 1), where F is the number of times a bird fed after the other and t is the total number of feeds in the observation. Repeatability of alternation at a nest was determined by regressing 2 alternation values from nests where 2 or more watches were conducted ( n = 54 nests, mean number of watches = 4). Where 3 or more watches were conducted, we randomly selected which alternation values were regressed using a random number generator.

A certain amount of alternation will occur by chance as a function of the similarity between parents’ provisioning rates and also the interval (henceforth “interfeed interval”) between successive feeds by each parent. For example, in the case of provisioning rates, only parents feeding at the same rate can have 100% alternation, and this metric must decrease as the difference between provisioning rates increases. Furthermore, the interfeed interval must inevitably decline as provisioning rate increases. To determine whether individuals alternate feeding more than expected by chance, we calculated expected alternation using a bootstrapping procedure based on these 2 factors.

We first extracted all observed interfeed intervals of individuals provisioning at rates between 7 and 19 feeds/h. In our observed data, provisioning rates below and above these values were rare (2–7 feeds/h: n = 12, 4.4% of all watches; 19–31 feeds/h: n = 10, 3.7% of all watches) and were excluded due to low sample size. Considering all provisioning rates together, interfeed intervals varied considerably (range 1–55min; mean = 6min). For provisioning rate x , the interfeed intervals of birds who provisioned at rate x were randomized so that they were no longer associated with particular nest watches or individuals, which meant that our simulated data would not be derived directly from the observed data. This randomization was repeated for each of the 13 possible provisioning rates (7–19) and separately for males and females. We used these pools of randomized interfeed intervals to create simulated nest watches for the 169 different possible combinations of male and female feeding rate (13 male provisioning rates × 13 female provisioning rates). To generate expected alternation values where the female provisions at rate x and the male provisions at rate y , we randomly selected x − 1 interfeed intervals from the female pool of intervals associated with rate x and y − 1 interfeeding intervals from the male pool associated with rate y . We made separate cumulative totals of the interfeed intervals from x and y and then combined the cumulative totals from x and y , in ascending order, into 1 time series over which all the feeds and interfeed intervals occur ( Supplementary Figure S1 ). According to the sex associated with each interfeed interval, we could then calculate alternation as for the observed nest watches described above (see Supplementary Figure S1 for a schematic of the described method). We repeated this process until all the interfeed intervals from the female provisioning rate x and male provisioning rate y pools were used up, then moved onto the next combination of provisioning rates. In each combination of provisioning rates, we ran 10000 bootstrap simulations of the resulting alternation values to produce our simulated dataset of expected alternation.

The difference between female and male feeding rates has an inevitable influence on the degree of expected alternation, as explained above. Rather than investigate separately the observed and expected alternation for each of our 169 provisioning rate combinations (each with a small number of observed cases), we grouped the bootstrapped alternations of the 169 provisioning rate combinations according to the difference between the feeding rates of the 2 parents. This process yielded 13 categories of provisioning rate difference (i.e., 0 when parents fed at the same rate, up to 12 when parents fed at 7 and 19 feeds/h). The observed alternation values calculated from provisioning watches were grouped in the same way. To determine whether observed alternation differed significantly from expected, we tested whether mean observed alternation for each provisioning rate difference was greater than the bootstrapped expected alternation (± 95% confidence intervals [CIs]).

Predictors and fitness correlates of alternation

To investigate potential predictors of alternation at a nest, we used a mixed model including mean breeder age (in years), hatch date (to nearest day), duration of the pair-bond (in years), brood size (number of chicks on day 11 of the nestling period), and nestling age (days since hatching) as fixed effects. We also included the difference between the respective provisioning rates of the male and female because the difference between their provisioning rates should have a strong negative effect on alternation (see above). The start time of the nest watch (to nearest minute) was included to control for potential differences in provisioning behavior across the day, and we used nest identity as a random variable to account for repeated measures across nests.

To explore the fitness correlates of parental alternation, we used a set of linear models to test the relationship between alternation (mean across watches at a given nest) and 4 response variables. 1) Mean total provisioning rate: this was modeled as a Gaussian response, with brood size included as a covariate to control for potential variation in provisioning rate with the number of chicks. Because alternation is expected to increase with provisioning rate by chance as the interfeed intervals become smaller (see above and Johnstone et al. 2014 ), we modeled alternation as deviation from that expected by chance, to account for the random influence of provisioning rate. Deviation scores were calculated by subtracting the mean expected alternation from the mean observed alternation across all watches at each nest. 2) Mean chick mass: this was modeled as a Gaussian response, including alternation, brood size, and mean tarsus length as covariates, the latter of which controlled for structural body size variation. 3) Parental survival: expressed as survival of parents to the year following observations of provisioning behavior. We modeled survival as a binary response in a mixed model including alternation as a predictor and year as a random effect to account for survival differences between years. Dispersal out of the study area occurs in an individual’s first winter and thereafter the probability of resighting is almost 100% in our study population ( McGowan et al. 2003 ), so we could reliably measure survival from resighting data. We did not account for adult age effects in the survival model because there is no discernible effect of age on survival in the study population ( Meade et al. 2010 ). 4) Nest fate: nests were categorized as either “depredated” or “fledged,” and nest fate was modeled as a binary response variable. Nestling age is expected to be lower on average in nests that fail before fledging, so here we restricted our analysis to nests watched on day 6 of the nestling period ( n = 64). We included alternation, provisioning rate, and brood size as fixed effects and year as a random effect.

Analysis of provisioning synchrony

For each nest watch, we calculated a synchrony score from the time intervals between alternating parental nest visits. Synchrony, s , was calculated as s = F / t , where F is the number of alternated nest visits where the second visit was within 1min of the first and t is the total number of visits in the observation period. We tested the relationship between alternation and synchrony using a generalized linear mixed model with Poisson error and nest identity as a random effect. The provisioning rate during an observation period inevitably influences synchrony because as the rate increases, a greater proportion of feeds occur within a minute of each other. We therefore also included provisioning rate as an explanatory variable.

To investigate potential fitness correlates of coordinated parental nest visits, we examined nest activity by estimating the time that parents spend in the immediate vicinity of nests per provisioning visit. Data for this analysis were obtained from filmed observations in 2012 ( n = 10 nest watches at 7 nests) where the nest and the surrounding 10 m was visible throughout a provisioning watch of 40–50min. From the start of each nest watch, we timed (to the nearest second) how long 1 or both parent(s) were visible (i.e., within 10 m of the nest) until a cumulative total of 5min with 1 or both parents in the vicinity of the nest was reached. The total number of feeds during this 5min provided a measure of the number of feeds per unit time spent near the nest; pairs that provision more often during this cumulative 5min logically spend less time near the nest per feed. Mean synchrony scores for each of these nests were also calculated, using provisioning data available from separate observations recorded in the same breeding attempt. We then investigated the relationship between the number of feeds per 5min around the nest and mean synchrony scores across nests.

All statistical analyses were carried out in R Studio, version 2.15.3 (R Development Core Team 2014). In each analysis, we sequentially removed nonsignificant terms in order of lowest significance until only significant terms remained. Mixed models were performed in the “nlme” package ( Pinheiro et al. 2015 ), and general linear models were performed using r base packages. Figures were produced using the “ggplot2” package ( Wickham 2009 ).

RESULTS

Alternation of provisioning visits

Across all provisioning rate difference categories, alternation occurred more often than expected by chance, alternation being significantly higher in all categories than the upper 95% CI for bootstrapped expected alternation ( Figure 1 ). As expected, the difference in provisioning rate between males and females was a strong predictor of observed alternation: Smaller differences in the feeding rates of the 2 parents corresponded with greater mean alternation (Anova: F11,451 = 457.21, P < 0.001; Figure 1 ). Mean alternation was significantly greater than expected for all 13 categories of provisioning rate difference (Tukey Honest Significant Difference, all P < 0.05). Observed alternation for a given pair of birds provisioning the same nest was significantly correlated across watches ( Figure 2 ).

Figure 1

Mean observed and expected alternation in relation to the difference in provisioning rate between the parents. Bars for observed values represent SE, and bars for expected values represent 5% and 95% CIs. All observed mean alternation values exceed the upper 95% CIs of expected scores and, as expected, alternation decrease as a function of increasing provisioning rate difference.

Figure 1

Mean observed and expected alternation in relation to the difference in provisioning rate between the parents. Bars for observed values represent SE, and bars for expected values represent 5% and 95% CIs. All observed mean alternation values exceed the upper 95% CIs of expected scores and, as expected, alternation decrease as a function of increasing provisioning rate difference.

Figure 2

Relationship between 2 randomly sampled measures of alternation values from repeated observations at the same nest. The line represents fitted values (linear model [LM]: F1,52 = 33.1, P < 0.001, R2 = 0.377) with its SE represented by shaded areas.

Figure 2

Relationship between 2 randomly sampled measures of alternation values from repeated observations at the same nest. The line represents fitted values (linear model [LM]: F1,52 = 33.1, P < 0.001, R2 = 0.377) with its SE represented by shaded areas.

Predictors of parental alternation

Alternation was not predicted by any of the variables we tested: hatch date, time of day, combined breeder age (mean = 3.70 years ± 0.12 standard error [SE]), pair-bond duration (mean = 1.19 years ± 0.03 SE), brood size (mean = 8.31±0.14), or nestling age (mean = 9.57 days ± 0.20 SE), but it was significantly negatively correlated with provisioning rate difference ( Table 1 ).

Table 1

Predictors (a) and fitness correlates (b) of parental alternation in provisioning visits, with significant terms in bold text

a) Predictors of alternation Parameter Estimate ± SE P value  
 Provisioning rate difference −2.55±0.45 <0.01 
Combined breeder age −0.26±0.60 0.67 
Hatch day 0.09±0.16 0.57 
Length of pair-bond 0.09±2.19 0.97 
Brood size 0.80±0.53 0.14 
Time of day −0.63±0.36 0.08 
Nestling age 0.03±0.31 0.92 
b) Fitness correlates Fitness measure Parameter Estimate ± SE P value  
 Provisioning rate Deviation from expected alternation  0.37±1.10  0.03 
Brood size 0.49±0.34 0.15 
Chick mass Mean chick tarsus 0.43±0.06 <0.01 
Brood size −0.05±0.002 <0.01 
Mean alternation <0.01 ± <0.01 0.11 
Male survival Mean alternation  0.02±0.02  0.34 
Female survival Mean alternation 0.02±0.02 0.29 
Nest fate Alternation  0.04±0.02  0.04 
Brood size 0.28±0.66 0.51 
Provisioning rate 0.03±0.05 0.54 
a) Predictors of alternation Parameter Estimate ± SE P value  
 Provisioning rate difference −2.55±0.45 <0.01 
Combined breeder age −0.26±0.60 0.67 
Hatch day 0.09±0.16 0.57 
Length of pair-bond 0.09±2.19 0.97 
Brood size 0.80±0.53 0.14 
Time of day −0.63±0.36 0.08 
Nestling age 0.03±0.31 0.92 
b) Fitness correlates Fitness measure Parameter Estimate ± SE P value  
 Provisioning rate Deviation from expected alternation  0.37±1.10  0.03 
Brood size 0.49±0.34 0.15 
Chick mass Mean chick tarsus 0.43±0.06 <0.01 
Brood size −0.05±0.002 <0.01 
Mean alternation <0.01 ± <0.01 0.11 
Male survival Mean alternation  0.02±0.02  0.34 
Female survival Mean alternation 0.02±0.02 0.29 
Nest fate Alternation  0.04±0.02  0.04 
Brood size 0.28±0.66 0.51 
Provisioning rate 0.03±0.05 0.54 

Parental alternation and reproductive success

Mean provisioning rate was not significantly related to brood size, but there was a weak significant positive correlation with deviation from expected alternation ( Table 1 ; Figure 3a ). This increase in provisioning rate with alternation was not reflected in mean chick mass, which was not significantly correlated with alternation ( Table 1 ). Removing tarsus as a covariate improved the relationship between alternation and chick mass, but it remained nonsignificant ( P = 0.07). Alternation was not significantly associated with parental survival to the following year for either sex ( Table 1 ).

Figure 3

Relationships between (a) total provisioning rate (feeds/hour) and alternation ( F1,245 = 7.447, P < 0.01) and (b) synchrony of provisioning visits between parents (number of alternated feeds that occurred within a minute of the previous feed) and alternation ( F1,245 = 35.98, P < 0.001). Lines show the predicted values.

Figure 3

Relationships between (a) total provisioning rate (feeds/hour) and alternation ( F1,245 = 7.447, P < 0.01) and (b) synchrony of provisioning visits between parents (number of alternated feeds that occurred within a minute of the previous feed) and alternation ( F1,245 = 35.98, P < 0.001). Lines show the predicted values.

Nest fate (i.e., fledged successfully or depredated) was not significantly related to total parental provisioning rate or brood size. The probability of fledging was significantly higher in broods where parents alternated more ( Table 1 , Figure 4 ).

Figure 4

Mean + SE proportion of nests that produced fledglings as a function of varying degrees of alternation. Alternation was modeled as a continuous variable but for visualization purposes is grouped according to the level of alternation achieved on day 6 of the nestling period.

Figure 4

Mean + SE proportion of nests that produced fledglings as a function of varying degrees of alternation. Alternation was modeled as a continuous variable but for visualization purposes is grouped according to the level of alternation achieved on day 6 of the nestling period.

Provisioning synchrony

The mean synchrony score of provisioning parents was 13.3% ± 0.51 SE of feeds occurring within the same minute ( n = 247 nest watches on 97 pairs). Synchrony increased with both alternation (generalized linear mixed model [GLMM]: β = 0.012±0.002, P < 0.001; Figure 3b ) and, logically, with total provisioning rate (GLMM: β = 0.041±0.002, P < 0.001). In the sample of 7 nests filmed in 2012, pairs that provisioned more synchronously fed broods more often per 5min of nest activity by 1 or both parents ( Figure 5 ). We interpret this pattern as evidence that when parents are more synchronized in their provisioning visits, the number of feeds per unit time of parental activity near the nest increases. In other words, synchronized visits reduced the amount of parental activity near the nest per feed and hence may reduce nest conspicuousness to predators.

Figure 5

Relationship between the mean nest activity (see Methods for details) and mean synchrony scores for nests in 2012. The regression line is derived from values predicted by a linear model and shows a significant relationship between nest activity and mean synchrony score (LM: F1,5 = 13.78, P = 0.014), with SE represented by shading.

Figure 5

Relationship between the mean nest activity (see Methods for details) and mean synchrony scores for nests in 2012. The regression line is derived from values predicted by a linear model and shows a significant relationship between nest activity and mean synchrony score (LM: F1,5 = 13.78, P = 0.014), with SE represented by shading.

DISCUSSION

Our study shows that long-tailed tit parents alternate their provisioning visits significantly more often than expected by chance and that pairs are consistent in the degree of alternation they exhibit over the nestling period. Moreover, higher rates of alternation were associated with higher total provisioning and lower depredation risk. The latter is probably the result of the fact that pairs who alternated their provisioning more also provisioned more synchronously and therefore reduced the total amount of parental activity near the nest per feed. We discuss these findings and their implications below.

Alternation of provisioning trips

Parental alternation of provisioning was recently suggested as a mechanism by which parents can reduce their conflict over care, but actual tests of this idea are largely missing. To our knowledge, the only previous study that investigated parental alternation is that of Johnstone et al. (2014) , who show that great tit parents take turns in feeding young more than expected by chance. However, some evidence suggests that parental coordination may occur in other species. First, 2 experimental studies have shown that parents match the effort of their partner, which suggests some form of tit-for-tat negotiation or bargaining ( Hinde 2006 ; Meade et al. 2011 ). Second, synchrony of provisioning visits has been reported in several biparental species (e.g., long-tailed finches Poephila acuticauda [ van Rooij and Griffith 2013 ] and zebra finches Taeniopygia guttata [ Mariette and Griffith 2012 , 2015 ]) and cooperative breeders (e.g., bell miners Manorina melanophrys [ McDonald, Marvelde, et al. 2008 ; McDonald, Kazem, et al. 2008 ] and pied babblers Turdoides bicolor [ Raihani et al. 2010 ]). Because synchrony in the current study appears to be closely linked with alternation, it seems likely that behavioral coordination of provisioning might be a common way in which parents reduce sexual conflict, as predicted by Johnstone et al. (2014) .

Our finding of a significant repeatability of alternation between observations of the same pairs provisioning at the same nest suggests that alternation may be associated with either properties of individuals or the nest environment. However, because long-tailed tits are single brooded and have high mortality ( Meade et al. 2010 ) and divorce rate ( Hatchwell et al. 2000 ), we had only very few observations from more than 1 nest belonging to the same pair, so we could not make pairwise comparisons of alternation values from the same pairs in different breeding attempts. If such observations were possible, pairs might be expected to alternate more in their first joint breeding attempt in order to establish response rules and parental effort levels ( Johnstone 2011 ; Lessells and McNamara 2012 ), then relax the degree to which they alternate in subsequent attempts. On the other hand, the coordination of pair activities may improve as the number of pair’s breeding attempts increases, thus providing a mechanism for the frequently observed positive relationship between reproductive success and pair-bond longevity ( Black 1996 ). It would be interesting to test these alternative predictions in species with longer pair-bonds.

Repeatable alternation of provisioning visits for individual pairs could also arise simply as a function of repeatable provisioning rate differences. If a pair deviates the same amount from expected alternation in each observation and also maintains a fairly constant feeding rate, alternation would be similar across observations. Provisioning rate difference might be stable if birds are following the negotiation rules of Lessells and McNamara (2012) , where the male and female each establish and maintain a negotiated parental effort early in the breeding attempt. Indeed, previous studies indicate that parental effort is repeatable in long-tailed tits ( MacColl and Hatchwell 2003b ), as is the effect of an individual’s care on the effort of others ( Adams et al. 2015 ), supporting the notion of individual consistency in provisioning behavior. Thus, it seems most likely that repeatability in alternation can be explained by an early negotiation of effort and consistent subsequent behavior by each pair member, rather than being determined by other (e.g., environmental) factors.

Fitness correlates of alternation

We found a positive relationship between parental alternation and total provisioning rate, as predicted by the model of Johnstone et al. (2014) , thus supporting the notion that alternation can reduce the costs of sexual conflict for offspring. It should be noted, however, that neither the current study nor previous ones are able to rule out potential confounds of parental quality, which might simultaneously affect provisioning effort and the ability to coordinate care.

Contrary to expectations, neither higher total provisioning rate nor alternation directly resulted in improved nestling condition. Previous studies have shown that helpers in this facultative cooperatively breeding species cause a substantial increase in provisioning rate with positive consequences for nestling growth and subsequent recruitment ( MacColl and Hatchwell 2002 ; Hatchwell et al. 2004 ). On the other hand, evidence from other biparental passerines suggests that increased parental provisioning does not necessarily result in greater chick mass ( Titulaer et al. 2012 ), especially in species with large broods where any increases in provisioning rate are likely to be diluted by the high demand for food by the offspring ( Bonneaud et al. 2003 ). It is possible that in this study the increase in provisioning rate with alternation ( Figure 3a ) was too weak to cause detectable differences in nestling mass. Therefore, subtle links between provisioning rate and alternation, such as long-term survival benefits for offspring, may remain undetected.

Alternation of nest visits could also allow parents to moderate the survival costs of reproduction through negotiation of parental investment, but we found no relationship between alternation and adult survival to the following year. This result is perhaps not surprising considering the high annual mortality (ca. 44%) in our study population ( McGowan et al. 2003 ; Meade et al. 2010 ). Furthermore, the fact that total provisioning rate increases with alternation suggests that any benefit of increased efficiency of care through negotiation may be masked by increased provisioning effort by parents.

The more marked relationship between alternation and reproductive success reported here is that alternation in successful nests was significantly higher than that in nests that were depredated. Predators are likely to be attracted to nests through the activity of the parents ( Lima 2009 ; Raihani et al. 2010 ), so we suggest that this finding may be linked to the positive relationship between alternation and synchrony, because overall, parents spent less time near the nest per feed when provisioning visits were more synchronized. Therefore, alternation and the associated synchrony of provisioning visits appear to confer a direct fitness benefit by improving the chances that offspring survive to fledging.

It should also be noted that we investigated the coordination of provisioning visits in long-tailed tit broods fed by parents only, omitting those broods where helpers assisted parents in caring for nestlings. A previous study found that the presence of helpers at the nest did not increase the risk of nest predation ( Hatchwell et al. 2004 ), a result that appeared counter-intuitive given that activity near the nest is likely to increase its conspicuousness and given that long-tailed tits are too small to defend the nest against their principal nest predators (Hatchwell BJ, personal observation). It would be interesting to extend the analysis to examine the provisioning behavior of carers at helped nests to investigate whether nest visits exhibit similar, or even greater, levels of coordination to avoid an increased risk of attracting nest predators.

GENERAL CONCLUSIONS

Long-tailed tit pairs alternated their provisioning visits more than would be expected by chance. This coordination of parental care was associated with an increase in total provisioning rate and a reduction in nest predation. The finding of a positive relationship between reproductive success and alternation, combined with the repeatable nature of alternation, is correlative in nature but strongly supports the contention of Johnstone et al. (2014) that alternation of visits provides a means of mitigating the cost of sexual conflict for offspring. The behavioral mechanisms underlying parental investment strategies are vital to our understanding of the evolution of stable biparental systems, and our results contribute substantially to the notion that coordination of nest visits is a behavioral adaptation to mitigate the costs of sexual conflict over care.

SUPPLEMENTARY MATERIAL

Supplementary material can be found at http://www.beheco.oxfordjournals.org/

FUNDING

Data collection for the analyses reported here was supported (as part of a long-term study) by NERC (Grants: GR3/12908, NER/A/S/2002/00779, NE/E006655/1, NE/1027118/1).

We thank D. Childs for statistical advice and also C. Napper, N. Khwaja, P. Gullett, S.A. Kingma, and 2 anonymous reviewers for their comments on the manuscript. We also thank the many fieldworkers who collected the data. We are grateful to Sheffield City Council, Yorkshire Water, Hallamshire Golf Club, and private landowners in the Rivelin Valley for allowing us access to their land.

REFERENCES

Adams
MJ
Robinson
MR
Mannarelli
ME
Hatchwell
BJ
.
2015
.
Social genetic and social environment effects on parental and helper care in a cooperatively breeding bird
.
Proc Biol Sci
  .
282
:
201550689
.
Black
JM
.
1996
.
Partnerships in birds: the study of monogamy
  .
Oxford
:
Oxford University Press
.
Bonneaud
C
Mazuc
J
Gonzalez
G
Haussy
C
Chastel
O
Faivre
B
Sorci
G
.
2003
.
Assessing the cost of mounting an immune response
.
Am Nat
  .
161
:
367
379
.
Clutton-Brock
TH
.
1991
.
The evolution of parental care
  .
Princeton (NJ)
:
Princeton University Press
.
Cockburn
A
.
2006
.
Prevalence of different modes of parental care in birds
.
Proc Biol Sci
  .
273
:
1375
1383
.
Dall
SR
Houston
AI
McNamara
JM
.
2004
.
The behavioral ecology of personality: consistent individual differences from an adaptive perspective
.
Ecol Lett
  .
7
:
734
739
.
Doutrelant
C
Covas
R
.
2007
.
Helping has signalling characteristics in a cooperatively breeding bird
.
Anim Behav
  .
74
:
739
747
.
Gächter
S
.
2007
.
Conditional cooperation: behavioral regularities from the lab and the field and their policy implications
. In:
Frey
BS
Stutzer
A
, editors.
Psychology and economics
  .
Cambridge (MA)
:
MIT Press
. p.
19
50
.
Gaston
AJ
.
1973
.
The ecology and behavior of the long-tailed tit
.
Ibis
  .
115
:
330
351
.
Harrison
F
Barta
Z
Cuthill
I
Székely
T
.
2009
.
How is sexual conflict over parental care resolved? A meta-analysis
.
J Evol Biol
  .
22
:
1800
1812
.
Hatchwell
BJ
Russell
AF
MacColl
ADC
Ross
DJ
Fowlie
MK
McGowan
A
.
2004
.
Helpers increase long-term but not short-term productivity in cooperatively breeding long-tailed tits
.
Behav Ecol
  .
15
:
1
10
.
Hatchwell
BJ
Russell
AF
Ross
DJ
Fowlie
MK
.
2000
.
Divorce in cooperatively breeding long-tailed tits: a consequence of inbreeding avoidance?
Proc Biol Sci
  .
267
:
813
819
.
Hinde
CA
.
2006
.
Negotiation over offspring care? A positive response to partner-provisioning rate in great tits
.
Behav Ecol
  .
17
:
6
12
.
Hinde
CA
Kilner
RM
.
2007
.
Negotiations within the family over the supply of parental care
.
Proc Biol Sci
  .
274
:
53
60
.
Houston
AI
Davies
NB
.
1985
.
The evolution of cooperation and life history in the dunnock, Prunella modularis
. In:
Sibly
RM
Smith
RH
, editors.
Behavioural ecology
  .
Oxford
:
Blackwell Scientific Publications
. p.
471
487
.
Houston
AI
Székely
T
McNamara
JM
.
2005
.
Conflict between parents over care
.
Trends Ecol Evol
  .
20
:
33
38
.
Johnstone
RA
.
2011
.
Load lightening and negotiation over offspring care in cooperative breeders
.
Behav Ecol
  .
22
:
436
444
.
Johnstone
RA
Hinde
CA
.
2006
.
Negotiation over offspring care—how should parents respond to each other’s efforts?
Behav Ecol
  .
17
:
818
827
.
Johnstone
RA
Manica
A
Fayet
AL
Stoddard
MC
Rodriguez-Gironés
MA
Hinde
CA
.
2014
.
Reciprocity and conditional cooperation between great tit parents
.
Behav Ecol
  .
25
:
216
222
.
Keser
C
van Winden
F
.
2000
.
Conditional cooperation and voluntary contributions to public goods
.
Scand J Econ
  .
102
:
23
39
.
Kokko
H
Johnstone
RA
Wright
J
.
2002
.
The evolution of parental and alloparental effort in cooperatively breeding groups: when should helpers pay to stay
.
Behav Ecol
  .
13
:
291
300
.
Lessells
CM
McNamara
JM
.
2012
.
Sexual conflict over parental investment in repeated bouts: negotiation reduces overall care
.
Proc Biol Sci
  .
279
:
1506
1514
.
Lima
SL
.
2009
.
Predators and the breeding bird: behavioral and reproductive flexibility under the risk of predation
.
Biol Rev Camb Philos Soc
  .
84
:
485
513
.
MacColl
ADC
Hatchwell
BJ
.
2002
.
Determinants of lifetime fitness in a cooperative breeder, the long-tailed tit Aegithalos caudatus
.
J Anim Ecol
  .
73
:
1137
1148
.
MacColl
ADC
Hatchwell
BJ
.
2003a
.
Sharing of caring: nestling provisioning behavior of long-tailed tit, Aegithalos caudatus , parents and helpers
.
Anim Behav
  .
66
:
955
964
.
MacColl
AD
Hatchwell
BJ
.
2003b
.
Heritability of parental effort in a passerine bird
.
Evolution
  .
57
:
2191
2195
.
Mariette
MM
Griffith
SC
.
2012
.
Nest visit synchrony is high and correlates with reproductive success in the wild zebra finch Taeniopygia guttata
.
J Avian Biol
  .
43
:
131
140
.
Mariette
MM
Griffith
SC
.
2015
.
The adaptive significance of provisioning and foraging coordination between breeding partners
.
Am Nat
  .
185
:
270
280
.
Martin
TE
Scott
J
Menge
C
.
2000
.
Nest predation increases with parental activity: separating nest site and parental activity effects
.
Proc Biol Sci
  .
267
:
2287
2293
.
McDonald
PG
Kazem
AJN
Clarke
MF
Wright
J
.
2008
.
Helping as a signal: does the removal of potential audience alter helping behaviour in the bell miner?
Behav Ecol
  .
19
:
1047
1055
.
McDonald
PG
Marvelde
LT
Kazem
AJN
Wright
J
.
2008
.
Helping as a signal and the effect of a potential audience during provisioning visits in a cooperative bird
.
Anim Behav
  .
75
:
1319
1330
.
McGowan
A
Hatchwell
BJ
Woodburn
RJW
.
2003
.
The effect of helping behavior on the survival of juvenile and adult long-tailed tits Aegithalos caudatus
.
J Anim Ecol
  .
72
:
491
499
.
McNamara
JM
Gasson
CE
Houston
AI
.
1999
.
Incorporating rules for responding into evolutionary games
.
Nature
  .
401
:
368
371
.
McNamara
JM
Houston
AI
Barta
Z
Osorno
J
.
2003
.
Should young ever be better off with one parent than with two?
Behav Ecol
  .
14
:
301
310
.
Meade
J
Nam
KB
Beckerman
AP
Hatchwell
BJ
.
2010
.
Consequences of ‘load-lightening’ for future indirect fitness gains by helpers in a cooperatively breeding bird
.
J Anim Ecol
  .
79
:
529
537
.
Meade
J
Nam
KB
Lee
JW
Hatchwell
BJ
.
2011
.
An experimental test of the information model for negotiation of biparental care
.
PLoS One
  .
6
:
e19684
.
Mrowka
W
.
1982
.
Effect of removal on the parental care behaviour of the biparental cichlid Aequidens paraguayensis
.
Anim Behav
  .
30
:
295
297
.
Paredes
R
Jones
IL
Boness
DJ
.
2005
.
Reduced parental care, compensatory behavior and reproductive costs of thick-billed murres equipped with data loggers
.
Anim Behav
  .
69
:
197
208
.
Parker
GA
.
1985
.
Models of parent-offspring conflict. V. Effects of the behavior of the two parents
.
Anim Behav
  .
33
:
519
533
.
Pinheiro
J
Bates
D
DebRoy
S
Sarkar
D
,
R Core Team
.
2015
.
nlme: linear and nonlinear mixed effects models
  . R package version 3.1-120. Available from: http: // CRAN.R-project.org/package=nlme .
R Core Team. 2012. R: A language and environment for statistical computing . R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0
Raihani
NJ
Nelson-Flower
MJ
Moyes
K
Browning
LE
Ridley
AR
.
2010
.
Synchronous provisioning increases brood survival in cooperatively breeding pied babblers
.
J Anim Ecol
  .
79
:
44
52
.
van Rooij
EP
Griffith
SC
.
2013
.
Synchronised provisioning at the nest: parental coordination over care in a socially monogamous species
.
PeerJ
  .
1
:
e232
.
Royle
NJ
Hartley
IR
Parker
GA
.
2002
.
Sexual conflict reduces offspring fitness in zebra finches
.
Nature
  .
416
:
733
736
.
Royle
NJ
Wiebke
S
Dall
SR
.
2010
.
Behavioral consistency and the resolution of sexual conflict over parental investment
.
Behav Ecol
  .
21
:
1125
1130
.
Schwagmeyer
PL
Mock
DW
Parker
GA
.
2002
.
Biparental care in house sparrows: negotiation or sealed bid?
Behav Ecol
  .
13
:
713
721
.
Shen
SF
Chen
HC
Vehrencamp
SL
Yuan
HW
.
2010
.
Group provisioning limits sharing conflict among nestlings in joint-nesting Taiwan yuhinas
.
Biol Lett
  .
6
:
318
321
.
Slagsvold
T
Lifjeld
JT
.
1988
.
Ultimate adjustment of clutch size to parental feeding capacity in a passerine bird
.
Ecology
  .
69
:
1918
1922
.
Stearns
SC
.
1992
.
The evolution of life histories
  .
Oxford
:
Oxford University Press
.
Titulaer
M
Spoelstra
K
Lange
CY
Visser
ME
.
2012
.
Activity patterns during food provisioning are affected by artificial light in free living great tits ( Parus major )
.
PLoS One
  .
7
:
e37377
.
Trivers
RL
.
1972
.
Parental investment and sexual selection
. In:
Campbell
B
, editor.
Sexual selection and the descent of man 1871–1971
  .
Chicago (IL)
:
Aldine Transaction
. p.
136
208
.
Wickham
H
.
2009
.
Ggplot2: elegant graphics for data analysis
  .
New York
:
Springer
.
Williams
GC
.
1966
.
Natural selection, the costs of reproduction, and a refinement of Lack’s principle
.
Am Nat
  .
100
:
687
690
.

Author notes

Handling editor: Nick Royle