Coordination of care by breeders and helpers in the cooperatively breeding long-tailed tit

Abstract

In species with biparental and cooperative brood care, multiple carers cooperate by contributing costly investments to raise a shared brood. However, shared benefits and individual costs also give rise to conflict among carers conflict among carers over investment. Coordination of provisioning visits has been hypothesized to facilitate the resolution of this conflict, preventing exploitation, and ensuring collective investment in the shared brood. We used a 26-year study of long-tailed tits, Aegithalos caudatus, a facultative cooperative breeder, to investigate whether care by parents and helpers is coordinated, whether there are consistent differences in coordination between individuals and reproductive roles, and whether coordination varies with helper relatedness to breeders. Coordination takes the form of turn-taking (alternation) or feeding within a short time interval of another carer (synchrony), and both behaviors were observed to occur more than expected by chance, that is, “active” coordination. First, we found that active alternation decreased with group size, whereas active synchrony occurred at all group sizes. Second, we show that alternation was repeatable between observations at the same nest, whereas synchrony was repeatable between observations of the same individual. Active synchrony varied with reproductive status, with helpers synchronizing visits more than breeders, although active alternation did not vary with reproductive status. Finally, we found no significant effect of relatedness on either alternation or synchrony exhibited by helpers. In conclusion, we demonstrate active coordination of provisioning by carers and conclude that coordination is a socially plastic behavior depending on reproductive status and the number of carers raising the brood.

INTRODUCTION

Parental care is observed in some form in most bird species (Cockburn 2006). In altricial species, much of the burden of care occurs postnatally (Godfray and Johnstone 2000) and typically involves a shared caring system, with either biparental or cooperative brood care, in which helpers assist with raising a brood (Cockburn 2006). The benefits of parental care to offspring are well documented (Trivers 1974; Godfray 1995; Godfray and Johnstone 2000; Hinde et al. 2010), as are the fitness costs to parents, including accelerated senescence (Gustafsson and Pärt 1990), reduced survival (Dijkstra et al. 1990; Visser and Lessells 2001), and lower future reproductive success (Nilsson and Svensson 1996). Therefore, in both biparental and cooperative breeding systems, there exists a fundamental conflict over individuals’ relative level of investment in the current brood. Shared benefits of increased offspring survival and condition must be traded-off against individual costs of reduced future fitness (Trivers 1974; Hinde et al. 2010). This conflict means that optimal parental care behaviors that maximize lifetime reproductive success are dependent on the actions of others, so carers should use information from their social environment to adjust their own behavior (Houston and Davies 1985; McNamara et al. 1999; Johnstone and Hinde 2006). Recent work has hypothesized that coordination of care may have a crucial function as a mechanism for negotiating investment between carers, gathering information about others’ effort, building trust, and therefore resolving this conflict so that carers more closely match their optimal level of (allo)parental investment (Johnstone and Hinde 2006; Johnstone et al. 2014; Johnstone and Savage 2019).

Coordination can take the form of two, non-mutually exclusive behaviors such as alternation, which is the act of feeding in turn with another carer(s) such that each carer avoids consecutive visits, and synchrony, which is the act of feeding within a short interval of another carer’s feed (Figure 1). Previous studies of parental coordination have investigated biparental (e.g., Bebbington and Hatchwell 2016; Iserbyt et al. 2017, 2019; Leniowski and Wegrzyn 2018; Baldan and Griggio 2019; Baldan et al. 2019; Ihle et al. 2019b; Iserbyt et al. 2017; Lejeune et al. 2019) and cooperative care Alternatively, synchrony may be a result of collective foraging behavior that causes carers to return to the nest synchronously (Mariette and Griffith 2012, 2015, Baldan 2019,: Chapter 4). (e.g., Raihani et al. 2010; Koenig and Walters 2016; Khwaja et al. 2017; Savage et al. 2017). The results, so far, are mixed, with many demonstrating a higher than expected level of alternation (Johnstone et al. 2014; Savage et al. 2017; Baldan et al. 2019; Ihle et al. 2019b), synchrony (Lee et al. 2010; Raihani et al. 2010; Mariette and Griffith 2015), or both (Bebbington and Hatchwell 2016; Koenig and Walters 2016; Leniowski and Wegrzyn 2018; Lejeune et al. 2019), whereas another reported no apparent coordination (Khwaja et al. 2017).

Figure 1

Illustration of sequence of feeding visits at a hypothetical provisioning watch. Breeding male visits in blue, breeding female visits in red and helper 1 visits in yellow. Alternated and synchronized visits denoted by asterisk and triangle, respectively.

An important message emerging from these studies is that researchers must account for a degree of passive coordination expected by chance due to common factors, such as localized predator risk, weather conditions, and resource abundance, that potentially influence all carers’ provisioning refractory periods (Schlicht et al. 2016; Ihle et al. 2019a; Santema et al. 2019). Refractory periods, which are the minimum times it takes carers to gather food and return to the nest, are hypothesized to inflate levels of alternation and synchrony because they create a short period of time after a feeding visit in which a consecutive visit by the same individual is not possible, but alternated and synchronized visits are possible (Ihle et al. 2019a). For example, if intervals between feeds were consistent and identical for all carers at a nest, the pattern of visits would resemble perfect alternation even in the absence of coordination behavior. To account for passive coordination, randomization and simulation techniques derived from observed behavioral parameters are required to evaluate the level of observed coordination relative to that expected by chance from passive processes (e.g., Johnstone et al. 2014; Baldan and Griggio 2019; Baldan et al. 2019; Khwaja et al. 2019). Ihle et al. (2019a) reviewed the different null models used to evaluate coordination. They showed that randomization at the scale of within-nest, within-individual, and inter-visit was the most conservative approach (Supplementary Figure S1), because these conserve provisioning refractory periods. The difference between observed and expected coordination can then be measured, hereafter termed “active” coordination.

In cooperative breeding systems, additional factors such as the number of carers, carer status, and relatedness of carers to the brood must also be considered when determining an individual’s optimal behavior (Crick 1992; Hatchwell 1999; Savage et al. 2013a, 2013b, 2015; Green et al. 2016). Most previous studies have identified some form of coordination, but few have investigated the role of variable numbers of carers on coordination behavior (Savage et al. 2017). Because alternation is hypothesized to facilitate cooperation between carers (Johnstone et al. 2014), variation in the level of coordination between nests with different numbers of carers may inform our understanding of how and why birds coordinate. For example, a change in active coordination between group sizes may represent: 1) a change in the importance of coordination, perhaps due to reduced costs of parental care resulting from load lightening in large groups (Crick 1992); 2) a change in the ability of carers to monitor one another; or 3) a change in the potential for analyses to detect active coordination behavior.

The status of individual carers within-groups might also influence their coordination. For example, fathers, mothers, and helpers may provision broods differently (Harrison et al. 2009; Green et al. 2016), and Savage et al. (2017) suggested that alternation was most prominent in breeders and helpers that invested more highly in broods. Synchronous feeding has also been proposed as a means of signaling effort to other carers (Doutrelant and Covas 2007; Koenig and Walters 2016; Trapote et al. 2021), so this hypothesis predicts that if signaling confers direct benefits to helpers, such as in a pay-to-stay system (Gaston 1978; Kokko et al. 2002), more active synchrony should be performed by helpers. Alternatively, synchrony may be a result of collective foraging behavior that causes carers to return to the nest synchronously (Mariette and Griffith 2012, 2015). Moreover, if coupled with a leader–follower relationship, for example, if helpers are more likely to follow a breeder back to the nest, this may result in greater synchrony by helpers.

In this study, we investigated how levels of coordination varied with the number and status of carers in the long-tailed tit, Aegithalos caudatus. Long-tailed tits are short-lived passerine birds, with a facultative cooperative breeding system in which failed breeders redirect their care to help raise the offspring of other breeders, to which they are typically related (Hatchwell et al. 2014; Hatchwell 2016). About half of all broods in our study population are raised by their parents alone, the remainder being fed by their parents assisted by helpers. Helping is a kin-selected adaptation that allows failed breeders to gain indirect fitness benefits by caring for their relatives’ offspring, thereby increasing relatives’ breeding success (Hatchwell et al. 2004, 2014). Previous studies have shown that the care provided by helpers varies with relatedness. First, helpers show an active preference for helping kin rather than non-kin (Russell and Hatchwell 2001; Leedale et al. 2018). Second, helpers provision at a higher rate when they are more closely related to a brood (Nam et al. 2010; Leedale et al. 2020).

Given that a helper’s relatedness influences their investment decisions, we might also expect that it would influence coordination behavior. For example, if carer coordination benefits the brood, less-related helpers may coordinate less due to their lower genetic investment in the brood (Savage et al. 2017). Alternatively, the shared interest of parents and helpers in the brood may be lower for more distantly related helpers, resulting in greater conflict and hence a greater need for coordination. This cooperative breeding system with variable numbers of carers and variable relatedness between carers and the shared brood is well suited for testing whether carers coordinate their care and the factors influencing the level of coordination.

Bebbington and Hatchwell (2016) reported that long-tailed tit parents provisioning at biparental nests coordinate their care so that observed alternation and synchrony were higher than expected by chance. That study, however, utilized a null model that did not fully account for expected alternation and synchrony caused by refractory periods (Ihle et al. 2019a). In this study, we build on the findings of Bebbington and Hatchwell (2016) by investigating the impact of the number of carers, carer status, and relatedness of helpers on coordination of care, using a more conservative approach to analyze a larger sample of biparental nests, as well as cooperative nests with up to three helpers. Our first objective was to investigate whether carers working in different group sizes coordinated their provisioning by comparing observed alternation and synchrony to that expected by passive processes (Ihle et al. 2019a). Second, we investigated individual variation in coordination, examining the extent of within-individual and within-nest repeatability in the level of active coordination, and whether levels of active alternation and synchrony varied in relation to the status of the carer (male breeder, female breeder, or helper). Finally, we examined variation in the degree of coordination by helpers to determine whether either alternation or synchrony was influenced by their relatedness to the brood.

METHODS

Study system and data collection

We used data from a long-term study of a population of long-tailed tits in the Rivelin Valley, Sheffield, UK (53°23ʹN, 1°34ʹW) from 1994 to 2019. The field site is ~3 km² with a population of 25–72 breeding pairs (Hatchwell 2016). Each year ~95% of adult birds were marked (under British Trust for Ornithology license) with a unique combination of two color rings on one leg and a BTO metal ring on the other. The adult annual mortality rate is ~50% (Meade and Hatchwell 2010), and ~20% of new recruits into the adult population were ringed as nestlings in the study site, although the remaining ~80% of new recruits were unringed adult immigrants that dispersed into the population. Unringed birds were captured in mist-nests during the nest-building period and DNA samples collected (under Home Office license) for genotyping and social pedigree reconstruction. Nests were found by following adults and once located, were monitored every 2–3 days, with daily visits around the expected hatch date. Median clutch size is 10 eggs (range: 4–12), which are incubated for ~15 days (Hatchwell 2016). Hatching is extremely synchronous within clutches, with all chicks typically hatching within 24 h of the first. Initial hatch date was recorded as day 0, and chicks were ringed and counted on day 11. Protocols for provisioning watches (hereafter “watches”) were broadly consistent throughout the study. In most cases, watches of duration ~60 min were carried out every other day, starting on day 2, either by direct field observation or by video camera, for later review (69% of watches were between 45 and 65 min). Watches were carried out between 04:00 AM and 06:00 PM, with 89% starting between 06:00 AM and 02:00 PM. Watches were performed until a nest was predated, abandoned, or chicks fledged, typically on days 16–18.

For ~5 days post-hatching, nestlings are brooded regularly by their mothers, who provision offspring only occasionally, whereas fathers either feed the offspring directly or give food to the mother, who then feeds the chicks. We restricted our analysis, therefore, to watches at day 6 and older, when both parents provision offspring directly. Long-tailed tits exhibit facultative cooperative breeding (Lack and Lack 1958; Hatchwell 2016), meaning nests may be uniparental (1 carer, in the rare event of a parent dying) or biparental (2 carers) or cooperative (>2 carers). For this study, we restricted analysis to watches of biparental and cooperative nests with up to 5 carers (i.e., social parents and up to 3 helpers). Our dataset contained 65% (516) of watches from biparental nests and 21% (171), 11% (88), and 3% (20) from nests with 3, 4, and 5 carers, respectively. Before starting a watch, ~10 min was usually allowed for birds to recover from observer disturbance and we restricted analysis to watches of total duration ≥30.0 and ≤180.0 min, with duration defined as the time between first and last observed feeds. Mean watch duration (±SD) was 54.8 ± 14.4 min (range 30–118 min, N = 795 watches). We omitted watches where the identity of any provisioning visit was unknown, and from nests that were manipulated for other behavioral studies (e.g., Meade et al. 2011). Watches were used from 24 years between 1994 and 2019, with 2007 and 2009 excluded because experiments conducted in those years meant that they contained no watches matching our criteria. In total, our dataset included 795 watches performed at 250 unique nests, involving 192 different breeding males, 203 breeding females, and 144 helpers.

Calculating coordination

We analyzed alternation and synchrony as the absolute number of alternated and synchronized feeding visits in a provisioning watch, respectively. We defined an alternated visit as any non-consecutive provisioning visit (i.e., a visit occurring after the provisioning visit of any carer other than itself) and a synchronized visit as an alternated visit occurring within 2-min of the previous feed (Figure 1). We chose an interval of 2-min in accordance with previous studies (Mariette and Griffith 2015; Bebbington and Hatchwell 2016; Ihle et al. 2019a), and further analyses revealed that number of synchronized visits was highly correlated for 1-, 2-, and 3-min intervals (Pearson correlations: 1 vs. 2 min, r = 0.97, df = 793, P < 0.001; 2 vs. 3 min, r = 0.97, df = 793, P < 0.001; 1 vs. 3 min, r = 0.94, df = 793, P < 0.001), and analyses of synchrony with different intervals produced qualitatively the same results.

We calculated observed alternation and synchrony directly from visit sequences and times recorded through field observation, generating coordination measures per watch and for each individual carer present in each watch (Figure 1). We generated expected data by null model randomization of observed data, with the binary factor “Data type” specifying whether data were observed or expected. In accordance with the most conservative method of calculating expected alternation and synchrony recommended by Ihle et al. (2019a), our null models used a within-watch, within-individual randomization procedure in which the order of provisioning visits within a watch was randomized in a manner that preserved the length and identity of each period between feeding visits (inter-visit intervals; Supplementary Figure S1). We calculated expected numbers of alternated and synchronized visits, both for group total and for individual carers, from the median of 1000 iterations of the null model applied to each provisioning watch. We used median values to preserve integer values for subsequent analysis in Poisson-distributed linear models; mean and median values were highly positively correlated (Pearson correlations: alternated visits, r = 0.99, df = 793, P < 0.001; synchronized visits, r = 0.99, df = 793, P < 0.001).

Calculating kinship

To calculate pairwise values of pedigree relatedness of helpers to parents, we constructed an additive relationship matrix using the R package NADIV (Wolak 2012), partially reconstructed using molecular genetic data from up to 17 microsatellite loci to perform offspring–parent reconstruction on CERVUS v.3.0.7 (Kalinowski et al. 2007) and sibling–sibling reconstruction on KINGROUP v.2 (Konovalov et al. 2004). Building on the social pedigree and protocol used in Leedale et al. (2018, 2020), we expanded the pedigree to include 2018 and 2019 data. Our study population is open, so even after reconstruction the social pedigree remained incomplete; therefore, where necessary we omitted data with incomplete pairwise relatedness metrics to either social parent.

Statistical analysis

All statistical analysis was performed in R version 4.0.2 (R Core Team, 2020). All models were built and analyzed using the lme4 package (Bates et al. 2015) and lmerTest (Kuznetsova et al. 2017), except for our repeatability models which were built and analyzed using the rptR package (Stoffel et al. 2017).

Collective coordination models (Alt-C and Sync-C)

To investigate collective alternation and synchrony performed by all carers at a nest, we defined two Poisson-distributed generalized linear mixed effects models (GLMM) named “Alt-C” and “Sync-C,” respectively. The response variables to these models were the number of alternated visits (collective) and synchronized visits (collective) by all carers at each watch, respectively. To control for observation and population structure, these models were built with the following random effects: “Year,” “Nest ID,” “Watch ID,” “Male ID,” “Female ID,” “Helper1 ID,” “Helper2 ID,” “Helper3 ID,” and “Row reference” (see Table 1 for explanation). The fixed effects tested were as follows: “Data type” (observed vs. expected values of alternation and synchrony), “Provisioning rate (collective),” “Carer number,” “Watch duration,” “Brood size,” “Time of day,” “Brood age,” “Hatch date,” and “AMax (or SMax)” (Table 1). We focused our analysis on “Data type” and two-way interactions with other fixed effect terms, as a disparity between observed and expected data represents the level of active coordination performed.

Table 1

Details of response variables, fixed effects, random effects used in models and non-model terms

	Description
Response variables
Alternated or synchronized visits (collective)	The absolute number of alternated or synchronized provisioning visits performed by all carers in a group during a watch.
Alternated or synchronized visit (individual)	The absolute number of alternated or synchronized provisioning visits performed by a given carer during a watch.
Active alternated or active synchronized visits (individual)	The absolute number of actively alternated or synchronized visits performed by a given carer during a given watch, calculated by subtracting a carer’s expected values from their observed values for each watch.
Random effects
Year	The unique identifier for which field season year the provisioning watch was performed in. Included to account for variation caused by site wide environmental factors which may vary between years.
Nest ID	The unique identifier for each nest. Because multiple watches were performed at the same nest, this was included to account for variation caused by factors specific to the local environment, for example, invertebrate abundance and microclimate.
Watch ID	The unique identifier for each nest watch. Included to account for similarity between observed and expected data from the same watch.
Row reference	Observation level random effect, included as per recommendation from Ihle et al. (2019a) to account for overdispersion in Poisson-distributed models.
Individual carer IDs	The unique identity of each individual carer present during a watch (“Male ID,” “Female ID,” “Helper1 ID,” “Helper2 ID,” and “Helper3 ID”). Included as random effects to account for between individual variation in provisioning behavior because some individuals were observed provisioning at multiple watches and nests (MacColl and Hatchwell 2003; Adams et al. 2015). However, 37/795 watches (4.7%) featured visits by a single unringed carer, these individuals could therefore not be distinguished with a unique Carer ID. Repeating analyses with these watches omitted produced qualitatively the same results in all cases.
Collective carer ID	The unique identifier for a particular combination of carers seen provisioning during a given watch. Few pairs persist across years due to high annual mortality (Meade et al. 2010) and divorce rate (Hatchwell et al. 2000) and carer combinations may change daily as helpers join and leave, so “Collective Carer ID” was included to account for between group variation in provisioning coordination.
Fixed effects
Data type	Factor designating whether data were observed (from field data) or expected (generated by null model). As we used expected values as our baseline level of coordination “Data type” was therefore a proxy for the level of active coordination performed. Our assessment of whether other terms significantly impact active coordination was investigated by looking at their interaction with “Data type.”
Provisioning rate (collective)	The total number of provisioning visits performed by all carers during a given watch per hour, modelled as continuous numerical values (mean: 24.1; range: 5.8–69.2).
Provisioning rate (individual)	The total number of provisioning visits performed by a given carer during a provisioning watch per hour, modelled as continuous numerical values (mean: 9.6; range: 0.5–34.3).
Brood size	The number of live chicks recorded at the nest on day 11, modelled as integer values (median: 8; range 1–12). Included because prior studies have shown that brood size influences provisioning behavior by mediating brood demand (Meade et al 2010). We used a single recorded measure of brood size, rather than recording brood size after each watch to minimize nest disturbance and because brood reductions are rare in long-tailed tits, having only a 0.2% daily chick mortality rate via starvation (Hatchwell et al. 2004).
Carer number	The number of adult birds that provisioned chicks during a given provisioning watch, modelled as an ordinal factor: 5>4>3>2.
Watch duration	The length of time in minutes between the first and last feeding visit during a provisioning watch, included as an interaction term with “Data type” to account for potential artificial coordination caused by setup disturbance, which would disproportionately affect shorter watches. Watch duration was modelled as a continuous numerical value.
AMax (or SMax)	To account for the upper limit of the number of alternated or synchronized visits we used the variables “AMax” and “SMax” which are the theoretical maximum percentage of provisioning visits that can be alternated or synchronized during a given watch, modelled as a continuous numerical value (mean: 87.5; range: 19.4–98.6). If one carer performed more than half the total visits then some visits exist which cannot be alternated or synchronized due to insufficient visits to alternate or synchronize with, this applies to all values of carer number. This term is used as a proxy for feed rate asymmetry, that is, the difference between provisioning rates of male and females, used by Bebbington and Hatchwell (2016), but is also applicable for cooperative nests. AMax and SMax must always be the same value for a given watch because synchrony was defined as an alternated feed occurring within 2 min of the previous feed. This time restriction can always theoretically be met, so the only remaining restriction on synchrony is whether a feed can be alternated.
Brood age	The number of days since recorded hatching (day 0) of chicks at a nest, modelled as integer values (range: 6–16).
Hatch date	The number of days after 1st of March each year, modelled as integer values (median: 3 May; range: 16 April – 6 June).
Time of day	The number of hours since the beginning of the day on which the provisioning watch was performed, modelled as a numerical value. Included because previous studies on the same population found higher provisioning activity immediately after sunrise (MacColl and Hatchwell 2002; Hatchwell et al. 2004).
Sex	Binary factor representing either male or female helpers.
Carer status	Factor designating the relationship each carer had to the brood with three categories: breeding female, breeding male, and helper. For our final analysis we did not sub-divide helpers by sex because our investigation on helper coordination (Alt-K, Sync-K, see results) found no effect of sex on helpers’ alternation or synchrony (Tables 6 and 7).
Kinship with father	Factor representing the pairwise relatedness of helpers to breeding male. Modelled as factors with three levels representing first-order kin (r = 0.50), second-order kin (r = 0.25) and unrelated individuals (r = 0.00). These levels were selected in accordance with results from previous studies demonstrating kinship of helpers with breeders (Leedale et al. 2018, 2020).
Kinship with mother	Factor representing pairwise relatedness of helper to breedingfemale.
Non-model terms
Percentage of AMax (or Percentage of SMax)	The overall percentage of visits which could be alternated or synchronized which were alternated or synchronized. These terms was used as the dependent variables in Figures 2a,b and 3a,b, being the best single metrics for visualizing active coordination as they factor in both total feed rate and what proportion of visits could be alternated or synchronized.

Individual coordination models (Alt-I and Sync-I)

To investigate the effect of carer status on alternation and synchrony performed by a given carer, we built two Poisson-distributed GLMMs named “Alt-I” and “Sync-I,” respectively. The response variables to these models were the number of alternated visits (individual) and synchronized visits (individual), respectively. These models were built with the following random effects: “Year,” “Nest ID,” “Watch ID,” “Carer ID,” and “Row reference” (Table 1). The fixed effects tested were as follows: “Data type,” “Carer status,” “Provisioning rate (individual),” “Carer number,” “Watch duration,” “Brood size,” “Time of day,” “Brood age,” “Hatch date,” and “Amax” or “SMax” (Table 1). In this analysis, the focus was on the interaction of “Data type” with “Carer status” because this term represents the disparity in active coordination between carers of different breeding status.

Repeatability models (Alt-R and Sync-R)

To investigate the repeatability of active alternation and synchrony within-nests and within-individuals, we constructed two Gaussian-distributed GLMMs named “Alt-R” and “Sync-R,” respectively. In these models, response variables were the number of actively alternated (individual) and actively synchronized (individual) visits by an individual during a watch, respectively (active alternation range: −3 to 6; active synchrony range: −7 to 9). We used these metrics because repeatability analyses required active coordination to be the response variable, rather than using interaction terms with “Data type” as in our other models. To control the effect of confounding factors on active coordination, we included all fixed effects previously found to significantly influence either individual alternation or synchrony (Alt-I, Sync-I) and, using the rptR function, ran models with 1000 bootstrapped simulations and 1000 permutations. We investigated both within-nest repeatability (“Nest ID”) and within-individual repeatability (“Carer ID”) in the same models. Additionally, we included “Year” as a random effect to account for between-year variation. As active coordination was the response variable and a Gaussian error distribution was used, “Watch ID” and the “Row reference” random effects were not required for these models. We present our repeatability results as values of R and extracted 2.5% and 97.5% confidence intervals (CIs) in addition to P-values.

In our dataset, many individuals were observed provisioning at only one nest, potentially confounding repeatability of an individual’s behavior with the potential effect of common nest factors. Therefore, we ran the repeatability analysis on a subset of data, restricted to carers observed provisioning at two or more nests (Supplementary Table S2). Results from these models were qualitatively the same as those for the full dataset for both within-nest and within-individual repeatability for both alternation and synchrony models.

Kinship models (Alt-K and Sync-K)

To investigate the effect of kinship to the breeding pair on alternation and synchrony performed by helpers, we constructed two Poisson-distributed GLMMs named “Alt-K” and “Sync-K,” respectively. Just as with “Alt-I” and “Sync-I,” the response variables to these models were the number of alternated visits (individual) and synchronized visits (individual) performed by an individual during a watch, respectively, however analysis was restricted to helpers whose pedigree kinship with breeders was known. These models were built with the same random and fixed effects as “Alt-I” and “Sync-I” but with the addition of three fixed effects: “Sex,” “Kinship with father,” and “Kinship with mother” (Table 1). We focused our analysis on the interactions of “Data type” with our kinship terms as these represent the relationship between the level of active coordination and relatedness.

RESULTS

Carer number

To test the hypothesis that carers exhibited behaviors resulting in alternated visits, model “Alt-C” compared observed alternation with that expected by chance from null model randomization. We found that observed alternation was indeed significantly higher than expected by chance, as indicated by the significance of the data type term (P < 0.001, Table 2). To investigate the effect of other terms on active alternation, we measured their effect on the difference between observed and expected data, that is, their interaction with data type. Carer number had a positive effect on both expected and observed alternation (Table 2), but the interaction term with data type was significant (P = 0.024; Table 2, Figure 2a), indicating that the difference between them, that is, active alternation, declined as carer number increased. The degree of active alternation was not significantly related to time of day, watch duration, brood size, or provisioning rate (Table 2, Figure 2b).

Table 2

Estimates and P-values for fixed effect terms on number of alternated visits per provisioning watch from generalized linear mixed effect model; N = 1590 from 795 watches at 250 nests, significant values in bold. “Obs’” is shorthand for data observed directly from field, as opposed to expected data generated from null model

Parameter	df	Estimate ± SE	F-value	P-value
Intercept	1	2.546 ± 0.014		<0.001
Data type (relative to expected)	1	Obs: 0.099 ± 0.017	38.11	<0.001
Provisioning rate	1	0.303 ± 0.010	3452.27	<0.001
Brood size	1	0.034 ± 0.011	22.88	<0.001
Carer number (relative to 2 carers)	3	3: 0.135 ± 0.024 4: 0.210 ± 0.029 5: 0.196 ± 0.050	152.88	<0.001
Watch duration	1	0.195 ± 0.008	1247.49	<0.001
Amax	1	0.131 ± 0.009	209.17	<0.001
Brood age	1	0.000 ± 0.007	0.01	0.973
Hatch date	1	0.012 ± 0.008	2.21	0.126
Time of day	1	−0.016 ± 0.010	6.76	0.055
Data type * provisioning rate	1	0.017 ± 0.012	0.43	0.133
Data type * brood size	1	0.004 ± 0.014	0.02	0.798
Data type * carer number	3	Obs, 3: −0.069 ± 0.031 Obs, 4: −0.094 ± 0.037 Obs, 5: −0.096 ± 0.067	2.68	0.024
Data type * watch duration	1	0.015 ± 0.010	2.11	0.134
Data type * time of day	1	0.005 ± 0.013	0.17	0.683

Figure 2

Plots of observed (orange) and expected (purple) percentage of maximum alternation achieved by carers versus: (a) the number of carers observed provisioning each nest during a watch and (b) the total provisioning rate by all carers at a nest during a given watch. Predicted relationships (±95% CI) are fitted from GLMMs, see Table 2. Total N = 795 watches on 250 nests for each data type.

To test the hypothesis that carers actively synchronized provisioning visits, we used model “Sync-C” to compare observed and expected synchrony. Just as for alternation, observed synchrony was greater than expected by chance, the data type term being significant (P < 0.001; Table 3). However, in contrast to our results for alternation, there was no significant interaction between data type and carer number (Table 3, Figure 3a), indicating that the level of active synchrony was similar at all group sizes. Investigation of the interaction between data type and other predictors of synchrony showed that provisioning rate was the only factor to influence the degree of active synchrony (P < 0.001; Table 3, Figure 3b), the difference between observed and expected synchrony declining with increasing provisioning rate. This result was expected because as provisioning rates increase, the probability that two birds feed within a 2-min period, even by passive process, inevitably increases. Neither brood size, time of day, nor watch duration was a significant predictor of the level of active synchrony (Table 3).

Table 3

Output values for fixed effect terms on number of synchronized visits per provisioning watch from generalized linear mixed effect model, N = 1590 from 795 watches at 250 nests, significant values in bold

Parameter	df	Estimate ± SE	F-value	P-value
Intercept	1	1.798 ± 0.055		<0.001
Data type (relative to expected)	1	Obs: 0.245 ± 0.024	100.94	<0.001
Provisioning rate	1	0.472 ± 0.013	3240.41	<0.001
Brood size	1	0.064 ± 0.016	36.85	<0.001
Carer number (relative to 2 carers)	3	3: 0.113 ± 0.034 4: 0.221 ± 0.062 5: 0.182 ± 0.089	44.41	<0.001
Watch duration	1	0.200 ± 0.012	667.33	<0.001
Smax	1	0.147 ± 0.013	127.52	<0.001
Brood age	1	0.014 ± 0.010	1.74	0.186
Hatch date	1	0.013 ± 0.012	1.08	0.231
Time of day	1	−0.023 ± 0.014	1.3	0.485
Data type * provisioning rate	1	−0.050 ± 0.015	19.79	<0.001
Data type * brood size	1	<0.001 ± 0.019	<0.01	0.969
Data type * carer number	3	Obs, 3: −0.030 ± 0.041 Obs, 4: −0.061 ± 0.048 Obs, 5: −0.057 ± 0.085	0.46	0.574
Data type * watch duration	1	0.017 ± 0.014	1.19	0.186
Data type * time of day	1	0.024 ± 0.018	2.11	0.146

Figure 3

Plots of observed (orange) and expected (purple) percentage of maximum synchrony achieved by carers versus: (a) the number of carers observed provisioning each nest during a watch and (b) the total provisioning rate by all carers at a nest during a given watch. Predicted relationships (±95% CI) are fitted from GLMMs, see Table 3. Total N = 795 watches on 250 nests for each data type.

Carer status

To investigate variation in alternation behavior by birds of different status (breeding male, breeding female, and helper), we used model “Alt-I.” Breeding females had higher overall levels of alternation than other categories of carer (P = 0.037; Table 4), but carer status did not influence the extent of active alternation because the interaction term with data type was non-significant (P = 0.975; Table 4, Figure 5a).

Table 4

Estimates and P-values for fixed effect terms on number of alternated visits per individual carer during a watch from generalized linear mixed effect model; N = 1997 samples from 487 unique carers from 795 watches at 250 nests, significant values in bold

Parameter	df	Estimate ± SE	F-value	P-value
Intercept	1	1.722 ± 0.019		<0.001
Data type (relative to expected)	1	Obs: 0.056 ± 0.024	18.07	<0.001
Carer status (relative to breeding male)	2	F: 0.030 ± 0.020 H: −0.012 ± 0.028	49.90	0.037
Individual provisioning rate	1	0.303 ± 0.010	2576.93	<0.001
Brood size	1	0.045 ± 0.011	55.89	<0.001
Carer number (relative to 2 carers)	3	3: 0.008 ± 0.026 4: 0.027 ± 0.032 5: −0.093 ± 0.053	31.05	0.018
Watch duration	1	0.181 ± 0.008	1089.08	<0.001
Amax	1	0.152 ± 0.009	280.17	<0.001
Brood age	1	0.024 ± 0.007	19.37	<0.001
Hatch date	1	0.015 ± 0.008	3.44	0.079
Time of day	1	−0.023 ± 0.010	15.61	0.003
Data type * carer status	3	Obs, F: 0.002 ± 0.028 Obs, H: 0.009 ± 0.039	0.92	0.975
Data type * individual provisioning rate	1	0.018 ± 0.012	4.38	0.137
Data type * brood size	1	0.007 ± 0.014	0.01	0.609
Data type * carer number	3	Obs, 3: −0.056 ± 0.033 Obs, 4: −0.078 ± 0.041 Obs, 5: −0.060 ± 0.070	1.28	0.207
Data type * watch duration	1	0.021 ± 0.010	3.93	0.044
Data type * time of day	1	0.005 ± 0.013	0.12	0.726

In contrast, in model “Sync-I,” the extent of active synchrony was influenced significantly by carer status, as indicated by the interaction term with data type (P = 0.024; Table 5, Figure 6a), with helpers performing the most active synchrony followed by breeding males then breeding females.

Table 5

Estimates and P-values for fixed effect terms on number of synchronized visits per provisioning individual carer during a watch from generalized linear mixed effect model; N = 1997 samples from 487 unique carers from 795 watches at 250 nests, significant values in bold

Parameter	df	Estimate ± SE	F-value	P-value
Intercept	1	0.905 ± 0.033		<0.001
Data type (relative to expected)	1	Obs: 0.301 ± 0.033	59.70	<0.001
Carer status (relative to breeding male)	2	F: 0.059 ± 0.033 H: 0.035 ± 0.042	8.89	0.210
Individual provisioning rate	1	0.444 ± 0.014	2132.76	<0.001
Brood size	1	0.089 ± 0.017	72.06	<0.001
Carer number (relative to 2 carers)	3	3: 0.097 ± 0.039 4: 0.236 ± 0.048 5: 0.164 ± 0.075	37.97	0.019
Watch duration	1	0.180 ± 0.012	472.16	<0.001
Smax	1	0.177 ± 0.015	162.39	<0.001
Brood age	1	0.071 ± 0.010	60.29	<0.001
Hatch date	1	0.011 ± 0.015	0.36	0.464
Time of day	1	−0.052 ± 0.014	9.46	0.005
Data type * carer status	3	Obs, F: −0.079 ± 0.037 Obs, H: 0.042 ± 0.051	1.56	0.024
Data type * individual provisioning rate	1	−0.084 ± 0.016	21.42	<0.001
Data type * brood size	1	0.010 ± 0.019	0.11	0.602
Data type * carer number	3	Obs, 3: −0.133 ± 0.044 Obs, 4: −0.234 ± 0.053 Obs, 5: −0.252 ± 0.089	7.41	<0.001
Data type * watch duration	1	0.023 ± 0.014	1.74	0.092
Data type * time of day	1	0.004 ± 0.017	5.80	0.016

The extent of individual active synchrony was also influenced significantly by carer number (P < 0.001; Table 5), a relationship which was not observed in the collective synchrony model “Sync-C” (P = 0.574; Table 2). We suspected that this trend may be due to covariances between carer number, individual, and total provisioning rate, coupled with load-lightening and the provisioning rate dependence of the null model (P < 0.001; Table 2, Figure 3b). Refitting the model with total provisioning rate and appropriate interaction terms revealed that the effect of carer number on active synchrony was contained within multiple significant three-way interaction terms which are probably a consequence of load-lightening behavior and the rate dependence of the synchrony null model (Table 2, Figure 3b). Importantly, however, the results for the effect of carer status on active synchrony remained qualitatively the same (Supplementary Figure S2).

Repeatability of coordination

Using model “Alt-R,” we assessed whether active alternation was consistent within-individuals and/or within-groups of carers working together at a nest. Active alternation of carers was significantly repeatable within-nests (R = 0.145, CI (2.5–97.5%) = 0.010–0.186, P < 0.001), but not within-individuals (R = 0.000, CI (2.5–97.5%) = 0.000–0.031, P = 0.500, Figure 4a), indicating that the degree of alternation was a property of social or nest-specific factors. In contrast, model “Sync-R” showed that active synchrony of carers was significantly repeatable within-individuals (R = 0.183, CI (2.5–97.5%) = 0.130–0.228, P < 0.001), but not within-nests (R = 0.000, CI (2.5–97.5%) = 0.000–0.009, P = 1.00, Figure 4b), indicating that the level of synchrony was a property of individual identity rather than the nest or social environment.

Figure 4

Estimates (R) of the within-nest repeatability (Nest ID) and within-individual repeatability (Carer ID) for the absolute number of actively (a) alternated and (b) synchronized visits. Estimates calculated from linear mixed effects models (2.5–97.5% CI).

Helper kinship

We found no significant effects of helper kinship to the helped breeders on any measures of coordination. Model “Alt-K” investigated variation in alternation behavior between helpers of varying kinship, but neither the overall level of alternation by helpers nor their degree of active alternation was influenced significantly by their kinship with either the breeding male (Table 6, Figure 5b) or breeding female (Table 6, Figure 5c). Similarly, model “Sync-K” showed that neither the overall level of synchrony exhibited by helpers, nor the extent of active synchrony was influenced by kinship with either the breeding male (Table 7, Figure 6b) or breeding female (Table 7, Figure 6c).

Figure 5

Boxplots of observed (orange) and expected (purple) percentage of visits alternated by a given individual during a watch versus: (a) the status of a carer; (b) the kinship between a helper and the breeding male; and (c) the kinship between a helper and the breeding female. Total N = 408 samples from 130 unique helpers from 273 watches at 95 nests for each data type.

Table 6

Estimates and P-values for fixed effect terms on number of alternated visits per individual helper during a watch from generalized linear mixed effect model; N = 408 samples from 130 unique helpers from 273 watches at 95 nests, significant values in bold

Parameter	df	Estimate ± SE	F-value	P-value
Intercept	1	1.486 ± 0.159		<0.001
Data type (relative to expected)	1	Obs: 0.038 ± 0.190	0.97	0.782
Kinship with breeding male (relative to r = 0)	2	0.25: 0.059 ± 0.088 0.5: −0.022 ± 0.057	1.47	0.837
Kinship with breeding female (relative to r = 0)	2	0.25: −0.079 ± 0.141 0.5: 0.012 ± 0.073	6.88	0.837
Sex (relative to female)		M: −0.024 ± 0.057	0.04	0.728
Individual provisioning rate	1	0.352 ± 0.022	786.89	<0.001
Brood size	1	0.024 ± 0.026	0.06	0.139
Carer number	3	3: 0.166 ± 0.147 4: 0.206 ± 0.155 5: 0.069 ± 0.164	8.63	0.088
Watch duration	1	0.228 ± 0.020	390.19	<0.001
Amax	1	0.070 ± 0.030	5.84	0.019
Brood age	1	0.019 ± 0.016	1.47	0.238
Hatch date	1	0.033 ± 0.018	3.44	0.078
Time of day	1	−0.022 ± 0.023	3.76	0.175
Data type * kinship with breeding male	2	Obs, 0.25: −0.066 ± 0.123 Obs, 0.5: 0.017 ± 0.075	0.22	0.818
Data type * kinship with breeding female	2	Obs, 0.25: 0.039 ± 0.196 Obs, 0.5: −0.017 ± 0.102	0.04	0.965
Data type * sex	1	Obs, M: 0.020 ± 0.080	0.16	0.804
Data type * individual provisioning rate	1	0.025 ± 0.028	1.30	0.377
Data type * brood size	1	0.008 ± 0.035	0.01	0.828
Data type * carer number	3	Obs, 3: −0.065 ± 0.172 Obs, 4: −0.077 ± 0.178 Obs, 5: −0.048 ± 0.191	0.07	0.972
Data type * watch duration	1	0.023 ± 0.025	0.92	0.348
Data type * time of day	1	−0.001 ± 0.032	0.00	0.984

Table 7

Estimates and P-values for fixed effect terms on number of synchronized visits per individual helper during a watch from generalized linear mixed effect model; N = 408 samples from 130 unique helpers from 273 watches at 95 nests, significant values in bold

Parameter	df	Estimate ± SE	F-value	P-value
Intercept	1	0.808 ± 0.229		<0.001
Data type (relative to expected)	1	Obs: 0.648 ± 0.245	25.37	<0.001
Kinship with breeding male (relative to r = 0)	2	0.25: 0.084 ± 0.128 0.5: 0.029 ± 0.084	1.60	0.683
Kinship with breeding female (relative to r = 0)	2	0.25: 0.208 ± 0.192 0.5: 0.067 ± 0.106	1.33	0.731
Sex (relative to female)		M: −0.036 ± 0.081	3.01	0.562
Individual provisioning rate	1	0.432 ± 0.030	558.01	<0.001
Brood size	1	0.064 ± 0.039	4.30	0.034
Carer number (relative to 2 carers)	3	3: 0.132 ± 0.211 4: 0.285 ± 0.221 5: 0.116 ± 0.237	5.17	0.047
Watch duration	1	0.187 ± 0.029	153.70	<0.001
Smax	1	0.130 ± 0.045	8.20	0.004
Brood age	1	0.047 ± 0.022	5.44	0.029
Hatch date	1	0.033 ± 0.030	1.26	0.276
Time of day	1	−0.073 ± 0.032	7.16	0.026
Data type * kinship with breeding male	2	Obs, 0.25: −0.016 ± 0.158 Obs, 0.5: 0.016 ± 0.097	0.01	0.982
Data type * kinship with breeding female	2	Obs, 0.25: −0.180 ± 0.246 Obs, 0.5: −0.122 ± 0.132	0.43	0.517
Data type * sex	1	Obs, M: 0.002 ± 0.100	0.50	0.982
Data type * individual provisioning rate	1	−0.050 ± 0.035	0.27	0.158
Data type * brood size	1	−0.001 ± 0.045	0.79	0.981
Data type * carer number	3	Obs, 3: −0.409 ± 0.222 Obs, 4: −0.487 ± 0.229 Obs, 5: −0.470 ± 0.244	1.73	0.190
Data type * watch duration	1	0.063 ± 0.031	3.52	0.043
Data type * time of day	1	0.039 ± 0.041	0.92	0.337

DISCUSSION

We found strong evidence for active coordination of care, with both alternation and synchrony being observed more than expected by chance. Active synchrony was detected across the full range of two to five carers (Figure 3a), whereas active alternation was detected only in biparental nests and cooperative nests with one helper (Figure 2a). Additionally, although breeding males, females, and helpers did not differ in their degree of active alternation (Figure 4a), helpers exhibited more active synchrony than breeders, and male breeders showed more active synchrony than female breeders (Figure 5a). We also found that the level of active alternation was linked to nest identity (Figure 4a), whereas active synchrony was linked to individual carer identity (Figure 4b), suggesting that alternation is a plastic behavior in response to social environment, and synchrony is influenced by both an individual’s identity and current carer status. Finally, contrary to our expectations, the degree of helper coordination was unaffected by their kinship with either breeding bird (Figures 5b,c and 6b,c).

Figure 6

Boxplots of observed (orange) and expected (purple) percentage of visits synchronized by a given individual during a watch versus: (a) the status of a carer; (b) the kinship between a helper and the breeding male; and (c) the kinship between a helper and the breeding female. Total N = 408 samples from 130 unique helpers from 273 watches at 95 nests for each data type.

The null hypothesis of a study seeking to quantify coordination of care is not that there is no apparent coordination, but rather that the observed level of coordination may be wholly explained by passive processes that affect provisioning, such as weather, predation threat, and resource distributions (Schlicht et al. 2016; Ihle et al. 2019a). We used the most conservative randomization approach (Ihle et al. 2019a), conserving individual refractory periods and hence controlling for much of the coordination that may be explained by passive processes. Our methods of data collection and analysis also accounted for potential observer disturbance effects that could enhance apparent coordination. On the other hand, the randomization process effectively decouples the refractory periods of carers at the same nest, so factors such as weather and predation threat that may impact all carers at the same time remain difficult to control for statistically. However, it can also be argued that highly conservative null models which re-order observed data retain a degree of active coordination that is reflected in refractory periods, thereby underestimating the true level of active coordination. Therefore, we conclude that our results support the case for active coordination of care in long-tailed tits.

The hypothesized function of alternation is that it facilitates conflict resolution between carers because conditional cooperation prevents exploitation by ensuring that carers match changes in one another’s provisioning rates (Johnstone et al. 2014). This enables carers to increase their investment to more closely match the brood’s optimum care level (Trivers 1974) without causing other carers to slacken their effort to increase their individual fitness pay-off. Our finding that active alternation declined as the number of carers increased may indicate a reduced need to monitor the investment of others when care is plentiful, especially as individual carers reduce their own costs by load-lightening when they have helpers (Hatchwell and Russell 1996; Meade et al. 2010; Adams et al. 2015). This result contrasts with findings from chestnut-crowned babblers Pomatostomus ruficeps (Savage et al. 2017), where active alternation was observed across the full range of carer numbers (2–6), using the same null model approach. This disparity may be due to differences in the ecology or social system of chestnut-crowned babblers and long-tailed tits. Babblers must gather food far from the nest despite not being proficient long-distance fliers, thus incurring substantial provisioning costs (Browning et al. 2012). Therefore, a strict and efficient allocation of effort between carers, with close monitoring, may remain important in this species even in large cooperative groups. In contrast, long-tailed tits are thought to suffer relatively modest costs of parental care (Meade and Hatchwell 2010; Hatchwell et al. 2014), so individual effort may be monitored less closely, resulting in a decline in active alternation with carer number.

Alternatively, the decline in active alternation with carer number may be a consequence of the null model failing to detect active alternation in large groups. In our study, expected alternation approached 90% in 4–5 carer nests, which contrasts with expected synchrony of just 50% or so in larger groups; thus, the scope for detection of active synchrony is greater than it is for active alternation. However, it is unlikely that detectability alone caused our result because Savage et al.’s (2017) study of chestnut-crowned babblers used the same null model approach across a greater range of group sizes (2–6), with expected alternation of >80% at large group sizes, and yet they did not observe the same trend.

We observed no significant difference in active alternation by carers of different status and subsequent analysis revealed that the level of active alternation was highly repeatable within-nests rather than within-individual carers. These results suggest that that if alternation is adaptive, it is performed by all carers at the nest to their collective benefit, rather than by certain individuals. However, we cannot disentangle whether this is a function of common nest factors or common social environment (Ihle et al. 2019a). For example, some nests may experience regular disturbance by predators that temporarily prevents feeding, causing the feeding cycles of carers to align, thus increasing alternation. Our finding that active alternation was unaffected by carer status could be explained by the interests of breeders and helpers being closely aligned. Long-tailed tit helpers gain only kin-selected benefits from their helping behavior (Meade and Hatchwell 2010; Hatchwell et al. 2014), and rates of extra-pair paternity and intraspecific brood parasitism are low (Hatchwell et al. 2000) so all carers have a shared interest in raising a related brood. In species where the dynamics of conflict are different, the extent of alternation may vary between carers of different status while still ultimately providing the adaptive function of conflict resolution (Johnstone et al. 2014; Johnstone and Savage 2019). This may explain why breeders alternate more to ensure the contribution of helpers in chestnut-crowned babblers (Savage et al. 2017) and our contrary finding does not necessarily invalidate conflict resolution as a function of alternation in long-tailed tits.

One proposed function of synchrony is that it facilitates accurate alternation via monitoring of other carers (Mariette and Griffith 2012, 2015; Bebbington and Hatchwell 2016), but there are other adaptive hypotheses for synchrony that do not require alternation per se. Synchrony may decrease parental activity at the nest, thereby reducing its conspicuousness and exposure to predators (Raihani et al. 2010; Mariette and Griffith 2012, 2015; Leniowski and Wegrzyn 2018; Khwaja et al. 2019). However, our finding that active synchrony was broadly consistent across group sizes does not support this hypothesis, because in larger groups, where the risk of exposing the nest to a predator is greater, active synchrony should increase. Alternatively, synchrony may ensure an even distribution of food between chicks by preventing monopolization (Shen et al. 2010; Mariette and Griffith 2012, 2015). However, contrary to our results, this hypothesis predicts that synchrony would decrease with group size as the increased rate of food delivery reduced the risk of monopolization. A detailed investigation of the consequences of synchrony for parental activity at the nest, the probability of predation, nestling growth, and survival is beyond the scope of the current article.

Helpers synchronized their nest visits with other carers more than breeders did. One explanation for this result is that helpers synchronize visits to signal their effort to other carers to increase their “prestige” (Zahavi 1977a, 1977b). Most studies have refuted this hypothesis (e.g., McDonald et al. 2008a, 2008b; Nomano et al. 2015; Raihani et al. 2010), but there is some limited empirical support (Doutrelant and Covas 2007; Trapote et al. 2021). For example, in carrion crows Corvus corone, subordinate female helpers overlapped their feeding visits with breeders more than either male helpers or breeders did. This was interpreted as a “pay-to-stay” system where female helpers, which are typically unrelated to breeders (unlike male helpers), signal their effort to remain within the group until they achieve breeding status in their own group (Trapote et al. 2021). Our results appear to support this hypothesis, but we think it is an unlikely explanation for the relatively high synchrony exhibited by long-tailed tit helpers. Helpers are expected to gain direct fitness via signaling when helping is payment of rent for living on the breeders’ territory (Gaston 1978; Kokko et al. 2002; Trapote et al. 2021), or if it increases an individual’s social status or perceived quality among other carers (Zahavi 1977a, 1977b; Lotem et al. 1999). However, studies have yet to detect any direct fitness benefits for helpers from their altruistic care in long-tailed tits (Hatchwell 2016). Therefore, unlike carrion crows, group membership, breeding opportunities and future direct fitness are not determined by helping behavior (Meade and Hatchwell 2010; Napper and Hatchwell 2016), so there seems to be no advantage for helpers from signaling their quality to other carers.

We suggest instead that variation in synchrony between carers of different status may be a consequence of group foraging. Collective foraging behavior may explain synchrony in zebra finches Taeniopygia guttata (Mariette and Griffith 2012, 2015), where it is thought to reduce predation risk for carers. This hypothesis would not necessarily predict that carers of different status should differ consistently in their degree of synchrony, nor that synchrony would be highly repeatable within-individuals, unless also coupled with a defined feeding order. If helpers tend to follow breeders in their visits to the nest, the way in which we measured synchrony means that they would also tend to have a relatively high synchrony score. Apparent following behavior could result from helpers shadowing breeders or from breeders delaying feeds until helpers are present. In the redirected helping system of long-tailed tits, helpers are likely to be less familiar with the brood and local area than breeders are, so the idea that helpers shadow foraging breeders is plausible. Furthermore, the suggestion that individuals may adopt specific roles, that is, as leader or follower, when foraging or when visiting nests may also explain why synchrony is individually repeatable. However, more detailed observations of the behavior of individuals as they approach the nest and the sequence in which they do so are needed to investigate these possibilities.

Active coordination by helpers was not influenced significantly by their kinship with the breeders they assisted. This result was unexpected because helper decisions in long-tailed tits, both in who to help and how much to help, are a function of their relatedness to the breeding pair (Russell and Hatchwell 2001; Nam et al. 2010; Leedale et al. 2018, 2020). Additionally, if alternation functions to resolve conflict between carers, we might expect greater conflict in less related groups, so we anticipated some effect of kinship on coordination. Comparisons with the repeatability results are potentially instructive. The kinship of a helper to breeders is a function of the group, that is, it is the dyadic relatedness between a helper and a specific male or female breeder, rather than a property of the helper per se. We suggested above that the repeatability of alternation within-groups could be a function of the social environment (e.g., group composition), which could include kinship. However, the absence of a kinship effect, suggests either that some ecological rather than social factor specific to a nest or group drives the repeatability of alternation, or that a social factor other than kinship (e.g., group familiarity) influences alternation. In contrast, repeatability in synchrony was at the level of individuals rather than groups. Therefore, it is perhaps unsurprising that synchrony of helpers was not predicted by their kinship with breeders, given that this is a property of two or more individuals rather than an individual helper. To our knowledge, this is the first study to explicitly test whether kinship influences coordination, and it would be interesting to explore this question more widely, and especially in species where the interests of helpers and breeders are not so closely aligned.

Several explanations for active alternation and synchrony have linked the two phenomena, with synchrony proposed as an adaptation for ensuring accurate monitoring of other carers, thus enabling alternation (Mariette and Griffith 2012, 2015; Bebbington and Hatchwell 2016). While studies have often found a correlation between alternation and synchrony, several of our findings suggest that alternation and synchrony may, in fact, fulfill separate functions. First, active alternation declined with increasing carer number, while active synchrony did not. Second, active synchrony varied between carers of different status, while active alternation did not. Finally, active alternation was repeatable between watches at the same nest, while active synchrony was repeatable between watches of the same individual. The independence of alternation and synchrony is also supported by a study of blue tits Cyanistes caeruleus which demonstrated that synchrony, but not alternation varied between different habitats (Lejeune et al. 2019). This is compatible with Johnstone et al.’s (2014) theory of conflict resolution for alternation and our suggestions of shadowing for synchrony.

Studies of coordination in parental care are still in their infancy, and much work remains to be done to fully understand its occurrence, function, and the causes of interspecific variation. Careful analysis of provisioning visits is essential to generate appropriate null models against which observed schedules of visits can be compared. In this study, adopting a conservative approach, we have shown that coordination is, to some extent, a function of group size in the cooperative breeding system of long-tailed tits. We have also shown that some measures of coordination vary with social role within-groups, but not with the kinship of helpers. In addition, we highlight the need for investigation of the proximate mechanisms by which individuals coordinate care, such as delaying feeding or shadowing others, as well as a need for experimental studies that can isolate and test social and environmental influences that are hard to take account of in observational studies. Finally, despite coordination of care being quite widely demonstrated in nature, the function of these behaviors remains poorly understood.

We are grateful to all field researchers who have contributed to the long-tailed tit project, and thank Sheffield City Council, Yorkshire Water, Hallamshire Golf Club, and private landowners of the Rivelin Valley for access to their land. Molecular analyses were conducted at the Natural Environment Research Council (NERC) Biomolecular Analysis Facility at the University of Sheffield, with support from Terry Burke and Deborah Dawson. We also thank James Savage for discussions. This work was supported by the Natural Environment Research Council (NE/S00713X/1 and NE/R001669/1).

Conflict of Interest: The authors declare no conflict of interest.

Data Availability: Analyses reported in this article can be reproduced using the data provided by Halliwell et al. (2022).