In cooperatively breeding species, group members may derive multiple benefits from helping to raise other individuals’ offspring, yet not all individuals do so. In this study, we tested several hypotheses to explain why group members feed offspring of breeding placid greenbuls (Phyllastrephus placidus). In accordance with the kin selection hypothesis, all helpers were first-order kin of the breeding female and the presence of helpers was associated with increased survival of the breeding pair. However, the propensity to help varied widely among group members, as 46% of group members related to the breeding female did not feed nestlings. Sex, relatedness to the breeding male, and group size did not explain additional variation in helping propensity or effort, however, younger helpers fed offspring more often than older ones. Overall, our results show mixed support for predictions of the group augmentation hypothesis, while predictions of the pay-to-stay hypothesis and skills hypothesis were not supported. We suggest that costs associated with providing food to nestlings, and benefits of other types of helping behavior (e.g. anti-predator behavior), may jointly explain why members of breeding groups often refrain from helping at the nest.

Study system and data collection

Data were collected during the breeding seasons of 2012-2018 in the Dabida forest archipelago of the Taita Hills (SE Kenya; 30°25'S, 38°20'E). The landscape consists of a heterogeneous mixture of indigenous cloud forest remnants (<1 ha to 120 ha), exotic plantations and small-scale subsistence agriculture. Eight of these forest remnants contain subpopulations of the placid greenbul, a long-lived insectivorous passerine from East Africa’s moist forests that occupies pair- or group territories year-round (DVL & BA personal observations). Their breeding season coincides with the onset of the short rainy season in November and lasts until March. Typically, breeding females lay and incubate two eggs that hatch synchronously, and pairs generally re-nest after breeding failure and (occasionally) also after successful breeding. Approximately 50% of greenbul offspring in our study area consist of extra-pair young sired by neighboring territory owners (Cousseau, Van de Loock, et al. 2020). Nest failure is mostly due to predation (Spanhove et al. 2014) and larger breeding groups are associated with lower rates of nest predation and higher post-fledgling survival (Van de Loock et al. 2017; Van de Loock 2019). Due to continued ringing and nest monitoring effort since 1996 and 2007, respectively, ca. 75% of the greenbul population is color-banded at any time (based on the ratio of color-banded individuals vs. total traps during mist-netting).

In our study area, the vast majority of subordinate group members are offspring from previous breeding season(s) that delay their natal dispersal, although non-natal individuals may occasionally join breeding groups (Cousseau, Hammers, et al. 2020). Both subordinate males and females show delayed dispersal, but males delay on average longer and eventually start breeding closer to their natal territory than females (Cousseau, Hammers, et al. 2020). While males disperse earlier when born in smaller forest patches compared to larger ones (Cousseau, Hammers, et al. 2020), this is not reflected in spatial variation in the likelihood that a breeding group contains subordinates, neither in the size of these groups (Van de Loock 2019).

Placid greenbul breeding behavior was monitored in all eight known subpopulations of the Dabida forests. Upon detection, each nest was visited every 3 to 5 days until all nestlings had fledged or the nesting attempt failed. At an age of ca. nine days, nestlings were individually color-banded, measured and a blood sample was taken. We recorded group size, identified breeding and subordinate group members, and quantified their helping behavior and individual nestling provisioning rates through a combination of focal observations, targeted mist-netting, and nest video recordings.

We conducted focal observations during incubation by hiding less than 10m from the nest (2012-15 only) as well as opportunistically at each nest visit (all years). We erected mist nets around the nest when nestlings were ca. 5 days old (range 3 – 8 days) or after nest depredation. Upon trapping, birds were measured, a blood and/or a feather sample was taken and the breeding status (cloacal swelling or brood patch) determined (see below). Unringed individuals were metal- and color-banded. Between 2012 and 2015, focal observations and mist-netting were conducted using playback of greenbul distress calls for a maximum duration of 10 min. As group members respond to these playbacks by approach and displays, presumably to distract and lead away potential predators, this method proved to be a rapid and efficient way to assess group membership and trap individuals. Playback was not used when a predator was present. When nestlings were ca. 8 days old (range 6 – 10 days), we video-recorded food provisioning for 5-6 hours continuously between 7 a.m. and 2 p.m. using a HD camera (Sony Corp.) installed on a tripod about 1.5 m from the nest (one video per nest; on average 25 videos taken per year). Besides helping behavior, video recordings allowed us to extract hourly provisioning rates (number of feeds per hour per nestling) of breeders and helpers. We only extracted provisioning rates when we could identify the visiting individuals in at least 70% of all visits over the total recording period.

Each breeding group consisted of the territorial breeding pair (henceforward called ‘breeders’) and all subordinates (if any) observed at a particular nest. Breeders were identified based on cloacal swellings (males), the presence of a brood patch or observed incubation (females). All other individuals present at a nest were assigned as subordinates. A subordinate was classified as a ‘helper’ if it was observed feeding nestlings at least once during the nest video recording. It follows that subordinates that were never observed feeding nestling during this period were classified as ‘non-helping subordinates’. Whether or not a subordinates helped during a breeding event is called ‘helping propensity’. We measured ‘helping effort’ as the number of feeds per hour per nestling. The age of each group member was assessed from its fledging year (when ringed as nestling) or its feather development at first capture (‘juvenile’, ‘immature’ or ‘fully-grown’ based on molt patterns of primaries, secondaries and coverts; Jackson 2005). Individuals were sexed using a set of sex-linked primers P2/P8 (Griffiths et al. 1998) and relatedness to the breeding pair was based on 12 microsatellite loci (see Husemann et al. 2015 and Cousseau, Van de Loock, et al. 2020 for primer details). We used the exclusion method (Jones and Ardren 2003; by using a self-written R script) to identify first-order kin (son or daughter). Previous analysis of known mother-offspring combinations in our study species showed genetic mismatches at a single locus in only 8% of the cases (Cousseau, Van de Loock, et al. 2020). Group members were therefore considered first-order relatives of a male or female breeder if they either showed no mismatch or a mismatch at a single locus, and were otherwise considered non-relatives.