Skip to main content

Females compensate for moult-associated male nest desertion in the Hooded Warbler

Cite this dataset

Mumme, Ronald (2020). Females compensate for moult-associated male nest desertion in the Hooded Warbler [Dataset]. Dryad.


Uniparental offspring desertion occurs in a wide variety of avian taxa and usually reflects sexual conflict over parental care. In many species desertion yields immediate reproductive benefits for deserters if they can remate and breed again during the same nesting season; in such cases desertion may be selectively advantageous even if it significantly reduces fitness of the current brood. However, in many other species parents desert late-season offspring when opportunities to renest are absent. In these cases, any reproductive benefits of desertion are delayed, and desertion is unlikely to be advantageous unless the deserted parent can compensate for the loss of its partner and minimize costs to the current brood. We tested this parental compensation hypothesis in Hooded Warblers Setophaga citrina, a species in which males regularly desert late-season nestlings and fledglings during moult. Females from deserted nests effectively doubled their provisioning efforts, and nestlings from deserted nests received just as much food, gained mass at the same rate, and were no more likely to die from either complete nest predation or brood reduction as young from biparental nests. The female provisioning response, however, was significantly related to nestling age; females undercompensated for male desertion when the nestlings were young, but overcompensated as nestlings approached fledging age, probably because of time constraints that brooding imposed on females with young nestlings. Overall, our results indicate that female Hooded Warblers completely compensate for male moult-associated nest desertion, and that deserting males pay no reproductive cost for desertion, at least through the point of fledging. Along with other studies, our findings support the general conclusion that late-season offspring desertion is likely to evolve only when parental compensation by the deserted partner can minimize costs to the current brood.


Study site and general methods

This study was conducted in northwest Pennsylvania, USA, at Hemlock Hill Field Station in Crawford County. A colour-ringed population of Hooded Warblers has been under continuous study at Hemlock Hill since 2010 (Mumme 2014, 2018) and Hemlock Hill has been the site of numerous previous studies of this species (Chiver et al. 2011). After spending the winter in Central America, Hooded Warblers arrive at Hemlock Hill in early May, with the first nests initiated in mid-May and the final nesting attempts of the season started in late July. Because breeding adults typically begin their annual moult in late July or August, moult can overlap extensively with late-season nesting and parental care, particularly for males (Mumme 2018). Intensive efforts are made each year to capture and colour-ring all breeding adults and to locate and monitor all nests. Most of the data reported here were collected from late-season nesting attempts in July and August, 2018-2019.


Provisioning of nestlings

We collected data on provisioning of nestlings from 37 different late-season nests when nestlings were 1-7 days old; no data were used from hatching day (day 0) or from nests with young older than 7 days old, as fledging typically occurs at age 8-9 d. We performed 90 individual nest observations (never more than 1 per day) from the 37 different nests, including 44 observation bouts conducted from hides established near some nests, and 46 observations obtained from high-definition video recordings. The mean duration of the 90 observation bouts was 96 min (range 22-221 min). For each observation bout we recorded the number of feeding visits made by the colour-ringed female and male, and converted the counts to total provisioning rates (feedings hr-1) and nestling per capita feeding rates (feedings nestling-1 hr-1). Most of the provisioning data were collected July 11 - August 8 in 2018 (32 observations) and 2019 (56 observations), but data from 2 additional late-season video nest recordings made in 2016 were also included in the dataset. All observation bouts were conducted during dry weather, with most conducted 0600-1000 hrs.

We collected data from 27 undeserted nests where males fed nestlings during all watches, 5 completely deserted nests where males never fed nestlings during the duration of the nestling period, and 5 partially deserted nests, where males were present and feeding nestlings at least 1 day during the nestling period, but were absent and not provisioning nestlings on other days. At 2 of the partially deserted nests, males initially fed nestlings but then deserted midway through the nestling period. At the other 3 partially deserted nests, moulting males did not feed young nestlings but fed older nestlings before permanently deserting immediately after the young fledged. Collectively, the 90 separate nest observations included 42 observations where males were not provisioning nestlings, and 48 where they were. 


Nestling growth and survival

We investigated the effects of male desertion on nestling growth by collecting data on body mass of nestlings from late-season nests, July 10 - Aug 9, 2018-2019. We weighed nestlings daily used a 20-g spring scale and recorded nestling mass to the nearest 0.1 g. Because Hooded Warbler nestlings more than 4 days old will sometimes fledge prematurely if handled by humans, no nestlings older than day 4 were weighed, and we report data for nestling mass for days 1-4. Our dataset includes 233 nestling weights collected from 34 nests, including 4 completely deserted nests (46 individual nestling weights), 5 partially deserted nests (44 nestling weights), and 25 biparental nests (143 nestling weights).

To examine the effect of male nest desertion on survival of nestlings from hatching to fledging, we used a sample of nests from the years 2013-2019 that were initiated after July 10 where at least some eggs hatched and male feeding status at the nest was unambiguously determined. A total of 38 nests, including 13 completely deserted nests, 5 partially deserted nests, and 20 undeserted nests, fulfilled these criteria. Eight of these nests were destroyed by predators before the nestlings fledged, but the remaining 30 nests were used in an analysis of brood reduction and survival to fledging. 


Statistical analysis

Data files used in the analyses are openly available in the repository Dryad (Mumme 2020). All statistical analyses were conducted using JMP Pro 12.2 (SAS Institute 2016). Data on provisioning rates were analyzed with general linear mixed models using the Mixed Model personality of JMP Pro’s Fit Model platform and focused on 3 different rates: total provisioning rate (total feedings hr-1), nestling per capita feeding rate (feedings nestling-1 hr‑1), and female provisioning rate (female feedings hr-1). Number of nestlings, nestling age (d), and male feeding status that day (male present and feeding nestlings or absent and not feeding young) were entered into models as fixed effects. To account for multiple observations from the same nest, a unique nest identifier was included in all models as a random effect. All provisioning rates were log-transformed to equalize variance, and models were validated by ensuring that residuals were approximately normally distributed, as determined by examination of normal quantile plots and application of the Shapiro-Wilk test for normality. 

Data on nestling body mass were also analyzed with a general linear mixed model. The number of nestlings present in the nest and nestling age were included as fixed effects, as was nest desertion type — no male desertion, partial male desertion, or complete male desertion. Because of the very rapid pace of nestling growth, the time of day that nestlings were weighed was also included as a fixed effect. A unique nest identifier was included as a random effect to account for multiple repeated measures from the same nest. The resulting model was validated by ensuring that residuals were approximately normally distributed, as described above.

To examine the effects of male desertion on brood reduction and survival of nestlings to fledging, the proportion of hatched young that fledged for each nest was used as the dependent variable. We compared the proportion fledged for biparental, partially deserted, and completely deserted nests with a nonparametric Kruskal-Walis test. 


Allegheny College