# Breeding under pressure: corticosterone is associated with reproductive investment under fluctuating predation risk in a long-lived sea duck [https://doi.org/10.5061/dryad.xd2547dpj](https://doi.org/10.5061/dryad.xd2547dpj) **Authors:** Bertille Mohring1,2, Markus Öst1,3, Kim Jaatinen4,5, Charline Parenteau2, Emmanuelle Grimaud2, Frédéric Angelier2 1Environmental and Marine Biology, Åbo Akademi University, 20500 Turku, Finland 2Centre d’Etudes Biologiques de Chizé, UMR 7372 CNRS – La Rochelle Université, 79360 Villiers-en-Bois, France 3Novia University of Applied Sciences, 10600 Ekenäs, Finland 4Nature and Game Management Trust Finland, 10160 Degerby, Finland 5Tvärminne Zoological Station, University of Helsinki, 10900 Hanko, Finland ## Description of the data and file structure Four datasets have been used in this paper, and the variables are detailed below and a broader description of the data is available in the paper. **Dataset 1 (dataset\_CORT.csv): Intrinsic and extrinsic determinants of female baseline corticosterone levels.** This dataset was used to assess population-level trends in baseline corticosterone levels (5.1), as well as the intra-individual consistency of baseline CORT levels (i.e., repeatability) (5.2) and the determinants of female baseline corticosterone levels (5.2). Variables: Year: Year (range: 2009-2022) Individual\_ID: Female ID (unique metal ring number) Age: Female minimum age or minimum number of years of breeding experience, starting at 1 (1 corresponds to the first time the female was trapped and ringed) BodyCondition: Female body condition index (see methods for further details on how this index was obtained) Island\_type: island type (open or forested island) ClutchSize: Number of eggs in the clutch RLD: relative laying date, in days (obtained by centering laying date around the yearly median laying date of the entire population to allow comparison of individual breeding timing irrespective of annual phenology) Adult\_pred: Island-specific adult predation risk, proportion of breeding females found killed on an island for a given year (see methods for further details on how this index was obtained) Nest\_pred: Island-specific nest predation risk, proportion of depredated nest on an island for a given year (see methods for further details on how this index was obtained) Age\_within\_std: Within-individual variation in age, standardised CS\_within\_std: Within-individual variation in clutch size, standardised RLD\_within\_std: Within-individual variation in relative laying date, standardised Adult\_pred\_within\_std: Within-individual variation in island-specific adult predation risk, standardised Age\_btw\_std: Among-individual variation in age, standardised CS\_btw\_std: Among-individual variation in clutch size, standardised RLD\_btw\_std: Among-individual variation in relative laying date, standardised Adult\_pred\_btw\_std: Among-individual variation in island-specific adult predation risk, standardised **Dataset 2 (dataset\_eagles.csv): Eagle index over time (5.1)** Variables: Year: Year (range: 2009-2022) Eagle\_index: annual index of white-tailed eagle abundance in the study area during the eiders’ reproductive season (from April 1 to June 15). This index was calculated, for each year, as the average daily number of white-tailed eagle sightings between April 1 and June 15 at the nearby (ca 20 km) Hanko Bird Observatory (HALIAS). **Dataset 3 (dataset\_island\_pred.csv): Population-level temporal trends in island-specific adult and nest predation risk (5.1)** Variables: Year: Year (range: 2008-2019) Island\_ID: Island ID NestPredationRisk: Island-specific nest predation risk, proportion of depredated nest on an island for a given year (see methods for further details on how this index was obtained) AdultPredationRisk: Island-specific adult predation risk, proportion of breeding females found killed on an island for a given year (see methods for further details on how this index was obtained) **Dataset 4 (dataset\_reproS.csv): Hatching success (5.1 and 5.3)** Variables: Year: Year (2009-2022) Individual\_ID: Female ID ReproS: Hatching success (0: failure, 1: success) OnNest: number of days of incubation CORT: Baseline corticosterone levels (in ng/mL) CORT\_within\_std: Within-individual variation in baseline corticosterone levels, standardised CORT\_btw\_std: Among-individual variation in baseline corticosterone levels, standardised ##

Phenotypic plasticity may enable individuals to cope with predictable and unpredictable environments during their life-cycle. In that context, studying glucocorticoids – corticosterone (CORT) in birds – is relevant because of their primary role in allostasis. Higher baseline CORT levels are classically associated with environmental constraints and lower fitness (the CORT-fitness hypothesis). However, in some environments, higher baseline CORT levels can promote reproduction, therefore being associated with higher fitness (the CORT-adaptation hypothesis). These two hypotheses have been tested in multiple systems, but rarely in a context of fluctuating predation threat. We used a long-term individual-based monitoring of baseline CORT levels in female common eiders Somateria mollissima (nCORT = 1537; nindividual = 790; 2009-2022) to disentangle the context-dependent links between environmental conditions, CORT and fitness. Importantly, the study population has been facing a drastic increase in predation pressure over the past decades, linked to the recovery of the white-tailed eagle Haliaeetus albicilla. Additionally, eiders breed on open or forested islands, further affecting adult and nest predation risk. This system allowed us to disentangle the relative contributions of within- and among-individual variation in baseline CORT levels under predation. Supporting the CORT-adaptation hypothesis, baseline CORT levels were positively associated with reproductive investment (clutch size), age and hatching success. By partitioning within- and among-individual effects, we showed that at the individual level, CORT flexibly increased with clutch size and age. Females displaying higher CORT levels were more successful, suggesting a link between CORT and individual quality. At both the population and individual levels, baseline CORT levels decreased over the study period. This decrease was correlated with an increase in predation risk. Females had reduced baseline CORT when nesting under high eagle abundance or adult predation risk (within-individual effect). Interestingly, apparent plasticity towards adult predation risk was only observed on open islands, likely reflecting habitat-dependent strategies. Consistent with the CORT-adaptation hypothesis, we show that changes in predation regime not only correlate with changes in reproductive investment, but also with rapid plastic adjustment of glucocorticoid levels and therefore individual strategies to cope with predation risk. Given the correlative nature of our study, we encourage further experimental studies testing for a causal relationship between predation and corticosterone levels.