Data related to "Sex differences in development alter the fledgling sex ratio in a lekking bird with strong sexual size dimorphism"

This study analyses how developmental parameters affect the sex ratio in juvenile ruffs (Calidris pugnax).

The analyses contain two major parts:

1) How developmental parameters differ between males and females and how they associate with adult body size. For this, we monitored wing growth, feather maturation and flight ability of captive ruffs from 2019 until 2023 and measured adult body mass and adult tarsus length in the same individuals. All measurements were taken at two locations, Burnaby (Canada) and Seewiesen (Germany). The population was originally established in Canada from eggs collected near Oulu, Finland. It was translocated to Seewiesen in 2019 and 2020 and supplemented with additional ruffs obtained from breeders and zoological gardens in the Netherlands, Belgium and Germany. Flight ability was experimentally determined conducting daily flight tests. We analyzed the data with two linear mixed models, both having fledging age as response variable. The first model contained 'sex' and 'adult body mass' as fixed effects, and 'cohort' (hatch year) and 'mother ID' as random effects. The second model contained 'adult tarsus length' instead of 'adult body mass'.
To compare wing length, wing growth and feather maturation in juveniles between sexes we first modelled posterior means of wing or emerged feather length (and their 95% credible intervals (CrIs)) in relation to age for each sex using measurements that were taken two to three times per week. For these growth models, we used a flexible Bayesian generalized additive model framework based on Markov chain Monte Carlo simulations that can include random factors. We included ‘mother ID’, ‘Individual ID’ and an interaction between ‘Individual ID’ and ‘age’ as a random factor in all models to allow for individual specific growth curves. In separate models, we estimated posterior means and 95% CrI of wing growth rates, the proportion of adult wing length, the rate, and proportion of feather emergence in relation to age. For all analysis of this first major part, we used R version 4.4.2 (R Development Core Team, 2024) and each variable is defined in the Readme file Readme_1.docx. The file formats are csv, R and rds files. The main correspondents for this part, i.e. Part 1, are Lina Giraldo-Deck and Clemens Küpper.

2) The analysis of how variation in development affects the ruff sex ratio from hatching to fledging, included three parts: 2a), 2b), and 2c).
2a) We estimated posteriors to inform the input parameters of our matrix model: probability of being male at hatch ρ, sex- and age-specific intrinsic survival ϕ_tS, sex- and age-specific probability of fledging ψ_tS, additive predation mortality D, flight survival advantage F. The posterior distribution of ρ, was estimated from molecular sexing data of newly hatched ruff chicks in Pitkänokka meadows (Finland, from 2016 until 2023). The model used to inform ρ had an intercept and two random effects: ‘Clutch ID’ and sample year (‘Cohort ID’). We estimated the posteriors to inform ϕ_tS and ψ_tS with data from captive ruffs described in Part 1) (from 2018 to 2023) and mixed-effects known fate survival models. The intrinsic survival and fledging models had two of the same fixed effects: ‘Age’ and ‘Sex’—the fledging model also included an Age and Sex interaction. The random effects were Mother ID (‘Mother’) and Cohort year (‘Year’), and Year for the fledging and survival models respectively. The posteriors distributions from survival data from wild marked chicks from Finland (from 2019 until 2022) and sex-averaged intrinsic survival from captive chicks described in Part 1) (2018 to 2023) informed D. For the wild chick data, we prepared another known fate survival model with an intercept and random effects (‘Cohort ID’ and ‘Brood ID’). We derived D from the two posterior distributions with the formula: D = 1 - (Wild survival posterior distribution)/sex- and age-averaged intrinsic survival). We derived F from previously published black-tailed godwits (Limosa limosa) data. We determined the period-specific daily survival rates by taking the difference in the nth root (where n is number of days in each period) of the cumulative survival for pre-fledging and fledging periods.
2b) To examine how intrinsic survival and fledging age alter the sex ratio from hatching through fledging, we constructed a two-sex single cohort matrix model, derived from classical life-cycle models. However, we limited it to only the chick stage. We informed the model with the parameter estimates from 2a): ϕ_tS, ψ_tS, ρ, D, and F. We simulated four chains of 3,500 model runs, burning 1,000 runs per chain, to obtain 10,000 simulations. Each simulation consisted of 100,000 ruff chicks that survived and fledged from ages 0 to 27 days, and a unique combination of values of ϕ_tS, ψ_tS, ρ, D drawn from their posterior distributions. F was an exception with only a single value (0.086). From these simulations we derived: the post-fledge sex ratio; change in the sex ratio from hatch to fledge; sex-specific relative fitness w and selection coefficients s; and their means and 95% CrI.
2c) To examine the relative influence our input parameters had on our responses (post-fledge sex ratio, w, and s), we used two life table response experiments (LTREs). First, we quantified the relative impacts of sex-specific differences in fledging age ΔAge and ϕ, with an alternative ‘unisex model’. This model only differed from the ‘two-sex model’, described in 2b), in that it used a sex-average rather than sex-specific intrinsic survival. We compared the models by dividing the mean and variances of the unisex model responses by their counterparts from the two-sex model. Second, we used a perturbation LTRE to measure the effect of each input parameter on the post-fledge sex ratio and s♂. From these perturbations, we calculated both a relative contribution and the effect of one positive standard deviation change in each parameter on the two responses.

For all the analyses of 2), we used R version 4.4.2 (R Development Core Team, 2024), file formats are csv, R, Rmd, and rds, and names for data files/variables are within the Readme_2.docx. The main correspondents for Part 2, are James D. M. Tolliver and Clemens Küpper.