Data from: Balancing moult, migration, and breeding in a long-lived partially migrant raptor
Data files
Mar 10, 2026 version files 165.41 KB
-
moult_EV_Raw_Database.csv
158.87 KB
-
README.md
6.54 KB
Abstract
Moult, breeding, and migration are the three major life-history events in the annual cycle of birds. All are energetically demanding processes that rarely overlap. In large raptors such as the Egyptian vulture Neophron percnopterus, completing a full moult may take more than one year, requiring birds to balance this process with other life-cycle events. We analysed 740 moult cards from across the species’ distribution, including migratory, sedentary, and mixed subpopulations, to test predictions about trade-offs between the physiological demands of migration and moult. Juveniles began moulting in January-February of their second calendar year and continued "year-round" until their fifth calendar year. Adults, however, suspended moult from the months preceding spring migration until the late breeding phase. Subadults from the Canary Islands started moulting earlier and completed it faster than those from Western Europe, while adults from Western Europe initiated moult later than adults from the Canary Islands, India, and Oman. Sedentary subpopulations, particularly those from the Canary Islands, showed a greater moult extent and replaced more primaries annually than migratory and mixed subpopulations, whereas the wintering subpopulation of southwestern Spain (Extremadura) exhibited the smallest moult extent. Moult symmetry decreased from the second plumage onwards and differed between the Canary Islands and Oman subpopulations. Geographic differences in moulting patterns likely arise from population-specific life-history trade-offs, reflecting local adaptations or responses to environmental constraints. These patterns highlight the influence of migratory strategy on moult progression and suggest that sedentary lifestyles may allow more extensive feather renewal in long-lived raptors.
Dataset DOI: 10.5061/dryad.c2fqz61qt
Description of the data and file structure
We analysed the timing and duration of moult using the models of Underhill and Zucchini (1988) and Underhill et al. (1990), implemented in the ‘moult’ package (Erni et al. 2013) for R (www.r-project.org). These models estimate moult duration, the average start date of moulting (from which the average end date can be derived), and the standard deviation of the start date. We modelled primary moult of each subadult age class plumage scores against Julian day from 1 January of one year to 31 December of the following year (730 days), to account for the nearly two-year continuous moult cycle of these age classes (Fig. 1 and 2;~~~~ Supporting information). We modelled primary moult scores for adults against Julian day fromMarch onwards (365 days, one year). Adults also require two years to complete the moult, but in two non-continuous periods, and therefore, for adults we only considered the moult scores of one year. To determine the best way to code moult data, we conducted a preliminary analysis comparing five approaches (type of moult data arguments) in null models. The most parsimonious method (type 1; Supporting information; Erni et al. 2013) categorized individuals within the population as pre-moult, in-moult, or post-moult (Underhill and Zucchini, 1988) and has been previously applied to modelling moult scores in large raptors (Zuberogoitia et al. 2016). Moult scores were standardized as a ratio between 0 and 1 to run the analysis. We ran beta regression models using the ‘betareg’ package to analyse continuous variables bounded between 0 and 1, such as rates and proportions (Cribari-Neto and Zeileis 2010, Zeileis et al. 2012), to examine geographic differences in moulting patterns. We first analysed moult start and duration to test whether birds (adults and subadults) from the long-distance migratory Western European population start moulting later or have shorter moult duration than the sedentary Canary Islands subpopulation. Plumage (i.e. subadult age-classes) was included as a factor to test for age-related differences and we added subpopulation (Western Europe, Canary Islands) as an additional factor to compare these two groups. A separate analysis was performed for adults, including subpopulation as the only factor. We then extended both models by including all four subpopulations (Western Europe, Canary Islands, India, and Oman), excluding those with small sample sizes.
Next, we modelled adult primary moult scores from the previous year to test whether moult extent differed 1) between continental and island subpopulations, 2) between migratory and resident populations, and 3) among individual subpopulations. Feather symmetry was then analysed in two ways.: First, symmetry indices between primaries and secondaries were modelled, using feather type (primaries versuss. secondaries) and month as predictors, with nested effects to account for temporal trends and variation between feather types. Second, primary feather symmetry was modelled among subpopulations using data from individuals with symmetry information available for both wings. The number of subpopulations included in this latter analysis differed from previous models due to incomplete symmetry data. Model fit and assumptions were evaluated using diagnostic plots of Pearson residuals versus fitted values and Q–Q plots (Cribari-Neto and Zeleis 2010). Competing models were compared using Akaike iInformation cCriteria (AIC), and significant differences between model performance were tested using the log-likelihood ratio test (Burnham and Anderson 2002). All analyses were performed in R software ver. 4.1.3 (www.r-project.org).
The data file moult_EV_Raw_Database.csv contains the following variables:
- ID – Individual identification code.
- Subpopulation1 – Subpopulation considered in this study. Some subpopulations are marked with “?” because migratory and resident populations may overlap.
- Subpopulation2 – Country or region.
- Locality – Specific locality information when available.
- Reference – Each individual is linked to one or more photographs obtained from online repositories or from field researchers and professional photographers. For each ID there may be one or several photographs, often including both wings as well as ventral and dorsal views. As a result, the total number of photographs amounts to several thousand, making it impractical to upload them to a repository. In addition, many photographs belong to professional photographers who granted permission only for use in this study and cannot be publicly distributed without their consent. However, specific photographs may be requested from the corresponding author.
- Date – Date when the individual was photographed.
- Transformed date – Date standardized according to the individual’s age (calendar years), used to generate the figures.
- Age – Age of the individual in calendar years (cy): 1cy, 2cy, 3cy, 4cy, 5cy, and adult (5cy+).
- Sex – Sex of the individual when known.
- The following columns correspond to the moult scoring sheet for Egyptian Vultures. The first half refers to the right wing and right side of the tail, and the second half to the left side.
- Flight feathers are numbered from the outermost feather towards the body: P10, P9, … P1, S1, S2, … S20. Tail feathers are numbered from the outermost towards the centre: R7, R6, … R1 (Egyptian Vultures have 14 rectrices).
- Moult scores- After each group of primaries, secondaries, or rectrices, the sum of the moult scores is provided.
- old feathers were scored as 0, fully grown new ones as 5, and growing ones as 1–4 for intermediate stages of feather development.
- Feather status code: In addition to the numerical code indicating the feathers moulted in the current season, a letter code is provided to indicate the feather status (i.e., feather age).
- Feathers scored as 0 were also recorded as ‘A’ (juvenile); ‘K’ were feathers moulted in the previous moult year, ‘M’ moulted two years previous and ‘O’ moulted three years previous. In some cases appear ‘K5’ that were feathers moulted at the end of the winter.
We created moult cards from high-definition and high-quality pictures (sufficient to compare feather wear and changes in tone and colour) of wild birds taken by research teams during trapping events (n = 265), wildlife photographers (n = 251), and two online citizen science repositories, the Macaulay Library (www.macaulaylibrary.org/) and iNaturalist (www. inaturalist.org/) (n = 81 and 143, respectively) (moult_EV_Raw_Database.csv). To avoid pseudoreplication, data on single individuals from different sources were not used, although we used different pictures of the same individual to cross-check the moult cards. In the few cases where individuals from the same area were selected, individual marks and moult state were checked to avoid resampling of the same bird. Scoring and recording moult on cards for all photographs was conducted by a single, trained researcher (IZ, see Zuberogoitia et al. 2013). Birds in hand were photographed with extended wings (normally taking two photos per wing, one per side). For 452 birds, we had photographs of both wings, and for 288, only one wing was photographed. A moult card was filled out for each wing with a photo, and feather generation was identified based on wear, shape, colour, age pattern, and growth of the remiges (i.e. help_outlineprimaries and secondaries; Zuberogoitia et al. 2018). This provided data for 10 primaries and 20 secondaries for each wing. Primaries were numbered in descending order (from inside to outside), and secondaries were numbered in ascending order (from outside to inside).To calculate moult scores, we followed Ginn and Melville (2000) and Newton (2009), using a standard recording system, based on the growth stage of primaries and secondaries: old feathers were scored as 0, fully grown new feathers as 5, and growing feathers as 1–4 for intermediate stages of development. These individual moult scores were then summed to give an overall moult score between 0 and 50 (for 10 primaries), and 0 and 100 (for 20 secondaries). However, moult in large birds, whichrequirese more than one year to replace all their remiges, occurs in ‘waves’ as a process termed ‘serial moult’ (Edelstam 1984) or ‘Staffelmauser’ (Zuberogoitia et al. 2018). Following Zuberogoitia et al. (2013), feathers scored as 0 were also recorded as ‘A’ (juvenile), ‘K’ (moulted in the previous moult year), ‘M’ (moulted two years earlier), and ‘O’ (moulted three years earlier) (Supporting information)