A widespread hypothesis for the ontogeny of behaviour and decision-making is the early-exploration-later-canalization hypothesis. It postulates that juveniles are more exploratory and adults more consistent in their behavior. In addition, it is often assumed that naïve juveniles could overcome the costs of individual experience building by copying more the decisions of others than adults (early-conformism-later-self-defining hypothesis). Here we compare the central-place-foraging movements of adults and post-fledging juveniles in their first flights around the colony before dispersal and migration, in two sympatric species of tropical seabirds: red-footed boobies and great frigatebirds. Using GPS records of individual movements, we analyzed the foraging directions of seabirds from the colony across successive trips. Juveniles of both species showed significant within-individual consistency in foraging direction but at lower levels than adults. Juveniles leaving the colony within the same time-window showed significant but low between-individual resemblance in foraging direction, at levels similar to adults. In both species, homing efficiency was lower in juveniles than in adults. Juvenile foraging directions were initially influenced by wind conditions, particularly in low-wing-loading frigatebirds. Wind conditions progressively lost influence on juvenile foraging directions during their first weeks of flights. In contrast within-individual consistency, between-individual resemblance and homing efficiency did not show signs of progression in juveniles. Our results support the early-exploration-later-canalization hypothesis but not the early-conformism-later-self-defining hypothesis. Relaxed constraints on self-feeding efficiency could favour high variability in post-fledging tropical seabirds. Our simple approach could be applied to further test these hypotheses by comparing strategies across a wide range of central-place foragers.

Birds were captured on nest and fitted with GPS-data loggers with adhesive tape. From the GPS tracks we identified each single foraging excursions from the colony, each one corresponds to one row in the provided dataset (TripMetadata.csv).

For each foraging trip we extracted statitistics such as the distance and direction from colony along diverse points of the track: maximum range from colony, different % of the maximum range from colony on both the way out and the way in, or the first point reaching 10, 20, etc. km from the colony both on the way out and the way in. From each of these points we also extracted a standardized straightness index.

Finally we also added information on wind (u,v as available from NOAA servers, but also transformed into other referentials e.g. wind direction and speed (polar referential) or in parallel/perpendicular to flight direction referential.

See M&M for more details on data collection and the analyses reported in the manuscript.

All analyses presented can be replicated following the script StatisticalAnalyses&Figures.R, as well as all figures except for Fig1 which requires original tracking data.

Notably, analyses on resemblance call on two specific scripts to establish resemblance between all possible pairs of trips (for juv and ad, respectively).

Sensitivity analyses can be carried by changing parameter values, or in the case of changing the definition of trip directions (e.g. first 10km) by using the script DirectionAt10km.