Social carry-over effects underpin trans-seasonally linked structure in a wild bird population

Abstract

Spatial structure underpins numerous population processes by determining the environment individuals' experience and which other individuals they encounter. Yet, how the social landscape influences individuals' spatial decisions remains largely unexplored. Wild great tits (Parus major) form freely moving winter flocks, but choose a single location to establish a breeding territory over the spring. We demonstrate that individuals' winter social associations carry-over into their subsequent spatial decisions, as individuals breed nearer to those they were most associated with during winter. Further, they also form territory boundaries with their closest winter associates, irrespective of breeding distance. These findings were consistent across years, and among all demographic classes, suggesting that such social carry-over effects may be general. Thus, prior social structure can shape the spatial proximity, and fine-scale arrangement, of breeding individuals. In this way, social networks can influence a wide range of processes linked to individuals' breeding locations, including other social interactions themselves.

Keywords: Carry-over effects; habitat selection; social networks; social relationships; spatial structure; territory choice.

Figure 1

a) Illustrative breeding positions and winter social network of great tits from 1 year (2012) in Wytham Woods. Black dividing lines show the inferred territory boundaries around each individuals breeding location. Points show breeding sites of birds recorded in winter social network (star = both parents, cross = male, ‘X’ = female). Adjoining lines represent social associations recorded in the previous winter, and line thickness illustrates strength of association (mean strength is displayed for boxes where both parents were included). A section has been enlarged for clarity. b) Mantel r test statistic assessing the relationship between winter social networks and subsequent breeding proximity matrices. Circles show the observed statistic and boxes show the 95% range of statistics calculated from the spatial null model (mid‐lines illustrate mean of these). Proximity is calculated from Euclidean distance. See Fig. S5 for ranked proximity and mantel tests for each class of individuals separately.

Figure 2

a) Relationship between winter social association strength and subsequent breeding distance for subsequent neighbours (solid lines) and non‐neighbours (dotted lines). Lines are based on the GLMM (see Table S4 for details) and colour denotes year. Box plots show each individual's winter social association to subsequent non‐neighbours (dotted – left) and neighbours (solid – right) within the range of their closest non‐neighbour and furthest neighbour (Fig. S2). Mid‐lines show median, box shows interquartile range (IQR), whiskers shows range (with values outside 1.5 times IQR excluded). b) The average winter social association strength individuals held to their subsequent neighbours at each sampling period (weekend) prior to it. Only social associations between subsequent neighbours are considered. X‐axis shows the number of days before the mean lay date in the following breeding season. Circles show the observed average winter association strength, and polygons show the 95% range calculated from the spatial null model.