Previous studies have reported functional integration between dispersal and other phenotypic traits allowing individuals to alleviate dispersal costs, and such associations can affect dispersal evolution in return. In sexually reproducing species, assortative mating according to dispersal can shape the maintenance of such trait associations. Despite the potentially crucial consequences of dispersal in natural populations, assortative mating for dispersal and its underlying mechanisms remain largely unexplored. Here, we assessed assortative mating for between-patch dispersal status in a fragmented population of a small passerine bird, the collared flycatcher, and explored whether such assortative mating could result from (i) direct mate choice based on dispersal-related behavioural (aggressiveness and boldness) and morphological traits (tarsus and wing length), (ii) biased mating due to spatio-temporal heterogeneity in the distribution of dispersal phenotypes and/or (iii) post-mating adjustment of dispersal phenotype or dispersal-related traits. We found intrinsic assortative mating (i.e. positive among-pair correlation) for current dispersal status (in the year of mating) but not for natal dispersal status, even though we could not exclude it due to limited power. We also found assortative mating for boldness and age category (yearlings vs. older adults), and the probability for pair members to be assorted for current dispersal status was higher when both pair members were of similar boldness score and of the same age compared to mixed-age pairs. Mate choice based on boldness and age thus appears as a possible mechanism underlying assortative mating for dispersal status. Our analyses however remained correlative and only an experimental manipulation of these traits could allow inferring causal links. Non-random mating for dispersal-related traits may affect the evolution of dispersal syndromes in this population. More work is nevertheless needed to fully assess the evolutionary implications of age- and behaviour-based assortative mating for dispersal.

The dataset was collected during several field word between 1986 and 2013 at the Island of Gotland (Sweden) where the population of collared flycatcher was monitored since 1980s.  We performed univariate and bivariate Bayesian model (MCMCglmm) to identify an assortative mating based on dispersal status and to investigate possible underlying mechanisms (e.g. assortative mating based on behavioural and morphological traits and age) that could be explain this type of assortative mating in order to produce a MS accepted for publication in Journal of Evolutionary Biology.