Contemporary phenotypic trends associated with global change are widely documented, but whether such trends always denote trait optimization under changed conditions remains obscure. Natural selection has shaped the wings of long-distance migratory birds to minimize the costs of transport, and new optimal wing shapes could be promoted by migration patterns altered due to global change. Alternatively, wing shape could vary as a correlated response to selection on other traits favoured in a changing environment, eventually moving away from the optimal shape for migration and increasing transport costs. Data from 20 years of monitoring of two Common Nightingale (Luscinia megarhynchos) populations breeding in central Spain, where environmental conditions for breeding have deteriorated during the last decades due to increased summer drought, show that birds have reduced wing length relative to body size over the period 1995-2014. However, long-winged nightingales survived their first round-trip migration better, and the shorter the average wing length of individuals, the stronger the survival-associated natural selection favouring longer wings. Maladaptive short wings may have arisen because the mortality costs of migration are outweighed by reproductive benefits accrued by short-winged nightingales in these populations. Assuming that the phenotypic integration of morphological and reproductive adaptations of migratory birds has a genetic basis, we hypothesize that the maladaptive trend towards shorter wings may be a correlated response to selection for moderate breeding investment in drying habitat. Our results provide evidence that contemporary phenotypic change may deviate average trait values from their optima, thereby increasing our understanding of the ecological constraints underpinning adaptation to rapid global change.

According to the journal (The Auk) policy about the ethical re-use of data, and considering the effort that the members of the SEO-Monticola ringing group put in assembling this database, we would appreciate that anyone interested in re-using the data to contact the corresponding author in order to agree to a collaboration when appropriate.

Analyses reported in this article can be reproduced using the data provided. In the README for databases.txt you can find a description of the variables included in each database file. Missing values are designated as NA.