Avian migrants may fly at a range of altitudes, but usually concentrate near strata where a combination of flight conditions is favourable. The aerial environment can have a large impact on the performance of the migrant and is usually highly dynamic, making it beneficial for the bird to regularly check the flight conditions at alternative altitudes. We recorded the migrations between northern Europe and sub-Saharan Africa of European nightjars Caprimulgus europaeus to explore their altitudinal space use during spring and autumn flights to test whether their climbs and descents were performed according to predictions from flight mechanical theory. The nightjars commonly operated at ascent rates below the theoretical maximum, and periods of descent were commonly undertaken by active flight, and rarely by energetically cheaper gliding flight, allowing the birds to sink at lower rates than possible during a best glide scenario. Spring migration across all regions was associated with more exploratory vertical flights involving major climbs, a higher degree of vertical displacement within flights, and less time spent in level flight, although flight altitude per se was only higher during the Sahara crossing. This study demonstrates a surprisingly frequent use of slow vertical displacements during migratory flights of a long-distance avian migrant and show how these behaviours vary between seasons and regions, presumably in response to different flight conditions. The results should inspire future studies on the potential costs associated with frequent altitude changes and their trade-offs against anticipated flight condition improvements for aerial migrants.

Data files in plain text containing detailed accelerometer, pressure and temperature data recorded on European nightjars Caprimulgus europaeus. Files containing activity data: For each timestamp the distribution of 12 activity registrations are presented in columns Act0 to Act10 (Act0 to Act5 for activity data 2018-2019) that represent a range of activity levels given a predefined sampling routine. See, methods and supplement in the related work. Column “check” is the sum of all activity counts and should be 12 if sampling worked as intended. Id represents each datalogger. The pressure data file contains a timestamp following the recorded temperature in °C and barometric pressure in mbar.