Directionality range in Emlen funnels

Abstract Emlen funnels can be used to study the birds’ ability to orient during the migratory seasons. Birds are so eager to migrate that they will jump in the direction in which they want to fly, even if they are placed in small cages during the night. Emlen funnels have therefore been used for decades to study the sensory capabilities and mechanisms that migratory birds use to find their way. A significant part of this research has focused on how night-migratory songbirds perceive the Earth’s magnetic field. Even though Emlen funnels have been proven quite successful in capturing the birds’ behavioural responses to different experimental conditions, the orientation behaviour of night-migratory songbirds in Emlen funnels is very noisy, i.e. tends to have a low signal to noise ratio. This noise makes Emlen funnel experiments very time consuming and limits the types of questions that can be studied to those which require very few different experimental conditions (permutations). Furthermore, the experimental design choices can be crucial, e.g. degree of blinding of the experimenters and data evaluators to the conditions tested, pre-testing of birds to make sure they are in a migratory state, and planning of the effective sample sizes. Hence, different traditions in experimental design choices can reduce reproducibility and comparability and render the interpretation of the results non-trivial. To better understand Emlen funnel data and the minimal requirements for good experimental design, we constructed and analyzed a large data set that we compiled by combining behavioural data from many previous Emlen funnel studies performed in Oldenburg. Our results provide realistic ranges for the expected orientation of the birds in Emlen funnels, which can be useful (and in some research crucial) for predicting the optimal sample sizes for future experiments. Our results thus offer concrete information for the design and analysis of statistically powerful future magnetic orientation experiments.