Abstract
Animals regularly use information from others to shape their decisions. Yet, determining how changes in social structure affect information flow and social learning strategies has remained challenging. We manipulated the social structure of a large community of wild songbirds by controlling which individuals could feed together at automated feeding stations (selective feeders). We then provided novel ephemeral food patches freely accessible to all birds and recorded the spread of this new information. We demonstrate that the discovery of new food patches followed the experimentally imposed social structure and that birds disproportionately learnt from those whom they could forage with at the selective feeders. The selective feeders reduced the number of conspecific information sources available and birds subsequently increased their use of information provided by heterospecifics. Our study demonstrates that changes to social systems carry over into pathways of information transfer and that individuals learn from tutors that provide relevant information in other contexts.
Highlights
Many animals use social information when making decisions related to fitness, such as for breeding and foraging [1,2,3,4,5]
Through experimentally manipulating the social foraging associations occurring among wild songbirds, we test how social information transmission is dependent upon social structure
We demonstrate that experimentally manipulating social structure in a wild animal community changes pathways of information transmission
Summary
Many animals use social information when making decisions related to fitness, such as for breeding and foraging [1,2,3,4,5]. If the s parameter is higher in matched dyads, it suggests that individuals disproportionately copied individuals who would access the same feeding stations as themselves over those they are associated with but could only access different feeding stations, i.e. the mismatched individuals they continued to co-occur with during the manipulation (see the electronic supplementary material). This inference is possible because the s parameter quantifies the increase in the rate of information transfer per unit of social association to knowledgeable individuals [16]. Following previous work [2,5], we took a conservative approach to inferring rates of social information use by removing discoveries between individuals that were within 10 min of each other to discard any occasions of birds discovering food together
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
