Wild cricket social networks show stability across generations

David N. Fisher, Rolando Rodríguez-Muñoz, Tom Tregenza* (Corresponding Author)

*Corresponding author for this work

Research output: Contribution to journalArticle

13 Citations (Scopus)
2 Downloads (Pure)

Abstract

Background
A central part of an animal's environment is its interactions with conspecifics. There has been growing interest in the potential to capture these interactions in the form of a social network. Such networks can then be used to examine how relationships among individuals affect ecological and evolutionary processes. However, in the context of selection and evolution, the utility of this approach relies on social network structures persisting across generations. This is an assumption that has been difficult to test because networks spanning multiple generations have not been available. We constructed social networks for six annual generations over a period of eight years for a wild population of the cricket Gryllus campestris.

Results
Through the use of exponential random graph models (ERGMs), we found that the networks in any given year were able to predict the structure of networks in other years for some network characteristics. The capacity of a network model of any given year to predict the networks of other years did not depend on how far apart those other years were in time. Instead, the capacity of a network model to predict the structure of a network in another year depended on the similarity in population size between those years.

Conclusions
Our results indicate that cricket social network structure resists the turnover of individuals and is stable across generations. This would allow evolutionary processes that rely on network structure to take place. The influence of network size may indicate that scaling up findings on social behaviour from small populations to larger ones will be difficult. Our study also illustrates the utility of ERGMs for comparing networks, a task for which an effective approach has been elusive.
Original languageEnglish
Article number151
JournalBMC Evolutionary Biology
Volume151
DOIs
Publication statusPublished - 27 Jul 2016

    Fingerprint

Keywords

  • exponential random graph models
  • gryllus
  • network comparison
  • population structure

Cite this