Linking phenotypic correlations from a diverse set of laboratory tests to field behaviors in the crayfish, Orconectes virilis. (25th March 2018)

• Record Type: Journal Article
• Authors: Edwards, David D.; Rapin, Kathryn E.; Moore, Paul A.
• Editors: Hebets, E.
• Is Part Of: Ethology. Volume 124:Number 5(2018)
• Page Start: 311
• Page End: 330
• Publication Date: 2018-03-25
• Subjects: behavioral phenotypes -- crayfish -- field behavior -- structural equation modeling; Animal behavior -- Periodicals; 591.5
• DOI: 10.1111/eth.12734
• Languages: English
• ISSNs: 0179-1613

Abstract

Abstract: The presence of phenotypic behavioral correlations and their connection to fitness consequences of organisms have received considerable debate within the literature. Yet, little work has been carried out to connect any behavioral correlates found within a set of laboratory studies to natural behavior observed under complex environmental conditions. To help fill this gap, individual crayfish, collected from the same local population, completed five different behavioral assays in a laboratory setting in a random order. These data were used to reveal any possible correlations for behavioral scores across all of the laboratory tests. Subsequently, these same individuals were placed into the field and video recorded for 24 hr. A separate set of field behaviors, related to the laboratory assays, were quantified from the field videos. The normalized laboratory and field behaviors were used in three stepwise statistical analyses. First, normalized data were loaded into a PCA to generate a priori hypotheses on potential behavioral correlates. These hypotheses were subsequently tested using general multiple linear regression. Finally, structural equation modeling was performed to elucidate any behavioral modules from the laboratory assays that correlated with behavioral patterns present from the fieldwork. Three laboratory‐based behavioral modules were connected to three separate field assays: exploration–avoidance, bold–shy, and aggressiveness. Yet, some behaviors exhibited in the laboratory assays were uncorrelated with any behaviors found in the field and vice versa. Results from this study provide evidence that although many different behavioral correlates may exist within laboratory settings, these same modules may not translate directly into predicting behavior under natural settings.