Hybridization may promote variation in cognitive phenotypes in experimental guppy hybrids

This is a Preprint and has not been peer reviewed. The published version of this Preprint is available: https://doi.org/10.1086/720731. This is version 3 of this Preprint.


Download Preprint


Catarina Vila Pouca, Sijmen Vedder, Alexander Kotrschal


Hybridization is an important mechanism of evolution. While hybrids often express inferior traits and are selected against, hybridization can promote phenotypic variation and produce trait combinations distinct from the parentals, generating novel adaptive potential. Among other traits, hybridization can impact behaviour and cognition and may reinforce species boundaries when hybrids show decreased cognitive abilities. However, the hypothesized role of hybridization in the diversification of cognitive phenotypes remains enigmatic. To test this idea, we compare the performance of female guppies (Poecilia reticulata), Endler’s guppies (Poecilia wingei), and their experimental hybrids in colour association and reversal learning. In addition, we introduce a new approach to compare multidimensional cognitive phenotypes. We found that hybrids showed intermediate learning abilities in both tasks compared to the parentals. Moreover, hybrids had slightly higher phenotypic dispersion, new trait combinations occurred in some hybrid individuals, and the mean phenotype of one hybrid group deviated away from the axis of variation of the parentals. Our method should hence be useful in further exploring how hybridization, and other evolutionary processes, impact behavioural and cognitive traits. Our results suggest that hybridization may promote cognitive variation and generate new trait combinations, even when learning performance at the group level is intermediate between parentals.




Behavior and Ethology, Ecology and Evolutionary Biology, Evolution, Life Sciences


associative learning, cognitive flexibility, kernel density estimation, phenotypic novelty, Transgressive segregation


Published: 2021-09-03 22:16

Last Updated: 2022-02-02 02:37

Older Versions

CC-BY Attribution-No Derivatives 4.0 International

Additional Metadata

Data and Code Availability Statement:
Data and R code used in analyses will be made available through FigShare after acceptance for publication.

Add a Comment

You must log in to post a comment.


There are no comments or no comments have been made public for this article.