Memory in trait macroevolution

Emma E. Goldberg, Jasmine Foo

Research output: Contribution to journalArticlepeer-review

9 Scopus citations


The history of a trait within a lineage may influence its future evolutionary trajectory, but macroevolutionary theory of this process is not well developed. For example, consider the simplified binary trait of living in cave versus surface habitat. The longer a species has been cave dwelling, the more accumulated loss of vision, pigmentation, and defense may restrict future adaptation if the species encounters the surface environment. However, the Markov model of discrete trait evolution that is widely adopted in phylogenetics does not allow the rate of cave-to-surface transition to decrease with longer duration as a cave dweller. Here we describe three models of evolution that remove this memoryless constraint, using a renewal process to generalize beyond the typical Poisson process of discrete trait macroevolution. We then show how the two-state renewal process can be used for inference, and we investigate the potential of phylogenetic comparative data to reveal different influences of trait duration, or memory in trait evolution. We hope that such approaches may open new avenues for modeling trait evolution and for broad comparative tests of hypotheses that some traits become entrenched.

Original languageEnglish (US)
Pages (from-to)300-314
Number of pages15
JournalAmerican Naturalist
Issue number2
StatePublished - Feb 1 2020

Bibliographical note

Funding Information:
We are grateful to Maria Servedio for organizing the symposium and imposing deadlines that enforced progress on this project. We thank members of the ?theory under con-struction? group at the Department of Ecology, Evolution and Behavior, University of Minnesota, for their comments in its early stages. Tanjona Ramiadantsoa implemented initial simulations of the renewal process. Will Freyman suggested testing inference with fossil tips and applying the renewal model to molecular sequence evolution. Michael Landis and Wayne Maddison pointed out important connections with previous work. Sally Otto caught a mathematical error of some hazard functions in an earlier version of the work. Matt Pennell and an anonymous reviewer provided additional constructive comments, including emphasizing the connection to hidden state models. The Minnesota Supercomputing Institute at the University of Minnesota provided computing resources. This work was supported by National Science Foundation grants DEB-1655478/1940868 to E.E.G. and DMS-1349724 to J.F.

Publisher Copyright:
© 2019 by The University of Chicago.


  • Comparative methods
  • Phylogenetics
  • Renewal process
  • Trait evolution


Dive into the research topics of 'Memory in trait macroevolution'. Together they form a unique fingerprint.

Cite this