Experimental Evolution: Concepts, Methods, and Applications of Selection Experiments

22. Laboratory Evolution Meets Catch-22: Balancing Simplicity and Realism

671 22 LABORATORY EVOLUTION MEETS CATCH-22 Balancing Simplicity and Realism Raymond B. Huey and Frank Rosenzweig Everything should be made as simple as possible, but not simpler. ATTRIBUTED TO A. EINSTEIN USING LNS TO TEST HYPOTHESES DERIVED FROM COMPARATIVE STUDIES PROBLEMS ASSOCIATED WITH LABORATORY ADAPTATION Selecting on Field-Fresh Lines Selecting on Laboratory-Adapted Lines Laboratory Environments Are Too Benign Laboratory Environments Are Too Stressful Simplicity Can Be Deceiving A CASE STUDY: A SELECTION EXPERIMENT AT ODDS WITH FIELD STUDIES ON MODIFYING LNS EXPERIMENTS CONCLUSION Experimental Evolution: Concepts, Methods, and Applications of Selection Experiments, edited by Theodore Garland, Jr., and Michael R. Rose. Copyright © by the Regents of the University of California. All rights of reproduction in any form reserved. This book lays out a clear and compelling message: selection experiments are remarkably powerful tools in the armamentarium of evolutionary biologists. We ourselves have often used selection experiments during our careers and certainly expect to use them in the future. In fact, the power and elegance of selection experiments applied to life-history evolution by Rose and colleagues (Rose and Charlesworth 1980; Service 1987) motivated one of us (R.B.H.) to switch from conducting descriptive evolutionary studies on lizards in the field to performing evolution experiments on Drosophila in the laboratory. In this chapter, we look critically at a particular type of experimental evolution, often called laboratory natural selection. In this protocol, stocks of organisms are reared chronically under different conditions (e.g., different thermal or life-history regimes) and allowed to evolve by natural selection over many generations (Rose et al. 1987; Garland 2003). At intervals, phenotypes of population members can be compared in a “common garden” (i.e., reared under identical environmental conditions; Garland and Adolph 1991). Differences between selected and control lines—at least if observed consistently among replicates—represent either direct or indirect responses to the selective regime. This is an old and venerable type of experimental evolution (Dallinger 1887; see box). Laboratory natural selection (LNS) is distinct from two other types of laboratory evolution (Rose et al. 1996, 1987; Garland 2003; Swallow and Garland 2005; Futuyma and Bennett this volume). In artificial selection, the experimenter actively measures and selects phenotypes to found the next generation. In laboratory culling, organisms are exposed to a lethal condition (e.g., no food, no water), and the longest-surviving individuals are used to found the next generation. LNS experiments can provide insight into genetic architecture and correlations underlying traits of interest (Rose et al. 1990). They can also be used to evaluate the rate, tempo, and repeatability of evolutionary trajectories (Lenski and Travisano 1994; Ferea et al. 1999; Dunham et al. 2002; Cooper et al. 2003, 2001; Fong et al. 2005; Woods et al. 2006) and to assess how historical contingency (Travisano et al. 1995), sex (Grimberg and Zeyl 2005; Zeyl et al. 2005), sexual selection (Rundle et al. 2006), ploidy (Paquin and Adams 1983; Zeyl and Bell 1997; Zeyl et al. 2003), and life history (Zeyl et al. 2005) influence the outcome of those processes. LNS experiments are especially useful for testing functional hypotheses (Rose and Charlesworth 1980; Bennett and Lenski 1999; Gibbs 1999), as derived lines yield experimental subjects that have “exaggerated” or “novel” phenotypes (Gibbs 1999; Bennett 2003; Garland 2003; Futuyma and Bennett this volume). LNS experiments are not only broadly applicable but also logistically advantageous (Rose et al. 1987; Futuyma and Bennett this volume). Effective population size can be manipulated over orders of magnitude, especially in microbes, largely eliminating genetic drift, if so desired. Experimental lines can readily be (should be!) replicated (Futuyma and Bennett this volume), and the intensity, frequency, uniformity, and duration of selection can carefully controlled. Selective agents of interest can be applied either singly or in concert. Moreover, selection is accomplished without direct intervention: 672 • C O N C L U S I O N E X P E R I M E N T A L E V O L U T I O N A N D C A T C H 2 2 • 673 THE FIRST LNS EXPERIMENT Probably the first experimental study of laboratory evolution was conducted in the 1880s by the Rev. W. H. Dallinger, who was President of the Royal Microscopical Society. Inspired by Darwin, Dallinger (1887) decided to determine experimentally “whether it was possible by change of environment . . . to superinduce changes of an adaptive character, if the observations extended over...