vol. 194, no. 4 the american naturalist october 2019 E-Article What Determines the Distinct Morphology of Species with a Particular Ecology? The Roles of Many-to-One Mapping and Trade-Offs in the Evolution of Frog Ecomorphology and Performance Daniel S. Moen* Department of Integrative Biology, Oklahoma State University, Stillwater, Oklahoma 74078 Submitted August 7, 2018; Accepted February 22, 2019; Electronically published August 26, 2019 Online enhancements: appendixes, supplemental material. Dryad data: https://doi.org/10.5061/dryad.n07742q. abstract: fi 1994; Losos 2011), for example, in plants (e.g., stem-succulent Organisms inhabiting a speci c environment often have fi fi distinct morphology, but the factors that affect this fit are unclear when plants in deserts; Arakaki et al. 2011), sh (e.g., sh in general; multiple morphological traits affect performance in multiple behav- East African cichlids in particular; Winemiller 1991; Mus- iors. Does the realized morphology of a species reflect a compromise chick et al. 2012), and lizards (e.g., Anolis, Agamidae; Losos in performance among behaviors (i.e., trade-offs)? Or does many-to- 1990; Melville et al. 2006). Such studies typically focus on one mapping result in morphological distinctness without compromis- the ways that ecomorphs are specialized for their particular ing performance across behaviors? The importance of these principles ecology, yet often such distinct forms share aspects of ecol- in organismal design has rarely been compared at the macroevolution- ary scale. Here I study 191 species of frogs from around the world that ogy and behavior, making it unclear which ecologies and be- inhabit different microhabitats, using models of phenotypic evolution haviors determine body form (Robinson and Wilson 1998). to examine how form-function relationships may explain the fitbetween Moreover, many morphological characters have multiple ecology and morphology. I found three key results. First, despite being functional roles (Endler 1995; Walker 2007; Bergmann and distinct in leg morphology, ecomorphs were similar in jumping perfor- McElroy 2014), and finding how morphology differentially mance. Second, ecomorphs that regularly swim showed higher swim- affects performance in those roles—or whether a difference ming performance, which paralleled the higher leg muscle mass in these even occurs—can be difficult. taxa. Third, many-to-one mapping of form onto function occurred at all but the highest levels of both jumping and swimming performance. The Two themes of organismal design are often seen in nature seemingly contradictory first two results were explained by the third: when multiple concurrent functional demands are placed on when one behavior occurs in all species while another is restricted to organisms and those functions are determined by multiple a subset, many-to-one mapping allows species with distinct ecologies morphological variables (Wainwright 2007; Bergmann and to have distinct body forms that reflect their specialized behavior while McElroy 2014). First, trade-offs in the effects of morphology maintaining similar performance in a more general shared behavior. on performance in multiple behaviors may affect the rate and Keywords: anura, biomechanics, jumping, Ornstein-Uhlenbeck model, direction of phenotypic evolution; selection may favor one swimming. body plan over another, depending on which behavior most affects fitness and how morphology affects performance in that behavior (Walker 2007; Holzman et al. 2011). In other Introduction words, the realized morphology of an ecomorph is expected fi The study of ecomorphology has shown remarkably consis- to be the one that maximizes tness (Parker and Maynard tent and often convergent relationships between morphol- Smith 1990), even though that morphology would result in ogy and ecology across the tree of life (Wainwright and Reilly suboptimal performance for at least one other behavior (Al- faro et al. 2005). Second, when several morphological variables affect a sin- * Email: [email protected]. fi ORCIDs: Moen, https://orcid.org/0000-0003-1120-0043. gle function, multiple morphological con gurations can pro- Am. Nat. 2019. Vol. 194, pp. E81–E95. q 2019 by The University of Chicago. duce the same functional value or performance (Wainwright 0003-0147/2019/19404-58659$15.00. All rights reserved. et al. 2005; Marks and Lechowicz 2006; Wainwright 2007), DOI: 10.1086/704736 alleviating potential trade-offs (Walker 2007; Holzman et al. E82 The American Naturalist 2011; Bergmann and McElroy 2014). This many-to-one map- 1996; Nauwelaerts et al. 2007; but see Gillis and Biewener ping may free morphology to adapt to various selective de- 2000; Richards 2010). This means that it may be possible to si- mands without compromising current function (Alfaro et al. multaneously optimize performance in both behaviors with 2004, 2005; Wainwright 2007). The result is that while many the same morphology (e.g., muscular legs). Alternatively, it different morphological forms could achieve equivalent per- may be advantageous for some types of frogs (e.g., climbing ar- formance in one function, the realized morphology of a spe- boreal species) to have less muscular legs because muscle tis- cies will likely be based on its other functions. This possibil- sue is heavy, instead achieving high jumping performance ity has seldom been explored (Alfaro et al. 2005; Holzman through having longer legs. In sum, it remains unclear whether et al. 2011), possibly because the many different functions many-to-one mapping, trade-offs, both, or neither affects the of morphological traits are often unclear. Overall, trade-offs importance of life in multiple environments in the evolution and many-to-one mapping are both potentially important of the distinctive morphology of frog ecomorphs. for explaining the evolution of body form in many organisms Here I test models of phenotypic evolution and relate body because performance in many functions is frequently deter- form to function to explain the distinctive leg morphology of mined by multiple morphological variables (i.e., potential for different frog ecomorphs. I conducted extensive novel anal- many-to-one mapping) and many morphological variables in- yses on previously published data sets of microhabitat use, fluence multiple functions (i.e., potential for trade-offs; End- morphology, and performance from 191 species of frogs from ler 1995; Marks and Lechowicz 2006; Walker 2007). Thus, a 12 sites around the world. First, I modeled the dynamics key to disentangling these two principles is the study of how of morphological evolution to test the distinctiveness of leg form relates to functional performance (Losos 1990; Wain- morphology across different microhabitats. Second, I used wright and Reilly 1994; Koehl 1996; Losos 2011). Such stud- the same set of models to test whether high performance in ies can identify the consequences of differing morphology on jumping and swimming is equally vital across various micro- performance and the ecological context of that performance. habitats. Third, I estimated the evolutionary relationships be- Anuran amphibians (i.e., frogs and toads; “frogs” here- tween leg morphology and both jumping and swimming per- after, for brevity) are an excellent group for examining the formance to explain contrasting results for morphology and importance of many-to-one mapping and trade-offs in the performance in their relationship to microhabitat use. Over- evolution of ecomorphology. Most frog species can be cate- all, I show that many-to-one mapping alleviates potential gorized on the basis of microhabitat use (e.g., aquatic, arbo- trade-offs in the morphological variables that affect perfor- real, terrestrial; fig. 1; Moen and Wiens 2017), which is clearly mance in multiple behaviors, allowing species to morpholog- reflected by their morphology (Moen et al. 2016; Citadini ically specialize in one behavior (e.g., swimming) without sac- et al. 2018). Moreover, most ecomorphs occur around the rificing high performance in another (e.g., jumping). world (Bossuyt and Milinkovitch 2000; Moen et al. 2013, 2016; Citadini et al. 2018), and convergent ecomorphs have evolved many independent times (fig. 1; Moen et al. 2016). Material and Methods This fit between ecology and morphology presumably relates Sampling and Data Collection to the different behaviors of species in these different micro- habitats (e.g., climbing in arboreal frogs). However, many of Moen et al. (2013) measured the locomotor performance and these ecomorphs live in multiple environments (e.g., semi- morphology of 44 species of anurans from Australia, China, aquatic frogs regularly spend time both on land and in wa- and Colombia. They assigned species to four ecomorph cat- ter), which means that in at least some cases the morphology egories based on adult activity and nonbreeding behavior: of an ecomorph must affect performance in multiple differ- arboreal (living in trees), burrowing (digging their own bur- ent behaviors—some shared across ecomorphs (e.g., jump- rows), semiaquatic (living at the interface of land and water), ing; Gans and Parson 1966; Zug 1978; Marsh 1994; Shubin and terrestrial (spending nearly all adult life outside water). and Jenkins 1995) and others more specialized (e.g., swim- Subsequent work (Moen et al. 2016; Moen and Wiens 2017) ming, climbing, burrowing, and walking; Emerson 1988; Rey- added four additional categories, two of which pertain to the naga et al. 2018). Importantly, each of these different