How Universal Is Natural Selection?

Special Book Article How Universal Is Natural Selection? Randomness in Evolution. John Tyler Bonner. Princeton University Press, 2013. 148 pp., illus. $27.95 (ISBN 9780691157016 cloth). A mong evolutionary biologists, there has been considerable argument about how universal and powerful natural selection is. Is every phenotypic and genetic characteristic of every species an adaptation, honed by a history of selection? Many biologists, especially those who study organisms, are prone to an affirmative, “adaptationist” answer (e.g., Reeve and Sherman 1993). Others have objected. After lengthy controversy, Kimura’s (1983) neutral theory of molecular evolution was accepted as accounting for some evolution of DNA sequences; Gould and Lewontin’s (1979) famous critique of adaptationism, especially in the morphological realm, remains controversial; Lynch (2007) proposed that much of genome evolution has resulted from random processes (e.g., genetic drift). Randomness in Evolution and Relentless Evolution are a study in contrast between the neutralist and adaptationist positions. In Randomness in Evolution, John Tyler Bonner, who has contributed to developmental and evolutionary biology for more than six decades, builds on two of his previous books, Size and Cycle (1965) and The Evolution of Complexity (1988). He argues that the evolution of larger size necessitates increased complexity, which entails prolonged development, during which “internal selection” operates to expunge most mutations. Therefore, the morphology of “large” (apparently meaning multicellular) organisms is governed by natural selection. In contrast, he proposes that “small” organisms 64 BioScience • January 2014 / Vol. 64 No. 1 Bonner cheerfully agrees that this is his “just-so” story, complementing similarly speculative adaptationist stories, and that his hypothesis is difficult to test. Although his short essay is largely free of data—or indeed of evidence—he cites two major observations that he interprets as support for his idea. One is that many species of related, morphologically diverse protists (e.g., radiolarians, diatoms) coexist, and he cannot imagine that their morphologies are adaptations to different niches or natural enemies; that is, they are “neutral” with respect to one another. The other observation, based largely on the fossil record, is that some of these morphologies are very old, and he interprets the lack of change in such cases as “consistent with the idea that they might be neutral phenotypes” (p. 46). As far as I know (i.e., not far), little research has been done on the functional morphology of protists such as radiolarians or on their adaptations or their ecology. Nevertheless, although Bonner’s thesis is intriguing, I am highly skeptical of it. Consider his argument from the coexistence of morphologically different species, in view of model studies on large organisms. When studied in detail, species of such organisms often prove to differ subtly in their ecology (as MacArthur famously showed in 1958 for species of warblers), and slight morphological differences are often shown to be affected by selection. Do any comparable studies of protists bear on Bonner’s proposition? Some ecologists are prepared to accept Hubbell’s (2001) neutral model for the coexistence of hundreds of species of trees in tropical rainforests, but few botanists would argue that their diverse morphologies have evolved entirely by genetic drift. A critical point, moreover, is that selection operates through fitness differences between ancestral and derived characters within species lineages; whether it produces competitively equal species does not bear on the question. Hubbell (2006), in fact, developed a model in which large numbers of coexisting species evolve to be ecologically equivalent! Bonner’s argument from long-term stasis is even weaker, I believe, because stasis actually contradicts his hypothesis. Population genetic theory tells us that if mutation generates neutral variation in a character, the mean will vary (drift) over time; it cannot remain fixed in a finite population. Indeed, fossil lineages of radiolarians show phenotypic variance within populations and change in the mean over time (Kellogg 1975, Lazarus 1983). If a character is capable of varying (as it clearly is, if it varies among related species), stasis requires a stabilizing factor, and some form of selection must be the leading hypothesis. (Of course, selection may act through pleiotropic effects of genes, not necessarily on the observed character itself; see Dobzhansky 1956.) Although I think Bonner’s hypothesis is wrong, if he stimulates research on the causes of the exquisite—and http://bioscience.oxfordjournals.org Relentless Evolution. John N.Thompson. University of Chicago Press, 2013. 512 pp., illus. $35.00 (ISBN 9780226018751 paper). (specifically, unicellular eukaryotes) are largely free of such internal selection, because they do not undergo such development, and so their morphologies evolve nonadaptively—by mutation and random genetic drift. Special Book Article http://bioscience.oxfordjournals.org populations to persist. Moving to longer- term evolution, he ventures that ecological speciation (i.e., the evolution of reproductive isolation driven by divergent natural selection on populations in different environments) is the most frequent mode of speciation and is often very rapid, that adaptive radiations are caused by adaptation to diverse interspecific interactions and can emerge from the selection mosaic, and that certain processes of coevolution can lead to the formation of larger networks—or webs—of interaction among species. For example, he suggests, mutualisms among free-living species (e.g., plants and their pollinators) may act as “vortices” as more and more species become adapted, often convergently, to interact with a given species or set of similar species. Predators and prey may undergo “coevolutionary alternation,” a frequency-dependent process at the species level, in which predators shift and become adapted to new, less welldefended prey species as their primary prey evolve more-effective defenses—a process that may continue indefinitely as predators shift in their specialization from one prey species to another. Large webs foster the evolution of new ways of life, taking advantage of the resources constituted by sets of similar species. Some of these conclusions recast familiar propositions in evolutionary theory. It is well known, for example, that maladaptation may result from gene flow between divergently selected populations, and a consumer or mutualist is likely to specialize on a resource that is abundant or poses little defensive barrier to exploitation. The idea that different populations may adapt to different interacting species—and give rise to an adaptive radiation if they speciate—sounds more novel than it is. As a supporting example, Thompson writes that “the radiation of pr (...truncated)