Punctuation

In punctuated evolution (Eldredge and Gould, 1972) periods of relatively little change (“stasis”) are punctuated by episodes of relatively rapid change in the rate of evolution of a quantitative morphological trait, as seen in the fossil record of morphology. According to Simpson (1984), the term quantum evolution, refers to the same thing. Like Eldredge and Gould, Simpson contrasted quantum evolution with phyletic change, or sustained directional evolution without branching; considered that it could be associated with speciation (though also with phyletic evolution; p. 206); and even mentioned interrupted equilibra “In phyletic evolution equilibrium of the organism-environment system is continuous, or nearly so, although the point of equilibrium may and usually does shift. In quantum evolution equilibrium is lost, and a new equilibrium is reached”. I use the term “punctuation” rather than “quantum” because it less ambiguously describes change in rate of evolution. In its original meaning (from the Latin quantus), quantum means quantity. But quantum change, as mentioned by Simpson, is identified with the “quanta” of physics, which are discrete units of energy. This could encourage mistaken identification of punctuated change with the origin of discrete novelties, not the intended meaning of punctuated evolution, which is periodically altered rate of change in a continuously variable, quantitative trait. Mayr, Eldredge, Gould, and others (e.g., Stanley, 1979, 1981) explain stasis and punctuation in terms of speciation. Speciational punctuation hypotheses see stasis as due to the characteristics of established biological species, such as gene flow within interbreeding populations, large population size, heterogeneity of the species environment that retards directional change, developmental integration, canalization, coadapted genomes, stabilizing selection, and frequently reversing evolution over time within established species (Eldredge and Gould, 1997). These factors have been summarized by the term “gene-pool cohesiveness” (Mayr, 1989) or “developmental coherences” (Gould, 1989b), though the causes of stasis under the speciational hypothesis are admittedly vague and debatable (for reviews of other possible causes of stasis, see Williamson, 1987; Coyne and Charlesworth, 1997; Van Valen, 1982a; Spicer, 1993).

Assessment

A book on developmental plasticity needs a chapter on assessment, if only to show that adaptive environmental assessment occurs. Skepticism regarding the ability of nonhuman organisms to assess conditions well enough to make adaptive decisions has a long history in evolutionary biology, and it has been an important barrier to understanding the evolution of adaptive developmental plasticity. It is worth briefly reviewing this history in order to understand certain preconceptions about assessment that still persist. In the nineteenth century, critics of Darwin’s theory of sexual selection (Darwin, 1871) balked at the idea of an “aesthetic sense” in lowly creatures that would enable female choice of mates (representative papers are reprinted and discussed in Bajema, 1984). Later, the barrier persisted for other reasons. Even though naturalists routinely used the condition-appropriate expression of phenotypic traits to support adaptation hypotheses—a practice that assumes adaptive assessment of conditions as it is defined here—theoretically inclined biologists paid little attention to the question of facultatively expressed traits. Part of the difficulty lay in the problem of explaining how adaptive assessment could evolve within the framework of conventional genetics. Theodosius Dobzhansky, one of the twentieth century’s leading evolutionary biologists, acknowledged this unresolved problem in remarks following a lecture by J. S. Kennedy on the phase polyphenisms of migratory locusts (Kennedy, 1961). Dobzhansky referred to the “challenge to a geneticist” of explaining the adaptive switch between the sedentary and the migratory phenotypes of the locusts, which had been shown to be largely independent of genotype. He suggested that an extrachromosomal factor may be involved, a symbiotic microorganism that acts as a “plasmagene” whose multiplication would eventually stimulate phase change. Although Dobzhansky’s proposal was no more preposterous than some of the regulatory devices that have actually been discovered, Kennedy (1961) minced no words in his reply to this suggestion: . . . [W]e need not feel obliged to invoke a second organism to explain [phase polymorphism] unless we are reluctant to concede an important part to the environment as well as to heredity in moulding development. . . .

Cross-sexual Transfer

Distinctive male and female traits are perhaps the most familiar of all divergent specializations within species. In cross-sexual transfer, discrete traits that are expressed exclusively in one sex in an ancestral species appear in the opposite sex of descendants. An example is the expression of brood care by males in a lineage where ancestral females are the exclusive caretakers of the young, as in some voles (Thomas and Birney, 1979). Despite the prominence of sexual dimorphism and sex reversals in nature, and an early explicit treatment by Darwin, discussed in the next section, cross-sexual transfer is not often recognized as a major factor in the evolution of novelty (but see, on animals, Mayr, 1963, pp. 435-439; Mayr, 1970, p. 254; on plants, Iltis, 1983). When more widely investigated, cross-sexual transfer may prove to rival heterochrony and duplication as an important source of novelties in sexually dimorphic lineages. For this reason, I devote more attention here to cross-sexual transfer than to these other, well-established general patterns of change. The male and female of a sexually dimorphic species may be so different that it is easy to forget that each individual carries most or all of the genes necessary to produce the phenotype of the opposite sex. Sex determination, like caste determination and other switches between alternative phenotypes, depends on only a few genetic loci or, in many species, environmental factors (Bull, 1983). There is considerable flexibility in sex determination and facultative reversal in some taxa. Among fish, for example, there is even a species wherein sex is determined by juvenile size at a critical age (Francis and Barlow, 1993). The sex determination mechanism, whatever its nature, leads to a series of sex-limited responses, often coordinated by hormones and not necessarily all occurring at once. A distinguishing aspect of sexually dimorphic traits in adults is that there is often a close homology between the secondary sexual traits that are differently modified in the two sexes.

The Nature and Analysis of Phenotypic Transitions

Part II is about origins: how do new traits arise from old phenotypes? People of all ages are fascinated by the question of origins. Origins are the common concern of evolutionists and creationists, of ethnic historians, of Mormon geneologists and the Daughters of the American Revolution, of adopted children searching for their biological parents— indeed, of all who have wondered where Johnny got his patience, his sense of humor, or his big nose. Darwin was a clever publicist when he titled his most famous book The Origin of Species. He touched deep human chords by discussing not only the origin of species but the origin of marvellously complex morphological and psychological traits—specialized limbs, sexual behavior, intelligence, heroism, and the vertebrate eye, to mention just a few. Research on selection and adaptation may tell us why a trait persisted and spread, but it will not tell us where a trait came from. This is why evolutionary biology inevitably intersects with developmental biology, and why satisfactory explanations of ultimate (evolutionary) causation must always include both proximate causes and the study of selection. Novel traits originate via the transformation of ancestral phenotypes during development. This transformational aspect of evolutionary change has been oddly neglected in modern evolutionary biology, even though it is an integral part of human curiosity about origins in other fields. From classical mythology to modern-day childrens’ books, origins are explained in terms of transformations of the phenotype, alongside attention to developmental mechanisms and adaptive functions. Consider this excerpt from The Apeman’s Secret (Dixon, 1980), a Hardy Boys adventure book: . . . [T]he Apeman hated cruelty of any kind. Whenever he saw crooks or villians do something nasty to a helpless victim, he would fly into a rage. This would change his body chemistry and cause him to revert to the savage state. . . .

Modularity

Modularity, like the responsiveness that gives rise to it during development and evolution, is a universal property of living things and a fundamental determinant of how they evolve. Modularity refers to the properties of discreteness and dissociability among parts and integration within parts. There are many other words for the same thing, such as atomization (Wagner, 1995), individualization (Larson and Losos, 1996), autonomy (Nijhout, 1991b), dislocation (Schwanwitsch, 1924), decomposability (Wimsatt, 1981), discontinuity (Alberch, 1982), gene nets (Bonner, 1988), subunit organization (West-Eberhard, 1992a, 1996), compartments or compartmentation (Garcia-Bellido et al., 1979; Zuckerkandl, 1994; Maynard-Smith and Szathmary, 1995; Kirschner and Gerhart, 1998), and compartmentalization (Gerhart and Kirschner, 1997). One purpose of this chapter is to give consistent operational meaning to the concept of modularity in organisms. Seger and Stubblefield (1996, p. 118) note that organisms show “natural planes of cleavage” among organ systems, biochemical pathways, life stages, and behaviors that allow independent selection of different ones. They ask, “What determines where these planes of cleavage are located” and suggest that a “theory of organic articulations” may give insight into the laws of correlation, without specifying what the laws of articulation may be. Wagner (1995, p. 282) recognizes the importance of modularity and proposes a “building block” concept of homology where structural units often correspond to units of function, but concludes (after Rosenberg, 1985) that “there exists no way to distinguish an adequate from an inadequate atomization of the organisms.” Here I propose that modularity has a specific developmental basis (see also West-Eberhard, 1989, 1992a, 1996; see also Larson and Losos, 1996). Modular traits are subunits of the phenotype that are determined by the switches or decision points that organize development, whether of morphology, physiology, or behavior. Development can be seen as a branching series of decision points, including those caused by physical borders such as membranes or contact zones of growing or diffusing parts (e.g., see Meinhardt, 1982; see also chapter 5, on development). Each decision point demarcates the expression or use of a trait—a modular set—and subordinate branches demarcate lower level modular subunits, producing modular sets within modular sets.

One Final Word: Sex

Sex transforms life. It affects morphology and behavior. It diverts enormous amounts of time and energy from the business of survival. It can even distract from the manufacture and safe packaging of offspring. The adolescent metamorphosis we each experience once, and thereafter view with amazement, in the relative calm of adulthood, has swept through nature on a grand scale, culminating in orchid flowers and peacock tails. All of this is due to chromosomal recombination—sex sensu strictu (Ghiselin, 1974)—and its organismal result, sexual reproduction or cooperation between two individuals to produce offspring. It is sex as sexual reproduction, the developmental side of sex that initiates the ontogeny of new individuals, that I mainly discuss here, though it is sex as recombination— the genetic side of sex—that has received most attention in discussions of the maintenance of sex. Of all the major transformations in the history of life, the evolution of sex is the most enigmatic. The question is not so much how sex got there as why it remains. Given the importance of genetic similarity, or kin selection (Hamilton, 1964a,b), for the maintenance of cooperation within and among organisms, sex seems designed to be disruptive. It requires the union of genetically dissimilar individuals, which dilutes the relatedness of mother and young, leaving the mother to invest in offspring genetically only half like herself. This has been called “the cost of meiosis” or the “twofold cost of sex” (Williams, 1975). It is a cost that usually falls to females, with their greater investment in eggs and care of offspring. By this view, the male is a parasite of his mate and participation in sexual reproduction is contrary to the best interests of females, who would do better to reproduce parthenogenetically on their own. Yet, among animals, only about one in one thousand species are thelytokous, that is, secondarily asexual, with no facultative or alternating sexual generation and no interaction with males. The prevalence of sexual reproduction in higher organisms is “inconsistent with current evolutionary theory” (Williams, 1975, p. v).

Macroevolution

Macroevolution, or trans-specific evolution, refers to two different things in the literature on evolution. In discussions of phylogeny, it means phylogenetic branching pattern, or trends, seen at relatively high taxonomic levels (e.g., Stanley, 1979; Brooks and McLennan, 1991; Sober, 1993)—”any patterns that transcend species boundaries” (Lynch, 1991)—such as births and deaths of species and higher taxa and the shapes and diversity of radiations (Valentine, 1990). In discussions of evolutionary phenotypic transitions like those of part II, it means major phenotypic change (Lincoln et al., 1982). Rensch (1960) defined macroevolution as “evolution above the species level.” Microevolution, by contrast, is evolution below the species level, such as adaptive phenotypic and genetic change within populations, and geographic variation within a species. According to Simpson (1953a), the terms “macroevolution” and “microevolution” were invented by Goldschmidt (1940 [1982]), who also claimed that they involve different kinds of evolution. This problematic idea dates back to antiquity (see Rensch, 1960, for a concise review). The macroevolution problem, with emphasis on phylogenesis, was among other things (see Vuilleumier, 1984) behind the skepticism regarding Darwinism promoted by the enormously respected and influential French zoologist P.-P. Grassé. Grassé was convinced that the neo-Darwinian approach, with its emphasis on microevolution, cannot account for the primary features of evolution, namely, the large-scale diversification of life into major phylogenetic branches separated by unbridged gaps (e.g., see Grassé, 1973). This challenge echoes in the writings of many other critics of neo-Darwinism (e.g., Ho and Saunders, 1984; Gould and Eldredge, 1977; Gould, 1994; see also below), especially those who wish to contrast multilevel selection (including species selection) with microevolutionary theories (see Gould, 1999). The two macroevolution concepts, like the homology concepts discussed in chapter 25, are used interchangeably without sufficient attention to potential confusions. The result is needless controversy. The phylogenetic definition, for example, implies that macroevolution cannot, by definition, occur within species, for it refers exclusively to patterns above the species level. The phenotypic, major-change definition, on the other hand, can include processes within species.

Divergence without Speciation

Part II discussed the developmental origins of novelty in terms of how the phenotype is reorganized during evolution. It did not deal extensively with the problem of adaptedness during evolutionary transitions. How are we to explain transitions from one well-adapted state to another? Many still-influential discussions of adaptive shifts, such as Simpson’s (1944) treatment of quantum evolution and Wright’s (1932) discussion of shifting balance, associate change with fitness cost. Speciational theories of change depict change as dependent upon reproductively isolated populations in new environments. This chapter discusses divergence without reproductive isolation of novel forms, where the presumed cost of change is sidestepped because of the presence of adaptive options in the population undergoing change. Darwin’s solution to the problem of maladaptation during change was strict gradualism in monomorphically adapted populations. Darwin (1859 [1966]) reasoned that transitions between specialized adaptive states need not be disruptive if they were to occur by a series of small steps. Wright’s (1932) shifting balance is another solution to the same problem, but in Wright’s theory, change is initiated by a chance combination of genes that happens to suit a population to a new adaptive mode. Without a gradual adaptive change or a lucky gene combination, a shift between two peaks on Wright’s adaptive landscape would imply passing through a valley of inferior adaptedness. Alternative phenotypes offer a third kind of solution, one that requires neither strict gradualism in a monomorphic population nor chance genetic events. In species with alternative phenotypes, a recurrent novelty that happens to prove advantageous to some individuals or in some circumstances can be refined via gradual genetic accommodation as an optional trait. Since this involves developmental diversification, not transformation or loss of existing traits, the new option develops as a new specialization alongside old ones. Shapiro notes that conditional expression of alternative phenotypes is a way of having two adaptive specializations “without carrying a genetic load,” or a cost of genotypes that oblige expression of phenotypes less favorable than the fittest one.

Recurrence

Recurrent phenotypes are similar or identical phenotypic traits with discontinuous phylogenetic distributions, which owe their similarity to common ancestry (homology). A recurrent trait may be found as a fixed trait, as an alternative phenotype (one morph of a polymorphism or polyphenism), or as a low-frequency developmental anomaly. Recurrence, then, is the phyletically disjunct appearance of homologous traits. An example is the repeated evolution of larviform (paedomorphic) adults in salamanders. The larviform morph is characterized by retention in the reproductive stage of homologous larval traits such as external gills and a tail. This has involved changes at various points in the hormonal mechanism that controls metamorphosis in all salamanders (chapter 25), perhaps under selection for accelerated reproduction in stressful environments (Whiteman, 1994). As is characteristic of recurrent phenotypes, the occurrence of the reproductive larviform adult morph varies in frequency from one species of salamander to another: it can be absent, an anomaly (<5% of population), a common (>5%) alternative to complete metamorphosis, or a predominant or fixed form. Even within the genus Ambystoma, the unmetamorphosed larviform adult occurs as an occasional anomaly in some populations, as a facultatively expressed alternative phenotype in others (e.g., A. tigrinum) and as a fixed form in others (e.g., A. dumerilii; Collins et al., 1993). All atavisms and reversions (see chapter 12) are examples of recurrence. Discontinuity of expression is expected in combinatorial evolution, where traits are turned off and on and expressed in different combinations due to regulatory change. The growing evidence of homoplasy in phylogenetic studies is important evidence that combinatorial evolution occurs and that homoplasy itself is worthy of study, not just a source of “noise” in cladistics (Wake, 1996a). Homoplasy has been defined as “possession by two or more taxa of a character derived not from the nearest common ancestor but through convergence, parallelism, or reversal”. More simply, homoplasy is the recurrence of similarity in evolution (Sanderson and Hufford, 1996).

Heterotopy

Heterotopy is the spatial analogue of heterochrony: it is evolutionary change in the site of expression of a phenotypic trait. Gould (1977) attributes the word “heterotopy” to Haeckel, who used it in a more specialized way, to mean evolutionary change in the germ layer from which an organ differentiates. Wray and McClay (1989, p. 810) list several examples of heterotopy, including the origin of muscles in tetrapod forelimbs from different somites, the origin of vertebrate primordial germ cells from different germ layers, and homeotic “heterotopic” mutations that transfer appendages from one body segment to another. A broad definition of heterotopy extends the concept to include spatial patterning, not only transposition from one location to another, but spatial organization of quantitative processes such as growth (Zellditch et al., 1992) or the location of precursors during the development of homologous traits (Wray and McClay, 1988, p. 313). As in other categorizations of transitions, heterotopies could as well be classified in other ways, such as duplication. Severtzoff signaled a general relationship between heterochrony and heterotopy when he wrote that “heterochrony in development is a means of topographic coordination; i.e., new adaptation of the parts to each other.” Many morphological heterochronies in plants produce heterotopic change, since the morphological ontogeny of a plant is recorded in its adult architecture. Thus, changes in timing of expression of juvenile and adult leaf forms result in heterotopic change in architecture of the mature plant, with the juvenile leaves appearing high on the stem, rather than only basally as before. A clear and oft-described example of environmentally mediated heterotopic change was demonstrated in early experiments on melanization in the Himalayan rabbit (Sturtevant, 1913; Iljin, 1927; Iljin and Iljin, 1930; see discussions in Schmalhausen, 1949 [1986]; Huxley, 1942; Levinton, 1988). In the Himalayan rabbit, as in the Siamese cat, pigment normally develops only in the extremities, where skin temperature is below the general body temperature. Individuals raised at temperatures above 30°C develop white extremities.