1. Introduction

Over the past decade or so a new discipline—evolutionary developmental biology or “evo-devo”—has been gaining visibility, largely because of its emphasis on the emergence of morphological novelty, not through the acquisition of de novo mutations, but through a “reshuffling” and/or activation or deactivation of existing genes and regulatory molecules that yields different signal transduction pathways (e.g., Raff Reference Raff1996; Schwartz Reference Schwartz1999a; Carroll, Grenier, and Weatherbee Reference Carroll, Grenier and Weatherbee2005). This construction of morphological (thus evolutionary) change and the genetics underlying it contrasts markedly with the traditional view of neo-Darwinism, in which mutation is viewed as always introducing new genetic material, and that, since there is assumed to be a direct connection between specific “genes” and an organism's form, small but ongoing changes at the genetic level will eventually transform “something” into “something else” (e.g., Morgan Reference Morgan, Sturtevant, Muller and Bridges1916, Reference Morgan, Sturtevant, Muller and Bridges1925; Dobzhansky Reference Dobzhansky1941; Mayr Reference Mayr1942). In addition, this model subsumes others: It is expected that (1) a mutation in one “direction” will be followed by another that continues the trajectory of change, and (2) geographical isolation and various kinds of selection eventually lead to reproductive isolation. One can summarize a major difference between evo-devo and neo-Darwinism thusly: In the former, morphological or evolutionary change results from reorganization internal to the organism, whereas, in the latter, external circumstances provide the basis for as well as orchestrate the details of both genetic and morphological change. One can also contrast evo-devo and neo-Darwinism with regard to the tempo with which change might occur. In neo-Darwinism, gradualism (slow, continuous transformation) on both the genetic and morphological levels is the rule, while an evo-devo perspective can lead to a model in which the spread of the potential for change at the “genetic level” would, of course, be continual (but not necessarily slow), while the emergence of morphological novelty would be abrupt (Schwartz Reference Schwartz1999a,Reference Schwartz1999b; Maresca and Schwartz Reference Maresca and Schwartz2006).

Interestingly, one can trace the essence of this intellectual difference—at least the different emphases on the “internal” versus the “external”—to the debate between Darwin (Reference Darwin1872) and Mivart (Reference Mivart1871) (see Schwartz, in press) and to the continued emphasis on “internal reorganization” leading to morphological novelty in the work of Bateson (Reference Bateson1894), Goldschmidt (Reference Goldschmidt1940), and Schindewolf (Reference Schindewolf1950). Here I shall focus on Thomas Hunt Morgan, who, more than anyone else in the early twentieth century, was in a position to combine an emergent understanding of genetics with a model of morphological and evolutionary change that would have foreshadowed by decades the rise of present-day evo-devo thinking.

2. Morgan: Anti-Darwinian and Anti-Mendelian

In the late nineteenth and early twentieth centuries, the embryologist Thomas Hunt Morgan (Reference Morgan, Sturtevant, Muller and Bridges1903, Reference Morgan, Sturtevant, Muller and Bridges1909) was a de Vresian mutationist/saltationist who embraced neither Mendelism nor Darwinism. He distrusted Mendelism because of the ease with which he thought scientists could invoke “factors” to explain one or another morphology. With regard to Darwinism, he was not convinced of the power of natural selection to provoke change, the gradual nature of evolution change, a “survival of the fittest” argument, or the applicability of artificial selection to an understanding of natural processes. As a saltationist/mutationist, and one who came from a background in the study of development, Morgan thought that sudden organismal change did occur, although not as frequently as the emergence of minor variations within species, and that change that is “perfectly definite and permanent in character” derives from factors internal to the organism (Morgan Reference Morgan, Sturtevant, Muller and Bridges1903, 339). Morgan (ibid.) provided multiple examples of the sudden origin of novel features in plants and animals (including the “monstrosities” Darwin [Reference Darwin1868] rejected as biologically relevant, e.g., ancon sheep, niata cattle) without transformation involving series of intermediate stages. In one case, the ancon race of sheep, Morgan described how the entire race, which is characterized by short, crooked legs and a long back, descended from the crossing with common sheep of the one ram that had emerged with this condition. In addition, when ancon sheep were crossed with other breeds of sheep, the offspring were typically of one or the other breed morphologically—not of an intermediate morphological condition, as Darwin (e.g., Reference Darwin1868, Reference Darwin1872) argued would have occurred, even though there was no evidence for this assertion.

Morgan also rejected the Darwinian practice of constructing transformation series of extant organisms from simple to complex as a surrogate for gradual evolutionary change: “It does not follow because we can arrange such series without any large gaps in its continuity, that the more complex conditions have been gradually formed in exactly this way from the simplest conditions” (1903, 415).

Morgan did not discuss fossils or the fossil record except in a few instances. When he did, however, it is clear that he accepted the sudden appearances of taxa, and the gaps between them and their presumed relatives, as real. For example, his quote (1903, 315) from de Varigny: “In Paraguay, during the last century (1770), a bull was born without horns, although his ancestry was well provided with these appendages, and his progeny was also hornless, although at first he was mated with horned cows. If the horned and the hornless were met in a fossil state, we would certainly wonder at not finding specimens provided with semi-degenerate horns, and representing the link between both, and if we were told that the hornless variety may have arisen suddenly, we should not believe it and we should be wrong.”

Morgan concluded his monograph of 1903, Evolution and Adaptation, thusly: “If we suppose that new mutations and ‘definitely’ inherited variations suddenly appear, some of which will find an environment to which they are more or less well fitted, we can see how evolution may have gone on without assuming new species have been formed through a process of competition. Nature's supreme test is survival. She makes new forms to bring them to this test through mutation, and does not remodel old forms through a process of individual selection” (464).

Given his clear rejection of Darwin's model of evolution, it would have been difficult to predict that, only seven years later, Morgan (Reference Morgan, Sturtevant, Muller and Bridges1910) would also reject his original field of embryology, along with comparative anatomy and paleontology, as having anything useful to contribute to the study and understanding of evolution. The latter endeavor was, he declared, only possible through the study of genetics, and it was through his experiments in fruit-fly population genetics that were predicated on Mendelian principles of inheritance (Morgan et al. Reference Morgan, Sturtevant, Muller and Bridges1915) that he convinced himself that Darwin's model—change occurring through the gradual accumulation of minor variations—was correct (e.g., Morgan Reference Morgan, Sturtevant, Muller and Bridges1916, Reference Morgan, Sturtevant, Muller and Bridges1925, Reference Morgan, Sturtevant, Muller and Bridges1935).

3. Toward Neo-Darwinism

One of the major experiments that influenced Morgan's theoretical shift was that in which his colleague, Muller, over a period of three years, gradually altered the average wing length of a fruit-fly colony from slightly longer than to slightly shorter than body length. The experiment began with a mutant whose wings were shorter than the colony's average normal wing length (Morgan Reference Morgan, Sturtevant, Muller and Bridges1916). Morgan (Reference Morgan, Sturtevant, Muller and Bridges1916) also cited a study in which crosses between a short- and a long-cobbed “race” of Indian corn (with nonoverlapping ranges of cob-length variation) produced an F₁ generation intermediate in cob length, and an F₂ generation in which the range of cob length spanned that of the P₁ generation.

As pointed out above, Morgan maintained that there was no evidence that a mutation producing a change or variation in a feature would be followed by another that affected the same trait—like tossing pennies, “the number of heads obtained has no influence on the number of heads that will appear in the next throw” (Morgan Reference Morgan, Sturtevant, Muller and Bridges1916, 190). Nevertheless, he now mounted an argument that would allow change to occur along a certain course: “Owing to this property of the germ plasm to duplicate itself in a large number of samples not only is an opportunity furnished to any advantageous variation to become extensively multiplied, but the presence of a large number of individuals of a given sort prejudices the probable future result” (Morgan Reference Morgan, Sturtevant, Muller and Bridges1916, 192). That is, and now in contradistinction to the reality of a coin toss, an increase in number of individuals bearing a beneficial trait can increase the likelihood that another mutation could or will arise that will also affect that trait. As he (1925, 150) declared elsewhere: “Evolution once begun in a given direction is in a favorable position to go on in the same direction rather than another, so long as the advance does not overstep the limit where further change is advantageous.”

What is interesting about Morgan's pursuit of this line of reasoning is that, from 1916 on, the illustration he provided of his version of directional evolutionary change, taken from an uncited work by the paleontologist R. S. Lull, is of reconstructions of snouts of differing lengths of extinct as well as recent proboscideans. How could Lull, or anyone for that matter, reconstruct with accuracy the length of an extinct proboscidean's trunk, which would have been soft tissue? Since the source of such speculation is paleontological, it would have to be based on time-sequencing fossils and, assuming that this endeavor represented a phylogenetic sequence, assume further that “trunkless” was the starting point of a transformation series that resulted in the extant long-snouted elephant. The fact that Morgan also succumbed to this series of assumptions is curiously at odds with his very own criticism of paleobiological practice (1925, 1–3):

Is, however, Morgan's explanation of why evolution would continue along a course that began essentially by chance any different than explanations he criticized paleobiologists for offering? I think not.

4. The Reconfiguring of Observation

As early as 1916 (86–87), Morgan began to articulate a model of gradual evolution: “Evolution of wild species appears to have taken place by modifying and improving bit by bit the structures and habits that the animal or plant already possessed.”

In 1925 (38–39), Morgan explicated further his model:

Although Morgan converted himself to a melded version of Darwinism and Mendelism (reducing the idea of discontinuity that characterized Mendelism to the graded continuity of Darwinism by diminishing the effect of mutation [Schwartz Reference Schwartz1999a]), how he came to that place was odd and unexpected. The observations available to him can be summarized thusly: First, mutations were spontaneous. Second, mutations produced all sorts of differences in fruit flies, for example, in wing shape and length; wing venation patterns; position and morphology of antennae; thoracic segment number and duplication, presence, absence, and number of thoracic bristles; eyelessness; and eye color. As he (Reference Morgan, Sturtevant, Muller and Bridges1916, 117) described the situation: “some of the changes [are] so slight that they would be overlooked except by an expert, others so great that in the character affected the flies depart far from the original species.” Importantly, as he emphasized throughout his many publications, the mechanism of altering genes (of mutation) was the same whether the effect proved, at least to the observer, to be profound or trivial. Third, many if not most mutations “appeared independently several times” (Morgan, Reference Morgan, Sturtevant, Muller and Bridges1916, 107). Fourth, many if not most of the mutations were in the recessive state and not all injurious when expressed (observations that Bateson [Reference Bateson1913] had also noted) (i.e., what were perceived as mutations behaved, following Mendelian rules of segregation, in such a way as to be expressed phenotypically in a manner consistent with the concepts of “recessive” and “dominant”). Fifth, being in the recessive state, a mutation (mutated or altered gene or factor) spread silently and unexpressed, increasing the number of heterozygotes until they were sufficiently numerous to produce progeny with the feature in question. Sixth, the variants or “conditions” (including striking conditions such as bithorax or absent or vestigial wings or eyes, or more benign conditions such as absence of two bristles on the side of the thorax) arise “at a single step” (Morgan, Reference Morgan, Sturtevant, Muller and Bridges1916, 109). Seventh, more than one individual with the “mutated” condition appears at a time. Eighth, even though many of these conditions are profoundly different from the parental norm, they breed true. Ninth, individuals that are profoundly different from the normal or wild condition may be crossbred with normal individuals. Tenth, the progeny of these crosses are not always sterile and usually display either the mutated or normal condition rather than an intermediate one.

These observations lead not to a model of gradual evolutionary change, with genotypes and phenotypes slowly changing in tandem, but to a model wherein the spread of a mutation in the recessive state may occur over several or more generations, while its phenotypic expression is sudden or abrupt. Within this framework also lies the realization that, whatever the underlying mechanism of genetic alteration or mutation, it applies equally to large- and small-scale change. Thus, the model of change that emerges from Morgan's accumulated facts accounts for cases of minor variation (such as thoracic bristle number in fruit flies) as well as for cases of profound morphological difference (such as stunted wings). Furthermore, Morgan's data reveal that, while offspring of crosses between “normal” and “profoundly different” individuals may be sterile, this is not a consistent or even expected occurrence. In short, whatever typically happens in the advent and spread of a minor variation, the same applies in the advent and spread of a major change. Consequently, it is not necessary, as Goldschmidt (Reference Goldschmidt1940) and others (e.g., Stanley, Reference Stanley1975) have argued, to invoke different genetic or mutational processes to deal with individual variation, on the one hand, and intertaxic differences, on the other. Rather, evolutionary change can be conceived as resulting from normal processes of mutation and phenotypic expression, not the accumulation of infinitesimally small variations. This having been said, the matter still remains as to what criteria one uses to identify a particular change as being evolutionarily significant—and this begs the questions of how one identifies species and whether species are, as has been traditionally believed, the “unit” of evolution, a question I shall not address here (but see discussion in Schwartz, Reference Schwartz1999c).

Why, then, did Morgan not embrace the obvious implications of his data? As is clear in one of the quotes above, Morgan's (Reference Morgan, Sturtevant, Muller and Bridges1925, 38) position was that, because “animals and plants are extremely complex machines that are highly adapted to the conditions of life in which they live … [a]ny change, and especially any great change in them, is far more likely to throw them out of balance with their environment than to bring to them an advantage.” Consequently, since something has to change in order for what we call evolution to occur, a slight change would appear to be preferable to a large one because it would be less likely to endanger the “fit” of an organism to its circumstances. No doubt experiments, such as Muller's shifting (interpreted in an evolutionary context as reducing) average wing length in a fruit fly colony, or the production of corn of graduated lengths, appeared to justify this conclusion. Hence, while not denying observations (including his own) of large-scale change, Morgan ignored them and restricted the playing field of evolution to the small end of the available spectrum of morphological (and thus also genetic) change.

Morgan's argument for preferring slight over profound change is, however, interesting in light of his general position on the relationship of an organism via its morphology to its environment. For instance, as he wrote in Evolution and Genetics (Morgan Reference Morgan, Sturtevant, Muller and Bridges1925, 150–151):

In this passage, as elsewhere in his publications, Morgan makes clear his position that the state of an organism is determined by internal, not external, factors. No doubt, this reflects his background in development and embryology, which he extended to embrace the concept of mutation. Although Morgan (e.g., Reference Morgan, Sturtevant, Muller and Bridges1916) acknowledges that morphology can be altered by external factors, such as heat, permanent change via mutation is an intrinsic property of organisms. In this way, he continues to keep from fully embracing Darwinism, which promotes the influence over the organism of that which lies “outside” it. Morgan also breaks with Darwinism in seeing the relationship between the organism and the environment in which it finds itself merely as fortuitous, not the result of eons of fine tuning an organism's morphology to fit its surroundings. In the Darwinian scenario, when the environment changes, the organism changes in response to it. Although Morgan did propose a model of gradual change, there is no causal relationship between change and the environment. Organismal change occurs only if the change is beneficial, or at least not harmful, to the organism. It is by chance that organisms find themselves in certain environments, and it is by chance that any change in their biology occurs. Within Morgan's framework, then, the chance of change proceeding along any perceived morphological continuum results from an organism having randomly started down it (which, in this scenario, is then supposed to increase the likelihood of change in that direction persisting). Thus, to use Morgan's example of proboscidean snout elongation, elephants have a long trunk because it did not interfere with their survival. Clearly, within this framework, a slight change is less likely to throw an organism out of whack with its surroundings than a more profound change. Consequently, even though most of Morgan's evidence of mutation pertained to large-scale changes from parent to progeny, his assumptions constrained his model of evolutionary change to a minor component of his data. At base, though, Morgan and de Vries's (Reference de Vries1910) mutation theories differed only with regard to the magnitude of change. In both theories, the theme was the same: Mutation provided the novelty upon which selection, however conceived, could then act.

Although Morgan did not articulate this, an expectation of his opting for change via the persistence of small mutation-based changes is that, as Darwin (e.g., Reference Darwin1859) had speculated, evolutionary change is additive or accumulative. As Morgan did characterize it, Darwinian evolutionary change emerges by the molding or reshaping of one form into another. With regard to the underlying genes, although Morgan (Reference Morgan, Sturtevant, Muller and Bridges1925) speculated that new genes are “added” to the genotype, he was quite forward thinking in various ways. For instance, he suggested that there might not be substantive difference between a dominant and a recessive gene, that mutant genes may be derivatives of already-present genes, and that genes can contribute to changes in the original feature (rather than there being a gene for a specific character).

Another interesting aspect of Morgan's theorizing is that, although his background in embryology led him to place the impetus or source for organismal change (via mutation) within the organism, it did not lead him to discuss evolution in terms of the development of, and developmental constraints upon, the organism. To the contrary, his experiments in fruit-fly population genetics and his discussion of results and their implications bypassed entirely the “unfolding” of the organism. Rather, as Darwin and thereafter population geneticists did (see review in Schwartz Reference Schwartz1999a, Reference Schwartz, Mueller-Wille and Reinberger2005a), Morgan emphasized the adult: What features does it have? How these features “got there” was not discussed in terms of any understanding of inter- or intracellular or any other kind of developmental interactions—in spite of the fact that Morgan (Reference Morgan, Sturtevant, Muller and Bridges1916, Reference Morgan, Sturtevant, Muller and Bridges1925) did speculate that the basis of organismal development was probably more chemical than not. However, beyond identifying the features that adults of a particular generation had, and seeing if they were similar to or different from features of the parental generation, the only other element of major concern to Morgan was the transmission of features: Who passed on what to whom? And the focus of the study of transmission is the adult. The only conceivable nod to development that a population geneticist like Morgan might make is acknowledging that adults pass on half their genetic material to form a zygote. Beyond that, however, Morgan's population genetics was a study of adults giving rise to adults—which continues to be the focus of neo-Darwinism, as it also was for Darwin (e.g., Reference Darwin1859, Reference Darwin1872) (see discussion in Schwartz Reference Schwartz, Mueller-Wille and Reinberger2005a).

5. Conclusion

Even while dismissing the significance of his experiments’ demonstration that large-scale morphological difference (a manifestation of large-scale mutation) arises abruptly and is transmitted to subsequent generations in the same fashion as small-scale morphological difference (a manifestation of small-scale mutation), without detrimental effects to the affected organism, either in terms of survival or reproductive viability, Morgan chose small- over large-scale change as the basis for organismal transformation. Thus, rather than recognizing that his very own observations provided a model of evolutionary (large-scale) change that was accommodated by the same mechanism of inheritance that applied to between-individual (small-scale) variation, Morgan constructed an argument against the relevance of the former for understanding the emergence of organismal novelty: Large-scale change would adversely affect the biological integrity of the organism; small-scale change would not.

But although he rejected large-scale mutation and change (and no doubt the saltationist model it suggested) in favor of a model of gradual change, Morgan did not fully reiterate Darwinism. He rejected “struggle for existence” and focused the basis of change on mutation, rather than on natural selection. Nevertheless, his unsubstantiated argument that evolutionarily significant change is only enacted through small-scale morphological changes, whose underlying small mutations would continue to emerge in a specific trajectory, would become the basis of the version of neo-Darwinism promoted in the evolutionary synthesis (e.g., Fisher, Reference Fisher1930; Dobzhansky Reference Dobzhansky1941; Mayr Reference Mayr1942). The synthesis did not, however, incorporate Morgan's ambivalences: whether “internal” and “external” forces direct change, mutation always adds “new” genes rather than altering already-present genetic material, and whether there were genes “for” specific structures. Emphasis on the “external” provoking change in the “internal” (the “vacuum effect” [Schwartz Reference Schwartz and Meyer2005b]), on mutation producing “something new,” and on there being genes “for” specific structure (“bean bag”) were specific to the synthesis.

But while population genetics and “bean-bag” thinking may have been historically important to the synthesis in providing the biological sciences with a common language, discoveries in developmental biology and genetics continually contradict its assumptions. For example, the notion of continual genetic change (producing continual variation) is contravened by the existence of DNA repair mechanisms (e.g., via heat shock or stress proteins) that seek to maintain “DNA homeostasis” (Maresca and Schwartz Reference Maresca and Schwartz2006), and that of “genes for structure” undermined by processes such as alternative genome splicing and frame reading (e.g., Ast Reference Ast2005; Stotz, Reference Stotz2006, in this issue).

With specific regard to Morgan, his suggestion that heat (temperature) or other external or environmental factors might play a role in provoking mutation, yet mutation remains an “internal” aspect of the organism, is borne out by studies on membrane physical states and heat shock (stress) protein response (see review in Maresca and Schwartz Reference Maresca and Schwartz2006)—which also lead to a theory of evolutionary change that follows from Morgan's (as well as Bateson's Reference Bateson1913) recognition that “mutation” arises in the “recessive” state (Schwartz Reference Schwartz1999a). That is, however one defines or conceives of “mutation” (e.g., due to improper DNA repair or protein folding, or to their larger effects via altering signal transduction pathways and consequent cascading influences on physical and epigenetic aspects of development), it will arise in the “recessive” (inactive) state and spread silently through the population until there are a sufficient number of heterozygotes to produce homozygotes. The spread of the potential for novelty may be relatively gradual, but the appearance of the novelty (whether or not it allows its bearers to survive) will be sudden and in multiple individuals (Schwartz Reference Schwartz1999a; Maresca and Schwartz Reference Maresca and Schwartz2006).

In short, some of the major intellectual underpinnings of evo-devo thinking were available to Morgan, but, in spite of his embryological background, he ignored their potential evolutionary significance. Thus, rather than furthering attempts to integrate development into a model of morphological change (e.g., Bateson Reference Bateson1894, Reference Bateson1913; de Beer Reference de Beer1930), breaking ground for others (e.g., Goldschmidt Reference Goldschmidt1940), and thus setting the stage for modern evolutionary developmental biology, Morgan contributed to a program in evolutionary biology—neo-Darwinism and the synthesis—that while profoundly affecting scientific inquiry, served at the same time to suppress the pursuit of alternative theories and to excoriate those who dared question what had quickly become received wisdom (Schwartz Reference Schwartz1999a).