The Structure of Evolutionary Theory

by Stephen Jay Gould

  7. Darwin erred in advocating natural selection for the origin of species. The Mutation Theory has corrected this basic mistake, and therefore repre­sents a novel direction for evolutionary thought. Darwin deserves our highest praise for his historical role, but he has now been superseded:

  The theory of descent aims at the scientific explanation of systematic relationship. It is Darwin's immortal service to have obtained general recognition for this generalization. By doing this he revolutionized the whole of biological, systematic, embryological and paleontological sci­ence. Tapping inexhaustible sources for new investigation and discover­ing everywhere mines where new facts were to be had for the picking up. The several propositions and hypotheses which Darwin employed as supports for this theory should be regarded now only as such, since their interest is mainly historical. They have served their purpose and are thereby fully justified... This is especially true of the theory of selec­tion, which now has served its time as an argument for the theory of descent; happily this theory no longer stands in need of such support (1909a, volume 1, p. 28).

  8. The theory of natural selection is erroneous. The theory of mutation is correct. “The mutation theory is opposed to that conception of the theory of selection which is now prevalent. According to the latter view the material for the origin of new species is afforded by ordinary or so-called individual varia­tion. According to the mutation theory individual variation has nothing to do with the origin of species. This form of variation, as I hope to show, cannot even by the most rigid and sustained selection lead to a genuine overstepping of the limits of the species and still less to the origin of new and constant char­acters” (1909a, volume 1, p. 4).

  One might conjecture that this full range of viewpoints represents a transi­tion from heart (in the early entries of the sequence) to mind — that is, from de Vries' need to express fealty with his personal hero to his recognition of the oppositional logic within his own system. (One might also be more cynical and interpret the early entries as diplomatic attempts to court favor with Dar­winians — but I don't think that this interpretation can be fairly defended, for [Page 443] Darwinism did not rank as a dominant philosophy at the time, and issues of heart hold such evident prominence for de Vries.) The Mutation Theory, with its explicitly saltational and nonadaptational origin of species, must be read as a confutation of Darwinism on the central question (the title of Darwin's book, after all) of “The Origin of Species.” And yet, by an interesting argu­ment (developed in the next section), de Vries did provide a genuine and am­ple field for Darwinian logic in another realm, even while he tried to extirpate natural selection without compromise on its own original turf.

  The logic of Darwinism and its different place in de vries' system. I have documented the psychological vacillation in de Vries's assessment of Darwin, but a stark contrast must be drawn between this frustrating emotional indefiniteness and de Vries' clear and subtle under­standing of selectionist logic. I think that only two other early evolutionists — Weismann and Darwin himself — ever grasped so fully, both in basic logic and expanding implications, the rich meanings of selectionism. De Vries' per­sonal dilemma lay in his unwillingness to tar his personal hero with the brush of selectionist errors (in his judgment), not in any softness or vacil­lation in understanding selection itself. Thus, he tried to distance Darwin from Darwin's own beliefs, grasping at straws (often of de Vries' own con­struction) in tortuous exegetical efforts to remake Darwin as a closet saltationist.

  One can hardly deny that de Vries' Mutation Theory represents, in princi­ple, about as anti-Darwinian a mechanism as anyone could construct at the crucial level of Darwin's own concerns (and chosen book title) — the origin of species. Neither selection nor adaptation can play a creative role in evolution­ary change if new species arise in single, fortuitous leaps.

  Yet de Vries insisted that his theory followed Darwinian principles at the larger scale of full unrolling of life's history — and here he presented a sound and fascinating argument that his contemporaries never understood (in their failure to grasp the generality of selectionist logic) and that later history therefore, and unfortunately, forgot. I argued in Chapter 2 that the operation of a selectionist mechanism makes three crucial demands upon the nature of internal “raw material”: that variation be copious, small in scope (relative to the unit of incremental change at the scale under consideration), and undi­rected (isotropic). At the level of speciation, de Vries' Mutation Theory be­comes decidedly anti-Darwinian by failing the second test — for single muta­tions generate new species in one step, and no creative role can be assigned to selection or adaptation at all.

  But suppose that we “promote” our gaze and consider evolutionary trends through geological time as the relevant scale of change. Then a species-form­ing mutational step might be considered sufficiently small (relative to the full trend) to fit into Darwinian logic — though not into Darwin's own theory, which explicitly requires, as its central tenet, that a process of organismal se­lection must govern the origin of species.

  But if we can regard speciational steps as small increments in macroevolution, then the applicability of Darwinian logic to trends would depend [Page 444] upon the validity of the remaining two criteria at this higher level: copious­ness and isotropy. The criterion of copiousness will surely be fulfilled in lin­eages undergoing a de Vriesian period of mutability. The full operation of Darwinian logic at this larger scale will then stand or fall upon the remaining criterion of isotropy. De Vries analyzed the issue in these appropriate terms and, by stating strong support for isotropy. He therefore identified a proper fealty with Darwinian logic (in a domain different from Darwin's own appli­cation, after denying the efficacy of the same general logic in Darwin's own favored realm).

  De Vries often states the principle of isotropy as one of his central conclusions: all characters may mutate in all directions. “Single variations [mutations] seem to be presented by all characters, to proceed in every direction and to be apparently without limit” (1909a, volume 1, p. 33). Moreover, de Vries expresses this view not only as an empirical conclusion but as a logical consequence of his theory: “The mutation theory demands that organisms should exhibit mutability in almost all directions” (1909a, volume 1, p. 204).

  Even more significantly, de Vries recognizes that isotropy must be asserted to validate Darwinian selection at the higher level of evolutionary trends. Tying isotropy directly to the efficacy of selection, de Vries writes (1905, p. 574): “Nearly all qualities vary in opposite directions and our group of mutants affords wide material for the sifting process of natural selection.” Se­lection can only operate as a sieve, but if variations are copious, isotropic and small (a good description for the status of de Vriesian mutations relative to the full extent of a geological trend), then these species-forming mutations can forge no large-scale cladal trend by themselves — and even a sieve, by ex­tended directional accumulation, becomes a creative mechanism. (This meta­phor beautifully restates Darwin's basic argument about the creativity of se­lection — see Chapter 2. But de Vries denies selection at Darwin's own favored level of organisms in populations, and grants power to Darwin's mechanism only at the higher level of sustained trends among species in clades.)

  According to Darwin, changes occur in all directions, quite indepen­dently of the prevailing circumstances. Some may be favorable, others detrimental, many of them without significance, neither useful nor injurious. Some of them will sooner or later be destroyed, while others will survive, and which of them will survive is obviously incumbent on the question, whether their particular changes agree with the existing conditions or not. This is what Darwin has called the struggle for life. It is a large sieve, and it only acts as such. Some fall through and are annihi­lated, others remain above and are selected, as the phrase goes. Many are selected, but more are destroyed: daily observation does not leave any doubt upon this point. How the differences originate is quite another question. It has not
hing to do with the theory of natural selection, nor with the struggle for life (1905, p. 571).

  De Vries strongly rejected any notion of directed variability (nonisotropy) in the production of trends — and therefore maintained no sympathy at all for [Page 445] the orthogenetic school. He even slipped into the philosophic fallacy that Whitman criticized (see p. 385), by equating nonisotropy with teleology in the strong sense of inherently unscientific assertion: “According to the Darwinian principle, species forming variability — mutability — does not take place in definite directions. According to that theory, deviations take place in almost every direction without preference for any particular one, and espe­cially without preference for that direction along which differentiation hap­pens to be proceeding. Every hypothesis which differs from Darwin's in this respect must be rejected as teleological and unscientific” (1909a, volume 1, p. 198).

  In a later, remarkable passage, de Vries gathered all these elements together, identifying nonisotropy as both a central Darwinian claim and a necessary bulwark in the struggle against supernaturalism. He also acknowledges that Darwin applied the resulting selectionist mechanism at the organismal level, whereas he favors the species level, thus devising a fundamentally different theory:

  We are strongly opposed to the concept of a definite “tendency to vary” which would bring about useful changes, or at least favor their appear­ance. The great service, which Darwin did, was that he demonstrated the possibility of accounting for the evolution of the whole animal and vegetable kingdom without invoking the aid of supernatural agencies. According to him, species forming variability exists without any reference to the fitness of the forms to which it gives rise. It simply provides mate­rial for natural selection to operate on. And whether this selection takes place between individuals, as Darwin and Wallace thought, or whether it decides between the existence of whole species, as I think; it is the pos­sibility of existence under given external conditions which determines whether a new form shall survive or not... The mutability of Oenothera Lamarckiana satisfies all these theoretical conditions per­fectly. Nearly all organs and all characters mutate, and in almost every conceivable direction and combination (1909a, volume 1, p. 257).

  I have presented this extended treatment of de Vries for a reason embedded in the plan of this chapter, and crucial to the logic of this book. I argue that “internalism” poses two separate challenges to pure Darwinian functionalism: saltational change arising from internal forces of mutability, and inher­ent directionality of variation (corresponding to facet-flipping and channeling on Galton's polyhedron). Most internalists (“structuralists,” “formalists,” “laws of form” theorists in other terminologies) emphasize the second theme of channels and preferred directionality of variation (now most often ex­pressed in the popularity of “constraints” as a subject in modern evolution­ary literature — see Chapters 10 and 11). This style of internalism represents the primary theme of Goethe, of Geoffroy, of Owen, and of the orthogeneticists. Fewer internalists emphasize the saltational theme — and those who do, like Bateson, tend to support channeling as well as facet flipping [Page 446] (for the two themes fuse well into a coherent anti-Darwinian philosophy, as Bateson recognized and articulated).

  When such basic themes commingle, our experimental traditions lead us to search for “pure” end-member cases — examples of one item without the other, so that we may assess the unadulterated contribution of each theme treated separately. We can identify several pure channeling theorists — biolo­gists, who extolled directional variation, but supported gradualism and re­jected saltation (several orthogeneticists fall into this category). But we rarely encounter a “pure” saltationist who accepts isotropy for large-scale changes and rejects all notions of preferred directionality. (Since saltation also implies an internal control upon change, the allied theme of preferred directions usu­ally gains assent as well — as in Bateson's arguments on homeotic variation.) But de Vries represents our instructive “pure” case of one without the other — a saltationist who accepted the isotropy of these immediate and substantial changes. Moreover, de Vries explicated the logic of his unusual commitment to recognize the interesting role that such a “nonstandard combination” must imply for Darwinism. (Saltationism precluded any role for selection in the or­igin of species; but mutational isotropy resurrected the Darwinian apparatus at the higher level of evolutionary trends.) As a “test case,” therefore, de Vries — all by himself — balances the rest of this chapter, with its primary em­phasis on internally channeled variation. De Vries's brilliance, particularly his clear attention to Darwinian logic, makes him a particularly attractive and in­structive figure in our quest to understand the components of internalist thinking.

  De Vries on macroevolution

  De Vries never placed primary emphasis upon macroevolutionary themes in his books, but these issues do receive more than passing attention. By his own intent and reckoning, de Vries' main contribution to macroevolution lay in his resolution of Kelvin's paradox — an earth too young to permit evolution by Darwinian gradualism (see Chapter 6). Obviously, if new species arise per saltum in a single generation, limits on the earth's age would not preclude the evolutionary work actually accomplished. “The demands of the biologists and the results of the physicists are harmonized on the ground of the theory of mutation” (1905, p. 712). “One of the greatest objections to the Darwin­ian theory of descent arose from the length of time it would require if all evolution was to be explained on the ground of slow and nearly invisible changes. This difficulty is at once met, and fully surmounted by the hypothe­sis of periodical but sudden and quite noticeable steps. This assumption re­quires only a limited number of mutative periods, which might occur within the time allowed by physicists and geologists for the existence of animal and vegetable life on the earth” (1905, p. 29).

  De Vries even made a semi-quantitative assessment, based again on Oenothera. Given Kelvin's 24 million year estimate for a habitable earth, even one mutation (adding a character) every 4,000 years would provide 6,000 new features for any extant lineage. Since neither Oenothera, nor any [Page 447] other creature, bears so many distinct features (“a number far higher than comparative and systematic science can by any means accumulate in its descriptions” — 1909a, volume 2, p. 665), geological time becomes positively bounteous for requirements of the mutation theory. De Vries epitomized his view in an “equation” (not really a mathematical formula, but a set of con­cepts epitomized by letters): M x L = BT, or number of mutations times length of intervals between mutations equals “biological time.” So long as bi­ological time doesn't exceed available geological time, Kelvin's paradox may be resolved.

  But de Vries' main contribution to macroevolution does not lie in this ex­plicit geological aid (soon made irrelevant by radioactive extension of the earth's age in any case). Rather, de Vries formulated a fully articulated macroevolutionary theory based on the application of Darwinian logic to the higher level of species and trends — a true theory of “species selection” (even so named by de Vries!). I do not think that de Vries ever recognized the full im­port of what he had done — for he never grants the theme real prominence, and pieces of the argument lie scattered throughout his writing. But de Vries developed all the parts, and they do cohere. His acute understanding of Dar­winian logic (a grasp that also led him to reject selection where he felt that such a mechanism couldn't apply) drove him on and informed all the dispa­rate elements of his thinking.

  De Vries developed his macroevolutionary concepts as a set of logical implications from the Mutation Theory. He recognized, first of all, that an ori­gin of new species by recurrent and effectively identical saltation provides no variation for the working of any Darwinian process of selection: “ [Perhaps] the mutants of one type ... would not be pure ... but would exhibit different degrees of deviation from the parent. The best would then have to be chosen in order to get the new type in its pure condition. Nothing of the kind, how­e
ver, was observed. All the oblonga mutants were pure oblongas. The pedi­gree shows hundreds of them in the succeeding years, but no difference was seen and no material for selection was afforded” (1905, pp. 559-560).

  All effective evolutionary variation exists only among the “elementary species” formed by mutation from a parental stock, not among individual organisms within these species. In principle, therefore, if selection works as an evo­lutionary force at all, Darwin's process can only sort elementary species, not organisms within populations. De Vries clearly recognizes the difference be­tween the conventional Darwinian form of organismal selection and his pro­posal for selection among elementary species: “The struggle for existence, that is to say the competition for the means of subsistence, may refer to two entirely different things. On the one hand, the struggle takes place between the individuals of one and the same elementary species, on the other between the various species themselves. The former is a struggle between fluctuations, the latter between mutations” (1909a, volume 1, pp. 211-212).

  Horticulturists, de Vries noted, recognize the two modes with different names (1905, pp. 604-605) — “race-breeding” for the limited and relatively ineffective Darwinian sorting based on fluctuating variation among organisms [Page 448] in a population; and “variety-testing” for sorting among discrete ele­mentary species produced by mutation. In a remarkable passage, de Vries then names this second mode “species selection,”* and also denies that this higher-level process can be equated with natural selection:

  The word selection has come to have more than one meaning. Facts have accumulated enormously since the time of Darwin; a more thorough knowledge has brought about distinction, and divisions at a rapidly increasing rate, with which terminology has not kept pace. Selection includes all kinds of choice... Selection must, in the first place, make a choice between the elementary species of the same systematic form. This selection of species or species selection [my italics] is now in general use in practice where it has received the name of variety testing. This clear and unequivocal term however, can hardly be included under the head of natural selection. The poetic terminology of selection by nature has al­ready brought about many difficulties that should be avoided in the fu­ture. On the other hand, the designation of the process as a natural selec­tion of species complies as closely as possible with existing terminology, and does not seem liable to any misunderstanding. It is a selection be­tween species. Opposed to it is the selection within the species (1905, pp. 742-744).