Bateson put his formalist and non-Darwinian thoughts together in one of the most interesting biological works of the late 19th century — Materials for the Study of Variation (1894). This work has been read primarily as a defense of saltational variation and change, a brief for structuralism using the “facet flipping” theme of Galton's polyhedron (see pp. 342–351). I will not challenge this primacy, but I do wish to demonstrate that Bateson's book integrates a broader set of formalist themes (including distrust of adaptation and suspicion of historical contingency) under a primary concern for saltation and discontinuity as a counter to Darwinism.
Bateson's Materials may be a famous book in retrospect (primarily, I suspect, because scholars want to grasp how the man who later invented the word genetics looked upon variation in the last pre-Mendelian decade). But the volume failed in its own time as a long compendium by an unknown author, and a financial disaster. Bateson's wife remembered (1928, pp. 57-58): “The book was not a success — the professors and lecturers of the day did not introduce their students to it. Perhaps, that they should not was to be expected. For a few years the annual arrival of the publisher's account was a dismal event, and the book was put “in remainder” and dropped out. The second volume as such was never written.”
As a rhetorical strategy, many comprehensive works begin with a small logical puzzle or anomalous observation. For example (see Gould, 1987b, for details), Burnet and Hutton presented their grand geological systems as solutions, in Burnet's case, to the problem of sources for water in Noah's flood; or, for Hutton, as a logical dilemma in final cause (why would a benevolent God, attuned to human needs, allow soil to be made by a process that must eventually erode the earth away). Bateson's Materials also poses its central argument as the solution to a particular puzzle: how can evolution produce a world of taxonomic discontinuity when environmental gradients, as potential impetuses for change, are generally continuous: “The differences between species ... are differences of kind, forming a discontinuous series, while the diversities of environment to which they are subject are on the whole differences of degree, and form a continuous series” (1894, p. 16). [Page 399]
Bateson recognized that both Lamarckism and Darwinism, as functionalist mechanisms, posited a flow of information from environment to organism as a basis of adaptive transformation. How then could a continuous environment yield our world of thinly populated morphospace, with vast gaps between realized designs? “According to both theories [Lamarckism and Darwinism], specific diversity of form is consequent upon diversity of environment, and diversity of environment is thus the ultimate measure of diversity of specific form. Here then we meet the difficulty that diverse environments often shade into each other insensibly and form a continuous series, whereas the specific forms of life which are subject to them on the whole form a discontinuous series. The immense significance of this difficulty will be made more apparent in the course of this work” (1894, p. 5).
For Bateson, a general solution could be derived from logical implications of the argument, prior to any search for causes: nature's discontinuity must arise from the internal workings of organisms:* “Such discontinuity is not in the environment; may it not, then be in the living thing itself?” (1894, p. 17).
Bateson, as a young Turk inspired by ideals of German mechanism and American experimentalism, but working in a more traditional world of descriptive natural history, knew what he didn't like about the inferential procedures of most Darwinians in his generation: the conjoined tactic of speculation based on embryology for phylogenetic reconstruction, and guesswork about utility for inferences about adaptation by natural selection. Bateson lists the two pitfalls of such sterile work: “The first of these is the embryo-logical method, and the second may be spoken of as the study of adaptation. The pursuit of these two methods was the direct outcome of Darwin's work” (1894, p. 7).
Bateson longed to apply the mechanistic style of experimental science to the causes of evolution. If guesswork about externalities had served the field so poorly, why not look to the intrinsic characters of organisms, features that might be resolved by manipulation and by understanding the mechanics of heredity. Variation itself must be taken as a primary phenomenon. Why, at least as an initial strategy, look beyond this palpable and measurable property of populations? Perhaps the causes of evolutionary change lie in variation itself and not in a superimposed external sorting, as the more complex Darwinian mechanism proposed: “Variation, in fact, is evolution. The readiest way, then of solving the problem of evolution is to study the facts of variation” (1894, p. 6). [Page 400]
Bateson notes the fascination of his colleagues for the causes of variation as expressed in the nature of heredity, but holds that the absence of any hard evidence has mired the subject in fanciful and speculative hypotheses, from Darwin's pangenesis to Haeckel's perigenesis. The quest should be postponed for now (though Bateson wouldn't have long to wait, as the Mendelian revival lay just around the corner of the coming century). Meanwhile, an empirical approach might yield great benefits, at least by providing an inductive entry to the difficult subject of causes. Why not, in short, simply gather the facts of variation: “It is especially strange that while few take much heed of the modes of variation or of the visible facts of descent, everyone is interested in the causes of variation and the nature of 'heredity,' a subject of extreme and peculiar difficulty. In the absence of special knowledge, these things are discussed with enthusiasm, even by the public at large. But if we are to make way with this problem, special knowledge is the first need. We must know what special evidence each group of animals and plants can give, and this specialists alone can tell us” (1894, p. ix).
Bateson chose to express this strategy of empirical compilation in the title of his book: “To collect and codify the facts of variation is, I submit, the first duty of the naturalist” (1894, p. vi). Brave words, to be sure, but Bateson recognized that such a complex and multifarious subject could not be resolved simply by toting the relative frequencies of an empirical list. He also understood that the very idea of a totally unbiased listing could only operate as a self-serving fiction to bolster a myth of perfect scientific objectivity. Bateson recognized a pervasive bias in traditional accounts of variation — strong preferences for continuity and gradualism, as expressed in the old Leibnizian and Linnaean aphorism, natura non facit saltum: “First there is in the minds of some persons an inherent conviction that all natural processes are continuous . . . Secondly, variation has been supposed to be always continuous and to proceed by minute steps because changes of this kind are so common in variation” (1894, p. 16). Bateson's list, therefore, would be a purposive account of a particular sort of variation: “If facts of the old kind will not help, let us seek facts of a new kind” (1894, p. vi).
Since Bateson sought the causes of evolution in variation itself, and since he viewed discontinuity as the primary fact of natural history, discontinuous variation among organisms within populations became his favored source of evolutionary change. Materials for the Study of Variation is not an unbiased compendium of all organic mutability, but rather an explicit attempt to catalog discontinuous variation as a source of insight into internally driven causes of evolution. The subtitle of the book explicitly refutes any claim to balance or comprehensiveness among styles: “Treated with Especial Regard to Discontinuity in the Origin of Species.”
Bateson divided variation into meristic (for serially and symmetrically repeated, countable and discontinuous structures) and substantive (for ordinary continuous variability). As an obvious ploy to promote his preference for locating the causes of evolution in discontinuity of variation, the meristic [Page 401] category commanded his primary attention. Bateson devotes the entire text of Materials to compiling examples of meristic variation. He planned, but never wrote or even seriously began, a second volume on substantive variation.
(Bateson left a substantial legacy of biolog
ical terminology. He is best known, of course, as the inventor of the word genetics, but Materials includes two new terms of later importance — meristic for this general style of variation, and homeosis for a subset that has since become central in modern evolutionary and developmental biology; see Chapter 10.)
The evolutionary import of Bateson's book may be summarized in three characteristic features of his argument for saltational variation and change, with a subsequent fourth theme then centered upon his discussion of implications for Darwinism.
Nature of the examples. Materials are, above all else, a compendium of examples of discontinuous variation in meristic characters. Bateson begins with a long sequence of chapters (pp. 87-422) on linear series, starting with arthropod segments, moving through vertebrae and ribs, where Bateson presents the “type” cases of homeosis, and proceeding to branchial openings, mammae, teeth and digits. The second sequence of chapters (pp. 423-566) treats symmetrically repeated structures under three headings: radial series, bilateral series, and secondary symmetry and duplication. Although Bateson adopted a convention of presenting facts in small type and interpretation in a larger font, he remained true to his own version of the Kantian dictum that percepts without concepts are blind — for even the factual small-type listings contain implicit interpretations for his worldview, and against Darwin's.
As his primary theme, Bateson emphasizes a basic implication of meristic variation. Segments must be conceptualized as discrete anatomical forms, and supernumeraries (or deletions) are therefore usually complete (or entirely suppressed). Half an added segment usually denotes a structural and functional absurdity, in principle. Merism, by its very nature, implies discontinuity in construction and change. Bateson writes, for example, about 12-jointed antennae within a group of normally 11-jointed beetles:
Would it be expected that the longicorn Prionidae, most of which have the unusual number of 12 antennary joints, did, as they separated from the other longicorns which have 11 joints, gradually first acquire a new joint as a rudiment which in successive generations increased? ... If anyone will try to apply such a view to hundreds of like examples in arthropods, of difference in number of joints and appendages of near allies ... he will find that by this supposition of continuity in variation he is led into endless absurdity. Surely it must be clear that in many such cases to suppose that the limb came through a phase in which one of its divisions was half-made or one of its joints half-grown, is to suppose that in the comparatively near past it was an instrument of totally different character from that which it has in either of the two perfect forms. But no such supposition is called for. With evidence that transition of this nature [Page 402] may be discontinuously effected the difficulty is removed (1894, pp. 410-411).
Whereas this basic argument contravenes gradualism, a second implication then disputes natural selection. Supposed anomalies of discontinuous variation often seem no less structurally “perfect” than normal forms. Why, then, do we so confidently ascribe the “normal” forms to long honing by natural selection working upon continuous variation? Should we not rather conclude that, since normal and anomalous variants may be equally well constructed, both categories arise by internal regulation? Of roaches with four-jointed tarsi (instead of the normal five), Bateson writes:*
The four-jointed tarsus occurring thus sporadically, as a variety, is not less definitely constituted than the five-jointed type, and the proportion of its several joints is not less constant. It is scarcely necessary to point out that these facts give no support to the view that the exactness or perfection with which the proportions of the normal form are approached is a consequence of selection. It appears rather, that there is two possible conditions, the one of five joints and the other with four, each being a position of organic stability. Into either of these the tarsus may fall; and though it is still conceivable that the final choice between these two may have been made by selection, yet it cannot be supposed that the accuracy and completeness with which either condition is assumed is the work of selection, for the “sport” is as definite as the normal (1894, pp. 64-65).
In a further move to complete the basic argument within a supposed compendium of objectively listed facts, Bateson emphasizes homology between the teratology of individuals in one population and the normal morphology of a related species — with an obvious implication of transformation by saltational variation. Fusion of bilaterally symmetrical organs in the mid line (or fission of singletons into pairs on the antimeres) establishes a major class of cases:
A normally unpaired organ standing in the middle line of a bilateral symmetry may divide into two so as to form a pair of organs; and conversely, a pair of organs normally placed apart from each other on either side of a middle line may be compounded together so as to form a single organ in the middle line. In animals and plants nothing is more common than for different forms to be distinguished from each other by the fact that an organ standing in the middle line of one is in another represented by [Page 403] two organs, one on either side. The facility therefore with which each of these two conditions may arise from the other by discontinuous variation is of considerable importance (1894, p. 448).
The physical basis of discontinuity. Early in his studies, Bateson developed an insight, a sort of epiphany in his own assessment, that would shape his views (haunt might be a better word) throughout his career. His emphasis on discontinuity, his dislike of Darwinism, his inability to come to terms with chromosomal theory, all reflect this central vision of his thinking. Bateson decided that discontinuously repeated organic structures bore isomorphic, and therefore common causal, similarity to physical phenomena produced by waves and vibrations. He therefore sought a physical cause for heredity in some wave-like form of energy — the “vibratory theory” in his own words — and he could therefore never fully accept a particulate basis for genetics. In 1891, he wrote with great excitement to his sister, stating that he dared not even hope to have an idea of such import ever again:
Did I tell you anything about my new vibratory theory of repetition of parts in animals and plants? I have been turning it over again lately, and feel sure there is something in it. It is the best idea I ever had or am likely to have — do you see what I mean? — divisions between segments, petals, etc. are internodal lines like those in sand figures made by sound, i.e. lines of maximum vibratory strain, while the midsegmental lines and the petals, etc. are the nodal lines, or places of minimum movement. Hence all the patterns and recurrence of patterns in animals and plants — hence the perfection of symmetry — hence bilaterally symmetrical variation, and the completeness of repetition whether of a part repeated in a radial or linear series etc. etc. I am, as you see, in a great fluster (in Bateson, 1928, p. 42).
In his next letter, he added: “You'll see — it will be a commonplace of education, like the multiplication table or Shakespeare, before long!”
Materials do not discuss the vibratory theme at length, if only because Bateson chose to organize the book as a compendium of data — and he could present no factual support for his suspicions about the production and inheritance of discontinuous variation. Still, he advanced several conjectures about the construction of phenotypic discontinuity from underlying continuity by a simple physical or chemical impetus. He attempts, for example, to analogize the discrete concentric rings of eyespots on lepidopteran wings with pond ripples (metaphorically) and, in greater hope of causal isomorphism, with chemical reactions:
A whole eyespot may come, or it may go ... leaving the field of the cell plain and without a speck. The suggestion is strong that the whole series of rings may have been formed by some one central disturbance, somewhat as a series of concentric waves may be formed by the splash of a stone thrown into a pool. It is especially interesting to remember that the [Page 404] formation even of a number of concentric rings of different colors from an animal pigment by the even diffusion of one reagent from a center occurs actually in Gmelin's t
est for bile-pigments. Bile is spread on a white plate and a drop of nitric acid yellow with nitrous acid is dropped on it. As the acid diffuses itself distinct rings of yellow, red, violet, blue, and green are formed concentrically around it by the progressive oxidation of the bile-pigment ... This example is merely given as an illustration of the possibility that a series of discontinuous chemical effects may be produced in concentric zones by a single central disturbance (1894, p. 292).
Rarely missing an opportunity for a dig at Darwinism, Bateson then adds that, with the chemical analogy, we at least hope to find a cause, whereas the standard adaptational speculation can make no such claim: “As to the function of the ocellar markings nothing is known, and I am not aware that any suggestion has been made which calls for serious notice” (1894, p. 294).
Writing more generally on the same theme, Bateson tries to attribute discrete color classes directly to the chemical stability of pigments — and to dismiss the alternative functional explanation of adaptational guesswork about the utility of discontinuous difference. (This passage occurs in Bateson's only short discussion of substantive rather than meristic variation — the category intrinsically less favorable to his preference for discontinuity. Thus, he strives for explanatory generality across all classes of variation): “It would, I think, be simpler to regard the constancy of the tints of the several species and the rarity of the intermediate varieties as a direct manifestation of the chemical stability or instability of the coloring matters, rather than as the consequences of environmental selection for some special fitness as to whose nature we can make no guess. For we do know the phenomenon of chemical discontinuity, whatever may be its ultimate causes, but of these hypothetical fitnesses we know nothing, not even whether they exist or no” (1894, p. 48).
The Structure of Evolutionary Theory Page 64
