The Structure of Evolutionary Theory

by Stephen Jay Gould

  In his most telling statement, and in response to de Vries's oft-repeated but invalid argument (see p. 445) that orthogenesis revitalizes teleology, Whit­man invokes the ontogenetic comparison to defend orthogenesis as the posi­tion most consistent with a mechanistic worldview:

  I take exception here only to the implication that a definite variation ten­dency must be considered to be teleological because it is not “orderless.” I venture to assert that variation is sometimes orderly and at other times rather disorderly, and that the one is just as free from teleology as the other. In our aversion to the old teleology, so effectually banished from science by Darwin, we should not forget that the world is full of order, the organic no less than the inorganic. Indeed, what is the whole devel­opment of an organism if not strictly and marvelously orderly? Is not ev­ery stage, from the primordial germ onward, and the whole sequence of stages, rigidly orthogenetic? ... If a developmental process may run on throughout life, . . . what wonder if we find a whole species gravitating slowly in one or a few directions?... If a designer sets limits to variation in order to reach a definite end, the direction of events is teleological; but if organization and the laws of development exclude some lines of varia­tion and favor others, there is certainly nothing supernatural in this (1919, p. 11).

  Darwin had begun the Origin of Species in a most honorable way that affirmed the necessary, and heretofore largely lacking, empirical foundation of evolutionary argument. Darwin's first chapter did not announce to the world his sweeping reform of all life and thought; instead, he wrote about pigeons (1859, p. 20): “Believing that it is always best to study some special group, I have, after deliberation, taken up domestic pigeons. I have kept every breed which I could purchase, or obtain, and have been most kindly favored with skins from several quarters of the world ... I have associated with several em­inent fanciers, and have been permitted to join two of the London pigeon clubs.”

  Darwin used pigeons to advance the two primary and distinct arguments of his book: (1) the factual claim that evolution had occurred, and represented the source of organic relationships, and (2) the theoretical assertion that nat­ural selection operated as the primary cause of evolutionary change. He sup­ported the first contention by proving that the full range of extensive diversity in modern domesticated breeds had descended from a common wild source, the rock-pigeon Columba livia. (Darwin then added the crucial analogical ar­gument that such intraspecific change could, by extension, serve as a model [Page 386] for evolution at all scales and times). He buttressed natural selection by not­ing that breeders had produced this extensive range of results by propagating favored forms from a cornucopia of essentially isotropic and undirected vari­ation.

  Whitman, of course, accepted the first contention, but refuted the second by challenging one of Darwin's smaller claims. Darwin had observed two ma­jor patterns of coloration within Columba livia — (1) “two-barred,” with two black bands on the front edges of the wings and uniform gray color elsewhere (Fig. 5-6); and (2) “checkered” (spelled “chequered” by both Darwin and Whitman), with black splotches on some or all wing feathers (Fig. 5-7), but also retaining the two bars (usually in more indistinct form). Darwin re­garded the two-barred state as ancestral, and the checkered pattern as de­rived. Whitman reversed this sequence, writing:

  The wild rock pigeons, universally regarded as the ancestral stock of all our domestic pigeons, exhibit two very distinct color patterns, one con­sisting of black chequers uniformly distributed to the feathers of the wing and the back, the other consisting of two black wing bars on a slate-gray ground. The latter was regarded by Darwin as the typical wing

  5-6. Whitman's figure of the two-barred wing pattern, which Darwin regarded as ancestral and Whitman interpreted as an advanced stage in his orthogenetic sequence. From Whitman, 1919.

  5-7. The checkered pattern, viewed by Darwin as derived and by Whitman as the primitive state in the evolution of pigeon wing colors. From Whitman, 1919.

  [Page 387]

  pattern for Columba livia; the former was supposed to be a variation arising there from, a frequent occurrence but of no importance. Just the contrary is true; the chequered pigeon represents the more ancient type, from which the two-barred type has been derived. ... The direction of evolution in pattern in the rock pigeons has been from a condition of rel­ative uniformity to one of regional differentiation (1919, p. 49).

  Whitman's inversion of Darwin's sequence lay embedded within a theory of evolutionary change that Darwin would also have rejected. Whitman based his reversal on a more general concept of directional change in color­ation from an initial homogeneity (checkers on all wing feathers) towards re­gional differentiation (elimination of checkers over most of the wing, with strengthening and coalescence to bars at the distal edge. The bars form by en­largement and alignment of checkers on adjacent feathers; a bar, in other words, arises from a row of checkers that “flow together in a single band” — Whitman, 1919, p. 99).

  Whitman then expanded his sequence of reduction plus regional differentiation beyond the patterns of domesticated pigeons to identify an ineluctable, orthogenetic tendency in the entire family Columbidae (with the portion dis­played by domestic pigeons as just a small part of a much more extensive trend). He identified a prototype for the entire series in the “turtle-dove” pat­tern, a homogeneous field of feathers, each with a dark spot in the center (see Fig. 5-8): “This ancestral mark is a dark spot rilling the whole central part of the feather, leaving only a narrow distal edge of a lighter color. This mark is still well preserved in some of the old world turtledoves — best in the Oriental turtledove of China and Japan. The chequer of Columba livia differs from the dark center of Turtur orientalis only in form and in having a lateral posi­tion” (1919, p. 23).

  From this beginning, the trend moved towards an inexorable end, guided

  5-8. The fully checkered turtle-dove pattern, regarded by Whitman as the original state in the orthogenetic sequence of all dove-like birds. From Whitman, 1919.

  [Page 388]

  by two criteria: general reduction in coloration, and concentration of remain­ing color into regionally differentiated bands or swaths. Thus, in the first stages, spots disappear from much of the plumage, while local areas may de­velop strong concentrations. Subsequently, these local intensities diminish in number and area (bars, for example, may become both narrower and fewer in number). Finally, the regional concentrations become effaced as well, and the bird turns light and monochromatic.

  In a bold move towards complete generality, Whitman then tried to extend this idea of an archetype, followed by an orthogenetic trend, beyond pigeons and doves to the entire field of avian plumage. He postulated that the uniform turtledove pattern should be regarded as ancestral for all coloration in birds. With differing degrees of heterochrony in the inexorable process of reduction, and varying places and styles of regional concentrations, all observed plu­mages might then be rendered as extensive variations upon a single orthogenetic trend. Even the ocelli of peacocks, for example, can be interpreted as altered spots of the turtle-dove pattern, while their restriction to limited areas of the plumage (however showy and conspicuous the result) denotes one form of participation in the universal trend: “With this [turtle-dove] pattern as an archetype it is possible to get an orientation of the whole field of avian patterns and to thread our way through what before seemed an impenetrable maze of multifarious variations, with no discoverable beginning or end of order” (1919, p. 58).

  The mere claim for a trend, even such a pervasive and inevitable series, does not of itself complete an argument for orthogenesis, or internally di­rected variation. After all, both the archetypal turtle-dove pattern and all sub­sequent stages of reduction might be adaptations, externally selected from isotropic Darwinian variation. But Whitman well understood the ingredients required to distinguish true orthogenesis from orthoselection or some other functionalist explanation of trends — namely, (1) evidence that
the trend pro­ceeds independently from (or even despite) adaptive pressures from local environments (selection may alter rates or add details, but cannot derail the basic route); and (2) data supporting an internally based directionality of variation available for shaping into evolutionary change (ontogenetic chan­neling in Whitman's view, as we shall see).

  On the first criterion, Whitman upheld the inexorable character of a trend towards local differentiation and final effacement, independent of what envi­ronment might favor in functional terms.

  The process of evolution in color patterns has been a sweeping one, involving the whole surface and taking the same general direction. The stages reached are various, ranging all the way from the full chequered to the wholely unchequered state; from chequers and bars combined in different proportions to bars alone; from many bars to three, two, one, a remnant, or none; and in all shades of brown, black, gray, red, to pure white. Nowhere in this field of variations do we find any indications that [Page 389] chequers originated in the form of bars at the posterior end of the wing and then spread from behind forward (1919, p. 55).

  The trend, Whitman argues, is pervasive and entirely general. He presented (1919) a remarkable vision of inexorable movement through the entire family of pigeons, from a uniformly spotted archetype to some idealized, albinized version of the Holy Ghost, depicted as a pure white dove in many medieval paintings: “When we see all these stages multiplied and varied through some 400 to 500 wild species and 100 to 200 domestic breeds, and in general tend­ing to the same goal, we begin to realize that they are . . . slowly passing phases in the progress of an orthogenetic process of evolution, which seems to have no fixed goal this side of an immaculate monochrome — possibly none short of complete albinism.” Can one conceive a more unpigeonlike state (in both appearance and deed) — at least in our metaphors — than “immaculate”? To the primary spotting agent of cities throughout the world, Whitman thus gave a higher aspiration and the promise of a purer form. As the Psalmist wrote, “Behold, I was shapen in iniquity . . . Wash me, and I shall be whiter than snow” (Psalm 51).

  The hypothesized putative inexorability of this trend allowed Whitman to explain or interpret many otherwise puzzling phenomena. He seemed most pleased about the unification thereby provided for the two main patterns of domestic pigeons — the empirical source of the entire study. If Darwin had been right, Whitman argued, then the origin of Darwin's postulated ancestral pattern (two-barred) would remain a mystery, and we would also lack any explanation for the subsequent evolution of checkers de novo. But if the trend begins with the turtledove pattern as ancestral, then the checkered state may be easily derived there from, and the two-barred condition becomes just a fur­ther step in reduction and concentration of pigment along an orthogenetic se­ries. (This argument, of course, leaves the origin of the turtledove pattern it­self as an unexplained “primitive term,” but methods of phyletic analysis can at least establish its ancestral status.) “We could not explain how two bars could arise de novo in a clear gray wing surface; but we can see how a sweep­ing reduction process, anteroposterior in direction, would leave two or more rows of chequers cut to dimensions that would coalesce in transverse bars at the posterior end of the wing” (1919, p. 61).

  To cite another example of variational puzzles resolved by the orthogenetic trend, Whitman notes that the stock dove, Columba aenas, develops weaker bars than the domestic pigeon and never exhibits any checkers. The trend, having so far surpassed the stage reached by domestic pigeons, no longer per­mits the development of any checkers, even as an occasional variant in highly colored individuals: “We can readily understand why the stock dove, which has, at least in many cases, a vestigial third bar, quite like that in domestic pi­geons, never appears in chequered dress. It is moving in the other direction, and no reversal of course is now open to it” (1919, pp. 60-61).

  On the second criterion of channeled variability, Whitman cites three lines [Page 390] of evidence for viewing his orthogenetic sequence as an extension of the ontogenetic pathway. First, in comparison with their own juvenile plumages, adult birds generally develop patterns of coloration that may be designated as “further along” the orthogenetic pathway. In a world of recapitulation, this palpable change in the course of weeks becomes a surrogate for the invisible alterations of past millennia. Whitman uses this ontogenetic evidence to as­sert orthogenesis against his two chief rivals, Darwin and de Vries:

  Moreover — and this is as close as we can hope to get to actual seeing — we find that progress of just the kind we are looking for is certainly made in passing from the juvenal [alternate spelling of juvenile, generally ar­chaic in English, but still occasionally used in ornithology] to the adult plumage. This is an ontogenetic change of a few weeks, which we can easily demonstrate by experiment to be progressive and continuous (p. 33). ... Even in the widest departures, when every spot has vanished in the adult plumage, the young bird frequently exhibits more or less per­fect traces of the old marking and sometimes requires several molts to reach its mature condition (p. 58). ... Juvenile phases of color patterns become luminous as recapitulations in the sense of the biogenetic law and do not stand as isolated prodigies of natural selection or as meaning­less exhibitions of mutations (p. 65).

  Second, as Eimer also maintained (and for complex reasons rooted both in cultural biases about sexual differences, and in the facts of embryology), Whitman viewed sexual distinctions as products of the same ontogenetic tra­jectory, with males “further along” than females. He then argued that plu­mages of adult male pigeons display more advanced stages of the ortho­genetic series than females of the same population. Third, Whitman argued that, within the adult plumage of individual birds, last-formed feathers devel­oped more advanced characters — as the biogenetic law required, based on its key principle of terminal addition for evolutionary novelty. Thus, three crite­ria, all interpreted as manifestations of ontogeny — differences among molts of a single bird, between sexes of adults, and within the adult plumage by or­der of formation — indicated a pervasive channel of variation, virtually com­pelling evolution into an extended ontogenetic pathway directed towards re­duced and concentrated coloration.

  To these categories of ontogenetic evidence, Whitman then added two additional sources of data to buttress his orthogenetic series: (1) From compara­tive anatomy, he asserted that phyletic series, established by criteria indepen­dent of coloration, illustrated the orthogenetic sequence in many parallel lineages. (I doubt the claimed independence in many cases, and Whitman may therefore have advanced a largely circular argument by basing phyletic infer­ences on the supposed trends in color themselves.) (2) From breeding, Whit­man found that selection along the orthogenetic trajectory could only move “forward” from checkers to bars, and never in Darwin's proposed order from bars to checkers. Selection must push, but the phyletic sequence can only pro­ceed in one direction — down the channel of orthogenetic variation: [Page 391]

  The conclusion supported by comparative study admits of experimental confirmation. We may take pigeons of the two-barred type, and try to advance from this condition to that of the chequered type, by selecting in each generation birds with the widest bars, and especially any that may have a trace of a third bar. This I have tried continuously for 6 years and with several different stocks. I have not been able to establish a third bar, or to extend chequers in front of the vestigial third bar, which is often found. With purebred birds, not allowed to mingle with chequered birds, I believe it is impossible to advance from bars to the chequered bird state. With chequered pigeons, on the other hand, it is fairly easy to advance in the opposite direction, gradually clearing the field and leaving two bars. The process has been carried to the point of completely elimi­nating the bars (1919, p. 60).

  These accumulated sources of evidence led Whitman to strong assertions about the primacy of orthogenesis among rival theories: “The orthogenetic process is the primary and fundamental one” (1919, p. 35).
In his boldest statement (1919, p. 191), Whitman advocated a model of inevitable evolu­tionary flow, and explicitly limited the role of natural selection to tinkering with the style and rate of a determined sequence:

  The steps are seriated in a causal, genetic order — an order that admits of no transpositions, no reversals, no mutation-skips, and no unpredictable chance intrusions. This series may conceivably be lengthened or short­ened, strengthened or weakened; indeed, we may multiply the number of steps at will; that is, we may provoke one or more steps to arise between any two normal steps; but in that case the new steps will be measured true to the time and place of introduction, and their direction will invari­ably coincide with that of the series as a whole, so that if the time and place of origin are noted, the nature and extent of the strides may be ap­proximately predicted.