For all these reasons, I suspect that selection among deme-individuals holds an importance as yet unrealized (and perhaps occurring in modes as yet unconceptualized) within our general picture of evolution. I have, in my career, witnessed three examples of widespread dismissal by ridicule as part of a professional ethos: the rejection of continental drift as physically inconceivable; [Page 703] the shunning of Goldschmidt's macroevolutionary ideas as dangerous to the Darwinian consensus; and the dismissal of group selection as addlepated nonsense (see pp. 553–556). Nothing in my intellectual life has made me feel more uncomfortable.
I take great pleasure in the comeuppance of the smug ridiculers in all three cases. Plate tectonics have validated continental drift to become a new paradigm for geology. Goldschmidt's particular genetics win no general plaudits, but his views on a conceptual break between micro- and macroevolution now enjoy substantial support. The vindication of group selection has been slower, but now moves on apace (see Sober and Wilson, 1998) — with a vigorous professional discussion finally occurring, and with general attention now accorded, both in the popular press (Lewin, 1996), and in the commentary sections of general professional journals (Morrell, 1996). Sic semper tyrannis.
THE SPECIES-INDIVIDUAL I propose, as the central proposition of macroevolution, that species play the same role of fundamental individual that organisms assume in microevolution. Species represent the basic units in theories and mechanisms of macroevolutionary change. In this formulation, the origins and extinctions of species become strictly analogous to the births and deaths of organisms — and just as natural selection works through differential proliferation based on schedules of organismal births and deaths, so too does species selection operate upon the frequencies and timetables of origins and extinctions. The next section of this chapter — entitled “the grand analogy” — shall complete this argument by attempting to cash out this comparison in detail, with all the intriguing differences that arise when disparate individuals at two such different levels work by the same abstract mechanism.
I will therefore confine this preliminary discussion to the three major objections that have been raised against the foundational idea that species can act as important evolutionary individuals. These objections treat, in reverse order, the three words in the key phrase, “important evolutionary individuals.” The first objection holds that species cannot be construed as proper individuals; the second admits that species are individuals, but argues that they cannot operate as interactors (as required for units of selection); while the third allows that species may be recognized as both individuals and interactors, but insists that they must remain effectively impotent in both roles.
Species as individuals. The classic argument of evolutionary gradualism denies real existence to species because they can only be defined as arbitrarily delineated segments of a lineage in continual anagenesis. Both Lamarck and Darwin, despite their maximally different views about proposed evolutionary mechanisms, strongly supported the nominalistic claim that only organisms exist as natural units, and that species must therefore represent abstractions, formally designated only for human convenience. (As many historians have remarked, Darwin chose an odd title for his revolutionary book — for he focusses upon the explanation of substantial change by anagenesis, and says little about speciation by branching of lineages.) [Page 704]
I presented the case for treating species as individuals in an earlier section of this chapter (pp. 603–608), noting that punctuated equilibrium greatly aids such a delineation, but then extending and generalizing the argument by holding that species can be individuated under any scheme that depicts their origin as an event of branching, rather than anagenetic transformation. Critics of this view, particularly Williams (who does not dispute the truly necessary claim for origin by branching), continue to raise standard objections, especially “an absence of a decisive beginning for a species” (Williams, 1992, p. 121). But Williams, in advancing this argument, commits the classic error of failure to appreciate proper scales. His claim for fatal fuzziness in origins views the question from a generational perspective at the scale of human lifetimes. The great majority of species, however, arise in geological moments (thousands of years, and thus overly long only at the inappropriate scale of our personal lives) — a shorter period of ambiguity (relative to later duration as a clearly separate entity) than we note for most asexual organisms that reproduce by budding (the proper organismic analog for the origin of a new species by branching)!
Most other published objections to species as individuals also express little beyond our psychological difficulty in making a transition to different criteria at unfamiliar scales. Some critics have argued, for example, that species can't develop the requisite property of heritability, because no mechanism can be analogized with the well-known Mendelian basis for this phenomenon at the organismic level. But heritability measures the correlation between parents and offspring based on direct transmission of formative properties — and daughter species surely inherit parental characteristics by this standard route. The required correlation arises by transmission of autapomorphic characters through retained homology — the appropriate mechanism of heritability at this higher scale, and in no way “worse” than Mendelian criteria for the construction of organisms. Moreover, species heritability can be measured in the same general way, and with the same potential accuracy, as standard organismic heritability — as Jablonski (1987) has done in our best-recorded case of species selection for the evolution of marine mollusks in normal times and episodes of mass extinction.
Other critics charge that species are too spatially diffuse, or too lacking in mechanisms of internal coherence, to count as individuals. But, again, these arguments only arise from failure to conceptualize this different scale in an appropriate manner — a mental foible rooted in our parochial allegiance to the particular (and poorly-scaling) criteria of individuality for organisms. Species don't build a physical skin, but reproductive isolating mechanisms maintain their borders just as sharply. Species don't evolve immune systems and other forms of “policing” against outside invaders, but the constant admixture among their parts via sexual reproduction maintains coherence with more than adequate force.
Species as interactors. This more interesting and challenging argument has unfolded among supporters of macroevolutionary theory as an “in-house” [Page 705] debate. Most discussants, including Brandon, Gould, Jablonski, Lloyd, Stanley, and Vrba, strongly support the concept of species as units of selection; while Damuth and Eldredge grant species a role as replicators, but not as interactors, and therefore not as agents of selection. Grantham (1995) has tried to mediate these positions with a compromise that will, I suspect, satisfy neither side.
Critics allow that species may be “fundamental units” of macroevolution in some sense — but they say, only as the replicators that serve as “atoms” of cladistic phylogeny, and not as interacting units that forge macroevolutionary change by active competition in natural environments. (Eldredge, for example, includes species in the genealogical column of his two-hierarchy scheme — see page 642 for a critique — but not in his economic column of interactors.) A species, the critics continue, may live in too broad a range of environments, and over too wide a geographic range — often discontinuous to boot — to serve as an interactor, or unit of selection. Moreover, although individual populations of two species may compete sympatrically over a well-delineated geographic range, entire species rarely maintain sufficient overlap to interact with each other as complete units.
To resolve this apparent dilemma, Damuth (1985) proposes that we define a new interactor corresponding most closely to the hierarchical level where species serve as replicators. Using a criterion of direct competition in sympatry, Damuth proposes the term “avatar” for such interactors, defined as sympatric populations in ecological competition, and therefore interpretable as alternatives subject to selection. Grantha
m's (1995) “compromise” position maintains allegiance to Damuth's insistence upon potential interaction in sympatry. Grantham defends species selection, and regards species as potential interactors — but he would restrict any particular study of species selection to members of clades living in the same broad region. He writes (1995, p. 311): “I suggest that paleontologists focus on geographically constrained portions of monophyletic clades.”
I would raise three arguments against this proliferation of terms and categories — and for the status of species as adequate interactors.
1. A standard mode of construing competition among organisms has beguiled us into thinking that interaction requires sympatry. As argued in Chapter 6 (pp. 470–477), Darwin strongly asserted the predominance of biotic over abiotic competition as the only promising path for a defense of progress in evolution. This preference has passed through the Victorian fascination with overt battle as a defining mode of competition, right into our present times, with continuing Tennysonian metaphors about “nature red in tooth and claw” (see Gould, 1992a), and newspaper stories about firms engaged in Darwinian struggles to the death as they vie directly for the allegiance of a limited population of consumers. (As I revised this chapter in the summer of 2000, a new magazine for “business evolving in the information age” made its debut under the name Darwin — also available on line at www.darwin-mag.com.) [Page 706]
But this focus on the biotic mode has always been indefensible as a claim for exclusivity, or even dominant relative frequency. In Darwin's own time, Huxley ridiculed this notion as “the gladiatorial theory of existence,” while Kropotkin (1902) and others constructed alternatives based on cooperation in sympatry and the prevalence of abiotic competition in most environments (see Todes, 1988; Gould, 1991b). Darwin himself clearly favored an expansive concept of interaction with environments in natural selection — as when he insisted, in a famous passage (1859, p. 62), that “a plant on the edge of a desert” struggles for existence against the drought and other features of the physical environment just as surely as “two canine animals in a time of dearth” struggle more overtly for a limited supply of meat.
This point becomes important when we try to translate this debate about organisms to a definition of higher-level interactors. Biotic competition does require sympatry for direct and literal struggle, while abiotic competition imposes no such conditions, and must often occur among organisms that never encounter each other, even while living in sympatry. If we use biotic competition as our (often unconscious) paradigm for the entire, and far broader, concept of interaction, then we too easily become unduly committed to the false restriction that interactors must be able to duke it out directly. In upward translation, this bias leads to the idea that species-individuals can't be interactors unless they live in the same place, and thus maintain a potential for engaging in some analog of overt battle.
But interaction at the canonical level of organisms doesn't demand direct contact, or even life in the same place — and no one has denied that organisms operate as quintessential interactors, and units of selection. If abiotic competition dominates the history of life — as many distinguished researchers insist (see references in Allmon and Ross, 1990), at least for many groups in many circumstances — then potential for direct contact cannot be invoked as a primary criterion for defining interactors.
Williams (1992) has strongly asserted the non-necessity of sympatry (and resulting potential for direct “struggle”) in defining higher-level interactors — and he uses the same analogy here advanced for asserting a similar non-necessity at the organismal level. I presented the full quote before, but repeat the operative line here (1992, p. 25): “One issue is whether the populations that bear the gene pools need be in ecological competition with each other. I believe that this is not required, any more than individuals within a population need interact ecologically to be subject to individual selection.” Later, Williams specifically criticizes Damuth's definition of avatars on this basis. Speaking of populations not in direct competition, but subject to similar stresses (a common predator in their separate environments in this hypothetical case), Williams writes (1992, p. 52): “I am inclined to recognize that clade selection is operating even here, unlike Damuth, who maintains that only sympatric avatars, populations in ecological competition, can be alternatives subject to selection. Allopatric forms may not be ecological competitors, for the inattention of a predator or anything else, but they compete for representation in the biota, the ultimate prize in clade selection.” [Page 707]
2. Although I recognize that some notion of a common environment must be invoked when we wish to define allopatric species as competing inter-actors, I do not view such a requirement as either rarely met or particularly difficult to specify. (I only mean, by the last phrase, “any more difficult to specify than for sympatric interactors.” We cannot know, in fully adequate detail, how individuals at any level react to all nuances of the environment, in all their horrendously complex and nonlinear interactions. Who can say whether two sympatric organisms, given their inevitable differences, perceive the same local change of environment — even such a linear effect as a falling temperature — in the same way? I am only arguing that we face the same difficulty for sympatric, as for allopatric, interactors in this respect.)
At least two strong arguments support the notion of adequate environmental similarity in allopatry:
(i) Environments cannot be conceptualized (or even operationalized) as objective places or circumstances in a world fully external to the organisms involved. First of all, environments include all interactions with other organisms, both conspecific and belonging to different taxa, and not just the climates, substrates, and other more measurable properties of a surrounding physical world. Second, and more important, as Lewontin has emphasized so forcefully (1978, 2000), environments are intrinsically referential, and actively constructed by the organisms in question. Environments, in short, are made, not found. Thus, important properties of the environment must be sufficiently comparable in a set of closely related and partly allopatric species engaged in a process of species selection. These species share key traits as autapomorphies of their clade — and since these traits help to construct the relevant environment, sufficient similarity becomes, in part, an active construction of related organisms, not only a happenstance of common externalities.
(ii) Organisms needn't occupy the same turf in order to be impacted in similar ways by the kinds of broad environmental changes that seem so important at geological scales. To choose an extreme example, when, 65 million years ago, a large bolide struck the earth in the region now occupied by the Yucatan peninsula, I suspect that Tyrannosaurus rex in the western United States, and its recently discovered sister taxon in Africa, experienced consequences sufficiently common and negative to influence their extinction (while some small-bodied mammals, living there and elsewhere, survived as a consequence of organismal or higher-level characters that also do not require sympatry with dinosaurs for meaningful comparison). Again, Williams (1992, p. 25) explains the issue succinctly and at more immediate scales: “Suppose a climatic change causes the brown trout of the upper Rhine to die out but lets the brown trout of the upper Danube survive. Suppose further that the difference in fate is attributable to some difference in gene frequency that causes a difference in vulnerability to the change. That is surely clade selection. The ultimate prize for which all clades are in competition is representation in the biota.”
3. In many cases of species selection, the success of one species over [Page 708] another cannot be explained by competition between their sympatric populations, but depends upon a species-level trait of the species's full range — in other words, species selection of the whole, not of avatars or sympatric subsections. I present in Figure 8-5 a hypothetical case developed by Robert N. Brandon (personal communication, 1988, Ohio State meeting).
The three species of a clade live on
four adjacent volcanic islands. Species 2 can move readily across small oceanic gaps and inhabits all four islands. Species 1 and 3 have limited mobility and live on only one island each. (Species 2 gains no necessary advantage of the moment thereby.) The population of Species 1 on Island A, and of Species 3 on Island C, may each exceed the total number of organisms in Species 2 on all four islands. In fact, on any individual island, either Species 1 or Species 3 may always fare better than Species 2. Each island maintains an active central volcano; when the volcano erupts, all life on the island dies, but the adjacent islands remain unaffected. One fine day, the volcanoes of Islands A and C erupt. As a consequence, Species 1 and
8-5. A hypothetical example of species selection based on traits that belong to entire species — in this case the full geographic range — and not to avatars or subpopulations thereof. See text for details of this verbal case developed by R. N. Brandon. Species 2 survives by virtue of its ability to spread among islands, even though any other species dominates over species 2 on any island of joint occurrence.
[Page 709]
Species 3 become extinct, but Species 2 survives thanks to populations on Islands B and D — that is, only by virtue of populations allopatric with Species 1 and 3.
The Structure of Evolutionary Theory Page 113
