Clearly, Species 2 has survived as a result of greater geographic range, caused by whatever organismal, deme, or species traits permitted the colonization of all islands. Geographic range may be either an emergent or aggregate trait of successful Species 2; but, in any case, this trait exists at the species level and confers an irreducible fitness based on superior range (obviously a property of the species, and not of any individual organism, deme, or avatar). This hypothetical case presents a potential and plausible example of species selection based on a trait of the entire species and its complete range — and explicitly not on any sympatric avatar, or any other subsection of the full entity.
Species selection as potent. Two separate arguments, one empirical and the other theoretical; have been raised against the efficacy of species selection. The first, which I regard as unfair, claims that a paucity of currently recorded empirical examples must indicate the rarity of the phenomenon. I would respond, first of all, that a few excellent (and elegant) cases have been well documented, so this process cannot rank as a distant plausibility waiting for an improbable verification, as some critics have charged. Jablonski (1987), for example, performed a pioneering study on species selection in Cretaceous mollusks during the long background interval preceding the mass extinction at the period's end. He found that species with planktotrophic larvae (defined as floating and feeding, and therefore remaining aloft for substantial time) generally have larger geographic ranges and longer geological durations than species with nonplanktotrophic larvae (defined as either never planktonic, or floating without feeding, and therefore aloft for only a short period).
Jablonski supplies good inferential evidence for the two key claims that a hypothesis of species selection requires. First, he presents a strong case that geographic range not only correlates with longevity, but also helps to cause the extended duration. Species tend to reach their maximal range soon after their origin, and to maintain this breadth thereafter as a potent hedge against extinction. Second, he calculated a strong heritability for geographic range by assessing the parent-offspring regression for this character. Geographic range surely constitutes a character of the species, not (obviously) of individual organisms. This trait confers an emergent fitness on species that gain increasing longevity thereby. All necessary attributes for an interpretation based on species selection have therefore been identified.
(The case also includes interesting complexities. As mentioned previously for similar examples in Tertiary mollusks, nonplanktotrophic species generally experience shorter longevity and maintain smaller populations in their more restricted ranges; but they also speciate more frequently, a presumed consequence of greater ease in forming isolated populations — for their evolution of larvae without extensive periods of flotation restricts gene flow among demes. Thus, the greater longevity of planktotrophic species need not [Page 710] imply increasing dominance of such species within the clade, for this positive trait can be counterbalanced by the higher speciation rates of shorter-lived nonplanktotrophic species. Moreover, Jablonski also showed that selective forces can change radically during episodes of mass extinction. In the great dying at the end of the Cretaceous period, geographic range of species shows no correlation with survivorship through the event. But, interestingly, geographic range of entire molluscan clades (though not of their component species) does correlate positively with persistence through the mass extinction — a potential example of clade selection.)
I freely admit that well-documented cases of species selection do not permeate the literature. But I regard this infrequency as a great opportunity, rather than a restrictive limitation or an indication that the phenomenon scarcely exists. We have barely begun to acknowledge (much less to define or operationalize) this process, and we have still not entirely agreed upon criteria for recognition. We face the tradition of a full century spent not considering causes at this level (indeed, actively denying the existence of such levels at all). We are just learning how to look — or, to state the issue more incisively; we have just begun to recognize that we should be looking at all! We face all the promise of a rich but unploughed field — and (to commit two literary barbarisms of mixed metaphors and parodied quotations at the same time), we should summon up the courage of John Paul Jones and recognize that we have not yet begun to think.
I regard the second, or theoretical, objection as even more unfair in its purely traditionalist grounding in the parochialism of viewing organisms as exclusive agents of evolutionary interest or importance — more an aesthetic defense about comfort or preference than an intellectual argument about mechanisms. Several Darwinian strict constructionists, Richard Dawkins and Daniel Dennett in particular, hold that almost everything of interest in evolutionary biology either inheres in, or flows from, natural selection's power to craft the intricate and excellent design of organisms — “organized adaptive complexity,” in Dawkins's favorite phrase. “Biology is engineering,” Dennett tells us again and again in his narrowly focussed book (Dennett, 1995).
I do not deny either the wonder, or the powerful importance, of organized adaptive complexity. I recognize that we know no mechanism for the origin of such organismal features other than conventional natural selection at the organismic level — for the sheer intricacy and elaboration of good biomechanical design surely preclude either random production, or incidental origin as a side consequence of active processes at other levels. But I decry the parochialism of basking so strongly in the wonder of organismic complexity that nothing else in evolution seems to matter. Yet many Darwinian adaptationists adopt this narrow and celebratory stance in holding, for example, that neutrality may reign at the nucleotide level, but still be “insignificant” for evolution because such changes impose no immediate effects upon organismal phenotypes; or that species selection can regulate longstanding and extensive trends in single characters, but still maintains no “importance” in [Page 711] evolution because such a process can't construct an intricate organismal phenotype of numerous, developmentally correlated traits.
Dawkins (1982, pp. 106-108), for example, damns species selection with faint praise in these terms:
I shall argue that a belief in the power of species selection to shape simple major trends is not the same as a belief in its power to put together complex adaptations such as eyes and brains ... The species selectionist may retreat and invoke ordinary low level natural selection to weed out ill-coadapted combinations of change, so that speciation events only serve up already tried and proved combinations to the sieve of species selection. But this “species selectionist” ... has conceded that all the interesting evolutionary change results from inter-allele selection and not from interspecies selection, albeit it may be concentrated in brief bursts punctuating stasis ... The theory of species selection ... is a stimulating idea which may well explain some single dimensions of quantitative change in macroevolution. I would be very surprised if it could be used to explain the sort of complex multidimensional adaptation that I find so interesting.
This statement commits the classic intentional fallacy of the prosecutor: attributing beliefs not held to adversaries, and then castigating them for apostasy (or praising them for good sense in recantation) — as illustrated by the paradigm for an opening thrust in a line of inquiry: “when did you stop beating your wife?” Dawkins finds the adaptive complexity of organisms uniquely interesting. I also regard the subject as fascinating, and I would never attribute this quintessential property of organisms to selection at some other level. I fully acknowledge, as do all species selectionists, that the adaptive complexity of organisms arises primarily by causal processes operating at the organismic level.
But this pluralistic principle applies equally well to other levels. If adaptive complexity marks “what organisms do,” and must therefore be explained at the organismic level — then “what species do” implies a consideration of causation at the species level. Species “do” two pri
mary things in macroevolution: they carry trends within clades across long geological stretches of time, and they stand as basic units (geological “atoms” if you will) for counting the waxing and waning of differential diversity through time (why does our current biota feature 500,000 named species of beetles, but fewer than 50 of priapulids?). As a paleontologist, I regard these two phenomena as surpassingly important, while I remain happy to grant Dawkins's commanding interest in the adaptive complexity of organisms. But just as I try not to impose my causes (for other scales and levels) upon his material a priori, I ask him to acknowledge the importance of my favored themes within a comprehensive evolutionary theory (even if they do not engage his personal concern), and therefore to recognize the efficacy of different appropriate causes at this paleontological level. In short, Dawkins and others commit a classic psychological [Page 712] fallacy in denying status to species selection by confusing personal interest with general importance.
Only one line of defense remains open to those who still wish to deny the importance of species-level processes after correcting this psychological fallacy, and admitting that trends and changing patterns in diversity rank as vital subjects in a complete evolutionary theory, and also represent “what species do.” Such a Darwinian stalwart must argue that all (or nearly all) phenomena at the species level find their causes in upward translation from ordinary natural selection on organisms. Thus, if current biotas feature half a million species of beetles, this plethora can only imply that beetle organisms maintain a particularly favorable adaptive design. And if geological trends privilege increasing body size, larger brains, more complex ammonite sutures, more symmetrical crinoid cups, fewer horse toes, and a thousand other documented patterns, these features must triumph by their adaptive value to organisms. I shall make no further arguments against such a narrow perspective here (to save my rebuttal for Chapter 9, pp. 886–893), and will only quote a great American character, Sportin' Life in Porgy and Bess, to remind us that received wisdom does not always prevail:
The things that you're liable to read in the Bible
It ain't necessarily so.
THE CLADE-INDIVIDUAL Although a logical space must exist in our structure of explanation for this highest level of the evolutionary hierarchy, I am not sure that clade selection plays a major role in evolution. Most clades contain so few parts (species) that their waxing and waning must often occur by processes that either operate as random inputs to the clade level, or result from selection among subparts (species selection, or lower-level selection), and therefore appear as drives at the clade level (and not as selection among entire clades treated as individuals). Secondly, while I have advocated a plurality of mechanisms for coherence of individuals at various levels in the hierarchy, I do have trouble in conceptualizing an adequate “glue” for clades, especially since their parts (species) may live in such complete independence, and in such different ecologies, on distant continents. Finally, clades maintain the peculiar property (perhaps only an odd “allometric” consequence of necessary structure at this highest level, and not any compromise in efficacy) of necessarily originating as a single subpart — the founding species, and gaining definition (as a full level) only retrospectively, after adding new parts (more species) sequentially.
How then, given all these difficulties, could clades compete, qua clades as discrete and integral evolutionary items, even under the broad definition (see p. 706) that does not require direct contact or even life in sympatry? Is a clade, uniquely among evolutionary individuals of the hierarchy, more a “holding firm” for subparts than a coherent entity frequently subject to selection at its own level?
One route to claiming a potential importance for clade selection remains [Page 713] open, but I am not confident that the argument can prevail (though Williams, 1992, despite his past as an ardent gene selectionist, has become a strong advocate of this view). What do we mean, for example, when we say that dinosaurs died and mammals survived, or that brachiopods dwindled to a remnant while clams continually expanded? Do these descriptive statements imply clade selection? A general argument would have to be framed in the following way: any distinct clade maintains defining autapomorphic characters expressed by all subparts (species). If a clade survives, while another living in roughly comparable habitats, dies — and if survival can be tied to autapomorphic characters held by the persisting clade (and absent in the extinct clade) — may we not speak of clade selection based on a range of variability that includes the key characters in the surviving case, but precludes their expression in the extinct clade?
For example, if mammals survived in part by virtue of small body sizes, and dinosaurs died for a set of consequences related to invariably (and substantially) larger body size, couldn't we say that mammals, as a clade, possessed genetic determinants (shared by homology in all subparts, with homology as the “glue” of cladal coherence) that all dinosaurs lacked as a result of their own evolved cladal distinctions? If such a scenario can count as clade selection (rather than just clade sorting, as an obviously valid description), then selection at this highest level becomes common in nature — for many clades yield in geological time to phylogenetically distant clades that share sufficient similarity in habitat and function to rank as genuine “replacements.”
I am not comfortable with this general argument, for no one has yet articulated firm and operational criteria for distinguishing true clade selection (based on irreducible fitness conferred by a clade-level property) from descriptive clade sorting (or differential survival as an effect of lower level properties belonging to species or organisms, but translating upwards to success or failure of a clade as a geologically persistent entity). Some examples probably do represent genuine clade selection — as in Jablonski's (1987) case of clade survival (through mass extinction), correlated with geographic range of the entire clade, but not with ranges of component species. Most other examples, however, may not invoke any genuine clade-level character (either aggregate or emergent), but only represent the death of each species, item by item (part by part in cladal terms, for this highest-level individual also maintains the peculiar property of relative immunity, especially in clades with large numbers of widely distributed subparts, to the fate of individual subparts). We may frame our best descriptions for such cases in terms of clade sorting, but do they also qualify as cases of clade selection?
At a minimum, however, such arguments illustrate a need for macroevolutionary accounts at all levels, even when causality arises from lower levels and merely affects the fate of higher-level individuals. Thus, the explicit study of macroevolution would remain vital even if traditionalists had been correct in ascribing all causality to organismic selection. But we needn't take refuge in this “minimalist” defense. Causal processes — and not only selection, as I shall demonstrate in the next section — do operate at substantial (often controlling) [Page 714] relative frequency at all levels (with the possible exception of some dubiety about the importance of clade selection as expressed above, and some recognition that organismic selection has effectively squashed most cell-lineage selection in many phyla of multicellular organisms).
I therefore end this section with two statements from George Williams (1992), who once rejected higher-level selection with such verve and skill (1966), but who (while properly reasserting his excellent arguments against the old form of so-called “naive group selection,” or interdemic selection in the Wynne-Edwards modality) now strongly defends both the importance of selection at the species level (“clade selection” of lowest rank in his terminology, because he rejects species as units), and our lamentable failure to consider this vital process in our previous theorizing. Echoing my methodological point that a rarity of recorded examples does not imply any actual weakness in nature, Williams writes (1992, p. 35): “Only the barest beginnings have been made in searching the fossil record for evidence of clade selection. The record can be searche
d for statistically significant trends in diversity and abundance of particular clades ... It can also be searched for consistent selection of certain characters.”
In an expansive and forceful plea for pluralism — representing the finest form of support that a paleontologist could obtain from colleagues engaged in the study of microevolution — Williams (1992, p. 31) then states that allelic change in populations cannot account for evolution because gene-pools function in nature through their entrapment within higher-level individuals operating and interacting as coherent and distinct entities in macroevolution.
The natural selection of alternative alleles, acting largely independently at each locus, is the only force tending to maintain or improve adaptations shown by the ephemeral organisms formed by the ephemeral genotypes. If one could look back to the evolution of our own or any other sexually reproducing species, back to well before the Cambrian, no other fitness enhancing process of any importance would be found. Having taken that position, I must take another. The microevolutionary process that adequately describes evolution in a population is an utterly inadequate account of the evolution of the earth's biota. It is inadequate because the evolution of the biota is more than the mutational origin and subsequent survival or extinction of genes in gene pools. Biotic evolution is also the cladogenetic origin and subsequent survival and extinction of gene pools in the biota.
The Grand Analogy: A Speciational Basis
for Macroevolution
PRESENTATION OF THE CHART FOR MACROEVOLUT1ONARY DISTINCTIVENESS
When Niles Eldredge and I first formulated the theory of punctuated equilibrium in the early 1970's (Eldredge, 1971; Gould and Eldredge, 1971; Eldredge [Page 715] and Gould, 1972; Gould and Eldredge, 1977), we had only the germ of an insight that its tenets could lend support to a generalized theory of macro-evolution, then entirely undeveloped. We did, however, dimly grasp the key notion that punctuated equilibrium might help to grant species a sufficient stability and coherence for status as what we would now call an evolutionary individual, or unit of selection. We developed this insight by groping towards an analogy that, when generalized and fully fleshed out (with apologies for another parochial organismic metaphor of common language!), sets a foundation for macroevolutionary theory. We dimly recognized, in short, that if species act as stable units of geological scales, then evolutionary trends — the fundamental phenomenon of macroevolution — could be conceptualized as results of a “higher order” selection upon a pool of speciational events that might occur at random with respect to the direction of a trend. In such a case, the role of species in a trend would become directly comparable with the classical status of organisms as units of change within a population under natural selection. We wrote (1972, p. 112):
The Structure of Evolutionary Theory Page 114
