The Phyletic Gradualism - Punctuated Equilibrium Debate

Punctuated equilibrium is a model of speciation as uncoupled from selection. Speciation is initiated by a radical mutation, probably affecting the organism's early development. If the mutants can survive and reproduce long enough to adapt to their environment, they are likely to be reproductively isolated from the parent species (they are more likely to make it this far if they are geographically isolated from the parent). Alternatively, hybridization between species may result in offspring with disrupted canalization, otherwise like the radical mutants of the original theory. The result is that new species appear suddenly and change relatively little otherwise. This pattern is the usual one observed in the fossil record.

Phyletic gradualism, also known as the sympatric speciation model, proposes that beneficial mutations occur once in a while and spread through the population, gradually rendering it completely different from what it once was. Also the environment is gradually changing and the population accrues mutations which help it deal with these changes. It works in concert with the allopatric speciation model, which creates a fossil pattern similar to that of punctuated equilibrium. According to this model, divergent selection and drift quickly render a peripheral population morphologically unlike and genetically separate from the parent. Evidence for the latter model has been found in the lab and in nature.

There is no reason, however, why both selective and non-selective forms of speciation should not occur. The two may be distinguishable by measuring fluctuating asymmetry in live or fossil organisms observed to be undergoing speciation.

Introduction

Punctuated equilibrium (Eldredge and Gould, 1972) is a theory about evolution which describes speciation as a intermittent and drastic process rather than a continuous and gradual one. Based on observation of the fossil record, Eldredge and Gould (1972) claim that most of a species' history is a long period of stasis, during which its morphology changes very little, if at all. Individual populations might change both genetically and morphologically due to drift or adaptation, but overall, a species exhibits a net stasis. Small reversible changes (relative to those associated with speciation), accompanied by environmental (depositional) changes are visible in fossil sequences (Lieberman et al., 1995). This stasis is interrupted by abrupt extinction and/or speciation events. So the punctualist "Tree of Life" is more angular and jagged than the traditional diagram showing the descent of species from a common ancestor (Figure 1). Once a new species comes into being, it may spread across the home range of the parent species and possibly outcompete it, which would result in a gradual transition, but with few or no intermediate specimens (depending on how complete reproductive isolation is between the new and old species).

Overall, Eldredge and Gould (1972) claim that changes within a species are either insignificant or so great that they warrant the definition of a whole new species; "most species are discrete at any moment in time. It [taxonomy] has no objective application to evolving continua." (p. 93). The reason that few fossil types are found that are transitional between related species, they argue, is that all the changes occur within one or a few generations. Eldredge (1985) perceives the taxonomic hierarchy as discontinuous. The distinction between populations and species (and between species and higher categories) is critical to the Punctualist viewpoint. The difference is that species are separated by full or partial reproductive isolation. Selective changes, resulting in adaptation, occur within species, but the change between species are not governed by selection (although selection and other factors determines which species survive).

Are Morphospecies and Biological Species the Same Thing?

The two opposing sides in this case are unevenly divided with respect to their perspective on evolution. Many gradualists study the genetics of living organisms whereas punctualism is more popular with paleontologists. This raises the question of whether the two factions are really discussing the same thing. Paleontologists distinguish species by morphology while neontologists separate them by crossing specimens to see if the hybrids are as fertile and as viable as purebreds (in which case they are not separate) and by looking for genetic markers which should be distinct in organisms from two different species in the same habitat (indicating that they have been unable to crossbreed).

Jackson and Cheetham (1994) examined this question by classifying modern bryozoans in the genus Stylopoma, which are found together in the Caribbean, into species using the same skeletal characterstics used for fossil bryozoa. They then checked the genetics of the organisms thus classified and found that they did indeed belong to separate species (p > .9999). So, for the moment (until some brave person runs such extensive tests on other kinds of organisms), its hould be assumed that morphospecies are actually biospecies.

Mechanism of Punctuated Speciation

It is not simply the tempo of speciation, but its mechanism (since that determines the tempo) which is bitterly debated by those who believe in sudden, episodic evolution (punctualists) and those who consider it to be a continuous process (gradualists). Eldredge and Gould (1972) initially invoke "allopatric theory", "for the vast majority of biologists, the theory of speciation." to explain how rapid speciation occurs. However, they are being misleading here. Mayr's version of the allopatric speciation theory, the one that was and is so popular among biologists, depends upon traditional Darwinian mechanisms, those of drift and selection to effect change in the population in question. Eldredge and Gould's (1972) "allopatric theory" resembles Mayr's only in the characteristics of populations that are likely to split off and become new species: small populations, initially on the periphery of the parent species' range, which are subsequently cut off from the parent population.

It is not until the last section of the paper that Eldredge and Gould (1972) explain the actual mechanism of change in their theory: a mutation which causes a temporary breakdown of developmental homeostasis. A developmental mutation might result in not just one big change at a time, but several, depending on how early in development it occurs. The mutant will be born very different from its parents, although possibly not yet reproductively isolated from its parental species. Loss of canalization is likely to be associated with a developmental mutation. Both the mutation and the developmental lability will be passed on, so the mutant's offspring will also be different from normal individuals of the parent species and highly variable. A period of intense selection afterwards wipes out most of these mutants if they have maladaptive or simply inviable phenotypes. Selection quickly restores canalization as well, but with the new morph as "normal", in which case the mutants will no longer be able to interbreed with members of the parent species and are forced to breed only with each other. Selection will reduce genetic variance, resulting in long period of stasis. Eldredge and Gould (1972) argue that it is developmental homeostasis that ordinarily maintains morphological stability and allows members of a species to interbreed.

Why the initial mutation needs to occur in a small population is unclear. Perhaps more important would be a habitat in which selection is relaxed during the period in which the first few mutants are trying to survive and reproduce. An event that might cause the failure of canalization described above would be hybridization between members of two related species or highly separated subspecies (Siems, pers. comm.), which would be more likely to occur in a peripheral population. Galapagos finches are hybridizing on one small island (Grant and Grant, 1992). Elsewhere, Geospiza fortis, G. scandens and G. fulginosa rarely interbreed, if ever. The hybrids found on the one island are morphological intermediates between the parent species with respect to beak shape and are not expected to survive during famine periods, but are actually surviving and reproducing better than purebred finches when food is plentiful. The hybrid offspring and backcrosses are genetically and phenotypically variable and may have the opportunity to speciate (Weiner, 1995). Hybridization, resulting in polploidy, is thought to be a common cause of speciation in plants.

Another non-selective mechanism postulated by Eldredge (1985) that would result in drastic changes and speciation is a shift in the species' mate recognition system. This would induce selection for a different (possibly radically different) phenotype and/or would promote hybridization with a related species (with the effects described above). Just like most developmental mutations, the effects of such a shift would be disastrous in most cases and quickly selected out, but every so often, it would result in a new species, immediately reproductively isolated from its parent (possibly completely isolated).

The reason that Eldredge (1985) thinks that mate recognition systems are important for organisms is that reproductive isolation is selected for in the face of environmental change. Organisms from distinct species have access to a gene pool that contains the variablity that will let their offspring adapt, yet is restricted to "good", co-adapted genes that are approriate for the organism (i.e. a gene for longer legs is more useful to rabbits than to birds). The tendency for similar species to have exaggerated differences in the traits that distinguish them (especially for mate recognition) in overlapping habitats is a case of a species adaptation to remain distinct. Speciation events result in "blurry" new species, which must isolate themselves if they are to survive.

Evidence for Punctuated Equilibrium in the Fossil Record

Although the geographic spread of fossil species may be gradual, their appearance usually is not. Jackson and Cheetham (1994) made a detailed study of Caribbean bryozoa (all species in the genera Metrarabdotos and Stylopoma) and found that new species appeared "fully formed" and remained unchanged for 2.5 million to 16 million years (p > .99 for sixteen species and p > .9 for three species). They found "no evidence that intraspecific rates of morphological change can account for differences between species."

Nehm and Geary (1994) examined detailed stratigraphic record of a speciation event between two snail species (Prunum coniforme gave rise to P. christineladdae and continued to exist alongside its descendent). The pattern is typically punctuated: the transition only lasts for .6 to 2.5% of P. conifome's lifetime and the two species exhibit morphological stasis outside of the transition, but intermediates were observed during the transition. However, selection may have played a role in the split. P. christineladdae was typically found in deeper water than its ancestor and the shelf on which it appeared was becoming deeper during the transition. Punctualists, however, might argue that this simply explains the survival of the new species in the face of competition from the ancestor and that its formation was a matter of chance.

Geary (1992) made an observation that would permit the Punctualists to discount many apparently gradual cases of evolution, complete with intermediates. She found specimens that were morphologically intermediate between Melanopsis fossilis and M. vindobensis (a new species descended from the first). The problem was that distinct specimens of both ancestor and descendent were found in the same sediments. Rather than claim that the M. vindobensis specimens had been washed in from younger sediments, she proposed the possibility that the intermediates were hybrids between the two species. If the descendent species is less likely to be preserved than the parent and the hybrids, due to a thinner shell or slightly different habitat, but can be found in younger sediments, a punctuated case like the one above might be mistaken for a case of gradual speciation.

Gradualist Theories of Speciation

Traditional models of evolution rely upon selection and/or genetic drift, both continuous processes, to induce the genetic changes that result in speciation. They predict that the genes which are being selected upon and which cause species to diverge are mostly additive genes affecting the phenotype directly and predictably. The changes in phenotype are small, but numerous. Two consequences of this is that change will be gradual when viewed over the timescale of many generations and that there should be intermediates in the fossil record.

Simpson (1953) argues that gradual change is necessary for adaptation because, although its genes act in isolation fron one another, the trait they produce do not. Selection is upon the entire phenotype, not upon isolated traits. If several genes, each with its own phenotypic effect, have changed within a single mutant, or one gene with a very drastic effect, it is unlikely that the creature will survive to reproduce. These new traits, in the context of its old phenotype, which was tuned by a long history of selection, will be liabilities.

Phyletic gradualism was described by Eldredge and Gould (1972) as the traditional Darwinistic belief that speciation is driven by the same changes in gene frequency (driven by drift and selection) that cause populations to vary within a species. New species usually result from a transformation of the parent species across its whole range due to an environmental change or to the appearance of an advantageous mutation. In either case, the adaptive genes would spread through the entire population and the older traits would gradually be eliminated by selection. During this time, there will also be selection on other genes to enable the rest of the organism to work in concert with the new adaptation. The problem with this model as an antagonist for punctuated equilibrium is that it is not intended as a complete explanation of speciation. What Eldredge and Gould call "Phyletic Gradualism" others call "Sympatric Speciation". Textbooks and articles which endorse the sympatric model usually present it side-by-side with Mayr's allopatric theory.

Mayr's theory, mentioned above, involves rapid change in peripheral populations resulting in speciation. The new species may reinvade the territory of the parent species. This theory would explain the sudden changes in the fossil record, but not stasis (and does not deny the possiblity or likelihood of sympatric speciation). However, the mechanism of change in Mayr's theory, like that in the sympatric theory, is change in gene frequencies. Selection pushes populations in different adaptive directions if they live in slightly different habitats, resulting in subspecies. Through gradual accumulation of mutations and lack of gene flow and/or through continued different selection, the subspecies continue to diverge until they can no longer interbreed with the sister species. Also, in small populations, genes which are rare in the parent population can drift to high frequency. If gene flow between the peripheral and parent populations is completely stopped, mutations that appear in one population will not spread to the other. Several such mutations might spread through the peripheral population and change it enough that it can no longer interbreed with the parent population.

The allopatric model would generate a punctuated equilibrium pattern in the fossil record using accepted mechanisms. A population which is as well-adapted to its environment as their pool of available genes will permit is unlikely to change. Fisher's Fundamental Theory about adaptation states that the process of adaptation will proceed until the environment changes, selecting on other genes or until a new and beneficial mutation crops up, as discussed above. This kind of change has been observed when directional selection was applied to organisms.

Evidence for Selective Speciation in the Modern World

Seeley (1986) compared modern specimens of a snail, Littorina obsata, to century-old specimens from a museum. The morphology of the two sets of specimens is radically different. About a century ago, a predatory crab moved into the home range of these snails, which rapidly developed a more durable shell. There are a few snails with pre-crab morphology living in areas isolated from the predator, and these are not prezygotically reproductively isolated from snails with the new, durable shell morphology. There has been no selection for reproductive isolation because the isolated snails have little opportunity to exchange genes with their conspecifics who face predation, but hybrids would be expected to be less fit in either environment.

According to both the allopatric and the sympatric models, speciation is simply a matter of time, an accumulation of continuously- accrued change, not the result of a most unlikely mutation. If this is so, speciation should be visible in the lab, given observation over a period of generations. Rice & Hostert (1993) conclude, after examination of laboratory experiments on the causes of allopatric speciation, that the loss of gene flow is less important in inducing reproductive isolation between populations than is divergent selection. Apparent complete reproductive isolation has been attained several times in Drisophila under disruptive selection regimes (Rice & Hostert, 1993). Selection causes rapid genetic change. Post-zygotic reproductive isolation is easy to induce under disruptive selection, because the hybrids are simply less likely to live. Prezygotic reproductive isolation was believed to be either a pleiotropic effect of one of the genes whose frequency changed under selection or the effect of some gene so tightly linked to a selected gene that it was highly unlikely to be separated by recombination. Rice & Hostert (1993) do not discuss developmental canalization in their review. None of the experiments they describe ran for more than a few years, although many achieved apparent speciation during that time.

The divergent selection model of speciation seems to be supported by field research on hybrid zones (Barton & Hewitt, 1989). Not only are hybrids less viable than purebreds in many species and subspecies, but the area in which they can be found, at the meeting point of the (sub)species' territories, is far narrower than one would expect given the organisms' capacity for dispersal. Apparently, the (sub)species are restricting themselves to separate home ranges and the hybrids are not surviving to disperse away from the zone. Barton & Hewitt (1989) propose disruptive selection, imposed not only by the variation in environment (which is usually not abrupt along the hybrid zone), but also by the phenotypes of the organisms themselves. Like Simpson (1953), they argue that a difference in one trait may require differences in several others to be as functional as the original trait.

Gradualism in the Fossil Record?

Punctualists would not deny that there are many examples of slow, continuous change in the fossil record. The problem is that well- documented examples do not include speciation events. Although incremental changes are visible in fossil lineages, they do not eventually result in a species transforming into another species. Many of the continuous changes seem to occur within species or across genera, plodding along as other, radical changes occur, marking speciation events. Eldredge (1985) remarks that throughout the history of genus Homo, there are many instances of increasing brain size and no cases of reduced brain size, resulting in a continuous trend if viewed across a long time-scale. But punctuated equilibrium does not deal with such changes.

Wei (1994a) observed continuous, non-reversed change in Globoconella forams. G. inflata differs from its ancestor, G. puncticulata, in its size, degree of peripheral roundness and the number of chambers in its test, but the older specimens resembled the ancestor in having similar number of chambers. If all three differences are considered, it took G. inflata almost a million years to differentiate. However, G. inflata is quite distinct from the parental species even if only the first two traits are considered. It appeared abruptly in the fossil record and coexisted with G. puncticulata for about a million years. During that time, no intermediates were visible. The change in chamber number occurred as the ancestral species was going extinct and was accompanied by size and roundness changes that made G. inflata the same size and shape as its extinct ancestor (which was no longer around to offer competition). The reason that Wei (1994b) did not define the specimens showing these changes as a different species is that G. inflata is highly morphologically variable. Wei (1994b) attributes both the speciation event and the later transformation to heterochrony, and the replacement of one type by another to selective climatic change.

Part of the dearth of fossil evidence may be due to the fact that many of the fossil records that can test questions about speciation can only come from marine environments. Continuous high-deposition records, in which speciation events would be visible, are available in some continental- shelf deposits and sometimes in the deep-sea, but not on land. It may be that punctuated evolutionary patterns are more common in near-shore organisms than in terrestrial ones, but this would be a difficult question to test.

The Punctualist Response

Punctuated equilibrium is considered non-Darwinian because one of its central tenets is that adaptation and speciation are uncoupled: separate and unrelated processes. A few speciation events may be driven by selection and some may even occur slowly, the punctualists will admit, but these are the exception and not the rule. Generally, Gould (1980) argues, selection is not a source of innovation, but rather the opposite, eliminating new mutations and gene combinations. The genetic variation found within modern species (and by uniformitarian argument, almost all species), is almost entirely selectively neutral, so continuous processes of adaptation have almost no raw material to work with. Gould (1980) does not propose that there is no adaptative change, but that it is, of necessity, small and reversible, with the necessary variability divided up among several demes. Gradualists would agree that adaptive change is small, but not necessarily reversible, if the organism adapts in a particular direction for long enough.

Nor is modern environmental change conducive to the possibility of adaptive speciation. Bennett (1990) argues that speciation is not the consequence of adaptation, at least since the onset of the Pleistocene two million years ago, because the timescale of climatic change is much shorter than that of organismic change. During the Pleistocene, the Earth was (and still may be) locked into a cycle of dramatic and (from a geologic point of view) short-term climate change. The cycle of ice ages, interglacials and interstadials is approximately a hundred thousand years long while the "lifetime" of a species is a million years long on the average. During the course of a species' existence, it will be forced to adapt to a new or changed habitat every few thousand years, undoing the centuries of adaptation to the old environment. Simply migrating south to find an environment similar to the old one is unlikely to work in most cases, as many ice age environments have no interglacial counterparts and vice versa.

Punctualists are also interested in mass extinction, another intermittant stochastic factor with large effects on the biota. Extinction is a matter of bad luck and speciation is a matter of opportunity, according to the punctualist worldview.

Combining the two

Charlesworth (1990) describes a combined mechanism for speciation using a version of Wright's adaptive landscape metaphor. Latitude and longitude would be the phenotypic values for two traits that are being studied. Each species tends to cluster around a different pair of mean values, separated from other species and held together by gene flow, selection and developmental canalization. The landscape also has topography, with elevation being the fitness, the degree of adaptation to the real environment, of the species located at a given point. Species have adapted as best they can and will remain where they are, as predicted by Fisher's Fundamental Theorem, so all points occupied by species are going to be hilltops. A species on a low hilltop cannot cross the valley separating it from a higher peak by selection, since selection drives a species uphill only, never down. Gradualists accept that drift can, little by little, move a small population downhill and possibly onto the slope of a higher peak, or environmental change can rearrange the landscape so that previously adaptive trait combination now has a lower fitness and selection will then get the population moving again.

Charlesworth's (1990) model starts with a random mutation with a large effect on the organism's phenotype to move that organism off the selective hilltop because its phenotypic value for one or both traits that define the landscape have changed dramatically from the population means. The area of the metaphoric landscape occupied by the species has therefore stretched out in one or more directions. If the organism, and for that matter, the population, is lucky, the new area includes the slopes of a nearby peak so that if the mutant interbreeds with the rest of the population, giving its offspring a chance to "cross over" to the new peak. In the meantime, selection will push the organisms in the vicinity of the peak upwards. In terms of the real world, selection will act on the mutated organisms to make them more coordinated and better adapted to their environment.

The example Charlesworth (1990) uses for his model is that of mimetic butterflies. A radical mutation renders a cryptic butterfly highly visible, but if it is lucky, it will be the same color as another, poisonous butterfly in the same area, so it may escape predators if it can fool them from a distance. Its offspring will be variable and some will resemble the poisonous butterfly more than others and selection will gradually (quickly on a geologic timescale) whittle a near-perfect imitation out of the mutants. Selection for reproductive isolation will also be in effect, as hybrids will be neither cryptic nor effective mimics.

This theory is very much like the punctualist theory described above, but Charlesworth (1990) emphasizes selection in his description and in fact calls it a classic Darwinian model in opposition to mutationist and punctualist models, which, he claims, dismiss selection entirely.

Testing Theories of Speciation: Fluctuating Asymmetry

Fluctuating assymetry would identify less canalized genomes (Jablonsky and Bottjer, 1990) and distinguish directional from stabilizing selection (Moller and Pomiankowski, 1993). Traits undergoing intense directional direction have decreased developmental canalization anyway and will show increased variability and asymmetry, which should be visible in fossils from the proposed punctuation event. If a dramatic change is noted without these characteristics, that would indicate a simple sedimentary gap. Traits subject to stablizing selection, as should be the case during periods of stasis, should be more symmetric and less variable. However, inbreeding, hybridization and environmental stress have also been found to increase fluctuating asymmetry (Moller and Pomiansky, 1993).

Distinguishing the effects of large random mutations from stabilizing selection might be difficult, but large mutations should cause asymmetry and variability in traits whose mean values have not changed. Additionally the asymmetry and the variablity found in traits whose mean values do change should not necessarily be proportional to the amount of change. Sadly, I can find no evidence in the literature of anyone actually having made these measurements on fossil specimens.

Conclusion

A serious problem is that punctuated equilibrium is a not well-defined theory. It may be impossible to define the differences between it and "gradualism" with respect to how fast speciation occurs and what happens in between speciation events, but the issue of developmental canalization and how it holds a species together or lets it fall apart may well be a good subject for study, given how far and how fast the study of development has progressed in the late 20th century.

Significant evidence has been offered to support the punctualist model and Darwinian allopatric/sympatric models of speciation. Potentially, several mechanisms of speciation have operated to generate Earth's biota. Some would be more important in certain environments at certain times. For example, now, in the late Holocene, with a global climate that has been relatively stable for ten thousand years and with well-defined and characterized environments, gradualistic forms of speciation should be in evidence, given the thousands of years of potential adaptation. However, at the end of the last glacial or after one of the many mass extinctions that Earth has undergone, new low-selection environments were opened up for colonization providing opportunities for radical low-fitness mutants to survive and adapt. Likewise, species ranges changed radically, allowing the possibility of hybridization. The question of which kind of speciation is most common and/or important can be raised again, but more data are needed to answer it (or determine if it really needs answering).

As for the politics of punctualism: it may well be time to admit large, non-selective mutations into the Darwinian synthesis as an important evolutionary force, just as drift was included earlier in the twentieth century, now that we know more about the developmental processes on which they act. Likewise, other massive and occasional stochastic events are important as extinction forces. These are necessary to explain group selection within species, another concept recently integrated into the Darwinian synthesis (Sloan Wilson, 1983). Over the very long term, say an era, all changes are gradual because selection continues to operate, whether or not speciation occurs and regardless of the mechanism. Both mass extinction and adaptation are going to have an effect at every level of phylogeny. Group selection was given a new lease on life in the 1980s (Sloan Wilson, 1983); mutationism may also be made sensible in the 1990s.

References

Barton & Hewitt. 1989. Adaptation, speciation, and hybrid zones. Nature 341: 497-503.

Bennett, K.D. 1990. Milankovitch cycles and their effects on species in ecological and evolutionary time. Paleobiology 16, 11-21.

Charlesworth, B. 1990. The evolutionary genetics of adaptations. In Evolutionary Innovations. Ed. by M.H. Nitecki. University of Chicago Press. Pp. 47-70.

Eldredge, N. 1985. Unfinished Synthesis: Biological Hierarchies and Modern Evolutionary Thought. Oxford University Press.

Eldredge, N. and Gould, S.J. 1972. Punctuated equilibria: an altenative to phyletic gradualism. In Models in Paleobiology, T.J.M. Schopf, ed. San Francisco: Cooper & Co. Pp. 82-115.

Geary, D.H. 1992. An unusual pattern of divergence between two fossil gastropods: ecophenotypy, dimorphism or hybridization. Paleobiology 18, 93-109.

Gould, S.J. 1980. Is a new and general theory of evolution emerging? In Evolution Now: A Century After Darwin. Ed. by J. Maynard Smith. W.H. Freeman & Co. Pp. 129-145.

Grant, P.R. and B.R. Grant. 1992. Hybridization of bird species. Science 256, 193-197.

Jablonsky, D. and D.J. Bottjer. 1990. The ecology of evolutionary innovation: the fossil record. In Evolutionary Innovations. Ed. by M.H. Nitecki. University of Chicago Press. Pp. 253-288.

Jackson, J.B.C. and A.H. Cheetham. 1994. Phylogeny reconstruction and the tempo of speciation in cheilostome Bryozoa. Paleobiology 20, 407-423.

Lieberman, B.S., C.E. Brett and N. Eldredge. 1995. A study of stasis and change in two species lineages from the Middle Devonian of New York state. Paleobiology 21; 15-27.

Moller A.P. and A. Pomiankowski. 1993. Punctuated equilibria or gradual evolution: fluctuating asymmetry and variation in the rate of evolution. Journal of Theoretical Biology 161, 359-367.

Nehm, R.H. and Geary, D.H. 1994. A gradual morphologic transition during a rapid speciation event in marginellid gastropods (Neogene: Dominican Republic). Journal of Paleontology 68, 787-795.

Rice, W.R. and E.E. Hostert. 1993. Laboratory experiments on speciation: what have we learned in 40 years? Evolution 27, 1673- 1653.

Seeley, R.H. 1986. Intense natural selection caused a rapid morphological transition in a living marine snail. Proceedings of the National Academy of Sciences, USA 83, 6897-6901.

Simpson, G.G. 1953. The Major Features of Evolution. Touchstone.

Sloan Wilson, D. 1983. The group selection controversy: history and current status. Annual Review of Ecology and Systematics 14, 159-187.

Wei. K.Y. 1994a. Stratophenetic tracing of phylogeny using SIMCA pattern recognition technique: a case study of the late Neogene planktic foramnifera Globoconella clade. Paleobiology 20, 52-65.

Wei. K.Y. 1994b. Allometric heterochrony in the Pliocene/Pleistocene planktic foraminiferal clade Globoconella. Paleobiology 20, 66-84.

Weiner, J. 1995. Evolution made visible. Science 267, 30-33.