Author's final comments.
If I am to speak to this issue, I want to make my focus very clear. This paper concerns the appearance of biological structure, not the tie of such appearance to biotic descent. Evidence for structural difference/ descent does not constitute evidence for the mechanism by which structural transformation took place. Therefore, the sorts of evidence that simply indicate relationship and/or descent from a common ancestor (e.g., molecular clock data, fossil sequences, chromosomal banding, and other measures of similarity) are not relevant to this question unless they indicate the nature of the creative mechanism that produced novelty during that descent. Evidence of ancestry does not imply knowledge of the morphogenetic mechanisms that are able to produce novelty.
This was perhaps better understood in the nineteenth century than it is today (Muller and Wagner, 1991). Indeed, by 1850, almost all researchers accepted common descent (Gillespie, 1979; Desmond, 1989). The unique implication of Darwin's theory was therefore not descent, but its suggestion that the source of bionic order was to be found in the natural (material) order. For the "Naturalist" (Materialist) of Huxley's Young Guard, natural selection was not simply a theory of mechanism, but a replacement for the Creator (Desmond, 1989; Moore, 1982).
It still is. From the time Darwin proposed it, the central hope of neo-Darwinian theory has been its supposed ability to remove the need for and to take the place of an immaterial designer. According to Stephen Gould (1982), "Natural Selection is a creator-it builds adaptation step by step." As G. G. Simpson (1967) put it,
It is already evident that all the objective phenomena of the history of life can be explained by purely naturalistic, or in the proper meaning of a much abused word, materialistic factors. They are readily explicable on the basis of differential reproduction in populations (the main factor in the modern conception of natural selection) and of the mainly random interplay of the known processes of heredity . . . Man is the result of a purposeless and natural process that did not have him in mind.Clearly, if the biosphere is self-realizing and Unguided, a designer without goals, Richard Dawkins was justified in his remark that "Darwin made it possible to be an intellectually fulfilled atheist " in that sense, Darwin's "scientific' theory forms a necessary support for the beliefs of the committed materialist. This does raise an immediate concern, since it is very difficult for true believers to be objective about proofs concerning the foundational assumptions of their faith.
But has the Darwinian hope proved justified? Or, have cracks in the explanatory plaster been papered over (by faith)? My thesis for this paper is that the plausibility of the neo-Darwinian hypothesis as an explanation for the appearance of biological novelty depends on an inadequately simple model of the genome.
The first problem is a matter of simple logic. How can natural selection be Gould's creator of new morphology when it does not write genetic messages, but only chooses between them? Rather than a creator, it is a critic. It brings no information into the genome, but only selects forms already "created" by the mutated genome. Michelangelo once said he did not carve an angel, he only released it from the rock. For the artist this was modesty, but for selection it is simple truth. The "grain" of the wood being carved, i.e., the informational characteristics of the genome itself and the probability structure of genetic phase space (Brooks et al., 1989, define GPS as the probability space of all possible genomes), determine what selection is able to produce. One cannot select a characteristic not already present; a horse breeder cannot produce Pegasus.
Clearly, then, the nature of the information encoded on the genome, the genetic programs that can be mutated, are central to understanding natural selection's ability to "create." The information structure, however, is far more complex than has been usually assumed. Specifically, the information encoded on the genome reveals hierarchy. It is a hierarchy of described reality, of blueprints for specific cell types, organs, organisms, hives-blueprints of immense stability. Further, these blueprints are organized in a form analogous to a linguistic hierarchy, in which the dehmitions of the "markers" are written in the code they define-e.g., amino acid code in the encoded description of the structure of the aminoacyl proteins. Nor are these simple descriptions of morphology, but a cybernetic hierarchy of controls, with the goals of the more comprehensive levels buffered by flexibility in the lower levels. Finally, those goals are realized through a temporal (developmental) hierarchy, in which the same goal may be achieved by different paths (Muller and Wagner, 1991). The information that dictates error-checked homeostatic/homeorhytic phenomena must itself be error-checked and cybernetic. Clearly, then, there are at least two classes of morphological information on the genome: adaptive and prescriptive. The implication for our topic is that the evidence for selective movement in the adaptive class does not prove that selection can move or formulate prescriptive information.
The other half of the Darwinian engine is mutation. Since the
environment can select only new variants tossed up by the genome (by
mutation, recombination, etc.), for morphology to be due to selective direction,
mutation would have to produce uniform additive phenotypic variation. Mutation,
however, seems to be an adaptive resource utilized by the existing hierarchical
genome, rather than a simple mechanism to "broaden" the phenotype. The existing
genome is the selective agent. Borstnik et al. (1987) report that random
mutation acts as a search process, and Parsons (1987, 1988, 1991) indicates that
mutational rates increase in stressed populations. Pakula and Saner (1989) base the
evaluation of new mutants on the function of the original sequence, and Wilkins
(1980) states that mutated genes must be understood within their higher level
constraints. Belyaev (1979) demonstrated that very high levels of genetic
variability are masked in fox populations, and Wilkins (1980) that assorted eye
mutants all affect the size of the entire eye. Turelli (1988) reports that mutation
rates per character are three-or four-fold higher than the rate per
Can natural selection provide the constraint for the genome? Models designed to explain the bases of such constraints have been problematic. Available genetic diversity is too high to be a constraint. Observed morphologies have moved as much as ten standard deviations in response to selection. Stabile environments (niche space), the ability of populations to track favorable environments, and stabilizing selection all predict low diversity. As Barton and Turelli (1989) put it, "the central paradox is that we see abundant polygenic variation, together with stabilizing selection that is expected to eliminate that variation." Direct replacement (Lande, 1975; Yoo, 1980) for the mutant alleles must be almost identical to those lost to selection (Barton and Turelli, 1989). Barton and Turelli's own stasis theory of pleiotropy predicts only short term stasis; since the dysfunctional effects are diffuse and randomized, the genes involved will be subject to slow replacement. Studies (Turelli, 1988; Wilkins, 1980) that show three-to-four fold higher mutation rate per character than per locus clearly show the constraint of the existing genomic blueprints. And again, the laboratory/field evidence (Endler, 1986; Boag, 1981) of simple changes in gene frequency due to selection, drift, etc., are not significant unless they can be shown to be capable of leading to true novelty, that is, to coherent new morphologies rather than just to shifts in the diversity pattern of the adaptive genome.
Such a complex genome is unlikely to be changed simply by random mutation. It is no block of uniform marble to be chipped away by random hammer blows, but a series of gnarled and knotted instruction sets that must be courted and wooed if change is to be achieved. This is clearly the meaning of the evidence given above for the complexity of mutational change. But mutation is not the only area of investigation that supports a cybernetic information structure on the genome.
For instance, the nature of species is an open question. The best definition of species seems to be based on a cohesion model rather than on reproductive isolation. Paterson (1985) and Templeton (1989) have pointed out inadequacies in the usual species definitions of reproductive coherence or isolation. For instance, parthenogenic species may remain morphologically coherent despite a total lack of interbreeding. Or, in syngameons, interbreeding species may remain morphologically stabile and separated despite millions of years of gene flow. Templeton suggests that species' identity is due to their possession of various genetically based cohesion mechanisms: thus species are characterized by specific individuated genetic controls. Burton and Hewitt (1989) Report that such mechanisms control species boundaries. Further, Vrba's (1984) work with the alcelaphine tribe of African antelope suggests that the whole tribe might be viewed as an entity controlled by such a common coherency.
Cybernetic models of coherency are also important in developmental biology. According to Wagner (1989), "Anatomy emerges at the level of the organ but not at the level of the parts." He refers to control by such sets of developmental constraints as "individuation" or entity formation. The organ is thus ontologically prior to the parts; it defines them and gives them a local '"purpose" or limited "final cause." Bryant and Simpson (1984) also speak of "emergence" as a characteristic of a group of cells committed to form an organ, and error-checked according to norms for that structure. Thus, an adequate understanding of embryonic tissues involves their purpose to the forming organ, and implies the existence of a genomic organ 'blueprint' (Wagner, 1989).
The tension between natural selection and developmental coherency is evident in two recent papers by Weber (1992) and Wake (1991). Weber states that ". . . very small regions of morphology (less than 100 cells across) can respond to selection almost independently ...." Wake states that the common phenomenon of amphibian homoplasy is due to "limited developmental and structural options," i.e., to design limitations. The power of developmental individuations (cybernetic coherencies) is shown in the fact that existing diversity is utilized to ensure morphological stasis rather than directional change.
Wagner (1989) also suggests that homology should be centered around shared entity formation. "Structures from two individuals or from the same individual are homologous if they share a set of developmental constraints, caused by locally acting self-regulatory mechanisms of organ differentiation. These structures are thus developmentally individualized parts of the phenotype." Such a view of homology would give a meaningful approach to several ongoing difficulties, including the phylogenetic reappearance of "lost" structures (e.g.. avian clavicles, Bakker, 1986); alternate inducers of the same organs (Hall, 1983); alternate paths of development in related species (Raff and Kaufmann, 1983); the use of same control genes in different developmental pathways (Marx, 1992) termed genetic piracy by Roth (1988); iterative homology (parallels in repeated organs) (Muller and Wagner, 1991), and the growth of "homologous" organs from different embryonic primordia or germ layers (Wagner, 1989).
Developmental individuation again demonstrates that the GPS is gnarled and knotted rather than uniform marble. Goodwin concludes that ". . . the organismic domain as a whole has a 'form' and is therefore, intelligible (which does not mean predictable) and that the 'content'-the diversity of living forms, or at least their essential features can be accounted for in terms of a relatively small number of generative rules or laws" (Webster and Goodwin, 1982). Such existing blueprints constrain selection into a few possible paths. Rieppel (1990) agrees that some sort of morphogenetic "generative principles" dictates the possibilities of biological form. Such structural rules would restrict living things to parts of GPS that contain permitted morphologies. As Goodwin (Webster and Goodwin, 1982) put it, "A 'generative structuralism' is required in order to solve the problem of the origin of structures." Again, ". . . living organisms are devices which use the contingent 'noise' of history as a 'motor' to explore the set of structures, perhaps infinitely large, which are possible for them." But, how are such curious devices first formed?
Individuations as coherent genetic entities are fundamental biotic realities. But that raises questions: What is the origin of such sets of cybernetic constraints/new individuations? How effectively can natural selection produce or modify such a coherency? Why is it that the evident structure of Genetic Phase Space is so convoluted? What are the density and distribution of coherent, viable blueprints in GPS?
Hard questions. Certainly GPS, which is the probability space of all possible genomes, and thus of all possible genomic coherences, is so large as to be beyond comprehension, much less prediction. The information content of GPS is 2n bits where n is the number of bases. The GPS of genomes of mammalian size (2.5 billion bases) contains around 101,000,000,000 binary bits of information. In contrast, Dawkins's Biomorph land (Dawkins, 1986) has a probability space of only 1015. What can we know of GPS? If Dawkins' little predefined universe contains a mysterious, unpredictable "Holy Grail," surely the probabilities of outcomes in GPS cannot be known. Nor can we demand that it have some specific probability structure so that neo-Darwinism will work. Or rather, we can demand that only if we first assume that neo-Darwinism works, but that competence is what we are trying to prove, is it not?
If direct knowledge of GPS total structure is impossible, all we have is inferences concerning its local structure, which can be drawn only from the pattern of the fossil Record, the record of the search. There is no other evidence. But the fossil record shows an unevenness of rate suggesting coherencies, and that evidence throws doubt on the adequacy of neo-Darwinism as a creative source of new morphology. If it cannot explain, why is it accepted? I note the following problem areas.
1. Life's origin. The origin of life requires the initial encoding of specified blueprints, a non-Darwinian process. Specification involves arbitrary definitions for the "letters" used to write the "messages." How then did specified complexity (blueprints and their described products/"machines") arise from any amount of nonspecified complexity (complex machines, but no blueprints)? Are we really making progress in explaining the source of the genetic code? "The holy grail is to combine information content with replication" (Orgel in Amato, 1992). That is, we need a machine that can write down its own specifications (Thaxton, 1984).
2. Origin of the first animals (Cambrian era). The Cambrian explosion illustrates the abrupt formulation of body-plan constraints (Erwin, et al. 1987). But how within 25 million years (impalas have remained unchanged longer than that) could the full complexity of 70plus metazoan phylum level body-plans arise, and be individuated with error-checking developmental cybernetic controls from protozoans? Remember that protozoans do not have encoded genetic information for morphology due to cellular interaction. How can code that does not yet exist be mutated? Further, given the appearance of new code, how are phylum level morphological "norms" generated, capable of holding for the remainder of the Phanerozoic? As David Jablonski put it, "The most dramatic kinds of evolutionary novelty, major innovations, are among the least understood components of the evolutionary process" (Lewin, 1988).
3. Species stasis. Species show morphological stasis in the face of high levels of selectable diversity (Stanley, 1979 & 1985). But what sort of genetic anchor can hold constant a species' morphological mean and variance for several million years (Michaux, 1989), when enough genetic diversity exists in such species to allow laboratory selection to cause a ten-fold movement of that morphological mean? Are current models of the informational organization of the genome adequate to explain this? This difficulty is reinforced by the still greater morphological stasis shown by the body-plans of the higher levels of the taxonomic system, a stasis that seems to shape, direct, and constrain lower level change in an almost "archetypic" manner. This is hardly the neo-Darwinian prediction.
4. Sudden individuation. New individuation, the appearance of adaptive complexes (morphological entities) is typically very abrupt-for instance, limb structure in Diacodexus (Rose, 1982 & 1987) or the Ichthyostegeds (Coates and Clack, 1991). New "type" forms usually appear suddenly, with the characteristic morphological systems already "individuated"-as defined and error checked entities. (Such definition will almost always require more "bytes' to encode.) Even if possible ancestors that lack the new complex seem to be present (usually at about the same point in time), where do the new control system norms come from? The appearance of new taxa seems to imply the sudden appearance of packages of individuated structural information, but how does closed, error-checked cybernetic feedback start? It may be relatively easy to show that a path across phenotypic space could be progressively adaptive (Kingsolver and Koehl, 1985), but explaining the necessary changes in the underlying genome is a different matter. The two seem identical only because neo-Darwinism has assumed the supply of sufficient additive variability.
The origin of individuation is not an easy question (Müller and Wagner, 1991). To make insects from centipedes, three segments must form a new individuated entity, the thorax. For that to happen, there must be a new set of constraints encoded on the genome for the thorax, rules that define the new entity. Such a rule-set requires a lot more information than did the original repeated structures (segments). The genomic change is far more complex than the phenotypic change. Wagner (1989) states that we have no way "to assess the plausibility of the internalization mechanism . . . the relevant type of data is not thus far available,"
5. Mosaic evolution at morphogenic transitions. Intermediate evidence, when it does exist, usually is mosaic in nature. Mosaic evolution (the movement of one character with stasis in another) indicates the constraints of existing genomic diversity. But, if the characteristic appearance of new suites of characters is similar to that seen in Archeopteryx, then an almost completely established (individuated) character set can be obtained for one organ/structure (flight feathers) with little movement in others (skeletal characteristics) (Wellnhofer, 1990; Sereno and Chenggang, 1992). This makes sense only if the complexity to be realized was already available in the genome. If large-scale morphological change depends on the appearance of a series of new mutations to be selected by a new adaptive niche, should not characters be mutated and move together at rates that are at least comparable?
6. Adaptive radiations. The speed, character, and commonness of adaptive radiations indicate the partitioning and exploration of an occasionally rich genome. Almost all groups at all taxonomic levels first appear in the record as "type" forms, and then "explode" into a number of different lineages with a mosaic of related but not identical potentials for adaptive morphological change (see #5 and the wealth of information in Carroll, 1988; MacFadden and Hulbert, 1988; Larson, 1989). This pattern suggests the partitioning of a very large common genetic package with a high number of alternate morphological potentials. But no known mechanism is available for generating such information-dense primordial genomes. Selection can act only on phenotype, not on hidden genetic potentials. The idea that a "key" innovation opens a "new" adaptive field assumes what needs to be proved about the ability of a genome to be reconfigured in multiple ways. As a matter of fact, a "key" adaptation would be more likely to produce a plethora of pleiotropic dysfunctions.
7. Parallel development in lineages. In adaptive radiations, the diverging lineages will frequently develop in a parallel fashion for a number of characteristics. Such parallels can be quite detailed, suggesting that distantly related species are relatively close. This implies that potentials for the parallel developments were already present in the parental genome as coherent potential blueprints. Thus, "convergent" evolution frequently looks as if it is due more to shared genomic constraints than to shared environments. To what extent can "random" mutations be expected to parallel each other?
So then, we have seen that selection does not "create" anything, but it must already be there for selection to find, and thus biological novelty must be generated by the entire genome. Further, we noted that numerous areas of biological investigation (the nature of mutation, species, development, and homology) point to a genome constructed as a hierarchy of cybernetic individuations. Finally, since the GPS is far too large to predict outcomes, the only way we have to evaluate even its local structure is the fossil record itself. The best evidence for selection appears to be the sorting of packages of existing genetic blueprints, not their creation (or location in GPS). Clearly, the GPS locale being searched by biotic lineages is extremely complex. But the mechanisms for the appearance of such novel packages (or the finding of such remote GPS locations/probabilities) remains mysterious. Thus, natural selection has not been shown to be an effective creator-substitute. It falls at just the point where it must succeed.
Of course, it is possible to postulate a structure for GPS that can he explored by random search processes, a structure that would (if we could see it) predict the world as it exists. In fact, there is no way to prove that GPS is not structured in this fashion. No empirical evidence can be raised against this possibility, because the necessary precursors of the evidence could be "programmed into" the model of GPS. It seems that a blind watchmaker properly programmed into GPS is capable of producing almost anything. But then, such an unknowable watchmaker is not much use in predicting outcomes, even if he is blind. Sounds rather like the creationists' problem.
In conclusion, it seems to me that there is indeed good reason to suppose that metaphysical assumptions have constrained vision in neo-Darwinian biology. Genomes that contain a high level of encoded morphological diversity in the form of error-checked coherent entities seem to appear with regularity. Neo-Darwinism can explain the exploration of such packages, but it has not proved that it can explain their origin. Based on uniform human experience, the simplest explanation for the appearance of a novel, dense pattern of information is an information-dense source. If available DNA templates seem inadequate, the alternative is a source of order exterior to the genome. Are there any known material sources of sufficient density to act as such sources other than human intelligence? Further, if no adequate material source suggests itself is not the remaining logical explanation an immaterial source? Such hypotheses are excluded by the methodological assumptions of science. But-think the unthinkable-is that an adequate reason to reject the possibility? One cannot logically exclude a hypothesis of material inadequacy on the basis of one's a priori assumption of material adequacy.
Neo-Darwinism has been constructed (1) under a metaphysical commitment to (global) materialism, (2) under the methodological commitment of science to use strictly material causal explanations, and (3) under the assumption that good science never lets a problem rest as 'presently unsolved." It follows that in places where material explanations of cause are thin (problematic), they should be treated as anomalies waiting for a more complete (material) explanation rather than as mysteries, or as reasons for reviewing the adequacy of the methodological assumption. And certainly, due to the key role played by neo-Darwinism in the apologetic of metaphysical materialism, thin spots in that theory can be expected to be frequently overlooked even as scientific problems. When recognized, such anomalies are still likely to be shelved with the "best" material explanation attached, not declared unsolved. Indeed, according to Lighuman and Gingerich (1992), such anomalies will probably not even be recognized as anomalies until a new paradigm able to explain them is proposed. If the assumption of global materialism is wrong, that might not happen until that assumption is rejected as necessary.
All of this may do as a working method for a materialist who has faith in God's absence. However, it does not justify telling the theist that, although God may exist (since science cannot prove otherwise), he is unemployed, since undirected material mechanisms have taken over his job. Assumed mechanisms are only assumptions, not proofs. (In any case, theists have never believed that any material event was undirected. How in the world could anyone demonstrate that any material event was not being directed?)
Science has proved neither that the material universe is undirected, nor that our material explanations are adequate. Therefore we should seriously re-examine the conclusions we have reached while working under the materialist agenda. Has anyone seen the emperor's new clothes recently?
Bakker, R. T. The Dinosaur Heresies. New York: William Morrow and Company, 1986.
Barton, N. H. and M. Turelli. "Evolutionary Quantitative Genetics: How Little Do We Know?" Annual Review of Genetics 23 (1989): 337-370.
Belyaev, D. K. "Destabilizing Selection as a Factor in Domestication. "The Journal of Heredity 70 (1979): 301-308.
Boag, P. T. and P. R. Grant. "Intense Natural Selection in a Population of Darwin's
Finches (Geospizine) in the Galápagos " Science 214(1981): 82-
Brooks, D. R., J. Collier, B. A. Maurer, J. D. H. Smith, and E O. Wiley, "Entropy and information in Evolving Biological Systems.'' Biology and Philosophy 4 (1989): 407-432.
Bryant, P. J. and P. Simpson. "Intrinsic and Extrinsic Control of Growth in Developing Organs. Quarterly Review of Biology 59 (1984): 387-415.
Burton, N. H. and G. M. Hewitt. "Adaption, Speciation and Hybrid Zones." Nature 341 (1989): 497-503.
Carroll, R. L. Vertebrate Paleontology and Evolution. New York: W. H. Freeman and Company, 1988.
Coates, M. I. and J. A. Clack. "Fish-like gills and breathing in the earliest known tetrapod." Nature 352.18 (1992): 234-236.
Dawkins, R. The Blind Watchmaker. New York: W. W. Norton and Co., 1986.
Desmond A. The Politics of Evolution. Chicago: University of Chicago Press, 1989.
Endler, J. A. Natural Selection in the Wild. Princeton, NJ: Princeton University Press, 1986.
Erwin, D., J. W. Valentine, and J. J. Sepkoski, Jr. "A Comparative Study of Diversification Events: The Early Paleozoic versus the Mesozoic." Evolution 41.6 (1987): 1177-1186.
Gillespie, N. C. Charles Darwin and the Problem of Creation. Chicago: University of Chicago Press, 1979.
Gould, S. J. "Darwinism and the Expansion of Evolutionary Theory. "Science 216 (1982): 380-386.
Hall, B. K. "Epigenetic Control in Development and Evolution." Development and Evolution. Eds. N. Holder, C. C. Wylie, and B. C. Goodwin. Cambridge: Cambridge University Press, 1983. 353-379.
Kingsolver, J. G. and M. A. Koehl. "Aerodynamics, Thermoregulation, and the Evolution of Insect Wings: Differential Scaling and Evolutionary Change." Evolution 39 (1985): 488-504.
Lande, R. "The Mauntenance of Genetic Variability by Mutation in a Polygenetic Character with Linked Loci." Genetic Research 26 (1975): 221-235.
Larson, A. "The Relation between Speciation and Morphological Evolution." Speciation and Its Consequences. Eds. D. Otte and J. A. Endler. Sutherland, Mass.: Sinauer (1989): 579-598.
Lewin, R. "A Lopsided Look at Evolution." Science 241 (1988): 292-293.
Lightrnan, A. and O. Gingerich. "When Do Anomalies Begin?" Science 255 (1992): 690-695.
MacFadden, B. J. and R. C. Hulben. "Explosive Speciation at the Base of the Adaptive Radiation of Miocene Grazing Horses." Nature 336 (1988): 466-468.
Marx, J. "Homeobox Genes Go Evolutionary." Science 255 (1992): 24-26.
Michaux, B. "Morphological Variation of Species Through Time."Biological Journal of the Linnean Society 38 (1989): 239-255.
Muller, G. B. and G. P. Wagner. "Novelty in Evolution: Restructuring the Concept." Annual Review of Ecology and Systematics 22 (1991): 229-256.
Pakula, A. A. and R. T. Saner. "Genetic Analysis of Protein Stability and Function." Annual Review of Genetics 23 (1989): 289-310.
Parsons, P. "Evolutionary Rates under Environmental Stress." Evolutionary Biology 21 (1987): 311-347.
Parsons, P. "Evolutionary Rates: Effects of Stress upon Recombination." Biological Journal of the Lineman Society 35 (1988): 49-68.
Parsons, P. "Evolutionary Rates: Stress and Species Boundaries," Annual Review of Ecology and Systemarics 22 (1991): 1-18
Paterson, H. E. H. "The Recognition Concept of Species." Species and Speciation. Ed. E. S. Vrba. Pretoria: Transvaal Museum Monograph (1985): 21-29.
Raff, R. A. and J. C. Kaufmann. Embryos, Genes and Evolution: The Developmental-Genetic Basis of Evolutionary Change. New York: Macmillan, 1983.
Rieppel, O. "Structuralism, Functionalism. and the Four Aristotelian Causes." Journal of the History of Biology 23.2 (1990): 291320.
Rose, K. D. "Skeleton of Diacodexis, Oldest Known Aniodactyl." Science 216 (1982): 621-623.
Rose, K. D. "Climbing Adaptations in the Early Eocene Mammal Chriacus and the Origin of the AniOdactyla,'' Science 236 (1987): 314-316.
Roth, V. L. "The Biological Basis of Homology." Ontogeny and Systematics. Ed. C. J. Humphries. New York: Columbia University Press (1988): 1-26.
Sereno, P. C. and R. Chenggang. "Early Evolution of Avian Flight and Perching: New Evidence from the Lower Cretaceous of China" Science 255 (1992): 845-848.
Simpson, G. G. The Meaning of Evolution. New Haven, CT: Yale University Press, 1967.
Stanley, S. M. Macroevolution: Pattern and Process. San Francisco: W. H. Freeman and Co., 1979.
Stanley, S. M. "Rates of Evolution." Paleobiology 11.1 (1985): 13-26.
Templeton, A. R. "The Meaning of Species and Speciations: A Genetic Perspective." Speciation and Its Consequences. Eds. D. Otte and J. A. Endler. Sunderland, MA: Sinauer (1989): 3-27.
Thaxton C B., W. L. Bradley, and R. L. Olsen. The Mystery of Life's Origin: Reassessing Current Theories. New York: Philosophical Library, 1984.
Turelli, M. "Population Genetic Models for Polygenetic Variation and Evolution." Proceedings of the Second International Conference on Quantitative Genetics. Sunderland, MA: Sinauer (1988): 601618.
Vrba, E' S. "Evolutionary Pattern and Process in the Sister-Groups Alcelaphini-Aepycerotini (Mammalia: Bovidae)." in Living Fossils. Eds. N. Eldredge and S. M. Stanley. New York: Springer-Verlag (1984): 62-79.
Wake, D. B. "Homoplasy: The Result of Natural Selection, or the Evidence of Design Limitations?" American Naturalist 138 (1991): 543-567.
Wagner, G. P. "The Biological Homology Concept." Annual Review of Ecology and Systematics 20 (1989): 51-69.
Weber, K. E "How Small Are the Smallest Selectable Domains of Form?" Genetics 130 (1992): 345-353.
Webster, G. and B. C. Goodwin. "The Origin of Species: A Structuralist Approach." Constructional Biology. Eds. H. Wheeler and J. Danielli. London: Academic Press, 1982.
Wellnhofer, P. "Archaeopteryx." Scientific American (1990): 70-77.
Wilkins, H. "Prinzipien der Manifestation Polygener Systeme." Z. Zool. Syst. Evolutions Forschung 18 (1980): 103-111.
Yoo, B. H. "Long-Term Selection for a Quantitative Character in Large Replicate Populations of Drosophila melanogaster .1. Response to Selection." Genetic Research 35 (1980): 1- 17.