© 2000 by The Society for Integrative and Comparative Biology
Axial versus Appendicular: Constraint versus Selection1
1 Biomechanics Laboratory, Department of Orthopaedic Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205
 |
SYNOPSIS |
|---|
TOP
 SYNOPSIS INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
SYNOPSIS.Constraint theory has suffered from its many and varying definitions and from repeated confusion with the action of selection. Constraints are difficult to isolate, more difficult to quantify, and their consequences may be simulated by the action of agents such as stabilizing selection. This paper presents an alternative method of constraint analysis using a simple one-dimensional measurement. A total of 29 skeletal elements from 15 species representing 11 orders of mammals were measured for length and normalized for body mass, after which means, standard deviations, and variances were generated. A coefficient of variation analysis was performed to normalize for mean element length. The axial skeleton was found to be less variable than the appendicular. The appendicular demonstrated a trend where the more distal elements were the most variable, and variation decreased with more proximal positions. The three most variable of the 29 elements were finger V, toe V, and metacarpal V. In summary, the axial skeleton was found to be more conservative in the lengths of its elements, the more distal appendicular elements were less constrained than proximal ones and these constraints were probably the results of genetic, developmental, and mechanical factors. It is also proposed that stabilizing selection played some role in maintaining the curb on length variance in these structures, based on mechanical performance. The results of this study are intended to promote discussion of alternative methods of constraint analysis.
|
INTRODUCTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The morphologies of individual organisms within any population are the result of complex interactions between genetic information, environmental influences during development, selection pressure, various constraints (physical, chemical, physiological, developmental, etc.), drift, phylogenetic inertia and historical contingency. Some of these factors are discrete, such as selection, and some are overlapping such as historical contingency and constraint. Perhaps the most controversial of these concepts is constraint. Evolutionary constraints are non-differential (non-selective) circumstances or processes that prohibit certain evolutionary trajectories.
Constraints are often posed as a question as in "Why are there no live bearing birds?" or "Why are there no species with high fecundity and large offspring?" The questions imply the presence of an evolutionary vacuum in species diversity and that some controlling factor(s) is/are disallowing potential evolutionary trajectories. Constraint may be as clear as a nerve conductance velocity being constrained by the speed of light or they may be as complex as selection for greater skull and brain volume being constrained by birth canal diameter (skulls may expand after birth).
Constraint and selection are often depicted as juxtaposed forces in competition and/or in concert for the fate of the evolution of organisms; selection passively driving morphology, and constraint confining or limiting it. Selection was explicitly defined by Darwin (1859)
as the differential survival of types, and is seen as a driving force in evolution, sifting populations in the direction of the fittest variants. Selection has since been the subject of numerous and refining landmark publications including Bumpus (1899), Dobzhansky (1947), Fisher (1930), Gould and Lewontin (1979), Huxley (1942); Kettlewell (1956), Lewontin (1970), and Wright (1949), to name a few. However, despite the fact that Darwin (1859, 1876) himself recognized the potential for the action of constraint in evolution (or something that could loosely be called constraint), this concept has not been so aggressively pursued as selection theory. Further, although it may be safely stated that the gaps in the existing morphospace of organisms are a result of the action of both extinction and constraint, the action of constraint is poorly understood and often confused with other factors. Thus, the significance of constraint in the evolution of life is widely recognized and yet constraint theory both suffers and benefits from its confusion with selection and its invocation by a diversity of fields and philosophies.
Forms of constraint
Many authors rightly attribute the revival of interest in constraint theory to the critique by Gould and Lewontin (1979) of the adaptationist viewpoint of the role of selection in evolution. Gould and Lewontin forced Neo-Darwinists to consider seriously the alternatives to selection theory when describing evolutionary phenomena, and this has generated an immense amount of interest in the action of constraint. One of the problems with the recent popularity of constraint is that different supporters often use alternative terminologies for similar concepts, many emphasizing one agent of constraint over others. For the purposes of this paper I have grouped the many forms of constraint into two basic types, historical and universal. Admittedly this grouping is an oversimplification of the literature but an extensive review of constraint theory is not the subject of this paper.
The identification of historical constraint is what Gould (1989) called "a problem of pure connectivity in decent." In other words, it is contingent upon the phylogenetic history of the organism and the resulting developmental pathways of that organism. Historical constraint proponents generally agree that the phylogenetic history of the organism and its resulting development are significant controlling factors in the evolution of an organism (Alberch and Gale, 1985; Alexander, 1985; Gellon and McGinnis, 1998; Gilbert et al., 1996; Goodwin, 1984; Gould, 1989; Hall, 1996; Gould, 1989; Maynard Smith et al., 1985; Muller and Wagner, 1996; Wake, 1991; Wake and Larson, 1987), and yet many of these proponents describe the same concepts but emphasize different agents and actions. In this way historical constraints are enigmatic and operate under many confounding factors.
Alternatively, universal constraints encompass the physical and chemical laws of the local universe of the organism (earth and its gravity or the oxygen affinity limits of hemoglobin for example). These constraints place limits on organisms via physiology, biomechanics, some aspects of development, and other basic law dependent processes and systems. These are universal because they are common to all organisms within the local universe, and no organism can directly avoid them. (Vertebrates can only grow to a certain body size, but they can circumvent the body size limits imposed on terrestrial life by evolving in a marine environment). But, it may also be argued for example, that an oxygen obligate population could evolve out of its oxygen dependencey, if given enough time. And therein lies the complex nature of this concept and its frequent confusion with selection theory.
Genetic limitations and stabilizing selection
Authors have sometimes pointed at a lack of genetic variation as being indicative of constraint, but this view ignores decades of identification of ample variation in populations, and is supported by the results of natural sampling errors (Charlesworth et al., 1982; Barker and Thomas, 1987; Chetverikov 1926; Lewontin et al., 1970; Lewontin, 1974; Mayr, 1963; Weber and Diggins, 1990; Fisher, 1923; Stebbins, 1966; Wright, 1931, 1951). Indeed, this is a difficult position to defend because "There is no reason for believing that this stability is brought about either by reduction of the size of the gene pool to such an extent that genetic variation is impossible, or by a drastic curtailing of the mutation rate." (Stebbins, 1966).
Similarly, evolutionary constraint is often invoked when in fact stabilizing selection is responsible for the limitation of variation in a population or species (Haldane, 1954; Lewontin et al., 1970), or as Haldane (1954) put it, "Stabilizing selection is widespread in natural populations" and thus, is often mistaken for constraint. Vermeij (1987) has called for a description of constraint that is free of selection and this call has been echoed at least in part by Gould (1989) and Schwenk (1995). Vermeij';s "forbidden phenotypes" are occurrences in reduced selection environments where new and unprecedented variants seem to appear because of a loss of a constraint. Vermeij attributes the appearance of these forbidden phenotypes to relaxed stabilizing selection, suggesting that selection and not constraint was the responsible agent for reduced variability. "In this case, limits to variation are set by minimal ecological standards, not by geometrical or ontogenetic constraints." (Vermeij, 1987). The result of this confusion between selection and constraint is perhaps best characterized by Stearns (1986), "One could describe all of biology as the consequence of constraints.... The meaning of the word would vanish."
Thus, constraints are sometimes a black box into which qualitative and observational data are fed and processed without regard to the ultimate origin of the data, which in turn leads to constraint being confused with selection. Recent authors (Futuyma, 1995; Schwenk, 1995; Vermeij, 1987; Schwenk and Wagner, 1999 manuscript, and others) have attempted to sift through the current constraint paradigm, and have begun what may be a renaissance of sorts for constraint theory, much like in selection';s "New Synthesis."
Larger scope studies
Evolutionary constraint is most often studied in a single species or population. This method may determine relative constraint levels within a population or species, but small populations and single species are more likely to be uniformly under stabilizing selection, drift, or weak directional selection, than would larger populations or multiple species. This is problematic because the results of any combination of these forces could mimic the action of constraint. As an example, the horseshoe crab Limulus has remained relatively unchanged morphologically for hundreds of millions of years, and it was believed that limited mutation rates or another constraining genetic factor must be responsible. It was found however, that temporal uniformity in the organism/environment relationship (and thus stabilizing selection) was probably responsible for its stasis, rather than constrained mutational variation (Lewontin et al., 1970). Therefore, it is likely that a larger, more wide ranging sample of species will reflect the morphological potential of constraint, and not the action of stabilizing selection or drift. It is proposed that a sample including a greater diversity of species will increase the likelihood of constraint detection and reduce the influence of stabilizing selection.
Summary
Constraint theory remains a very subjective topic, with as many views of constraint as there are disciplines in the study of evolution. These different views of constraint have lead to confusion in its identification, mode of action, and quantification. In particular, the results of selection can mimic constraint and thus morphological vacuums may simply be the result of a lack of an appropriate selection pressure or conversely, the action of stabilizing selection. Thus, constraints of the physiological, developmental, and phenotypic nature remain a black box, and only through direct empirical investigation can we assess the likelihood that the course of evolution is guided by patterns of genetic and phenotypic variation (Futuyma, 1995).
It is proposed that the successful quantification of constraint may be obtained through the use of a greater diversity of taxa, measuring constraint in relative terms, and with a conscious and persistent separation of selection and constraint (Schwenk, 1995; Schwenk and Wagner, 1999 manuscript). It may be that we can only infer the presence of constraint, but if this inference is translated to a relative scale, then at least constraint will be semi-quantifiable. However, the need for a more precise characterization and empiricization of constraint needs to be addressed.
The Study
The purpose of this study was to stimulate discussion of alternative methods for the analysis of constraint in the evolution of organisms. This was an attempt to quantify constraint in a relative fashion by comparing coefficients of variation among structural element lengths in the mammalian skeleton. Admittedly, this was an unconventional approach and has its weaknesses, but it was felt that an alternative approach was justified in order to estimate relative constraint levels and further understand the action of constraint during the evolution of organisms. An attempt was also made to recognize the effects of agents other than constraint, for it is certainly true that limitations to skeletal length variation are not completely defined by constraint, and that selection has simply not fully explored the available potential morphospace. So not only is stabilizing selection likely to have some role in evolutionary stasis, but the efficacy of selection to populations may have its limits and that potential needs to be recognized.
Thus, the efficacy of selection may always be in question, and it may have to be enough that we recognize its potential for underestimating the volume and scope of potential morphospace explored by evolutionary trajectories. Additionally, no attempt was made to completely separate the actions of constraint and selection: it is believe that they will be acting in tandem at all times on the many genetic, developmental, and structural interconnections of any particular character. It is the proportion of constraint and selection in any particular case that needs to be investigated, and within those proportions, the relative amount of constraint in different circumstances (taxa). This paper was simply an attempt to create a metric for measuring the presence of constraint on a relative scale.
|
MATERIALS AND METHODS |
|---|
|
TOP
SYNOPSISINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Length measurements were taken from 29 axial and appendicular skeletal elements (Table 1) representing 15 species of mammals (Table 2), and included only adult specimens. At least three individuals were measured to represent each species (when possible). Appendicular elements are chosen based upon the possession of a consistently predominant dimension, or length. These included the humerus, femur, radius, ulna, tibia, metacarpals, metatarsals, and both fore and hind proximal phalanges. The fibula was not chosen because in many taxa it appeared in a non-weight bearing state and its association with the tibia made clear measurements difficult. Axial elements were represented by the axis, cervical vertebrae 3–7, the middle five thoracic vertebrae (biased toward anterior if the total number of thoracic vertebrae is even), and the last five lumbar vertebrae. Values for each of these regions were averaged to provide single cervical, thoracic, and lumbar means, and individual data was taken for the axis.
|
|
Skeletal elements were measured from proximal articular surface to distal articular surface, and articular endpoints are defined as the perpendicular to the long axis of the bone. Semi-lunar articulations such as on the proximal ulna were measured at the midpoint of the articulating surface. Vertebral bodies are measured without the intervertebral discs, and length is determined by the anterior-posterior axis of the centrum (superior-inferior in bipeds). Like the larger appendicular elements, metacarpals, metatarsals, and phalanges were measured from proximal to distal articular surfaces. To preserve homology these structures were ordered I–V from medial-most to lateral-most. When a species lacked a particular structure, or a certain specimen lacked a particular element, this information was incorporated into the data as a no score.
Linear measurements for each of the 15 species were made on a minimum of three adult specimens, and were averaged for each species. Recorded body mass measurements were preferred, but means for species were taken from Walker (1975) when data was unavailable. The means from all specific data were then established for length measures and body mass. In using the term "axial skeleton" in this study, I will be referring to the cervical, thoracic, and lumbar vertebrae only. By appendicular skeleton I refer to the limbs, with all rectangular elements distal to and including both the humerus and femur. The pelvis and scapula will not be considered in either category here because their associations with both systems complicate the issue.
Size differentials among species were a concern so the species were normalized for the lengths of their bones using body mass in the equation:
|
|
This procedure developed a dimensionless metric, adjusted for body mass, which can be used to compare a "population" of humeri for example, and determine the inherent variation in that structure, across species lines, regardless of body size. This was the basis by which variation was estimated and compared, within and between the axial and appendicular skeletons. Further allowance for the difference in mean size of each skeletal element, relative to the others (femur versus axis for example) was accomplished by generating a coefficient of variation for each of the 29 elements. This was performed using the formula:
|
|
Thus, the length variance, which is proportional to the mean length of each skeletal element and its variance is transformed to a metric scaled for mean length. These adjustments make proportional the different sized species and compensate for mean skeletal element length differences so that variance in the elements is comparable on a relative scale. Thus, skeletal elements that are typically smaller but vary a great deal relative to their length will demonstrate more variability than larger elements that do not vary as much, proportional to their length. This coefficient of variation is the relative scale for constraint in the skeletal elements with the lowest coefficients of variation corresponding to the highest levels of constraint.
Specimens were made available by the National Museum of Natural History';s Division of Mammals, their satellite support center in Maryland and from private collections. Measurements were made using digital calipers, a ruler, and a tape measure. Statistical analyses of the data were completed using Statview.
|
RESULTS |
|---|
|
TOP
SYNOPSISINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The coefficient of variation analysis enabled the generation of a table of skeletal elements ordered from lowest to highest variation (Table 3, Fig. 1) and a figure representing group averages (Fig. 2). The vertebrae were found in the low variability positions, with lumbar (0.197), axis (0.343), cervical (0.349), and thoracic (0.399) in positions 1, 2, 3, and 6 respectively. The long bones were similarly conservative, with radius (0.373), femur (0.388), tibia (0.429), humerus (0.452), and ulna (0.524) in positions 4, 5, 7, 8 and 10. The parallel-structure groups were the most variable among the 29 elements, and were randomly shuffled amongst themselves, with only metatarsals V (0.645), II (0.647) and I (0.678) in three consecutive positions, and metatarsals II (0.647) and IV (0.529) in two consecutive positions. Groups of elements were also averaged together. The axial elements were the least variable, the most proximal appendicular elements were in the middle range, and the more distal appendicular elements were the most variable. Metacarpals (0.644) were slightly more variable than metatarsals (0.605), first toe elements (0.702) were more variable than first finger elements (0.662), and the meta-structures (0.624) were less variable than the more distal first finger and toe elements (0.682). The skeletal elements with the highest variability were metacarpal V (1.01), toe V (0.889) and finger V (0.884), while metacarpal I (0.478) was less variable than any other parallel element.
|
|
|
|
DISCUSSION |
|---|
|
TOP
SYNOPSISINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The spine is, in most respects, less morphologically variable than the appendicular skeleton (McShea, 1993
; Thomas and Reif, 1993). This is why a collection of mammalian vertebrae is not so useful for taxonomic identification as are limb bones or the skull and teeth. The mammalian skull and especially the teeth are an exceptionally variable portion of the axial skeleton. These are separable from the vertebrae because they are an unique interface between the organism and its environment and thus should be subject to robust selection. Reduced variability in the vertebrae is presumed to be due to selection for mechanical performance (Schwenk and Wagner, manuscript), historical constraints via conserved genetic programming (Maynard Smith and Sondhi, 1960), and due to the mechanical dependence of the axial skeleton on the appendicular skeleton. The axial skeleton also protects the CNS from damage, and is thus constrained morphologically in its design (intervertebral joint angle excursion or centrum diameter or length). These constraining factors are beyond the ability of selection to overcome and thus limitations are placed on morphological variation. Further, the thoracic may be the more variable of the regions because of its role in accommodating the forelimbs and the ribs. The lumbar vertebrae are likely the least variable because they act as the load transducers from the hindlimbs to the anterior (superior in bipeds) of the skeleton, and because those loads are relatively large. Thus, the lumbar vertebrae are morphologically stabilized by selection as well as being constrained by conservative genes.
Development of the limb and axial elements follow conserved, predetermined patterns (Erwin, 1999; Holland et al., 1994; Olsson and Hall, 1999), and the production of variation in skeletal elements as well as innervation patterns can both be determined early in embryogenesis (Alberch, 1982; Alberch, 1989; Alberch 1985; Alberch and Gale, 1985; DiSilvestro, 1997; Lande, 1978; Shubin et al., 1997). In the limbs, the fifth elements were the most variable among the fingers, toes, and metacarpals, and were the most variable of all 29 elements. This finding may reflect the lower mechanical dependency on the fifth element during locomotion and general body mass support. It may also reflect a differential constraint on variability due to the generation of AP polarity during development and the limitations it must observe.
The appendicular skeleton is also dependent on the stability of the axial skeleton, upon which it operates. Not only does the axial skeleton not encounter the selection pressure of the environment/locomotion interface as intensely as the appendicular (the intensity of which would depend on the species'; locomotor behavior), but the axial skeleton is constrained by the functional reliance of the appendicular skeleton. It can also be considered that the appendicular system acts as a buffer between the axial system and the environment, so that the axial is not under equivalent selection pressure to be altered for mechanical performance during locomotion. This is not to suggest that the axial skeleton is unimportant in locomotion, but that it simply cannot vary as much morphologically. It may then be predicted that limbless or reduced limb species would have more variable axial skeletons.
Surely historical and universal constraints are active in the differential retardation of variation in the elements of the mammalian skeleton, as surely as is natural selection. What particular part is played by developmental limitations, physiological limitations, mechanical laws, or selection is not known. Indeed, it is not yet clear what analysis method is appropriate to measure and compare morphological variability or constraint levels in skeletal elements. It is believed however that the extremes of anatomical form best represent the outer limits of potential morphospace and thus the outer limits of potential form regulated by constraints. Therefore, this approach should be given some consideration. Perhaps future studies will be successful in incorporating these methods, or other new methods into constraint analysis. What is clear, is that evolutionary constraint needs to be distilled into a more functional and universally applicable concept.
|
ACKNOWLEDGMENTS |
|---|
I would like to thank John Long and Tom Koob for the great honor of presenting with this impressive group, Nozomu Inoue for assistance, Kurt Schwenk for assistance, guidance, and access to unpublished material, Mari Nakamura for morph-video assistance, S. J. Gould for statistical advice, R. C. Lewontin for advice, Doug Futuyma for references, David Wake for reprints, Linda Gordon of the NMNH kindly gave me access to the Smithsonian';s mammal collections, E. Y. S. Chao for my post-doctoral fellowship and the flexibility to conduct this project, and all the members of the Northeast Regional Meeting of the SICB Division of Vertebrate Morphology. I also thank the two anonymous reviewers for their time, effort, and excellent advice.
|
FOOTNOTES |
|---|
1 From the Symposium on The Function and Evolution of the Vertebrate Axis presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 6–10 January 1999, at Denver, Colorado.
2 E-mail: dmckmc{at}hotmail.com
|
References |
|---|
|
TOP
SYNOPSISINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Alberch, P., and E.A. Gale. 1985. A developmental analysis of an evolutionary trend: Digital reduction in amphibians. Evolution, 39:8-23.[Medline]
Alberch, P. 1982. Developmental constraints in evolutionary processes. In J. T. Bonner (ed.)Evolution and development,, pp. 313-332Springer-Verlag, Berlin.
Alberch, P. 1985. Developmental constraints: Why St. Bernards often have an extra digit and poodles never do. Am. Natur., 126:430-433.[CrossRef]
Alberch, P. 1989. The logic of monsters: Evidence for internal constraint in development and evolution. Geobios (Mem. Spec.), 12:21-57.
Alexander, R. McN. 1985. The ideal and the feasible: Phylogenetic constraints on evolution. Biol. J. Linn. Soc., 26:345-358.
Barker, J.S.F., and R.H. Thomas. 1987. A quantitative genetic perspective on adaptive evolution. In V. Loeschcke (ed.)Genetic constraints on adaptive evolution,, pp. 3-23Springer, Berlin.
Bumpus, H.C. 1899. The elimination of the unfit as illustrated by the introduced sparrow. Biol. Lectures, Marine Biol. Lab. Woods Hole, 11:209-226.
Charlesworth, B., R. Lande, and M. Slatkin. 1982. A Neo-Darwinian commentary on macroevolution. Evolution, 36:474-498.[CrossRef][ISI]
Chetverikov, S.S. 1926. On certain aspects of the evolutionary process from the standpoint of modern genetics. J. Exp. Biol (Russian) A2:3–54. Eng. transl. (1961), Proc. Amer. Phil. Soc., 105:167-195.
Darwin, C.R. 1859. On the origin of species.. J. Murray, London.
Darwin, C.R. 1876. The Variation of Animals and Plants Under Domestication. Vol. 2. 2nd ed. D. Appleton and Co., New York.
Dobzhansky, T. 1947. Adaptive changes induced by natural selection in wild populations of Drosophila.. Evolution, 1:1-16.
Erwin, D.H. 1999. The origin of bodyplans. Amer. Zool., 39:617-629.
Fisher, R.A. 1923. Darwinian evolution by mutations. In Eugenics, genetics and the family.. Sci. Papers 2nd Intern. Congr. Eugenics, New York, 1921. Williams and Wilkins, Baltimore, 1:115-119.
Fisher, R.A. 1930. The genetical theory of natural selection.. Oxford Univ. Press, Oxford.
Futuyma, D.J., M.C. Keese, and D.J. Funk. 1995. Genetic constraints on macroevolution: The evolution of host affiliation in the leaf beetle genus Ophraella. Evolution, 49:797-809.[CrossRef]
Goodwin, B.C. 1984. Changing from an evolutionary to a generative paradigm in biology. In J. W. Polard (ed.)Evolutionary Theory: Paths into the future,, pp. 99-120Wiley, New York.
Gould, S.J. 1989. A developmental constraint in Cerion, with comments on the definition and determination of constraint in evolution. Evolution, 43:516-539.[CrossRef][ISI]
Gould, S.J., and R.C. Lewontin. 1979. The spandrels of San Marco and the Panglossian paradigm: A critique of the adaptationist programme. Proc. Royal Soc. Lond. B, 205:581-598.
Haldane, J.B.S. 1954. The measurement of natural selection. Caryologia, Suppl. 6:480-487.
Hall, B.K. 1996. Bauplane, phylotypic stages and constraint. Why are so few types of animals In M.K. Hecht, R.J. McIntyre, and M.T. Clegg (eds.)Evolutionary biology Vol. 29, pp. 215-261Plenum Press, New York.
Holland, P.W.H., J. Garcia-Fernandez, N.A. Williams, and A. Sidow. 1994. Gene duplications and the origins of vertebrate development. Development, 1994: Supp125-133.
Huxley, J.S. 1942. Evolution the modern synthesis. Allen and Unwin London.
Kettlewell, H.B.D. 1956. Further selection experiments on industrial melanism in the Lepidoptera. Heredity, 10:287-301.
Lande, R. 1978. Evolutionary mechanisms of limb loss in tetrapods. Evolution, 33:73-92.
Lewontin, R.C. 1970. The units of selection. Ann. Rev. Ecol. System., 1:1-18.[CrossRef]
Lewontin, R.C. 1974. The genetic basis of evolutionary change.. Columbia Univ. Press, New York.
Lewontin, R.C., R.K. Selander, S.Y. Yang, and W.E. Johnson. 1970. Genetic variation in the horseshoe crab (Limulus polyphemus), a phylogenetic "relic". Evolution, 24:402-414.[CrossRef][ISI]
Maynard Smith, J., and K. C. Sondhi. 1960. The genetics of a pattern. Genetics, 45:1039-1050.
Maynard Smith, J., R. Burian, S. Kauffman, P. Alberch, J. Campbell, B. Goodwin, R. Lande, D. Raup, and L. Wolpert. 1985. Developmental constraints and evolution. Quart. Rev. Biol., 60:267-287.
Mayr, E. 1963. Animal species and evolution.. Belknap, Harvard Univ. Press, Cambridge.
McShea, D.W. 1993. Evolutionary changes in the morphological complexity of the mammalian vertebral column. Evolution, 47:730-740.[CrossRef]
Olsson, L., and B.K. Hall. 1999. Introduction to the symposium: Development and evolutionary perspectives on major transformations in body organization. Amer. Zool., 39:612-616.
Schwenk, K. 1995. A utilitarian approach to evolutionary constraint. Zoology, 98:251-262.
Schwenk, K., and G.P. Wagner. 1999. The reconciliation of selection and constraint. Manuscript. .
Stearns, S.C. 1986. Natural selection and fitness: adaptation and constraint.. In D.M. Raup, D. Jablonski (eds.)Patterns and processes in the history of life,, pp. 23-44Springer-Verlag, New York.
Stebbins, L. 1966. Process of organic evolution.. Prentice Hall, New Jersey.
Thomas, R.D.K., and W.E. Reif. 1993. The skeleton space: A finite set of organic designs. Evolution, 47:341-360.[CrossRef][ISI]
Vermeij, G.J. 1987. Evolution and escalation: An ecological history of life.. Princeton Univ. Press, New Jersey.
Wake, D.B. 1991. Homoplasy: The result of natural selection, or evidence of design limitations?. Amer. Natur., 138:543-567.[CrossRef]
Wake, D.B., and A. Larson. 1987. Multidimensional analysis of an evolving lineage. 238:42-48.
Walker, E.P. 1975. Mammals of the world.. Johns Hopkins Univ. Press, Baltimore.
Weber, K.E., and L.T. Diggins. 1990. Increased selection response in larger populations. II. Selection for ethanol vapor resistance in Drosophila melanogaster at two population sizes. Genetics, 125:585-597.[Abstract]
Wright, S. 1931. Evolution in Mendelian populations. Genetics, 16:97-159.
Wright, S. 1949. Adaptation and Selection. In G. L. Jepsen (ed.)Genetics, Paleontology, and Evolution,, pp. 365-389Princeton Univ. Press, Princeton, New Jersey.
Wright, S. 1951. The genetical structure of populations. Ann. Eugenics, 15:323-354.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||