Systematics and the Study of Organismal Form and Function Advances in systematics are defining new directions for functional morphology and comparative physiology

George V. Lauder, Ray B. Huey, Russell K. Monson, and Richard J. Jensen

he study of form and function Growth and Form (1917) is a clas- in organisms has a long and A broadly comparative, sic earlv attemDt to understand Tdistinguished history, dating physical causes of biological form. at least from the sixth century B.C., historical approach to The second new area, an expan- the time of the first recorddd me: sion of physiological studies of or- thodical investigations of animal the study of "eanismal functidn. was associated structure and function (Russell with interest in mechanistic aspects [I9161 1982). Investigations into the organismal function is of organismal function. The physi- design of plants and animals gained ological tradition of mechanistic increased visibility in the eighteenth just beginning research on organisms, begun in and nineteenth centuries with the nineteenth-century Europe, diverged work of German poet Johann rapidly from morphological and Wolfgang von Goethe, French natu- flourished as the study of the mecha- structural work. Morphologists and ralist Georges L. C. Cuvier, and nistic basis of animal and plant life anatomists were focusing on the use Swiss zoologist Louis Agassiz. Their began in earnest. A British mor- of morphological data for phyloge- investigations of structure captured phologist of that time, Richard netic analysis, and mechanistic bi- the attention not only of profes- Owen, formulated the concepts of ologists eschewed phylogenetic goals sional naturalists and scientists but homology and analogy-the central for experimental and manipulative also much of the general public. theme underlying all of compara- investigation (Allen 1975, Coleman Physiological investigation also tive biology (Hall 1994). 1977). The discipline of physiology, With the start of the twentieth even comparative investigations of George V. Lauder is a functional and century, the comparative analysis of physiological differences among in- evolutionary morphologist in the De- organismal design expanded into dividuals, populations, or species, partment of Ecology and Evolutionary two new areas. while research in remained strangely divorced from Biology, University of California, Irvine, historical biology continued to be systematics and phylogeny for many CA 92717. He works on vertebrate feed- ~erformedlargely by paleontolo- years. ing and locomotor systems. Ray B. Huey gists, systematists, and comparative In the last 30 vears. with the rise is an evolutionary physiologist in the anatomists. The first new area em- of integrative research areas such as Zoology Department, University of biomechanics and experimental ! Washington, Seattle, WA 98195. He is phasized physical principles that interested in the evolution of locomotor govern the structure of plants and functional morphology, compara- performance in lizards and of thermal animals. This focus on physical bi- tive physiology and comparative sensitivity in flies. Russell K. Monson is ology included mechanistic analy- anatomy have begun to overlap a physiological ecologist in the Depart- ses of developmental patterns by broadly in their subject matter and ment of Environmental, Population, and biologists such as T. H. Morgan, E. techniques of analysis. But even more Organismic Biology, University of Colo- G. Conklin, and E. B. Wilson (build- significant has been the explicit move rado, Boulder, CO 80309. He works on ing on the earlier work of German toward the incorporationof system- the evolution of photosynthetic mecha- anatomist and zoologist Wilhelm atic concepts and hypotheses into nisms in plants. Richard J. Jensen is a Roux in the late nineteenth century; both comparative physiological and plant systematist in the Department of morphological research (Burggren Biology, Saint Mary's College, Notre Allen 1975). There were also gen- Dame, IN 46556. He uses morpho- eral investigations into possible and Bemis 1990, Emerson 1988, metrics to address problems in oak tax- physical determinants of morpho- Huev 1987. Lauder 1990). Within onomy. 0 1995 American Institute of logical features by workers such as the iast dec'ade, many studies have Biological Sciences. D'Arcy Thompson. Thompson's On appeared that integrate compara-

696 BioScience Vol. 45 No. 10 tive morphological and physiologi- cal analyses of organismal design with phylogenetic methods and con- cepts (e.g., Garland and Carter 1994, Huey and Bennett 1987). In many ways, the advances in systematics, phylogeny reconstruction, and com- parative methodology are now de- fining new directions for functional morphology and comparative physi- ology. In the future, an understand- ing of phylogenetic principles and practices is likely to be a prerequi- site for research in comparative physiology and morphology. This new perspective brings increased precision to the selection of species for structural and physiological analysis, to research on the evolu- tionary transformation of form and function, and to the statistical analy- sis of comparative data. Figure 1. Diagram of the coiled shell morphospace. The central cube shows the The comparative study of volume defined by the three axes used by Raup (1966) to model the geometry of the coiled shell. Regions of the cube labeled A, B, C, and D denote the portion of organismal form and function the morphospace occupied by gastropods, ammonoids, pelecypods, and brachio- Research in comparative morphol- pods respectively. Representative computer-generated shell shapes from areas of ogy and physiology lies at the heart the cube are shown also; note that much of the theoretically possible morphospace is unoccupied by either extant or extinct forms. From: Principles of Paleontology of many of the most pressing scien- by D. M. Raup and S. M. Stanley (1971). O W. H. Freeman and Company. Used tific questions in comparative biol- with permission. ogy today, and the recent integra- tion of phylogenetics into the conceptual core of these disciplines occur regularly as new geographic resulted in new analyses of taxo- has redefined many fundamental areas are explored and as extinct nomic diversity and has engendered questions. In this article we con- taxa are discovered by paleontolo- a healthy controversy, rooted in sys- sider several examples in which gists applying new techniques for tematics. of the extent to which the analyses of organismal morphology recovering fossils from previously novel ~&ess Shale morphologies and physiology (largely separately) studied areas. represent fundamentally new groups have provided new approaches or The discovery of novel taxa is an of organisms (Wills et al. 1994).For insights into biological diversity. We important (and yet often underval- example, Briggs and colleagues also consider the integration of these ued) component of research in quan- (1992) have argued that the appar- two disciplines and examples of how titative and evolutionary morphol- ent diversity of morphology (often a phylogenetic framework and more ogy. New discoveries not only termed disparity to avoid confusion rigorous systematic underpinnings contribute to the inventory of bio- with taxonomic diversity) in Bur- have shaped current research on both logical diversity (as new species) but gess Shale taxa is in part 'a'n artifact form and function. they also allow current theories of inadequate phylogenetic knowl- about the evolution of form and edge. Taxa are considered to be dif- Comparative morphology. An im- function to be tested using- these ferent when they cannot be placed portant area in which morphologi- fresh data. into extant clades. As structural cal studies contribute to the knowl- Additional discoveries also may knowledge of Burgess Shale taxa edge of biodiversity is in the overturn previous conceptions of the increases, species previously deemed discovery of previously unknown or- diversity of life in the past, forcing a sufficiently disparate as to consti- ganisms. Organisma] structure pro- reevaluation of models of bioloaicalu tute new higher taxa (and hence vides much of the record of the his- diversification through time. An evidence for so-called explosive tory of life and constitutes the most outstanding recent example of novel adaptive radiation early in the Cam- common means by which new spe- structures discovered in fossil taxa brian) are being placed within es- cies are recognized. Such discover- is the Burgess" Shale fauna. in which tablished monophyletic clades. ies often redefine previously ac- a large number of morphologically Although the discovery of taxa cepted limits of organismal design, distinct species have been found with novel structures provides data and they challenge scientists to ex- (Conway Morris 1992, Gould 1989). on organismal diversity, without a plain novel structures and evolu- The extensive morphological diver- general model of how organisms are tionary patterns. These discoveries gence shown by these species has constructed one would have little

November 1995 from three of these parameters by considering x-, y-, and z-axes to be defined, respect~vely,by the transla- tion rate of the shell coil down the coiling axis, the rate of expansion along the coiling axis, and the dis- tance of the generating curve from the coiling axis (Figure 1).Each point in this three-dimensional sDace marks a theoretical shape deiived from appropriate parameters for each axis. Raup noticed that living Environment taxa occupy asmall region of the (Height of surface vegetation -*) cube of possible shapes. Why is so much of the morpho- space unoccupied? Perhaps, as sug- gested by Raup and Stanley (1971), some areas i-epresent biologically- impossible morphologies or just re- gions in which morphologies are inefficient at performing required functions and have thus been se- lected against. An alternative his- torical (phylogenetic) explanation might" be that the earlv evolution of shell sha~ein these clades may have begun in one direction, and once developmental programs and func- tional relationships among struc- F EDCBA tures became established in the clade. changing to a radically different shell shape was not possible. This type of analysis is important in its use for generating hypotheses about the nature of constraints on morpho- logical evolution and in its defini- tion of potential, not just actual, boundaries to morphological evolu- tion. More recent analyses have used a Figure 2. The interpretation of a structure-environment correlation depends on phylogeny to determine the histori- the phylogeny (Lauder 1981). (a)Plot of six theoretical taxa (A-F) illustrating one cal ~aththat individual clades fol- possible relationship between a structural feature of the taxa (e.g., leg length) and low through the morphospace. An an environmental variable (e.g., vegetation height): Longer length limbs might be advantage of combining phyloge- found in species that live in habitats with higher surface vegetation. (b) Two netic analysis with the definition of possible patterns of genealogical relationship among the taxa. (c) Predicted pattern of environmental change through time if phylogeny 1 correctly depicts the a morphospace is that a filling of the genealogical relationships among taxa A-F. No such prediction can be made if morphospace may be followed, and phylogeny 2 is correct. the location of primitive clades com- pared with that of phylogenetically derived taxa. One such example is idea of the range of possible biologi- possible range of organismal designs, the definition of a shape morpho- i cal designs. Developing a general and one may examine the volume of space for cottid fishes using multi- model into which existing (as well this morphospace that is actually variate morphometrics by Bookstein as yet-to-be-discovered) organisms occupied by living and fossil taxa in and colleagues (1985). A phylogeny can be placed is of immense value in an attempt to understand the pat- was then superimposed onto the 1 analyzing the evolution of biologi- tern of structural diversification in positions of taxa in the morph- i cal design. Thus, one important con- a clade. ospace. tribution of research on organismal The classic example of this ap- In addition to allowing genera- form has been the precise definition proach is the work of Raup (1966), tion of a theoretical morphospace, of a theoretical morphospace within in which coiled invertebrate shells the mathematical modeling of plant which a diversity of forms may be were modeled using four parameters. and animal structure has additional placed. A morphospace defines the A morphospace can be generated benefits. For example, the produc-

BioScience Vol. 45 No. 10 tion of a mathematical model of In contrast to the paucity of (Graves and Somero 1982). Minor organismal design not only abstracts knowledge about the diversity of changes in amino acid sequence un- salient features of structure into a functional attributes, understanding derlie the observed differences in precisely defined set of relationships, of the mechanistic asDects of these kinetic properties and thermal sta- but the relationships among parts of attributes has increased tremen- bility of these muscle lactate dehy- the model can be manipulated to dously over the past 50 years. For drogenases. These comparative bio- 1 generate new forms or to simulate example, comparative physiologists chemical-molecular analyses provide change in morphology either in on- have extensively analyzed how information relevant not only to I togeny or through phylogeny. A physiological processes scale with biogeographers and systematists 1 good example of mathematical mod- body size (Calder 1984, Schmidt- concerned with the factors that I eling of plant design is the work of Nielsen 1984), and they often inter- might limit species distribution pat- Niklas (1992) on plant growth pat- pret residuals from allometric re- terns, but also to protein chemists terns. Plant structure may be de- gressions in an ecological context and molecular biologists interested fined by branching angles and rota- (McNab 1966. Peters 1983). Such in elucidating protein structure- tion angles around a central axis gnalyses have irovided insights into function relationships. Comparative and a branching probability func- many features of organismal design biochemists thus provide data on tion used to generate different plant and have been instrumental in the so-called natural evolutionary ex- morphologies. Plant shapes gener- spread of new ways of thinking about periments at the -molecular level. ated using different models may then how organisms are built. For be compared to see the effect of example, the use of dimensional Systematics and the analysis of morphology on performance in tasks analysis and scaling to evaluate organismal design such as light interception and fluid mechanistic hypotheses about mus- I conduction within the plant. Niklas culoskeletal function has contrib- The study of organismal form and (1986) examined major evolution- uted a number of significant insights function may contribute important ary trends in plant morphology in into how birds fly (Pennycuick conceptual tools for the analysis of the light of results from the model- 1992). organismal design. Indeed, such ing and simulation studies and con- ~lio,the now wides~readaware- studies are the raw material for his- cluded that a number of possible ness of the importance of funda- torical analyses of evolutionary pat- geometric patterns may serve equally mental physical relationships such terns and processes. But most such well to meet demands of any one as surface-area-to-volume ratios for analyses, until recently, have lacked environmental task such as light in- understanding organismal design is an explicitly phylogenetic frame- terception. in large measure due to the contri- work. The increasing use of system- butions of comparative physiolo- atic concepts and methods in com- Comparative physiology. The study gists. Although scaling analyses have parative morphology and physiology of the function of structural fea- in the past usually been conducted reDresents more than a minor shift tures of organisms opens up a new outside a phylogenetic framework, in thinking- or research methodol- realm in the analysis of biodiversity. several workers (e.g., Heusner 1982) ogy and more than a new and short- Much attention has been focused on have recognized that the regression lived phase of what might be called the morphological parts of organ- slope determined from an analysis phylogenetic correctness. We believe isms and on the geographic distri- of separate monophyletic clades that the integration of phylogenetic bution of species, but in comparison within a larger group may produce a methods into disciplines tradition- I1 relatively little is known about the different value than a slope deter- ally involved with the mechanistic 1 diversity of function. Although" the mined from a single overall analysis analysis of organismal design has form-function relationship is one of ignoring phylogenetic structure. begun to revolutionize not only the the oldest areas of concern in biol- Comparative biochemistry has daj-to-day analyses conducted on ogy, relatively little attention has also contributed important insights data but also the key conceptual been paid by systematists to the func- into the potential physiological, bio- foundations and questions in these tion side of this dichotomy (Lauder chemical. and molecular determi- disciplines (Huey 1987, Lauder 1990). The primary and traditional nants of biogeographic patterns. 1991, Wake 1992). goals of comparative physiology and Examples include studies by Watt One important problem with biochemistry are to explore the di- (1983), Koehn (1987), and Powers many past analyses of organismal versity of physiological and bio- (1987). An interesting example design is that species have been I chemical processes and to take ad- comes from a study of enzymes treated as statistically inde~endent vantage of that diversity to elucidate (muscle lactate dehydrogenases) components of the analysis. Tradi- I I fundamental mechanistic principles from related species of barracuda tional studies of morphological and (Burggren 1991, Feder 1987, Hochach- (Sphyraena) found in different ther- physiological scaling are particularly ka and Somero 1984). Accordingly, mal environments. Differences in subiect to this assum~tion.but the disciplines analyzing organismal average body temperature of only independence of species is an issue function are traditionally reduction- several degrees apparently favor se- that underlies comparative analyses ist, mechanistic, and experimental: lection for different enzyme vari- of all kinds. Because organisms are They seek to understand how or- ants that accord with the distribu- related in a hierarchical fashion. ganisms work. tion patterns of the studied species closely related species are more likely

November 1995 Species A B c D E of habitats is thus temporally re- species (Figure 3). The magnitude lated to structural changes in these and significance of a potential evo- Character 1 15 9 18 35 28 taxa, and the phylogeny allows pre- lutionary correlation between the Character 2 1 2 3 4 5 diction of the sequence of environ- two characters can be tested in sev- mental invasion. On the other hand, eral ways. First, a conventional another possible phylogeny would Pearson product-moment correla- have closely related taxa correspond- tion, ignoring phylogeny and assum- ing to two major clusters of taxa. In ing independence, suggests a strong this case, without an outgroup taxa evolutionary relationship (correla- one cannot determine which of the tion 0.789) between the characters. two groups of environments repre- Second, Felsenstein's (1985) inde- sents the primitive condition. Fur- vendent-contrast method. which in- thermore, if the six species represent corporates phylogenetic information only two major clades, then the in- and corrects for nonindependence, Figure 3. Phylogeny of imaginary spe- dividual species clearly are not inde- suggests instead that the correlation cies A-E with the numerical values of pendent points for regression analy- is weak (0.016). Third. a minimal- two physiological characters (1 and 2) sis. Using a computer program such evolution corrklation,' which also shown under each species. Because of as that described in Martins and incorporates phylogenetic informa- the historical (genealogical) relation- Garland (1991), one can compare tion and which assumes gradual ships among the taxa, a correlation of branch tip values may give a misleading the correlation observed in the ab- character evolution a art ins and picture of the actual evolutionary trans- sence of a phylogenetic analysis with Garland 1991), yields an intermedi- formational relationship between the a phylogenetically standardized cor- ate value (0.498). two characters. The numbers to the left relation. Ignoring phylogeny, the Thus, the estimate of the correla- of the branches indicate branch lengths correlation is 0.935 (P<0.01).Using tion itself depends critically on the in units of expected variance of change phylogenetically standardized con- phylogeny of the species studied. (Felsenstein 1985), which would ap- trasts for the first phylogeny consid- And the apparent statistical signifi- propriately be estimated as divergence ered gives a correlation of 0.824 (PC cance of the correlation may depend times (e.g., from fossil information) if 0.05), while the second phylogeny on the assumed model of evolution characters evolved as by a gradual, generates a correlation of 0.792 (not (Harvey and Page1 1991, Page1 and clock-like model. significant). Harvey 1989; see Martins and Gar- Because species cannot be as- land 1991 for methodology) as well to share aspects of their phenotype sumed to represent independent as on the specific test used for esti- than are distantly related species. datapoints for statistical analyses, mation and hypothesis testing. Ex- An illustration of this point is this degrees-of-freedom problem amples of these methods applied to given in Figure 2, which shows six prevents traditional statistical meth- real physiological data are found in species (A-F) that show a correla- ods from being properly applied to Garland and colleagues (1991) and tion between some aspect of their comparative datasets (Clutton- Walton (1993). design (e.g., leg length) and a fea- Brock and Harvey 1977, Felsenstein ture of their environment (e.g., 1985, Harvey and Page1 1991, Mar- Reconstructing historical patterns height of surface vegetation). If one tins and Garland 1991). Computer- and sequences of trait differentia- ignores phylogenetic relationships, simulation studies show that ignor-- tion. A maior area of interest in one might conclude that the species ing phylogenetic relatedness, and comparative biology involves at- studied show a good correlation hence the possibility of resemblance tempts to reconstruct and analyze between leg length and vegetation due solely to relationship, leads to the historical patterns and sequences height. One might further be in- inflated Type I error rates (P-val- of the diversification of traits clined to make a causal or efficiency ues), reduced power to detect sig- (Brooks and McLennan 1991, argument that longer leg length nificant relationships, and inaccu- Donoghue 1989, Greene 1986, Huey could confer a selective advantage rate estimates of evolutionary and Bennett 1987, Monson 1989). in those habitats in which it is found. relationships (Grafen 1989, Mar- A phylogenetic basis is essential for However (and this point is a key tins and Garland 1991).Several phy- reconstructing the evolution of issue in phylogenetic approaches), logenetically based comparative physiological ind biochemical di- the interpretation given to the pat- methods correct for these problems versity in a lineage and understand- tern depends critically on the phylo- (reviews in Harvey and Page1 1991, ing how different types of structural genetic relationships of the taxa Martins and Garland 1991). and functional traits evolve. Many studied. Consider two alternative The importance of phy~bgenetic studies have focused on evolution- phylogenies for these taxa. The first considerations for statistical analy- ary patterns in one class of traits phylogeny implies that environments ses may be illustrated by a hypo- (e.g., skull structure or the amino with low vegetation heights were thetical case presented by Martins acids in a protein). One area of cur- the earliest ones inhabited by this and Garland ( 1991) in which a com- rent research is the historical rela- group, and speciation subsequently parative hioldgist seeks to determine tionship among different types of occurred into habitats with higher whether two characters are evolu- organismal traits (both structures vegetation. The progressive invasion tionarily correlated in a set of five and functions).

BioScience Vol. 45 No. 10 For example, in order to under- Table 1. One possible set of structural and physiological classes (or levels) of stand why several species of bird characters that could be studied using a phylogenetic framework. Evolutionary differ in the movements of their change might proceed at different rates in each level; change among levels is not wings during mating, one might ex- necessarily correlated. Thus, a fruitful avenue of research in morphology and amine several different classes of physiology is the examination of ontogenetic or phylogenetic transformations structural and physiological char- among levels. acters (Table 1)-changes in any one Class of character (organismal trait) Sample character of which could result in a different behavior. Changes in neuronal out- Behavioral Display behavior during mating put patterns in the nervous system, physiological properties of wing Functional/physiological Kinematics of bone movement, physiologi- (at the level of peripheral tissues) cal properties of muscles, biomechanical muscle fibers, or reorganization of tissue properties muscle attachments might all (ei- ther singly or together) result in a Structural Topographic arrangement of muscles and different movement pattern being (at the level of peripheral tissues) bones, tissue histology generated by a species. One can ex- Functional/physiological Neuronal spiking patterns, motor patterns, tend the simple hierarchical view of (at the level of the nervous system) membrane properties, modulation by neu- different classes of characters shown rotransmitters in Table 1 to a phylogenetic/histori- cal view of structure and function Structural Neuronal morphology, topology of neu- (at the level of the nervous system) ronal interconnection, wiring of sensory illustrated in Figure 4. By compar- and motor pathways ing both the structural design and physiological patterns of several taxa of known phylogenetic relationship, one can begin to examine parallel ate models of the first amphibians. group, revealed highly significant evolutionary changes in structure In effect, "...the frog is a red her- correlations (p<0.05) between acorn and function and hence address ques- ring" (Gans 1970). volume and geographic range, sug- tions relating to the evolutionary Similarly, although neurophysi- gesting a cause-effect relationship. timing of changes in design. For ologists have long assumed that heat- Jensen's reanalysis, in which the two example, do novelties occur first in sensitive pit organs of the pit vipers monophyletic groups of species were neuronal components, or are periph- (e.g., rattlesnakes) represent adap- treated separately, revealed much eral musculoskeletal elements altered tations to feeding (i.e., ability to lower correlations between acorn first? Are some classes of characters detect warm-blooded prey at night), volume and geographic range (Table 1)more phylogenetically con- a recent phylogenetic analysis (p>O.OS) and provided evidence that servative than others? Or do all com- (Greene 1992) suggests that the ori- the two groups have evolved differ- ponents of a complex system tend to gin of facial pits was more likely ently with respect to morphological change together? A phylogenetic correlated with the evolution of sta- features that might influence breadth approach to reconstructing ances- tionary defensive behavior and only of geographic range. tral characteristics of form and func- secondarily was used for feeding. Friedman (1990, 1992) was able tion can help answer these questions. Also, although the accumulation of to trace the evolutionary origins of glycinebetaine in response to salin- polyploid endosperm tissue in the Major patterns of character evolu- ity and water stress was once thought seeds of angiosperms. Endosperm tion. Comparative physiologists to have evolved widely among plants functions as the principal nutritive sometimes try to reconstruct the native to saline habitats, broader tissue of angiosperm seeds, support- origin of major physiological inno- comvarative studies have shown that ing the growth and maintenance of vations (e.g., origin of endothermy, not all halophytic species accumu- the embryo. Polyploid endosperm is of anaerobic glycolysis, and of C4 late glycinebetaine. The occurrence a defining trait of the angiosperms photosynthesis) or the changes in of the compound appears to have a and might be one of the advantages physiological or biochemical traits strong phylogenetic component allowing angiosperms to dominate associated with major evolutionary (Wyn Jones and Storey 1981). most terrestrial habitats. Friedman's transitions, such as from water to Another example is provided by studies were conducted in the land (Bennett and Ruben 1986, Jensen (1992), who demonstrated Gnetales, a group of nonflowering Burggren and Bemis 1990, Carrier that failure to consider phylogenetic seed plants that have been proposed 1987). A phylogenetic and histori- relationshivs led to inflated estimates through several phylogenies to rep- cal perspective is crucial here. For of the correlation between morvho- resent the closest extant ancestor to example, although some compara- logical and biogeographic param- the angiosperms. Using the Gnetales tive physiologists had used the frog eters. Other workers had reported as an outgroup, Friedman demon- as a physiological model of amphib- that oaks with large acorns have strated that endosperm probably ians that first invaded the land, Gans broader geographic ranges than evolved its nutritive role from su- (1970) showed in a classic study those with small acorns. Their analy- pernumerary embryos. In the Gnet- that frogs are highly specialized ani- ses, conducted by treating all east- ales, such secondary embryos de- mals and thus are hardly appropri- ern North American oaks as a single generate as they nourish the primary

November 1995 70 1 of species has been guided by spe- cific features (physiological, mor- phological, and environmental) of the species or by its tractability and accessibility for physiological stud- ies (the August Krogh Principle- the idea that specific physiological problems can be matched to a spe- cies in which that problem can be most easily studied; see Krebs 1975). For example, studies of urine con- centration bv the mammalian kid- ney might focus on a species of desert rodent, under the assumption that such a species is likely to be adapted GFE DCB to that environment and will thus show with special clarity the rela- b \ / \ 0 \ \ /* Y y pattern b \ \ tionship between structure, function, Motor pattern b for muscle 3 and environment. However widespread use of the Motor pattern a Krogh principle has been in guiding for muscle 4 physiological research, it is funda- mentally a nonphylogenetic ap- Muscle 4 0Physiological character proach to organismal design. The Morphological character choice of a single species in a single Motor pattern a for muscle 3 environment does not allow any Muscle 3 judgment to be made about the his- torical origin of the traits under consideration and is in realitv an Figure 4. (a) Schematic diagram of a musculoskeletal system in taxon A composed eauilibrium (nonhistorical) mAhod of bones (circles), muscles (black lines), and ligaments (shaded lines) to illustrate fir analyzing organisms and their the role that a phylogenetic analysis might play in understanding evolutionary patterns of structural and physiological traits. (b) Phylogenetic relationships of current environments. seven related taxa showing the origin of morphological and physiological traits of A maxim of comparative phylo- this musculoskeletal system. An analysis of structure and function could be done genetically based research is that at for each terminal taxon (A-G) and the historical pattern of character change least a three-species comparison is mapped onto a phylogeny. The phylogeny shows the sequence with which physi- best for clarifying structure-func- ological and morphological characters have arisen. Note that the structural and tion-environment relationships functional traits have different evolutionary histories. For example, the presence (Brooks and McLennan 1991, Gar- of muscle 3 is primitive for this clade along with a particular pattern of muscle land and Adolph 1994, Huey 1987). activation (motor pattern a). The clade A+B is characterized by a novelty in The use of at least three species physiology (motor pattern b) for muscle 3 but no change in muscle structure. The allows historical sequences of char- presence of muscle 4 characterizes the clade A-E, but two different physiological patterns of activation are present within this clade (motor patterns a and b). acter change to be reconstructed and is in itself test of the generality of findings on one taxon alone. Thus, embryo in an apparent act of altru- phyletic clades of lizards promoted if a second species of rodent closely ism-a phenomenon that brings to a false dichotomy between taxa pur- related to the s~eciesstudied is ex- the surface several hypotheses con- ported to use vision for feeding ver- amined, one might find that its kid- cerning the role of kin selection in the sus taxa that supposedly rely primar- ney is modified for urine concentra- evolution of developmental traits. ily on chemoreception to find food. tion even though this species does Finally, Schwenk (1994) has re- Had use of these two sensory mo- not live in the desert. (An exam~le cently shown how the use of a dalities been mapped onto mono- of supposed physiolog'ical adasa- nonphylogenetically based classifi- phyletic taxa, 40 years of false gen- tions in one species that are later cation of squamate reptiles, has hin- eralizations about the evolution of found to be widespread is discussed dered understanding of the evolu- sensory systems in lizards could have in Dawson et al. 1977). Further- tion of sensory systems. Schwenk been avoided. Squamate reptiles pro- more, if a third species of rodent, demonstrated that use of a classifi- vide an excellent case study of the one that inhabits an arboreal habi- cation that did not reflect genea- predictive and analytic consequences tat, is studied, one might discover logical relationships led to incorrect of phylogenetic hypotheses of or- that its kidney is unable to concen- generalizations about squamate vi- ganismal design. trate urine. With these data and a sion and chemoreception and their phylogeny, one can reconstruct a environmental correlates. For ex- Choice of species for comparison. In historical sequence of physiological ample, recognition of nonmono- many comparative studies the choice and environmental change that is

BioScience Vol. 45 No. 10 not possible with a one- or two- outlines as well as discrete land- being based increasingly on phylo- taxon study. marks. and traditional multivariate genetic information. The diversity In many other comparative stud- statistical methods have been ex- of organismal function (and its rela- ies the explicit intent is to address tended to deal more effectively with tionship to structure) is a vast area evolutionary issues such as physi- such biological problems as size and of unexplored biology. With the in- ological adaptation to an extreme shape. In particular, the traditional fusion of historical methods and environment. For such studies, the bivariate approach to allometry has concepts from systematics, the fu- choice of species for analysis can be been extended into the multivariate ture promises exciting advances as guided by systematic information domain, with the benefit that many physiologists and morphologists (Burggren 1991, Garland et al. 1991, studies of form today, whatever their bring new tools to bear on the analy- Huey 1987). A comparison of close purpose, are multivariate in nature. sis of structural and functional di- relatives reduces, for instance, the It would be difficult to overesti- versity. probability that observed differences mate the impact that the develop- between taxa are an artifact of long- ment of both data acquisition and Acknowledgments separate phylogenetic history rather statistical tools for the study of than of adaptation to a particular morphology has had on research on We gratefully acknowledge the con- environmental feature. organismal form; in truth, only in tributions of the manv~, members of the last decade has the promise of the original two committees (espe- Detecting evolutionary anachro- Thompson's approach to structure cially T. Garland) for Systematics nisms. A systematic and historical been realized. Earlv attemvts to Agenda 2000 (Quantitative and Evo- study the deformatioh of oneshape lutionary ~or~holo~~,Comparative Lversvective 1 also is useful in detect- ing evolutionary anachronisms (i.e., into another, for example, were Physiology and Biochemistry) who traits that evolved in response to largely qualitative and inaccurate assisted in the preparation of the conditions no longer existing). For and not readily amenable to statisti- original reports that formed the ba- example, the giant fruits of some cal study. sis of this article. This article is a Neotropical trees might have evolved Although vigorous debate con- product of the initiative Systemat- as adaptations for dispersal by Pleis- tinues on the assumptions of the ics Agenda 2000, which was gener- tocene megafauna (Janzen and Mar- analvtical methods used to studv ously supported by the National tin 1982), and the presence of flight form. there is no doubt that the Science Foundation through grant motor neurons (currently nonfunc- recent rapprochement of morpho- DEB-9396035. Preparation of the tional) in flightless crickets probably metrics and phylogenetics has manuscript was supported by: NSF reflects the evolution of flightless grass- opened up a broad array of new IBN-9119502 to George Lauder, hoppers from ancestors that could fly auestions. It has also raised the stan- NSF DEB-9301151 to Ray Huey, (Dumont and Robertson 1986). dards to which answers will be held and NSF BSR-8604960 and BSR- Thus, an appreciation of phyla- (Zelditch et al. 1992, 1995). 8911433 to Russell Monson. genetic history sometimes clarifies the function-or the lack thereof- Conclusions References cited of an otherwise puzzling physiologi- cal trait (Huey 1987). A phyloge- Physiological and functional traits Allen G. 1975. Life science in the twentieth netic analysis may reveal that the of plants and animals are not well- century. New York: John Wiley. preserved in fossils. Unfortunately, Bennett AF, Ruben J. 1986. The metabolic traits thought to represent novel evo- and thermoregulatory status of ther- lutionary responses to current envi- much of the effort toward biodiver- apsids. Pages 207-218 in Hotton N, ronments may have been retained sity recognition and preservation has MacLean PD, Roth JJ, Roth EC, ed. The from an ancestral condition that had focused on biogeographic and taxo- ecology and biology of mammal-like rep- little in common with current envi- nomic concerns rather than on the tiles. Washington (DC): Smithsonian In- stitution Press. ronmental conditions. analysis of functional diversity. For Bookstein F, Chernoff B, Elder R, Humphries most clades, a comparative analysis J, Smith G, Strauss R. 1985. Morpho- Morphometrics. Systematics has had of physiological traits is in its in- rnetrics in evolutionary biology. Phila- considerable influence recently in fancy. In part, perhaps because of delphia (PA): Academy of Natural Sci- the past tendency of physiologists to ences. the quantitative description of mor- Briggs DEG, Fortey RA, Wills MA. 1992. phology. Two distinct avenues of focus on a few organisms consid- Morphological disparity in the Cambrian. progress are evident. First, many ered to be good models, a broadly Science 256: 1670-1673. biometricians have developed new comparative historical approach to Brooks DR, McLennan DA. 1991. Phylog- methods for archiving the shape of the study of organismal function is eny, ecology and behavior. Chicago (IL): University of Chicago Press. biological structures (and whole or- just beginning. This beginning oc- Burggren WW. 1991. Does comparative res- ganisms). These methods are rigor- curs at an opportune time, however, piratory physiology have a role in evolu- ous, quantitative, and lend them- because methodologies for quanti- tionary biology (and vice versa)? Pages selves to multivariate statistical tative comparative methods are blos- 1-13 in Woakes AJ, Grieshaber MK, analysis (reviewed in Rohlf and soming, organismal biologists are Bridges CR, ed. Physiological strategies for gas exchange and metabolism. Cam- Bookstein 1990). becoming interested in expanding bridge (UK): Cambridge University Press. Some of these methods deal with beyond one-taxon analyses, and the Burggren WW, Bemis WE. 1990. Studying organismal shape data presented as choice of species to be analyzed is physiological evolution: paradigms and

November 1995 pitfalls. Pages 191-228 in Nitecki MH, and functional enzymic evolution in four Niklas KJ. 1986. Computer simulations of ed. Evolutionary innovations. Chicago species of eastern Pacific barracudas from branching patterns and their implication on (IL): University of Chicago Press. different thermal environments. Evolution the evolution of plants. Lectures on Math- Calder WA. 1984. Size, function, and life his- 36: 97-106. ematics in the Life Sciences 18: 1-50. tory. Cambridge (MA): Harvard University Greene HW. 1986. Diet and arboreality in the . 1992. Plant biomechanics. Chicago Press. Emerald Monitor, Varanus prasinus, with (IL): University of Chicago Press. Carrier DR. 1987. The evolution of locomotor comments on the study of adaptation. Pagel MD, Harvey PH. 1989. Comparative stamina in tetrapods: circumventing a me- Fieldiana Zoology n.s. 31: 1-12. methods for examining adaptation de- chanical constraint. Paleobiology 13: . 1992. The ecological and behavioral pend on evolutionary models. Folia 326-341. context for pitviper evolution. Pages Primatologica 53: 203-220. Clutton-Brock TH, Harvey PH. 1977. Primate 107-117 in Campbell JA, Brodie ED, eds. Pennycuick CJ. 1992. Newton rules biology: A ecology and social organization. Journal of Biology of the pitvipers. Tyler (TX):Selva physical approach to biological problems. Zoology (London) 183: 1-39. Publications. Oxford (UK): Oxford University Press. Coleman W. 1977. Biology in the nineteenth Hall BK. 1994. Homology: the hierarchical Peters RH. 1983. The ecological implications century: problems of form, function, and basis of comparative biology. San Diego of body size. Cambridge (UK): Cambridge transformation. Cambridge (UK): Cam- (CA): Academic Press. University Press. bridge University Press. Harvey PH, Pagel MD. 1991. The comparative Powers DA. 1987. A multidisciplinary approach Conway Morris S. 1992. Burgess Shale-type method in evolutionary biology. Oxford to the study of genetic variation within faunas in the context of the "Cambrian (UK): Oxford University Press. species. Pages 102-134 in Feder ME, Bennett explosion": a review. Journal of the Geo- Heusner AA. 1982. Energy metabolism and AF, Burggren WW, Huey RB, eds. New logical Society (London) 149: 631-636. body size. I. Is the 0.75 mass exponent of directions in ecological physiology. Cam- Dawson WR, Bartholomew GA, Bennett AF. Kleiber's equation a statistical artifact? bridge (UK): Cambridge University Press. 1977. A reappraisal of the aquatic special- Respiration Physiology 48: 1-12. Raup DM. 1966. Geometric analysis of shell izations of the Galapagos marine iguana Hochachka PW, Somero GN. 1984. Biochemi- coiling: general problems. Journal of Pale- (Amblyrhynchus cristatus). Evolution 31: cal adaptation. Princeton (NJ): Princeton ontology 40: 1178-1 190. 891-897. University Press. Raup DM, Stanley SM. 1971. Principles of DonoghueMJ. 1989. Phylogenies and theanaly- Huey R. 1987. Phylogeny, history, and the paleontology. San Francisco (CA): W. H. sis of evolutionary sequences, with examples comparative method. Pages 76-101 in Feder Freeman and Co. from seed plants. Evolution 43: 1137-1 156. M, Bennett AF, Burggren WW, Huey RB, Rohlf FJ, Bookstein FL. 1990. Proceedings of Dumont J, Robertson RM. 1986. Neuronal eds. New directions in ecological physiol- the Michigan Morphometrics Workshop. circuits: an evolutionary perspective. Sci- ogy. Cambridge (UK): Cambridge Univer- Ann Arbor (MI): University of Michigan ence 233: 849-853. sity Press. Museum of Zoology. Emerson S. 1988. Testing for historical pat- Huey RB, Bennett AF. 1987. Phylogenetic stud- Russell ES. [I9161 1982. Form and function: a terns of change: a case study with frog ies of coadaptation: preferred temperature contribution to the history of animal mor- pectoral girdles. Paleobiology 14: 174-186. versus optimal performance temperature of phology. Chicago (IL): University of Chi- Feder ME. 1987. The analysis of physiological lizards. Evolution 41: 1098-1115. cago Press. diversity: the prospects for pattern docu- Janzen DH, Martin PS. 1982. Neotropical Schmidt-Nielsen K. 1984. Why is animal size mentation and the general questions in eco- anachronisms; the fruits the gomphotheres so important? Cambridge (UK): Cambridge logical physiology. Pages 347-351 in Feder ate. Science 214: 19-27. University Press. ME, Bennett AF, Burggren WW, Huey RB, Jensen RJ. 1992. Acorn size redux. Journal of Schwenk K. 1994. Comparative biology and ed. New directions in ecological physiol- Biogeography 19: 573-579. the importance of cladistic classification: a ogy. Cambridge (UK): Cambridge Univer- Koehn RK. 1987. The importance of genetics case study from the sensory biology of squa- sity Press. to physiological ecology. Pages 170-185 in mate reptiles. Biological Journal of the Lin- Felsenstein J. 1985. Phylogenies and the com- Feder ME, Bennett AF, Burggren WW, Huey nean Society 52: 69-82. parative method. American Naturalist 125: RB, eds. New directions in ecological physi- Thompson DW. 1917. On growth and form. 1-15. ology. Cambridge (UK): Cambridge Uni- Cambridge (UK): Cambridge University Friedman WE. 1990. Double fertilization in versity Press. Press. Ephedra, a non-flowering seed plant: its Krebs HA. 1975. The August Krogh Principle: Wake MH. 1992. Morphology, the study of bearing on the origin of angiosperms. Sci- for many problems there is an animal on form and function, in modern evolutionary ence 247: 951-954. which it can be most conveniently studied. biology. Oxford Surveys in Evolutionary . 1992. Evidence of a pre-angiosperm Journal of Experimental Zoology 194: Biology 8: 289-346. origin of endosperm: implications for the 221-226. Walton BM. 1993. Physiology and phylogeny: evolution of flowering plants. Science 255: Lauder GV. 1981. Form and function: struc- the evolution of locomotor energetics in 336-339. tural analysis in evolutionary morphology. hylid frogs. American Naturalist 141: Gans C. 1970. Respiration in early tetrapods- Paleobiology 7: 430-442. 26-50. the frog is a red herring. Evolution 24: . 1990. Functional morphology and sys- Watt WB. 1983. Adaptation at specific loci. 11. 723-734. tematics: studying functional patterns in an Demographic and biochemical elements in Garland T, Adolph SC. 1994. Why not to do 2- historical context. Annual Review of Ecol- the maintenance of the Colias PGI poly- species comparisons: limitations on infer- ogy and Systematics 21: 317-340. morphism. Genetics 103: 691- 724. ring adaptation. Physiological Zoology 67: . 1991. Biomechanics and evolution: Wills MA, Briggs DEG, Fortey RA. 1994. Dis- 797-828. integrating physical and historical biology parity as an evolutionary index: a compari- Garland T, Carter PA. 1994. Evolutionary in the study of complex systems. Pages 1-19 son of Cambrian and recent arthropods. physiology. Annual Review of Physiology in Rayner JMV, Wootton RJ, ed. Biome- Paleobiology 20: 93-130. 56: 579-621. chanics in evolution. Cambridge (UK):Cam- Wyn Jones RG, Storey R. 1981. Betaines. Garland TJ, Huey RB, Bennett AF. 1991. Phy- bridge University Press. Pages 172-205 in Paleg LG, Aspinall D, logeny and coadaptation of thermal physi- Martins E, Garland T. 1991. Phylogenetic ed. Physiology and biochemistry of ology in lizards: a reanalysis. Evolution 45: analysis of the correlated evolution of con- drought resistance in plants. New York: 1969-1975. tinuous characters: a simulation study. Evo- Academic Press. Gould SJ. 1989. Wonderful life: the Burgess lution 45: 534-557. Zelditch ML, Bookstein FL, Lundrigan BL. Shale and the nature of history. New York: McNab BK. 1966. The influence of food habits 1992. Ontogeny of integrated skull growth W. W. Norton. on the energetics of eutherian mammals. in the cotton rat Sigmodon fulviventer. Grafen A. 1989. Phylogenetic regression. Philo- Ecological Monographs 56: 1-1 9. Evolution 46: 1164-1 180. sophical Transactions of the Royal Society Monson RK. 1989. On the evolutionary path- Zelditch ML, Fink WL, Swiderski DL. 1995. of London B Biological Sciences 326: ways resulting in C4 photosynthesis and Morphometrics, homology, and phylo- 119-157. Crassulacean acid metabolism. Advances genetics: quantified characters as synapo- Graves JE, Somero GN. 1982. Electrophoretic in Ecological Restoration 19: 58-110. morphies. Systematic Biology 44: 179-189.

BioScience Vol. 45 No. 10