|
Main | Message Board | Evolution Basics | Within Evolution | About Evolution | Information | Sign Guestbook | View Guestbook | Credits |
The study of dinosaurs is a branch of the study of paleobiology. Paleobiologists have fewer tools at their disposal for use in defining species and in determining phylogenies than biologists who study present-day species have. The fossil record allows only inferences based solely on morphologic criteria to be made; therefore, paleospecies, which are defined strictly on morphologic grounds, do not necessarily correspond to actual biological species, or interbreeding populations, whose differences may not be morphologic. Once species are defined, two methods for their classification, cladistics and morphologic distance measures, are available to paleontologists. Of these two, only cladistics can give unambiguous, objective classifications, and thus is the only method widely accepted in paleontology. Even using the methods of cladistics, which involve outgroup comparison to demonstrate phylogenies, paleontologists do not agree on how to classify dinosaurs, as illustrated by a number of examples in which cladistic method is applied but in which conclusions are not compatible. These disagreements come about due to the imperfection of the fossil record.
Dinosaurs have an ability to fascinate in a way that few other life forms, present or past, can exceed. Yet there are unique problems in their study, due to the nature of the material available to work with. Nothing more than bones and footprints remain of these animals, leaving paleontologists to answer questions about their classification and evolution without the benefit of observation of behavior, soft-tissue morphology, et cetera. This paper will address some of the problems encountered by paleobiologists interested in the phylogenies and taxonomies of dinosaur groups, and illustrate some solutions using examples of conclusions drawn by these researchers.
Chaline (1990) explains that as defined by Mayr, a biological species consists of a population of interbreeding individuals. The criteria which differentiate a species may be morphologic, biochemical, chromosomal, ecologic, behavioral, et cetera. Except for morphologic criteria, none of these characters fossilize, so paleontologists cannot define species in the biological sense with the same accuracy as biologists studying modern-day species. Paleospecies can only be defined based on certain morphologic criteria (bones and teeth, which fossilize), and a paleospecies may or may not correspond to an actual biological species. A single paleospecies, therefore, may refer to a single biological species, or to a group of several biological species, which in life did not interbreed but were identical morphologically.
Distinguishing among species is one step in the process of classifying dinosaurs, but classifications extend above the species level. Tracing phylogenies is the object of dinosaur taxonomists.
Chaline states that the classification scheme of Linne assumed a hierarchy based on phenetic characters. Species were considered immutable but were arranged into a hierarchical, branching system. Linne's classification system has been adapted by evolutionary biologists, who use the hierarchy to represent not morphologic similarity but rather phylogenetic relationships. Therefore, a modern (evolution-based) classification system traces lineages; and in order to classify an organism, its phylogeny must be determined. Again, paleontologists have only one tool -- morphology -- to work with when determining phylogenies.
Two types of morphologic characters can be defined for use in determining phylogenies: apomorphies (derived, or evolved, characters) and plesiomorphies (primitive, or ancient, characters). Chaline (1990) explains that there are three methods in paleontology to infer whether a character is an apomorphy or a plesiomorphy. They can be inferred from the ontogeny of modern day species (or, if fossil embryos are available, from the ontogeny of fossilized species), using a version of the theory of recapitulation: the earlier in development a character appears, the more primitive it is assumed to be. This method is not preferred, however, because it makes assumptions about the accuracy of the theory of recapitulation.
A second method is termed the stratophenetic method; relatively older fossils are considered to have more plesiomorphies, while relatively younger fossils are considered to have more apomorphies. These measures of what is primitive and what is derived are then used to infer phylogenies from the fossil record. This method is inaccurate because it ignores the fact that evolution can move at different rates; for example, a primitive character may be retained in a later form because it evolved slowly. This character, found in the later form, will then be termed an apomorphy due to the age of the specimen in which it was found, when it is in fact a plesiomorphy which has been preserved.
The third, and most reliable, method of determining whether a character is an apomorphy or a plesiomorphy is by outgroup comparison. A group must be compared with a sister group (a group sharing a common ancestor); if the character being studied is absent in the sister group, then the character is an apomorphy. These determinations are relative, and dependent on the level of organization being considered; apomorphies in relatively more ancient groups become plesiomorphous with time.
Sereno (1990) explains that specifying whether a character is apomorphous or plesiomorphous is important in taxonomy because the patterns of the distribution of apomorphies among different groups can be used to infer phylogenies. Chaline (1990) continues that if several groups share a number of apomorphies which are not present in other groups, then these characters are considered synapomorphies. Plesiomorphies which are shared by several groups are considered symplesiomorphies. Sereno points out that, just as with apomorphies and plesiomorphies, whether a character is synapomorphic or symplesiomorphic depends on the taxonomic level; a synapomorphy in a high-ranking (generalized) taxonomic level becomes a symplesiomorphy in lower ranking (more specialized) taxonomic levels nested within the higher-ranking level.
Classification schemes based solely on phylogeny rely on measures of apomorphic characters because phylogenies can only be inferred from synapomorphies. Groups with a number of synapomorphies constitute, along with their common ancestor, a monophyletic group (including all descendants of the ancestor).
Sereno (1990) explains that another factor, morphologic distance, can be used to classify organisms. Distance in terms of morphology refers to similarity; the more morphologically similar two groups are, the lower the taxonomic rank in the hierarchy of their common group. Measures of morphologic distance are distinguished from measures of synapomorphies because morphologic distance measures do not consider the presence of morphologic characters in outgroups (i.e. they do not distinguish between synapomorphies and symplesiomorphies); rather, they count synapomorphies as well as symplesiomorphies between groups and assign distance measures based on their number. By this method, paraphyletic groups, which exclude descendant groups which are sufficiently distant morphologically from the ancestor, are admitted into a classificatory scheme.
Since, in a scheme admitting morphologic distance measures, both synapomorphies and symplesiomorphies are used, a certain degree of arbitrariness must be used in deciding which characters will be utilized, leading to the creation of "unreal" (non-objective) taxa, which may be defined differently by independently working researchers.
Morphologic distance measures reach another pitfall in paleobiology. Sereno (1990) explains that due to the character of the fossil record, in which there are a number of gaps due to unexposed or unfossiliferous strata, any morphological gap between types of organisms may be artificial, i.e. due to poor sampling. These gaps could represent as-yet unrecorded rapid evolution rather than showing actual distance between groups. A third problem in measures of morphologic distance involves the instability of taxa that it introduces; for example, some groups (such as the iguanodontoids in Sereno's example below) are destroyed because its member groups have fewer characters in common with each other than with their common ancestor. Due to these problems with morphologic distance measures, Sereno dismisses them on the grounds that "a reasonably explicit methodology for the measurement of morphologic distance and its incorporation into a classification is not available. . . . The time has passed when the construction and maintenance of paraphyletic taxa can proceed unchallenged."
The rejection of morphologic distance measures leaves phylogenetic measures, using cladistics, as the only way to systematically classify organisms. Using comparisons of synapomorphies, branching lineages called clades can be created for species. In a strict cladistic sense, only monophyletic groups can be considered; paraphyletic groups are unnatural because they obscure phylogenetic relationships (and are based on morphologic distance measures), while monophyletic groups reflect lineages accurately. In taxonomy, symplesiomorphies are ignored; Sereno states that "modern cladistic method employs synapomorphy as the sole systematic criterion for the delineation of monophyletic taxa."
These theories regarding cladistic method are routinely employed by scientists researching the paleontology of dinosaurs. A great deal of work, particularly within the last few decades, has been published regarding the phylogenies of the dinosaurs; some of those findings will be reviewed later. The problem of species definition must be addressed first; several examples of studies regarding the definition of dinosaur species appear below. For example, Coombs (1990) reports on the identification of several species of ankylosaurs based on morphologic variation of their teeth. In studies of fossilized teeth, five methods of identifying morphological variation are available: positional, ontogenetic, intraspecific, taxonomic, and chimeric. Positional variation refers to the location of teeth in the jaw, and to their absolute sizes and sizes relative to one another. Ontogenetic variation refers to variation among specimens at different stages of development (for example, juveniles are compared with mature adults). This method requires a range of specimen ages, and thus cannot always be analyzed. Intraspecific variation is concerned with such factors as sexual dimorphism and clinal variation along a species' geographical range. Taxonomic, or interspecific, variation refers to differences among individuals of different species; it should apply only to the species level (rather than being used to distinguish among larger groups). The amount of variation between species is variable; there may be a great deal of overlap in morphology, which leads to ambiguities in species determinations, or morphologic separation may be nearly complete, making species determinations definite. Chimeric variation is concerned with ontologically deformed or damaged teeth, and is generally ignored in studies of variation. Some of the ankylosaurs in Coombs' study cannot be identified based on tooth morphology. This difficulty is due to similarity among teeth of species which are members of highly similar groups. Family Nodosauridae is an example; Coombs explains that in this group, "teeth of Sauropelta edwardsi . . . from the Lower Cretaceous Cloverly Formation differ only in minute details from Edmontonia longiceps and E. rugosidens from the Upper Cretaceous Judith River formation." Coombs goes on to say that "Edmontonia longiceps and E. rugosidens cannot be consistently distinguished by dental characters." The similarities in the teeth of these two species represent symplesiomorphies.
Much of the method of identifying dinosaur species based on their teeth can be applied to the study of other morphologic characters as well. Raath (1990), in the study of a group of small theropod fossils, concludes that the variations among these individuals are intraspecific, and that the population comprises a single species; more precisely, he concludes that the variation is due to sexual dimorphism.
Raath explains that the bones of small theropods are relatively fragile compared to those of other dinosaurs because theropod bones are small and have thin walls. Because of this fragility, data on these animals is hard to accumulate. Raath's study focused on a small theropod which lived in the early Jurassic period, Syntarsus rhodesiensis, recovered from a formation in Zimbabwe, Africa. Raath records that the fossil specimens were found grouped together, as if they "were members of a single cohesive, interacting group. This alone implies that they were members of a single species." He catalogues the morphologic variation among the individuals and explains it as intraspecific rather than interspecific. Though the population appears to belong to a single species, two distinct morphs, gracile and robust, are defined by the study. These two morphs are not graded. In addition, robust morphs are not found in specimens below a certain size. Raath postulates that these phenomena show that the differences were due to sexual dimorphism. This theory explains that lack of grading between the two morphs, and the lack of small robust morphs appears to be due to the fact that smaller individuals are juveniles and are not yet sexually mature. Sexual dimorphism appears to be present in other theropod species, including Tyrannosaurus Rex (Carpenter 1990) and Coelophysis bauri (Colbert 1990).
Above the species level, dinosaur taxonomists are concerned with classifying dinosaurs according to their phylogenies. Sereno (1990) uses the classification of ornithopod dinosaurs to illustrate cladistics versus morphologic distance measures. Due to morphologic differences, four groups of these dinosaurs (fabrosauroids, hypsylophodontoids, iguanodontoids, and hadrosauroids) were at one time considered four discrete groups. This classification has since been proven, based on analysis of synapomorphic characters, to represent these groups in a paraphyletic scheme. Sereno lists a monophyletic scheme in which each group is a descendant of the one preceding it. By inference, then, all hadrosauroids are iguanodontoids, all iguanodontoids are hypsylophodontoids, and all hypsylophodontoids are fabrosauroids. A recently discovered genus, Ouranosaurus, is placed between iguanodontoids and hadrosauroids in Sereno's scheme.
Horner (1990) elaborates on this scheme for the classification of ornithopods. He divides the group defined by Sereno as "hadrosauroids" into two branches, the hadrosaurs and the lambeosaurs. He places the genus Iguanodon into the hadrosaur group, along with the Hadrosauridae family, and places the genus Ouranosaurus into the lambeosaur group, along with the Lambeosauridae family. He justifies this scheme with a number of synapomorphies for both the hadrosaurs and the lambeosaurs. He shows that there is no evidence that either Iguanodon or Ouranosaurus is ancestral to either group, contradicting Sereno, who claimed that in a classification such as Horner's, the grouping of these two genera was paraphyletic. Horner suggests that the Hadrosauridae and Lambeosauridae share similarities due to convergence rather than to common ancestry and should be classified in separate groups, whereas Sereno places them into a single group, the hadrosaurs, with Ouranosaurus ancestral to the hadrosaurs and Iguanodon ancestral to Ouranosaurus.
Thus, even when paleontological classifications such as those of dinosaurs are based only on cladistics and not on morphologic distance measures, there are disagreements due to ambiguities in the fossil record.
There are other, less controversial examples of paraphyletic classification in dinosaur taxonomy, the most obvious involving the separation of class Aves from class Reptilia. Currently, it is generally accepted that birds descended from theropod dinosaurs (though some paleontologists claim that they merely share a common ancestor with dinosaurs, making them sister groups). Consequently, Aves is more and more frequently given as a subgroup of Dinosauria (or else more specifically of Saurischia or of Theropoda). Chaline (1990) classifies Aves as equivalent hierarchically with the two main dinosaur lineages, Saurischia and Ornithischia; he demonstrates these latter two groups as being monophyletic by listing a number of synapomorphies (including, in the Saurischia, a triradiate pelvis and extension of the teeth along the jaws; and in the Ornithischia, a tetraradiate pelvis, a predentary bone, and a toothless premaxilla). In his classification scheme, Class Reptilia is done away with entirely, to be replaced by the group Amniota, including birds, mammals, and animals traditionally classified as reptiles. According to Benton (1990), the consideration of Dinosauria as a monophyletic group is a fairly recent classification; originally, the lines Saurischia and Ornithischia were considered to have evolved from separate ancestors and were ranked as sister groups along with Crocodilia and Pterosauria, but recent studies have shown a number of synapomorphies which group them together.
For example, Bakker (1986) cites the morphology of the manus, which is modified into a "twist-thumb," as a synapomorphy between theropods and anchisaurs (which are the ancestors of sauropodomorphs). The theropods and the sauropodomorphs together constitute the Saurischia. He then cites a prong on the ilium of Anchisaurus as a possible synapomorphy between Ornithischians and Saurishichia. By this and seven other synapomorphies compiled by Bakker along with Peter Galton, Dinosauria is demonstrated to be a natural, monophyletic group.
The Ornithischia line was considered separate from the Saurischia line by Bakker based on the lack of the "twist-thumb" character in the Ornithischia; subsequently, however, the genus Heterodontosaurus was discovered. This genus possesses a number of characters, including a predentary bone, which define it as an ornithischian, yet it also has a "twist- thumb." Bakker concludes that the "twist-thumb" character was lost in later ornithischians, and based on this synapomorphy proposes a classificatory scheme which is different from that given by Benton (1990), in which Sauropodomorpha is grouped with Ornithischia as a monophyletic group descended from the anchisaurs. These two lines comprise a group Bakker names Phytodinosauria ("plant-dinosaurs," referring to their herbivorous habits), which is a sister group to the Theropoda. Benton acknowledges Bakker's studies by listing the eight possible synapomorphies between Sauropodomorpha and Ornithischia in his review.
Despite the well-defined methods of cladistics, both species definitions and phylogenies in paleontology remain ambiguous. The fossil record provides data which are imperfect at best and misleading at worst, and though paleobiologists use the strict logic of cladistics to reach their conclusions, ambiguities will always remain present in classifications of extinct organisms such as dinosaurs.
© 1997, 1998 by Robyn Conder. All rights reserved.
E-mail comments to Robyn Conder at rconder@ou.edu