Models And The Locomotor Strategies Of Extinct Taxa

Central to paleobiological research that aims to explain both aspects of behavioral ecology of extinct forms and patterns of historical factors (these efforts are usually limited to dietary and locomotor strategies) is the analysis of skeletal remains. Living species models with their rigorously analyzed form-function attributes and their ecological causality lay the foundation for not only character analysis in systematics (as opposed to taxic analysis), but also for analyzing, through the use of convergence and matching, the form-function of the fossils as well (Szalay, 2000). Biomechanical generalities, such as occlusal mechanics of teeth or the loading of joints are paramount, but because, due to the uniqueness of lineages, there are no living species that match exactly the habitus of fossil entities.

It is not unusual that a living analog is used to find similarity (a concept fundamentally context- and paradigm-driven) for some sort of fossil morphology without functional, and therefore, causal reasons. The lack of a causal analysis (i.e., ecological, real-time) in the process of modeling does not allow one to conclude that selected matching morphologies indicate adaptive (ecological) similarities between the model and the fossil. Nevertheless, this approach can supply some meaning for paleobiological assessment if a whole skeleton is available for the fossil. Without complete specimens, however, the modularity-based and well-corroborated patterns of mosaic evolution render such assessments problematic for functional units of the skeleton. Such a general similarity evaluation lacks, as its basis, the necessary character analysis that functional-adaptive approaches provide and which test the nature of similarities before these are used either for paleobiological assessment or phylogenetics.

Modeling relies heavily on theoretical perspectives, as well as the experiences of the modelers with the subjects that they are focused on. A far more desirable procedure than mere similarity matching is the construction of mechanically and adaptively meaningful relationships in character complexes in a number of distantly related species that display attributes which are more likely convergent than homologous (e.g., Szalay, 1981a). One may call this a convergence-based "modular-function" approach. It is important to have some strong ecologically compelling evidence that certain recurrent attributes are (given a similar level of basic mechanical organization of the skeletal biology) under strong selectional imperatives for their recurrent development. An understanding of functional-adaptive significance (and consequently the probability of convergence versus homology of properties) is decisive in establishing a list of tested taxonomic properties. This approach has both an inductive component in using the recurrent correlations between morphology and mechanics and the ecological context, as well as a deductive one in applying the correlations to the fossil taxa. Uncovering consistently convergent, biome-chanically significant, features that have strong functional associations with either feeding or locomotor strategies in the skeleton of extant mammals does supply us with powerfully modeled "postdictors" for adaptations in the fossil record.

Furthermore, if the probability is high that one or more aspects of properties in two or more taxa examined are the result of phyletically independent adaptive responses (rather than ancestral constraints), then, such convergent attributes (not to be considered taxonomic properties at a level higher than species) become excellent indicators of ecologically meaningful aspects of the fossils under study. Once the initial and boundary conditions (both phyletic and adaptive in a morphotype) are established for extant model species, and the fossils can be placed in a particular ecologically meaningful framework, then further analysis of the attributes of these fossil taxa becomes properly constrained for phylogenetically useful character analysis.

Models are particularly significant as they represent results of judiciously chosen surrogate evolutionary processes for a particular set of adaptive transformations. These selected extant models are chosen based on form-functional considerations with the heritage attributes often necessarily de-emphasized! These tested models (i.e., whose causal correlations with their various biological roles are well understood), as noted above, like all models, can never be a complete match for extinct organisms, or their aspects, that are subjected to analysis. Nevertheless, when size is controlled for, and functional (mechanical) attributes are correlated with some well-understood adaptations in the living models, many behaviors can be inferred for those fossils that share these features (for detailed examples see Court, 1994, for assessing Numidotherium, or Cifelli and Villarroel, 1997, for an interpretation of Megadolodus). Such procedures provide a corroborated level of character explanation (to varying degrees), both functional (nomothetic, nomological in essence) and phyloge-netic (evolutionary, i.e., unique, idiographic, historical). Szalay and Sargis (2001) have demonstrated this to be the case in their use of selected osteo-logical attributes of four extant model species of metatherians (boosted by numerous other examples examined there in less detail) for interpreting adaptive strategies in fossil marsupials.

In light of the foregoing I should comment here on the use of Caluromys, and various concepts of the didelphid ancestry, as models for interpreting the origins of the euprimates or their relatives, the plesiadapiforms. Morphology is the only point of reference that fossils can offer for analysis, and similarly, the assessed morphotype locomotor mode of a group is grounded in osteology. This should be connected with functionally well understood similar, or instructively contrasting, morphology in proper models that represent aspects of extant species, whose biological roles have been well investigated. Explaining fossil morphology should not consist of picking a living species based on some behavioral criterion, and stating categorically that its behavioral or physiological state (or another attribute) was probably similar to that in a postulated fossil taxon or an inferred common ancestor. Unfortunately, sometimes this has been done in primatology (not frequently, fortunately) even when the morphology of the designated extant "model" is singularly dissimilar to the inferred fossil condition. This dissimilarity is not only phyletic (as expected) but functional as well. The use of some marsupials is a case in point. For example, Rasmussen (1990) chose the didelphid Caluromys as a "model" for the protoeuprimate. Some of the factors he recognized, regarding arboreal adaptations of the euprimates, were no doubt correct, but these are not diagnostic of the stem of that clade. The type of arboreality displayed by arboreal didelphids, however, is a very good approximation of what the emerging evidence suggests for plesiadapiforms. Caluromys, therefore, may be a very good model for the origins of arboreality for the archontan or plesiadapiform stem.

Rasmussen (1990) posited that the relatively large brain and eyes, small litters, slow development (meaning postparturition because preparturition development is nearly uniform in all didelphids and fundamentally different from the universally "accelerated" condition of eutherians when these are compared to metatherians), and agile locomotion (compared to clumsier similar-sized arboreal didelphids such as the not infrequently terrestrial and scansorial species of Didelphis) represent a suite of attributes that is convergent to the euprimate ancestor. He stated (p. 263) that these "analogous...selection pressures, represent an independent test of the arboreal hypothesis,...the visual predation hypothesis,...and the angiosperm exploitation hypothesis of primate origins." Regrettably, the prehensile-tailed Caluromys does not have special similarity in its osteological properties to the diagnostic conditions of early euprimates, and therefore, cannot support the consensus of views envisaged by Rasmussen. Nevertheless, this was a useful analysis in that it resignaled the importance of marsupials for the study of archaic primates. However, among its numerous critical attributes the protoeuprimate, unlike the clawed Caluromys (which occasionally indulges in small leaps), had nails (for details see later section) and had a hindleg superbly adapted for leaping. No extant and arboreal marsupial comes close to the level of bio-mechanical attributes displayed by the Eocene euprimates. There are no osteological attributes of Caluromys that parallel euprimate osteological features, and therefore, this genus (or any didelphid) is an inappropriate model for interpreting euprimate ancestry. But a strong case can be made that, oste-ologically, Caluromys probably approximates a good model for the arboreal protodidelphid (but not for the didelphidan or sudameridelphian ancestry)— one that significantly differed in its advanced arboreal abilities from the post-cranially more primitive sudameridelphians of the Paleocene (Szalay, 1994; Szalay and Sargis, 2001) whose stem, in a departure from Cretaceous metatherians, may have been more terrestrial. The well-known agility of Caluromys (and other didelphids as well) compared to Didelphis, which is quite scansor-ial and is at home on terrestrial substrates, does not provide evidence for the argument that the agile arboreality of Caluromys is a derived condition within the Didelphidae. Many smaller species of didelphids are also quite agile and quick in an arboreal environment (see discussion of the Didelphidae in Szalay, 1994). Although a proposed model species like Caluromys tells us little about the origins of euprimate skeletal morphology (and therefore the inferred habits from that), it does, however, as noted, may be very useful for comparisons with archontans and plesiadapiforms. The stem euprimate lineage was likely transformed, via a still poorly understood arboreal archontan stage, from an essentially terrestrial placentalian heritage into an ancestor with a relatively well-understood primitive euprimate postcranial state whose obligate leaping behaviors were not unlikely (Dagosto, 1988; Dagosto et al., 1999; Szalay et al., 1987).

In attempting to explain arboreal attributes of the inferred common ancestry of euprimates, Lewis (1989, and references to his previous articles therein) has derived the various primate attributes from an essentially didelphid condition—the latter standing in as a surrogate for a "marsupial stage" prior to eutherian arboreality. Neither the phylogenetically troubling details that primates are eutherians with their own highly taxon specific constraining heritage that circumscribes their morphology, nor the fact that didelphids appear to be a particularly derived arboreal clade among South American Metatheria, have constrained Lewis' explanation. His transformational analysis lacked the necessary and appropriate phylogenetic context. Furthermore, many of the problems with his proposed transformations were also due to a lack of ecomorphologi-cally meaningful assessment of details. The general notion that some aspects of marsupials are probably primitive (e.g., their reproductive or developmental patterns) compared to their eutherian homologues does not mean that there is a functional similarity between eutherian skeletal attributes and those of didelphid marsupials (Szalay, 1984, 1994). Hence, the same applies even more emphatically to any attempt to understand euprimate origins based on didelphids.

Another inappropriate use of various modeled conditions of metatherian and eutherian skeletal adaptations was made by Martin (1990). He provided narratives, based on the contributions of Lewis (summarized in 1989), that were supposed to connect (historically!) metatherian morphology to the Paleocene plesiadapiform evidence, certainly well understood by that time in Plesiadapis. The explanations advanced by Martin heavily relied on implicit assumptions about the relevance of didelphid attributes for evaluating fossil eutherians. Martin confused the application of modeled properties in his text. He presented a lengthy, literature-based analysis of selected osteological attributes of euprimates and their possible closest relatives, specifically the ple-siadapiforms, colugos, and tupaiids. In writing about the evolution of mammalian locomotion, primate arboreality, and the specifics of the osteological evidence retrieved from the literature, a number of issues that relate to modeling and phylogenetic analysis of the metatherian-eutherian dichotomy framed his account. His views on the alleged homology of arboreality in marsupials and protoplacentalians, on the supposed "primitiveness" of the cheirogaleid primates within the euprimates, and the use of the various didelphid attributes for an arboreal habitat preference have provided confusing examples of modeling. Additionally, gross mistakes were committed when critical morphological details were misperceived or mistakenly reinterpreted from the literature.

It needs to be emphasized how important unexamined assumptions can be in any search for causal explanations of euprimate origins. Martin interpreted morphology in light of his assumption that ancestral placentalians were arboreal—a view which framed his ideas on the origin of the euprimate radiation. Interestingly, one who believed that the stem placentalian was arboreal (and who categorically continued to dismiss the relevance of the Plesiadapiformes) could accept the Archonta in spite of the fact that the modern rebirth of that concept (Szalay and Decker, 1974) was largely based on diagnostic arboreal adaptations (albeit taxon specific ones). Martin's published illustrations do not represent the actual morphology that he used to support his views. He overlooked, and missed the significance of the fact that, unlike the relatively free upper ankle joint adjustments in such primitive living marsupials as didel-phids (with their meniscus mediated fibular contact that puts little restraint on the upper ankle joint laterally), the protoplacentalian condition has evolved considerable tibial and fibular restraint for the upper ankle as reflected by the astragalus.

Similar, but taxon specific and independently evolved ankle restriction patterns can be found in obligate terrestrial marsupials like peramelids and macropodids. Martin and others failed to recognize (even though this has been painstakingly detailed in the literature) that the extensive lower ankle joint adjustments of plesiadapiforms, euprimates, and all other obligate arboreal placentalians became constrained by the protoplacentalian adaptation, and that the most extensive adjustments to pedal inversion have invariably occurred in these taxa in the lower ankle joint. As a result, evolution of a morphological complex in the lower ankle joint that facilitates inversion is invariably a derived condition among early placentalians that show such morphotypic attributes, albeit convergently, such as archontans, some lipotyphlans, creodonts, carnivorans, and rodents.

0 0

Post a comment