Taxonomic Implications of the Current Analysis

In light of these results the issue of whether or not to include plesiadapiforms becomes one entirely of taxonomic philosophy. If one wishes to emphasize adaptive cohesiveness there is no question that the euprimate clade is more adaptively cohesive than any grouping that includes plesiadapiforms. The common ancestor at the euprimate node can be reconstructed as possessing a long list of newly derived characters which, together, seem to be associated with improvements to the visual system (reduced snout; complete postorbital bar; large optic foramina) and modifications of the postcranial skeleton for leaping (third trochanter at the same level as the lesser trochanter; deep distal femur; long astragalar neck; humerofemoral index less than 70; distal cal-caneus elongate; lateral side of the femoral trochlea more anteriorly projecting than the medial side). The plesiadapiform + euprimate node is largely supported by dental features that are less easily interpreted as a functional complex (e.g., postprotocingulum present on P4, enlarged M3 hypoconulid)—this is hardly surprising, however, in light of the fact that the basal-most plesiadapiforms are known only from teeth. A crown-clade approach would also exclude plesiadapiforms from Primates, since all plesi-adapiforms appear to lie outside of the clade that would include modern Primates (although no modern Primates were actually included in the list of taxa studied). An apomorphy-based approach would include or exclude ple-siadapiforms depending on which feature was chosen. As the only character that has been identified as being unique to Primates (Wible and Covert, 1987), the presence of a petrosal bulla is likely the only choice that could hope to garner a consensus as such a key "indicator apomorphy". In spite of this, the petrosal bulla is a manifestly impractical choice. First, we do not yet know exactly how a petrosal bulla was acquired, but it seems unlikely that it appeared in an evolutionary instant. Wible and Covert (1987), for example, highlight similarities in scandentians and modern Primates in the expansion of part of the petrosal, the tegmen tympani, which might indicate part of the pathway leading from a nonpetrosal bulla to one formed exclusively by this bone. Specifying just "presence of a petrosal bulla" as the indicator apomor-phy has the potential, therefore, to open up questions about how much of a petrosal bulla is adequate, particularly if fossils representing intermediate stages do become available. Second, and more importantly, it has been pointed out that it is impossible to demonstrate conclusively whether a fossil has a petrosal bulla. The morphology of both Plesiadapis (Russell, 1959, 1964; Silcox, 2001; Szalay, 1972; Szalay et al., 1987) and Carpolestes (Bloch and Silcox, 2003, 2006; Silcox, 2001) is consistent with the presence of a petrosal bulla. As MacPhee et al. (1983; see also Beard and MacPhee, 1994; MacPhee and Cartmill, 1986) point out, however, this identification cannot be confirmed without developmental evidence, which will likely never be available for these taxa. A character that cannot be confidently identified in the taxa of greatest interest and dispute is singularly inappropriate as a delimiter for a taxonomic group. It is unlikely, however, that any other single character would be supported as a "key apomorphy" for Primates, making this approach generally impractical for delimiting the order.

In any case, none of these options are informative about how plesiadapi-forms should be classified if they are not to be considered Primates. One option is to place plesiadapiforms in a wastebasket Insectivora. Cartmill (1972, 1974, 1992) has advocated this approach in the past, in spite of dental (Gidley, 1923; Simpson, 1935a), cranial (Russell, 1959, 1964) and postcranial (Russell, 1964; Szalay et al., 1975) features that were already well known at that time to be shared by plesiadapiforms and Primates, and that are missing in insectivorans. The end result of this re-classification might make Primates easier to define, but makes Insectivora a meaningless, polyphyletic assemblage. This also ignores the fact that Insectivora is not, in fact, "available" to be a wastebasket for taxa that anthropologists do not want to deal with. The makeup and evolutionary relationships of forms traditionally included in Insectivora (more properly Lipotyphla) are a subject of very intense current debate. This debate has been stirred up by molecular discoveries (e.g., Stanhope et al., 1998), which suggest multiple origins of Insectivora, and their apparent conflict with morphological data (e.g., Asher, 1999). If one's only goal is to provide a clear definition of Primates these debates may seem irrelevant—what an insectivore is, and who they are related to, becomes an unimportant question. However, if the order is to be understood in the broader context of mammalian evolution, these debates are vitally important. A common finding of molecular results is that primates are part of a clade with dermopterans and scandentians (Euarchonta; Liu and Miyamoto, 1999; Liu et al., 2001; Pumo et al., 1998; Waddell et al., 1999) that may be closely related to Glires (rodents + rabbits) and only very distantly related to any traditional insectivorans (Waddell et al., 1999; Murphy et al., 2001a,b; Madsen et al., 2001; Springer et al., 2006). As such, Insectivora would be an entirely unsuitable place to put taxa that are closely related to modern primates (or dermopterans).

Also, we cannot simply ignore or dump into a wastebasket taxon those forms that do not already display all the features present in extant groups, rendering them by implication unimportant to questions of primate evolution. It is these forms that will be most crucial, in fact, in helping us understand which unusual euprimate features did actually arise as part of a particular adaptive complex. Recent discoveries highlight this fact. It is now clear from novel discoveries of plesiadapiform postcranial material that most of the features associated with grasping predate the common ancestor of modern Primates (Bloch and Boyer, 2002, 2003 and 2006), and evolved much earlier than characteristics such as convergent orbits or the postorbital bar. As such, any adaptive scenario for the origin of euprimates that links together grasping and visual features in a single pattern of change must be incorrect, because the evolutionary transitions involved occurred at significantly different points in time.

Another option would be to use Plesiadapiformes (or Proprimates; see Gingerich, 1989) itself as a separate order. However, the fact that Plesiadapiformes does not appear to be a monophyletic grouping implies that it is not possible to classify that cluster as a separate order, unless one is willing to provide a formal taxonomic label for a non-monophyletic group. I would argue against this approach for two reasons. First, modern taxonomic practice frowns on the use of non-monophyletic groupings. It is likely that any such taxonomy would be subject to rapid revision, or at least derision, along these lines. Second, and more importantly, dumping these forms into a different, common wastebasket (Plesiadapiformes or Proprimates rather than Insectivora) still obfuscates the central point that some plesiadapiforms are more closely related to euprimates than others. The structure of the tree, in terms of the branches leading up to the group that is of central interest to most (Euprimates), is actually very important, since it documents what steps precede that node and what forms are the best models for the common eupri-mate ancestor. Failing to recognize this taxonomically by dumping all plesi-adapiforms into a group together, therefore suggesting that they can be treated as a unit, would lead to obfuscation of this important point. Although I think continuing to use "plesiadapiforms" informally is useful, I would not advocate applying this term as a formal taxonomic label.

Applying a stem-based definition is also problematic. Primates could be defined as the clade consisting of Euprimates and all organisms that share a more recent common ancestor with Euprimates than with the common ancestor of Volitantia + Scandentia. This definition assumes, however, that the Volitantia + Scandentia clade is well supported, which is really not the case. Particularly, some of the results of this study suggested that Scandentia might be the basal-most group in Archonta. This would imply that a common ancestor of Volitantia and Scandentia not shared with Euprimates never existed. Also, this possibility means that the stem-based definition suggested earlier in this article (i.e., the clade consisting of Haplorhini and all organisms that share a more recent common ancestor with Haplorhini than with Scandentia) could mandate the inclusion of Volitantia in Primates. I believe that this would stretch the adaptive boundaries of Primates in a way unacceptable to most. Finally, the assumption of the monophyly of Archonta made for this analysis does not take into consideration recent molecular results (e.g., Miyamoto et al., 2000; Pumo et al., 1998; Waddell et al., 1999;) that suggest chiropterans belong in a clade with carnivores and ungulates, rather than as part of a monophyletic Volitantia with dermopterans. In light of these uncertainties, and although I am generally an advocate of a stem-based approach to delimiting groups, it appears to be an impractical approach for Primates at the current time.

I suggest that a node-based approach is the most appropriate here. If Primates is defined as the clade stemming from the most recent common ancestor of Purgatorius and Euprimates (or Haplorhini, etc.), we have a solution that provides a satisfactory and historically consistent ordinal designation for both plesiadapiforms and Euprimates. What is more, because Purgatorius is a very primitive form, most future discoveries of euprimate stem taxa or plesiadapiforms will be easily accommodated in this definition. As such, this is essentially equivalent to a stem-based definition, while avoiding the problems created by uncertainty over the relationships of the rest of Archonta (or Euarchonta). This definition does not rely on any characters that cannot be assessed in fossils (e.g., the petrosal bulla), and avoids the defeatist and shortsighted approach that dumps all difficult to classify fossils into a meaningless wastebasket taxon.

0 0

Post a comment