Taxonomy

The order Scandentia is represented by the single family Tupaiidae, which includes the subfamilies Ptilocercinae and Tupaiinae (Table 1). Ptilocercinae is represented only by Ptilocercus lowii, while Tupaiinae consists of Tupaia (14 species), Dendrogale (2 species), Urogale everetti, and Anathana ellioti (see Table 1; Wilson, 1993). Differences in postcranial morphology are often split down the subfamilial line, as Ptilocercus' postcranium is adapted for arboreal locomotion, while that of the tupaiines is adapted for terrestrial locomotion (Sargis, 2000, 2001, 2002a,b,c).

Table 1. Classification of tree shrews (Wilson, 1993)

Order Scandentia Family Tupaiidae Subfamily Tupaiinae

Tupaia (14 species) Dendrogale (2 species) Anathana ellioti Urogale everetti Subfamily Ptilocercinae Ptilocercus lowii

Supraordinal Relationships of Tupaiids

Tree shrews were originally included in the order Insectívora by Wagner (1855), and Haeckel (1866) later grouped them in the insectivoran suborder Menotyphla with the elephant shrews (see Table 2; Butler, 1972). A close relationship between tupaiids and primates was first suggested in 1910 when Gregory proposed the superorder Archonta, which included Chiroptera, Dermoptera, Primates, and Menotyphla (which he recognized as an order). Carlsson (1922) moved tupaiids to the order Primates, and Le Gros Clark (1924a,b, 1925, 1926) strongly supported this grouping with his studies of tupaiid anatomy (Table 2). Simpson (1945) considered Archonta to be an unnatural group, but at the same time he supported the inclusion of tupaiids in the order Primates (Table 2). Tupaiids continued to be classified as Primates until the 1960s (see Napier and Napier, 1967), when they were removed from

Table 2. History of tree shrew ordinal designations

Order Insectivora Wagner (1855)

Removed from order Primates

Order Scandentia

Order Menotyphla Order Primates

Haeckel (1866) Gregory (1910) Carlsson (1922)

Le Gros Clark (1924a,b, 1925, 1926)

Simpson (1945)

Napier and Napier (1967)

Van Valen (1965) Jane et al. (1965) McKenna (1966) Campbell (1966a,b) Martin (1966, 1968a,b) Szalay (1968, 1969)

Butler (1972)

the order by Campbell (1966a,b), Jane et al. (1965), Martin (1966, 1968a,b), McKenna (1966), Szalay (1968, 1969), and Van Valen (1965) (Table 2). Most similarities between the two groups were deemed to be erroneous observations, shared primitive characters, or convergences found only in derived representatives of the groups rather than ancestral morphotypes. Once tupaiids were removed from the order Primates, Butler (1972) classified them in their own order Scandentia (Table 2)—a name used (at the family level) by Wagner (1855) for tupaiids. In 1975, McKenna accepted Butler's (1972) classification of tupaiids as an independent order Scandentia, and he revised Gregory's (1910) superorder Archonta by including the orders Scandentia, Primates, Chiroptera, and Dermoptera and excluding the Macroscelidea (elephant shrews). Szalay (1977) supported this grouping with evidence from the tarsus, and within the Archonta, he recognized a Primate-Scandentia clade and a Chiroptera-Dermoptera clade (called Volitantia; see Figure 1A; Table 3).

Szalay and Drawhorn (1980) found further support for the archontan hypothesis in the tarsus of fossil plesiadapiforms and other archontan mammals. Novacek and Wyss (1986) came to support the archontan hypothesis (Table 3) based on one tarsal character (from Szalay and Drawhorn, 1980) and one penial character (from Smith and Madkour, 1980), despite Novacek's (1980, 1982, 1986) previous agreement with Cartmill and MacPhee (1980) that this hypothesis was not supported by cranial or post-cranial evidence. Since 1986, Novacek has become one of the major proponents of both the Archonta and Volitantia hypotheses (Novacek, 1989, 1990, 1992, 1993, 1994; Novacek et al., 1988). Wible (Wible and Covert, 1987; Wible and Novacek, 1988) has also supported these hypotheses, and has provided evidence for a Scandentia-Euprimates clade (Table 3) that does not include the "archaic primates" (Plesiadapiformes) (also supported by Kay et al., 1992). Some subsequent studies have continued to bolster support for the Archonta hypothesis (Johnson and Kirsch, 1993; Shoshani and McKenna, 1998; Szalay and Lucas, 1993, 1996), while others have rejected it (e.g., Kay et al., 1990, 1992). Recently, McKenna and Bell (1997) have reconfirmed support of the Archonta in their classification of mammals.

Within the Archonta, however, interordinal relationships have not been agreed upon. For instance, Beard's (1989) detailed functional morphological analysis of archontan postcranial morphology led him to reject the Volitantia hypothesis. Beard argued that plesiadapiforms should be included in the order Dermoptera (a hypothesis also supported by Kay et al., 1990, 1992)

Volitantia

Primatomorpha

Scandentia Pnmates CMroptera termoptera Scandentia Chiroptera Primates Dermoptera

"Primatomorpha"

Volitantia

Scandentia Primates Chiroptera Dermoptera

Scandentia Primates Chiroptera Dermoptera

Euarchonta

Dermoptera

Dermoptera

Primates

Scandentia

Euarchonta

Primatomorpha

Dermoptera

Volitantia

Chiroptera Dermoptera

Dermoptera

Primates

Chiroptera Dermoptera

Scandentia

Primates

Figure 1. Selected views of archontan phylogenetic relationships supported by: (A) Novacek (1992), Szalay (1977), Wible and Covert (1987), Wible and Novacek (1988); (B) Beard (1993b); (C) Shoshani and McKenna (1998); (D) Liu and Miyamoto (1999), Liu et al. (2001), Murphy et al. (2001a,b), this study (Analysis 3); (E) Waddell et al. (1999); and (F) Silcox (2001a,b, 2002), this study (Analysis 2).

Primates

Scandentia

Scandentia

Primates

Scandentia

Primates

Table 3. Clades supported by various phylogenetic analyses discussed in the text

Archonta or

Euarchonta Within Archonta

Szalay, 1977

Archonta

Volitantia, Primates +

Scandentia

Novacek and Wyss, 1986

Archonta

Volitantia

Wible and Novacek, 1988

Archonta

Volitantia, Euprimates +

Scandentia

Beard, 1993b

Archonta

Primatomorpha

Shoshani and McKenna, 1998

Archonta

Volitantia

Waddell et al., 1999

Euarchonta

Primatomorpha

Liu and Miyamoto, 1999

Euarchonta

Dermoptera + Scandentia

Liu et al., 2001

Euarchonta

Dermoptera + Scandentia

Murphy et al., 2001a,b

Euarchonta

Dermoptera + Scandentia

Silcox, 2001a,b, 2002

Archonta

Volitantia, Primates (sensu lato)

and that it is Dermoptera and Primates, rather than Dermoptera and Chiroptera that form a natural group, which he named Primatomorpha (Figure 1B; Table 3). The evidence for this grouping includes similarities between the morphology of the intermediate manual phalanges of der-mopterans and paromomyids that may be related to gliding (Beard, 1990, 1993a), but other characters have been used to support this hypothesis as well (Beard, 1993b). This grouping of Dermoptera with Primates contrasts sharply with the grouping of Dermoptera with Chiroptera in Volitantia, which has been strongly supported in numerous morphological studies using extremely varied databases and phylogenetic methods (see Bloch et al., 2002; Johnson and Kirsch, 1993; Kriz and Hamrick, 2001; Novacek, 1982, 1986, 1989, 1990, 1992, 1993, 1994; Novacek and Wyss, 1986; Novacek et al., 1988; Sargis, 2002d; Shoshani and McKenna, 1998; Silcox, 2001a,b, 2002; Szalay, 1977; Szalay and Lucas, 1993, 1996; Thewissen and Babcock, 1991, 1993; Wible, 1993; Wible and Covert, 1987; Wible and Novacek, 1988). Simmons (1995), Simmons and Quinn (1994), and Thewissen and Babcock (1992) have also supported Volitantia rather than Primatomorpha, but they incorporated Beard's (1990, 1993a) results by including the paromomyids in the order Dermoptera (but see Bloch and Silcox, 2001). McKenna and Bell (1997), while not recognizing Beard's Primatomorpha, also included the Paromomyidae in Dermoptera and further recognized Beard's (1993a,b) results by including Dermoptera as a suborder of the order Primates (but see Szalay, 1999, for criticisms). Most of the other plesiadapiform families were placed by McKenna and Bell (1997) in the order Primates, but not specifically in the suborder Dermoptera. A curious exception was the placement of the Carpolestidae within the Tarsiiformes in the suborder Euprimates (see Szalay, 1999). Shoshani and McKenna (1998) recognized both Volitantia and Primatomorpha, but their "Primatomorpha" was a grouping of Volitantia and Primates (Figure 1C; Table 3). This concept of Primatomorpha has surely lost the meaning that Beard (1993a,b) intended because Shoshani and McKenna (1998) argued that Chiroptera was the closest relative of Dermoptera, while Beard (1993a,b) supported a Dermoptera-Primates clade. It is significant that when Beard's (1993b) data set was incorporated into the much larger data set of Shoshani and McKenna (1998), the signal for a Dermoptera-Primates clade was lost, whereas a Dermoptera-Chiroptera clade was supported. Recently, Bloch et al. (2002) and Silcox (2001a,b, 2002) supported both a Volitantia-Scandentia clade (Figure 1F) and a Plesiadapiformes-Euprimates clade (i.e., Primates, sensu lato; Table 3). Their studies rejected Beard's (1993a,b) Primatomorpha, and Silcox's (2001a,b, 2002) classification included plesiadapiforms in Primates rather than in Dermoptera (contra Beard, 1989, 1993a,b).

Beard (1989, 1990, 1991, 1993a,b) has clearly advanced the debate about archontan phylogenetics, but his studies have been criticized on numerous grounds. Krause (1991) questioned the identifications and associations of the paromomyid specimens that Beard (1989, 1990) analyzed. Szalay and Lucas (1993, 1996) questioned and reevaluated many of the postcranial characters that Beard used to support both his concept of Primatomorpha and his hypothesis concerning the "mitten" or finger-gliding capabilities of paro-momyids (based on intermediate phalangeal proportions). Simmons (1994) showed that 2 of Beard's (1993b) 29 characters included erroneous observations, while Stafford and Thorington (1998) showed that 2 additional characters included erroneous observations and another character was misinterpreted (see Sargis, 2002d). Shoshani and McKenna (1998) used only 12 of Beard's (1993b) 29 characters in their phylogenetic analysis because they said that Beard himself stated that the others were questionable. Hamrick et al. (1999) rejected the "mitten" or finger-gliding capabilities of paromomyids based on their analysis of phalangeal proportions. They did recognize several similarities between the phalanges of dermopterans and paro-momyids; however, they interpreted these features not as gliding adaptations but as adaptations for vertical clinging and climbing on large arboreal supports. They also showed that a phalangeal feature previously believed to be unique to paromomyids and dermopterans (Beard, 1993b) is also found in chiropterans (see below; Thewissen and Babcock, 1992), and they identified an additional derived phalangeal character shared by chiropterans, dermopterans, and paromomyids (Hamrick et al., 1999). These characters support the amended volitantian concept (where paromomyids are included in Dermoptera; but see Bloch and Silcox, 2001) of Simmons (1995), Simmons and Quinn (1994), and Thewissen and Babcock (1992) rather than Beard's (1993a,b) Primatomorpha. Lemelin (2000) further supported Volitantia rather than Primatomorpha with a unique feature of the volar skin that is shared by dermopterans and chiropterans. Finally, Sargis (2002d) demonstrated that 12 of Beard's (1993b) 22 postcranial characters should be interpreted differently when Ptilocercus, rather than Tupaia, is used to represent Scandentia, which greatly reduces the evidence for Primatomorpha.

In contrast to many of these morphological studies, molecular studies have consistently supported a group that includes Dermoptera, Scandentia, and Primates to the exclusion of Chiroptera (Adkins and Honeycutt, 1991, 1993; Allard et al., 1996; Cronin and Sarich, 1980; Honeycutt and Adkins, 1993; Liu and Miyamoto, 1999; Liu et al., 2001; Murphy et al., 2001a,b; Porter et al., 1996; Waddell et al., 1999). This group has been called Euarchonta by Waddell et al. (1999). Despite this apparent consensus regarding the exclusion of bats from the Archonta based on molecular evidence, there has been little agreement concerning which order represents the closest relative of the remaining archontan orders (see Allard et al., 1996). Several orders, including Macroscelidea, Lagomorpha, Rodentia, and occasionally both Lagomorpha and Rodentia (placed in the supraordinal grouping Glires), have been proposed to be more closely related to the remaining members of the Archonta than are bats (Allard et al., 1996; Bailey et al., 1992; Goodman et al., 1994; Honeycutt and Adkins, 1993; Madsen et al., 2001; Miyamoto, 1996; Murphy et al., 2001a,b; Porter et al., 1996; Stanhope et al., 1993, 1996; Waddell et al., 1999). Recently, a Euarchonta-Glires clade has received the most support (Murphy et al., 2001a,b), and this clade has been named Euarchontoglires by Murphy et al. (2001b).

Some molecular studies have specifically supported a Scandentia-Lagomorpha clade. Graur et al. (1996) argued that lagomorphs are very closely related to primates and tree shrews, and they tentatively concluded that Lagomorpha represents the sister taxon of Scandentia. This hypothesis was also supported by Schmitz et al. (2000). However, Liu and Miyamoto

(1999) recently found the most support for a Dermoptera-Scandentia clade (also supported by the molecular data of Liu et al., 2001; Madsen et al., 2001; Murphy et al., 2001a,b; see Figure 1D; Table 3), while Waddell et al.

(1999), in the same volume, grouped Dermoptera with Primates in the Primatomorpha (also supported by the molecular analyses of Teeling et al., 2000; Killian et al., 2001; see Figure 1E; Table 3). The latter conclusion is particularly significant with regard to the analysis conducted by Graur et al. (1996), who stated that the "phylogenetic position of Dermoptera relative to Primates and Lagomorpha could not be resolved with the available data" (p. 335). Liu and Miyamoto's (1999) conclusion is also noteworthy because Graur et al. (1996) never tested a Scandentia-Dermoptera relationship and Schmitz et al. (2000) did not include Dermoptera in their analysis. Perhaps Graur et al. (1996) and Schmitz et al. (2000) would not have supported a Scandentia-Lagomorpha relationship if they had included a test of a Scandentia-Dermoptera clade in their studies (see Liu and Miyamoto, 1999; Liu et al., 2001; Madsen et al., 2001; Murphy et al., 2001a,b). Similarly, it is possible that Primatomorpha would not have been supported by Teeling et al.

(2000) if they had included Scandentia in their analysis, thereby testing a Scandentia-Dermoptera relationship.

The exclusion of chiropterans from Archonta based on molecular data is not only a revision of the morphological concept of Archonta, but it is also a rejection of the Volitantia hypothesis. Hence, the molecular concept of Euarchonta is incompatible with the morphological concepts of Archonta and Volitantia. It seems, therefore, that there is little congruence between morphological and molecular data concerning these alternative phylogenetic hypotheses. Beard's Primatomorpha hypothesis (1989, 1993a,b), however, is based on morphological evidence and is concordant with the molecular concept of Euarchonta. It is the competing hypotheses of Volitantia and Primatomorpha that will be considered here in a reexamination of some of the postcranial evidence.

Significance of Ptilocercus

A study of the postcranium of the least well-known order in the Archonta, Scandentia, was undertaken by Sargis (2000) in order to provide additional information to contribute to an understanding of the relationships among archontan mammals. The inclusion of Ptilocercus in this study was critical because Ptilocercus has long been considered to be the living taxon most closely resembling the ancestral tupaiid in both its ecology and morphological attributes (Butler, 1980; Campbell, 1974; Emmons, 2000; Gould, 1978; Le Gros Clark, 1926; Martin, 1990; Sargis, 2000, 2001, 2002a,b,c,d; Szalay, 1969; Szalay and Drawhorn, 1980; Szalay and Lucas, 1993, 1996) and thus must play a paramount role in any supraordinal phylogenetic analysis that includes the Tupaiidae. A better understanding of tupaiid supraordinal relationships is likely confounded by the common use of Tupaia, a relatively derived tupaiid (see Martin, 1990), to represent Scandentia in studies of mammalian supraordinal relationships (e.g., Beard, 1989, 1993b), and these relationships would likely be better understood if Ptilocercus was included in the analysis (Sargis, 2002d). Most previous studies have also used Tupaia, rather than Ptilocercus, as an outgroup when the relationships among various groups of primates were being examined (e.g., Shoshani et al., 1996).

In this chapter, I reanalyzed Beard's (1993b) data specifically in light of the fact that Beard (1993b) did not use Ptilocercus to represent Scandentia in his analysis. I also recoded characters based on an a priori character analysis (Sargis, 2002d), and added additional postcranial characters to see how they affect the results.

Was this article helpful?

0 0

Post a comment