Rodentia Bowdich 1821

Wilson, Don E. & Reeder, DeeAnn, 2005, Order Rodentia - Family Aplodontiidae, Mammal Species of the World: a Taxonomic and Geographic Reference (3 rd Edition), Volume 2, Baltimore: The Johns Hopkins University Press, pp. 753-753 : 745

Rodentia Bowdich 1821

Families: 33 families with 481 genera and 2277 species in 5 suborders and 2 infraorders:

Suborder SCIUROMORPHA Brandt 1855

Family Aplodontiidae Brandt 1855 (1 genus with 1 species and 7 subspecies)

Family Sciuridae Fischer de Waldheim 1817 (51 genera with 278 species and 879 subspecies)

Family Gliridae Muirhead 1819 (9 genera with 28 species)

Suborder CASTORIMORPHA A. E. Wood 1955

Family Castoridae Hemprich 1820 (1 genus with 2 species)

Family Heteromyidae Gray 1868 (6 genera with 60 species and 303 subspecies)

Family Geomyidae Bonaparte 1845 (6 genera with 40 species and 351 subspecies)

Suborder MYOMORPHA Brandt 1855

Family Dipodidae Fischer de Waldheim 1817 (16 genera with 51 species)

Family Platacanthomyidae Alston 1876 (2 genera with 2 species)

Family Spalacidae Gray 1821 (6 genera with 36 species)

Family Calomyscidae Vorontsov and Potapova 1979 (1 genus with 8 species)

Family Nesomyidae Major 1897 (21 genera with 61 species)

Family Cricetidae Fischer 1817 (130 genera with 681 species)

Family Muridae Illiger 1811 (150 genera with 730 species and 5 subspecies)

Suborder ANOMALUROMORPHA Bugge 1974

Family Anomaluridae Gervais 1849 (3 genera with 7 species and 2 subspecies)

Family Pedetidae Gray 1825 (1 genus with 2 species)

Suborder HYSTRICOMORPHA Brandt 1855

Infraorder CTENODACTYLOMORPHI Chaline and Mein 1979

Family Ctenodactylidae Gervais 1853 (4 genera with 5 species)

Infraorder HYSTRICOGNATHI Brandt 1855

Family Bathyergidae Waterhouse 1841 (5 genera with 16 species and 16 subspecies)

Family Hystricidae G. Fischer 1817 (3 genera with 11 species and 7 subspecies)

Family Petromuridae Wood 1955 (1 genus with 1 species and 15 subspecies)

Family Thryonomyidae Pocock 1922 (1 genus with 2 species and 2 subspecies)

Family Erethizontidae Bonaparte 1845 (5 genera with 16 species and 14 subspecies)

Family Chinchillidae Bennett 1833 (3 genera with 7 species and 25 subspecies)

Family Dinomyidae Peters 1873 (1 genus with 1 species)

Family Caviidae Fischer de Waldheim 1817 (6 genera with 18 species and 27 subspecies)

Family Dasyproctidae Bonaparte 1838 (2 genera with 13 species and 29 subspecies)

Family Cuniculidae Miller and Gidley 1918 (1 genus with 2 species and 5 subspecies)

Family Ctenomyidae Lesson 1842 (1 genus with 60 species and 27 subspecies)

Family Octodontidae Waterhouse 1839 (8 genera with 13 species and 3 subspecies)

Family Abrocomidae Miller and Gidley 1918 (2 genera with 10 species and 2 subspecies)

Family Echimyidae Gray 1825 (21 genera with 90 species and 33 subspecies)

Family Myocastoridae Ameghino 1904 (1 genus with 1 species and 4 subspecies)

Family Capromyidae Smith 1842 (8 genera with 20 species and 7 subspecies)

Family Heptaxodontidae Anthony 1917 (4 genera with 4 species)

Discussion: Rodentia is the largest order of living Mammalia, encompassing 2277 species as recognized herein, or approximately 42% of worldwide mammalian biodiversity. Following the mid-1900s era of uncritical application of the biological species concept and consequent obscuration of species richness, views on the size of the Order continue to appreciate substantially (1591 species— Corbet and Hill, 1980; 1719— Honacki et al., 1982; 1738— Corbet and Hill, 1986; 2015— Wilson and Reeder, 1993; 2277—this volume). Thus, "a checklist of species" considered valid is an appropriate taxonomic focus in the current work, as emphasized by its title and that of its predecessors ( Honacki et al., 1982; Wilson and Reeder, 1993). In contrast to the period covered by the second edition ( Wilson and Reeder, 1993), however, systematic research on Rodentia since 1993 has been equally as prolific, multifaceted, and informative at taxonomic levels above the species and requires introductory comment. Issues of monophyly, phyletic relationship, and corresponding classification at the genus- and family-group ranks are addressed to varying depths in the following chapters. This brief foreword offers remarks on the Order as a whole, particularly evidence that bears on its monophyly and on suprafamilial relationships. Although 20 years old, the 1984 Paris symposium on rodent evolutionary relationships, edited by Luckett and Hartenberger (1985 a), today remains an excellent and invaluable primer to these same issues.

Rodent monophyly and ordinal relationships.—The evolution of living rodents from some Paleocene common ancestor and their acceptance as a monophyletic order of Mammalia have not been seriously challenged until recently. The provocative inquiry "Is the guinea-pig a rodent?" ( Graur et al., 1991) leapt precipitously to the conclusions that "The guinea-pig is not a rodent" (D’Erchia et al., 1996), that, by extension, hystricognaths represent a separate mammalian order, and that Rodentia as conventionally viewed is polyphyletic (also Frye and Hedges, 1995; Graur et al., 1992; Li et al., 1992). Attention in molecular investigations just as swiftly turned to, foremost, the adequacy of taxon sampling persuant to the taxonomic question posed, to the appropriateness of models that account among-site rate heterogeneity, and to a preference for nuclear over mitchondrial genes (and for multiple over single) in illuminating deeper divisions of mammalian phylogenesis (e.g., see Adkins et al., 2001; Cao et al., 1997; Corneli, 2002; Lin et al., 2002; Luckett and Hartenberger, 1993; Philippe, 1997; Sullivan and Swofford, 1997). The subsequent wave of sleeves-up research, characterized by wider sampling of taxa and genes, has made the case for rodent monophyly vastly more secure ( Adkins et al., 2001, 2003; Amrine-Madsen et al., 2003; Debry, 2003; Debry and Sagel, 2001; Huchon et al., 1999, 2002; Lin et al., 2002; Nedbal et al., 1996; Waddell and Shelley, 2003). And in those molecular studies that have focused on hystricognaths in general or cavioids in particular ( Huchon and Douzery, 2001; Kramerov and Vassetzky, 2001; Lara et al., 1996; Leite and Patton, 2002; Marivaux et al., 2002; Mouchaty et al., 2001; Rowe and Honeycutt, 2002; and see references under Hystricognathi ), the guinea pig nests unremarkably within the kinship web so long predicated by so many past studies and classifications.

An implicit but faulty premise in these early molecular studies perhaps engendered the over-eager acceptance of rodent polyphyly: that is, the evolution of large, chisel-like incisors in other mammalian groups (e.g., the primate Daubentonia and diprotodont marsupials) underscores the likelihood for the independent acquisition of enlarged incisors in other distantly related lineages, including and especially those that had been mistakenly lumped under Rodentia based on this single, homoplasious trait. The point is that the hypertrophied incisors in Rodentia represent retained deciduous second incisors (dI2/2), whereas the gliriform teeth in those other orders consist of permanent first incisors (I1/1). The embryogenesis of rodent incisors from the dI2 was appreciated by the late 1800s and constitutes a persuasive synapomorphy for the Order (see Landry, 1999, Luckett, 1985, and references therein). To demonstrate that the adult incisors in one or more other rodent groups develop from an anlage other than the dI2 would more severely cripple the case for monophyly of Rodentia . The morphological character information historically used to argue the common ancestry of rodents is wonderfully richer than this single feature, however, as anyone who peruses the plates of Tullberg’s (1899) monograph will readily appreciate. See Carleton (1984) for general morphological characterization of the Order; Hartenberger (1985), Landry (1999), and Luckett and Hartenberger (1993) provide phylogenetic interpretation of derived characters and integrated character suites that support monophyly of Rodentia . In retrospect, we are fortunate that Tullberg based his phylogenetic study on more than a single species each of Cavia , Mus , and Rattus .

Rodentia and Lagomorpha have long been advocated as sister taxa, with the two groups initially placed in the same order, Rodentia or Glires , usually as the suborders Simplicidentata and Duplicidentata ( Alston, 1876; Brandt, 1855; Gregory, 1910; Thomas, 1897 c; Tullberg, 1899). After elevation to separate orders ( Gidley, 1912; Miller and Gidley, 1918), indication of cognate affinity was still conveyed, e.g., as the Cohort or Superorder Glires ( Landry, 1974, 1999; Luckett and Hartenberger, 1985 b; Simpson, 1945). For some time, a paleontological foundation for their supposed shared ancestry proved disconcertingly elusive ( Gregory, 1910; Simpson, 1945, 1961), but such hard fossil evidence has proliferated rapidly in the later 1900s ( Bryant and McKenna, 1995; Dashzeveg et al., 1998; Li and Ting, 1985, 1993; Li et al., 1987). The current picture of earliest gliriform radiation and classification is complex, and the cladistic position of various groups, notably eurymylids in the broad sense, as basal to or subtended by the crown-clade Rodentia has been variously interpreted (see Meng and Wyss, 2001, and Meng et al., 2003, for phylogenetic analysis, review, and discussion). Sister-group stature of Lagomorpha and Rodentia is decisively sustained in modern phylogenetic studies, both those drawing on morphological ( Ade, 1999; Frahnert, 1999; Giere et al., 1999; Landry, 1999; Luckett and Hartenberger, 1993; Martin, 1999; Meng and Wyss, 2001; Meng et al., 2003; Mess, 1999; Novacek, 1985, 1986; Novacek and Wyss, 1986; Novacek et al., 1988; Shoshani and McKenna, 1998) and molecular data sources ( Amrine-Madsen et al., 2003; Delsuc et al., 2002; Eizirik et al. 2001; Huchon et al., 2002; Lin et al., 2002; Murphy et al., 2001 a, b; Waddell and Shelley, 2003). Other orders besides Lagomorpha have been identified as closely related to Rodentia , in particular Primates ( Patterson and Wood, 1982; Wood, 1962). Such proposals merit some vindication, albeit phyletically more remote, in the emerging recognition of the superorder Euarchontoglires (Dermoptera-Scandentia-Primates + Lagomorpha-Rodentia— Archibald, 2003; Delsuc et al., 2002; Helgen, 2003 a; Huchon et al., 2002; Lin et al., 2002; Murphy et al., 2001 a, b; Scally et al., 2001; Waddell and Shelley, 2003). Zalambdalestids from the late Cretaceous of Asia have also been claimed as sister-group to Rodentia-Lagomorpha ( Archibald, 2003; Archibald et al., 2001), but their close link is unsupported by analyses employing a greater range of relevant taxa and characters ( Ji et al., 2002; Meng et al., 2003; Wible et al., 2004).

If Eurymylidae , including Heomys , is considered a family of Rodentia (e.g., Li and Ting, 1993), then the geological range of the Order extends from the early-middle Paleocene. If not (e.g., Meng and Wyss, 2001; Wyss and Meng, 1996), then several ischyromyid genera from the late Paleocene appear to be the earliest true rodents yet recorded ( Hartenberger, 1998; McKenna and Bell, 1997). Explosive diversification into principal lineages (suborders) transpired in the early Eocene, and examples of most modern families are encountered by the middle to late Oligocene ( Hartenberger, 1996, 1998; Wood, 1959). Based on the earliest fossils so far discovered and their cladistically basal stature within Glires or Rodentia , "Asia" (a vast and diverse region to be sure) is presently accepted as the area of origin for the Order ( Beard, 1998; Bryant and McKenna, 1995; Hartenberger, 1996; Meng and Wyss, 2001).

Rodent suborders?—The now classical and familiar rodent suborders— Sciuromorpha , Myomorpha , and Hystricomorpha —issue from Brandt (1855), who eponomously based his names on Waterhouse’s (1839) characterizations of sciuromorphous, myomorphous, and hystricomorphous zygomasseteric morphologies. Thereafter, the question of rodent suborders has become a century and a half long work-in-progress. The most important conceptual landmark since Brandt is Tullberg’s (1899) Eine Phylogenetische Studie, in which he integrated mandibular conformation (sciurognathy versus hystricognathy) and Brandtian zygomasseteric criteria to delineate two major rodent groups (each called a Tribus), the Sciurognathi (Sciuromorphi and Myomorphi) and Hystricognathi (Bathyergomorphi and Hystricomorphi). Subsequently, most mammalian classifiers and rodent systematists have adopted either the Brantian tri-subordinal ( Alston, 1876; Lavocat, 1978; Miller and Kellogg, 1955; Simpson, 1945; Wilson, 1949; Wood, 1965) or Tullbergian dual subordinal themes ( Chaline and Mein, 1979; Ellerman, 1940; Landry, 1999; Lavocat, 1973; Patterson and Wood, 1982; Woods, 1972), some in more or less their orthodox forms and others with due amendment of lesser ranks and reallocation of taxa among them. Other variations have recognized more primary divisions within Rodentia , whether or not called a suborder per se, anywhere from five to 16 ( Hartenberger, 1985, 1998; McKenna and Bell, 1997; Miller and Gidley, 1918; Thaler, 1966; Thomas, 1897 c; Weber, 1904; Wood, 1955, 1959). Within each of these classifications, problematic and-or poorly understood groups have been regularly acknowledged through the time-honored qualifiers of ‘ incertae sedis ’ or ‘suborder indeterminate,’ some to an extent that little sense of higher-level relationship within the Order is conveyed (e.g., Carleton, 1984). See Carleton (1984), Ellerman (1940), Hartenberger (1985), Landry (1999), Simpson (1945), Wilson (1949), and Wood (1955, 1959, 1965) for additional review and commentary on rodent classifications, in particular the historical treatment of rodent suborders.

Compared with the vacillation and disagreement over suborders, the number of families of living rodents considered valid has remained fairly stable over the past half-century, around 30 to 35 ( Anderson and Jones, 1967, 1984; Corbet and Hill, 1986; Grassé and Dekeyser, 1955; Hartenberger, 1985; Honacki et al., 1982; Simpson, 1945; Wilson and Reeder, 1993). Variation among these, and other, works pivots on the authors’ decisions concerning the rank of certain groups as subfamily versus family. Thirty-three families are acknowledged herein (Table 1), and recent molecular investigations have generally sustained the monophyletic status of most of these, where the taxonomic sampling is suitably robust and critically focussed (see various family-group Comments). Moreover, certain kinship hypotheses linking these families recur in the proliferation of gene-sequencing investigations over the past decade. The stability of these genetic clades, superposed upon the morphological evidence mustered and classifications promulgated over the past 150 years, lends some empirical confidence that agreement about rodent suborders will be eventually realized, perhaps in less than a decade. Five family-group phyletic associations are briefly reviewed below as a basis for the provisional subordinal groupings observed here.

(1). Aplodontidae-Sciuridae + Gliridae : Aplodontids and sciurids form the core of Sciuromorpha , as reflected in classifications generated from the late 1800s to the present (e.g., Alston, 1876; Hartenberg, 1998; Landry, 1999; McKenna and Bell, 1997; Simpson, 1945; Thomas, 1897 c; Tullberg, 1899). Their common evolutionary origin has been historically postulated ( Wilson, 1949), and this phyletic link has been reinforced by a suite of morphological traits ( Landry, 1999; Luckett and Hartenberger, 1985 b; Tullberg, 1896, 1899), in particular craniodental features ( Vianey-Liaud, 1985), cranial foramina ( Wahlert, 1985), middle ear anatomy ( Lavocat and Parent, 1985; Meng, 1990), and fetal membrane development (Luckett, 1985). That Aplodontidae and Sciuridae are each others closest relative (among living rodents) is also overwhelmingly indicated by albumin immunology ( Sarich, 1985) and by phylogenetic comparisons of several mitochondrial and nuclear genes ( Adkins et al., 2001, 2003; DeBry, 2003; DeBry and Sagel, 2001; Huchon et al., 1999, 2000, 2002; Montgelard et al., 2002; Nedbal et al., 1996; Waddell and Shelley, 2003).

Association of Gliridae with aplodontids and sciurids is not novel, but their union in an expanded clade (Sciuroidea or Sciuromorpha ) has recently garnered additional confidence. Most classifications and phylogenetic reconstructions have allied dormice with Myomorpha because the zygomasseteric structure of extant glirids (excluding graphiurines) was characterized as myomorphous ( Alston, 1876; Chaline and Mein, 1979; Ellerman, 1940; Landry, 1999; Simpson; 1945; Thomas, 1897 c; Wahlert et al., 1993; Wilson, 1949; Wood, 1965). However, the "myomorphy" common to most living dormice is now recognized as convergent to the zygomasseteric configuration in true Myomorpha (Dipodidae-Muroidea): i.e., "pseudomyomorphy" as per Maier et al. (2002) and Vianey-Liaud (1985) or "pseudosciuromorphy" as per Landry (1999).

Exclusion of dormice from Myomorpha is further bolstered by certain morphological traits, especially middle ear anatomy ( Lavocat and Parent, 1985; Meng, 1990) and the cephalic arterial pattern ( Bugge, 1971 a, 1985), and notably by the rich fossil record that traces glirids back to the early Eocene ( Hartenberger, 1971, 1994, 1998; Vianey-Liaud, 1974, 1985, 1989, 1994; Vianey-Liaud et al., 1994; Vianey-Liaud and Jaeger, 1996). These inquiries identify glirids as sciuromorph rodents, not myomorphs. Sequence analyses based on both mitochondrial and nuclear genes similarly represent dormice as sister-group to Sciuridae or to a sciurid-aplodontid clade ( Adkins et al., 2001, 2003; Bentz and Montgelard, 1999; DeBry and Sagel, 2001; Huchon et al., 1999, 2002; Kramerov, 1999; Kramerov and Vassetzky, 2001; Kramerov et al., 1999; Lin et al., 2002; Montgelard et al., 2001, 2002; Murphy et al., 2001 a; Nedbal et al., 1996; Reyes et al., 1998). See Gliridae for expansion.

(2). Castoridae + Heteromyidae-Geomyidae: Castorids have been traditionally allied with groups placed in Sciuromorpha or Sciurognathi, whether in a Brantian or Tullbergian scheme ( Alston, 1876; Chaline and Mein, 1979; Ellerman, 1940; Hartenberger, 1985, 1998; Landry, 1999; McKenna and Bell, 1997; Miller and Gidley, 1918; Miller and Kellogg, 1955; Simpson, 1945; Thomas, 1897 c; Wilson, 1949). However, their sciuromorphous zygomasseteric anatomy, sciurognathus mandible, primitive cranial morphology but highly derived dentition, and deep and rich fossil history beginning in the late Eocene of North America ( Korth, 2001) have spawned other views that conflict with the traditional arrangement. Some have divorced beavers from any close phyletic link with sciurids (Schaub, 1953; Wood, 1955), or frustratingly viewed their affinities as intractable ( Bugge, 1985:347 —"one of the most isolated groups, near to no other Recent family"; Lavocat and Parent, 1985; Wood, 1959, 1965). Meng (1990:27) also emphasized the uncertain kinship of castorids and instead speculated that they are "derived from an ancestral stock giving rise to muroids than [sic] to sciurids."Origin of Geomyoidea (Heteromyidae-Geomyidae) is documented by fossils from the North American Eocene ( McKenna and Bell, 1997), and the subsequent evolutionary diversification of the group has unfolded in the New World. Although Thaler (1966) isolated geomyoids as a separate suborder, the Geomorpha, their sciuromorphous and sciurognathus crania have influenced many researchers to retain them within Sciuromorpha (e. g., Ellerman, 1940; Hartenberger, 1998; Miller and Gidley, 1918; Simpson, 1945), and others to suggest, based on different morphological traits, a closer phylogenetic link with myomorph rodents ( Alston, 1876; Hill, 1937; Landry, 1999; McKenna and Bell, 1997; Wahlert, 1985; Wilson, 1949; Wood, 1955). Nevertheless, affiliation with myomorphs, cautioned Fahlbusch (1985:627), is not strongly supported by morphology, and as "long as zygomasseteric structure is not replaced by a more reliable feature in classification, there is scarcely any reason to transfer the Geomyoidea from Sciuromorpha to Myomorpha ." Recent phylogenetic analyses of mitochondrial and nuclear gene sequences regularly disclose a castorid-geomyoid clade ( Adkins et al., 2001, 2003; DeBry, 2003; Huchon et al., 2002; Montgelard et al., 2002; Murphy et al., 2001 a; Waddell and Shelley, 2003). These results, while surprising in view of the usual subordinal separation of castorids ( Sciuromorpha ) and geomyoids ( Myomorpha ), mirror the evolutionary tree developed by Tullberg (1899:481), who more than a century ago depicted the common ancestry of "Castoroidei" (fossil and extant beavers) and "Geomyoidei" ( Geomys , Heteromys , Perognathus , Dipodomys ) within his Sciuromorphi.

Molecular data generally support inclusion of the castorid-geomyoid clade in a larger group that variously contains muroids, dipodoids, anomalurids, and Pedetes (e. g., Adkins et al., 2003; DeBry, 2003; Eizirik et al., 2001; Huchon et al., 2002; Montgelard et al., 2002). Integrity of such an assemblage requires further testing, as urged especially by Adkins et al. (2003). They considered this large clade to be poorly resolved based on their sampled genes and suggested that "the base of the rodent phylogeny may be resolved by a much larger amount of sequence data … or a unique class of molecular characters." Further paleontological study may also clarify the reputed alliance between the castorid-geomyoid and dipodoid-muroid (Myodonta) clusters. For example, North American Eocene Sciuravidae has been suggested as ancestral to both geomyoids ( Vianey-Liaud, 1985; Wilson, 1949) and myodonts (see Dipodidae for discussion and references), and Walton (1993) identified a basic occlusal pattern shared by some Eocene Eutypomyidae (extinct sister-group to Castoridae — Korth, 2001; Wahlert, 1977), Eocene Eomyidae (extinct sister-family to geomyoids—Falbusch, 1985; Wahlert, 1985), and early sciuravids.

(3). Dipodidae + Muroidea ( Platacanthomyidae through Muridae ): Close phylogenetic affinity between Dipodoidea and Muroidea has long been recognized, notably as developed by the insightful arguments of Tullberg (1899), Wilson (1949), and Klingener (1964). Schaub (1958) formalized their cognate relationship as Myodonta, used as a suborder or infraorder, and a very substantial body of recent paleontological, morphological, and molecular results also ratifies this phyletic union (see Dipodidae and Muroidea for references).

The clade Dipodidae-Muroidea corresponds to the nucleus of Myomorpha in the usually accepted sense, and their joint kinship with Geomyoidea (Heteromyidae-Geomyidae), and sometimes with Gliridae , has been commonly affirmed by classifications advanced in the middle to late 1900s (e.g., Carleton, 1984; Chaline and Mein, 1979; McKenna and Bell, 1997; Simpson, 1945; Wood, 1955, 1965). The purported relationship to Gliridae finds no support in recent phylogenetic studies, but the evolutionary connection to Geomyoidea receives moderate corroboration, particularly in certain molecular results (see above commentary under 1 and 2).

(4). Anomaluridae + Pedetidae : Along with ctenodactylids, the hystricomorphous and sciurognathus anomalurids and pedetids have been often treated as incertae sedis, whether either one or both families, in past classifications (e.g., Carleton, 1984; Chaline and Mein, 1979; Hartenberger, 1998; Simpson, 1945; Wood, 1965). Although recent molecular studies have credited Simpson (1945) for perception of their cognate affinity (e.g., Montgelard et al., 2002), the idea properly dates at least to Winge (1887) and Tullberg (1899), both of whom provided characters and arranged the two groups in the same family or superfamily. This phyletic association has earned recent support based on arterial patterns ( Bugge, 1974, 1985; George, 1981), middle ear anatomy ( Lavocat and Parent, 1985; Meng, 1990), and mitochondrial genes ( Montgelard et al., 2001, 2002); McKenna and Bell (1997) acknowledged their possible sister-group affinity as the Suborder Anomaluromorpha . Not all investigators agree and some specifically dispute any special relationship between Anomaluridae and Pedetidae (e.g., Jaeger, 1988 b; Landry, 1999), especially given fundamental microstructural differences in their incisor enamel ( Martin, 1993, 1995).

Evidence for close relationship to any of the other four suprafamilial groups is similarly mixed. In molecular studies that have sampled only Anomalurus or only Pedetes , the represented family usually emerges basally in a large clade encompassing Castoridae-Geomyoidea and-or Dipodidae-Muroidea ( Adkins et al., 2001, 2003; Debry, 2003; Huchon et al., 2002; Nedbal et al., 1996); in those that have sampled both, using two mitochondrial genes, Anomaluridae-Pedetidae is depicted as sister group to Sciuromorpha-Hystricognathi ( Montgelard et al., 2001, 2002). The possible link with Dipodidae-Muroidea recalls the older and broader notion of "Myomorphi" ( Ellerman, 1940; Tullberg, 1899). On the other hand, some morphological interpretations have implicated early genealogical ties with Ctenodactylidae-Hystricognathi ( Bryant and McKenna, 1995; Meng, 1990); George (1985), Landry (1999), Martin (1993, 1995), and Thomas (1896) cautiously endorsed this kinship hypothesis for Pedetidae but not Anomaluridae . This alternative association recalls the formative views of " Hystricomorpha " before the character weighting accorded to mandibular morphology, a feature emphasized by Tullberg (1899).

Although researchers are again probing into the relationships of Anomaluridae and Pedetidae , resolution of their phylogenetic placement has improved little since the summary of Luckett and Hartenberger (1985 b). Two interrelated and systematically critical questions remain at issue: do anomalurids and pedetids constitute a natural clade and where does this putative sister-group cladistically fit among the other major rodent associations listed here, in particular Myomorpha versus Hystricomorpha ?

(5). Ctenodactylidae + Hystricognathi ( Bathyergidae through Heptaxodontidae ): The African gundis, Ctenodactylidae , represent another group whose phylogenetic placement has perplexed systematists: early placed in Hystricomorpha (Thomas, 1897 c); arranged with Sciurognathi after Tullberg (1899) stressed the cardinal importance of sciurognathy and hystricognathy ( Chaline et al., 1977; Ellerman, 1940); isolated (among living families) in its own suborder ( Hartenberger, 1998; McKenna and Bell, 1997); or left as infraorder or suborder incertae sedis (Carleton, 1984; Simpson, 1945; Wood, 1955, 1965). Of these prior interpretations of relationship, or concessions to indeterminate relationship, a substantive and jointly impressive body of information supplies support for Thomas’ association of ctenodactylids with Hystricomorpha sensu lato (He actually arranged gundis as a subfamily of Octodontidae , a family assignment contraindicated by this same literature). Such corroborative studies broadly draw upon morphology, both individual character systems and multicharacter cladistic analyses ( Bryant and McKenna, 1995; Bugge, 1971 b, 1985; Flynn et al., 1986; George, 1985; Jaeger, 1988 b; Landry, 1957, 1999; Lavocat and Parent, 1985; Luckett, 1985; Luckett and Hartenberger, 1985 b; Martin, 1993, 1994 a; Meng, 1990), paleontology ( Flynn et al., 1986; Hussain et al., 1978; Marivaux et al., 2002; Marivaux and Welcomme, 2003), and mitochondrial and nuclear DNA sequences ( Adkins et al., 2001, 2003; Huchon and Douzery, 2001; Huchon et al., 2000, 2002; Montgelard et al., 2002). Certain of these analyses posit southern Asia as the continental area of origin of the common ancestor of Ctenodactylidae-Hystricognathi ( Huchon and Douzery, 2001; Marivaux et al., 2002). Landry (1999) introduced the name Entodacrya (reflecting the internal course of the nasolacrimal duct) to identify the Ctenodactylidae-Hystricognathi clade, and Huchon et al. (2000) soon after coined Ctenohystrica, a contraction whose taxonomic meaning is self evident.

The story of the hystricognathus and hystricomorphous rodents is a suspenseful and exciting read in systematic mammalogy. Ever since Tullberg (1899) defined the Hystricognathi , the explicit union of New and Old World groups, respectively neo- and paleotropical in principal distribution, has sparked lively debate over issues of monophyly and polyphyly, character homology and homoplasy, biogeography and areas of origin, evolutionary and phylogenetic criteria for classification. It would be pretentious to encapsulate the details and nuanced conclusions of these exchanges in a brief foreword (see summaries in Huchon and Douzery, 2001, Jaeger, 1988 b, Landry, 1999, and Martin, 1993, as well as remarks and references under Hystricognathi ). Important bookmarks in the long dialogue include Wood (1950, 1965, 1974, 1980), Landry (1957), Wood and Patterson (1959), Lavocat (1969, 1971, 1974, 1980), Hoffstetter and Lavocat (1970; Hoffstetter, 1975), Woods (1972, 1982), Patterson and Wood (1982), and many thoughtful papers in Luckett and Hartenberger (1985 a). As summarized by Luckett and Hartenberger (1985 b:691), a wealth and heterogeneity of data provide "overwhelming support for hystricognath monophyly, and all indicate that it is virtually impossible to distinguish between Old and New World hystricognaths as separate taxa." As fitting denouement to their assessment, ensuing hypothesis-testing using a variety of mitochondrial and nuclear genes has only reinforced the phylogenetic perspective for monophyly of Hystricognathi sensu Tullberg ( Adkins et al., 2001, 2003; Catzeflis et al., 1995; Huchon and Douzery, 2001; Huchon et al., 1999, 2000, 2002; Murphy et al., 2001 a; Nedbal et al., 1994, 1996).

Provisional subordinal classification.—Although this systematic compendium is aimed at the species, a higher level classification is an obvious, and natural, means to collate and convey such alpha-level information. With regard to suborders, we provisionally use the names Sciuromorpha , Castorimorpha , Myomorpha , Anomaluromorpha , and Hystricomorpha , respectively, to designate suborders for the five clades reviewed above (Table 1). Based on our impressionistic consensus, that research indicates that Sciuromorpha , Myomopha, and Hystricomorpha , as arranged, receive good to excellent support as monophyletic taxa. The evidence for Castorimorpha and Anomaluromorpha in this regard is overall less persuasive. Future phylogenetic investigation of these groups must address both the question of their naturalness and the possibility that they belong within one of the other three suborders. Ranks between the suborder and family were avoided, except where already assigned authorship at certain higher levels (namely the Superfamily Muroidea and Infraorder Hystricognathi ) required coordinate groupings for consistency and balance of classificatory arrangement. Proper exposition of the nomenclatural and taxonomic background and of phylogenetic and biogeographic issues that bear on these intermediate ranks (superfamily to infraorder) is beyond the scope of a species checklist (and would certainly overtax the patience of our editors, whose kind indulgence to add this chapter we much appreciate). See McKenna and Bell (1997) for one classificatory view of those ranks, their past usage, and allocation of synonyms.

We retain Brandt’s (1855) names for three of the five suborders ( Sciuromorpha , Myomorpha , and Hystricomorpha ), whereas some foremost rodent authorities have pointedly advised against continued employment of these Brantian epithets, especially at the subordinal rank (e.g., Landry, 1999:283; Wood, 1985). They have argued that the descriptive meaning of the terms does not strictly concord with the morphologies of included members, and that this contradiction will only continue to engender confusion. That the logical connotation of a taxon’s name should uniformly correspond to its taxonomic intention strikes us as puzzling. Not all members of the Superorder Afrotheria live in Africa, nor are all species of the Order Carnivora carnivorous. The species- to family-group ranks are rife with valid names whose logical meaning, as intended by the original descriptor, is partially inconsistent or actually misleading in terms of the currently accepted contents of the taxon, whether in strict morphological accuracy, indication of distribution, or implication of phylogenetic alliance. In reviewing 150 years of rodent classifications, we are impressed that few of our predecessors applied the features of infraorbital configuration and jaw shape in an overridingly typological touchstone of all rodent classification. Most made the distinction between, e.g., Myomorpha in its taxic sense and myomorphy as a morphological condition and routinely consulted additional traits. Thus, the hystricomorphous Dipodidae have been nearly always included within Myomorpha (in retrospect, correctly it would seem), and the sciuromorphous Geomyoidea have been considered by many to also fit within Myomorpha (in retrospect, perhaps incorrectly). And while Brandt’s taxonomic names qua morphological descriptors may not perfectly correspond to all members, they nonetheless do conform very well. The core family members of Sciuromorpha , Myomorpha , and Hystricomorpha have remained largely the same, and most, not all, do exhibit those fundamental zygomasseteric structures. More importantly, these core assemblages have survived recent scrutiny using ever more explicit principles for systematic classification and sophisticated methods for inferring phylogenetic relationship and assessing descent from a common ancestor.

Brandt’s names, intended as a suprafamilial rank, do not merit any special consideration as far as the International Code of Zoological Nomenclature (Int. Comm. Zool. Nom., 1999) is concerned. Aside from requirements of formal publication and availability, the Code does not prescribe standard suffices to denote relative rank or stipulate strict adherence to priority for taxa recognized above the family-group. Such novel names for recently apprehended clades are beginning to appear within Rodentia (e.g., Bryant and McKenna, 1995; Huchon et al., 2000; Landry, 1999; McKenna and Bell, 1997). Criteria such as familiarity, historical usage, and nomenclatural priority, however, persuaded us to retain Brandt’s terms, particularly when recent evidence for monophyly has not substantively altered their earlier taxonomic meaning. Perhaps an argument could be developed for a new subordinal name to contain the apparent clade Castoridae-Geomyoidea; however, the case for common ancestry is still tenuous, and in the meanwhile, we conventionally applied nomenclatural priority in adopting Castorimorpha Wood, 1955, over Geomorpha Thaler, 1966. For other viewpoints on naming, or the need to rename, order-level taxa, see Archibald (2003:356-357), Bryant and McKenna (1995:32-34), and McKenna and Bell (1997:513-514).