RattusDivision. Rattus Frisch, 1775, is unavailable. Corbet and Hill (1992:334) noted that Fischerís original spelling is Ruttus but ". . . there is no evidence within this publication that the spelling Ruttus was an error. However it has, very sensibly, been universally accepted as an error for Rattus and it would serve no useful purpose to revert to Ruttus." Sody (1941) proposed the genera Christomys, Cironomys, Geromys, Mollicomys, Octomys, and Pullomys for various species we list in Rattus; Corbet and Hill (1992) doubted the validity of Mollicomys and Octomys since they are based on mammary formulae shared with other murines (but see generic account of Taeromys). Togomys is based on R. exulans (Dieterlen, in Ansell, 1989). Taxonomic changes altering the definition of Rattus as understood by systematists working in the middle 1900s (Chasen, 1940; Tate, 1936; Ellerman, 1941, 1949a, 1961; Ellerman and Morrison-Scott, 1951; Laurie and Hill, 1954; Simpson, 1945) have been documented and summarized by several specialists (Misonne, 1969; Musser, 1981b; Musser and Newcomb, 1983; Musser and Holden, 1991; Musser and Heaney, 1992). Species in many Asian and African genera recognized today were once embraced by the broad definition of Rattus (e. g., Chasen, 1940; Ellerman, 1941, 1949a; Ellerman et al., 1953): Cremnomys, Millardia, Madromys, Berylmys, Sundamys, Kadarsanomys, Komodomys, Maxomys, Leopoldamys, Niviventer, Srilankamys, Margaretamys, Bunomys, Paruromys, Taeromys, Apomys, Bullimus, Limnomys, Tarsomys, Aethomys, Micaelamys, Praomys, Mastomys, Hylomyscus, and Myomyscus. A few of these genera are phylogenetically allied to species currently arrayed in Rattus (members of Rattus Division; see Table 1), but others represent very distant lineages having separate origins from the early ancestral murine stock.
Even divorced of species in these genera, Rattus remains a heterogeneous accumulation of species. The wide range of morphological variation present disallows an objective generic diagnosis and indicates the present contents to consist of several monophyletic clusters that may or may not prove to be embraced by a single genus. This polyphyletic pattern is reflected in our sorting of species into the following six species groups, and a seventh assemblage containing unaffiliated species for want of adequate information or study.
1. Rattus norvegicus species group. Rattus norvegicus, the type species of the genus, is traditionally placed in subgenus Rattus with R. rattus and its close allies (e. g., Ellerman, 1941, Ellerman, 1949a), but the two diverge significantly in morphology (Musserís studies), allozymic variation, albumin immunology, and L1 (LINE-1) retrotransposons and amplification events, and mtDNA cytochrome b sequences (Aplin, in litt., 2004; Chan, 1977; Chan et al., 1979; Baverstock et al., 1983a, c, 1986; Verneau et al., 1997, 1998; Watts, in litt.). Rattus nitidus and R. pyctoris also belong here as indicated by molecular data and some morphological traits (see those accounts).
2. Rattus exulans species group. Another species usually included within subgenus Rattus along with R. argentiventer, R. nitidus, R. norvegicus, R. rattus, and others that have, in the past, formed the core of that group (e.g., Ellerman, 1941; Misonne, 1969). Some electrophoretic and chromosomal data have supported this allocation (Chan, 1977; Chan et al., 1979; Raman and Sharma, 1977), but a larger body of morphological, other biochemical data, mtDNA cytochrome b sequences, and L1 (LINE-2) retrotransposons and amplification events indicate that R. exulans is phylogenetically distant from species in our R. rattus group as well as from those in our R. norvegicus group (K. Aplin, in litt., 2004; Gemmeke and Niethammer, 1984; Medway and Yong, 1976; Pasteur et al., 1982; Verneau et al., 1997, 1998). Based on analyses of mtDNA cytochrome b sequences, R. exulans is as phylogenetically distant from the R. rattus group as are the R. leucopus (New Guinea-Moluccan lineage) and R. fuscipes (Australian lineage) assemblages from the R. rattus cluster (K. Aplin, in litt., 2004).
3. Rattus rattus species group. Species usually regarded to represent the core of species related to R. rattus (R. adustus, R. andamanensis, R. argentiventer, R. baluensis, R. blangorum, R. burrus, R. hoffmanni, R. koopmani, R. losea, R. lugens, R. mindorensis, R. mollicomulus, R. osgoodi, R. palmarum, R. rattus, R. satarae, R. stoicus, R. simalurensis, R. tanezumi, R. tawitawiensis, and R. tiomanicus; Musser and Holden, 1991). Whether or not this cluster, R. norvegicus, and R. exulans, form the monophyletic contents of Rattus proper is a possibility that has to be assessed by systematic revision and phylogenetic study (Musser and Heaney, 1985, 1992; Musser and Holden, 1991). Comparative chromosomal data for many of these species summarized by Musser and Holden (1991) and Rickart and Musser (1993), additional chromosomal descriptions for Vietnam samples reported by Bulatova et al. (1992), and silver-stained karyotypes of Chinese species compared by Chen et al. (1992). Schwarz and Schwarz (1967) offered an idiosyncratic revision of the group.
4. Rattus fuscipes species group. Native Australian species (R. colletti, R. fuscipes, R. lutreolus, R. sordidus, R. tunneyi, and R. villosissimus), revised by Taylor and Horner (1973) and reviewed by Watts and Aslin (1981); one (R. sordidus) also occurs on New Guinea. Allozymic variation and chromosomal studies indicate that the species form a monophyletic cluster to the exclusion of either R. rattus or R. norvegicus (Baverstock et al., 1977d, 1983a, 1986), but "the extent of biochemical divergence between the Australian Rattus on the one hand, and either R. rattus or R. norvegicus on the other is no greater than that between R. rattus and R. norvegicus, thus supporting the notion that the Australian forms do indeed belong to the genus Rattus" (Baverstock et al., 1986:29). Baverstock et al. (1983a, 1986) proposed a phylogenetic hypothesis based on cladistic analyses of electrophoretic, immunologic, and chromosomal data, and another proposed set of relationships assessed by isozyme electrophoresis (Baverstock et al., 1986). Seddon and Baverstock (2000) compared evolutionary history of the second exon of RT1.Ba and intron b with the hypothesis suggested by Baverstock et al. (1986). Mahoney and Richardson (1988) cataloged taxonomic, distributional, and biological references.
Preliminary phylogenetic analyses of mtDNA cytochrome b sequences from exemplars in the R. fuscipes, R. leucopus, and R. xanthurus groups indicate that the R. fuscipes cluster (Australian endemics) is sister-group to the R. leucopus assemblage (New Guinea and Moluccan endemics), and that R. xanthurus and its allies (Sulawesi endemics) is sister to those two assemblages (K. Aplin et al., 2004). In Aplinís view, members of the R. fuscipes and R. leucopus groups each may represent ". . . discrete radiations albeit closely related and with some exchange across the Torres Strait." Should some supraspecific nomenclatural designation be warranted, Sodyís 1941) Geromys is available for the Australian lineage, and Thomasís (1910e) Stenomys for the New Guinea-Mollucan radiation.
5. Rattus leucopus species group. Species indigenous to New Guinea and adjacent archipelagos (R. arfakiensis, R. arrogans, R. giluwensis, R. jobiensis, R. leucopus, R. mordax, R. niobe, R. novaeguineae, R. omichlodes, R. praetor, R. richardsoni, R. steini, R. vandeuseni, and R. verecundus); one (R. leucopus) also occurs in NE Australia. Most have been subjects of systematic revisions (Taylor et al., 1982, 1983). Known endemics of the Moluccas (R. elaphinus, R. feliceus, and two undescribed Moluccan species [Musser and Holden, ms]) belong in this group. Included here also are species Musser and Carleton (1993) and Flannery (1995a) placed in Stenomys. All former Stenomys are members of a Rattus clade as determined by albumin immunology (Watts and Baverstock, 1994b, 1995b, 1996; R. niobe was the only species sampled), skin and skull morphology (specimens in AMNH and BMNH), and spermatozoal traits (Breed and Aplin, 1994). Multivariate analyses of morphometric traits brought together R. niobe, R. verecundus, and R. richardsoni in a cluster separate from other native species of New Guinea Rattus (Taylor et al., 1982), but phylogenetic analyses of mtDNA cytochrome b sequences indicate R. verecundus to be very closely related to R. leucopus (K. Aplin, in litt., 2004) and Aplin (2004) has suggested that Stenomys should be subsumed again within Rattus. RŁmmler (1938), on morphological grounds, appreciated the alliance between leucopus and verecundus and treated them as species of Stenomys, along with R. niobe and the Seramese Nesoromys ceramicus. Flannery (1990b, 1995a, b) provided photographs of animals along with distributional and biological summaries.
6. Rattus xanthurus species group. Species native to Sulawesi, adjacent Pulau Peleng, and Pulau Sangir in Kepulauan Sangir off tip of NE peninsula of Sulawesi (R. bontanus, R. marmosurus, R. pelurus, R. salocco, R. xanthurus, and an undescribed species from Kepulauan Sangir, off tip of NE Sulawesi). Musser and Holden (1991) speculated this monophyletic cluster may not belong in Rattus, but preliminary phylogenetic analyses of mtDNA cytochrome b sequences indicates the Sulawesi endemics are sister to the R. leucopus (New Guinea and Moluccas) and R. fuscipes (Australia) groups, and all three fall within the bounds of Rattus if such distinctive clusters as the R. rattus and R. exulans groups are included (K. Aplin, in litt., 2004).
7. Rattus species group unresolved. Phylogenetic affinities of R. annandalei, R. enganus, R. everetti, R. hainaldi, R. hoogerwerfi, R. korinchi, R. macleari, R. montanus, R. morotaiensis, R. nativitatis, R. ranjiniae, R. sanila, and R. timorensis are unresolved; some will eventually be excluded from the genus (see individual species accounts).
Reports presenting nonmorphological information pertinent to understanding evolutionary relationships among species of Rattus include chromosomal data (Baverstock et al., 1883a; Capanna and Corti, 1991; Gadi and Sharma, 1983; Milyutin, 1990; Raman and Sharma, 1977; Rickart and Musser, 1993; Yosida, 1980), allozymic variation (Baverstock et al., 1986; Watts and Baverstock, 1994b), and L1 (LINE-1) retrotransposons and amplification events (Verneau et al., 1997, 1998); additional studies are referenced in the species accounts. Mahoney and Richardson (1988) summarized taxonomic, distibutional, and biological references for the Australian species.
Descriptions of nine undescribed species of Rattus are being prepared by various researchers. One is from Pulau Bisa (Bisa Rat) in Kepulauan Obi of the Moluccas that is represented only by a skull (Flannery, 1995b). Another was collected on Mt Kemenagi in the Kikori River Basin of S Papua and the Mt Karimui area (Leary and Seri, 1997). Two additional new species were found in archaeological deposits on Morotai Isl in the N Moluccas (Flannery et al., 1998; Helgen, 2003b), a fifth (possibly not referable to Rattus) has been found on Timor by K. Helgen (see account of R. timorensis), two occur on Pulau Taliabu in the Sula Arch. (Musser, 1981c; Musser and Holden, ms), one is known from upland habitat near Loei in N Thailand (Aplin, in litt., 2004), and the last, a chromosomally and morphologically distinct member (a close relative of R. villosissimus) of the R. fuscipes group, is found in C Queensland, Australia (Aplin, in litt., 2004).
Evolutionary history of Rattus extends to late Pliocene in Asia. Two species, identified only as Rattus sp., are represented by fossils recovered from middle to late Pleistocene cave sediments in the Sichuan-Guizhou region (Zheng, 1993) and Guangxi Province (Chen et al., 2002) of S China. In Thailand, the extant R. andamanensis (reported as sikkimensis) has been documented back to early Pleistocene, the extinct R. jaegeri is represented by fragments from late Pliocene and early Pleistocene cave sediments, and specimens identified as Rattus sp. come from middle Pleistocene strata (Chaimanee, 1998). See also Chaimanee and Jaeger (2000a) for a look at diversity of Thai Rattus during the Plio-Pleistocene. The extinct R. trinilensis and at least five other species of Rattus have been recovered from Pleistocene sediments in C and E Java and represent the earliest records of the genus on the Sunda Shelf (already present by 1.6 million years ago; Musser, 1982d; Van der Meulen and Musser, 1999). In the Hondo region of Japan, R. norvegicus is represented by Holocene and late Pleistocene fossils, and a form identified as R. aff. norvegicus from the middle Pleistocene that may have been ancestral to R. norvegicus in the Japanese Isls (Kowalski and Hasegawa, 1976; Kawamura, 1989). The extinct R. haasi comes from late Acheulian (middle Pleistocene) cave sediments in Israel (Tchernov, 1968, 1996). Rattus casimcensis and R. dobrogicus are represented by fragments from the late Pleistocene of Romania (Kowalski, 2001).
Some fossils originally identified as Rattus are examples of other genera. The m1 reported as "cf. Rattus" from the Chinese Pliocene of Shaanxi (Jacobs and Li, 1982) is an example of extinct Chinese Saidomys (Cai and Qiu, 1993). Rattus sp., also based upon an m1, was recorded from Plio-Pleistocene Pinjor beds in the Siwaliks of NW India (Gaur, 1986), but the cusp pattern resembles those in Cremnomys and Millardia rather than Rattus (Musserís observations). Finally, late Pleistocene isolated molars from N Pakistan Siwaliks identified as "cf. Rattus" by Jacobs (1978) are examples of Hadromys loujacobsi (Musser, 1987b). No fossils of true Rattus have yet been recovered from Plio-Pleistocene Siwalik strata in either N Pakistan or NW India (Musser, 1987b; Patnaik, 2000, 2001).