StenocephalemysDivision. Historical taxonomic use of Mastomys as a genus or subgenus of either Rattus or Praomys summarized by Meester et al. (1986), Chevret et al. (1994), and Granjon et al. (1997b). Morphological (Lecompte et al., 2002a; Van der Straeten, 1979; Van der Straeten and Robbins, 1997), protein electrophoretic analyses (Iskandar and Bonhomme, 1984), DNA/DNA hybridization experiments (Chevret et al., 1994), analyses of complete mtDNA cytochrome b sequences (Lecompte et al., 2002b), and chromosomal data (Britton-Davidian et al., 1995; Granjon et al., 1997b) support union of most of the species listed below in a monophyletic group different from those clades containing species of Praomys, Myomyscus, and Hylomyscus. Mastomys pernanus is the exception (see that account). Important to studies of specific diversity, biogeography, population and community ecology, and medicine, species of Mastomys have been examined in several contexts (see reviews by Granjon et al., 1997b; Keogh and Price, 1981; Leirs, 1992; Skinner and Smithers, 1990; and Taylor, 2000b). Until the late 1990s, definition of species by chromosomal and biochemical traits from some regions (Duplantier et al., 1990a; Green et al., 1980; Hubert et al., 1983) proceeded faster than definitions based on morphology, resulting in a new view of species-diversity in the genus, but also an ignorance of morphological limits of those species and their real geographic distributions. Recently, however, integration of chromosomal, molecular, and morphometric data has allowed sharper resolution of species limits, research summarized by Granjon et al. (1997b), who also presented the most current and best review of Mastomys systematics. Except for adding new data and various comments, we follow their conception of species contents.

Monophyly of M. erythroleucus, M. natalensis, M. huberti, and M. coucha is supported by cladistic analyses of chromosomal traits (Britton-Davidian et al., 1995; Volobouev et al., 2002b) and variation in skeletal, dental, and soft tissue characters (Lecompte et al., 2002a). Analyses of complete mtDNA cytochrome b sequences support monophyly of the same four species and adds M. kollmannspergeri (reported as verheyeni) to that cluster (Lecompte et al., 2002b). Lavrenchenko et al. (1992) provided a review of chromosomal and biochemical polymorphism (including hemoglobin electrophoresis) in samples of Mastomys and the significance of these data in distinguishing species and other topics. Pattern of karyotype evolution among several Mastomys, its correlation with phylogenetic patterns indicated by analyses of gene sequences, and relevance to cladogenesis in the genus discussed by Volobouev et al. (2002b). Spermatozoal morphology of selected species documented by Breed (1995a) and Baskevich and Lavrenchenko (1995). Analysis of microcomplement fixation of albumin grouped Mastomys with Otomys, Myomys (=Myomyscus), Hylomyscus, and Praomys (Watts and Baverstock, 1995a).

Robbins and Van der Straeten (1989) and Van der Straeten and Robbins (1997) studied holotypes that are the basis of most names associated with the Mastomys-Myomys (= Myomyscus) Hylomyscus-Praomys group, and identified the taxa they definitely regard as representing Mastomys but unfortunately did not allocate any of them to species. One of these names is Mus colonus Brants (1827), which was made the type species of Myomys by Thomas (1915). Because some workers claimed colonus was unidentifiable and therefore Myomys is invalid, Myomyscus was proposed by Shortridge (1942) to replace Myomys. The problem and historical opinions have been thoroughly reviewed (Roberts, 1951; Ellerman et al., 1953; Rosevear, 1969; Van der Straeten and Verheyen, 1978b; Meester et al., 1986). Recently, Van der Straeten and Robbins (1997) located the specimen (an adult male) described by Brants, which represents the holotype of colonus, and to them the skull and teeth "shows all Mastomys characters" (p. 153); they did not assign the holotype to a definite species of Mastomys. Brants (1827) did not indicate the type locality, but Van der Straeten and Robbins reported that "Zondagsrivier" is written on the specimen label and that place is listed as "Uitenhage, Port Elizabeth (33E46'S, 25E24'E)" in the gazetteer they consulted. This is far eastward of the geographic range of Myomyscus verreauxii, very close to if not within the western margin of the distribution of M. coucha, but not within the range of M. natalensis (as mapped by Skinner and Smithers, 1990, and de Graaff, 1997q; but see Taylor, 2000b). Musser studied the specimen set aside by Van der Straeten in Berlin and agrees with the identification by Van der Straeten and Robbins. It is not related to Myomyscus verreauxii as Roberts (1951) indicated (he treated it as a species of Myomyscus, but did not have the benefit of examining the holotype), but is a Mastomys.

Identifying the holotype of colonus as a Mastomys affects the taxonomy of Mastomys and Myomys in three ways. First, Myomys has to be arranged as a synonym of Mastomys (both were named by Thomas and preliminarily diagnosed on the same page with Myomys on the line above Mastomys). Second, Myomys must be replaced by Myomyscus to cover the species formerly assigned to it. Finally, colonus may be the earliest name for what is now called M. coucha. Here we allocate Myomys to the synonymy of Mastomys but await future study to determine the specific identity of the holotype of colonus. That could be accomplished by including the holotype in a multivariate discriminant function analysis among Mastomys samples from southern Africa, and by extracting a piece of dry tissue, bone or tooth from it for analysis by mitochondrial or nuclear DNA sequencing.

Despite accumulation of systematic information through the last decade, the species of Mastomys still require careful taxonomic revision; geographic ranges have to be resolved for most species, and associating the many scientific names proposed with currently defined species remains a vexing problem. Living species of Mastomys occur primarily in subsaharan Africa (isolated population in WC Morocco is the North African exception) and the genus is represented in late Pliocene, early Pleistocene, and Holocene deposits of East and South Africa (Avery, 1998, 2000; Denys, 1987a, 1999; Jaeger, 1976; Wesselman. 1984), but at least three once occurred in the S Levant (Israel). Mastomys batei, described from middle and Late Pleistocene (Bate, 1942b; Tchernov, 1968), is absent from sediments dated later than 80,000-70,000 years before present (Tchernov, 1986, 1992, 1994, and references cited therein). Fossils identified as Mastomys galilensis and M. levantinus were recovered from the early Pleistocene ‘Ubeidiya Formation in the Central Jordan Valley of Israel (Tchernov, 1986).