rodentia: muridae - Oxford Journals - Oxford University Press

Journal of Mammalogy, 85(5):911–923, 2004

CRANIAL VARIATION AND GEOGRAPHIC PATTERNS WITHIN THE DASYMYS RUFULUS COMPLEX (RODENTIA: MURIDAE) S. K. MULLIN,* N. PILLAY,

AND

P. J. TAYLOR

Ecophysiological Studies Research Group, School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, WITS 2050, South Africa (SKM, NP) Durban Natural Science Museum, Mammal Department, P.O. Box 4085, Durban 4000, South Africa (PJT)

This study examines inter-operational taxonomic unit (OTU) relationships within a morphologically defined species complex of the African water rat Dasymys rufulus that comprises West, North, Central, and East African populations. Based on both traditional skull measurements and skull coordinates, 5 possible cryptic species within the rufulus complex were identified: western West Africa (Senegal to western-southeastern Ivory Coast), eastern West Africa (eastern-southeastern Ivory Coast to Nigeria, Chad), Central Africa (Cameroon, Congo, Democratic Republic of Congo), Central and East Africa (northern Angola, Zambia, southern Malawi, western Mozambique, Sudan) and southern Central and East Africa (southern Tanzania, northern Malawi, Mozambique). Data also confirmed that 2 recently identified species, D. foxi (restricted to Jos Plateau, Nigeria) and D. longipilosus (restricted to Mt Cameroon, Cameroon) are morphologically distinct. Each of the 3 datasets provided similar results, but the ventral skull coordinates were the most useful in separating OTUs with respect to a broad geographic pattern. Key words:

Africa, Dasymys rufulus, Muridae, Rodentia, skull morphology

Although the African water rat genus Dasymys is widespread throughout sub-Saharan Africa, it has a discontinuous distribution, with some populations considered relict (Davis 1962; De Graaff 1981; DuPlantier et al. 1997; Gordon 1991; Lavrenchenko et al. 2000). The latest International Union for the Conservation of Nature and Natural Resources Red Data List (Hilton-Taylor 2000) states that Dasymys has a Data Deficient status because so little information has been collected for this species, whereas Dasymys from South Africa was classified as Near Threatened (Y. Friedmann et al., in litt.). Fossil evidence indicates that members of this taxon are less numerous today than in the Pleistocene (Avery 1991, 1998; De Graaff 1961; Denys 1999; DuPlantier et al. 1997; Kingdon 1971; Lavocat 1956; Misonne 1969); this might be because of its low vagility, low numbers in nature, low reproductive potential, and habitat restriction (Avery 1991; Crawford-Cabral 1998; Gordon 1991; Hanney 1965; Happold 1987; Meester 1988; Pillay 2003; Rosevear 1969). Dasymys is found in several different biome types, but appears to prefer wetter areas

such as vleis (or swamps), areas with lush grasses and ephemeral river systems (Carleton and Martinez 1991; Rautenbach 1982; Rosevear 1969; Taylor 1998). It seems plausible that Dasymys disperses along riverbanks, making historical river patterns and connections relevant to any study of this genus. As little is known about Dasymys, this rodent group has been receiving much attention from researchers recently, particularly with regard to chromosomal (Volobouev et al. 2000) and morphometric studies (Carleton and Martinez 1991; CrawfordCabral and Pacheco 1989; Mullin et al. 2002). A recent study of Dasymys (Mullin et al., in press) suggested the existence of 11 distinct morphological species in sub-Saharan Africa, including a D. rufulus species complex. The latest classification of D. rufulus (Musser and Carleton 1993) listed this species as occurring throughout West Africa, encompassing 11 countries: Senegal, Guinea, Sierra Leone, Liberia, southern Mali, Ivory Coast, Burkina Faso, Ghana, Togo, Benin, and Nigeria. However, data presented by Mullin et al. (in press) indicated that the rufulus complex as described included material from a larger geographic area that ranges from Senegal, east to southern Sudan, south to northern Angola (excluding material from Kenya, Uganda, Burundi, Rwanda, and eastern Democratic Republic of Congo), and east to southern Tanzania. The lack of clear taxonomic groups warranted further research of this taxon.

* Correspondent: [email protected]

2004 American Society of Mammalogists www.mammalogy.org 911

912

JOURNAL OF MAMMALOGY

A preliminary morphological study indicated the presence of 5 subgroups within the rufulus complex. These subgroups included western West Africa (Senegal to western-southeastern Ivory Coast); eastern West Africa (eastern-southeastern Ivory Coast to Nigeria, Chad); Central Africa (Cameroon, Congo, Democratic Republic of Congo); Central and East Africa (northern Angola, Zambia, southern Malawi, western Mozambique, Sudan); and southern Central and East Africa (southern Tanzania, northern Malawi, Mozambique). Therefore, the primary aim of this study was to identify possible morphological species or subspecies within the large geographic area occupied by the D. rufulus complex, as outlined in Mullin et al. (in press). A secondary aim was to examine the relationship between D. rufulus and D. foxi, a species restricted to the Jos Plateau in Nigeria, because both were once considered to be subspecies of the nominate species D. incomtus (described from South Africa). Carleton and Martinez (1991) concluded on the basis of morphometric studies that D. rufulus and D. foxi represent 2 separate species. However, Mullin et al. (in press) suggest that this taxonomic grouping is still inconclusive, and found some evidence that D. foxi might represent a subspecies within D. rufulus. A 3rd aim was to consider current and historical biogeographical records in conjunction with the dataset.

MATERIALS AND METHODS Five hundred and thirty-six specimens from sub-Saharan Africa were examined in this study (Appendix I). Morphological data were analysed hierarchically, starting at an individual level and then pooled by localities and finally operational taxonomic units (OTUs; i.e., pooled localities to increase sample sizes). Throughout the text, the names of the OTUs that represent more than 1 locality are referred to by the 1st locality name listed for that OTU in Appendix I instead of listing all of the localities associated with each OTU each time that particular OTU is mentioned. At each stage, subsets were screened for outliers and tested for normality. Data were examined at a locality level only once it was determined that each locality was represented by the same phenon (defined as a group of morphologically similar individuals). This was done in order to ascertain interlocality relationship on a more manageable scale and to decide whether or not neighboring localities could be pooled for an inter-OTU analysis. In addition to examining D. rufulus specimens from West Africa (sensu Musser and Carleton 1993), we included material from Cameroon, Congo, Central African Republic, Democratic Republic of Congo, Angola, Zambia, Malawi, Mozambique, and Tanzania, since Mullin et al. (in press) suggested that these were part of the rufulus complex based on morphological characteristics. Four geographic appellations were used to generally describe specimens: West Africa (comprising material from Senegal to Nigeria), North Africa (Chad, Sudan), Central Africa (Cameroon, Central African Republic, Congo, Democratic Republic of Congo, and northern Angola) and East and Central Africa (Zambia, Malawi, Mozambique, and Tanzania). Material representing D. foxi was also included in this study (for reasons explained in the introduction) as was D. longipilosus, from Mt Cameroon, Cameroon, primarily to confirm its morphological distinctness from other Cameroon material. The definition of OTU was made with reference to vegetation, geology, soil, altitude, and phytogeographic maps. Localities that were represented by less than 3 individuals were pooled into OTUs, while

Vol. 85, No. 5

respecting geographical barriers such as mountain ranges and rivers, which delineated geographically close localities found in the same biome type (after Taylor and Meester 1993; Appendix I). If a sample from a given locality consisted of less than 3 individuals and it was not possible to combine that locality into an OTU, data were treated separately. In these cases, canonical scores of individuals were projected onto the scores of the OTUs to produce a comprehensive plot. At each step of the analysis, the 3 morphological datasets (traditional skull measurements and geometric morphometric data collected from both the dorsal and ventral skull) produced similar results. Given the repetitive nature of presenting 3 diagrams for each step, we chose the plots that presented data most clearly. Results not presented here may be obtained from the 1st author. Traditional morphometrics.—Four hundred and eighty-six adult individuals were used. The character suite consisted of the following 13 cranial characters (see Mullin et al. 2002): greatest skull length; crown length of upper (maxillary) toothrow; hard palate width; nasal width; zygomatic arch width; internal diameter of zygomatic arch; least breadth of interorbital constriction; foramen magnum width; least width of zygomatic arch; zygomatic plate width; greatest mandible length; mandibular tooth row; greatest height of skull. A previous study of nongeographic variation revealed that specimens from age classes III, IV, V, and VI (defined by the amount of wear on the teeth) represent adult D. incomtus individuals and that sexual dimorphism was negligible in D. incomtus populations (Mullin et al. 2001), therefore data for males and females were pooled in the present analyses. The dataset was normally distributed, nonkurtotic, nonskewed, and homoscedastic. Standardized data were subjected to a principal component analysis and a multigroup canonical variates analysis, which tested multivariate differences among OTUs. Both the principal component analysis and canonical variates analysis subroutines were used for each subgroup to obviate bias by including only 1 approach. Because the different approaches provided similar results, we chose scatterplots with the most clearcut patterns for the results section for ease of presentation. Statistical analyses were performed with the NTSYS-pc program (Rohlf 1996). A multivariate analysis of variance (MANOVA) and a Student Neuman-Keuls multiple range test completed on OTUs to determine significant differences in cranial measurements was performed with the Statistica (2001) software package (StatSoft, Tulsa, Oklahoma). Geometric Morphometrics.—A 5-mm Leica DC100 camera attached directly to a computer was used to capture images of the dorsal and ventral skull. Each digital file was saved in JPEG format and was analyzed using the specific software programs created for geometric morphometric data (i.e., the TPS series of programs and GRF-nd) available from F. J. Rohlf at the website http://life.bio.sunysb.edu/ morph. In total, 484 dorsal and 366 ventral views were collected from specimens representing the localities described in Appendix I. Landmarks (Appendix II; Fig. 1) were placed on an image using the software program tpsDig (ver. 1.31, Department of Ecology and Evolution, State University of New York, Stony Brook, New York). The software program TpsSmall (ver. 1.17, Department of Ecology and Evolution, State University of New York) confirmed that the correlation between tangent and shape space was high enough to carry out subsequent analyses. Data were then scaled, aligned and transformed using the Procrustes procedure in the software program TpsRelw (ver. 1.24, Department of Ecology and Evolution, State University of New York). The centroid size for each specimen was extracted and subjected to a one-way ANOVA to test for differences between centroid sizes for each group. The software program GRF-nd (Slice 1999) was used to extract eigenvectors and landmark residuals

October 2004

MULLIN ET AL.—CRANIAL VARIATION WITHIN THE DASMYS RUFULUS COMPLEX

913

and to examine data for outliers. The program TpsRelw was also used to compute a partial weights matrix (W) based on the consensus configuration for the data, as well as to complete a relative warps analysis, which included the affine components, U1 and U2. Dorsal and ventral weight matrices were subjected to a canonical variates analysis, principal component analysis (based on among-OTU correlations), and cluster analysis (based on Mahalanobis distances) using the software program NTSYS-pc (Rohlf 1996). The software program TpsRegr (ver. 1.25, Department of Ecology and Evolution, State University of New York) was used to regress the 1st and 2nd canonical variates onto both the dorsal and ventral weight matrices (in order to examine the thin-plate splines associated with each axis) and to regress the centroid size on to both the dorsal and ventral data.

RESULTS Although the one-way MANOVA showed an overall significant inter-OTU variation (k ¼ 0.002, d.f. ¼ 806, 4,766, P , 0.001), the Student Neuman-Keuls multiple range test provided few, if any, patterns or significant differences (P , 0.05) between OTUs. F values ranged from 1.59 (least breadth of interorbital constriction) to 5.81 (greatest height of skull), indicating a strong similarity between specimens. Greatest height of skull and greatest mandible length (5.72) had the highest F values, indicating that these characters best separated the OTUs, which was confirmed in the principal component analysis results (below). Least breadth of interorbital constriction and internal diameter of zygomatic arch (1.90) had the lowest F values, suggesting that these characters had similar values (sizes) in the dataset. With regard to the multivariate analyses, each of the 3 datasets (traditional measurements, dorsal, and ventral skull coordinates) provided similar results, although the dorsal skull data consistently produced the weakest patterns. This suggested that the dorsal skull shape of all of the specimens examined was similar. Similarly, principal component analysis plots showed loosely defined groupings that were more strongly delimited in the canonical variates analysis plots and therefore the plot of the principal component axes is not provided here. However, the principal component analysis loadings from the analysis of the traditional measurements indicated 4 important characters—greatest skull length (0.916), greatest mandible length (0.890), zygomatic arch width (0.871), and greatest height of skull (0.748) on the 1st axis (which comprised 42.74% of the variation) that distinguished between OTUs (Table 1). Foramen magnum width (0.791) had the highest value on the 2nd axis (accounting for 11.13% of the variation), which was positively correlated with least width of zygomatic arch (0.440). The 3rd axis, explaining 10.95% of the variation, was positively correlated with least breadth of interorbital constriction (0.476) and negatively correlated with nasal width (0.675). Only greatest skull length, greatest mandible length, zygomatic arch width, and greatest height of skull had high percent variance contributions to both the 2nd and 3rd axes confirming their importance. Results of the scatterplot of the first 2 variates in the canonical variates analysis were similar between the 3 datasets analyzed so only 2 of the scatterplots (traditional morphometric

FIG. 1.—Diagram of the landmark positions used in this study on the a) dorsal and b) ventral Dasymys skull. The landmarks are defined in Appendix II.

measurements and ventral skull coordinates) are presented (Figs. 2 and 3, respectively). Fig. 3, based on the ventral skull coordinate data provided the clearest results, accentuating patterns seen in Fig. 2, which was included primarily to emphasize the separation seen between D. longipilosus and material from the D. rufulus complex on the 2nd variate. Several common patterns were seen in the canonical variates analysis plots for all 3 datasets (Figs. 2 and 3). The 1st was that the OTU representing D. foxi was separate from other OTUs representing West Africa. A 2nd pattern was that West African individuals seemed to be split into 2 separate groupings that corresponded with geographic regions. Group 1 comprised material from Senegal, Guinea, Sierra Leone, Liberia, and Ivory Coast and Group 2 contained material from Ivory Coast, Ghana, Burkina Faso, Benin, Togo, and Nigeria. In all 3 datasets, a separation within Ivory Coast was seen with individuals from Kong, Lamto, Sienso, and Tiegbe (Ivory Coast) more similar to material from Group 2. Material from Malawi (the Zomba and Mulanje Plateaus), Mozambique (Zambue), and Zambia appeared to be morphologically separate from West African individuals. The 3 remaining Mozambican OTUs (OTU 62-Beira, 59-Furancungo and 60Vila Coutinho) in addition to 1 Malawian OTU (56-Kasungu) and the single Tanzanian OTU (39-Mlali) were at times more similar to West African material (in particular Group 2) and at other times appeared to form a separate group. Thin-plate splines, depicting skull deformations, are shown along the 1st and 2nd variates of the canonical variate analysis (Fig. 4). Splines 1L and 1R correspond to the positive and negative values (respectively) along the 1st variate, while splines 2L and 2R represent deformations from negative and positive values (respectively) from the 2nd variate (Fig. 4). The splines along the 1st variate indicated differences in skull shape

914

Vol. 85, No. 5


TABLE 1.—Principal component loadings of variables for components 1, 2, and 3. The percent variance contribution is provided in parentheses. Character Greatest skull length Crown length of upper (maxillary) toothrow Hard palate width Nasal width Zygomatic arch width Internal diameter of zygomatic arch Least breadth of interorbital constriction Foramen magnum width Least width of zygomatic arch Zygomatic plate width Greatest mandible length Mandibular tooth row Greatest height of skull Variance explained

1 0.916 0.670 0.583 0.429 0.871 0.686 0.402 0.085 0.567 0.682 0.899 0.439 0.748

(83.82%) (44.96%) (33.99%) (18.41%) (75.94%) (47.05%) (16.16%) (0.73%) (32.13%) (46.51%) (80.74%) (19.27%) (55.89%)

42.74%

between individuals from Zambia, Malawi, and Mozambique and West African material. Spline 1L was characterized by a shorter palatal foramen, narrower nasal width and zygomatic arch width, a smaller skull length, narrower upper tooth width, and less flared bullae, with the opposite observed in Spline 1R. Splines from the 2nd axis separated material from Angola, Tanzania, and Mozambique from the remaining material, which was similar on the second axis. Spline 2L was characterized by flared bullae, a longer skull, shorter palatal

2 0.129 0.012 0.167 0.040 0.024 0.162 0.264 0.791 0.440 0.371 0.064 0.547 0.209

(85.48%) (44.97%) (36.77%) (18.57%) (75.99%) (49.67%) (23.12%) (63.30%) (51.48%) (60.29%) (81.16%) (49.16%) (60.25%)

11.13%

3 0.003 0.420 0.413 0.675 0.066 0.390 0.476 0.158 0.143 0.167 0.100 0.380 0.106

(85.48%) (62.57%) (53.79%) (64.17%) (76.43%) (64.86%) (45.75%) (65.81%) (53.53%) (63.07%) (82.15%) (63.63%) (61.38%)

10.95%

foramen, and wider upper molar tooth width, with the opposite observed in Spline 2R. Individual scores from the canonical variates analysis (from localities with ,3 specimens) were projected onto the canonical scores of the entire dataset, so that individuals were matched to OTUs with similar skull structures (Table 2). In general, specimens retained geographic order and usually fell into 1 of the 5 broad groups outlined by canonical variates analysis scatterplots, although there were several individuals

FIG. 2.—Canonical variates analysis plot based on traditional linear skull measurements for Dasymys. n ¼ West Africa Group 1 (Senegal, Sierra Leone, Guinea, Liberia and eastern-southeastern Ivory Coast); h ¼ West Africa Group 2 (northwestern-northeastern-central Ivory Coast, Ghana, Burkina Faso, Benin, Togo, and Nigeria); ¼ Cameroon, Central African Republic, Congo, and Democratic Republic of Congo; * ¼ D. foxi (Panyam, Nigeria); ¼ Zambia; m ¼ Malawi; n ¼ Mozambique; ¤ ¼ Tanzania; ) ¼ Sudan; v ¼ Angola; þ ¼ D. longipilosus (Mount Cameroon, Cameroon). The operational taxonomic units are demarcated on the figure and are described in Appendix I.

October 2004


915

FIG. 3.—Canonical variates analysis plot based on ventral skull coordinates for Dasymys. Symbols as in Fig. 2. The operational taxonomic units are demarcated on the figure and are described in Appendix I.

that appeared to be exceptions (e.g., L, U, and V; Table 2). These anomalies were most likely because data for individuals, rather than OTUs, were analyzed. Material from Mali (A) and Sierra Leone (B and C) grouped with the West African Group 1, whereas material from Ivory Coast (D), Nigeria (E and F), Chad (G), and northwest Sudan (I and J) were more similar to Group 2. Three of the 5 Democratic Republic of Congo individuals (K, M, O) were either more similar to other Democratic Republic of Congo or Cameroon OTUs, whereas L was closer to a Liberian OTU and N to a Malawian OTU. Results indicated an East and Central African subgroup with close links to specimens from southern Sudan (Table 2). Of the 5 Tanzanian specimens examined, 2 individuals (P and Q) were more similar to material from Sudan, 2 were grouped with a Malawian OTU (R and S), and 1 (P) with an OTU from Mozambique. The 2 Malawian specimens (Y) were both most similar to material from Mozambique. One of the Sudanese individuals (H) was similar to Zambian material. Material from Zambia was either similar to OTUs from Zambia (V and X) or Malawi (W), with 1 specimen (X) more similar to an OTU from Sudan. Two of the 3 Angolan individuals examined (T) appeared to be part of the Zambian group with the exception of an individual from Mombolo (U) which was more similar to West African material. Finally, variation within the West African group (see Figs. 2 and 3) was examined in more detail with a 2nd series of canonical variates analyses. A 3-dimensional canonical variates analysis plot (Fig. 5), based on traditional measurements,

showed a distinct split between individuals from Senegal, Guinea, and Liberia (Group 1) and material from Ghana, Burkina Faso, Benin, Togo, and Nigeria (Group 2), consistent with the pattern observed in Figs. 2 and 3. Within the Ivory Coast, there was again a separation between localities, with Dasymys specimens from OTUs 10-Bolo, 12-IDERT, 13Adiopodoume´, 14-Dabou, and 15-Mopoye´me within Group 1 and material from OTUs 7-Sienso, 8-Tiebge, 9-Kong, and 11Lamto more similar to Group 2.

DISCUSSION With the aid of either morphometric or genetic data, it has been common in recent years to revoke the monotypic status of many African rodents, as originally presented by Ellerman et al. (1941) and Misonne (1969), in favor of the original descriptions of what were once polytypic genera. Dasymys is no exception to this generalization, and is certainly represented by more than 1 species throughout its extensive range (Carleton and Martinez 1991; Crawford-Cabral 1998; Mullin et al., in press; Musser and Carleton 1993). The patterns seen here in the rufulus complex were weak, implying that any morphological groups outlined most likely represent either subspecies or cryptic species. Although not uncommon for mammals with a wide distribution to retain a monotypic status (Grubb 1999), it seems more plausible that Dasymys OTUs within the rufulus complex are not homogenous. This statement is based on its habitat requirements, the high likelihood of population

916

Vol. 85, No. 5


TABLE 2.—Projection of individuals onto operational taxonomic units (OTUs) using the canonical variates module in the program NTSYS (Rohlf 1996). DRC ¼ Democratic Republic of Congo. Designation in Appendix I A

Specimen Kangaba, Mali Kangaba, Mali

B

C

D

E

FIG. 4.—Thin-plate splines from the first and second variate of Fig. 3. 1L ¼ left side of variate 1, 1R ¼ right side of variate 1; 2L ¼ left side of variate 2; 2R ¼ right side of variate 2.

fragmentation (Avery 1991; Gordon 1991), the large geographic area considered here, and the numerous geographical barriers within this range. Morphological data alone will not resolve taxonomic questions in the Dasymys genus, but will be able to provide indications of affinities as well as important areas where future genetic studies should focus. D. foxi and D. longipilosus both appeared to represent valid species that were separate from D. rufulus. Although D. foxi was not distinct on any of the axes in any of the scatterplots, it was consistently more similar to material from the Zambian group than to other West African OTUs. The clear morphological separation documented between D. foxi and material that Musser and Carleton (1993) recognize as D. rufulus (i.e., from Senegal to Nigeria) in the canonical variates analysis plots corroborates the findings of Carleton and Martinez (1991), as well as the conclusions reached by Mullin et al. (in press).

Rokupr, Sierra Leone Rokupr, Sierra Leone Sandaru, Sierra Leone Sandaru, Sierra Leone Bouake´, Ivory Coast Bouake´, Ivory Coast Gudi, Nigeria Gudi, Nigeria

F

Ashaka, Nigeria

G

Bekao, Chad Bekao, Chad

H I J K

Raga, Sudan Jebel, Sudan Mt Baginzi, Sudan M’boko, DRC

L

Brazzaville, DRC

M N

Likouala, DRC Lukolela, DRC

O P

Ngombe, DRC Usuhilo, Tanzania Usuhilo, Tanzania

Q R S T

Dabaga, Tanzania Igale, Tanzania Illolo, Tanzania Dundo, Angola Dundo, Angola

U V

Mombolo, Angola Kasungu, Zambia Kasungu, Zambia

W

Abercorn, Zambia

X

Hot Springs, Zambia Hot Springs, Zambia Nyika Plateau, Malawi Nyika Plateau, Malawi

Y

Similar to OTU

Remarks

Banfora, Burkina Faso Lake Retba, Senegal Dabou, Ivory Coast Adiopodoume´, Ivory Coast Bolo, Ivory Coast Kasinga, Zambia Kafue River, Zambia Tiegbe, Ivory Coast Pulima, Ghana Furancungo, Mozambique Manyinga, Zambia Ibadan, Nigeria Kong, Ivory Coast Chibale, Zambia Padori, Togo Mt Coffee, Liberia Lubumbashi, DRC Mt Nimba, Liberia Kinshasha, DRC Mulanje Plateau, Malawi Mie´ri, Cameroon Torit, Sudan Vila Coutinho, Mozambique Kagelu, Sudan Mulanje, Malawi Mulanje, Malawi Limalunga, Zambia Limalunga, Zambia Padori, Togo Mt Nimba, Liberia Ft Jameson, Zambia Zomba Plateau, Malawi Ft Jameson, Zambia Torit, Sudan Furancungo, Mozambique Vila Coutinho, Mozambique

Type D. i. shawi

Type D. i. edsoni Type D. i. bentleyae

Type D. i. alleni

October 2004


Dasymys longipilosus from Mt Cameroon was certainly separate from the rufulus material on the 2nd variate in the canonical variates analysis plot based on traditional measurements, emphasizing the morphological differences between the 2 species. Several endemic rodent species have previously been identified on Mt Cameroon (Hutterer et al. 1992; Petter 1982; Taylor and Kumirai 2001), including D. longipilosus (Eisentraut 1963). The remaining data suggest the existence of 5 possible groups in the material examined, corresponding to both current and historical biogeographical patterns (e.g., paleovegetation patterns, changes in river flow, or the presence of geographic barriers—Denys 1999; Grubb 1999; Happold 1996), that are known to have influenced the distributions of several taxa, in addition to rodents. The 1st group comprised western West African localities (Senegal, Guinea, Sierra Leone, Mali, Liberia, and the localities Adiopodoume´, Bolo, Dabou, IDERT, and Mopoye´me, all from southeast Ivory Coast). Group 2 represented eastern West African localities (Ghana, Burkina Faso, Benin, Togo, Nigeria, and the localities Sienso, Lamto, Kong, and Tiebge from northern and northeastern Ivory Coast). The 3rd group encompassed Central Africa (Cameroon, Congo, Democratic Republic of Congo, and possibly northern Angola), whereas the 4th was restricted to Zambia, southern Malawi (Zomba and Mulanje Plateaus), and western Mozambique (Zambue), again possibly including northern Angola. Finally, the 5th group was confined to Tanzania, northern Malawi (Kasungu), and both western and eastern Mozambique (Beira, Furancungo, Vila Coutinho, Vila Vasco de Gama). Sudanese material appeared to have a stronger affinity with the Tanzania group, perhaps corresponding with a link between the Sudanian and Zambezian vegetation zones present in the Pleistocene (Denys 1999) or perhaps to a historical vegetation connection between the Juba (Sudan) and Zambezi (Zambia) Rivers (Grubb et al. 1999). Species fragmentation within West Africa has long been attributed to the Dahomey Gap (Robbins 1978), a savanna belt in Ghana and Togo that is located between the rainforests found elsewhere in West Africa. The Volta and Niger Rivers, which delineate the western and eastern borders of the Gap, also might have limited mammal distributions rather than the Gap itself (Dobigny and Volobouev 2000; Happold 1996; Robbins 1978). In several other rodent genera, new West African species have been named, dividing what was once considered a continuous distribution throughout the area into smaller, and more localized, ranges (Capanna et al. 1996; Dobigny and Volobouev 2000; DuCroz et al. 1997; Granjon et al. 1997; Hubert et al. 1983; Lecompte et al. 2001; Robbins and Van der Straeten 1989; Swanepoel and Schlitter 1978; Van der Straeten and Dieterlen 1987; Volobouev et al. 2002). Ivory Coast fauna has been a focal point for research since several endemic rodent species (particularly at Lamto) have been identified in this species-rich area (Corti et al 2000; Happold 1996; Swanepoel and Schlitter 1978; Van der Straeten and Verheyen 1978a, 1978b). The behavior of the Ivory Coast OTU data was interesting and suggests that the skull morphology of Dasymys from this region is diverse. Five of

917

FIG. 5.—Canonical variates analysis based on traditional measurements from Dasymys skulls from West African specimens encompassing both Groups 1 (Senegal, Sierra Leone, Guinea, Liberia, and eastern-southeast Ivory Coast) and 2 (northwestern-northeasterncentral Ivory Coast, Ghana, Burkina Faso, Benin, Togo, and Nigeria).

the 10 OTUs (Adiopodoume´, Bolo, Dabou, IDERT, and Mopoye´me; from eastern and southeastern Ivory Coast) were more similar to the western West Africa group (Group 1), whereas the remaining 5 OTUs (Bouake´, Kong, Lamto, Sienso, and Teigbe; from central, northwestern, and northeastern Ivory Coast) grouped with the eastern West Africa group (Group 2). Individuals from Lamto were somewhat of an anomaly, compared to material from the 4 other proximate localities: Adiopodoume´, Dabou, IDERT, and Mopoye´me. Instead of grouping with Group 1 with 4 other localities, Lamto was more similar to material from Group 2. The 3rd Dasymys group delineated here appears to be confined to Central Africa (i.e., Cameroon, Congo, and the eastern and northern Democratic Republic of Congo), although all of the specimens do share strong cranial similarities with West African individuals. A link between Cameroon and Nigeria has long been documented in rodents (Dieterlen and Van der Straeten 1992; Grubb 1999; Happold 1996; Hutterer et al. 1992; Volobouev et al. 2002) in spite of the Sanaga River, found in eastern Nigeria, which might have helped cause disjunct distributions (Happold 1996). The lowland forests of Cameroon, Congo, the Central African Republic, and the Democratic Republic of Congo were apparently refuges in the Pleistocene and consequently acted as centers of endemism for mammals, as a consequence of climate fluctuations, and water level changes in the Congo River basin helped to isolate populations of both fish and mammals (Grubb 1999; Turpie and Crowe 1994). The similarity between Angolan, Zambian, and East African material coincides with zoogeographical zones in Angola (Feiler 1990) and the Zambezian savanna zone outlined by Denys (1999). Hall (1960) postulated that the Nigerian-Central African link continued as far south as northern Angola via a one-time continuous forest system, but this was not corroborated in this study because Angolan material appeared to be more similar to specimens from Zambia than from Central African. Chitau, Duque de Braganca, and Dundo are found in

918

Vol. 85, No. 5


northern Angola in an area where forest refuges for mammals have been identified, and which also marks the southern limit of the Congo basin, potentially limiting species distributions (Happold 1996). Mombolo is found in the coastal escarpment zone that is close to the Congo basin. The Congo basin has obviously been isolated for a substantial period of time, as evidenced by the high number of endemic fish species there that indicate that at 1 time there was a link between the Congo Basin and Zambia as well as the Nile River system (Hamilton 1982). Feiler (1990) indicated that a connection between Angolan and East and Central Africa fauna was possible via the Katanga-Zambezi Plateaus, which extend between Zambia and central Angola and is connected to the Angolan Plateau by the Angolan escarpment. The species richness of the Zambia (Burda 2001; Filippucci et al. 1994, 1997; Kawalika et al. 2001; Van der Straeten 1980) have been attributed to the effects of a possible isolation on the Katanga-Zambezi Plateau as well as the presence of the barriers such as the Muchinga Escarpment or the Luangwa and Zambezi Rivers (Burda 2001; Grubb et al. 1999). Zambian OTUs were linked with 1 of the Mozambican OTUs from the Tete Corridor (Zambue) and 2 OTUs from southern Malawi (from the Mulanje and Zomba Plateaux). The 3rd Malawian OTU examined in this study, from the Nyika Plateau, was more similar to material from southern Tanzania and the remaining OTUs from Mozambique, including material from the coast (Beira) and the Tete Corridor (Furancungo, Vila Coutinho, and Vila Vasco de Gama). Happold and Happold (1989) postulated that a grassland corridor, which allowed migration, once connected Zambia, Malawi, and southern Tanzania, a distribution pattern that apparently occurs in Acomys (Barome et al. 2001). Northern Mozambique was once inundated with flood plains, which under the correct conditions could have allowed migration between Mozambique and Tanzania (Axelrod and Raven 1978; Bowmaker et al. 1978; Gaigher and McPott 1973; Stuart et al. 1993; van Zinderen Bakker 1978; Werger 1978). The taxonomic relationships within Dasymys are proving difficult to diagnose. Patterns produced by morphometric data are complex, and gaps in the data collection somewhat restrain attempts to resolve taxonomic issues, not only in the entire genus but more specifically in the rufulus complex. Additional material should be collected and analyzed from Sudan, Cameroon, Congo, the Democratic Republic of Congo, Angola, Tanzania, Malawi, and Mozambique, all of which are countries that are traditionally associated with limited specimen collections in museums (Schlitter and Delany 1985). Despite the fact that this study produced somewhat ambiguous results, this information increases our knowledge about the Dasymys genus and indicates that the latest classification for Dasymys provided by Musser and Carleton (1993) needs to be updated.

ACKNOWLEDGMENTS We thank the following institutions and people for loaning us material and for allowing us to examine the collections: Amathole Museum, King Williamstown (FK); American Museum of Natural

History, New York (BR); Bulawayo Museum, Bulawayo (WC, AK); Carnegie Museum of Natural History, Pittsburgh (SM); Field Museum of Natural History, Chicago (BS, BP); Museé Naturalle d’Histoire, Paris (JC); Museé Royale d’Afrique Centrale, Antewerp (WNV); Museum of Comparative Zoology, Harvard (MR, TM); Natural History Museum, London (PJ); Staatliches Museum, Stuttgart (FD); Transvaal Museum, Northern Flagship Institution, Pretoria (CC, DM); United States Smithsonian Museum, Washington, D.C. (MC, HK); South African Museum, Cape Town (DD). This study was funded by the University of the Witwatersrand (NP, SKM), and the National Research Foundation (NP, PJT). ASM guidelines were followed with respect to all animal care and use related to this study.

LITERATURE CITED AVERY, D. M. 1991. Late Quaternary incidence of some micromammalian species in Natal. Durban Museum Novitates 16:1–11. AVERY, D. M. 1998. An assessment of the Lower Pleistocene micromammalian fauna from Swartkrans Members 1–3, Gauteng, South Africa. GEOBIOS 31:393–414. AXELROD, D. I., AND P. H. RAVEN. 1978. Late Cretaceous and Tertiary vegetation history of Africa. Pp. 77–130 in Biogeography and ecology of southern Africa (M. J. A. Werger, ed.). Dr W. Junk bv Publishers, The Hague, The Netherlands. BAROME, P. O., ET AL. 2001. Phylogeny of Acomys spinosissimus (Rodentia, Muridae) from north Malawi and Tanzania: evidence from morphological and molecular evidence. Biological Journal of the Linnean Society 73:321–340. BOWMAKER, A. P., P. B. N. JACKSON, AND R. A. JUBB. 1978. Freshwater fishes. Pp. 1181–1230 in Biogeography and ecology of southern Africa (M. J. A. Werger, ed.). Dr W. Junk bv Publishers, The Hague, The Netherlands. BURDA, H. 2000. Determinants of the distribution and radiation of African mole-rats (Bathyergidae, Rodentia). Ecology or geography? P. 31 in Small African mammals (C. Denys, L. Granjon, and A. Poulet, eds.). Colloques et Se´minaires—Editions de l’ IRD, Paris, France. CAPANNA, E., ET AL. 1996. A multidisciplinary approach to the systematics of the genus Arvicanthis Lesson, 1842 (Rodentia, Murinae). Mammalia 60:677–696. CARLETON, M. D., AND C. MARTINEZ. 1991. Morphometric differentiation among west African populations of the rodent genus Dasymys (Muroidea: Murinae), and its taxonomic implications. Proceedings of the Biological Society of Washington 104:419–435. CORTI, M., C. DI GIULIOMARIA, AND W. VERHEYEN. 2000. Geometric morphometrics of three new species of Lophuromys. P. 49 in Small African mammals (C. Denys, L. Granjon, and A. Poulet, eds.). Colloques et Se´minaires—Editions de l’ IRD, Paris, France. CRAWFORD-CABRAL, J. 1998. The Angolan rodents of the superfamily Muroidea: an account of their distribution. Ministaêrio da Cia˜encia e da Tecnologia, Instituto de Investigaðcaäo Cientaîfica Tropical, Luanda, Angola. CRAWFORD-CABRAL, J., AND A. PACHECO. 1989. A craniometrical study on some water rats of the genus Dasymys (Mammalia, Rodentia, Muridae). Garcia de Orta, Serie Zoologique 15:11–24. DAVIS, D. H. S. 1962. Distribution patterns of southern African Muridae, with notes on some of their fossil antecedents. Annals of the Cape Provincial Museum 2:56–76. DE GRAAFF, G. 1961. On the fossil mammalian microfauna collected at Kromdraai by Draper in 1895. South African Journal of Science 57:259–260. DE GRAAFF, G. 1981. The rodents of South Africa. Butterworth and Company, Durban, South Africa.

October 2004


DENYS, C. 1999. Of mice and men. Pp. 226–252 in African biogeography, climate change, and human evolution (G. T. Bromage and F. Schrenk, eds.). Oxford University Press, Oxford, United Kingdom. DIETERLEN, F., AND E. VAN DER STRAETEN. 1992. Species of the genus Otomys from Cameroon and Nigeria and their relationship to East African forms. Bonner Zoologische Beitra¨ge, 43:383–392. DOBIGNY, G., AND V. VOLOBOUEV. 2000. Comparative cytogenetics and phylogeography of the west African species of Taterillus (Rodentia, Gerbillinae). P. 33 in Small African mammals (C. Denys, L. Granjon, and A. Poulet, eds.). Colloques et Se´minaires— Editions de l’ IRD, Paris, France. DUCROZ, F., L. GRANJON, P. CHEVRET, J. M. DUPLANTIER, M. LOMBARD, AND V. VOLOBOUEV. 1997. Characterization of two distinct species of Arvicanthis (Rodentia: Muridae) in West Africa: cytogenetic, molecular and reproductive evidence. Journal of Zoology (London) 241:709–723. DUPLANTIER, J.-M., L. GRANJON, AND K. BA. 1997. Re´partition biogeógraphique des petits rongeurs au Seńe´gal. Journal of African Zoology 111:17–26. EISENTRAUT, M. 1963. Die Wirbeltiere des Kamerungebirges. Verlag Paul Parey, Hamburg, Germany. ELLERMAN, J. R., R. W. HAYMAN, AND G. W. C. HOLT. 1941. The families and genera of living rodents. British Museum (Natural History) Publications, The Trustees of the British Museum, London, United Kingdom. FEILER, A. 1990. Distribution of mammals in Angola and notes on biogeography. Pp. 221–235 in Vertebrates in the tropics (G. Peter and R. Hutterer, eds.). Alexander Koenig Museum, Bonn, Germany. FILIPPUCCI, M. G., H. BURDA, E. NEVO, AND J. KOCKA. 1994. Allozyme divergence and systematics of common mole-rats (Cryptomys, Bathyergidae, Rodentia) from Zambia. Zeitschrift fu¨r Saügetierkunde 59:42–51. FILIPPUCCI, M. G., M. KAWALIKA, M. MACHOLAN, A. SCHARFF, AND H. BURDA. 1997. Allozyme differentiation and systematic relationship of Zambian giant mole-rats, Cryptomys mechowi (Bathyergidae, Rodentia). Zeitschrift fu¨r Saügetierkunde 62:172–178. GAIGHER, I. G., AND R. MCPOTT. 1973. Distribution of fishes in southern Africa. South African Journal of Science 69:25–27. GAUTUN, J.-C., M. TRANIER, AND B. SICARD. 1985. Liste pre´liminaire des rongeurs du Burkina Faso (ex Haute-Volta). Mammalia 49:537– 542. GORDON, D. H. 1991. Chromosomal variation in the water rat Dasymys incomtus. Journal of Mammalogy 72:411–414. GRANJON, L., J. DUPLANTIER, J. CATALAN, AND J. BRITTON-DAVIDIAN. 1992. Karyotypic data on rodents from Senegal. Israel Journal of Zoology 38:263–276. GRANJON, L., J. M. DUPLANTIER, J. CATALAN, AND J. BRITTON-DAVIDIAN. 1997. Systematics of the genus Mastomys (Thomas, 1915) (Rodentia: Muridae). A review. Belgian Journal of Zoology 127 (Suppl.):7–18. GRUBB, P. 1999. Evolutionary processes implicit in distribution patterns of modern African mammals. Pp. 150–164 in African biogeography, climate change, and human evolution (G. T. Bromage and F. Schrenk, eds.). Oxford University Press, Oxford, United Kingdom. GRUBB, P., O. SANDROCK, O. KULLMER, T. M. KAISER, AND F. SCHRENK. 1999. Relationships between eastern and southern African mammal faunas. Pp. 253–267 in African biogeography, climate change, and human evolution (G. T. Bromage and F. Schrenk, eds.). Oxford University Press, Oxford, United Kingdom.

919

HALL, B. P. 1960. The faunistic importance of the scarp of Angola. Ibis 102:420–442. HAMILTON, A. C. 1982. Environmental history of East Africa. Academic Press, London, United Kingdom. HANNEY, P. 1965. The Muridae of Malawi (Africa: Nyasaland). Journal of Zoology (London) 146:577–633. HAPPOLD, D. C. D. 1987. The mammals of Nigeria. Clarendon Press, Oxford, United Kingdom. HAPPOLD, D. C. D. 1996. Mammals of the Guinea-Congo rain forest. Proceedings of the Royal Society of Edinburgh 104B:243– 284. HAPPOLD, D. C. D., AND M. HAPPOLD. 1989. Biogeography of montane small mammals in Malawi, Central Africa. Journal of Biogeography 16:353–367. HILTON-TAYLOR, C. 2000. IUCN Red List of threatened species. IUCN, Gland, Switzerland. HUBERT, B., A. MEYLAN, A. PETTER, A. POULET, AND M. TRAINER. 1983. Different species in genus Mastomys from western, central and southern Africa (Rodentia, Muridaea). Annales Museé Royal de l’Afrique Centrale, Sciences Zoologiques 237:143–148. HUTTERER, R., F. DIETERLEN, AND G. NIKOLAUS. 1992. Small mammals from forest islands of eastern Nigeria and adjacent Cameroon, with systematical and biogeographicl notes. Bonner Zoologische Beitra¨ge 43:393–414. KAWALIKA, M., H. BURDA, AND D. BRU¨GGERT. 2001. Was Zambia a cradle of the genus Cryptomys (Bathyergidae, Rodentia)? A further new ancestral (?) species of Cryptomys from Zambia. Pp. 253–261 in Small African Mammals (C. Denys, L. Granjon, and A. Poulet, eds.). Colloques et Se´minaires—Editions de l’ IRD, Paris, France. KINGDON, J. 1971. East African mammals. An atlas of evolution in Africa. Academic Press, London, United Kingdom. LAVOCAT, R. 1956. La faune des rongeurs des grottes a` australopithe`ques. Palaeontologica Africana 4:69–75. LAVRENCHENKO, L. A., A. N. MILISHNIKOV, AND A. A. WARSHAVSKY. 2000. Allozymic phylogeny: evidence for coherent adaptive patterns of speciation in Ethiopian endemic rodents from an isolated montane massif. Bonner Zoologische Monographien 46:245–253. LECOMPTE, E., C. DENYS, AND L. GRANJON. 2001. An identification key for species within the genus Praomys (Rodentia: Muridae). Pp. 127–139 in Small African mammals (C. Denys, L. Granjon, and A. Poulet, eds.). Colloques et Se´minaires—Editions de l’ IRD, Paris, France. MEESTER, J. 1988. Chromosomal speciation in Southern African small mammals. South African Journal of Science 84:721–724. MISONNE, X. 1969. African and Indo-Australian Muridae: evolutionary trends. Museé Royale de l’Afrique Centrale, Teruven, Belgique, Series Zoologie 172:1–219. MULLIN, S. K., N. PILLAY, AND P. J. TAYLOR. 2001. Non-geographic morphometric variation in the water rat Dasymys incomtus (Rodentia: Muridae) in southern Africa. Durban Museum Novitates 26:38–44. MULLIN, S. K., N. PILLAY, P. J. TAYLOR, AND G. K. CAMPBELL. 2002. Genetic and morphometric variation in populations of South African Dasymys incomtus incomtus. Mammalia 66:381–404. MULLIN, S. K., P. J. TAYLOR, AND N. PILLAY. In press. Skull size and shape of Dasymys (Rodentia: Muridae) from sub-Saharan Africa. Mammalia. MUSSER, G. D., AND M. D. CARLETON. 1993. Family Muridae. Pp. 501–755 in Mammal species of the world: a taxonomic and

920


geographic reference (D. E. Wilson and D. M. Reeder, eds.). Smithsonian Institution Press, Washington, D.C. PETTER, F. 1982. Les parente´s des Otomys du Mont Oku (Cameroun) et des autres formes rapporteés a` O. irroratus (Brants, 1827) (Rodentia, Muridae). Bonner Zoologische Beitra¨ge 33:215–222. PILLAY, N. 2003. Reproductive biology of a rare African rodent, the water rat, Dasymys incomtus. Journal of Mammalogy 84:505– 512. RAUTENBACH, I. L. 1982. Mammals of the Transvaal. Ecoplan Monograph, Pretoria, South Africa. ROBBINS, C. B. 1978. The Dahomey gap—a reevaluation of its significance as a faunal barrier to West African high forest mammals. Bulletin of Carnegie Museum of Natural History 6:168–174. ROBBINS, C. B., AND E. VAN DER STRAETEN. 1989. Comments on the systematics of Mastomys Thomas 1915 with a description of a new West African species (Mammalia: Rodentia: Muridae). Senckenbergiana Biologica 69:1–14. ROHLF, F. J. 1996. NTSYS-pc. Exeter Software, Setaouket, New York. ROSEVEAR, D. R. 1969. Rodents of West Africa. Trustees of the British Museum (Natural History) Publications, London, United Kingdom. SCHLITTER, D. A., AND M. J. DELANY. 1985. Geographic areas of Africa poorly covered in mammal research collections. Acta Zoologica Fennica 170:47–48. SLICE, D. E. 1999. GRF–ND: Generalized rotational fitting of ndimensional data. Department of Ecology and Evolution, State University of New York, Stony Brook. STUART, S. N., F. P. JENSEN, S. BRO¨GGER-JENSEN, AND R. I. MILLER. 1993. The zoogeography of the montane forest avifauna of eastern Tanzania. Pp. 203–228 in Biogeography and ecology of the rain forests of eastern Africa (J. C. Lovett and S. K. Wasser, eds.) Cambridge University Press, Cambridge, United Kingdom. SWANEPOEL, P., AND D. A. SCHLITTER. 1978. Taxonomic review of the fat mice (genus Steatomys) of West Africa (Mammalia: Rodentia). Bulletin of Carnegie Museum of Natural History 6:53–76. TAYLOR, P. J. 1998. The smaller mammals of KwaZulu-Natal. University of Natal Press, Durban, South Africa. TAYLOR, P. J., AND A. KUMIRAI. 2001. Craniometric relationships between the Southern African Vlei rat, Otomys irroratus (Rodentia, Muridae, Otomyinae) and allied species from North of the Zambezi River. Pp. 161–181 in Small African mammals (C. Denys, L. Granjon, and A. Poulet, eds.). Colloques et Se´minaires—Editions de l’ IRD, Paris, France. TAYLOR, P. J., AND J. MEESTER. 1993. Morphometric variation in the yellow mongoose, Cynictis penicillata (Cuvier, 1829) (Carnivora:

Vol. 85, No. 5

Viverridae), in southern Africa. Durban Museum Novitates 18: 37–71. TRANIER, M., AND J. C. GAUTUN. 1979. Recherches caryotypiques sur les Rongeurs de C^ ote d’Ivoire: re´sultats pre´liminaires pour les milieux ouverts. Le cas d’Oenomys hypoxanthus ornatus. Mammalia 43:252–254. TURPIE, J. K., AND T. M. CROWE. 1994. Patterns of distribution, diversity and endemism of larger African mammals. South African Journal of Zoology 29:19–30. VAN DER STRAETEN, E. 1980. A new species of Lemniscomys (Muridae) from Zambia. Annals of the Cape Provincial Museum 13:55–62. VAN DER STRAETEN, E., AND F. DIETERLEN. 1987. Praomys misonnei, a new species of Muridae from Eastern Zaire (Mammalia). Stuttgarter Beitra¨ge zur Naturkunde, Serie A (Biologie), 402:1–11. VAN DER STRAETEN, E., AND W. N. VERHEYEN. 1978a. Karyological and morphological comparisons of Lemniscomys striatus (Linnaeus, 1758) and Lemniscomys bellieri Van der Straeten, 1975, from Ivory Coast (Mammalia: Muridae). Bulletin of Carnegie Museum of Natural History 6:41–47. VAN DER STRAETEN, E., AND W. N. VERHEYEN. 1978b. Taxonomical notes on the West-African Myomys with the description of Myomys derooi (Mammalia—Muridae). Zeitschrift fu¨r Sa¨ ugetierkunde 43:31–41. VAN ZINDEREN BAKKER, E. M. 1978. Quarternary vegetation changes in southern Africa. Pp. 131–143 in Biogeography and ecology of southern Africa. (M. J. A. Werger, ed.). Dr W. Junk bv Publishers, The Hague, The Netherlands. VOLOBOUEV, V. T., ET AL. 2002. Chromosomal characterisation of Arvicanthis species (Rodentia, Murinae) from western and central Africa: implications for taxonomy. Cytogenetic and Genome Research 96:250–260. VOLOBOUEV, V. T., B. SICARD, V. M. ANISKIN, J. C. GAUTUN, AND L. GRANJON. 2000. Robertsonian polymrphism, B chromosomes variation and sex chromosomes heteromorphism in the African water rat Dasymys (Rodentia, Muridae). Chromosome Research 8:689–697. WERGER, M. J. A. 1978. Biogeographical division of southern Africa. Pp. 145–170 in Biogeography and ecology of southern Africa. (M. J. A. Werger, ed.). Dr W. Junk bv Publishers, The Hague, The Netherlands.

Submitted 8 November 2002. Accepted 10 October 2003. Associate Editor was Penny S. Reynolds.

October 2004


921

APPENDIX I Locality and operational taxonomic unit (OTU) names for the specimens examined. The specimens are housed at the following institutions: Amathole Museum, King Williamstown, South Africa; American Museum of Natural History, New York; Bulawayo Museum, Bulawayo, Zimbabwe; Carnegie Museum of Natural History, Pittsburgh; Durban Museum of Natural History, Durban, South Africa; Field Museum of Natural History, Chicago, Illinois; Museé Naturalle d’Histoire, Paris, France; Museé Royale d’Afrique Centrale, Antwerp, Belgium; Museum of Comparative Zoology, Harvard, Cambridge, Massachusetts; Natural History Museum, London, United Kingdom; Staatliches Museum, Stuttgart, Germany; Transvaal Museum, Northern Flagship Institution, Pretoria, South Africa; United States Smithsonian Museum, Washington, D.C.; South African Museum, Cape Town, South Africa. The species names follow the current classification under Musser and Carleton (1993). Under the heading OTU, a letter designation instead of a number indicates that these specimens were analyzed at an individual level only (not OTU). 2n ¼ diploid numbers of karyotyped populations, TM ¼ traditional morphometrics, GM ¼ geometric morphometrics, (n) ¼ number of individuals. DRC ¼ Democratic Republic of Congo, CAR ¼ Central African Republic. The superscript letters found under the column 2n refer to the specific references that document the diploid numbers for that locality. Species D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D.

rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus rufulus foxi foxi rufulus rufulus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus

Country

Locality

OTU

Coordinates

2n

TM(n)

GM(n) dorsal

GM(n) ventral

Senegal Senegal Senegal Mali Guinea Guinea Sierra Leone Sierra Leone Liberia Liberia Liberia Liberia Ivory Coast Ivory Coast Ivory Coast Ivory Coast Ivory Coast Ivory Coast Ivory Coast Ivory Coast Ivory Coast Ivory Coast Burkina-Faso Burkina-Faso Ghana Ghana Ghana Togo Togo Togo Benin Benin Benin Nigeria Nigeria Nigeria Nigeria Nigeria Nigeria Nigeria Nigeria Chad Sudan Sudan Sudan Sudan Sudan Sudan Sudan Sudan Cameroon

Niayes Lake Retba Lake Mbaouane Kangaba Macenta Yale´ Rokupr Sandaru Mt Nimba Mt Coffee Begwai Harbel Sienso Tiegbe Kong Bolo Lamto Bouake´ IDERT Adiopodoume´ Dabou Monpoyeme´ Banfora Bobo-Dioulasso Pulima Pirisi Wenchi Padori Pewa Pagola Zizonkame´ Ke´tou Nikki Dada Gudi Ibadan Lagos Panyam Ugar Jabar Umuahia Ashaka Bekao Mt. Kinyeti Torit Juba Iwatoka Kagelu Raga Jebel Mt. Baginzi Mie´ri

1 2 2 A 3 3 B C 4 5 6 6 7 8 9 10 11 D 12 13 14 15 16 16 17 18 19 20 20 21 22 22 22 23 E 24 24 25 25 26 F G 27 27 28 28 28 H I J 29

148529N, 158419W 148509N, 178159W 148409N, 178269W 118589N, 88249W 88339N, 98289W 78399N, 88309W 88409N, 128239W 88249N, 108429W 78449N, 88289W 68309N, 108359W 68139N, 108069W 68199N, 108209W 98259N, 78319W 98429N, 58209W 98099N, 48379W 88299N, 78349W 68439N, 58159W 78419N, 58029W 58209N, 48079W 58209N, 48079W 58259N, 48339W 58189N, 48279W 108369N, 48459W 118119N, 48189W 108519N, 28039W 108079N, 28279W 78429N, 28079W 108139N, 08259E 98179N, 18149E 88119N, 08589E 78559N, 28019E 78219N, 28379E 98569N, 38139E 118349N, 48299E 88549N, 88179E 78179N, 38309E 68279N, 38239E 58329N, 78299E 58389N, 68249E 98279N, 98129E 98319N, 88239E 88349N, 168059E 38579N, 328549E 48279N, 328319E 48529N, 318309E 38459N, 308389E 48039N, 308379E 88279N, 258359E 88359N, 248429E 78469N, 278409E 48149N, 138589E

36a 36b

3 5 1 2 23 2 2

2 4 1 0 17 2 2

2 4 0 0 14 2 2

3 3 2 0 7 7 5 10 7 2 9 7 12 13 6 1 5 5 3 6 1 3 1 1 2 7 2 3 2 11 2 0 1 2 1 2 2 1 1 1 1 1 5

3 3 1 2 9 6 5 8 7 3 9 7 12 14 4 1 7 4 3 6 1 2 1 1 2 8 2 5 4 10 3 3 1 0 1 1 0 0 0 1 1 0 4

3 3 1 2 7 6 5 11 4 2 7 6 8 11 4 1 8 4 3 6 1 2 1 1 2 7 2 4 4 9 2 0 0 1 1 1 0 0 0 0 0 0 4

3639a

36a,c

36d

Type locality

D. rufulus

D. foxi

D. i. palustris

D. i. shawi

922

Vol. 85, No. 5

JOURNAL OF MAMMALOGY APPENDIX I.—Continued.

Species D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D. D.

incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus incomtus

Country

Locality

OTU

Coordinates

Cameroon Cameroon Cameroon CAR CAR Congo Congo DRC DRC DRC DRC DRC DRC DRC DRC DRC DRC DRC Tanzania Tanzania Tanzania Tanzania Tanzania Angola Angola Angola Angola Angola Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Zambia Malawi Malawi Malawi Malawi Mozambique Mozambique

Yaounde´ Mt Cameroon Tongo Bangui La Maboke´ M9boko Brazzaville Bagbele Faradje Nambira Nambirima Medje Niangara Likouala Lukolela Ngombe Lumbumbashi Kinshasha Usuhilo Mlali Dabaga Igale Ilolo Duque du Braganca Dundo Chitau Dondi Mombolo Kasungu Abercorn Kasama Ft Rosebery Lake Chiyaya Ntambo’s Area Mpika Solwezi Boma Kabompo-Lunga Confluence Luansongwe River Nsangi-Kabompo Confluence Kasempa Boma Temwa Laha Lusiwashi Chibale Kasinga Fort Jameson Kabombo-Zambezi Confluence Manyinga River Mankoya Mayau Limalunga Senanga Chilanga Kafue River Choma Hot Springs Kasungu Zomba Plateau Mulanje Plateau Nyika Plateau Furancungo Vila Vasco De Gama

30 31 32 33 33 K L 34 34 35 35 36 36 M N O 37 38 P 39 Q R S 40 T 41 41 U V W 42 43 43 44 45 46 47 47 47 47 47 48 48 49 50 51 51 52 53 54 54 55 55 55 X 56 57 58 Y 59 59

38529N, 118319E 48129N, 98119E 48119N, 168149E 48229N, 188359E 38549N, 178569E 08369N, 148539E 48149N, 158149E 48219N, 298179E 38449N, 298439E 48209N, 298169E ca. 48209N, 298169E 28259N, 278189E 38429N, 278529E 18379N, 188049E 18109S, 178119E 68409S, 208579E 118409S, 278289E 48179S, 158189E 68249S, 338579E 68589S, 378339E 88079S, 358559E 98049S, 338239E 98109S, 338369E 98069S, 158579E 88589S, 218369E 118259S, 178099E 128319S, 168259E 118359S, 148259E 88279S, 298229E 88509S, 318209E 108139S, 318109E 118119S, 288539E 118159S, 298459E 118449S, 248269E 118529S, 318269E 128109S, 268249E 128309S, 248539E 128309S, 248539E 128309S, 248539E 138259S, 258509E 138259S, 258509E 138009S, 308479E 138369S, 308089E 138339S, 238089E 138399S, 328409E 148179S, 238129E 148179S, 238129E 148489S, 248489E 148559S, 278409E 158169S, 238089E 168079S, 238159E 158359S, 288189E 158529S, 278459E 168379S, 268589E ?? 138019S, 338309E 158209S, 358169E 168039S, 358319E 108359S, 338429E 148269S, 338239E 148549S, 328159E

2n

TM(n)

GM(n) dorsal

GM(n) ventral

3 3 0 1 3 1 1 6 2 10 2 6 0 1 1 1 4 4 2 6 1 1 1 5 2 7 0 1 2 1 15 1 3 5 4 12 4 1 2 1 1 1 2 11 7 1 2 1 3 51 1 2 4 2 2 3 3 7 2 2 2

3 0 3 1 3 0 1 6 4 10 2 5 0 1 0 0 4 4 1 6 0 1 0 5 1 7 2 0 0 0 6 0 3 4 3 13 0 0 4 0 0 0 2 9 5 1 0 0 0 39 0 2 5 2 0 2 4 5 0 2 2

0 0 3 1 2 0 1 4 3 8 1 2 1 1 0 0 3 3 1 6 0 1 0 5 1 7 0 0 0 0 4 0 0 3 3 6 3 0 0 0 0 0 1 8 5 0 0 0 0 30 0 2 5 2 0 2 3 6 0 2 2

Type locality D. i. longipilosus

D. b. edsoni D. i. bentleyae

D. i. alleni

October 2004


923

APPENDIX I.—Continued. Species

Country

Locality

OTU

Coordinates

D. incomtus D. incomtus D. incomtus

Mozambique Mozambique Mozambique

Vila Coutinho Zambue Beira

60 61 62

148449S, 348229E 158109S, 308509E 198519S, 348549E

2n

TM(n)

GM(n) dorsal

GM(n) ventral

4 3 5

5 3 5

5 3 5

Type locality

a

Volobouev et al. 2000. Granjon et al. 1992. Tranier and Gautun 1979. d Gautun, et al. 1985. b c

APPENDIX II Descriptions of landmarks recorded on the dorsal skull (represented in Fig. 1). The landmark number is followed by a description of the landmark and the landmark type (in parentheses). Type 1 ¼ discrete juxtaposition of tissues; Type 2 ¼ maximum curvatures and bony process tips; Type 3 ¼ extremal points. Dorsal view. Landmark 1—anterior tip of nasals (2). Landmark 2— anterior point at suture between nasals and premaxilla (1). Landmark 3—narrowest point of rostrum (3). Landmark 4—anterior point of upper maxillary process (2). Landmark 5—anterior point of interior orbit (2) Landmark 6—widest point of zygomatic arch (3). Landmark 7—posterior point of interior orbit (2). Landmark 8—exterior tip of external auditory meatus (2). Landmark 9—edge of supraoccipital ridge (2). Landmark 10—posterior point of supraoccipital (2). Landmark 11—Junction between interparietal, parietal, and midline (1). Landmark 12—junction between parietal, frontal, and midline (1). Landmark 13—junction between frontal, nasals, and midline (1).

Ventral view. Landmark 1—anterior tip of nasals (2). Landmark 2— widest point of rostrum (3). Landmark 3—anterior point of upper maxillary process (2). Landmark 4—anterior point of internal orbit (2). Landmark 5—posterior point of internal orbit (2). Landmark 6— maximum anterior curvature of tympanic bulla (2). Landmark 7— exterior tip of external auditory meatus (2). Landmark 8—maximum external curvature of posterior tympanic bulla (2). Landmark 9— maximum interior curvature of posterior tympanic bulla (2). Landmark 10—posterior tip of foramen magnum (2). Landmark 11—anterior tip of foramen magnum (2). Landmark 12—posterior tip of sphenopalatine vacuities (2). Landmark 13—junction of presphenoid, palatine and midline (1). Landmark 14—posterior tip of palatine foramen (2). Landmark 15—anterior tip of palatine foramen (2). Landmark 16— anterior edge of m1 (2). Landmark 17—interior junction between m1 and m2 (2). Landmark 18—interior junction between m2 and m3 (2). Landmark 19—posterior edge of m3 (2). Landmark 20—exterior junction between m1 and m2 (2). Landmark 21—exterior junction between m2 and m3 (2).