Grammomys dolichurus (Smuts, 1832)

Bryja, Josef, Kerbis Peterhans, Julian C., Lavrenchenko, Leonid A., Nicolas, Violaine, Denys, Christiane, Bryjová, Anna, Šumbera, Radim & Mikula, Ondřej, 2025, Integrative taxonomic revision of the African thicket rats (Murinae: Grammomys): how genomics decreases the number of currently recognized species, Zoological Journal of the Linnean Society (Statistics in Society) 203 (2), pp. 1-20 : 14

Species: Grammomys dolichurus (Smuts, 1832)

Synonymy:

M [us] dolichurus Smuts, 1832

Thamnomys baliolus Osgood, 1910

Grammomys tongensis Roberts, 1931

Type locality: South Africa, near Cape Town .

Distribution: Eastern part of South Africa, from Limpopo Province along coast through KwaZulu-Natal and Eastern Cape provinces to Port Elizabeth ( Fig. 5 ; de Graaff 1981, 1997, Taylor 1998) and Eswatini (A. Monadjem, pers. comm.); the presence in southern Mozambique and, possibly, Zimbabwe should be confirmed. All previous records reported under this name in other parts of Africa belong to other species, mainly to G. surdaster .

Comments: One syntype skull of dolichurus from the Cape region is very damaged and could not be included in our morphometric analyses; the other syntype is very young, hence it was also not included. The species name has been attributed to many other taxa, and a number of scientific names were synonymized with it ( Musser and Carleton 2005). We restrict it to the South African genetic clade, which seems to be different morphologically ( Fig. 4 ; Supporting Information, File S3), as also suggested by previous studies. For example, Musser and Carleton (2005) mention that ‘specimens of true dolichurus from South Africa have duller pelage and more inflated bullae than animals from East and West Africa; should these prove to be diagnostic specific differences, the northern populations should be identified as G. surdaster ’. Our genomic analysis supports this view unequivocally. Here, as first proposed by Roberts (1951), we keep only two synonyms of G. dolichurus : baliolus and tongensis . They were both assigned to the dolichurus group with PP =.2 only, with a higher probability of belonging to the surdaster group ( Table 1 ), but the two groups are overlapping morphologically ( Fig. 4 ), and the genetic material close to both type localities clearly belongs to G. dolichurus ( Fig. 6 ).

Karyotype: The karyotype of specimens from the Eastern Cape province and from Port St. Johns is 2 n = 52, FN = 62, FNa = 58 (Kryštufek etal. 2008; notethatallspecimensreportedas G.cometes in their karyotypic and CYTB phylogenetic analyses in fact represent G. dolichurus ; see Bryja et al. 2017). Dippenaar et al. (1983) reported two different karyotypes from South Africa, 2 n = 52 and 2 n = 44, but the latter is most likely to be G. cometes (see above).

Figures

Figure 4. The scatter plot of the first three between-group principal component scores. The large, filled points mark group mean shapes, and small open points are projections of individual shapes onto the axes. Four major groups of Grammomys (sensu Bryja et al. 2017) are shown by different colours. The projections of type skull shapes are indicated by grey triangles labelled by the respective names (‘_p’ means paratypes).

Figure 5. Taxonomic summary of the cometes group (= selousi group of Bryja et al. 2017). A, the distribution of mtDNA lineages.The different colours of triangles correspond to different lineages (se1–se5 described by Bryja et al.2017; se6 and se7 are new discoveries in this study). B, delimitation of MOTUs based on ddRAD data. All studied populations in the cometes group (purple triangles) were grouped in the single clade no. 8 by a branch-cutting approach (indicated by a number in a coloured rectangle).Likewise, they clustered together by INFOMAP as a single MOTU IX (see also Fig.3). The proposed taxonomy is shown on the right side of the maps.Stars show type localities for the cometes group. The proposed species name for the whole group, Grammomys cometes, is shown in larger bold letters, while smaller letters correspond to synonyms.

Figure 6. Taxonomic summary of the dolichurus and surdaster groups (sensu Bryja et al. 2017). A, the distribution of mtDNA lineages for the dolichurus (grey circles) and surdaster (triangles) groups.The different colours of triangles correspond to different lineages (su1–su10 described by Bryja et al. 2017; su11 and su12 discovered in this study). The short sequences from the type series of G. caniceps (su13) cluster unequivocally with the surdaster group as a sister to su11 (see Fig.2). We show the only two known localities of this taxon by green rhombuses (arrow ‘su13’ shows the type locality Malindi in Kenya; arrow ‘Roche et al.(1984)’ indicates the second known locality, close to Mogadisho in Somalia).B, delimitation of MOTUs based on ddRAD data. The numbers in coloured rectangles show one MOTU in the dolichurus group and three MOTUs in the surdaster group, delimited by a branch-cutting approach (numbers correspond to Fig. 3), with circles (dolichurus) and triangles (surdaster) in the same colours indicating localities of ddRAD-genotyped specimens. The proposed taxonomy is shown below the maps, with Roman numerals showing MOTUs from the INFOMAP clustering (VIII for the dolichurus group, and I and II for the surdaster group).Note that we failed to genotype G. caniceps (i.e. su13 mtDNA lineage) by ddRAD.Stars show type localities for the dolichurus (red) and surdaster (green) groups. The proposed species names for the two groups (i.e. Grammomys dolichurus and Grammomys surdaster), are shown in larger bold letters, while smaller letters correspond to synonyms.