Praomys taitae, (HELLER, 1912)

Carleton, Michael D. & Stanley, William T., 2012, Species limits within the Praomys delectorum group (Rodentia: Muridae: Murinae) of East Africa: a morphometric reassessment and biogeographical implications, Zoological Journal of the Linnean Society 165 (2), pp. 420-469 : 444-447

PRAOMYS TAITAE ( HELLER, 1912) ( FIGS 6–9 ; TABLES 5, 6)

Epimys taitae Heller, 1912: 9 ; type locality – ‘ Mt. Mbololo , Taita Mountains [Taita Hills], British East Africa [ Kenya], 5000 feet altitude’ [1524 m]; type specimen – USNM 181797 , an old adult male collected 5 November 1911 by E. Heller .

Rattus [( Praomys )] taitae, Hollister, 1919: 79 (name combination).

Rattus (Praomys) delectorum, Allen & Loveridge, 1927: 435 (part, provisional identification of series from the Ululguru Mts, Tanzania).

Praomys taitae, G. M. Allen, 1939: 410 (name combination).

Praomys jacksoni octomastis Hatt, 1940: 2 ; type locality – ‘ 6000 feet [1829 m] in the Old Mbulu Reserve [Kainam], Tanganyika Territory’ [ Tanzania] ; type specimen – AMNH 55718 , an adult female collected 28 November 1928 by A. L. Moses .

Rattus [( Praomys )] jacksoni octomastis, Swynnerton & Hayman, 1951: 316 (name combination, retention as valid subspecies).

Praomys delectorum, Davis, 1965: 131 (octomastis and taitae allocated as synonyms without indication of rank).

Praomys jacksoni, Misonne, 1974: 27 (octomastis and taitae allocated as synonyms without indication of rank).

Praomys (Hylomyscus) denniae anselli, Bishop, 1979: 528 (part, misidentification of series from Ngorongoro, Tanzania).

Emended diagnosis: A species of the Praomys delectorum group characterized by small size ( Table 5), delicate molars (CLM ª 4.0–4.2 mm), and relatively heavy incisors; tops of hind feet whitish, lacking dusky swath across metatarsum.

Distribution: Chyulu Hills and Taita Hills of southern Kenya, through the Northern Highlands and Eastern Arc Mountains of northern and central Tanzania, as far south as the western Udzungwa Mts ( Fig. 10 ); known elevation 230–2895 m, the majority of locality records falling between 1000 and 2000 m.

Morphological description and comparisons: Heller’s (1912: 9) characterization of the pelage of P. taitae was brief: ‘Dorsal area russet, darkest medially; sides lighter cinnamon, and sharply contrasted with light underparts; ears and tail broccoli-brown; feet white; underparts whitish, with a cream-buff suffusion; the hairs plumbeous basally’. His chromatic synopsis captures the reddish-brown colours and brighter tone that dominate the pelage in samples of P. taitae , especially compared with those of P. melanotus ( Figs 8A–C , 9 ). Some shade of reddish brown covers the crown, back, and rump, tending toward fuscous brown or umber over the middle dorsum and grading to cinnamon or fulvous brown on the cheeks and flanks. The junction of the dorsal–ventral pelage is sharply delineated, without any ochraceous lateral line. Relatively long terminal bands of white or cream partially obscure the lead-grey bases of the ventral hairs, imparting a medium grey to greyish white appearance to the underparts; patches of entirely white hairs occur over the throat and chest in some specimens. Series of P. taitae lack the strong overwash of pinkish-buff that commonly embellishes the underparts of P. delectorum ; compared with series of P. melanotus , the long white tips of ventral hairs give a brighter tone to the grey underparts of P. taitae ( Fig. 8A′ –C′). Tops of hind feet of P. taitae are uniformly white to dirty white, from the ankles to the tips of toes, standing apart visually from the brown dorsal pelage; this trait usefully contasts with the pes of P. delectorum and P. melanotus , both of which commonly possess a swath of dusky hairs that extend over the tarsus onto the metatarsus ( Fig. 9 ).

Overall small size is the defining quality of P. taitae among the three species of the P. delectorum group. Its diminuitive size and delicate proportions are attested by the univariate statistics for many variables ( Table 5) and by variable loading coefficients in the multivariate analyses ( Tables 3, 4). Size contrast, unrelated to age, is easily appreciated between samples of the smaller P. taitae and larger P. melanotus ( Fig. 6 ), but discrimination is less evident between those of P. taitae and P. delectorum . Their cranial differentiation rests upon the narrower zygomatic plate, larger molars, and finely built incisors in P. delectorum compared with relatively broader zygomatic plate, dainty molars, and stout incisors in P. taitae ( Fig. 7 ).

Remarks: Like Thomas’ (1910) description of Epimys delectorum, Heller (1912: 9) placed his new species, the ‘Taita Forest Mouse,’ within Epimys . Heller (1912: 9) contrasted taitae only with the geographically nearby form peromyscus ( Heller, 1909) – ‘much smaller in size, with shorter ears; skull differs decidedly in lacking beads to interorbital edges, which are rounded, and in the shorter palatal foramina, which reach only anterior edge of first molar’ – the latter currently treated as a synonym of Praomys jacksoni (e.g. Musser & Carleton, 2005). Such trenchant differences served to easily diagnose taitae as distinct from the very different P. jacksoni (peromyscus), but Heller’s description evidenced no awareness of Epimys delectorum Thomas (1910) , then recently described from southern Malawi. Later authors would associate taitae as a syononym of P. delectorum ( Davis, 1965; Musser & Carleton, 1993), and our morphometric evidence indicates its status as a species within the P. delectorum group.

Hatt (1940) considered his new subspecies octomastis to be a distinctive regional form of Praomys jacksoni . At the time (i.e. Allen, 1939), the construct of P. jacksoni was broadly defined to include populations and taxa that subsequent authors have segregated as P. delectorum . As reflected in his choice of the specific epithet octomastis, Hatt (1940: 2) emphasized the mammary formula (2 + 2 = 8) as the ‘only noteworthy point of divergence from surrounding subspecies’. Surrounding subspecies in this case were populations of P. jacksoni peromyscus , which Hatt recognized as possessing only three pairs of mammae (1 + 2 = 6). Hatt (1940: 2) was aware of the description of melanotus by Allen & Loveridge (1933) – octomastis differs from jacksoni melanotus in ‘its generally lighter colour, particularly marked on the under surface’ – but he curiously overlooked their mention (1933: 107) of a nursing female with ‘four pairs of nipples’. All three females of the five specimens forming the type series of octomastis, including the type itself (AMNH 55718), do exhibit eight prominent mammary glands, but we also have verified this count on numerous females within series of P. delectorum , P. melanotus , and P. taitae . The trait is appropriately considered a characteristic of the P. delectorum species group, a contrast that distinguishes it from most members of the P. jacksoni group ( Lecompte et al., 2002a; P. degraaffi also possesses eight teats – Van der Straeten & Kerbis Peterhans, 1999).

Nor did Hatt’s (1940) description of octomastis address Heller’s (1912) taitae , a geographically closer form and one whose morphology essentially resembles octomastis. The type specimen of octomastis (AMNH 55718) agrees closely with that of taitae (USNM 181797) in most craniodental measurements ( Table 6) and nests comfortably within the variation observed for P. taitae , as attested by the mathematical certainty of its a posteriori classification with taitae in DFA ( Fig. 4C ). The two types exhibit the core qualitative traits of the P. delectorum group – interorbital morphology, supraotic fenestration, long incisive foramina, and pronounced t3 on M1; moreover, the type series of octomastis conforms to the subtle diagnostic traits of P. taitae , notably the wholly white metatarsum, intermediately sized zygomatic plate, smaller molars, and comparatively robust incisors.

As we alluded to above, Allen & Loveridge’s (1933) initial assignment of the Kigogo specimen (MCZ 26498), a juvenile from the Udzungwa Mts, EAM, to melanotus was a dubious allocation. Series from the Udzungwa Mts available to us (see Specimens examined) morphologically fit with other populations in the EAM, the senior name for which is Heller’s (1912) taitae . Although a juvenile, measurements of the unworn molars of MCZ 26498 (CLM = 4.17 mm, WM1 = 1.24 mm) fall at the lowest limits recorded for P. melanotus but within the midrange of values obtained for P. taitae ( Table 5). We regard this past record of melanotus from the EAM as an initial misidentification.

Bishop (1979) allocated specimens from Ngorongoro to the type series of his new subspecies Hylomyscus denniae anselli . Those that we have examined (BMNH 65.3612 –65.3617) prove to be examples of Praomys taitae as indicated by their size, long rostrum and relatively narrow braincase, broad zygomatic plate with deep notch, and well-defined t3 and t9 on the M1.

Although the taxa delectorum , melanotus , and octomastis eventually became associated as geographical representatives of a widely distributed P. jacksoni ( Allen, 1939; Swynnerton & Hayman, 1951; Misonne, 1974), the small Praomys named from the Taita Hills was regularly maintained as a species over this period ( Hollister, 1919; Allen, 1939; Allen & Loveridge, 1942; Swynnerton & Hayman, 1951). Within Tanzania, Swynnerton & Hayman (1951) assigned specimens from the Usambara and Uluguru Mts in the eastern EAM to Rattus taitae , which they maintained as distinct from R. jacksoni melanotus . With reallocation of the populations from the Udzungwa Mts to P. taitae , contra Allen & Loveridge (1933) and Swynnerton & Hayman (1951), the southern distributional occurrence of P. taitae is now well documented, approximately delimited by the Makambako Gap ( Fig. 10 ). The northern extent of its range is less clear. While the species certainly occurs in the Northern Highlands of Tanzania, we have not found P. taitae in the Gregorian Rift mountains of central Kenya (Mau Escarpment, Aberdare Mts, Mt Kenya), at least in North American collections. Search of other natural history collections is recommended to clarify whether it inhabits those Kenyan highlands.

Natural history notes: Samples of Praomys taitae are generally found in submontane, montane, and upper montane habitats (sensu Lovett, 1993b), but have been documented as low as 230 m in the Kwamgumi Forest Reserve, in lowland forest at the base of the East Usambara Mountains ( Stanley et al., 2005b). The abundance of this species at this elevation (as measured by trap success) was low relative to sites in the same mountain range at higher elevations (Stanley & Goodman, 2011). The Udzungwa Scarp Forest Reserve, Udzungwa Mts, contains undisturbed forest that ranges from 300 to 2000 m, one of the last natural and continuous elevational gradients that can be sampled in the Eastern Arc Mountains. Praomys taitae (reported as P. delectorum ) was recorded from 600 to 2000 m, but in abundance greater than expected at 1460 m; none of the other elevations (600, 910 and 2000 m) yielded numbers of P. taitae greater than expected ( Stanley & Hutterer, 2007). Notably, P. taitae was not found during a survey of the same forest at 300 m of this escarpment a year earlier (W. T. Stanley, unpubl. data). The highest elevation where P. taitae has been documented is 2900 m in forested habitats on Mt Kilimanjaro. Significantly, this species was not observed above the tree line during a faunal survey of this mountain (W. T. Stanley, unpubl. data). Thus, P. taitae is restricted to forests, but ranges lower in elevation than Hylomyscus arcimontensis , with which it is commonly sympatric.

Praomys taitae shares its habitat with several other rodent species. Based on intensive faunal surveys over the past two decades, three species have been recorded in every montane setting where P. taitae has been recorded: Lophuromys aquilus , Grammomys ibeanus , and Graphiurus murinus . Other species found with P. taitae in the Eastern Arc Mountains are Beamys hindei and Hylomyscus arcimontensis . The latter two rodents are unknown from the northern highland massifs. The calls for critical systematic evaluation of the various isolated populations of Grammomys ibeanus and Graphiurus murinus in East African mountains ( Holden, 2005; Musser & Carleton, 2005) advise caution about sympatry among various species until such revisionary studies are conducted.

Data collected over three dry season surveys in the Usambara Mountains indicate that populations of P. taitae have limited reproductive activity between July and September. Only 13.5% of females examined (N = 74) were pregnant and only one in 44 (2.3%) females assessed was lactating (Stanley & Goodman, 2011). Between 2004 and 2006, Makundi, Massawe & Mulungu (2006) found a significant density increase in P. taitae (reported as P. delectorum ) in the Magamba Forest, West Usambara Mts, in March– June compared with September–December. The greatest number of scrotal testes and perforate vaginas were scored from the period April to June. Based on these data, little reproductive activity occurs during the dry period (July–August) and P. taitae apparently breeds annually within a relatively short period. This conclusion is also borne out by data collected during surveys of other Eastern Arc and northern highland populations conducted every year since 1993. In 251 females examined during these collective surveys, only 31 (12.3%) were pregnant. A notable exception is the sample of nine specimens examined in 1996 in the Uluguru Mountains where five (55%) were pregnant. The number of embryos averaged 4.0 (N = 31, range 3–6), the embryos in a single uterine horn ranging from 1 to 6. The embryos averaged 10.5 mm in crown–rump length (range = 2–24 mm, N = 20).

Unlike the scansorial proclivities of Hylomyscus arcimontensis , individuals of Praomys taitae are typically captured in traps set on the ground or in other terrestrial settings, such as on rocks or on logs easily reachable from the ground. Arboreal traps set by Stanley and colleagues commonly caught examples of Hylomyscus arcimontensis , but not those of Praomys taitae . Kingdon (1974) stated that P. jacksoni is partially arboreal.

Specimens examined (365, as follows): KENYA: Nairobi area , Chyulu Hills, 6000 ft ( USNM 344947 ) ; Coast Region, Taita Hills, Mt Mbololo , Summit , 5000 ft ( FMNH 43458–43460 ; MCZ 32124, 32125 , 32131 ; USNM 181797 , 183411–183420 , 183422–183426 , 183429–183432 , 184508–184524 , 184525 ) ; Taita Hills, Mt Umengo , Summit ( USNM 183433–183441 ) ; Taita District, Taita Hills, Ngangao Forest , 4.25 km N and 1.75 km W Wundanyi, 1700 m ( CM 102635 –102643 , 102645–102647 , 102649–102652 ) .

TANZANIA: Arusha Region, Arumeru District, Arusha National Park, Mount Meru , 2300 m ( FMNH 208574–208580 , 208583–208585 , 208595–208597 , 208599 , 208601 , 208607 , 208611 , 208612 , 208615 , 208619 ) ; Ngorongoro North , 7500 ft ( BMNH 65.3612 – 65.3617) ; Old Mbulu Reserve , 5500 ft ( AMNH 55702 ), 6000 ft ( AMNH 55715–55718 ) . Dodoma Region, Mpwapwa District, Rubeho Mts, Mwofwomero Forest Reserve , 1923 m ( FMNH 197908–197911 , 197916 , 197918 , 197919–197924 , 197926 , 197927 , 197929 , 197930 , 197934 ) ; Mpwapwa District, Rubeho Mts, Mwofwomero Forest , near Chugu Peak, 1900 m ( FMNH 197943 , 197945 , 197950–197952 , 197955 ) . Iringa Region, Iringa District, Udzungwa Mts, Ndundulu Forest , 9 km E Udekwa, 1900 m ( FMNH 177709 , 177711–177715 , 177717 , 177719 , 177720 ) . Kilimanjaro Region, Moshi District, Mt Kilimanjaro, Kilimanjaro National Park , 7 km N and 2.5 km W Maua, 8100 ft ( FMNH 173881 , 174269–174271 , 174273–174283 ) ; Moshi District, Mt Kilimanjaro, Kilimanjaro National Park , 10.5 km N and 3.5 km W Maua, 9500 ft ( FMNH 174285 , 174286 ) ; Moshi District, Mt Kilimanjaro, 4 km N and 1.5 km W Maua, 6700 ft ( FMNH 174287– 174294 , 174296 , 174297 , 174299–174309 ) ; Mwanga District, North Pare Mts, Kindoroko Forest Reserve , 1688 m ( FMNH 192692 , 192696 , 192697 , 192701 , 192704 , 192705 , 192709 , 192711 , 192719 , 192721 , 192722 , 192728 , 192730 , 192733 , 192734 ) ; Mwanga District, North Pare Mts, Minja Forest Reserve , 1572 m ( FMNH 192737 , 192738 , 192741 , 192743 , 192744 , 192748 , 192749 , 192750 , 192752 , 192755 , 192762–192764 , 192769 , 192773 ) ; Same District, South Pare Mts, Chome Forest Reserve , 7 km (by air) S Bombo, 1100 m ( FMNH 151311 , 151312 , 151320 , 151335 , 151336 , 151338 , 151339 , 151344 , 151348 , 151349 , 151351 , 151360 ) ; Same District ( FMNH 153972–153974 , 153976–153987 , 153993 , 153996 ) . Morogoro Region, Udzungwa Mts , Kigoro ( MCZ 26498) ; Kilombero District, Udzungwa Mts , 3.5 km W and 1.7 km N Chita, along Chita-Ihimbo trail, 910 m ( FMNH 155466 , 155467 , 155625 , 155627 , 155630 ) ; Kilombero District, Udzungwa Mts , 4.5 km W Chita, along Chita-Ihimbo trail, 600 m ( FMNH 155478 , 155653 ) ; Kilombero District, Udzungwa Mts , 19.5 km N and 0.5 km W Chita, 2000 m ( FMNH 155480 , 155654 , 155658–155660 ) ; Kilombero District, Udzungwa Mts , 4 km W and 5 km N Chita, along Chita-Ihimbo trail, 1460 m ( FMNH 155631 , 155637– 155647 , 155649 , 155651 , 155652 ) ; Kilosa District, Malundwe Mts, Mikumi National Forest , 985 m ( FMNH 187296–187301 , 187303–187311 , 187313– 187318 ) ; Kilosa District, Rubeho Mts, Ilole Forest , 1878 m ( FMNH 197956 , 197957 , 197960–197964 ) ; Kilosa District, Ukaguru Mts, Mamiwa-Kisara Forest, 1.5 km S Mt Munyera , 1840 m ( FMNH 166687 , 166688 , 166944 , 166946–166949 , 166951 , 166952 , 166954 , 166956 , 166957 , 166959 ) ; Kilosa District, Ukaguru Mts, Mamiwa-Kisara Forest, 1 km E and 0.75 km S Mt Munyera , 1900 m ( FMNH 166685 , 166960–166962 , 166965 , 166966 , 166968 ) ; Morogoro District, Nguru Mts, Manyangu Forest Reserve , 8 km N and 3 km W Mhonda, 1000 m ( FMNH 161281 , 161282 , 161284–161286 , 161289 , 161291 , 161294– 161297 ) ; Morogoro District, Nguru Mts, Nguru South Forest Reserve , 6 km N and 6 km W Mhonda, 1500 m ( FMNH 161301–161305 , 161307 , 161310 ) ; Morogoro District, Uluguru Mts, Uluguru North Forest , 3 km W and 1.3 km N Tegetero, 1345 m ( FMNH 158366 , 158368 , 158370 , 158371 , 158373–158375 ) ; Morogoro District, Uluguru Mts, Uluguru North Forest , 5.1 km W and 2.3 km N Tegetero, 1535 m ( FMNH 158379– 158381 , 158384–158386 , 158388 , 158563 , 158574 ) ; Uluguru Mts, Bagilo ( MCZ 22504, 22506 ) ; Uluguru Mts, Vituri ( MCZ 22509, 22512 , 22513 ) . Tanga Region, Hadeni District, Nguu Mts, Nguru North Forest Reserve , 5.6 km S and 3 km E Gombero, 1180 m ( FMNH 168350 , 168354 , 168355 , 168357 , 168358 , 168365 , 168366 , 168368 , 168370 , 168373 , 168374 , 168376 , 168377 , 168379 , 168383 , 168384 , 168386 , 163387 , 168393 , 168394 , 168396 , 168399– 168401 ) ; Hadeni District, Nguu Mts, Nguru North Forest Reserve, 3.6 km S and 4.7 km E Gombero , 1430 m ( FMNH 168402 , 168406 , 168408 , 168412 , 168413 , 168416 ) ; Korogwe Distrtict, West Usambara Mts, 12.5 km NW Korogwe, Ambangulu Tea Estate , 1300 m ( FMNH 147309 , 147311 , 147313 , 147317– 147320 , 147327 , 147332 , 150158 , 150280 , 150281 , 150283–150290 , 150352–150354 , 151272 , 151275 , 151513 , 151514 ) ; West Usambara Mts, 11 km NW Korogwe, Ambangulu Tea Estate , 1120 m ( FMNH 147335 ) ; West Usambara Mts, Sunga , 35 mi N Lushoto at Resthouse ( USNM 340787 ) ; Muheza District, East Usambara Mts, 6 km NW Amani, Monga Tea Estate , 1100 m ( FMNH 150217 , 150218 , 150221 , 150294 , 150297 , 150307 , 150311 ) ; Muheza District, East Usambara Mts, 4.5 km WNW Amani, Monga Tea Estate , control site, 1100 m ( FMNH 150236 , 150240 , 150245 , 150250 , 150319 , 150324 , 151280 , 151298 , 151300 ) ; Muheza District, East Usambara Mts, 8 km WNW Amani, Bulwa Tea Estate , 1000 m ( FMNH 150257 , 150332 , 150337 ) ; Muheza District, East Usambara Mts, 4.5 km ESE Amani, Monga Tea Estate , control site, 900 m ( FMNH 150262 , 150268 , 150274 , 150279 , 150339 , 150346 , 150349 , 151305– 151307 ) ; Muheza District, Kwamgumi Forest Reserve , 4.4 km W Mt Mhinduro, 2 km S Kwamtili Estate offices, 230 m ( FMNH 153998 ) ; Magrotto Mt , Magrotto Estate ( MCZ 39057) .

Figures

Figure 4. Three scatter plots depicting results of discriminant function analysis performed on 18 log-transformed craniodental variables, as measured on 541 intact adult specimens representing 21 OTUs of the Praomys delectorum complex. A, projection of individual specimen scores onto the first two canonical variates (CV) extracted. B, projection of centroids for the 21 OTUs onto the first two canonical variates (see Material and Methods for general locality and sample sizes; bars indicate ± 1 standard error of the mean for a sample centroid); OTU numbers 1–15 are placed approximate to the entire cluster in view of the extensive overlap and visual congestion among those analytical samples. C, multivariate disposition of the four type specimens entered as unknowns and based on a posteriori classification using the discriminant function coefficients; confidence envelopes represent 1 standard deviation around a grand centroid and enclose approximately 68% of specimen scores for each. See Table 4 for variable correlations and variance explained.

Figure 6. Dorsal (top row) and ventral (bottom row) views (about 2¥) of adult crania of the three species of the Praomys delectorum group recognized here: A, P. delectorum (FMNH 181214; ONL = 29.0 mm), a male from Lichenya Hut, Mt Mulanje, 1840 m, Malawi; B, P. melanotus (FMNH 163682; ONL = 29.9 mm), a female from 5 km east of Ilolo, Rungwe Forest Reserve, Mt Rungwe, 1870 m, Tanzania; C, P. taitae (FMNH 153979; ONL = 28.7 mm), a male from 3 km east and 0.7 km north of Mhero, Chome Forest Reserve, South Pare Mts, 2000 m, Tanzania.

Figure 7. Lateral view (about 2¥) of adult crania of the three species of the Praomys delectorum group (same specimens as illustrated in Fig. 6): A, P. delectorum; B, P. melanotus; C, P. taitae.

Figure 8. Dorsal (left set) and ventral (right set) views of adult round skins (about 0.5¥), illustrating ‘typical’ pelage patterns of the three species of the Praomys delectorum group recognized here: AA′, P. delectorum (FMNH 181237; Total Length = 223 mm), a male from Chisongeli Forest, Mt Mulanje, 1575 m, Malawi; BB′, P. melanotus (FMNH 205299; Total Length = 234 mm), a male from Ngozi Crater, Poroto Mts, 2250 m, Tanzania; CC′, P. taitae (FMNH 150346; Total Length = 240 mm), a male from 4.5 km ESE of Amani, East Usambara Mts, 900 m, Tanzania.

Figure 9. Examples of the Praomys delectorum group as photographed in the field: Top, P. melanotus from Ngozi Crater in the Poroto Mountains, Southern Highlands; Bottom, P. taitae from Mount Malundwe, central Eastern Arc Mountains. Darker pigmentation of the epidermal surfaces is typical of individual P. melanotus compared with P. taitae; also note the extension of dark hairs onto the metatarsum as characteristic of P. melanotus versus a wholly white metatarsum and phalanges in most P. taitae. Hindfoot length averages about 23–25 mm; photographs by W. T. Stanley.

Figure 10. Distributions of the three species of the Praomys delectorum group based on specimens documented herein (see species accounts). The patchy occurrence of Praomys across this region is closely linked to equally patchy associations of wet montane forest (areas shaded pale grey, as extracted from World Wildlife Fund Ecoregions, USGS Digital Atlas of Africa).