ebook img

Morphometric differentiation among West African populations of the rodent genus Dasymys (Muroidea, Murinae), and its taxonomic implications PDF

17 Pages·1991·7 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Morphometric differentiation among West African populations of the rodent genus Dasymys (Muroidea, Murinae), and its taxonomic implications

PROC. BIOL. SOC. WASH. 104(3), 1991, pp. 419-435 MORPHOMETRIC DIFFERENTIATION AMONG WEST AFRICAN POPULATIONS OF THE RODENT GENUS DASYMYS (MUROIDEA: MURINAE), AND ITS TAXONOMIC IMPLICATIONS Michael D. Carleton and Christine Martinez Abstract.—CrdtmodQnXdX variation among West African samples o^Dasymys is reassessed based on the larger museum series now available. Two morpho- logical kinds can be discerned in this region and are referable to the locally availableepithetsDasymysrufulusMiller(1900),describedfromMountCoffee, Liberia, andDasymysfoxiThomas (1912), named from Panyam, Nigeria. The degree ofcraniodental differentiation between these two phena matches that documentedforothercongenericpairsofmuridrodentspeciesinhabitingparts ofWestAfrica. Dasymys rufulusdiffers from D.foxibyits overall smallersize, especiallyas observedin its smaller, narrowercranium and less robust molars. In West Africa, samples ofD. rufulus are broadly distributed from east ofthe Niger River, Nigeria, west to northern Sierra Leone; whereas, the known ex- amples ofD.foxi are restricted to the Jos Plateau in Nigeria. Dasymys rufulus and D.foxi are provisionally accorded specific status until taxonomic studies ofDasymyspopulationsfromotherpartsofAfricaareundertakentodetermine theirrelationships and distributionallimits. Itis doubtful thateitherform, but especially D. rufulus, is synonymous with D. nudipes or D. incomtus. Shaggy-furred or marsh rats ofthe genus pears very unlikely that there is more than Dasymys are broadly ifintermittently dis- one species ofthis genus." His conviction tributed over much ofAfrica south ofthe in this regard may have been influenced by Sahara Desert, inhabiting a variety ofwet- Osgood's(1936:256)earlierobservationthat landareaswiththickgroundcover. Current "Bothexternalandcranialcharactersinthis classifications recognize a single species, genus are elusive andthe range ofvariation Dasymys incomtus (Sundevall, 1847), throughout is comparatively slight," which throughoutthisbroadrange(Misonne 1974, he offered in the description ofZ). incomtus Honacki et al. 1982, Corbet & Hill 1986). griseifrons. Except for the occasional reten- The genus has not always been viewed as tionofD. nudipesasaspecies(Roberts 1951, monotypic, for 13 ofthe 20 species-group Crawford-Cabral 1983), subsequent classi- taxa associated with Dasymys have been fications and regional catalogues have uni- named as species. As of 1939, Allen listed formlyacknowledgedthe one speciesD. in- 10 species, four of them containing from comtus (for example, Ellerman et al. 1953, two to four subspecies. Hanney 1965, Rosevear 1969, Delany 1975, The current perception ofspecies diver- De Graff 1981, Meester et al. 1986, Hap- sity in Dasymys originated with Ellerman's pold 1987). A notable omission in arriving (1941) treatment of the genus. Ellerman at the current species arrangement ofshag- (1941:121) regarded all named forms, ex- gy-furred rats has been the presentation of cept orthos Heller (1911) which he did not datasupportive ofthesetaxonomic changes. examine, as races of D. incomtus: "It ap- The basis forrecognition ofa single species 420 PROCEEDINGSOFTHEBIOLOGICALSOCIETYOFWASHINGTON has hinged on Ellerman's surmise that "it ciatedskulls,wereexamined(seeTaxonom- appears very unhkely" that more than one icSummary)fromthefollowingcollections, exists. theiracronymslistedinparentheses: Amer- The present study contributes distribu- icanMuseumofNaturalHistory,NewYork tional and morphological information that City (AMNH); Museum of Comparative bears on the status ofDasymyspopulations Zoology, Harvard University, Cambridge inhabiting portions ofWest Africa, namely (MCZ); and the U.S. National Museum of from Nigeria west to Sierra Leone. Com- Natural History, Smithsonian Institution, paredtothenumberofformsdescribedfrom Washington, D.C. (USNM). eastern and southern Africa, the taxonomy Preparation ofthe distribution map was ofshaggy-furred rats is less problematic in based on locality information and geo- West Africa, where only two forms have graphical coordinates basically as given by been described: Dasymys rufulus Miller collectors on specimen tags. We also con- (1900) from Mount Coffee, Liberia, and sulted the gazetteers provided by Davis & Dasymys foxi Thomas (1912) from Pan- Misonne(1964), Rosevear(1965)andHap- yam, on the Jos Plateau, Nigeria. Clarifi- pold (1987), to verify the plotting ofthese cation ofpatterns ofmorphological varia- sitesandtoobtaingeographicalcoordinates tion on a regional basis, set within the not listed by collectors. context of local taxonomic designations, Several dimensions ofthe skin and skull marks a necessary step toward refining the were recordedin millimeters (mm) to char- species taxonomy of Dasymys as a whole acterizethemorphologyofthesamples. To- and gaining insight into its historical zoo- tal length, tail length (TL), and hindfoot geography. length (HFL) were extracted to the nearest mm LikeotherformsofDasymys, Miller'sru- whole from the skin tag, as was field fulusandThomas'ybx/wereviewedas sub- weight to the nearest whole gram (gm). species ofD. incomtus by Ellerman (1941), Length of the head and body (HBL) was an arrangement which has persisted to the obtained by subtracting the tail from total presentday(Rosevear 1969,Misonne 1974, length. Fourteen cranial and two dental Happold 1987). Rosevear (1969:311) even measurementsweretakentothenearest0.01 mm questioned the validity ofmaintaining any by means ofhand-held digital calipers formal racial division among West African accurateto 0.02 mm. Thesemeasurements, populations: ". the differences in size be- and their abbreviations as used herein, in- . . tweenthisrace [thatis, rufulus] 3.ndfoxiare clude: occipitonasal length (ONL); greatest really very slight and it is open to some zygomatic breadth (ZB); breadth of the doubt whether the distinction can be re- braincase (BBC); least interorbital breadth tained." Happold(1987:132)formedasim- (lOB); length ofthe rostrum (LR); breadth ilar impression based on his review ofNi- of the rostrum (BR); postpalatal length gerian samples, noting that, in general, the (PPL); length of the bony palate (LBP); smaller rufulusoccurs in southern localities length ofthe incisive foramen (LIF); length andthelarger^bjc/innorthernones,butthat ofthe diastema (LD); breadth ofthe bony "... the subspecific distinction may not be palate acrosstheupperfirstmolars (BM1s); valid." These assessments deserve review breadthofthezygomaticplate(BZP);length baseduponthelargermuseumseriesofWest oftheauditorybullae(LAB);breadthacross African Dasymys now available. theexoccipitalcondyles(BEC);crownlength of the maxillary toothrow (LMl-3); and Materials and Methods width ofthe upper first molar (WMl). The limits of most ofthese dimensions on the Approximately 300 specimens, consist- murine skull are defined and illustrated by ing ofconventional study skins with asso- Musser(1979). VOLUME NUMBER 104, 3 421 Recording ofmeasurements was limited yses ofvariances were applied to the three to animalsjudged to be adult based on the largest OTUs (G3, N4, and NIO) to assess possessionoffullyeruptedthirdmolarsand the effect ofgender on the variation ofthe adult pelage. "Adult" animals were further 16 craniodental variables. Principal com- segregated into crude age classes based on ponentswereextracted from avariance-co- upper molar wear as follows. variance matrix and computed using the Young adult.—Tricusp arrangement of craniodental variables after transformation first and second chevrons ofMl-2 well de- totheirnaturallogarithms. Loadingsareex- fined, the enamel borders ofthe medial (t1 pressed as Pearson product-moment cor- and t3) and lateral (t3 and t6) cusps deeply relation coefficients of the principal com- constricted and contiguous or virtually so; ponents with the original cranial variables. dentinal lakes ofchevrons thus narrow and Allanalytic procedureswere carriedout us- M3 discontinuoustransverselyornearlyso; ing Systat (Version 4.1, 1989), a series of erupted and slightly worn; tls on M2-3 statistical routines programmed for micro- whollyisolatedfromenamelbordersofsec- computers. ond chevron; transverse enamel valleys be- tween chevrons deep and extend across the Morphometric Results width ofthe molar. Full adult.—Enamel indentations of the Sexual dimorphism in cranial size is un- M medialandlateralcuspsonchevronsof 1- appreciablewithinpopulationso^Dasymys, 2notincontact, theirdentinbasinsbroader afindingin agreementwithbiometric stud- andcontinuousacrossthewidthofthetooth; ies ofother African murines (for example, enamel, lessoftenthedentin, oftlsonM2- Van der Straeten & Verheyen 1981, 1982; 3 connected to that ofthe second chevron Carleton&Robbins 1985). Amongthethree (anterior border oft5). largest population samples analyzed, only Oldadult.—Occlusalwearheavysuchthat sixcranialdimensionsvarysignificantlywith thedentinofchevrons,particularlythemid- regardtogender(Table 1),withmaleslarger dle one and t8, cross-connected at their in each instance, and each significant vari- medial and lateral edges; enamel valleys able is unique to a particular sample. We separating chevrons appear as islands sur- attribute the sporadic occurrence ofsignif- rounded by dentinal basins. icant findings across the three samples to a Although the above data were recorded combination ofType I samplingerrors and for all specimens examined, the multivar- unequalagerepresentationamongthesexes iateanalyseswereappliedtothe largerpop- of each sample. For example, within the ulation samples with the highest incidence Panyam sample (NIO), males are dispro- ofintact crania. These are: portionately represented in the adult (12 to OTU G3.-Ghana, Pulima, n = 37. 9) and old adult (2 to 0) age groups. As OTU II.-Ivory Coast, Kong, n = 10. discussed below, age can contribute sub- OTU 12.—Ivory Coast, Sienso, « = 15. stantial intrapopulation variation. In view OTU LI.—Liberia, Mount Coffee, « = 5 ofthe negligible orinconsistenteffects, par- (type series ofrufulus). ticularly with regard to the magnitude of OTU N4.-Nigeria, Dada, n = 24. interlocalitydifferentiation, bothsexeswere OTU N 10.-Nigeria, Jos Plateau, Pan- included in the multivariate analyses. yam Fish Farm, « = 43 (near type locality With the large series available from near o^foxi). the type locality ofD. foxi, populations in OTU T3.-Togo, Padori, « = 1 1. Nigeria have provided critical insight to Standard descriptive statistics (mean, the morphological differentiation among range, standard deviation) were derived for West African Dasymys. Thomas (1912) di- the OTUs (see Appendix 1), One-way anal- agnosed D. foxi on the basis of its larger 422 PROCEEDINGSOFTHEBIOLOGICALSOCIETYOFWASHINGTON Table L—/values from one-way anovas (effect = size,especiallyasobservedinthelengthand sex)ofthreepopulationsamplesofWestAfricanDas- breadth ofthemolarrows. Suchdifferences ymys. were borne out in principal component Cranial Ghana:Pulima Nigeria:Dada Nigeria:Panyam analyses of the two largest Nigerian sam- variables (14S,239) (83, 159) (223, 169) ples, Dada and Panyam Fish Farm, whose ONL 0.31 1.27 3.96 scoresoverlappedhttleinmultivariatespace ZB 0.01 2.11 2.65 (Fig. 1). The uniformly positive and high BBC 0.43 6.17* 1.54 correlations ofmost craniodental variables lLORB 05..5720* 00..1016 03..6730 with the first principal component suggest BR 0.48 0.29 5.46* a general size factor (Table 2) and empha- PPL 0.54 0.49 2.47 size the overall larger size of the Panyam LBP 0.10 0.62 6.83* specimens. The dispersion of individuals LIF 0.61 0.01 0.83 along the second principal component re- LD 0.70 0.63 3.17 flects certain breadth measurements (BEC, WLMMll-3 1.69 2.06 0.01 lOB)andsizeofthemolars(LMl-3,WMl), 0.13 0.84 0.06 BMls 0.01 0.51 0.02 whichagainreflectthebroadercraniumand BZP 0.11 1.41 0.16 more robust molars recorded for the spec- LAB 0.22 13.68** 0.42 imens from the vicinity ofPanyam. BEC 0.61 2.83 11.62** Some age-related sizevariation undoubt- P < 0.05; ** = /'< 0.01. edly contributes to the elliptical spread ob- served for each population sample in the bivariateplotofthefirsttwoprincipalcom- Fig. 1. Projection ofindividual specimen scores on the first two principal components for the two largest samples ofNigerian Dasymys, Dada (« = 23) and vicinity ofPanyam (« = 38). Ellipses correspond to 95% confidenceenvelopesforscoresofindividualswithineach sample. SeeTable 2. VOLUME 104, NUMBER 3 423 Panyam C2) Old Adult Dada C6] Panyam C21] Adult Dada C17] Young'=* „Panyam C15] A^ddult 1_ -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 First Principal Component Fig. 2. Boxplots offirstprincipalcomponentscores(see Fig. 1)ofDadaand Panyam samplesarrangedby three adult age classes. Mean (vertical line), one standard error ofthe mean (closed rectangle), one standard deviation(openrectangle),andtherange(horizonalline) are indicated. ponents (Fig. 1). The range offactor scores results are here reported only for the Dada viewed by age class suggests such an influ- and Panyam samples, all Nigerian speci- ence, especially for the first principal com- mens with intact skulls were subjected to ponent (Fig. 2). Still, a marked separation the same analyses and revealed the same between the Dada and Panyam samples is pattern ofdual clouds ofpoints. Specimens apparent for a given age class. The absence from the Jos Plateau (Panyam Fish Farm, of specimens classified as young adults in 2 mi N Panyam, and Ugar Jabar), which the Dada sample probably accounts for its represent Thomas' (1912) D. foxi, formed widerconfidence ellipse as well as the mar- oneofthoseconstellations,whereastheoth- ginally significant difference (P = 0.052) in ercorrespondedtospecimensfromallother the slopes oftheir regression equations. localities in Nigeria (see Specimens Exam- Rotation(varimaxprocedure)oftheprin- ined). cipal components underscored the same The distinctiveness ofthe animals from disparityin cranialbreadth (ZB, BBC, JOB, the vicinity of Panyam is still apparent in BEC) and molar size (LMl-3, WMl) be- principal component analyses including tween the Dada and Panyam samples, as samplesfromfourotherWestAfricancoun- wellasintheLAB(Table 2). One-wayanal- tries (Fig. 3). Again, overall larger size—in yses ofvariance, using locality as the cate- particular as expressed by the lOB, BEC, WMl— gory of effect, generally corroborated the LAB, LMl-3, and contributed to highervariabilityofthesedimensions, with the separation ofthis Nigerian sample from WM1,BEC,andLM1-3 exhibitingthelarg- other West African Dasymys (Table 3). est / values, respectively (Table 2). Al- Confidence ellipses for the centroids ofthe though the principal component and anova latterpopulations, includingthe Dada sam- 424 PROCEEDINGSOFTHEBIOLOGICALSOCIETYOFWASHINGTON Table 2.—Resultsofprincipalcomponentanalysesandone-wayanovas(effect= locality)forDada(« = 23) andPanyam (n = 38) samplesofDasymys(seetextandFig. 2). Unrotated Rotated Cranialvariables PCI PCII PCI PCII /(locality)' ONL 0.92 0.29 0.88 0.34 37.0 ZB 0.90 0.04 0.68 0.52 54.8 BBC 0.79 -0.21 0.45 0.65 59.8 JOB 0.52 -0.53 0.02 0.73 37.5 LR 0.79 0.44 0.84 0.14 15.8 BR 0.74 0.29 0.81 0.24 9.9 PPL 0.88 0.29 0.86 0.32 29.3 LBP 0.59 0.11 0.40 0.27 12.2 LIF 0.71 0.29 0.76 0.22 14.9 LD 0.81 0.47 0.89 0.13 12.6 LMl-3 0.76 -0.46 0.25 0.83 119.7 WMl 0.75 -0.61 0.15 0.94 273.7 . BMls 0.84 0.04 0.51 0.47 27.7 BZP 0.77 0.25 0.41 0.24 20.2 LAB 0.79 -0.27 0.47 0.71 76.0 BEC 0.77 -0.43 0.35 0.82 194.5 Eigenvalue 0.027 0.006 0.016 0.014 % Variance 60.4 14.0 34.1 31.2 All/values significantatP < 0.01. 0.0 PC I Fig. 3. Projectionofgroup centroids onthe first two principal components forsix largest samples ofWest AfricanDasymys(seeTable3; samplecodesandsizesaregiveninMaterialsandMethods).Ellipsescorrespond to95%confidencelimitsofagroup'scentroid. EnclosedstarrepresentstheholotypeofDasymysrufulusMiller (1900). VOLUME 104, NUMBER 3 425 pie, were broadly congruent and circum- Table 3.—Results ofprincipal component analysis scribed the data point that represents the usingsixlargestOTUsofDasymys{n = 130)(seetext type specimen (USNM 83844) of Miller's and Fig. 3). (1900)D. rufulus. Althoughwe didnotper- Cranialvariables PCI PCII PCIII sonallyexaminethetypematerialofD.foxi, the cranial measurements of the holotype ONL 0.96 -0.18 0.09 (BMNH 12.4.3.15 from Panyam) supplied ZBBBC 00..7869 00..0139 00..0156 by Dr. Erik Van der Straeten clearly depict lOB 0.30 0.58 0.16 arobustskullthatfitswiththeUSNMseries LR 0.87 -0.33 0.08 from the Jos Plateau. BR 0.81 -0.24 0.13 PPL 0.92 -0.09 0.05 LBP 0.68 -0.14 -0.17 Discussion LIF 0.80 -0.15 0.22 LD 0.91 -0.27 0.07 Several ofthe commonnames applied to LMl-3 0.61 0.60 0.17 WMl Dasymys—namely, "marsh," "swamp," or 0.51 0.74 0.16 BMls 0.87 0.05 -0.01 "water" rats—underscore the regular asso- BZP 0.78 0.16 -0.59 ciation ofthese rodents with various kinds LAB 0.71 0.41 0.12 of wetland environments. Typical habitat BEC 0.68 0.37 0.16 references include stream and river banks, Eigenvalue 0.033 0.005 0.003 marshyareaswithreedbeds, swampyedges %Variance 63.5 10.4 5.9 along rivers, lush grass covering moist ground, matted semi-aquatic grasses and sedges,wetterpartsoffloodplains,andhigh- variation in habitat and habit reflects pop- land bogs (Rosevear 1969, Smithers 1971, ulational or specific differences. Sheppe 1972, Kingdon 1974). In addition In West Africa, the pattern of morpho- to preferring generally wet conditions, the logical variation supports the existence of presenceofdensegroundcoverappearscrit- two separate species, to which the closest ical to theirecological occurrence in afford- regional epithets Dasymys rufulus Miller ing protected runways and nesting sites. (1900) and Dasymys foxi Thomas (1912) Their diet is thought to consist principally apply. The marked discrimination between ofgreen vegetable matter, yet some insects these two sets of populations clearly dis- remains have been identified in stomach putes the notions that the traits of rufulus contents(Hanney 1965,Smithers 1971,Coe andfoxi are insufficiently marked for even 1975). Shaggy-furred rats appear to be subspecific recognition or that they inter- mostly terrestrial, although semi-aquatic grade imperceptibly in size along a north- habits have been noted, and their activity southcline (Rosevear 1969, Happold 1987). has been described as principally nocturnal Instead,thecraniodentaldifferentiationob- but to a limited extent diurnal (Thomas served between populations ofeach resem- 1912, Roberts 1951, Smithers 1971, De bles that demonstrated for other pairs of Graaff 1981). congeneric species inhabiting portions of Knowledge ofthelifehistoryandecology West Africa—for example those of Hybo- ofDasymys, however, remains sketchy and mys, Lemniscomys, Malacomys, and Prao- stems largely from trapping experiences re- mys (Van der Straeten & Verheyen 1978, latedinfaunalaccountsandfromcollectors' 1979, 1981, 1982; Carleton & Robbins notations on specimen tags. Thorough field 1985). studiesonthebiologyofthisgenushaveyet How these two species in West Africa re- to be conducted, a situation which leaves late to forms of Dasymys described from unanswered whether some ofthe apparent central, eastern, and southern Africa will 426 PROCEEDINGSOFTHEBIOLOGICALSOCIETYOFWASHINGTON ^- 4 Fig.4. ViewsofthecraniumandmandibleofanadultmaleDasymysfoxi(USNM403644)fromthevicinity ofPanyam,Nigeria. Approximately x1.75. require careful revisionary attention to the Africa). His enumeration ofthe differences entire genus, which is beyond the scope of between D. rufulus and D. incomtus is apt our paper. The primary descriptive htera- and emphasized the generally smaller size ture on Dasymys lends few clues. and more rufous pelage ofthe former rel- Miller (1900) contrasted his new species ative to the larger skull, robust molars, and Dasymys rufulusto D. bentleyaeand D. in- dark-toned pelage ofthe latter. These mor- comtus, but had at hand for examination phological contrasts are apparent in more only a single specimen ofthe latter (prob- recently collected examples ofD. incomtus USNM ably 1 1884fromPortNatal, South availableto us from South AfricaandZim- VOLUME 104, NUMBER 3 427 Fig. 5. Views ofthecranium andmandibleofanadultmaleDasymysrufulus(USNM 83901) from Moimt Coffee, Liberia. Approximately xl.75. babwe. The eastern distributional limits of as a close relative ofD. foxi, based princi- rufulusareuncertain;examplesthatwehave pally on the former's smaller molar rows, seen from Cameroon (AMNH specimens and instead focussed his comparisons to D. from Tongo and Ndokayo) are referable to medius from central Africa. rufulus. Our principal component analyses cor- In his description of Dasymys foxi, roborateThomas'(1912)emphasisofmolar Thomas (1912) had available for exami- size as a cardinal trait for distinguishing D. nation examples of D. rufulus, as well as foxi from D. rufulus. The distinction rests specimens ofDasymys from other regions notonlywiththegreaterlengthofthetooth- ofAfrica contained in the British Museum rows in D. foxi but also with their greater (Natural History). With only passing men- width (Tables 2, 3; Appendix 1). Breadth tion,hedismissedMiller's(1900)D. rufulus measurements across the zygoma and 428 PROCEEDINGSOFTHEBIOLOGICALSOCIETY OFWASHINGTON braincasecontribute substantiallytothero- sent a variety of taxa including incomtus bust appearance ofthe skull ofD. foxi ver- proper. Only examples ofD. nudipes seem sus that ofZ). rufulus (Figs. 4, 5), in general to typically lack a sixth pad (Hill & Carter relativelymoresothanlengthvariables(Ta- 1941, Crawford-Cabral 1983), but we have ble 2), although specimens of D. foxi do examined too few specimens ofthis taxon average larger in most cranial dimensions tojudge the constancy ofthe trait. In other (Appendix 1). Field weights and external featuresaswell—morebrightlycoloreddor- measurements ofthe skin similarly reflect sal and ventral pelage, large size (hind foot the larger size of D. foxi, but overlap in lengthtypically38to42mm), andtheusual ranges is appreciable (Appendix 1). Exam- presence ofa tiny t3 on the upper second ples ofD.foxiandD. rufulusare practically molar—nudipes represents a well-marked inseparable on the basis of pelage colora- form that probably warrants specific status tion:bothexhibitagrayishtoreddishbrown (also see Crawford-Cabral 1983 and Craw- dorsumandalightto darkgrayventerwith ford-Cabral & Pacheco 1989). varying concentrations of a creamy-buff" Available data on chromosomal varia- wash. Red tones seem more predominant tioninthegenusDasymysislimitedtostan- in the dorsum ofD. rufulus and gray in D. dard karyotypic preparations. Matthey foxi, but this impression is undermined by (1958)reportedadiploidnumberof38,with the range ofchromatic variation observed fourpairsofbiarmedautosomes, forasam- within series ofeach form. ple (number of individuals unknown) of We have not yet examined series of D. Dasymys from Ivory Coast. Later, Tranier medius to form an opinion on its degree of & Gautun (1979) obtained a similar figure similarity ofD.foxi. Inexternal andcranial forIvory CoastDasymys, 2N = 36 with six size, however, specimens ofD.foxi closely pairs of metacentrics and submetacentrics resemblethose ofZ>. incomtus, as known to based on two males. Both records are pre- us from several localities in South Africa sumablyreferabletoD. rufulus,butthevari- and Zimbabwe. The agreement is particu- ation implied by these studies should be larly strong in their possession of heavily reconfirmed using larger samples of fresh constructed,longtoothrows.LikeD. rufulus material. Theonekaryotypicexamplefrom thepelagecolorofD.foxicontrastswiththe eastern Africa, a single male from Burundi, darker, grayish-black upper parts typical of exhibited a 2N = 40, with six pairs of D. incomtus. Moreover, the texture of its biarmedautosomes(Maddalenaetal. 1989). fur seems finer relative to both D.foxi and The authors provisionally identified their D. rufulus. The taxonomic significance of specimen as D. incomtus but stressed the thesepelagedifferencesmustawaitrigorous lackofchromosomaldatabasedontrue in- evaluation ofvariation using large series of comtus from South Africa. Although these D. incomtus and determination ofits rela- studiessuggestthepotentialvalueofkaryo- tionship with forms described from eastern typic information in Dasymys systematics, and central Africa. the evidence is, as yet, too fragmentary to Miller(1900)erredinreportingfiveplan- offer insight on species boundaries. tar pads as characteristic of D. rufulus. A Unlike some other rodent species, sam- sixth pad, the hypothenar, is present and, ples ofDasymys from West Africa are not althoughsmall, isclearlydiscemable onthe restricted to a particular faunal region or dried hind feet of most specimens com- biome. Instead their geographic distribu- A prisinghis type series. total ofsix plantar tion spans several major biotic realms, in- pads also appears characterisitic ofD. foxi cluding (per Delany & Happold 1979) low- and of most other Dasymys from central, landrainforest, Guineasavanna,andSudan eastern, and southern Africa, which repre- savanna (Fig. 6). Moist living conditions

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.