BULLETIN DE L'INSTITUT ROYAL DES SCIENCES NATURELLES DE BELGIQUE, BULLETIN VAN HET KONINKLIJK BELGISCH INSTITUUT VOOR NATUURWETENSCHAPPEN, BIOLOGIE, 67: 163-186, 1997 BIOLOGIE, 67: 163-186, 1997 Systematics and zoogeography of the small mammal fauna of Cameroun: Description of two new Lophuromys (Rodentia: Muridae) endemic to Mount Cameroun and Mount Oku by W.N. VERHEYEN, 1. HULSELMANS, M.COLYN & R. HUTTERER Abstract The mountain forest archipelago of south-east Nigeria and west Cameroun is known to be an important biogeographical region. A constantly increasing number of endemic taxa is being described. The present morphometrical revision of the Lophuromys in museum collections from this mountain-range allowed us to identify and describe two new endemic species. Lophuromys roseveari sp.n. belongs craniometrically to the "unspeckled" L. sikapusi species complex and seems to have a distribution restricted to the mountain forest belt on Mt Cameroun. Lophuromys dieterleni sp.n. is morphologically and morphome- trically related to the L. jlavopunctatus species complex but has an "unspeckled" pelage. This taxon is endemic to the mountain foref:ts of Mt Oku and seems to be sympatric with L. sikapusi. Its closest relative is L. eisentrauti, endemic to the nearby Mt Lefo mountain. It is likely that new taxa will be described whenever a generic revif;ion is undertaken but the present list of endemic mammals is already sufficiently impressive to classify the mountain forests of !vItCameroun and Mt Oku as high priority regions for the conservation of African mammal diversity. Key-words: Rodentia, Muridae, Lophuromys, systematics, morphology, morphometries, zoogeography, Africa Resume La ~;uccessiondes forets de montagne du sud-est du Nigeria et de I'ouest du Cameroun constitue une region bien connue pour son importance biogeographique. En effet, Ie nombre de taxons end{~miquesdecrits de cette region est en constante augmen- tation. La presente revision morphometrique des Lophuromys des collec- tions des musees, provenant de cette chaine de montagnes, nous a p~rmis d'identifier et de decrire deux nouvelles especes end'~miques; Le Lophuromys roseveari sp.n. appartient, craniometriquement, au groupe d'especes L. sikapusi «non-tachetees», et sa distri- bution semble etre limitee a la zone des forets de montagne du Mt Cameroun. Lophuromys dieterleni sp.n. est apparente, morphologiquement et morphometriquement, au groupe d'especes L. jlavopunctatus, mais possede un pelage «non-tachete». Ce taxon est endemique des fon'!ts de montagne du Mt Oku et semble etre sympatrique avec L. sikapusi. Son plus proche parent est L. eisentrauti, espece endemique de la montagne Mt lefo, non loin de la. II est probable que de nouveaux taxons seront decrits des qu'une revision generique sera entreprise, mais la liste actuelle des mammiferes endemiques est deja suffisamment impressionnante pour que les forets de montagne des Mt Cameroun et Mt Oku soient cIassees comme hautement prioritaires pour la preser- vation de la diversite des mammiferes africains. Mots-cles: Rodentia, Muridae, Lophuromys, systematique, morpho logie, morphometrie, zoogeographie, Afrique. Introduction The biogeography of the small mammal fauna of the mountain forest archipelago of south-east Nigeria and western Cameroun was recently discussed by HUTTERER et al. (1992). They argued that only a careful systematic approach can lead to solid conclu- sions, and that "Much more sampling in these mountains and subsequent rigid systematic analyses have to be done before we can fill the empty space in the puzzle we have before us. (p. 411)". However, the small mammal collections already available from these isolated mountain forests are, at least for certain genera, sufficient to arrive at adequate taxonomic insights, on the condition that they are integrated in regional or preferably continent-wide revisions. We here describe two new Lophuromys taxa within such a context. It is becoming increasingly clear that taxonomic studies of the small mammal fauna of these isolated mountain forests will often reveal new endemic taxa. We are aware that collecting and morphologically describing new "endemics" alone will not suffice to fill in the above-mentioned "puzzle". DNA-sequencing of selected genes of adequately choosen marker-taxa will prove essential to unravel the biogeographical history of this mountain range. Material and methods Pending that such an approach becomes possible, we continue to apply the classic morphological and morpho- metrical methodology. The description of the craniolo- gical measurements, age-classes, acronyms of museums and institutes, as well as the statistical methods, follow VERHEYEN, COLYN&HULSELMANS (1996). Where needed, we provide more details in the legends of the graphs. The data sets used in the descriptions are concentrated
16
Embed
Systematics and zoogeography of the small mammal fauna of ...projects.biodiversity.be/africanrodentia/descriptions/african... · Systematics and zoogeography of the small mammal fauna
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
BULLETIN DE L'INSTITUT ROYAL DES SCIENCES NATURELLES DE BELGIQUE,
BULLETIN VAN HET KONINKLIJK BELGISCH INSTITUUT VOOR NATUURWETENSCHAPPEN,
BIOLOGIE, 67: 163-186, 1997
BIOLOGIE, 67: 163-186, 1997
Systematics and zoogeography of the small mammal fauna of Cameroun:
Description of two new Lophuromys (Rodentia: Muridae) endemic toMount Cameroun and Mount Oku
by W.N. VERHEYEN, 1. HULSELMANS, M.COLYN & R. HUTTERER
Abstract
The mountain forest archipelago of south-east Nigeria and westCameroun is known to be an important biogeographical region.A constantly increasing number of endemic taxa is beingdescribed.
The present morphometrical revision of the Lophuromys inmuseum collections from this mountain-range allowed us toidentify and describe two new endemic species.Lophuromys roseveari sp.n. belongs craniometrically to the"unspeckled" L. sikapusi species complex and seems to have adistribution restricted to the mountain forest belt on MtCameroun.
Lophuromys dieterleni sp.n. is morphologically and morphome-trically related to the L. jlavopunctatus species complex but hasan "unspeckled" pelage. This taxon is endemic to the mountainforef:ts of Mt Oku and seems to be sympatric with L. sikapusi.Its closest relative is L. eisentrauti, endemic to the nearby MtLefo mountain.
It is likely that new taxa will be described whenever a genericrevif;ion is undertaken but the present list of endemic mammalsis already sufficiently impressive to classify the mountain forestsof !vItCameroun and Mt Oku as high priority regions for theconservation of African mammal diversity.Key-words: Rodentia, Muridae, Lophuromys, systematics,morphology, morphometries, zoogeography, Africa
Resume
La ~;uccessiondes forets de montagne du sud-est du Nigeria etde I'ouest du Cameroun constitue une region bien connue pourson importance biogeographique. En effet, Ie nombre de taxonsend{~miquesdecrits de cette region est en constante augmen-tation.
La presente revision morphometrique des Lophuromys des collec-tions des musees, provenant de cette chaine de montagnes, nousa p~rmis d'identifier et de decrire deux nouvelles especesend'~miques;Le Lophuromys roseveari sp.n. appartient, craniometriquement,au groupe d'especes L. sikapusi «non-tachetees», et sa distri-bution semble etre limitee a la zone des forets de montagne duMt Cameroun.
Lophuromys dieterleni sp.n. est apparente, morphologiquementet morphometriquement, au groupe d'especes L. jlavopunctatus,mais possede un pelage «non-tachete». Ce taxon est endemiquedes fon'!ts de montagne du Mt Oku et semble etre sympatriqueavec L. sikapusi. Son plus proche parent est L. eisentrauti, espece
endemique de la montagne Mt lefo, non loin de la.II est probable que de nouveaux taxons seront decrits des qu'unerevision generique sera entreprise, mais la liste actuelle desmammiferes endemiques est deja suffisamment impressionnantepour que les forets de montagne des Mt Cameroun et Mt Okusoient cIassees comme hautement prioritaires pour la preser-vation de la diversite des mammiferes africains.Mots-cles: Rodentia, Muridae, Lophuromys, systematique,morpho logie, morphometrie, zoogeographie, Afrique.
Introduction
The biogeography of the small mammal fauna of the mountainforest archipelago of south-east Nigeria and western Camerounwas recently discussed by HUTTERER et al. (1992). They arguedthat only a careful systematic approach can lead to solid conclu-sions, and that "Much more sampling in these mountains andsubsequent rigid systematic analyses have to be done beforewe can fill the empty space in the puzzle we have before us.(p. 411)".However, the small mammal collections already available fromthese isolated mountain forests are, at least for certain genera,sufficient to arrive at adequate taxonomic insights, on thecondition that they are integrated in regional or preferablycontinent-wide revisions. We here describe two new Lophuromystaxa within such a context.It is becoming increasingly clear that taxonomic studies of thesmall mammal fauna of these isolated mountain forests will
often reveal new endemic taxa. We are aware that collectingand morphologically describing new "endemics" alone will notsuffice to fill in the above-mentioned "puzzle". DNA-sequencingof selected genes of adequately choosen marker-taxa will proveessential to unravel the biogeographical history of this mountainrange.
Material and methods
Pending that such an approach becomes possible, wecontinue to apply the classic morphological and morpho-metrical methodology. The description of the craniolo-gical measurements, age-classes, acronyms of museumsand institutes, as well as the statistical methods, followVERHEYEN,COLYN& HULSELMANS(1996). Where needed,we provide more details in the legends of the graphs.The data sets used in the descriptions are concentrated
164 W.N. VERHEYEN, J. HULSELMANS, M.COLYN & R. HUTTERER
FIG.1.1
00° N
05° N
FIG.1.2 \j
05° N
NIGERIA
07° N
06° N
CAMEROUN
04° N
0-500. 500-1000. 1000-1500. >1500
. 9°E 100E 11°E 120EFig. 1. -
Schematic representation of the geographic distribution of the collecting localities of the Lophuromys specimens used in this study. Thenumber in the symbols refers to each locality as mentioned in table 2 where coordinates and altitudes are mentioned.
1.1. The geographic distribution of the OTU's of Lophuromyssikapusi as defined in this study. The following symbolscharacterize the localities of. OTU:SANAGA N. = (NIG-CAM.N). OTU: GAB-CON-RCA.. OTU: CAM.S.The boundaries of the tropical forest are based on mapsedited by l.A. SAYERet a\. (1992).
1.2. Schematic representation of the mountain range on theborder between Cameroun and Nigeria with indication ofthe more important mountains and plateaus (based upon amap published by HUTTERERet a\., 1992)(MC): Mt Cameroun; (MK): Mt Kupe; (MM): Mt Manenguba;(ML): Mt Lefo; (MO): Mt Oku; (MP): MambillaPlateau; (AP):Adamaoua Plateau; (MG): Gotel Mts; (OP) Obudu Plateau.Ii Lophuromys roseveari sp. n.o Lophuromys dieterleni sp.n.o Lophuromys sikapusio Lophuromys eisentrauti (type locality Mt Lefo)
Systematics and zoogeography of the small mammal fauna of Cameroun 165
in annexes (App. 1-1;1-2;2-1;2-2;3;4) to this paper. Onlyskulls with fully erupted M3were retained in the analyses(age classes 1-2-3-4); specimens with severely erodedteeth were excluded (cl. 5). Since our experience withlarge series of Lophuromys sikapusi and 1. flavopunctatushas shown that sexual dimorphism is insignificant, wemade no attempt to evaluate sexual dimorphism in ournew taxa.Out of the total set of 24 available cranial measurements,19 were selected for multivariate analyses (see table I).This selection was imposed by the need to plot a numberof essential type-specimens and some geographicallyunique specimens on the multivariate graphs.Our analyses are based upon the following 5 operationaltaxonomical units (OTU's):- UGANDA : - Lophuromys ansorgei DE WINTON, 1896
- from Crater-track (Q.E.Park - Uganda)- 84 complete adult skulls
- CAM.S. : - Lophuromys sikapusi species-complex- Cameroun south of the Sanaga River- 16 complete adult skulls
For more details on the composition of these OTU's, werefer to App. 5.1. and 5.2.The corresponding metrical data sets can be obtainedthrough e-mail ([email protected]).The necessity to split the available 1. sikapusi materialfrom the study area into three separate OTU's becameapparent through a series of preliminary analyses. Wefound that all specimens from Cameroun south of theSanaga (CAM.S) appear to be craniometrically sufficientlydifferent from N.SANAGA and GAB-CON-RCA to beconsidered incertae sedis; it is even possible that they
Table I. - Recapitulation and short description of the measurements as used in this study. For a full description werefer to VERHEYENet al. (1996). Only measurements marked with * were retained for the multivariateanalyses.
NUMBER
MI
M2
M3
M 4 *
M 5 *
M 6 *
M 7 *
M 8 *
M 9 *
MlO*
Mil *
MI2 *
M13*
MI4 *
M15
M16
M17 *
M18 *
MI9 *
M20 *
M21 *
M22 *
M23 *
M24 *
ACRONYMS
GRLS
PRCO
HEBA
HEPA
PAFL
DIAl
DIA2
INTE
ZYGO
PAL A
UPTE
UPDA
MIBR
ZYPL
BNASLNAS
LOTE
CHOB
BULL
BRCA
DINC
ROHE
ROBR
PCPA
MORPHOMETRICALCHARACTERS
greatest length of skull
condylobasal lengthhenselion-basion
henselion-palation
length of palatal foramen
length of diastema
distance alveolus M I and upper incisorsmallest interorbital breadth
zygomatic breadth
smallest palatal breadth
length of upper cheekteeth
breadth of upper dental arch
greatest breadth of first upper molar
smallest breadth of zygomatic plategreatest breadth of nasals
greatest length of nasals
length of mandibular teeth
greatest breadth of choanae
length of auditory bullagreatest breadth of braincase
depth of upper incisor
mediosagittal projection of rostrum heightgreatest rostrum breadth
distance coronoid and angular processes
166 W.N.VERHEYEN,J. HULSELMANS,M.COLYN& R. HUTTERER
will prove to be related to L. ansorgei, the eastern repre-sentative of the L. sikapusi species-complex. Our resultssuggest that the Sanaga River has been an effectivezoogeographical barrier between the sikapusi populationsof GAB-CON-RCA and N.SANAGA. However, morevoucher-specimens from southern Cameroun as well asa number of karyotyped specimens and DNA sequencesare needed to confirm this.Fig.l gives an overall view of the geographical distrib-ution of the species, specimens and types involved. Thecoordinates of the localities and their approximate altitudescan be found in table 2.
The Lophuromys of Mount Cameroun
In a key to the species of Lophuromys in his excellentbook "The Rodents of West Africa", ROSEVEAR(1969:202-203) mentions a form of "brush-furred rat" from MountCameroun that he tentatively considers to be "possiblyan independent species". He even gives an extendeddescription (p.305) of the skin and skull of specimenBMNH 34.6.6.3 from Onyanza (alt. 8500 ft), the onlyspecimen at that time available to him. On p.306 hewonders why EISENTRAUT(1963) " ... amongst all thenumerous Lophuromys (36) he collected on the Cameroun
Table 2. - Alphabetical gazetteer of the collecting localities of the Lophuromys specimens included in this study. Thelocalities are followed by their co-ordinates and approximate altitudes (m). The numbers preceding thelocalities refer to fig. 1 illustrating the distribution of the species.
Systematics and zoogeography of the small mammal fauna of Cameroun 167
Mountain, has not recorded any markedly long-hairedexamples corresponding to the British Museum specimen,since he generally draws attention to such character ifany skin displays it" .EISENTRAUT(1973), reacting to these remarks, admits that,after re-examining the rather important Lophuromys serieshe collected on Mount Cameroun, all of his specimensunmistakably show a long, soft and dense dorsal pelageaveraging 13-14mm in hairlength. He concludes: "Soferndie Langhaarigkeit nicht als blosse phiinotypischeErscheinung und damit als Anpassung an das kiihlereMontanklima auszusehen ist, diirfte es sich wohl urn eineMontanrasse von sikapusi handeln, da die Masse und dieSchiidelmerkmale weitgehend sikapusi-Charakter zeigen" (p.85).DIETERLEN(1976, 1979:296) agrees with these observa-tions and adds, after close scrutiny of the specimen BMNH34.6.6.3. from Onyanza, that (I) the external measure-ments given by ROSEVEARare probably the result ofshrinkage of the skin due to preparation, (2) the otherpossibly important characters invoked by ROSEVEARarenot relevant when the other Lophuromys-material of MtCameroun is taken into consideration. He arrives at theconclusion" ... dass die Diskussion urn dieses Stiick mitRosevears (1969) eigenen Siitzen abgeschlossen werdenkann ..." it would appear that this form agrees more withsikapusi than nudicaudus, and it may prove to be nothingmore than a montane race of that species. In most respectsit is just a little smaller.We were able to study not only the above mentionedLophuromys specimens collected by Eisentraut and hiscolleagues on Mount Cameroun (West-Afrika-Reise1966/67)but also material from Mount Cameroun availablein other museums, and studied these within the scope ofa bigger continent-wide revision of the Lophuromyssikapusi species-complex. This comparison leads us tothe conclusion that the Mount Cameroun population differssufficiently to warrant species rank.
Description of Lophuromys roseveari sp.n.
HOLOTYPE
ZFMK 69.297; ad. female; round skin; skull complete;collected by M. EISENTRAUT(18 march 1967) at MUSAKE(04.08N - 09.12E) on the slope of Mount Cameroun atan altitude of 1850-2200m; collecting number 656.
PARATYPES
15 specimens, all collected on the slopes of MountCameroun but at different localities and altitudes.
The holotypes and paratypes were collected on the slopesof Mount Cameroun at altitudes between approximately1200 and 3000m.
The collector's notes indicate that the specimens weremostly caught in "grassland" or "grassland near forest-fringes" or "among small patches of wood" or in "highforest". In a few instances "native farm" , "garden","deforested zone in forest", "tree-plantation" and"elephant-grass" are mentioned.
ETYMOLOGY
We dedicate this new species to D.R.ROSEVEARin appre-ciation of his taxonomical and zoogeographical insightinto the Western and Central African small mammal fauna.
DIAGNOSIS
L. roseveari is a new species of "unspeckled and short-tailed brush-furred rat" belonging to the sikapusi-ansorgeispecies-complex. Consequently, it is differentiatedcranially and dentally from the nudicaudus-huttererispecies-complex, from L. rahmi, the smaller endemicspecies from the mountainous region of Kivu (Zaire),
168 W.N.VERHEYEN,1. HULSELMANS,M.COLYN& R. HUTTERER
A B c
Fig. 2. - Schematic drawings of the ventral side of adult crania of (A) Lophuromys roseveari (ZFMK69.297-type), (B) Lophuromyssikapusi (RUCAIO.051-Gabon) and (C) Lophuromys dieter/eni (ZFMK69.289-type).The skulls are aligned with the anterior borders of the first molars.
4 5 6
Fig. 3. -Comparative drawings of the right maxillary teeth of Lophuromysroseveari (4) and some specimens representative for theLophuromys sikapusi species-complex:
Fig 5. - Comparative drawings of the right maxillary teeth ofLophuromys dieterleni (2) and some selected specimensfrom the Lophuromys flavopunctatus species-complex:
Systematics and zoogeography of the small mammal fauna of Cameroun 169
and from 1. eisentrauti, the smallish endemic species ofMt Lefo in the Cameroun highlands related to thejlavopunctatus species-group. Within the sikapusi-ansorgeispecies-complex, 1. roseveari can immediately be recog-nized by its (I) remarkably slender, narrow skull, (2)somewhat larger bullae, (3) bigger ears and (4) somewhatsofter and longer pelage (hair length on dorsum (::I:15mm).
Discussion of the morphological characters(table 3; figs 2 and 3)
A statistically valid comparison of the external metricalcharacters (table 3) is not possible since these data havebeen collected by different researchers almost certainlyusing differing measuring methods. However, we canconclude safely that, whereas most of the externalmeasurements suggest that 1. roseveari is just a fractionsmaller than 1. sikapusi, its ear length scores clearlyhigher.Concerning the pelage, we confirm the findings ofROSEVEAR(1969) who notes that the middorsal region hasa softer feel than in 1. sikapusi.Confusion with 1. nudicaudus can be excluded for bothexternal measurements and pelage texture (VERHEYENetaI., 1996).The general habitus of the skull (fig. 2) shows a clearresemblance to 1. sikapusi-ansorgei. However, the skullof 1. roseveari has a notably more slender and fragilebuild but in profile has the same rostrum and skull heightas 1. sikapusi-ansorgei.In addition, the morphology of the molars of the upperand lower jaw (fig.3) clearly relates 1. roseveari to the1. sikapusi-ansorgei species-complex.Compared to typical sikapusi and ansorgei the skull of1. roseveari has
(1) strikingly narrower choanae;(2) a more reclining anterior border of the zygomaticplates;(3) weakly developed supra-orbital ridges and notch;(4) more slender zygomatic arches;(5) a relatively longer and more slender processus angularisof the mandibula.
Discussion of the craniometrical data
UNIYARIATE ANALYSIS (table 4 and 5)
The basic statistics show the variation coefficients to befar below 10% except for the "choanae breadth" (nr.18)which is an unreliable measurement.When comparing 1.sikapusi samples from GAB-CON-RCA, N.Sanaga and the sample from Mt Cameroun bymeans of one way ANOYA, only 8 out of 24 variablesare not significantly different between the samples (MI,M2, M3, M5, M6, M8, M16, M24). Most of thesevariables are skull-length measurements.The other 16 variables are highly significantly different.In most cases, tested by SNK a posteriori tests, the threesamples are significantly different, the Mt Camerounsample being the smallest (M9, MIO, MIl, M12, M13,M14, M15, M17, M18, M21, M22). These are mostlywidth-measurements. In four more cases the MtCameroun sample scores the lowest, but differs neitherfrom sikapusi N.Sanaga (M7, M20, M23) nor fromsikapusi GAB-CON-RCA (M4). For only one variable(M19 = length of the bulla) the Mt Cameroun sampleis the largest.Summarizing, the Mt Cameroun is characterized by aclearly narrower skull and somewhat heavier bullae.
Table 3. -External measurements of Lophuromys roseveari sp. n. compared with populations of adult (cI.2-3-4)Lophuromys sikapusi captured in Mopoyem (Ivory Coast) and Odzala (Congo).(W: weight; Tol: total length; HB: head + body length; Tl: tail length; HF: hindfootlength-nail; El: earlength).
W Tol HB Tl HF EI
L. roseveari 63.5 193.5 127.5 66.0 22.5 18.5 xMt. Cameroun 49 - 88 164 -219 104 - 141 50 - 78 20,0 - 25,1 16 - 21 min - max
24 29 32 29 32 28 n
L. sikapusi 68.5 206 136.8 69.5 23.9 16.2 xMopoyem 50 - 95 189 - 231 122 - 152 60 - 82 22 - 25 15 - 18 min - max
165 116 167 116 132 112 n
L. sikapusi 68.7 198.3 130.1 68.1 22.1 15.8 xOdzala 49 - 92 179 - 222 112 - 145 59 - 80 20,4 - 24,3 14,4 - 17,5 min - max
24 20 24 20 24 20 n
170 W.N. VERHEYEN, J. HULSELMANS, M.COLYN & R. HUTTERER
Table 4. - Basic statistics of Lophuromys roseveari sp. n. compared with the basic statistics of Lophuromys ansorgei(OTU=UGANDA), Lophuromys sikapusi (N.SANAGA) and Lophuromys sikapusi (GAB-CON-RCA).Only specimens from age classes 2-3-4 were retained.
For the complete set of metrical data of L. roseveari we refer to App. 2.1. and 2.2.
The composition of the craniological series of the other species can be consulted in App. 5.1. and 5.2.;the entire craniometrical data sets can eventually be obtained through e-mail.
In this analysis MI, M3, M9, M16, M18, M24 wereexcludedto maximize the number of specimens, especiallyin the Mt Cameroun sample. After forward selection alsoM5 was discarded. Wilk's Lambda is significally lowerthan I, indicating a good discrimination between thegroups. 58% of the total variation is expressed in root I.The scatter diagram of root 1 versus root 2 shows thethree groups to differentiate clearly, the Mt Cameroungroup being separated along root 2. When looking at thefactor structure matrix M 19 (bulla length) is mostpositively correlated with root 2 : this measure is thelargest in the Mt Cameroun sample (see univariateanalysis). Most negatively correlated with root 2 are MI7(lower teeth length), M21 (upper incisor depth), MIl(upper cheekteeth length) and MI3 (first molar breadth).The squared Mahalanobis distances show the Mt Cameroungroup to be somewhat more separated from the twosikapusi groups.A number of selected specimens are plotted on fig.4 inorder to clarify their position in relation to L.roseveari.
Both type specimens (L.sikapusi and L.afer) clearly fallin the N.SANAGA sector, suggesting that this OTU willprove to be rather closely related to typical west AfricanL.sikapusi. The CAM.S population identifies itself clearlyas sikapusi-related and overlaps with both N. SANAGAand GAB-CON-RCA.The skulls from Manenguba (ZFMK 69.294) and Mt Oku(ZFMK 91223) coming from high altitudes ( 1800m and2050m respectively) fall well within the sikapusi-rangeand are thus not related to L.roseveari. On the other hand,the skull SMNS 5355 from Musake that we judged tobe somewhat aberrant and in our opinion not to beincluded in our paratypical series of L.roseveari, whenplotted, situates itself well within the L.roseveari range.Finally, skull SMNS 41328 from Gangirwal (Gotel Mts;alt. 2300m) situates itself morphologically and metricallyclearly within L.sikapusi (N.Sanaga).Also in this analysis, it is more than obvious that the MtCameroun specimens are well differentiated from theother sikapusi groups, their probability of correct classi-fication being 100%.
Systematics and zoogeography of the small mammal fauna of Cameroun 171
Table 5. -Results of ANOVA (one way) analyses performed on 24 craniodental measurements of age classes (2+3+4)of Lophuromys roseveari sp. n. and Lophuromys sikapusi (OTU:GAB-CON-RCA, N.SANAGA). A posterioritests (Student, Newman, Keuls) are used to evaluate the differences between OTU's (SOKAL& ROHLF,1969)indicated by the column <OTU>.
172 W.N.VERHEYEN,1. HULSELMANS,M.COLYN& R. HUTTERER
Conclusions
The Lophuromys rats living on the higher flanks of MtCameroun are sufficiently differentiated to be consideredspecifically distinct from their sikapusi relatives inhab-iting both the lowland forests and also the rest of themontane forest-archipelago of the surrounding Nigeria-Cameroun region. However, in view of the considerablemorphometrical differences also encountered between thestudied OTU's (GAB-CON-RCA - N.SANAGA), weexpect that the exact systematical position of the MountCameroun form will only become clear within a continent-wide revision of the Lophuromys sikapusi-ansorgeicomplexwhich should include also DNA-sequences andcaryological data.Finally, we point out that the fauna of Mount Camerounis already characterized by some endemic mammals suchas Sylvisorex morio (Gray, 1862), Crocidura eisentrautiHeim de Balsac, 1957 (HUTTERER,1993), and thesubspecies Otomys irroratus burtoni (Thomas, 1918) whichmay equally represent a good species (DIETERLENandVANDERSTRAETEN,1992). Lophuromys roseveari is thusthe fourth mammal taxon known only from this moun-tain.
The Lophuromus of Mount Oku
In his publication on the zoogeographical importance ofMount Oku (Bamenda-Banso highlands of west Cameroun)EISENTRAUT(1968) mentions that he and his collaborators
7
6
5
4
3
2
collected five Lophuromys on Mount Oku during his 1966-67 expedition to Cameroun. He provisionally assignedthese specimens to L. sikapusi subsp. indicating, however,that they are characterized by somewhat aberrant skullcharacteristics when compared with typical L. sikapusi.EISENTRAUT(1973) expatiates over this statement bycomparing this Oku material also to his Mount Camerounseries, by adding a number of external and cranialmeasurements, and arrives at the conclusion that hisMount Oku material is clearly differentiated. He adds,however, that the two specimens collected in Manengubaclosely resemble his Oku material and thus suggest aclose relationship between both samples.DIETERLEN(1979:296), reviewing the same material,considers the differences described by EISENTRAUT(1973)to be very slight when compared to the other sikapusiforms. On the same page he describes the new taxoneisentrauti from Mount Lefo (Bamenda highlands) thathe considers to be a subspecies of L. sikapusi.We finally draw attention to the paper by HUTTERERetal. (1992) where they discuss the literature concerningthe Lophuromys of the forest islands of eastern Nigeriaand adjacent Cameroun (DIETERLEN,1979; HAPPOLD,1987;HUTTERER& JaGER,1982; EISENTRAUT,1975). They arriveat the conclusion that the observed size differencesbetween the smallish Mount Lefo specimen and theLophuromys representatives from Mt Cameroun, Rumpi-Hills, Mt Kupe, Mt Oku, Gotel Mts and the AdamaouaPlateau are important enough to elevate the Mount Lefopopulation to full species rank as Lophuromys eisentrautiDIETERLEN,1979.
L.ROSEVEARI
[] MtCarneroon
1:1 TypeL.SIKAPUSI
[] Gab.Con.RCA
o N.Sanaga+ Types (plot)PLOT. Carn(S). MtOku. Musake... Manenguba
....................
SIKAPUSI. . ............................. . .
...................................
root 1
2 7 83 4 5 6
FigA. - Canonical analysis on a selected data set of Lophuromys roseveari sp.n. (MT CAMEROUN),Lophuromys sikapusi (S.SANAGA)and Lophuromys sikapusi (GAB-CON-RCA)
0
-1
-2
-3
-4
-5
-6-6 -5 -4 -3 -2 -1 0
Systematics and zoogeography of the small mammal fauna of Cameroun 173
Description of Lophuromys dieter/eni sp.n. PARATYPES
HOLOTYPE 4 specimens from the same locality as the type-specimenand collected by EISENTRAUTbetween 17 and 31 January1967.
ZFMK 69.286 (ad. female; skull + skin; coIl. nr. 459) -ZFMK 69.287 (ad. female; skull + skin; coIl. nr. 347) -ZFMK 69.288 (ad. male; skull + skin; coIl. nr. 369) -ZFMK 69.290 (ad. male; skull + skin; coIl. nr. 389).
ZFMK 69.289; ad. male; round skin; skull complete;collected by M. EISENTRAUT(21 January 1967) on theborder of the crater-lake of Mount Oku (Lager IV; 06.12N- 1O.32E;alt. 2100m) in the Bamenda-Banso highlands;collecting number 384.
Table 6. - Summary of the main results of the discriminant function analyses on a selected craniometrical data setbetween Lophuromys roseveari sp.n. and Lophuromys sikapusi (GAB-CON-RCA) and Lophuromys sikapusi(N.SANAGA).
"Wilk's Lambda = 0,0679 F (34,206 ) = 17,1836 p<O,OOOO
174 W.N. VERHEY EN, J. HULSELMANS, M.COLYN & R. HUTTERER
TYPE LOCALITY
The holotype and paratypes were collected on Mt Okuat 2100m altitude in the rather narrow high forest beltthat encircles the crater lake.
ETYMOLOGY
We name this new species after our friend and colleagueDr. Fritz DIETERLENto underline the importance of hisecological work on the small mammal fauna of CentralAfrica and more specifically on the genus Lophuromys.
DIAGNOSIS
Externally L. dieterleni is an "unspeckled and short-tailedbrush furred" rat. However, the study of the skull andteeth demonstrates that it is neither related to the sikapusi-ansorgei species-complex nor to the nudicaudus-huttererispecies-complex, nor to L. rahmi. Cranially, it is clearlyrelated to the L. flavopunctatus species-complex and thusforms part of the so called "speckled and short-tailedbrush furred" rats.Craniometrically it is easily distinguished from itsgeographically closest relative L.eisentrauti (Mt Lefo) andfrom a representative population of L. flavopunctatuslaticeps THOMAS& WROUGHTON,1907from Kivu (Zaire).
Discussion of the morphological characters(figs 2 and 5)
A straightforward description of the differences basedsolely on morphological characters between the skulls ofthe sikapusi and flavopunctatus species-groups remainsdifficult. Generally speaking, the skulls of the sikapusi-group are somewhat bigger and heavier than typicalflavopunctatus (including dieterleni); on the other handflavopunctatus skulls and consequently dieterleni skullshave a somewhat broader silhouette and a shorter rostrum.A similar difficulty is encountered when comparing theupper-tooth morphology (see figs. 3 and 5) and yet thereis an unmistakable but subtle difference in the form andstructure of the MI. A conspicuous difference in MImorphology is that inflavopunctatus-related species thereis often a clear ridge between t6 and t9 which is almostalways absent in the sikapusi-group.
Discussion of the craniometrical data
UNIVARIATE ANALYSIS (table 7)
When comparing the small sample (n=5) of Mt Okuspecimens to a large sample (n=118) of Lflavopunctatus
from Kivu, most minima and maxima of Mt Oku spe-cimens fall within the observed range for Lflavopunctatus,except for some measurements (M8,M9,MI4,M2I).Despite the small size of the Mt Oku sample, somevariables indeed differ significantly between the twosamples. These measures include the longer snout (M2and M4), the wider orbital region and larger zygomaticplate (M8, M9, MIO, MI4), the heavier upper incisors(M2I) and the shorter upper molar row (M 11).When compared to the type-series of L.dieterleni, theskull of the type of L.eisentrauti described from MountLefo proves to be clearly smaller in all the observedmeasurements except for the choanal breadth (M18) whereit scores higher. Although the type of L.eisentrauti is ayoung specimen (age class 1), most of these observeddifferences cannot be explained by growth alone.
MULTIPLE DISCRIMINANT ANALYSIS (table 8, fig. 6)
In order to maximize the number of specimens, a set of17 variables on 279 specimens were analysed, includingtwo samples of L.sikapusi, one from north of the SanagaRiver and one from south of the Sanaga, one sample ofL.ansorgei from Uganda and one sample of Lflavopunc-tatus from Kivu. The discriminant scores of the specimensfrom Mt Oku were calculated a posteriori and plotted onthe scattergram.Wilk's Lambda is significantly different from 1,approaching zero, indicating a good discrimination betweenthe groups. Figure 6 shows a clear separation of the threespecies in the plane of rootl/ root 2, which expresses91% of total variation, the two samples of L.sikapusi forthe greater part overlapping each other. The Mahalanobis'squared distances confirm the more distant positions ofL.flavopunctatus and of L.ansorgei with respect toL.sikapusi.When plotting the Mt Oku specimens, four of the fiveclearly fall outside the 95% equal probability ellipses ofthe groups concerned, the fifth specimen lying on theellipse of L.sikapusi (S.Sanaga). Another Mt Oku specimen(ZFMK 91.223) showing unmistakable skull and teethcharacteristics of L. sikapusi (figs.3.6 and 5.2) clearlyplots within the area of that species. The types ofLflaticeps and of Lfzena fall within the flavopunctatusarea. The type of L.eisentrauti lies clearly outside all ofthe groups considered but is closest to L.dieterleni.Since the Mt Oku specimens were not incorporated inthe analysis because of the small sample size it is difficultto interprete these results in terms of canonical coeffi-cients or of factor structure. However, the Mt Oku sampleis differentiated from Lflavopunctatus along root 2, whichis influenced most by interorbital breadth on the positiveside and by the greatest rostrum breadth on the negativeside. This seems to confirm the trend of the univariateanalysis. Future introduction in the analysis of a biggerpopulation of Mt Oku-specimens as a separate group maywell change the factor structure.
Systematics and zoogeography of the small mammal fauna of Cameroun 175
Fig. 6. -Canonical analysis on a selected data set of Lophuromys ansorgei (OTU = UGANDA), Lophuromys sikapusi(OTU = GAB-CON-RCA and OTU = N.SANAGA) and Lophuromys jlavopunctatus laticeps (OTU = KIVU)in order to situate by plotting Lophuromys dieterleni sp. n. and some possibly related types and geographi-cally critical specimens.
Table7.- Basic statistics of Lophuromys dieterleni sp.n. compared with the basic statistics of Lophuromys jlavopunc-tatus laticeps (Kivu province - Zaire).On the right side the results of the T-Student test are shown.For the complete set of metrical data of 1.dieterleni we refer to App.4.The composition of the craniological series of the 1. jlavopunctatus laticeps population can be consultedin App.5.1. The entire craniometrical data set can eventually be obtained through e-mail.
L.flavopunctatus Kivu L.dieterleniMtOku
Variables MEAN N MIN MAX sm CV% MEAN N MIN MAX sm CV% TStud DF P SIGN
o . ..I.....o::.Y...AQo.oy.:.:.::':-:..L\.....A...'..........,...1 . Mt Oku (plot)
-1 . ./.; ....'fJ'>,W..O ..:1 .J A. .75... ..... ;'\.. I 0 Type (plot)
-2
-3
-4I
-51.....
root 1-6
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
176 W.N.VERHEYEN,J. HULSELMANS,M.COLYN& R. HUTTERER
Table 8.- Summary of the main results of the discriminant function analyses on a selected craniometrical data setbetween Lophuromys jlavopunctatus laticeps (KIVU) - Lophuromys sikapusi (GAB-CON-RCA) - Lophuromyssikapusi (N.SANAGA) - Lophuromys ansorgei (UGANDA).
Conclusions Both EISENTRAUT(1968, 1973) and DIETERLEN(1979)overstressed the "unspeckledness of the pelage" and under-estimated the importance of the skull and molar charactersthat were correctly observed and described by the formerauthor.
In our opinion L. dieterleni is an endemic species probablyrestricted to the narrow high forest-belt encircling thecrater-lake of Mt Oku and belonging to the L. jlavo-punctatus species-group in spite of its "unspeckled" dorsalpelage. The specimen ZFMK 91223, collected by local
The brush-furred rats living in the high forest surroundingthe crater-lake of Mt Oku occupy a somewhat confusingposition in the "classical" taxonomy of the genusLophuromys.Indeed, on the one hand their "unspeckled"dorsal pelage suggests a closeness to the sikapusi-speciesgroup, but on the other hand they possess a skull andmolar morphology characteristic of the jlavopunctatus-specIes group.
Systematics and zoogeography of the small mammal fauna of Cameroun 177
people in a village at the foot of Mt Oku on 5 January1991, clearly belongs to 1. sikapusi and shows that itpossiblyis sympatric with 1. dieterleni. Also the specimenswe have seen from Manenguba are clearly 1. sikapusi.Weconsider 1. eisentrauti from Mt Lefo to be the speciesclosest to our new species 1. dieterleni. Indeed, both taxaare"unspeckled" but with skull and molar characters typicalfortheflavopunctatus-group.It is also worthwhile mentioningthat both type localities are separated by a stretch of 40kmof montane grassland and that both type specimens werecollected in isolated patches of montane high forest.By inclusion of Lophuromys dieterleni, the mammal faunaof Mt Oku contains a number of endemic taxa, namelyone genus (Lamottemys), 4 species (Lamottemys okuensisPetter, 1986, Lemniscomys mittendorfi Eisentraut, 1968,Hylomyscusgrandis Eisentraut, 1969, Lophuromys dieterlenisp.n.), and one subspecies (Chrysochloris stuhlmannibalsaciLamotte & Petter, 1981). Other species (MyosorexokuensisHeim de Balsac, 1986, Sylvisorex camerunensisHeimde Balsac, 1968, Sylvisorex isabellae Heim de Balsac,1968,Praomys hartwigi Eisentraut, 1968, Hybomys eisen-trautiVan der Straeten & Hutterer, 1984, Otomys occiden-taUsDieterlen & Van der Straeten, 1992) are shared bysome neighbouring mountain forests only (EISENTRAUT,1968; LAMOTTE & PETTER, 1981; PETTER, 1986; HUTTERER
et aI., 1992). It is likely that more new taxa will bedescribed in the future, but the present list of localmammals is impressive enough already to classify theMt Oku forest as a high priority region for the conser-vation of African mammal diversity.
Acknowledgments
We wish to express our gratitude to the curators of themammalcollections of a number of museums who allowedus to study the type-specimens and collections in theircare.The following colleagues have been most helpful andunderstanding:P.JENKINS(British Museum of Natural History, London,UK)G. MussER (American Museum of natural History, NewYork,USA)P.MYERS(University of Michigan, Museum of Zoology,Ann Arbor, USA)B. PATTERSON(Field Museum of natural History, Chicago,USA)M. TRANIER(Museum d'Histoire Naturelle, Paris, France)W.VANNEER(Koninklijk Museum voor Midden-Afrika,Tervuren, Belgium)Our gratitude goes further to T. DIERCKX,A. FONTAINE,R. VANTICHELENand N. WOUTERSfor their technicalassistance.The support of the European Community Commission(ECOFAC and BioFAC DGVIII projects) and AGRER(Brussels) allowed us to do our fieldwork at the ECOFACand BioFAC sites in Congo and Gabon.Finally, this work was supported by the EK.EO. (Grant
2/0004/91/N) of the National Foundation for ScientificResearch of Belgium (Brussels).
References
BATES, G.L., 1905. Notes on the mammals of southernCamerouns and the Benito. Proceedings Zoological SocietyLondon: 65-85.
DIETERLEN, F., 1976. Die afrikanische MuridengattungLophuromys Peters, I 874: Vergleiche an Hand neuer Daten zurMorphologie, Okologie und Biologie. Stuttgarter Beitriige zurNaturkunde, Ser.A, 285: 1-96.
DIETERLEN, F., 1979. Beitrage zur Kenntnis der GattungLophuromys (Muridae; Rodentia) in Kamerun und Gabun. Bonnerzoologische Beitriige, 29: 287-299.
DIETERLEN,F. & VAN DER STRAETEN,E., 1992. Species ofthe genus Otomys from Cameroon and Nigeria and theirrelationship to East African forms. Bonner zoologische Beitriige,43: 383-392.
DOWSETT,R.l ed., 1989. A preliminary natural history surveyof Mambilla Plateau and some lowland forests of eastern Nigeria.Tauraco Research Report, I: 1-56.
EISENTRAUT,M., 1963. Die Wirbeltiere des Kamerungebirges.Parey, Hamburg und Berlin.
EISENTRAUT,M., 1965. Rassenbildung bei Saugetiere und Vogelnauf der Inseln Fernando Po. Zoologischer Anzeiger, 174: 37-54.
EISENTRAUT,M., 1968. Die tiergeographische Bedeutung desOku Gebirges im Bamenda-Banso-Hochland (Westkamerun).Bonner zoologische Beitriige., 19: 170-175.
EISENTRAUT,M., 1970. Eiszeitklima und heutige Tierverbreitungim tropischen West-Afrika. Umschau in Wissenschaftund Technik,70-75.
EISENTRAUT,M., 1973. Die Wirbeltierfauna von Fernando Poound Westkamerun. Bonner Zoologische Monographien, 3: 1-428+ 5 pis.
EISENTRAUT,M., 1975. Weitere Beitrag zur Saugetierfauna vonKamerun. Bonner zoologische Beitriige, 26: 76-93.
ELLERMAN,J.R., 1941. The families and genera of living rodents.With a list of named forms (1758-1936) by R.W.HAYMANandG.W.C.HOLT,vo1.lI, Family Muridae, London, 690 pp.
FEILER,A., 1988. Die Saugetiere der Inseln im Golf von Guineaund ihre Beziehungen zur Saugetierfauna des westafrikanischenFestlandes. Zool. Abhandlungen Staatkundiges Museum TierkundeDresden , 44: 83-88.
GOOD, A.l, 1947. Les rongeurs du Cameroun. Bulletin de laSociete d'Etudes Camerounaises , no 17-18: 5-20.
GRAY, J.E., 1862. List of Mammalia from the CamerounMountains, collected by Capt. BURTON,H.M. Consul, FernandoPo. Proceedings of the Zoological Society of London, 180-181.
HAPPOLO,D.C.D., 1987. The mammals of Nigeria. ClarendonPress, Oxford, 402 pp.
HILL, lE., 1982. Records of bats from Mount Nimba, Liberia.Mammalia, 46: 116-120.
HUTTERER, R. & JOGER, u., 1982. Kleinsauger aus demHochland von Adamaoua, Kamerun. Bonner zoologischeBeitriige,33: 119-132.
r
178 W.N. VERHEYEN, 1. HULSELMANS, M.COLYN & R. HUTTERER
HUTTERER,R., DIETERLEN,F. & NIKOLAUS,G., 1992. Smallmammals from forest islands of eastern Nigeria and adjacentCameroon,with systematical and biogeographical notes. BonnerzoologischeBeitriige. 43: 393-414.
HUTTERER,R., 1993. Order Insectivora. - Pp. 69-130, inMammal Species of the World, a taxonomic and geographicreference, second ed. (D.E. Wilson and D.M. Reeder, eds.).Smithsonian Institution Press, Washington D.C., 1206 pp.
NIKOLAUS,G. & DOWSETT,R.I., 1989. Small mammals collectedin the Gotel Mounts and on the Mambilla Plateau, easternNigeria. TauracoResearch Report, I: 42-47.
PETTER,F., 1986. Un rongeur nouveau du Mont Oku (Cameroun)Lamottemysokuensis, gen. nov., sp. novo (Rodentia, Muridae).Cimbebasia,ser. A8 (12): 97-105.
SAYER,lA., HARCOURT,C.S. & COLLINS,N.M., 1992. TheConservation Atlas of Tropical Forests. AFRICA. MacMillan,U.K., 288 pp.
VERHEYEN,W.N., COLYN,M. & HULSELMANS,1, 1996. Re-evaluation of the Lophuromys nudicaudus HELLER,1911species-complex with a description of a new species from Zaire (Muridae- Rodentia). Bulletin de /' Institut royal des Sciences naturellesde Belgique, biologie. 66: 241-273.