Altitudinal distribution and anthropogenic influence on small mammal assemblages on Mount Kupe, SW Cameroon

2009·1.7 MB·

by  Diverse

© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
Bonnerzoologische Beiträge Band 56 F^eft 3 Seiten 159-173 Bonn, September2009
Altitudinal distribution and anthropogenic influence on small mammal
assemblages on Mount Kupe, SW Cameroon
Christiane Denys', Alain Didier Missoup' -, Bartlielemy Tchiengue-\ Gaston Achoundong\ Atanga
Ekobo-*, Charles Felix Bilong Bilong-, Dieudonné Massoma Lembé'' & Violaine Nicolas'
' Muséum national d'Histoire naturelle, Dept. Systématique et Evolution, Paris, France
2Department ofBiology ofAnimal Organisms, University ofYaoundé I, Yaoundé, Cameroon
HerbierNational du Cameroun, Yaoundé, Cameroun
4WWF, Coastal Forests Programme, Limbé, Cameroon
5Département de Biologie des Organismes Animaux, Université de Douala, Douala, Cameroun
Abstract.WeconductedataxonomicalinventoryofsmallmammalbiodiversityonMountKupe, SWCameroon,these-
cond inventory on that mountain after M. Eisentraut's more than 50 years ago. Our survey yielded a total catch of 19
species ofmammals: 16 rodents, two bats, and one shrew. Foreach species we summarize data on natural history and
discusstaxonomicalanddistributionalaspects.Weobservedadifferencebetweentherodentassemblagesofthelowland
(below 1000 m) (including Praomys misonnei) and the upper sub-montane and montane forest (1500-2000 m) (inclu-
dingHylomyscus cf. walterverheyeiii, identified forthe first time from the Cameroon Mountains). This seems to corre-
spondtopreviouslydescribed vegetation changesbetween 1000 and 1500 m. The inhabited andcultivated zones atthe
baseofMt. Kupeyieldedeightwidely-distributedrodentspecies.Anumberoftaxonomicalproblems concerningmon-
tane forest speciesremain tobe solved.
Keywords.Africa, Cameroon, Chiroptera, Rodentia, Soricomorpha, taxonomy, diversity.
INTRODUCTION
1.
Tropical regions are well known to have a higher biodi- but the base and the top hold a higher diversity than the
versity than temperate regions. The forests ofwest-cen- intermediatezones (Brown 2001; Lomolino2001). Oth-
tralAfricahavebeendescribedasoneofthetropical "hot er authors found a continuous decrease ofdiversity with
spots" (Myers et al. 2000; Kuper et al. 2005). African elevation(Diazetal. 2002; Kasangakietal. 2003)orcon-
mountains seem to harboura very important endemic di- trastingpatterns in differenttaxonoiuic groups (Hofer et
versity and are of exceptional conservation importance al. 2000). However, a positive correlation between
(HeimdeBalsac 1968; Eisentraut 1973; Gentry 1992; speciesturnoverandelevationwasrecentlyevidencedfor
Stanley et al. 1998, 2005; Kasangaki et al. 2003; Her- rodentsofdifferenttropicalmountains(Mena&Vasquez-
rmann et al. 2005; Burgess et al. 2007). Both lowland DOMINGUEZ 2005). The authors considered climate, pro-
and mountain forests are presently subject to intense de- ductivity and landscape heterogeneity as well as histori-
forestation. Several authors have emphasized the role of cal factors as possible explanations forelevational gradi-
AfricanforestrefiigiainpromotingspeciationduringQua- ents.
temary climatic changes (Maley 1987, 1996; Querouil
et al. 2003, Plana et al. 2004; Behnin 2006, Huhndorf There are many other vertebrate groups for which zona-
etal. 2007). However, the knowledge ofthe mechanisms donhasbeendescribedasafunctionofaltitude inAfrican
by which so many species arise and differentiate in such mountains (e.g. Eisentraut 1968; Amiet 1987; Stuart
regions depends directly on aprecise characterization of et al. 1987; Kityo & Clausnitzer 2001; Kasangaki et
theirbiodiversity in relation to altitudinal and vegetation al. 2003) but the observed changes seem to vary locally
shiftsalongelevational gradients.Thealtitudinaleffecton andtheircausesarenotwellunderstood. It is well known
small mainmal diversity is not well known and available thatthecompositionofplantcommunities changesalong
studies differ in theirresults. According to some authors altitudinalgradientsbuttherespectiverolesofvegetation
species-richnessisnotuniformlydistributedonmountains. oravailable streamsversusaltitudinal change remain un-
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
160 ChristianeDenys et al.: Small mammales ofMt. Kupe, Cameroon
clear (Hofer et al. 2000). Because most terrestrial small With this infonnation as a background we concentrated
mammal species are primary consumers or have limited our small mammal survey on the montane and sub-mon-
dispersal capacities(restrictedrangesand/orhabitats)they tane species. Our study should be ofconservation value
aregoodmodels forstudyingaltitude-diversitygradients. andprovidefurtherbiogeographicalhypotheses. Inthispa-
per we describe the composition of the small mammal
Situated in the Central African forests blocks of the community ofMt. Kupe along an altitudinal gradient. In
Guineo-Congolesehotspotofbiodiversityandextending ordertobetterdescribetheMt. Kupe sinallmammalcom-
for about 1600 km, the Cameroon Volcanic Line (CVL) munity we made a morphological and molecular inven-
is a succession ofhigh volcanoes and plateaus ranging in tory ofsmall mammal species to accurately describe the
age between 30 and one Myr (Ubangoh et al. 1998). At degreeofendemism oftheKupefauna,andeventuallyal-
a distance ofabout 3000 km from the Western Rift Val- so look at a possible altitudinal difference between sub-
ley ofEastAfrica, it represents the second highest range montane and montane communities. This should further
ofmountains inAfrica, with MountCameroon at4100m allowusto evaluatethe importanceoftheCVLforsmall
andMountOkuat3011 m. Becauseofdifficultaccessthe mammal biodiversity consei"vation and diversification.
Cameroon Highlandsecoregionhasbeenpoorlysampled
for its small mammal fauna. Small mammal expeditions
were launched by Eisentraut 1957, 1961, 1963, 1968), 1.1. Background
(
Lamotte(1950-1970, unpublished), andSchlitteretal.
(1999), among others. These studies showed a high en- Previous informationontheMt. Kupesmallmammal fau-
demism ofthe region, particularly in Afro-alpine grass- na: Small mammals from the Cameroon Volcanic Line
lands (above 2000-2500 m) and in mountain forests (CVL) were described by Sanderson (1940) and Eisen-
(Eisentraut 1963; Heim de Balsac 1968; Lamotte & traut(1957, 1961, 1963, 1968, 1969). Theymainlystud-
Petter 1981; Petter 1986; Dieterlen & Van der iedthesmall mammals fromMt. CameroonandBioko Is-
Straeten 1992; Hutterer et al. 1992; Verheyen et al. land,but also collected specimens fromthemassifofMt.
1997). Studies on other vertebrates and plants have also Kupe. M. Lamotte's surveys (1970-1975) followed and
describedmanyendeinics,emphasizingthe importanceof his collections allowed the description and rehabilitation
thisregion(Stuart 1986; Amiet 1987; Sosef 1994; Lar- ofendemic species in the BamendaHighlands (Lamotte
isoN et al. 2000; Cheek et al. 2004; Herrmann et al. & Petter 1981; Petter 1986; Musser & Carleton
2005). Mostpreviousbiodiversity inventoriesfocusedon 2005), but not the Bakossi Highlands. In an unpublished
the two highest mountains (i.e. Mount Cameroon and report Dowsett-Lemaire & Dowsett (1998) provided a
Mount Oku). The biodiversity ofthe other mountains in faunal list ofthe Bakossi Mountains. This list contained
the region is poorly known (summarizedby Hutterer et a great number ofunidentified species.
al. 1992). Among the CVL highlands, Kupe Mountain
(2064 m) harboursawell preservedmountain forest in the Thestudyarea: Mt. Kupe issituatedonthewestern slopes
Bakossi Highlands. This forest is conserved on the west- oftheCVL. It isasteep-sidedmassivehorstblockformed
em sideofthemountain,thankstothe Bakossi peopleand duringCameroon'sthirderuptivephase intheQuaternary
a nature resei"ve project monitored by the World Wildlife (Wild 2004). Mt. Kupe is the youngest mountain ofthe
Fund (WWF). While its vegetation has been extensively Bakossi Highlands. The mountains ofthis highland are
studied (Cheek et al. 2004; Tchiengue 2004), its small amongthewettestareas inAfrica(aroundsixtosevenme-
mammal diversity islesswellknown; Eisentraut(1958, tresannual rainfall. Cheeketal. 2004).Approximately80
%
1963) published a faunal list forthis mountain. Mt. Kupe ofitoccurs intheApril-Octoberwetseasonbutingen-
is famous forthe presence often primate species, among eral every month ofthe yearexperiences rainfall. Never-
theindrill{Mandrillusleucophaeus)andchimpanzee(Pan theless, there is a dry season between November and
troglodytes)populations, whicharethreatenedwithextinc- March, and a single week ortwoofbright sunshinewith-
tion (Wild et al. 2004, 2005). The forest elephant (Lox- out rain usually occurs in July (the "short dry season").
odonta africana cyclotis) is also present on Mt. Kupe Humidityisexceptionallyhigh. Theminimumrelativehu-
(Wild 2004). Numerous woody plant species ofthe low- midity (81 %) is recorded during the dry season.
land forest exhibit distinct sub-montane forest 'fonns'
(Letouzey 1985). Most ofthem still have no taxonomic Descriptions ofthe Kupe forest vary from one author to
status and were not included in assessments ofdiversity, another.All authors agreethatthe lowlandevergreenfor-
endemism and the impacts ofecological change. Several est occurs between 150 to 800 m and is restricted to the
studies (e.g. Breteler 1999; Mc Key 2000) havedemon- westernandsouthernpartofMt. Kupe. The sub-montane
strated the distinctness ofsub-montane forest species, as forest is found from 800 m asl to the summit; Cheek et
well as their key role in understanding the history ofthe al. (2004)considerittobeconstitutedbymid-altitudefor-
lineages concerned. est and Afro-montane rain forest. Previously, Thomas
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
Bonnerzoologische Beiträge 56 161
(1986)recognizedonlytwo foresttypes: the lowland for- urements (weight [W], head and body length [HB], tail
estbelow 1600 mandthesub-montane forestabove. This length [TL], car [E] and hindfoot lengths [HF]) and the
classification isnotfollowedbyLane(1994)orbyCheek sex ofeach animal were taken. All animals were autop-
%
et al. (2004). The sub-montane forest is characterized by sied. The liver was kept in 70 ethanol for molecular
%
thepresenceofPodocarpiismHanjianiis (Podocarpaceae) identifications andthe carcasses preserved in 4 forma-
around 1600-2000m,Santiiiu trimera (Burseraceae)and lin, and the skulls cleaned. All this material was includ-
MNHN
numerous large Guttiferae (Clusiaceae) ofthe generaAl- edinthe collections, Paris. Comparisonswereper-
lanblackici, Peiitadesma,andSymphoniabetween 1000m formed with specimens previously collected from
and2000 m, and Cola (Sterculiaceae) also between 1000 Cameroon (Nditam and Dja regions), Gabon (Makokou,
m m
and2000 (Letouzey 1968). Letouzey(1968)divides Belinga) and Central African Republic (La Maboké). All
the sub-montane forest into three blocks. According to specieswere identifiedbasedon infonnationpublished in
Cheek et al. (2004) this point ofview was based mainly Rosevear (1969), Vivien (1991) and Wilson & Reeder
on aerial observation and it is possible that the band of (2005).
vegetation influenced by human activity in this area was
misinterpretedas sub-montanethicket. Cheeketal. (2004) Forspecies identificationwcperformedclassical morpho-
dividedthe sub-montane forestintotwoaltitudinaltypes: metric analysison both external andcranial distances, us-
the lower sub-montane forest (800-1400 m), harbouring ingadigitalcaliper(MITUYOTO,precision0,01mm)and
ahighepiphyteandsaprophytediversitywith someofthe theprotocol described by Denys &Tränier (1992). Due
Bakossi Highlandsendemics, andtheuppersub-montane to the abundance ofsibling species in tropical Africa, 81
forest(1400-1900m) which is less diverse.Accordingto specimens were sequenced by one ofus (ADM) for con-
Cheek et al. (2004) there is also a montane forest (from finnation of identification (Cytochrome b and/or 16S
m
1900 to the summit), characterized by the absence of genes). ThesespecimenswerecomparedusingtheNeigh-
Claiisena anisata, Laesa lanceolata, and Puvetta hooke- bour Joining Tree Analysis and the sequences available
riana, all otherwise frequent in montane forests of the in Genbank.
CameroonHighlands. Ontheotherhand, itcontainsmany
endemic species such as Leptoderris fasciaculata, Bate- To study the small mammal communities, we used the
santhuspurpureus (also sub-montane), Ficusoreodfyad- Shannon-Wienerdiversity index H= -X Pjlnp, as a direct
um, Zanthoxylum leprieiihi, and Aframomum zambezi- measure ofspecies diversity in each habitat, as well as a
acum. A patch ofmountain grassland is found on each correspondence analysis on presence-absence data. The
ofthe three peaks ofthe summit which are up to 100 m correspondence analysis CFA (Benzecri & Benzecri
wideonly(Cheeketal. 2004). Thesemountain grasslands 1984) was performed on the presence-absence ofspecies
areconstitutedbytwospeciesonly: Paniciim hochstetteri indifferenthabitats. Byreducingthenumberofvariables
and Gladiolus aeqiiinoctialis. and usingthe chi-square distance, this method allowsthe
representationoftherelationshipbetweenvariables (here
habitats)andindividuals(herespecies)onthesamegraph
MATERIALSAND METHODS
2. (scatterplot type) and gives an overview ofthe distribu-
tion ofspecies diversity.All statistical analyseswereper-
Small mammals were trapped during three periods of formed by using XLSTAT 7.1 software (Addinsoft).
fieldwork. December 2005: around Nyasoso (N 04° 50'
/E 09° 40') and onthe mountain below at 950 m; Febru-
ary 2006: between 1500 m and 2000 m; August 2006: 3. RESULTS
around 1000m(Table 1). Linesof250 Shennantraps bait-
edwithpeanuts,flour,oilanddryfishwereusedeachtime 3.L Trapping success, faunal lists
and were set in six habitats: cultivated lowland and sec-
ondary forest (100-800 m), lower sub-montane forest In total, 98 small mammals belonging to 19 species were
(900-1000m), uppersub-montaneforest(1500m),mon- capturedonMt. Kupe. Overalltrapsuccesswith Shennan
tane forest and montane grassland (2000 m). The traps traps was 3,03 %. This varied among habitats: trap suc-
%
werecheckedtwiceperday.Additionally,anumberoftra- cess was always low in the forest, varying from 1,86
ditionaltraps setbyhunterswereputrandomlyinthesur- to2,88 %, dependingon thealtitude; itwashigherinhabi-
roundings ofNyasoso village and at the summit. We al- tatsunderanthropogenic influence(5,51 %;Table2). The
m
soputtwo mist nets atthe entrance ofacave (900 asl) Shannondiversity index indicatesquiteanequal distribu-
duringtwonightsinDecember2005 inordertocatchChi- tion ofspecies abundances in the habitats under anthro-
roptera. Due to logistic constraints the number of trap pogenic influence and the low diversity ofthe sub-mon-
nights varied between habitats. Standard external meas- tane and montane forests.
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
162 Christiane Denys et al.: Small mammales ofMt. Kupe, Cameroon
Table 1. Characteristics ofthe differenthabitats investigated andperiods whentheywere sampled.
Sampling period Elevation range (m) Main vegetation Main habitats GPS coordinates Collecting method
investigated
Dec 2004
begimiing of 850 anthropogenic farmland — Traditional traps
dry season lowland forest closed canopy - Traditional traps
anthropogenic fallow, farmland N 04° 50" 10,1"/ Sherman traps (60)
linesA,B,C,F E 09° 40'48,5"
Dec 2005
beginningof 850-950 anthropogenic home garden N 04° 50' 09.6"/ Sherman traps (40)
dry season E 09° 40" 51,2'"
InwlanH forp'it rlo'ipH fiinnnv N 04° 49' ^6 0"/ Sherman traps (160)
lines D,E,T,U,V E 09° 41" 11,9"
850 anthropogenic fallow, farmland. N 04° 50"/ Shermantraps (100)
line C home garden E 09° 40'
Febr2006
endofdry season 1500-1600 sub-montane forest closed canopy. N04° 49"08,7""/ Sherman traps (250)
lines 0,P,Q,R,S dense under story E09° 42" 27,1"'
1900-2000 montane forest closed canopy N 04° 47" 51,1"/ Sherman (250) and
lines K,L,M E09° 42' 11,2" traditional traps
montane grassland woodland. N 04° 47' 52,7"/ Shermantraps (250)
lines 1,J savannah E 09° 42'23,9"
August 2006
littledry season 850 lowland forest closed canopy N 04° 49'26,0"/ Shermantraps (250)
E09°41' 11,9"
Table 2. Details oftrapping efforts and success in each habitat sampled. The trapping success corresponds only to animals cap-
tured m Shennan traps and the Shannon Index was calculatedon pooled Sherman +Traditional trap captures in all habitats.
Species/Trap Anthropogenic Lowland forest Sub-montane forest Montane
success/Divcrsity/H' forest 2000m
Sher Trad Sher Trad Sher Trad Sher Trad
Crocidiirapoensis complex 4 0 1 0 0 0 0 0
Cricetoiiiys cf. emini 0 1 0 0 0 0 0 3
Deiulwmiis sp. 0 1 0 0 0 0 0 0
Deomysferrugineiis 0 0 0 0 2 0 0 0
Grammomyspoensis 2 1 0 0 0 0 0 0
Hybomys cf univittatus 0 0 1 0 0 0 0 0
Hylomysciis walterverheyeni 0 0 10 0 12 0 12 0
Hylomyscus sp. 0 0 1 0 0 0 0 0
Lemniscomyssiriulus 2 0 0 0 0 0 0 0
Lophwomys cf. sikapusi 0 3 0 0 0 0 0 0
Malacomys longipes 0 0 3 0 0 0 0 0
Mastomys iiatalensis 1 0 0 0 0 0 0 0
Mus musculus 0 2 0 0 0 0 0 0
Mussetulosus 4 1 1 0 0 0 0 0
Oenomys hypoxantlnis 8 3 0 0 0 0 0 0
Praomysjacksoni 2 1 0 0 0 0 0 0
Praomysmisonnei 0 0 13 0 0 0 0 0
Sherman trap nights 47] 1041 751 637
Shennan captures 23 13 30 0 14 0 12 3
Trap success per night (%) 5.51 - 2.88 1.86 1.88
Diversity 7 8 7 0 2 0 1 1
H'(pooled Sher+Trad/habitat) 0,921 0,456 0,178 0,217
) 1
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
Bonnerzoologische Beiträge 56 163
Table3. Means andranges forthe standardexternalmeasurements [mm] ofthe Mt. Kuperodents.
IMeasurements HB TL HF E
Osnoyiivs hvpoxciuthus {n—11) 149 177,3 31 18
(110-168,5) (151-194) (23-36) (13-20)
AAnlnrniw^ lnn<yinp'^ in— 137,7 161 35,3 20,3
(130-142) (151-167) (32-35) (20-21)
LcTnfiiscoTnysstricitus (n=2) 109-122 111-115 25-26 15-16
Mussetulosus (n—5) 74,9 53,5 14,8 11,7
(63,5-84) (50-57,5) (14-16) (10,5-14)
Hybofnys cf. luíivitütiis (n—1) 160 93 28 15
GrüfywíOiiíyspocnsis (n—3) 122 178,7 25 15,3
(116-133) (171-189) (21-^8) (15-16)
Hylofnyscuswcihci^'crhcycfici 89,2 129,34 18,22 14,56
(n-28) (74-103) (112-150) (12-21) (13-16)
Hylovtivscus sp. (n—1) 73 112 19 16
Prüotnvsjacksoni (n=3) 120,3 125 24,3 18
(100-132) (113-139) (24-25) (16-19)
Praomys misonnei (n=10 117,9 133 23,65 17,25
(100-197) (115-160) (21,5-28) (15,5-19)
Dendromus sp.(n=l) 55 84 19 10
Cricetomys emini(n=3) 347,7 371 66 39,7
(340-358) (350-391) (65-67) (39^0)
Lophiiromys cf sikapusi (n=2) 130-142 ' 59-70 20 17
Deomysfernigineus (n=2) 119-137 172-191,5 33-36 25
Mastomysnatalensis (n=l) 113,5 112 20 17
Intotal, 16rodentspecieswerecaptured. Themostabun- OrderRodentia
%
dant species wasHylomyscus walten'erheyeiii (35, 80
ofcaptures),followedbyPraomysmisonnei(13,68%)and Family Muridae Illiger, 1811
Oenomyshypoxanthus(11,58 %). Onlyoneshrewspecies
{Crocidura cf. poensis) and two bat species (Epomops Oenomys hypoxanthus Pucheran, 1855
franqueti andMyonycteris torqiiata) were captured. The
maximum diversity was found in the habitats under an- Six males and five females (onejuvenile) ofthis species
thropogenic influenceandin lowlandforest. Eightspecies were recorded. Standard measureinents are provided in
weretrappedintheformerhabitattype, withOenomyshy- Table 3. Our specimens displayed the characteristic red
poxanthus and Mus setulosus being the most abundant. nose ofthe genus and an underside ofwhite-orange col-
TheabsenceofHylomyscuscf. walten'erheyenifromthis ors ventrally, with a more intense orange-red part in the
habitat type is noteworthy, comparedto its general abun- genital region. The back fur displays a mixture of long
dance in the natural types. In the lowland forest Hy- black guard and short red-yellow hairs. The tail is cov-
lomyscus walterverheyeni and Praomys misonnei domi- eredwithsmallblackscalesandisnakedexceptforafew
nate the assemblages. The montane grassland yielded no scattered,veryshorthairs.TheKupespecimens fallwith-
captures. The sub-montane and the montane forest habi- in the size range of speciinens from Mount Cameroon
tats'diversityarelowwithtwo specieseachandarechar- (Eisentraut 1963)andSNigeria(Happold 1987). Oeno-
acterized both by the presence ofHylomycus walterve- myshypoxanthus shows appreciablegeographic variation
heyeni. in fur color and body size; in Cameroon it was also cap-
turedinNyasoso, Buea, Musake, Malende andMueli (1
specimens) (Eisentraut 1963). During this inventory of
3.2. Species records, characteristics and systematic the Mt. Kupe small mammal fauna, this species was on-
notes lytrapped at low altitude in Nyasoso village and in a fal-
low areawithAgeratum spp. Theseresults are congruent
In this species account we provide morphological and/or with the observations ofKingdon (1997) and Happold
habitat characteristics foreach species. Where confusion (1987), according to which O. hypoxanthus commonly
with another species is possible we also provide results nests in fallow ground and village gardens. It is general-
DNA
from skull analysis and cyt b or 16S sequences ly restricted to forestbelow 2000 m.
analysis.
2
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
164 Christiane Denys et al.; Small mammales ofMt. Kupe, Cameroon
Malacomys longipes Milne-Edwards, 877 Mus musculus Linnaeus, 1758
1
Three specimens (two males and one sub-adult female) Two M. musculus males were trapped by inhabitants in
were trapped in the forest along a river at an altitude of Nyasoso village. Cyt b analysis confirmed their identifi-
mm
850 m. They have a hindfoot longer than 34 (Table cation.
3). The specimens from Nyasoso have a shorter ear
(20-21mm) than those from Efulan (Cameroon) or Hybomys cf univittatus (Peters, 1876)
Makokou/Belinga (Gabon) (25 mm), and their skull and
body sizes are also smaller. These observed differences One sub-adultmale was recordedin lowlandforest. Both
may have been a resuU ofthe low sample size ofspeci- skull morphology and cyt b analyses identified it as Hy-
mens captured from Nyasoso. Musser & Carleton bomys (Table4). Eisentraut (1968, 1973) describedon-
(2005)recognizedtheexistenceofthreespecieswithinthe lyH. univittatusbadiusfromNyasosoandMt. Cameroon.
genusMalacomys,amongwhichonlyM. longipesismen- Butaccordingto the same authorHybomys u. badius and
tioned in Central and South Cameroon. However, M. H. u. univitattus coexist in Mamfe region. Musser &
longipes and M. edwardsi occur both in South Nigeria Carleton (2005) mentioned taxonomic problems con-
(Happold 1987). Eisentraut (1963) found only seven cerning the Central African Hybomys species, but con-
specimens ofM. longipes in Batoki, Malende and Isobi finned three other forms in the Cameroon Highlands: H.
(Caineroon)andonlyoneonMt. Kupeat900 m.butmen- badius on the Bakossi Highlands (Mt. Cameroon, Rumpi
tionedagreatvarietyofpelagecoloursinhissample. Sim- Hills), H. eisentrauti (described from Mt. Lefo, syn-
ilarly Happold (1987) also mentioned that Cameroon onymizedbyMusser& CarletonwithH. badius)onthe
specimens oflongipes show considerable variation, and BamendaHighlands, andH. basiliiontheIslandofBioko.
some individuals are almost the size ofthe smaller ed- Further revision and discoveries ofnew specimens from
wardsi. However, the comparison ofourcytb sequences both lowland and highland forests ofthe CVL pending,
to those available in Genbank allows the definite identi- we attribute the Mt. Kupe specimen from lowland forest
fication ofthe three specimens as M. longipes. provisionally to the univitattus complex.
Lemniscomys striatus Linnaeus, 1758 Grammomyspoensis (Eisentraut. 1965)
Twomalestripedratsweretrappedinamaize fieldatNya- Three specimens were caught in lowland forest and cul-
soso village, 800 m asl. They fallwithinthe sizevariabil- tivatedareasatNyasosovillage(Table2). Oneofthetwo
ity and color ofL. striatus from Central African Repub- males was sexually active and the female was pregnant,
lic (CAR), Congo and Cameroon and have molecular carryingtwoembryosanddisplaying0+2mammae. Spec-
affinities with representatives from the region (Nicolas imenswere identifiedbasedontheirverylongtail (Table
etal. 2008a). Afew earlier specimens had been collected 3), ending with a small tuft, and by typical skull charac-
inNyasoso, Mt. Manengouba(Eisentraut 1973) andthe teristics, notablytheglobularaspect ofthebraincase, mi-
Mamfe region (Sanderson 1940). crodonty, aswell as theexistence ofastephanodont crest
M
onthemolarcuspsandpresenceofa t7ontheupper 1
Mus (Nannomys) setulosus Peters, 1876 (seereviewinPetter&Tränier 1975). Comparisonwith
G. gazellae (put in synonymy with G. macmillani by
Eisentraut(1957)attributedtheNyasosoMus specimens Musser & Carleton 2005) specimens from CAR and
toM. musculoides, but mentioned that they have a brow- with G. "rutilans"(asynonymofG.poensis)fromGabon
nerpelagethanspecimens from BueaandMt. Cameroon.
and with the holotype from Fernando Poo allows assig-
Later, Eisentraut(1968) con-ectedthe identificationand nationofournewKupespecimensto G.poensis.Yet,they
attributedthe M. musculoides specimens toM. setulosus. are slightly smaller than the holotype which is a young
which we confirm here on morphological and molecular adult. Two specimens were sequenced (cytb) but a revi-
grounds. The type specimen ofM. setulosus comes from
Limbe (formerly Victoria), less than 100 km S of Mt. sion ofthewhole genus in WestAfrica is requiredbefore
definite conclusions can be drawn. According to Eisen-
Kupe. We obtained six pygmy mice (two males and four traut (1968) this species is absent from Mount
females) attributed toMus (Nannomys) setulosus. based Cameroon, butpresent in Mamfe (Sanderson 1940) and
on their dark pelage, rather large size (Table 3), and the on Bioko Island (Eisentraut 1965).
presence ofa second lobe on the lower M3 (Kouassi et
talh.e2i0d0e8n)t.ifTichaetiyonha(vKeoubaesesnimeotlealc.u2l0a0r8l)y:tTypLedistsoocmoenwfhiantn Hylomyscus waltei-verheyeniNicolas et al. 2008
shorterthan HB.All six specimensweretrapped in zones
Thirty-fourspecimens ofthis rodentwere trappedonMt.
isnufblaudeunlctefdebmyalheuhmaadn2a+ct2ivmitaimesmaaned,inthleowotlhaenrdffeormeaslte. Ownaes Kupe at different altitudes (Table 4). Among the 12 fe-
males we found twojuveniles, the otherbeing adultwith
ajuvenile.
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
Bonnerzoologische Beiträge 56 165
2+2 mammae. Three females were each carrying four to to us by Nyasoso villagers (Table 4). The mammae for-
five embryos. Ten ofthe 20 males were sexually active. mula was 1+2. This species is recognized by its well-de-
Our identifications were confirmed through molecular veloped supra-orbital crests and the presence ofa clear-
analyses (cyt b and 16S gene sequencing). According to ly separatedt3 on Ml (Lecompte etal. 2001).Acompar-
ourpreliminary molecular analyses, specimens from Mt. ison ofour 6S andcytb sequences withthose published
1
Kupe are the sister clade ofthe newly described species by Nicolas et al. (2005) and by Lecompte et al. (2002)
H. waltef-verheyeni(Nicolasetal. 2008a,Missoup2009). confinned our identification.
This species was described from the region between the
Sanaga River and the Oubangui and Congo rivers in Praomys misonneiVan der Straeten & Dieterlen, 1987
Cameroon, Gabon, CentralAfrican Republic and Repub-
licofCongo.However,theauthorsstressedtheimportance Five femalesandeightmalesweretrappedinthe lowland
ofconductingadditional sampling inWesternCameroon, forest (Table 2). Based on the absence of supra-orbital
Nigeriaandthe DahomeyGap region in ordertoprecise- crests and their molar moq^hology they clearly belong to
lydefinethewestern limitofits distributional range. Re- the so-calledtullhergicomplex. The femaleshaveamam-
fined molecular analyses and multivariate morphometri- mae formula of 1+2. A comparison ofour 16S and cyt b
cal analysisofcranio-dental measurements arenecessary sequences with those published by Nicolas et al. (2005)
to confirm whetherthe Kupe individuals represent a dis- and Lecompte et al. (2002) confirmed ouridentification.
tinct montane forest clade.
Subfamily Deomyinae Thomas, 888
Hylomyscus sp. 1
OneadultfemaleofHylomyscuswascapturedinthezone Deomysferrugiueus Thomas, 1888
underanthropogenic influence andwas molecularly sim-
ilar to four specimens called ''Hylomyscus taxon 2" by Two specimens were captured at c. 1500 m asl on Mt.
Nicolasetal. (2006).Thesefourspecimensallcamefrom Kupe (Table 2). This is surprising since Eisentraut
the KompNational Park in western Cameroon. They un- (1963)onlyobtainedthisspeciesinthe lowlandforest, six
doubtedly belong to a new species. Waiting for a gener- specimens in Nyasoso and at "LagerV" (western part of
al revision ofthe genus and especially ofthe taxa found Mt. Cameroon) at 600 m.
alongtheCVL,weprovisionallycall it""Hylomyscussp."
This female had a mammae fonnula of2+2. Lophuromys cf. sikapusi (Temminck, 1853)
Mastomys natalensis (Sm\th, 1834) Wetrappedonlyonemaleandone female;onebadlydam-
aged specimen (sex unknown) was brought to us by the
OneadultmalespecimenwastrappedinNyasosovillage. villagers.All specimensweretrapped inthe village.They
Ananalysis ofthe cytb gene confirmedits identification are characterizedby a redbellyanda shorttail (Table4).
asM. natalensis. Becauseofthehighmorphological vari- Thefemalehad 1+1 mammae. Recently,Lophuromystax-
ability observed in this group and the low degree ofspe- onomy and systematics in Central, West and EastAfrica
cific distinctiveness(Denysetal. 2007), wedidnotcom- hasbeenunderdebate.AccordingtoMusser& Carleton
parethisspecimenwithotherspecimensmoiphometrical- (2005), the Lophuromys sikapusi complex may be repre-
ly. Previously only one specimen of this species was sentedin SCameroonbydifferentspecies, followingVer-
trapped(Eisentraut 1968), andtherefore itdoesnotseem heyen et al. (1997, 2000). Numerous sibling species are
to be common in the CVL while it is abundant in West found within this complex. A molecular revision is nec-
Africa(Denysetal.2005)andinNigeria(Happold 1987). essary in order to precisely identify the specimens from
Mt. Kupe.
Praoffijs Thomas, 1915
Basedonmolecularanalysesandthe identificationkeyof Family Nesomyidae Major, 1897
Lecompte et al. (2001) the Mt. Kupe specimens belong
totwo differentspecies. Eisentraut (1968) only record- Subfamily Dendromurinae G. M. Allen, 1939
ed one Praomys species, P. morio, from Mt. Cameroon.
Dendromiis sp.
Praomysjacksoni (de Winton, 1897)
Onlyoneadultmalewascaptured inacropfield.Accord-
Inourstudythis species isrepresentedbyonefemale,one ing to MussER & Carleton (2005) and Eisentraut
maleandonejuvenilewhichwerealltrappedandbrought (1968, 1973),D. messorius in the Cameroonhighlands is
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
166 Christiane Denys et al.: Small mammales ofMt. Kupe, Cameroon
known from Manengouba and Nyasoso, and from the Order Soricomorpha Gregory. 1910
holotype from Efulen. The species is a typical inhabitant
of tropical lowlands and lower mountain slopes. In the Family Soricidae G. Fischer, 1814
same region. Eisentraut foundDendroiniis oreas at high-
er altitudes (1850 m and 3000 m on Mt. Cameroon). Crocidura ci.poensis (Eraser, 1862)
MussER& Carleton (2005)mentionedthehighvariabil-
ity inD. oreasfromWestCentralAfrica. Our"new"spec- Fivemedium-sizedshrewswerecapturedinShermantraps
imen is smaller than D. messorius and D. oreas of the (Table 2). We recovered four males and one female with
CVL. In size and skull morphology it is close to speci- 0+2 mammae. At least eleven species of Crocidura are
mens in the Paris Museum from Belinga, Gabon, identi- known from SWCameroon(Hutterer2005).According
fied asD. mystacalis. The specimen from Mt. Kupe isal- to size and morphology ournew specimens fitwithin the
so smallerthanD. messorius from CAR. Pending further C. poensis complex (P. Barriere, pers. comm.).
comprehensive revision of the genus in West-central
Africa, we leave this specimen as unidentified. It should be noted that Eisentraut (1973) recorded 5
species ofshrews, Sylvisorex megalura, Sylvisorex ollu-
la, Crocidura nigeriae/poensis. C. attUa, and C. hilde-
Subfamily Cricetomyinae Roberts, 1951 gardeae from Nyasoso at 850 m.
Cricetomys cf emini Wroughton, 1910 Order Chiroptera Blumenbach, 1779
Four Cricetomys individuals (three males and one unde- We used mistnets to sample the bat population ofa cave
termined) were brought to us by hunters. They were ob- 950 m asl in the lowland forest ofMt. Kupe. Two fruit
tained by traditional traps placed in Nyasoso village and bat species were collected.
in the forestnearthetopofthemountain, atapproximate-
ly 1950masl. Eisentraut(1968)described 11 specimens Family Pteropodidae Gray, 1821
from Buea, Nyasoso, Kumba and Mt. Cameroon as C.
gambiamts poensis. MussER & Carleton (2005) put Epomopsfranqiieti (Tomes. 1860)
poeiisis in synonymy with emini and recognize C. gam-
bianus only in the sub-Saharan savannah belt. Our spec- This species displays white basal eartufts, and males al-
imens fit within the external measurements range ofvari- so have white epaulets. We collected two adult females
ability provided by Eisentraut (1968), except for the ofthis species (Table5).This specieswas quitecommon-
shorter tail length. According to Olayemi (2008) both C ly trapped by Eisentraut (1968) on Mt. Cameroon and
gambiamisand C. eminiarepresent in SouthNigeria. Fur- atLakeBaroinbi Mbo, andhe attributedthesespecimens
thermolecularworkwillhelptorefinethetaxonomywith- to the subspecies/ra/ií/ííe//, mentioning the considerable
in this genus. variabilityofthespeciesoverallofWestAfrica. Ourspec-
imens fromNyasosoaremuch largerthantheaveragefor
Table4. External measurements [mm] ofMt. Kupe Crociclii- his specimens (FA=57,4,TL=12, HB=9I, HF=15; Eisen-
ra. traut 1968) but are smaller than Nigerian specimens
HB TL HF E W (FA=82-9I; Happold 1987).
Myonycteris torqiiata (Dobson, 1878)
mean 94,2 53,4 14,6 8,6 16.7
range 82-105 48-62 10-17 7-11 13-19
One male ofthe little collared fruitbatwas trapped. This
species has also been sampled by Eisentraut (1968) at
Table 5. External measurements [mm ofMt. Kupe Chiropte- "LagerV"and"LagerVI"ofMt. Cameroonandwas list-
ra. ed in an unpublished report by Dowsett-Lemaire &
DowsETT (1998). Our specimen has a smaller foreann
W
Species HB TL HF+ claw FA length (44 mm) than specimens from Mt. Cameroon (av-
erage 57,4mm; Eisentraut 1968) andNigeria(Happold
Epomops 1 66 122 10 20+ 2 80 1987) (Table 5).
Epomops 2 50,5 113 16 20 + 2 77
Myonycteris 14 65 10+ 0,5 44
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
Bonner zoologische Beiträge 56 167
4. DISCUSSION LikeDowsett-Lemaire& Dowsett(1998),we found the
bat species Epomops franqueti and Myonycteris torc/ua-
4.1. Rodent species diversity on Mt. Kupe compared ta inacave in thelowlandforestofMt. Kupe. Eisentraut
to results ofprevious studies (1973) and Fedden & McLeod (1986) collected both
species at the same altitude. They seem to be quite com-
OurstudyconfinnsthepresenceofL. cf.sikapusi,L. stria- mon on Mt. Bakossi and Mt. Cameroon.
tus, M. setulosus, O. hypoxanthiis, H. cf. iinivittatus and
D.femigineusonMt. Kupe(Table6). Thedifferencesbe- The yield of shrews was poor as compared to previous
tweenourspecies listandthoseofpreviousstudiesmain- studies (Eisentraut 1973) and does not allow any con-
ly concern genera forwhich the taxonomy is underrevi- clusion.
sion: Praomys, Hylomyscus, Mus, Nanuomys and Crice-
tomys. ConcerningthegenusPraom'i,wecouldnotcon- Ourstudy indicatesthattherodent faunaon Mt. Kupehas
firm the presence ofP. nioiio on Mt. Kupe, as suggested notchanged much since the pioneerstudy ofEisentraut
by Eisentraut (1968). This species was described from (1957). This maybe a sign ofthe quality ofconservation
Mt. Cameroon where it ranges at altitudes up to 3000 m. ofthis forest. It has also shown that the rodent diversity
The validity ofP. morio was never assessed by molecu- is quite high in comparison with other mountains ofthe
lar data. Our results show that the specimens from Mt. CVL (Mt. Bakossi, Mt. Manengouba, Mt. Lefo, Mt. Go-
KupecorrespondtoP. misonnei, nottoP. morio. Iffuture tel, Mt. Cameroon - see Missoup et al. 2006). With 16
molecularstudiesconfinnthevalidityofP. morio itshould species, however, it is lower than on Mt. Oku (Missoup
be restricted to the high altitudes ofMt. Cameroon. We et al. 2006). This may be due to the fact that few inven-
also confirmed the presence ofP.jacksoni, as mentioned torieshavebeenperformedonthesemountains.All moun-
byEisentraut(1973)andDowsett-Lemaire& Dowsett tains may harbour inore species than are currently
(1998). These authors also cited Hylomyscus Stella from known. However, our results may also be biased by the
Mt. Kupe. However, Nicolas et al. (2008b) showed that common species because ofthe low numberofindividu-
H. Stella is restricted to East Africa, and that the speci- als captured despite ourtrapping efforts. All the possible
mens fi-om West Central Africa (from the Sanaga River habitats ofthe lowland and montane forest were investi-
to the Oubangui-Congo River) represent a new species, gated but our trapping success remained very low in all
H. waltei-verheyeni. Our results suggest that this species lines and in all seasons (Table 2). Trapping success was
couldalsobepresenton Mt. Kupe. More detailedmolec- lower than that obtained for Mount Elgon montane for-
ularand morphometrical analyses, though, are necessary est (7,7-10%) (KiTYO & Clausnitzer 2001), Kilimand-
toconfirmthisconclusion. Eisentraut(1968) attributed jaro (4,5-36%)(Mulungu etal. 2008)ormontane forests
Mus specimens fi-omNyasoso toM. setulosus, which we oftheAlbertine Rift mountains, SW Uganda (3,41-11,4
confirm here on morphological and molecular grounds. %) (Kasangakietal. 2003). Inall these studies, thehigh-
Concerning Cricetomys, Dowsett-Lemaire & Dowsett esttrapping success was obtained indisturbedordegrad-
(1978) mentioned the presence of C. emini at 1500 m. ed habitats.
However,molecularanalyses inprogress seemtoindicate
quite a large genetic variability. 4.2. Community analysis
Our inventory adds three species to the fauna ofthe Mt.
Kupe: Grammomys poensis, Mastomys natalensis, and It is generally assumed that the diversity ofsmall mam-
mals decreases with altitude, or at least changes with
Hylomyscus walterverheyeni (previously recorded under
changes in vegetation type. Vegetation types may play a
thenameH. Stella) (Table 6). Despiteitspresenceon Mt. crucial role in structuringsmall mammal assemblages. In
Cameroon, the genus Mastomys was not previously
ordertotestifthesehypothesesapplytoMt. Kupeweper-
recorded on Mt. Kupe. Similarly, Eisentraut (1957,
formedacorrespondenceanalysisonpresence-absenceda-
1968) didnot list Grammomyspoensisbut indicatedthat
itwastrappedby Sanderson(1940)intheMamferegion. ta forrodents and shrewsas a functionofaltitudinal veg-
We also confinn the presence of Malacomys lougipes, etation levels.
Oenomyshypoxanthus,Deomysferrugineus, andMusse-
tulosus inNyasoso. We failedto collectStochomys loiig- The graph ofaxis Fl x F2 ofthe coiTespondence analy-
icaudatus (4 specimens collected in Nyasoso by Eisen- sis displays a quite good separation between all habitats
(Fig. 1). It is clearthat zones under direct human impact
traut)andHylomyscusacta, botheasily identifiablebased
are well characterized by a cohort ofcommensals (like
onmorphologicalcharacters.Stochomys longicaudatus is Mastomys, Mus and Rattus), but also species which are
known to be present in fragmented populations through-
o(ourKtimnitagsdydoinestvr1ei9nb9u7ht)ai.ovnIeanldtirhseaanKpguep,epaearneaddresamiatnhcyiesbs1ep9e6lc3oi.ceasllmyaaybubnedraanrte Lfharobepiqhtuuaetrnsotmdiyinsssp,elcaDoyenanddadrriyoffmfeuorrseen,sttPda.invjdearccskrisotopynsiw)(ie.t.hgT.mhLaeermktnreiudsecdofimofrfveessr,-t
© Biodiversity Heritage Library, http://www.biodiversitylibrary.org/; www.zoologicalbulletin.de; www.biologiezentrum.at
168 Christiane Denys et al.: Small mammales ofMt. Kupe, Cameroon
Table 6. Comparisons ofthe faunal lists forMt. Kupebasedon differentstudies (x=presence).
Species This study Dowsett-Lemaire Eisentraut Eisentraut 1968, 1973
&
Dowsett 1998 1957
Crocidurapoensis x
Crocidiira attila
Crocidura sp.
Cricetomysgamhicunis
C. emini x
Dendromus sp. x
Grammomyspoensis x
Lophuromys sikapusi x
Deomysferruguineus x
Lemniscomys striatiis x
Mastomys natcdensis x
Mus musculus x
Mus setulosus x
Mus musculoides
Oenomys hypoxanthus x
Praomysjacksoni x
Praoinys morio
Praomys misonnei x
Hybomys cf. univitattus x
Hylomyscus walten'erheyeni x
Hylomyscussp. x
Hylomyscus acta
Hylomyscus Stella?
Malacoinys longipes x
Stochomys longicaudatus
Rattus rattus
Micropteruspusillus
Epomops franquea x
Myonycteris torquata x
Nannonycteris veldkampi?
Lissonycteris angolensis
Roussettus aegyptiacus
Megaloglossus woermanni
Kerivoula smithii
Hipposideros caffer
Hipposideros ruber
Hipposideros cyclops
Rliinolophys alcyone
Pipistrellus eisentrauti?
Tadarida thersites?
enees between the lowland forests on the one hand, and montane and montane forests) on Mt. Kupe. Eisentraut
the sub-montane andmontane forests (1500-2000 m) on (1957)showedforMt. CameroonthatMalacomysandDe-
the other along the F2 axis. Hybomys, Hylomyscus sp., omys are lowland forest taxa (< 900 m) and that only
Malacomys, andP. misonneiarecharacteristicofthe low- PraomysmorioandHylomyscusactareachthehigherlev-
land forest and are tme forest taxa also found in the low- els(upto3100mand2000m,respectively). Otherspecies
land forests of Gabon (Duplantier 1982; Nicolas & oíHylomyscusare coinmonlytrappedathigh altitudes in
COLYN2003).H. walterverheyeniandDeomysfenngineus EastAfrica: OnMt. ElgonH. denniae is found inthe for-
aretheonlyspecies foundat 1500and2000m (uppersub- est and bamboo zones up to 2875 m and 3100 in (Kityo