Patterns of morphological and genetic variation in western Indian Ocean members of the Chaerephon ‘ pumilus’ complex (Chiroptera: Molossidae), with the description of a new species from Madagascar Author GOODMAN, STEVEN M. Author BUCCAS, WAHEEDA Author NAIDOO, THESHNIE Author RATRIMOMANARIVO, FANJA Author TAYLOR, PETER J. Author LAMB, JENNIFER text Zootaxa 2010 2010-07-28 2551 1 1 36 http://dx.doi.org/10.11646/zootaxa.2551.1.1 journal article 10.11646/zootaxa.2551.1.1 1175-5326 10832899 Chaerephon atsinanana sp. nov. ( Figures 3 -7) Madagascar free-tailed bat Holotype . FMNH 185259 , adult female (not in reproductive condition), body preserved in 12% formalin and subsequently transferred to 70% ethanol, skull removed and cleaned. Original field number Fanja H. Ratrimomanarivo ( RHF ) 1060. Muscle tissue preserved in EDTA and housed in the FMNH under the same catalog number. This specimen was used in both the morphological and molecular analyses. Type locality . Madagascar : Province de Fianarantsoa , Farafangana , Collèges d'Enseignement Général ( CEG ) Fenoarivo , 22°49.275'S , 47°49.860'E , 10 m . Captured in a village on 26 April 2005 with a mist net placed between buildings . Paratypes . Includes eight additional specimens from the type locality ( FMNH 185260-185268 ). All of these specimens were used in the morphological analyses and FMNH 185260 in the molecular analyses . Referred specimens (specimens not used in morphological analysis in bold and those used in the molecular analysis underlined). Province d’Antsiranana : Andapa , 500 m ( FMNH 154064 ) . Province de Fianarantsoa : Vangaindrano (ville), bureau circonscription scolaire, 23°21.300'S , 47°35.763'E , 10 m , ( FMNH 185229 , 185230 , 185231-185238 ) ; Vangaindrano (ville), Dispensaire public, Ampasy , 23°21.426'S , 47°35.813'E , 10 m ( FMNH 185239-185240 , 185241 , 185242-185243 , 185244-185256 , 185257 , 185258 ) ; Commune Mahabo , chalet de marché, 23°11.316'S , 47°40.745'E , 30 m ( FMNH 185269-185271 , 185272 , 185273-185278 ) ; Commune Ampahatelo , maison du domaine d'Akamasoa , 23°20.678'S , 47°35.778'E , 20 m ( FMNH 185279-185282 ) ; Vohipeno , Commune Vohitrindry , Quartier Fenoarivo , grenier N. 1, 22°21.997'S , 47°50.206'E , 35 m ( FMNH 185283-185285 , 185286 , 185287-185292 ) ; Vohipeno , quartier Ambohimanarivo , bureau de Fivondronana , 22°21.242'S , 47°50.421'E , 30 m ( FMNH 185293-185296 , 185297 , 185298-185305 ) ; Manakara , Manakara be, EPP Tanambe , salle A, 22°09.418'S , 48°01.009'E , 15 m ( FMNH 185306-185311 , 185312 , 185313 , 185314 , 185315 , 185316-185318 ) ; Ifanadiana (ville), Hôpital CSB II Mazavatakona , 21°18.394'S , 47°38.144'E , 459 m ( FMNH 185319-185321 , 185322 , 185323-185328 ) ; Ifanadiana (ville), école Lycée , salle I, 21°17.904'S , 47°38.264'E , 460 m ( FMNH 185329 , 185330-185335 , 185336 , 185337- 185338 ) ; Ranomafana ( Ifanadiana ), maison d'habitation, 610 m ( FMNH 188081-188082 , 188083 , 188084- 188087 , 188088 , 188089 , 188090-188091 ) ; Ifanadiana ( MNHN 1985.437 - 1985.440 ) . Province de Toamasina : Périnet-Analamazaotra , 140 km E Tana ( BMNH 76.1896 , 76.1897 ); Moramanga , 950 m ( MCZ 45098 -45099 , 45107 , 45582-45583 ) ; Ambodiriana , 20 km SW Périnet , 975 m ( ROM 42071 -42072 ) ; Périnet ( Andasibe ), CIBA , 18°53.747'S , 48°24.907'E , 950 m ( FMNH 184491 , 184492 , 184493-184495 ) ; Périnet ( Andasibe ), cité CIBA , 18°55.410'S , 48°25.246'E , 980 m ( FMNH 184496-184499 , 184500 , 184501-184508 ) ; Périnet ( MNHN 1985.475 ) ; Beforona , bureau de poste, 18°53.349'S , 48°34.683'E , 560 m ( FMNH 184509-184518 ) ; Beforona , ex hotel villa Martin'son , 18°53.320'S , 48°34.711'E , 560 m ( FMNH 184519-184522 ) ; Beforona , école CEG , 18°53.334'S , 48°34.792'E , 560 m ( FMNH 184523-184524 ) ; Ambatondrazaka , Andrarabarikely , 17°49.784'S , 48°25.112'E , 1000 m ( FMNH 184651-184652 , 184653 , 184654-184656 , 184657 , 184658-184659 , 184660 ) ; Sabotsy Anjiro , maison d'habitation, 18°53.672'S , 47°58.401'E , 850 m ( FMNH 184661-184666 , 184667 ) ; Sabotsy Anjiro , maison d'habitation No. 2, 18°53.711'S , 47°58.391'E , 850 m ( FMNH 184668-184672 , 184673 , 184674-184676 ) ; Sabotsy Anjiro , Andranoalina , maison d'habitation No. 3, 18°52.945'S , 47°58.245'E , 850 m ( FMNH 184677 , 184678 , 184679-184685 , 184686 ) ; Toamasina (ville), menuiserie Mangarano , 18°08.441'S , 49°22.670'E , 10 m ( FMNH 187797—187799 , 187800-187803 ) ; Toamasina (ville), EPP Ambohijafy , 18°07.569'S , 49°24.128'E , 10 m ( FMNH 187804—187805 , 187806 , 187807-187813 ) ; Toamasina (ville), menuiserie Mangarano , 18°08.441'S , 49°22.670'E , 10 m ( FMNH 187814-187815 ) ; Fanandrana , EPP, 18°15.122'S , 49°16.067'E , 40 m ( FMNH 187816-187822 ) ; Brickaville , bureau commune, 18°49.317'S , 49°04.343'E , 10 m ( FMNH 187823 ) ; Ranomafana , hôpital CSB II, 18°57.661'S , 48°50.655'E , 50 m ( FMNH 187824-187830 , 187831 , 187832-187833 ) ; Ranomafana , église catholique, 18°57.636'S , 48°50.845'E , 90 m ( FMNH 187834 , 187835 , 187836 , 187837 ) ; Moramanga (ville), Lycée technique, 820 m ( FMNH 188112-188118 , 188119 , 188120-188121 ) ; Moramanga (ville), grande salle de l'EPP près du Lycée , 820 m ( FMNH 188122-188126 , 188127 , 188128 , 188129 , 188130 ) ; Moramanga (ville), école primaire privée, 820 m ( FMNH 188131 ) ; Anjiro , Marozevo , maison d'habitation II, 815 m ( FMNH 188142 , 188143 , 188144 ) ; Moramanga (ville), Sahanofata , 835 m ( FMNH 188132 , 188133 , 188134 , 188135-188137 , 188138 , 188139-188140 , 188141 ) ; Station Forestière d'Ivoloina , près du gîte Mahatsinjo , 18°03.406'S , 49°21.635'E , 15 m ( UADBA RB-20 , RB-21 ) ; Andreba , Railroad Station ( USNM 328771-328772 ) ; 3 km E Périnet ( USNM 341734-341743 ) . FIGURE 6 . Lateral view of central portion of external right ear and antitragus of different Chaerephon spp. : A – C . atsinanana sp. nov. (FMNH 188144) from Madagascar: Province de Toamasina, Commune Anjiro, Marozevo; B – C . leucogaster (FMNH 188644) from Madagascar: Province d’Antsiranana, Nosy Komba, Ampangorinana; C – C . pumilus sensu stricto (BMNH 19.7.7.334) from Massawa, Eritrea, the type locality of this taxon. Etymology . The name atsinanana is derived from the Malagasy word meaning “from the east.” Diagnosis . A relatively small molossid bat with a forearm of 37-42 mm . The ears are notably shorter than the head and they are united by a short band of skin. The antitragus is large and broad, the anterior margin slightly angled and posterior margin rounded, and blunt tip ( Figure 6a ). The tragus has a broad attachment. Skull and mandible relatively large for a small member of Chaerephon . The nasal is inflated and in dorsal view, the rostrum is expanded anterior-laterally ( Figure 3 ). The cranial portion of the frontal, parietal, and supraoccipital, as well as the squamosal, are inflated. The basisphenoid pits are large, deep, slightly oval, and approaching the basioccipital septum. Basioccipital pits are either not present or indistinct. Palatal foramen is open, but not forming a deep cleft. The individual teeth, particularly molariform, relatively robust. The P 3 is prominent and aligned towards the outer margin of the toothrow. Lingual portion of upper molariform teeth elongated, particularly the hypoconal flange of M 1 , the paracone and hypoconal flange of M 2 , and the hypoconal flange in M 3 . Given present taxon and character sampling, the species is further diagnosed by the following strict synapomorphies in the cytochrome b gene (the first nucleotide given is the ancestral state, followed by the nucleotide position in the cytochrome b gene, followed by the derived state; substitutions that result in an amino acid change are indicated in bold text): Hap 15: C108T, A126T, A147G, A238T, T282C, T309C, T351C, C352T, A516G, C573T, T579C, G709A, G711A, A753G, A1020G, G1038A, C1101T, G1104A. Hap 16 and 19: C108T, A238T, T282C, T309C, T351C, C573T, T579C, G711A, A1020G, G1038A, C1101T, G1104A. Hap 17: C108T, A147G, A238T, T282C, T309C, T351C, C352T, C573T, T579C, G711A, A753G, A1020C, G1038A, C1101T, G1104A. Hap 18: C108T, A238T, T282C, T309C, T351C, C573T, T579C, G711A, A1020G, G1038A, C1101T, G1104A. Sympatrically occurring Molossidae . At the holotype site in Farafangana near the Collèges d'Enseignement Général (CEG) Fenoarivo, only C . atsinanana was captured, but Mops leucostigma was also found in this village. Across the range of C . atsinanana , it is known to occur in sympatry, in virtually all cases in synanthropic settings, with C . leucogaster (at Manakara [ Ratrimomanarivo et al . 2009a ]), Mormopterus jugularis ( Ratrimomanarivo et al . 2009b ) , and Mops leucostigma ( Ratrimomanarivo et al . 2008 ) . Description. The lectotype of nominate pumilus (SMF 4311, an adult female based on the examination of the specimen’s external genitalia) from Massawa, Eritrea , has a partially damaged skull, while other specimens from the type series collected in Massawa, particularly SMF 11918, are in excellent condition. In the Senckenberg Museum, these specimens were compared to USNM 38032 (an adult female), also from Massawa, and all are morphologically similar with the exception of some subtle differences that are best considered falling within the range of intraspecific variation. In subsequent comparisons, USNM 38032 was used to represent the cranial and dental character states in nominate pumilus . External characters. The forearm length of the lectotype of nominate pumilus (SMF 4311) taken from the outstretched wing of the mounted specimen measured 38.6 mm . This falls within the range of C . atsinanana . Given the style of preparation and the state of the specimens, it was difficult to compare USNM 143167 or individuals of C . atsinanana to the type series of nominate pumilus in the SMF. Herein, comparisons are based on the specimens from Eritrea , Saati - USNM 143167 (an adult male) and from Massawa - USNM 143168, USNM 38032, and BMNH 19.7.7.3347 (all females) from Massawa. On the basis of the material available, in fluid preserved specimens of nominate pumilus and C . atsinanana , the fleshy portion of the ears, including the band connecting them, have similar morphology. However, the antitragus of C . atsinanana is notably broad, with an angular anterior edge, and terminating with a rectangular-blunt tip, while in nominate pumilus the structure is less thickset, with a less angular anterior margin, more angular posterior margin, and slightly rounded tip ( Figure 6 ). Further, the short tragus in C . atsinanana is distinctly broader and with an extensive portion attaching to the ear, compared to nominate pumilus that has a narrower articular surface (not illustrated). In C . leucogaster , the antitragus is small, proportionately not wide, the two lateral margins being largely similar in shape, and terminates with a rounded edge and in C . atsinanana it is notably larger and broader, anterior margin slightly angled and posterior margin rounded, and blunt tipped ( Figure 6 ). The tragus of C . leucogaster is slightly longer and narrower at the base than the shorter and broader structure in C . atsinanana . The dorsum, throat, and chest in C . leucogaster is dark brown, as compared to C . atsinanana with a blackish-brown dorsum, brown throat, and dark brown chest; in the latter taxon there is considerable variation (Figure 7). The abdomen in C . leucogaster tends to have a large whitish area, although in a few individuals this is not extensive and restricted to the mid-ventral area, while in C . atsinanana the venter is a dark brown and in a few rare cases, there is a small white mid-ventral patch. In general, C . atsinanana have a distinct whitish or beige strip of hairs at the base of the wings (plagiopatagium), also found in most African animals assigned to C . pumilus , although considerable intraspecific variation can be found on Madagascar within adults of the same population (Figure 7). This pelage trait is largely unknown from Malagasy C . leucogaster . Skull and dentition. The skull of C . atsinanana is notably more massive than in nominate pumilus ( Figure 3 ). In all cases, the average skull measurements of C . atsinanana are larger than those of nominate pumilus and for certain variables (GSL, CON INCI, MASTOID), there is no overlap in measurements between these two species ( Table 3 ). In dorsal view, there is a notable inflation of the nasal in C . atsinanana as compared to nominate pumilus ( Figure 3 ), resulting in a distinct antero-lateral expansion of this portion of the rostrum. When examined from a lateral view, the lacrimal ridge has a more angular shape in nominate pumilus , particularly along the facies orbitalis of the frontal. Further, in C . atsinanana the cranial portion of the frontal, parietal, and supraoccipital, as well as the squamosal, are notably more inflated than in nominate pumilus . In ventral view, the basisphenoid pits in C . atsinanana are larger, deeper, slightly oval, and approaching the basioccipital septum; this is in contrast to the notably smaller, shallower, rounded, and separated structures in nominate pumilus ( Figure 3 ). Further, in nominate pumilus shallow basioccipital pits are present and these structures are very indistinct or absent in C . atsinanana . The form of the palatal foramen in the two taxa is similar, being open (state 2 or 3 in character C65 of Freemen 1981), but not with a deep cleft. Some caution is needed when assessing this character, as certain skulls, which have not been thoroughly cleaned, particularly with Dermestidae beetles, retain remnant tissue and cartilage. A case in point is the holotype of C . atsinanana ( Figure 3 ), in which the remaining attachment between the premaxillae is cartilaginous. This character is also highly variable in African populations of C . pumilus ( Taylor 1999b ) . The condylocanine length (CON CANI) in C . atsinanana is on average longer than in nominate pumilus , but there is broad overlap in measures of the temporal and masseter moment arms of the mandible (MOM1 COR, MOM2 ANG) and lower toothrow. The lower molariform teeth in these two taxa have similar morphology and cusp structure, but the width of individual teeth is wider in C . atsinanana than in nominate pumilus . With one exception, the average measurements for the dental variables are larger in C . atsinanana than in nominate pumilus ; the exception is UP CANIN in nominate pumilus (sexes combined), which on average is slightly longer ( 2.6 mm ) than in female C . atsinanana ( 2.5 mm ). Hence, individual molariform teeth in C . atsinanana are distinctly more robust than in nominate pumilus , particularly when comparing adult individuals of the same sex. The P 3 in nominate pumilus is notably reduced in size, being peg-like, and is aligned in the middle of the toothrow, as compared to in C . atsinanana in which this tooth is notably larger and aligned towards the outer margin of the toothrow. In general, the cusp morphology of the upper molariform teeth of these two species are similar, although in C . atsinanana there is a distinct elongation of the portion of the tooth lingual to the commissures, particularly the hypoconal flange of M 1 the, the paracone and hypoconal flange of M 2 , and the hypoconal flange and more open commissure of M 3 . Patterns of morphological variation in members of the C. pumilus complex in the western Indian Ocean. In order to provide further insight into patterns of morphological variation in a multivariable sense of C . atsinanana with other western Indian Ocean islands and east African populations of C . pumilus sensu lato , a PCA analysis was conducted separately for males and females for different cranial and dental variables ( Table 5 ). In order to maximize the number of specimens that could be used in these comparisons, a few variables with numerous cases of missing data were excluded from the analysis. In general, two different groups occur in the projections of PC 1 plotted against PC 3 (Figure 8): a cluster of points composed of C . atsinanana from Madagascar and C . pumilus sensu lato from Kenya and a second cluster composed of C . pusillus from the Comoros and Aldabra, which in most cases included C . pumilus sensu stricto from Eritrea . For the various comparisons, with the sexes and types of variables separated, the total percentage of the explained cumulative variance for the first three axes surpasses 94%, with the first axis explaining a minimum of 82% ( Table 5 ). The second axis adds an additional 6.4—7.8 %, indicating that size is the major factor separating the two groups. FIGURE 7. View of the ventrum of three individuals of Chaerephon atsinanana sp. nov . captured the same day and from the same colony at Andasibe, showing the range of variation in ventrum coloration (left to right): Male (FMNH 184504, RHF 068), female (FMNH 184506, RHF 070), and female (FMNH 184508, RHF 072). TABLE 5 . Factor loadings of principal component analyses of cranial and dental characters of male and female Chaerephon pusillus from Aldabra and the Comoros (excluding Mayotte), C . atsinanana sp. nov . from Madagascar, C . pumilus sensu stricto from Eritrea, and C . pumilus sensu lato from Kenya. Several variables were removed from the analyses in cases of numerous missing values to maximize the number of specimens used. See Figure 9 for plots of these analyses. _____________________________________________________________________________________ Cranial variables Male Female PC 1 PC 2 PC 3 PC 1 PC 2 PC 3 GSL -0.961 0.120 0.016 -0.961 0.102 0.015 PALATE -0.813 0.556 -0.109 -0.821 0.561 -0.049 LACR WID -0.860 -0.317 -0.382 -0.910 -0.123 0.309 POST ORB -0.900 -0.153 0.311 -0.895 -0.200 -0.354 ZYGO BR -0.962 -0.038 0.021 -0.958 -0.101 0.060 MASTOID -0.949 -0.128 0.107 -0.939 -0.183 0.005 Eigen value 4.961 0.465 0.266 5.026 0.423 0.227 % total variation explained 82.7 90.5 94.9 83.8 90.9 94.7 Dental variables Male Female PC 1 PC 2 PC 3 PC 1 PC 2 PC3 C 1 -C 1 -0.891 0.442 0.090 -0.926 -0.326 -0.120 M 3 -M 3 -0.944 -0.039 -0.324 -0.948 -0.122 0.277 UP MOL -0.936 -0.264 0.169 -0.921 0.353 0.038 MTR -0.962 -0.115 0.070 -0.955 0.097 -0.196 Eigen value 3.488 0.280 0.146 3.517 0.255 0.131 % total variation explained 87.2 94.2 97.9 87.9 94.3 97.6 _______________________________________________________________________________________ On the basis of morphological characters and different cranio-dental measurements, as well as molecular genetics, it has been shown that animals referable to nominate pumilus obtained in the vicinity of the type locality in Eritrea are distinct from Malagasy specimens. The PCA analysis of cranial and dental measurements is concordant with this conclusion (Figure 8; Table 5 ), with a broad separation for both males and females for individuals collected in Eritrea and Madagascar . The single exception is a male from Eritrea , which shows, overlap for the dental variables with individuals from Madagascar and Kenya . Further, the PCA analysis closely groups individuals from the Comoros and Aldabra, which is in agreement with the molecular analysis, and gives morphological support to these populations being united under the name pusillus . The overlap between individuals from Kenya and Madagascar , which are in different clades of the C . pumilus complex, is presumed to be associated convergence in aspects of craniodental size, although, as outlined above, several morphological characters separate African and Malagasy members of this species complex. 4 3 3 2 2 1 1 0 3 3 PC PC 0 -1 -1 -2 -2 -3 -3 -4 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 -2 -1 0 1 2 3 4 PC 1 PC 1 4 4 3 3 2 2 1 1 3 3 PC 0 PC 0 -1 -1 -2 -2 Madagascar , C . atsinanana -3 -3 Kenya , C . pumilus sensu lato Comoros , C . pusillus Aldabra, C . pusillus -4 -4 Eritrea , C . pumilus sensu stricto -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 -3 -2 -1 0 1 2 3 4 PC 1 PC 1 FIGURE 8 . Projection of PC factors 1 and 3 for Chaerephon pusillus from Aldabra and the Comoros (excluding Mayotte ), C . atsinanana sp. nov . from Madagascar , C . pumilus sensu stricto from Eritrea , and C . pumilus sensu lato from Kenya . A: cranial variables - males, B: cranial variables - females, C: dental variables - males, and D: dental variables - females. See Table 5 for the PCA factor loadings. Distribution, biology and conservation status . Chaerephon atsinanana is known from numerous localities across the eastern half of Madagascar from near sea level to over 1100 m ( Figure 1 ). Of the 44 day roost sites located over the past few years, all are in synanthropic settings (schools, churches, occupied houses). Most of these buildings have distinct architectural styles, with the roofs 4—6 m off the ground, with or without air ventilation holes in the gables, and metal roofs. All of these sites are in urban or at least rural areas and outside of natural forest. During our exploration of Madagascar , we have never located a natural day-roost site of this species. Hence, we assume that since the construction of fixed permanent structures on the island, within the past hundred years or so, local populations of this species have increased and its distribution may have concordantly expanded. While there is some evidence of hunting of this species for bush-meat (Goodman et al . 2008), it is remarkably adaptable to human transformation of the natural landscapes of the island ( Harper et al . 2007 ) and occupying buildings for day roosts. Hence, there is no immediate conservation concern associated with the short or medium-term future of C . atsinanana . The same holds true for C . pusillus from the Comoros , which particularly on Mohéli , Grande Comoro, and Anjouan are abundant in certain architectural styles of abandoned or occupied buildings. In contrast, C . pusillus from the western Seychelles , specifically Aldabra Atoll, is not known to occupy the few buildings on the island, indicating that day roost sites might be more limited. Taxonomic conclusions Several authors have underlined that the current taxonomical classification of the Chaerephon pumilus sensu lato complex, as a single species, does not reflect the evolutionary history of this group (e.g., Simmons 2005 ; Taylor et al . 2009 ). Here, based on the isolation of DNA from a tissue sample recovered from an older specimen collected at the type locality (Massawa, Eritrea ), we were able to establish which clade represents C . pumilus sensu stricto . Thereafter, with a combination of molecular genetics and morphological characters several different aspects could be established: 1) C . leucogaster , which is genetically nested within C . pumilus sensu lato , was diagnostically distinct to animals classically referred to C . pumilus from eastern Madagascar and at one site on the island they are known to occur in sympatry; 2) subsequently, to refer all of the animals within the C . pumilus sensu lato clade to a single species was not concordant to phylogenetic patterns and it was necessary to consider the Malagasy population of C . ‘ pumilus ’ as a separate evolutionary unit; and 3) a range of morphological characters allow the separation of Malagasy animals from nominate pumilus and the Malagasy animals are described herein as a new species to science, C . atsinanana . Another important aspect is that size alone, specifically associated with cranio-dental measurements, should not be used as a taxonomic character, as an overlay of the phylogenetic results and morphological results indicate, for example, that similarities in size between C . atsinanana and C . pumilus from Kenya is convergent. The research presented herein is the first step in resolving a portion of the evolutionary history of the C . pumilus species complex. Movements of volant animals between continental areas and western Indian Ocean islands Even though numerous islands in the western Indian Ocean are several hundred kilometers from the nearest continental landmass, there is evidence for different animal groups of regular migration and irregular movements. This is at least in part associated with seasonal wind patterns of the Inter-Tropical Convergence Zone (ITCZ) ( Anderson 2009 ; Dijkstra 2007 ; Pedgley et al . 1995 ). Evidence for movements of bats between Africa and Madagascar can be found in the molossid Mops midas , which was formerly thought to have distinctive African and Malagasy populations. Recent research shows no morphological or genetic differentiation between these populations ( Ratrimomanarivo et al . 2007 ) and either movements of these bats across the Mozambique Channel or recent colonization of Madagascar can best explain this observation. Although the Comoros Archipelago is separated from the nearest portion of Madagascar by about 300 km of open water, several molossids are shared between these two areas ( M . leucostigma and Chaerephon leucogaster ) and show no genetic differences. Hence, these taxa either only recently colonized this archipelago or there is at least occasional dispersal movements between Madagascar and the Comoros that maintain the genetic similarity. Another example is two small species of Miniopterus (Family Miniopteridae ) bats that are shared in common between northern Madagascar and two Comorian islands (Grande Comore and Anjouan) that are morphologically and genetically similar ( Goodman et al . 2009 ; Weyeneth et al . 2008 ). None of the bat taxa mentioned in this paragraph is known from the outer western Seychelles island of Aldabra, which forms a point of a nearly equilateral triangle between northern Madagascar and the Comoros ( Figure 1 ). The case presented here for animals of the C . pumilus species complex shows a different pattern in the western Indian Ocean from those mentioned above for other bat species. The Malagasy animals, named herein as a new species endemic to the island, are morphologically and genetically distinct from regional islands and mainland Africa. The population of C . pusillus from the western Seychelles atoll of Aldabra cluster with that of the Comoros , rather than Madagascar . For numerous other species of volant vertebrates (bats and birds), the origin of the Aldabra fauna is mixed between the Comoros and Madagascar (e.g., Goodman & Ranivo 2008 ; O’Brien et al . 2009 ; Pasquet et al . 2007 ; Warren et al . 2003 ). Hence, in the case of these 9-17 g Chaerephon bats, these water barriers have been associated with the isolation and subsequent differentiation of populations after successful dispersal and colonization events. In contrast, within the Comoros , there is no apparent genetic structure between with the individual islands, which are separated by 40 and 80 km , indicating that this distance is regularly traversed by these bats and giving rise to intraarchipelago panmixia of populations. The next step Our definitions of morphological and molecular characters associated with C . pumilus sensu stricto and C . leucogaster sensu stricto ( Ratrimomanarivo et al . 2009a ; herein), as well as the delimitation of taxa in the C . pumilus - leucogaster complex from the western Indian Ocean, sets the stage for further resolution of the species-complex from the African continent. This is notably complicated by the various named forms within this complex that have been placed in synonymy and the need to either obtain sequence data from type series or topotypic material to resolve in a taxonomic sense the names that should be applied to the different clades. For example, the status of various small-sized taxa assigned to C . leucogaster from west Africa (see Rosevear 1965 ) needs to be critically evaluated. Additionally, the taxonomic status of the different C . pumilus sensu lato clades identified in Taylor et al . (2009) needs to be assessed in the light of wider molecular sampling and comparison of other types of data, such as echolocation and karyotypes.