Rhinolophus monticolus, Soisook & Karapan & Srikrachang & Dejtaradol & Nualcharoen & Bumrungsri & Lin & Aung & Bates & Harutyunyan & Buś & Bogdanowicz, 2016
Soisook, Pipat, Karapan, Sunate, Srikrachang, Mattana, Dejtaradol, Ariya, Nualcharoen, Kwan, Bumrungsri, Sara, Lin, Sai Sein, Aung, Moe Moe, Bates, Paul J. J., Harutyunyan, Margarita, Buś, Magdalena M. & Bogdanowicz, Iesław, 2016, Hill forest dweller: a new cryptic species of Rhinolophus in the ‘ pusillus group’ (Chiroptera: Rhinolophidae) from Thailand and Lao PDR, Acta Chiropterologica 18 (1), pp. 117-139 : 119-133
treatment provided by
Rhinolophus monticolus sp. nov.
PSUZC-MM.2008.1, field no. PS080123.2, adult ♂, body in alcohol, skull and baculum extracted, collected by Pipat Soisook, Piyawan Niyomwan and Tuanjit Srithongchuay on 23 January, 2008.
Pu Nam Ron Stream, Mae Nam Pha Chi Wildlife Sanctuary, Ratchaburi Province, Thailand, 13º16’N, 99º20’E, 800 m a.s.l ( Fig. 1 View FIG ).
PSUZC-MM.2013.14, field no. PS130516.1, adult ♂, body in alcohol, skull and baculum extracted, collected from Chong Yen , Mae Wong National Park, Kamphaeng Phet Province, Thailand (16°05’N, 99°06’E, 1,320 m a.s.l.) by Sunate Karapan and Pipat Soisook on 16 May, 2013; PSUZC- GoogleMaps MM.2013.16, field no. PS130816.1, lactating ♀, body in alcohol, skull extracted, collected from Chong Yen , Mae Wong National Park, Kamphaeng Phet Province, Thailand (16°06’N, 99°06’E, 1,223 m a.s.l.) by Sunate Karapan and Pipat Soisook on 16 August, 2013 GoogleMaps ; PSUZC-MM.2006.22, field no. SB060520.9 and PSUZC-MM.2006.23, field no. SB060520.10, adult ♂ ; PSUZC-MM.2006.7 field no. SB060520.6, adult ♀, body in alcohol, skull extracted, collected from Phu Soun Sai National Park , Loei Province, Thailand (17°28’N, 100°58’E, 620 m a.s.l.) by Sara Bumrungsri and Pipat Soisook on 20 May 2006 GoogleMaps ; PSUZC-MM.2007.256, field no. AD070911.9, adult ♂, body in alcohol, skull extracted, collected from Phu Soun Sai National Park , Loei Province, Thailand (17°30’N, 100°57’E, 1,300 m a.s.l.), by Ariya Dejtaradol and Pipat Soisook on 11 September 2007 GoogleMaps .
Three specimens identified as ‘ Rhinolophus cf. pusillus ’ in BOLD (not examined by authors) are referred to the new species based on DNA barcodes, namely ROM MAM 118356; INECOL M0145; IN- ECOL M0147, from Ban Nam Ya, Louang Namtha, Lao PDR, 21°3’N, 101°13’E, approx. 1,500 m a.s.l. ( Fig. 1 View FIG ), collected by Antonio Guillén-Servent on 24 April 1998.
Fifteen individuals of lactating female, which are believed to represent the new species, were captured and measured from this site. Of these, eight were taken as specimens (ROM MAM 118356; 118357; 118358; 118359; 118360; INECOL M0145; M0146; M0147) but only three, as mentioned above, were barcoded. The forearm ranged from 40.1–44.6 mm (mean 42.9 mm). Body mass ranged from 7.0– 8.2 g (mean 7.5 g) (C. M. Francis, personal communication).
This is a small-medium size horseshoe bat with a FA of 41.2–44.1 mm and SL of 17.44–19.71 mm. The anterior noseleaf is broad, 7.1–8.6 mm in width, with a long and pointed connecting process and a concave lancet. The sella is broad with almost parallel sides and a squared-off tip. The dorsal pelage comprises hairs with whitish bases (half the length) and dark brown tips. The ventral pelage is whitish for about 80% of the hair length with brown tips. The anterior and posterior median swellings are relatively large and inflated. The upper canine is large, twice the height of the second upper premolar (P 4) and with the same crown area. In occlusal view, the upper canine and the second upper premolar are somewhat rounded. The first upper premolar (P 2) is small and lies in the middle of the axis of the toothrow. The second lower premolar (P 3) is small and lies in the axis of the toothrow or slightly extruded.
The new species name, monticolus , is Latin meaning ‘mountain dweller’, which refers to the habitat where the new species is found. The proposed English name is ‘Mountain horseshoe bat’.
This is a small-medium size Rhinolophus in the ‘ pusillus group’ with a FA of 41.2–44.1 mm, HB 42.3–48.5 mm ( Table 1). The ear is relatively long, 15.5–19.4 mm. The noseleaf is broad, with the anterior noseleaf width (NL) of 7.1–8.6 mm, but does not completely cover the muzzle ( Fig. 2 View FIG a–2b). The supplementary leaflet is clearly present. The lancet is tall with slightly concave sides and a bluntly pointed tip. The connecting process is relatively long, projecting anteriorly, and sharply pointed. The sella is broad with almost parallel sides that are only slightly convex in the mid-part; the sella tip is squared-off ( Fig. 2c View FIG ). The colour of the noseleaf is brown. However, it is noteworthy that in four specimens from Thailand the noseleaf is concealed by a layer of an orange, unknown liquid ( Fig. 2 View FIG ), which can be washed off easily. The upper lip has three mental grooves. The dorsal pelage is whitish on the basal half and dark brown towards the hair tips ( Fig. 2d View FIG ). The ventral pelage is slightly paler, with whitish hair bases up to about 80% of the hair length and brown tips ( Fig. 2e View FIG ). In the wing, the shortest metacarpal is the third (3MET: 31.2–34.9 mm), whose first (3D1P) and second phalanges (3D2P) are 12.2–14.1 mm and 16.3–19.9 mm respectively in length. The fourth metacarpal (4MET) is the longest (32.0– 35.1 mm), slightly longer than the fifth (5MET) (31.9–34.8 mm). The plagiopatagium is dark brown on both sides and attached at the ankle. The uropatagium is dark brown with a relatively short tail (19.7–25.6 mm). The hindfoot (HF: 7.4– 8.6 mm) is 42.0–46.7% of the tibia length (TIB: 16.7–18.6 mm).
The skull is relatively large and elongated, with a SL of 17.44–19.71 mm and CCL of 15.30–16.52 mm ( Table 2). The sagittal crest is strong and well developed, particularly in the mid-part of the braincase; posteriorly it is attached to the lambda ( Fig. 3a View FIG ). The anterior rostral chambers are relatively large and inflated, with an AMSW of 3.22–3.63 mm and the ALSW of 4.69–5.06 mm. The posterior rostral compartment is also large and slightly inflated, in the lateral view. When viewed from above, the supraorbital depression is very shallow and the supraorbital ridges are not strong, but clearly visible. The postorbital constriction (PC) is 2.30– 2.70 mm in width. Each zygoma is robust and the jugal bone measures 1.18–1.64 mm in height at the mid-part; the zygomatic breadth (ZB) is 8.72–9.22 mm. The braincase breadth (BB) is 7.09–8.02 mm, and the mastoid width (MW) is 8.44–8.93 mm.
The upper canine is relatively large, about twice that of the second upper premolar (P 4) in height. In occlusal view, it is rounded in shape and is about equal or slightly exceeds that of the P 4 in crown area ( Fig. 4a View FIG ). The first upper premolar (P 2) is small and situated in the main axis of the toothrow. The P 4 is rounded. The maxillary toothrow length (C 1 –M 3) is 6.36–6.94 mm; the toothrows slightly converge anteriorly, with a C 1 –C 1 of 64.9–70.6% of the M 3 –M 3. The palatal length is short, 24.6–29.3% of the maxillary toothrow length. The mandible length (M) is 10.99–12.36 mm. The mandibular toothrow length (C 1 –M 3) is 6.90–7.42 mm. The lower incisors, I 2 and I 3, are about equal in size and both tricuspidate. The first lower premolar (P 2) is half the height of the third (P 4) and about one-third that of the lower canine. The P 4 is rounded and relatively small, about equal to that of the P 2 in crown area. Its height is about two-thirds that of the lower canine. The second premolar (P 3) is small and is in the main axis or slightly extruded from the toothrow. The talonids of the all three lower molars slightly exceed the crown areas of their respective trigonids.
The baculum is elongated with a cylindrical shaft. It has a distinct dorsal bend and a ventral bend, near the base and tip, respectively ( Fig. 5a View FIG ). The dorsal margin of the basal cone is shorter than the ventral margin. The basal cone of the ventral side is deeply concave. The total length of the baculum (BL) is 3.6–3.7 mm.
Rhinolophus monticolus emits typical long, narrow band ‘constant frequency’ (CF) signals, with short starting and ending ‘frequency modulated’ (FM) components. The CF segment of the five hand-held male individuals has an fmaxe between 83.6 kHz (holotype) and 87.0 kHz. The two female individuals emit a higher fmaxe of 91.6 and 93.0 kHz. One female from Lao PDR had an fmaxe of 92.0 kHz (C. M. Francis, personal communication).
Comparison with similar species
The new species is separated from other similar size species in the other Rhinolophus species groups by its triangular, pointed connecting process and the shape of the rostral swellings. It is distinguished from other species in the ‘ pusillus group’ generally by its body size. It is intermediate between the smaller R. pusillus , R. lepidus and R. shortridgei (see Table 1), and the larger species R. acuminatus . In comparison to R. lepidus , although there is some overlap, the new species is larger in forearm length (FA) particularly between specimens of the same sex ( Fig. 6 View FIG ). It is distinctly larger than R. lepidus in skull measurements; for example skull length (SL) is 17.44–19.71 mm in R. monticolus and 15.74–17.31 mm in R. lepidus ( Table 2). It should be noted that the specimens of the taxon feae, which was described from Biapo, SE Myanmar (Andersen, 1907) and currently regarded as a synonym of R. lepidus ( Csorba et al., 2003; Simmons, 2005), housed in the BMNH (see Appendix I) were examined. The measurements and skull morphology of these specimens agree with other specimens referred to R. lepidus from Vietnam, Cambodia and Thailand, and therefore are included here in R. lepidus .
The taxon refulgens found in the Thai-Malay Peninsula, however, is here considered to be a separate species from R. lepidus based on its distinctive DNA barcodes (see below) and craniodental morphology. In R. refulgens , for instance, the supraorbital region of the skull is broad with poorly defined supraorbital ridges and a very shallow depression. In R. lepidus , the supraorbital region is narrower and deeper with well defined supraorbital ridges. In the dentition, the two species differ in the length of the upper canine (C 1); in R. lepidus it is twice the height of the second upper premolar (P 4); it is shorter in R. refulgens and only slightly exceeds P 4 ( Table 3). As with R. lepidus , R. refulgens is smaller and with a higher fmaxe than the new species ( Fig. 6 View FIG , Tables 1 and 2).
Rhinolophus subbadius , another member of the pusillus species group, which was described from Nepal and is known from N Myanmar and Vietnam ( Corbet and Hill, 1992; Csorba et al., 2003), is very small with a FA of 33.5–38.0 mm and SL 14.45– 15.01 mm ( Csorba et al., 2003). Csorba et al. (2003) also noted that the narrow C 1 –C 1 of 2.7–3.2 mm and M 3 –M 3 of less than 5 mm distinguish this taxon from all other species, a view followed here.
Rhinolophus pusillus has a very wide distribution. Following Koopman (1994), eight subspecies are currently recognised, namely: blythii (NW India), gracilis (S India), szechuanus (Szechuan, China), calidus (Fujian, China), parcus (Hainan Is., China), minutillus (Anamba Is., Indonesia), pagi (Mentawai Is., Indonesia) and pusillus (Java). Yoshiyuki (1990) described the taxon lakhanae, as a subspecies of R. pusillus from Chiang Mai, N Thailand on the basis of a shorter tail and skull length and broader zygomatic width than that of szechuanus. However, the validity of the taxon lakhanae is highly questionable ( Csorba et al., 2003), and all specimens of pusillus from NE India, Myanmar, SW China and Thailand are currently referred to the subspecies szechuanus ( Lekagul and McNeely, 1977; Koopman, 1994; Csorba et al., 2003; Simmons, 2005). Specimens referred to R. pusillus , collected from Myanmar, Vietnam and Thailand ( Fig. 1 View FIG , Appendix I), are all much smaller than R. monticolus , with a FA of 34.9–39.4 mm ( Table 1), and SL of 14.93–16.09 mm ( Table 2). Although the genetic analyses suggested that they are closely related to R. monticolus (see below), the two species are clearly distinct in terms of overall size ( Tables 1 and 2) and craniodental morphology ( Table 3) and there geographical ranges overlap. The fmaxe of the two species do not overlap ( Table 1).
The new species is smaller than R. acuminatus , which has a FA 45.7–50.0 mm and a SL 18.99–22.03 mm (n = 7). Rhinolophus acuminatus also differs from R. monticolus in having a very short lancet with concave sides and a very narrow pointed tip; more distinctly inflated anterior median swellings; very well developed sagittal crest; and extruded second lower premolar (P 3) (see also, Csorba et al., 2003). Unlike R. monticolus , R. acuminatus has a distinct musky smell.
Rhinolophus monticolus most closely resembles R. shortridgei , which was described from Bagan (= Pagan), Myanmar, and is currently known from India, Myanmar and China ( Chiozza, 2008). Although there is some overlap, the new species can be distinguished from R. shortridgei by a combination of its larger size (e.g., forearm and skull length — Fig. 6 View FIG ), craniodental and bacular morphology, and echolocation and genetics (see below). In R. monticolus , the FA averages 42.3 mm (41.2–44.1 mm), which is larger than R. shortridgei (mean 40.5 mm; range 38.2–42.1 mm, Table 1). In the noseleaf, the sella shape of R. shortridgei , as described in Csorba et al. (2003), is parallel-sided with a broadly rounded tip, but it is squared-off in R. monticolus ( Fig. 2c View FIG ). In the skull, the profile of the rostrum differs between R. monticolus and R. shortridgei ; the anterior lateral chambers of R. monticolus are more inflated and clearly distinguishable ( Fig. 3 View FIG ). In R. shortridgei , the posterior compartments are smaller and less inflated. Therefore, the supraorbital depression is deeper with more distinct supraorbital ridges than in R. monticolus ( Fig. 3 View FIG ). The skull length (SL) of R. monticolus (17.44–19.71 mm), although with some overlap, is larger than that of R. shortridgei (17.17–18.60 mm). The upper canine of R. monticolus is rounded and is about equal or slightly exceeds that of the second upper premolar (P 4) in crown area, but in R. shortridgei , it is more massive, more elongated and considerably exceeds the size of the P 4, such that the crown area of P 4 is less than two-third that of the C 1 ( Fig. 3 View FIG ). In R. monticolus , the second (middle) lower premolar (P 3) is in the main axis or only slightly extruded from the toothrow; however, it is half-way or fully extruded in R. shortridgei ( Fig. 4 View FIG ). The shape and size of the third lower premolar (P 4) is also different; it is more rounded and about equal in size to the first (P 2) in R. monticolus , but in R. shortridgei , it is more elongated and exceeds the size of P 2, such that P 2 is only about two-thirds the crown area of P 4 ( Fig. 3 View FIG and Table 3).
According to the PCA of 11 craniodental measurements, the new species overlaps with R. shortridgei but not with other species in the same group ( Fig. 7 View FIG ). The first three components of the PCA explain 96.4% of the total variance ( Table 4). The PC1 explains 88.5% of the total variance, and discriminates species having a short skull (SL and CCL), short upper toothrow (C 1 –C 3) and narrow rostral swelling (ALSW). The PC2 (5.7%) discriminates species on the basis of the width of the postorbital constriction (PC), whereas the PC3 (2.3%) differentiates species having broad palatal width (C 1 –C 1).
The baculum of R. monticolus is slightly shorter than that of R. shortridgei (BL 3.6–3.7 mm versus 4.0 mm), with a straighter shaft. In R. shortridgei , the baculum has a distinctly curved shaft particularly in the mid part. The dorsal margin of the basal cone of R. monticolus is higher and more inflated. Therefore, the basal cone of R. monticolus is wider in lateral view ( Fig. 5 View FIG ). The basal cone of R. monticolus is also different from that of R. lepidus and R. pusillus as illustrated in Csorba et al. (2003).
In terms of echolocation, based on individuals collected from Bagan and South Sagaing Division, Myanmar, male specimens of R. shortridgei have a fmaxe of 94.5–97.9 kHz (n = 4) and females, 98.8– 100.7 kHz (n = 2). This is higher than that of R. monticolus from Thailand and Laos ( Table 1).
Based on the habitat, R. monticolus is likely to be a forest-dwelling bat (see below) at higher elevations and may roost in hollow trees or rock crevices. This contrasts to the habitat of R. shortridgei in Myanmar where specimens were collected from pagodas, limestone caves or karst habitats at lower elevations. The features that discriminate between R. monticolus and other species in the pusillus group in Thailand are summarised in Table 3.
Genetic relationships and taxonomic remarks
The phylogenetic trees inferred from both cyt b and COI are very similar whether using Bayesian or Maximum Likelihood methods. Therefore, only the results of Bayesian analysis are presented here ( Figs. 8 View FIG and 9 View FIG ). The genetic analyses, especially of COI, suggest that R. pusillus and its various subspecies may represent a species complex ( Fig. 9 View FIG ), but addressing that complex is beyond the objectives of this study, so for convenience, all of these specimens are here referred to R. pusillus , followed by a number, where appropriate, to indicate different genetic branches that may or may not represent distinct taxa. Both Bayesian phylogenetic analyses of cyt b ( Fig. 8 View FIG ) and of COI gene sequences ( Fig. 9 View FIG ) indicate that R. monticolus groups with other species in the ‘ pusillus- group’. The new species clusters in a monophyletic clade with R. shortridgei from Myanmar and is closely related to several branches of the R. pusillus species complex from elsewhere ( Figs. 8 View FIG and 9 View FIG ). It differs genetically from R. shortridgei from Myanmar by 2.36% based on cyt b ( Table 5) and by 1.55% based on COI ( Table 6). It is interesting to note that although clearly different in term of body and skull size, it is more similar genetically to R. pusillus from Thailand and Lao PDR (cyt b 2.20%, Table 5; COI 1.35%, Table 6) than to R. shortridgei . Moreover, the genetic divergence between R. shortridgei from Myanmar and R. pusillus (from Thailand) is also very small, only 1.76% based on cyt b ( Table 5). The same pattern is also found in COI, as R. shortridgei differs from R. pusillus 4 from Lao PDR and Thailand by 1.88% ( Table 6).
The clade of R. pusillus 5 from Lao PDR differs from the new species by 1.55% based on COI and is clustered in a sister clade to the new species ( Fig. 9 View FIG and Table 6). The genetic divergence is similar to that between the new species and R. shortridgei ( Table 6). It is possible that these specimens ( R. pusillus 5 from Lao PDR) belong to R. monticolus but it cannot be confirmed without direct comparison of specimens. This supports the view of Francis et al. (1999) that specimens referred to R. pusillus from Lao PDR represent at least two or three species.
In comparison with small Rhinolophus from East Asia, R. monticolus differs from R. pusillus and R. blythii from China by 4.35% and 4.14%, respectively. It differs from R. monoceros from Taiwan by 3.89%, and from R. cornutus from Japan by 3.90%. In general, genetic differences between taxa analysed in this group range from as little as 1.04% ( R. pusillus [ China] vs R. blythii ) to 4.35% ( R. monticolus vs R. pusillus [ China]). It is noteworthy that the difference among East Asian taxa, R. pusillus / R. blythii from mainland China and the two island taxa, R. monoceros from Taiwan and R. cornutus from Japan, are also relatively small, only 1.04– 2.81% ( Table 5). Li et al. (2006) suggested that this monophyletic group of taxa, with little genetic divergence and overlapping echolocation call frequencies, should be considered as subspecies (i.e., they suggest the taxa cornutus and monoceros are island subspecies of R. pusillus ). However, Wu et al. (2012), based on morphometric analysis, showed distinct morphological differences and considered R. cornutus in Japan and R. monoceros in Taiwan to be distinct from R. pusillus found in mainland China and Hainan Island.
As already highlighted in Francis et al. (2010), the genetic analyses of this study also suggests that there is a need for a future investigation of the taxonomy and phylogeography of the R. pusillus species complex, particularly to examine subspecific affinities, and whether it is actually a species complex. Currently, it is known that there is considerable variation in morphometrics, genetics and call frequency between geographical populations of R. pusillus from throughout its range.
According to the COI analysis of this study, it is interesting to note that R. lepidus (sensu lato) is separated into two clades, the Indochinese and Sundaic. The population of the latter clade is usually referred to the subspecies ‘ refulgens ’ (e.g., Corbet and Hill, 1992; Simmons, 2005; Francis, 2008). The genetic divergence between the two clades is of 2.85% ( Table 6). In comparison with other species, this divergence is greater than that between R. monticolus and the most similar species R. shortridgei (1.55% — Table 6). Although ‘ R. lepidus ’ and ‘ refulgens ’ are very similar in terms of morphometrics ( Fig. 7 View FIG ), the female specimens of ‘ refulgens ’ average smaller in forearm and skull lengths ( Fig. 6b View FIG ), whereas the males are very similar in size ( Fig. 6a View FIG , Tables 1 and 2). They also differ in morphology of the supraorbital region in a way that is comparable to R. monticolus versus R. shortridgei . Based on the distinctive DNA barcodes and morphology, we here regard the taxon ‘ refulgens ’ as a separate species.
Distributional, ecological and conservation notes
Rhinolophus monticolus is known from evergreen forest in the mountainous areas of Mae Nam Pha Chi Wildlife Sanctuary, Mae Wong National Park along the Thai–Myanmar border; and from Phu Suan Sai National Park near the Thailand and Lao PDR border. Specimens were collected at elevations between 620–1,320 m a.s.l. However, further surveys in higher elevation forests elsewhere in Southeast Asia may extend the known range of this species.
The specimen of R. monticolus from Mae Nam Pha Chi WS was collected in a harp trap, which was set in a valley across a seasonal streamlet, near the headwater of the Pu Nam Ron Stream. The area is surrounded by dense evergreen forest and steeply scarped rocky mountains. Rhinolophus monticolus was found together with another five species of insectivorous bats (see Materials and Methods section). A male specimen of R. monticolus from Mae Wong NP was captured by hand in an underground pipe and another lactating female was captured in a mist net set across a streamlet in evergreen forest with many wild banana trees. The latter two specimens were both found on a cold, windy and rainy evening, near the top of a hill surrounded by hill evergreen forest. Other bat species that were found at this site included Rhinolophus affinis , Sphaerias blanfordi and Macroglossus sobrinus .
The new species appears to be restricted to forest habitats. Since there were no caves near capture sites, its diurnal roosts probably comprise hollow trees and/or rock crevices. Although it is found in protected areas, the status of this species could potentially be at risk owing to its low abundance (reflected in its low capture rate) and the rapid rate of forest conversion in Thailand.
University of Montpellier
No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.