A new Pipistrelle bat from the oceanic Island of Príncipe (Western Central Africa)

Abstract

We describe a population of pipistrelle-like bats from Príncipe Island (Gulf of Guinea, Western Central Africa) as a new species based on the molecular and morphological characteristics of six specimens collected more than 30 years ago. The description of this new species was not possible until the traditionally entangled systematics of the whole pipistrelle group was clarified in recent years with the inclusion of molecular techniques and adequate species sampling. In this new taxonomic framework, the new species was clearly included within the dark-winged group of the recently described genus Pseudoromicia. The pipistrelles from Príncipe Island present a moderately inflated skull in lateral view with inner upper incisors that are moderately bicuspids and a baculum distinctly long with expanded tips. Besides these morphological characters, the new bat species is distinguished by its dwarfism, being the smallest species recognized within the genus. The ecology and conservation status of this endemic island species are unknown and field studies are urgently needed to evaluate the situation and conservation threats to this new species in its natural habitat.

The systematics of the vespertilionid bats (Family Vespertilionidae) has been entangled in unstable taxonomic arrangements for many years (Hoofer et al. 2003; Monadjem et al. 2021b). During the last two centuries, the taxonomic position of a number of species—all showing a conservative and similar morphology—has varied considerably and only recently has the application of molecular techniques produced a more stable overall picture for the entire family (Hoofer and Van Den Bussche 2003; Hoofer et al. 2006; Rohers et al. 2010).

The situation was particularly confused within the group of small pipistrelle-like forms. These bats were placed in dozens of taxa differentiated by subtle details in external characters. Many characteristics were subjective (like the color of the pelage) and/or of poor taxonomic value (like the presence or absence of a tiny second upper premolar used to distinguish Pipistrellus from Eptesicus). The comparative study of baculum morphology allowed Hill and Harrison (1987) to establish the first modern arrangement of the entire subfamily Vespertilioninae, with quite radical results in relation to former traditional arrangements. Hill and Harrison (1987) dismantled genera, described a genus (Nycticeinops) and new subgenera (e.g., Neoromicia), and changed the species assignment of numerous taxa (e.g., the species Eptesicus capensis was changed to Pipistrellus (Neoromicia) capensis). They maintained the generic separation of Eptesicus and Pipistrellus but modified significantly their species composition and recognized seven subgenera within Pipistrellus. In particular, they defined the subgenus Neoromicia within Pipistrellus to include all African pipistrelle-like bats that did not have a second upper premolar (e.g., Pipistrellus (N.) tenuipinnis), and that had been previously placed in Eptesicus. Interestingly, most of the changes they proposed that were based on similarities in baculum morphology were later supported by comparative cytogenetic studies (e.g., Volleth et al. 2001; Kearney et al. 2002). In fact, Neoromicia was not only recognized, but later raised to generic rank based on several unique karyotype characters (Kearney et al. 2002). This arrangement of the African pipistrelle-like forms was included in general mammalian studies (Moratelli and Burgin 2019), although admitting it was still “jumbled and confusing.”

Recent molecular approaches focusing on the Vespertilionidae (e.g., Hoofer and Van Den Bussche 2003; Koubínová et al. 2013), and particularly a series of recent papers focusing on the African pipistrelle-like bats (Monadjem et al. 2013; Decher et al. 2015; Monadjem et al. 2021a, 2021b), have established a stable taxonomic reorganization of these forms and their systematics at a supra-specific level. These studies, using molecular markers summarized in Monadjem et al. (2021a), suggest: (1) a sister relationship between the groups Neoromicia, Nycticeinops, and Hypsugo; (2) the split of the former genus Neoromicia into four differentiated genera (Neoromicia, Laephotis, Pseudoromicia, and Afronycteris); and (3) the description of several new species within these new genera, also supported by the shape of skulls and bacula.

In a survey of the bats of the small oceanic island of Príncipe, Gulf of Guinea, Western Central Africa (Fig. 1), which we carried out in the late 1980s, a few specimens of pipistrelloid bats (sensu Hoofer and Van Den Bussche 2003) were collected and deposited in the EBD-ICTS collections of the Doñana Biological Station (CSIC; Juste and Ibáñez 1994). According to external characters and baculum morphology, these bats could not be assigned to any of the recognized forms at that time (Juste 1990), and it could only be verified that they belonged to the Neoromicia group of the genus Pipistrellus, as defined by Hill and Harrison (1987). More than 30 years later, and with the molecular and morphological information now available, we have been able to complete the characterization of the Príncipe’s pipistrelles and compare their DNA with the now available homologous sequences from the mainland African species. These analyses and the comparisons with morphologically similar species have confirmed that the pipistrelles of Príncipe Island belong to a different lineage, one that is molecularly and morphologically distinct than any other known so far. They are therefore described here as a new species.

Fig. 1.

Map of Príncipe Island and distances to the mainland and other islands in the Gulf of Guinea (Bioko and São Tomé).

Materials and Methods

Specimens.

A total of six pipistrelle-like bats captured in Príncipe Island and deposited at the EBD-ICTS (CSIC) collections (EBD 17358M, EBD 17475M, EBD 29604M, EBD 29605M, EBD 29606M, EBD 29607M) were captured in 1988 and 1992 in two localities along the banks of the Papagaio river in Santo António, Príncipe Island (Fig. 1). Sequences of three specimens (EBD 29604M, EBD 29605M, EBD 29606M) were compared to homologous sequences obtained from GenBank that represent six out of the seven species known thus far that form the new genus Pseudoromicia, together with representatives of closely related genera (Table 1). Príncipe Island’s specimens were compared morphologically with specimens of the pipistrelle species of the nearby island of Bioko (EBD 19107M), and with the following African pipistrelles-like taxa: Hypsugo (cf.) musciculus, P. brunnea, P. tenuipinnis, Laephotis capensis, Nycticeinops crassulus, and Afronycteris nana. These specimens were collected in the mainland lowland rainforest and are housed at the EBD-ICTS (CSIC) collections (see Appendix I). They were also compared with related species within the genus Neoromicia (sensu lato) housed at the Natural History Museum (BMNH) in London, now considered as: P. tenuipinnis (BMNH 75.2801, BMNH 26.7.6.81), P. rendalli (BMNH 61.50), and P. brunnea (BMNH 98.5.4.18), L. capensis (BMNH 35.1.6.39), Neoromicia zuluensis (BMNH 75.2149), N. guineensis (BMNH 53.89), and N. somalica (BMNH 75.2802).

Table 1.

Pipistrelle-like bats of the genera Afronycteris, Hypsugo, Laephotis, Neoromicia, Nycticeinops, Pipistrellus, and Pseudoromicia used in the molecular analyses of this study, along with their associated GenBank accession numbers for the used Cytb marker. EMNH, Eswatini National Museum of Natural History, Kwalusen; FMNH, The Field Museum of Natural History, Chicago; IVB, Institute of Vertebrate Biology, Brno; NMK, National Museums of Kenya, Nairobi.

Species	GenBank number	Country	Voucher	Reference
Afronycteris nana	EU797428	Ethiopia	AK21161	Trujillo et al. (2008)
Afronycteris nana	MT777844	Democratic Republic of Congo	FMNH 173297	Monadjem et al. (2021a)
Afronycteris nana	MT777848	Guinea	EMNH ZWW202	Monadjem et al. (2021a)
Hypsugo savii	AJ504450	Switzerland	-	Stadelmann et al. (2004)
Laephotis botswanae	EU797444	South Africa	TM 38153	Trujillo et al. (2008)
Laephotis capensis	KM886073	Botswana	TM 48485	Goodman et al. (2015)
Neoromicia somalica	MT777973	Kenya	FMNH 221017	Monadjem et al. (2021a)
Neoromicia somalica	MT777976	Kenya	FMNH 221026	Monadjem et al. (2021a)
Neoromicia zuluensis	MT777997	Kenya	NMK185143	Monadjem et al. (2021a)
Neoromicia zuluensis	MT778004	Kenya	FMNH 220919	Monadjem et al. (2021a)
Nycticeinops grandidieri	MT778013	Kenya	FMNH 216112	Monadjem et al. (2021a)
Pipistrellus hesperidus	AJ841968	South Africa	MR-M987	Stadelmann et al. (2004)
Pseudoromicia brunnea	MT777937	Guinea	ZWW136	Monadjem et al. (2021a)
Pseudoromicia brunnea	MT777939	Guinea	ZWW282	Monadjem et al. (2021a)
Pseudoromicia kityoi	MT777940	Uganda	FMNH 223211	Monadjem et al. (2021a)
Pseudoromicia kityoi	MT777941	Uganda	FMNH 223555	Monadjem et al. (2021a)
Pseudoromicia nyanza	MT778043	Kenya	FMNH 215628	Monadjem et al. (2021a)
Pseudoromicia nyanza	MT778044	Kenya	FMNH 215629	Monadjem et al. (2021a)
Pseudoromicia rendalii	MT778054	Kenya	FMNH 216159	Monadjem et al. (2021a)
Pseudoromicia rendalii	JX276206	Senegal	IVB S1212	Koubínová et al. (2013)
Pseudoromicia roseveari	MT777942	Guinea	ZWW251	Monadjem et al. (2021a)
Pseudoromicia roseveari	MT777943	Guinea	ZWW263	Monadjem et al. (2021a)
Pseudoromicia tenuipinnis?	JQ956446	Central Afr. Rep.?	Isolate RCA322	Maganga et al. (2014)

Molecular analyses.

Total genomic DNA of three specimens of Príncipe’s pipistrelles was extracted from tissue samples preserved in 70% ethanol after an isopropanol DNA precipitation with saline purification, following Gemmells and Akiyama (1996). Since the mitochondrial gene cytochrome b (Cytb) was the marker selected by most of the previous studies, this gene was amplified with the primers MOLCIT-F (5ʹ-AATGACATGAAAAATCACCGTTGT-3ʹ) and CYTb-H (5ʹ-CTTTTCTGGTTTACAAGACCAG-3ʹ) from Ibáñez et al. (2006) and Weyeneth et al. (2008), respectively. The PCR reaction (20 µl final volume) included 2 µl of DNA extract, 1 µl of each primer (10 µM), 0.8 µl of MgCl₂ (50 mM), 0.16 µl of dNTP (25 mM), 0.2% BSA, 0.5 units of Taq Polymerase. Thermo-cycling consisted of 4-min initial denaturation at 94°C followed by 35 cycles of 45 s at 94°C, 45 s at 52°C, and 60 s at 72°C and a final extension of 5 min at 72°C. Products were sequenced in an ABI 3100 automated sequencer (PE Biosystems, Warrington, United Kingdom) following the manufacturer’s protocols. The sequences were checked visually with Geneious software (Kearse et al. 2012). The molecular sequences generated by this study have been deposited in GenBank under the accession numbers (MW287571, MW287572, and MW287573).

The sequences were aligned with the Clustal W algorithm using MEGA 6.0 (Tamura et al. 2013) together with the homologous fragments recovered from GenBank, for a total of 23 African pipistrelles-like species (Table 1) related to the Neoromicia group as defined by Hill and Harrison (1987). They were selected according to their availability and to their similarity index using nucleotide–nucleotide standard BLAST in the GenBank database (Altschul et al. 1990). The sampling covered six related bat genera according to the recent phylogenetic proposals summarized in Monadjem et al. (2021a). Evolutionary relationships were inspected according to Bayesian and maximum likelihood (ML) criteria, using Pipistrellus hesperidus as outgroup and a GTR+I+G evolutionary model selected by jModelTest2 v0.1.10 (Darriba et al. 2012) based on BIC criterion. Bayesian phylogenetic trees were obtained using MrBayes v3.2.1 (Ronquist et al. 2012) after running 10⁷ generations with four chains, sampling every 200th generation in two simultaneous runs, and discarding the first 25% of trees as burn-in. Under ML, trees were obtained after heuristic search with an initial BioNJ tree and under tree and length default optimization criteria. Bootstrap resampling was used with 1,000 replicates to evaluate the tree nodes (Felsenstein 1985) and run in PhyML 3.0 (Guindon 2010). Trees and posterior probabilities were visualized with Figtree v1.3.1. (http://tree.bio.ed.ac.uk/software/figtree/).

Codon-based partitions were not used in reconstructions (despite the potential gain of resolution especially for deeper nodes), in order to make our results directly comparable with previously published phylogenies of the group and that did not use this analysis. Finally, levels of genetic differentiation between groups were estimated using MEGA 6.0 (Tamura et al. 2013) according to Kimura 2-parameter (K2P) evolution model since this estimate of genetic distance is conducive to comparison from other studies.

Morphological analyses.

Qualitative variation in skull and external morphology was determined by directly comparing specimens. We measured 14 external and 16 craniodental characters with a digital caliper (to the nearest 0.1 mm for external and 0.01 mm for skull characters). The morphological traits and their abbreviations are as follows: body weight (W, expressed in g); whole body length (HBL); tail length (TL); forearm length (FA); hind foot length, including claws (HF); tibia length (TIB); ear length (EAR); tragus length (TRA); third finger metacarp (3FM); third finger phalange (3FPH1); forth finger metacarp (4FM); forth finger phalange (4FPH1); fifth finger metacarp (5FM); fifth finger phalange (5FPH1); greatest length of skull, excluding incisors (GLS); condylo-basal length (CBL); condylo-canine length (CCL); palatal length (PL); maxillary toothrow length (CM³); width across the upper molars (M³M³); width across the upper canines (C¹C¹); zygomatic breadth (ZB); greatest braincase width (GBW); greatest mandible length excluding incisors (MAND); postorbital width (POB); rostrum width (RW); maxillary toothrow length (CM³); distance from the 3rd upper molar to the upper incisor (IM³); mandibular toothrow length (CM₃), lower dental raw (IM₃); upper molar series (molars + premolar; SPM). More detailed definitions of these measurements are found in Ruedi et al. (2012) and Tungaluna et al. (2013). Dentition morphology was studied in detail and the teeth for each individual were carefully measured with a Zeiss 2 stereo microscope under 1.6× to 6× magnifications. The complete set of body, cranial, and dental measurements are reported in Supplementary Data SD1 and SD2.

The baculum of one of the two males (EBD 17358) was removed and stained with alizarin red after a 5% KOH maceration following Hill and Harrison (1987) and Kearney et al. (2002). The stained baculum was dissected and preserved in 100% glycerol after being drawn, measured, and photographed under a Zeiss 2 stereo microscope.

Echolocation analyses.

A total of 40 echolocation call sequences from free-flying Príncipe Pseudoromicia bats were recorded in October 1992 in the city of Santo António and surroundings. These calls could not be misidentified with those from any other bat because the Príncipe Pseudoromicia is the only vespertilionid bat on Príncipe Island. Moreover, many of the calls were recorded from bats that we saw flying in the open at dusk, while we were walking along the Papagaio riverbanks to the state “Roça Bela Vista” (3 km SSW from Santo António). Calls were recorded with a D960 ultrasound detector (Pettersson Elektronik AB, Uppsala) in time expansion mode (10×) onto metal-XR Sony tapes with a Sony Professional Walkman WM-D6C cassette recorder. The recordings were analyzed with BatSound Pro 4.00 software (Pettersson Elektronik AB, Uppsala) using a 1,024 pt. FFT Hanning window. A single pulse from each call sequence was randomly selected and close call sequences were avoided in time to minimize the risk of pseudo-replication. The sonograms and spectrograms of each of the 40 selected pulses were eye-inspected and manually measured.

Results

Phylogenetic analyses.

The final data set consisted of sequences of the nearly complete (1,134 bp) mitochondrial Cytb for the three specimens of pipistrelles from Príncipe Island and 23 sequences downloaded from GenBank (Table 1). Four highly supported clades, corresponding to the new genera (Pseudoromicia, Afronycteris, Neoromicia, and Laephotis), were shown in the reconstructions by both reconstruction criteria (Fig. 2). The topologies supported a sister relationship between Afronycteris and Pseudoromicia on one hand, and between Neoromicia and Laephotis on the other, as suggested in other studies (e.g., Monadjem et al. 2021a). The samples from Príncipe Island clearly fall into their own well-supported group within a clade, corresponding to the genus Pseudoromicia. This clade includes the species P. roseveari and P. kityoi and more basally P. brunnea (Fig. 2). The sequence belonging to P. tenuipinnis, according to GenBank (JQ956446), clustered between two sequences of other species (P. rendalli). In fact, the two taxa were differentiated by a genetic distance of only 1.7%. Since both taxa are well-accepted and recognized species, we suspected that the original specimen (Maganga et al. 2014), from an unknown locality, was misidentified. Thus, we decided to remove the Cytb sequence from this particular P. tenuipinnis, which is deposited in GenBank (JQ956446), from further analyses. Genetic distances among the four genera range between 15.7 and 18.3% according to Monadjem et al. (2021a; Table 2). Within the genus Pseudoromicia, the species we studied show values of K2P genetic distances (Table 2) that vary from 5.6% between P. kityoi and P. roseveari to 18% between P. kityoi and P. nyanza, values slightly larger than those presented by Monadjem et al. (2021b). The shortest genetic distance of the sequences from Príncipe Island with any of the known species was 9.9% (with P. brunnea) and averaged 13.1% for all species studied. This distance value clearly supports species recognition of this isolated Príncipe’s population, whereas the topology (Fig. 2) points to the closest relationship to the recently described species P. roseveari and P. kityoi (Monadjem et al. 2013, 2021a) and P. brunnea, all members of the dark-winged group of species recognized within Pseudoromicia (Monadjem et al. 2021a). This arrangement gives phylogenetic support to the morphological distinction between dark-winged and white-winged bats within Pseudoromicia (Fig. 2).

Fig. 2.

Bayesian phylogenetic reconstruction based on the mitochondrial gene cytochrome b (Cytb). Values above branches represent Bayesian posterior probabilities and values below branches are bootstrap values of maximum likelihood (support values < 50 are not shown). The right column presents sequences GenBank accession numbers. The sequence JQ956446 is listed in GenBank as Pseudoromicia tenuipinnis but it was probably misidentified. See Table 1 for a complete list of the specimens used and main text for details of the molecular analyses.

Table 2.

Genetic distances (Kimura 2-parameter [K2P] distances) based on the complete mitochondrial Cytb gene among the species studied within the genus Pseudoromicia.

	P. principis sp. nov.	P. kityoi	P. nyanza	P. rendalli	P. roseveari
P. kityoi	0.112	—	—	—	—
P. nyanza	0.161	0.180	—	—	—
P. rendalli	0.165	0.163	0.154	—	—
P. roseveari	0.101	0.056	0.158	0.153	—
P. brunnea	0.099	0.089	0.148	0.143	0.084

Morphological analyses.

The bats from Príncipe Island exhibit a rounded head (Fig. 3) and ears, with a narrow tragus with a notch at the middle of the posterior margin. Their skulls are moderately inflated in lateral view (Fig. 4) and differ from the flat silhouette of the genera Laephotis and Neoromicia (sensu stricto). Additionally, the baculum of the male we examined from Príncipe Island is distinctly long with expanded tips (Fig. 5) and completely different from the spatulate-like baculum reported for Neoromicia (Hill and Harrison 1987; Monadjem et al. 2013; Van Cakembergue and Happold 2013). These characters indicate that they belong to the recently described genus Pseudoromicia (Monadjem et al. 2021a), which agrees with our molecular data, although the pipistrelles from Príncipe Island are slightly smaller in all measurements (external, cranial, and dental) than the any other species of Pseudoromicia (Tables 3 and 4).

Fig. 3.

Portrait of Pseudoromicia principis sp. nov. showing a typical pipistrelle-like appearance. Pelage is chocolate brown dorsally and ventrally although tips of the ventral pelage present a distinct light creamy tinge. Photograph by Jorge Palmeirim.

Fig. 4.

Skull and mandible of the holotype of Pseudoromicia principis sp. nov. (EBD 17475M) in dorsal, ventral, lateral, front and back views as well as mandible top view. Photographs by Joxerra Aihartza.

Fig. 5.

(A) Dorsal; (B) lateral; (C) ventral; (D) apical (above), and basal (bellow) views of the baculum of the paratype Pseudoromicia principis sp. nov. (EBD 17358M). Photographs by Laura Torrent.

Table 3.

Descriptive statistics (mean ± SD, minimum and maximum values, and number of measured specimens) at six external measurements recorded for Pseudoromicia principis sp. nov. and other members of the black-winged group of the genus Pseudoromicia. Pseudoromicia tenuipinnis is also included for comparisons. See text for definitions of the variables. Measurements for P. brunnea, P. kityoi, P. roseveari, and P. tenuipinnis after Monadjem et al. (2021a).

External variable	P. principis sp. nov.	P. brunnea	P. kityoi	P. roseveari	P. tenuipinnis
FA	31.42 ± 0.74 (30.5–32.3) N = 6	34.6 ± 1.06 (32.8–36.7) N = 15	37.5 (37–38) N = 2	36.5 ± 1.54 (32.6–38.0) N = 10	29.4 ± 1.18 (28.1–32.0) N = 6
W	4.36 ± 0.38 (3.8–4.8) N = 5	6.0 ± 1.04 (4.8–9.4) N = 15	7.95 (7.9–8.0) N = 2	6.5 ± 0.50 (6.0–7.0) N = 10	4.0 ± 0.29 (3.5–4.4) N = 6
HBL	71.56 ± 4.33 (64.5–75) N = 5	84.8 ± 3.14 (80–89) N = 15	88.5 (88–89) N = 2	87.3 ± 2.95 (83–93) N = 10	74.2 ± 2.17 (72–77) N = 5
TL	34.3 ± 3.23 (30–37.5) N = 5	35.9 ± 1.55 (33–38) N = 15	35.0 (34–36) N = 2	39.7 ± 2.65 (34–44) N = 9	29.2 ± 0.82 (28–30) N = 5
EAR	7.62 ± 0.93 (6.5–8.7) N = 6	12.4 ± 0.83 (11–14) N = 15	10.0 (10– 10) N = 2	12.9 ± 0.64 (12–14) N = 8	12.7 ± 0.91 (12–14) N = 7
HF	5.88 ± 0.38 (5.4–6.5) N = 6	7.79 ± 0.97 (7–10) N = 14	9.5 (9–10) N = 2	9.9 ± 1.10 (8–11) N = 10	6.5 ± 0.80 (5–7) N = 7

Table 4.

Descriptive statistics (mean ± SD, minimum and maximum values, and number of measured specimens) at eight cranial measurements recorded for Pseudoromicia principis sp. nov. and the other members of the black-winged group of the genus Pseudoromicia. The small P. tenuipinnis is also included for comparisons. See text for definitions of the variables. Measurements for P. brunnea, P. kityoi, P. roseveari, and P. tenuipinnis after Monadjem et al. (2021a).

Skull variable	P. principis sp. nov.	P. brunnea	P. kityoi	P. roseveari	P. tenuipinnis
GLS	12.68 ± 0.24 (12.3–12.9) N = 6	13.73 ± 0.35 (13.10–14.24) N = 15	14.84 (14.7–14.99) N = 2	14.13 ± 0.45 (13.4–14.5) N = 10	12.43 ± 0.28 (12.00–12.78) N = 10
ZW	7.2– N = 1	8.68 ± 0.49 (7.70–9.61) N = 14	9.6 (9.54–9.63) N = 2	8.85 ± 0.5 (8.00–9.50) N = 9	7.48 ± 0.5 (6.8–8.19) N = 9
POB	3.33 ± 0.05 (3.3–3.4) N = 6	3.92 ± 0.23 (3.60–4.45) N = 15	3.95 (3.91–3.99) N = 2	3.80 ± 0.19 (3.50–4.09) N = 10	3.79 ± 0.2 (3.47–4.25) N = 9
GBW	6.12 ± 0.26 (5.8–6.5) N = 6	7.50 ± 0.24 (6.65–7.69) N = 15	7.59 (7.51–7.67) N = 2	7.31 ± 0.35 (6.87–7.40) N = 10	6.51 ± 0.13 (6.40–6.81) N = 10
MAND	8.93 ± 0.21 (8.6–9.2) N = 6	10.06 ± 0.32 (9.40–10.66) N = 14	10.98 (10.76–11.2) N = 2	10.4 ± 0.3 (9.8–10.7) N = 9	8.77 ± 0.38 (8.2–9.33) N = 10
CM3	4.23 ± 0.1 (4.1–4.3) N = 6	4.86 ± 0.13 (4.60–5.07) N = 15	5.16 (5.12–5.20) N = 2	5.03 ± 0.2 (4.80–5.30) N = 10	4.23 ± 0.14 (3.90–4.35) N = 10
C1C1	3.78 ± 0.15 (3.7–4.0) N = 6	4.22 ± 0.24 (3.63–4.63) N = 15	4.67 (4.61–4.74) N = 2	4.34 ± 0.25 (3.8–4.7) N = 10	3.81 ± 0.29 (3.3–4.14) N = 9
M3M3	5.15 ± 0.19 (4.9–5.4) N = 6	5.86 ± 0.26, (5.50–6.55) N = 15	6.19 (6.09–6.1) N = 2	6.00 ± 0.25 (5.5–6.3) N = 10	5.03 ± 0.18 (4.70–5.29) N = 9
CM3	4.78 ± 0.29 (4.4–5.3) N = 6	5.30 ± 0.33 (4.80–5.86) N = 15	5.61 (5.59–5.63) N = 2	5.25 ± 0.32 (5.0–5.97) N = 9	4.63 ± 0.23 (4.4–5.24) N = 10

Therefore, the combination of unique characters, together with the level of genetic differentiation, supports assigning species status to the isolated and unnamed pipistrelles from Príncipe Island that we formally describe here.

Pseudoromicia principis, new species

Príncipe’s Pipistrelle

Figs. 3–6

Fig. 6.

(A) Detail of the upper incisors (top left) and drawings of the (a) lower incisors and (b) upper incisors (bottom) of the holotype of Pseudoromicia principis sp. nov. (EBD 17475M). Photograph by Joxerra Aihartza and drawings by Joaquín López-Rojas. (B) Details of the rhinarium of (c) the holotype of P. principis sp. nov. (EBD 17475M) and of (d) Nycticeinops happoldorum (ZFMK-MAM-2009.0029), from Hutterer et al. (2019). Photograph by Laura Torrent. (C) Drawing of the tragi of: (e) P. principis sp. nov. (EBD 17358M, paratype); (f) P. brunnea from Van Cakenberghe and Happold (2013) and (g) P. brunnea (DM13229) from Monadjem et al. (2013) and (D) dorsal and ventral views (from left to right) of bacula of: (h) P. principis sp. nov. (EBD 17358M, paratype); (i) P. rendalli from Hill and Harrison (1987); and (j) P. brunnea (DM13229) from Monadjem et al. (2013). Drawings of P. principis sp. nov. baculum by Joaquín López-Rojas.

• Pipistrellus sp.: Juste and Ibáñez 1994:841.

• Pipistrellus (N.) sp.: Juste and Ibáñez 1994:841

• Pseudoromicia sp.: Rainho et al. 2022.

Holotype.

Female (EBD 17475M, field number: B6718), stuffed preserved specimen with skull and skeleton removed), mist-netted near Papagaio River in Santo António, Príncipe Island, at sea level on 26th March 1988 by JJ and CI. External measurements (in mm except for weight in g) are: W: 3.8; FA: 32.2; HF: 6; TIB: 11.9; EAR: 8.4. Skull measurements are: GSL: 12.8; CBL: 11.9; CCL: 11.44; PL: 4.5; GBW: 5.9; ZW: —; POB: 3.4; CM³: 4.1; IM³: 4.8; SPM: 3.4; M³M³: 5.4; C1-C¹: 3.8; MAND: 9; CM₃: 4.8; IM₃: 5.5. Tissue from the holotype has not been sequenced.

Type locality.

Santo António city, Príncipe Island (Fig. 1), República Democrática de São Tomé e Príncipe (1.635°N, 7.419°W), at sea level.

Paratypes.

A subadult male (EBD 17358M) from Santo António, Príncipe Island, preserved in ethanol (70%) with skull and baculum extracted and captured on 20th March 1988 by JJ and CI. Two adult females (EBD 29604M, EBD 29605M) captured at the “Roça Bela Vista” over the Papagaio River, 3 km apart from Santo António on the 3rd November 1992. One pregnant female (EBD 29606M) and one male (EBD 29607M) captured on the 12th November 1992 in a small lagoon near the bank of the Papagaio River in the city of Santo António. These last four specimens were preserved in ethanol (70%) with skull extracted and tissue samples preserved separately. The Cytb sequences of three of them are deposited in GenBank with accession numbers: MW287571, MW287572, and MW287573.

Description.

Pseudoromicia principis sp. nov. is a small pipistrelle-like bat (Fig. 3) with tail fully enclosed in interfemoral membrane; forearm length 30.5–32.3 mm (n = 6); greatest skull length 12.3–12.9 mm (n = 6), and length of the upper tooth row ranging from 4.2 to 4.4 mm (n = 6; Tables 3 and 4). The ears are rounded, with a relatively small tragus (less than 40% length of the ear). The tragus has a short and straight anterior margin, and a posterior margin that has a sharp angle, giving the general appearance of a parallelogram diagonally truncated, but pointed at the anterior corner (Fig. 6e). The posterior edge of the tragus decreases gently at a slope of 260° as it moves toward the exterior of the ear, with a notch at the middle of its margin (Fig. 6e). An antitragus is absent. The skull is small, with a short rostrum and rounded braincase.

The pelage is dense, and the texture is very soft, being 6–7 mm long at its longest. Dorsally, the pelage of P. principis sp. nov. is unicolored and a dense dark chocolate brown. The ventral pelage is bicolored: chocolate brown at the base and cream at the tips. The pelage does not extend onto the wings or interfemoral membrane. The ears and tragus are also chocolate brown, as is the skin around the mouth, whereas the skin around the eyes is paler (Fig. 3). The dorsal pelage and wing membrane are chocolate brown, as are the hairless areas covering the tail, femur, tibia, fibula, ankle, hind foot, forearm, and fingers. The external nose (rhinarium) is broad, and has a shape similar to the rhinarium described for NycticeinopsHutterer et al. (2019), in which “Pars supranarica” and “Pars internarica,” based on Ade (1998) terminology, cover most of the “Pars supralabialis.” However, the fenestrae of the rhinarium of P. principis sp. nov. are larger and more centered than in Nycticeinops (Figs. 6c and 6d).

The cranium of P. principis sp. nov. is small and robust (Fig. 4). The rostrum is low, short, and broad and the braincase is broad and high. The rostrum and braincase are united by a moderate concave curve. The sagittal and lambdoid crests are absent; thus, there is no occipital helmet. The zygomatic arches are wide and delicate. The supraorbital crests are minimally developed and the postorbital process is absent. The auditory bullae are moderate in size. The palate is short and broad. The anterior palatal emargination is deep and wide and forms a U-like shape. The superior incisors displace the “U” curve slightly inward. The dental formula is as follows: I 2/3, C 1/1, P 1/2, M 3/3, total 32. The upper inner incisors are bicuspid and pointed, and the outer incisors show an accessory cusp at the posterior base of the tooth (Fig. 6b). The height of the upper outer pair of incisors is 45–70% of the height of the inner pair. The lower incisors are trilobed and rounded (Fig. 6a). The mean maximum height of the upper canine is 1.4 mm, and the inferior is 1.0 mm (Supplementary Data SD2). The mean maximum width of the upper premolar is 0.6 mm and the lower is 0.9 mm. Upper premolars are aligned in the tooth row and are in contact with the canines and the molars (Fig. 4). Both the maxillary and mandibular molars are well developed. The third molar is massive and myotodont. The distance between the mesostyle to mesocone of the third molar is 55–79% of the distance between the parastyle and the paracone.

The baculum shows the same general morphological pattern as the bacula described for P. rendalli (Hill and Harrison 1987) and P. brunnea and P. roseveari (Monadjem et al. 2013) but differs in having a wide base connected through a long shaft to an expanded and lobed, plate-like structure (Figs. 5 and 6). These similarities support a close phylogenetic relationship between these species and P. principis sp. nov. However, in the new species, the base (0.85 mm) shows two distinct basal lobes. The shaft is very long (1.8 mm), slender, and slightly curved at the tip. It is shelf-like (0.5 mm), and flattens and broadens with two characteristic circular bulges (Figs. 5 and 6).

Echolocation.

Despite intense sampling efforts across Principe Island during recent years, no other vespertilionid bat has been captured or recorded on the island (Juste et al. 1994; Rainho et al. 2022) nor can any echolocation calls could be confused with those of P. principis sp. nov. This fact together with the fact that many calls were recorded from pipistrelle-like bats that were seen flying at dusk make us confident that all analyzed calls correspond to the new species. The echolocation calls of P. principis sp. nov. show the typical pipistrelle-like call structure (Fig. 7). The first part is characterized by a steep downward frequency modulation (FM), followed by a quasi-constant frequency (QCF) component (Kalko and Schnitzler 1993). The relative importance of each component varies depending mainly on the surrounding environment (e.g., Kazial et al. 2001). As a result, pipistrelles typically show wide individual variation in shape and structure between calls. Figure 7 illustrates examples of three calls with different combinations of the FM and QCF elements. The mean ± SD and range of the analyzed sequences of calls (N = 40) are as follows: Start Frequency, 55.47 ± 8.43 (44.75–74.05) kHz; End Frequency, 44.71 ± 1.44 (41.92–47.83) kHz; Maximum Energy Frequency or Peak Frequency, 45.69 ± 1.26 (43.9–48.3) kHz; Duration, 5.82 ± 1.39 (5.32–9.41) ms; Inter-pulse interval, 90.55 ± 8.89 (72–114.7) ms. As expected, the Peak Frequency value found is similar to the value (45 kHz) of the Palearctic Pipistrellus pipistrellus, which is approximately the same size (Van Cakenbergue and Happold 2013), and higher than the values reported for the larger pipistrelles Laephotis kirinyaga (43.9 kHz) and P. nyanza (40.4 kHz). These comparisons must be considered tentative because the last values were obtained from bats flying in a cage instead of from free-flying bats (Monadjem et al. 2021a).

Fig. 7.

Selected echolocation calls of Pseudoromicia principis sp. nov. with three examples (A, B, C) of calls showing decreasing frequency modulation (FM) and increasing quasi-constant frequency (QCF) elements. Calls were directly recorded from free-flying bats at dusk in Santo António city and surroundings.

Distribution.

The new species is endemic to the 128 sq km area of the island of Príncipe, in the Gulf of Guinea (Western Central Africa). This island lies just north of the Equator. As an oceanic island, it has never been in contact with the mainland, from which it is separated by a distance of 220 km (Fig. 1). Pseudoromicia principis sp. nov. is known from several localities, mainly from the northern half of the island (Juste et al. 1994; Rainho et al. 2022), suggesting that it is common across the lowlands of the northern half of Príncipe Island. The original forest of this part of the island was transformed into cocoa plantations and orchards, with few houses and other human constructions, none of which appear to negatively affect the population of P. principis sp. nov. On the other hand, it is possible that the current distribution pattern on the island is the result of biased sampling, since the forested southern half is much less accessible. Thus, the species could be abundant in this part of the island as well (Rainho et al. 2022).

Etymology.

The epithet “principis” is derived from the Latin princeps and refers to Príncipe Island (Gulf of Guinea, Western Central Africa) from where the bat is native and endemic. As a vernacular name we propose “Príncipe’s Pipistrelle.”

Diagnosis.

A small pipistrelle-like bat (forearm length 30.5–32.3 mm) with a long baculum with unique flattened tip with three circular bulges. Dark brown pelage with uncolored dorsal hair places the species in the genus Pseudoromicia. Dark chocolate brown pigmentation of hairless parts of the body and membranes (including wings) distinguishes this species from all the members of the translucent white-winged group of species of Pseudoromicia. This is the smallest bat so far described within the genus. It has a tiny moderately inflated skull (12.3–12.9 mm) with dimensions that do not overlap those of any other species of Pseudoromicia (Tables 3 and 4).

Comparisons.

The skull (size and shape), dentition, baculum (shape and length), pelage coloration, and forearm length distinguish P. principis sp. nov. from all other African pipistrelle-like bats. Its moderately inflated skull in lateral profile differs from the flattened skull of Laephotis and the highly inflated skull in Afronycteris (Van Cakenbergue and Happold 2013). It is also distinguished from Afronycteris by the absence of P¹. The dorsal unicolor fur contrasts with the bicolor (or even tricolor) pattern of the fur characteristics of the genus Neoromicia (Monadjem et al. 2021a). The wide and centered openings (fenestrae) of the external nose distinguish the new species from Nycticeinops (Fig. 7), as well as the absence of a secondary cuspid in its upper outer incisor (I²) and the different shape of the baculum. In fact, the shape of the baculum in the new species—with its curved, long, and slender shaft—indicates that it belongs to the genus Pseudoromicia, as confirmed by molecular evidence. The wing membranes and pelage are both chocolate brown, indicating it is part of the dark-winged group of Pseudoromicia, in contrast to the translucent white-winged species, including P. tenuipinnis, which, though similar in size, has a much smaller baculum with a distinct flattened tip (Monadjem et al. 2021a). There are three known dark-winged species within Pseudoromicia. Two of them, P. roseveari and P. kityoi, can be differentiated from P. principis sp. nov. by the smaller size of the new species (forearm more than 5 mm smaller in P. principis sp. nov.; Tables 3 and 4). Pseudoromicia principis sp. nov. more closely resembles the remaining dark-winged species, P. brunnea, since both are relatively similar in size, and show chocolate brown ventral and dorsal pelage. However, the tips of the ventral pelage have a distinct light creamy tinge in the new species. Additionally, P. brunnea is consistently larger in all body, cranial, and dental measurements (Tables 3 and 4). The main differences are to be found in the dentition: P. brunnea shows unicuspid upper incisors with the outer incisor very small (Fahr 2013), whereas the incisors in P. principis sp. nov. are bicuspid with the outer one relatively large with its tip reaching over the height of the cingulum of the inner (Fig. 7). The bacula are also different, as the shaft is shorter and more curved in P. brunnea (Fig. 6). Finally, the molecular analyses show that the two species differ by 9.9% in their Cytb sequences and that they are not considered to be sister clades (Fig. 2).

Discussion

The description of this new species was only possible after the systematics of the entire group of vespertilionid bats—traditionally entangled—has been recently settled by including a variety of characters. The first step towards an ordered and stable classification began with the finding by Hill and Harrison (1987) that the variation in the shape and size of the bacula was meaningful and extremely useful as a tool for establishing taxonomic units at the species and supra-specific levels within the group. The increasing cytogenetic information (e.g., Volleth 1987; Volleth and Heller 1994; Heller et al. 1994; Volleth et al. 2001; Kearney et al. 2002) complemented and supported the new arrangements. Later, the application of molecular techniques to the entire group (Hoofer et al. 2003, 2006) revealed paraphyly of Neoromicia (sensu lato) and suggested that the genus be split. Later, the existence of four clearly supported groups within Neoromicia was confirmed (Koubínová et al. 2013). Finally, more recent studies focusing primarily on African taxa (Monadjem et al. 2013, 2021a, 2021b) led to the recognition of four genera (two already named) within what once was defined as Neoromicia. Pseudoromicia principis sp. nov. clearly fit in one of these new-named genera (Pseudoromicia). Similarly, other new pipistrelle-like bats have been recently described and taxonomically assigned. These include P. roseveari (Monadjem et al. 2013), P. isabella (Decher et al. 2015), Pipistrellus dhofarensis (Benda et al. 2016), Nycticeinops happoldorum (Hutterer et al. 2019), Pipistrellus simandouensis (Monadjem et al. 2021b); and, in Madagascar, Pipistrellus raceyi (Bates et al. 2006), and Laephotis robertsi (Goodman et al. 2012). The fact that the new systematics has incorporated all these new taxa without requiring further adjustments suggests that the African vespertilionid bats are currently organized in a robust and lasting taxonomic arrangement. On the other hand, the arrangement so far is based only on mitochondrial markers. The definitive test will be to see in the future how it will be supported by nuclear markers such as the set of introns recently developed for mammals (Igea et al. 2010). These nuclear markers have been tested successfully to establish phylogenies for other bats (e.g., Salicini et al. 2013; Demos et al. 2019). Finally, we know virtually nothing about the ecological requirements of the new species. Given the fragility and high risks of island endemic forms and of bats in particular (Conenna et al. 2017), field studies are urgently needed to establish the habitat requirements and conservation status of this new species that, at present, should be listed as “Data Deficient” following the IUCN categories.

Appendix I

Specimens of pipistrelle bats of the genera Afronycteris, Hypsugo, Laephotis, Nycticeinops, and Pseudoromicia used in the morphologic comparison. They are housed in the EBD-CSIC collections.

Species	Country	Voucher
Laephotis capensis	Equatorial Guinea	EBD 19107M
Afronycteris nana	Equatorial Guinea	EBD 15714M
Afronycteris nana	Equatorial Guinea	EBD 13938M
Laephotis capensis	Equatorial Guinea	EBD 16827M
Hypsugo (cf.) musciculus	Equatorial Guinea	EBD 15554M
Nycticeinops crassulus	Equatorial Guinea	EBD 17743M
Pseudoromicia brunnea	Equatorial Guinea	EBD 15706M
Pseudoromicia brunnea	Equatorial Guinea	EBD 20520M
Pseudoromicia tenuipinnis	Equatorial Guinea	EBD 14163M
Pseudoromicia tenuipinnis	Equatorial Guinea	EBD 15308M

Version of Record, first published online January 11, 2023, with fixed content and layout in compliance with Art. 8.1.3.2 ICZN.

Nomenclatural statement: A Life Science Identifier (LSID) number was obtained for this publication: urn:lsid:zoobank.org:pub: 252DBECE-70A7-4FC6-BC1D-053569170FD5

Supplementary Data

Supplementary data are available at Journal of Mammalogy online.

Supplementary Data SD1.—Measurements (mm, g) used. Acronyms: forearm length (FA); body weight (W); ear length (EAR); tragus length (TRA); tibia length (TIB); hind food, including claws (HF); tail length (TL); whole body length (HBL); third finger metacarp (3FM); third finger phalange (3FPH1); forth finger metacarp (4FM); forth finger phalange (4FPH1); fifth finger metacarp (5FM); fifth finger phalange (5FPH1); greatest length of skull, excluding incisors (GLS); condylo-basal length (CBL); condylo-canine length (CCL); palatal length (PL); greatest braincase width (GBW); zygomatic breadth (ZB); postorbital width (POB); rostrum width (RW); maxillary toothrow length (CM³); distance from the 3rd upper molar to the upper incisor (IM³); mandibular toothrow length (CM₃), lower dental raw (IM₃); upper series (molars + premolar) (SPM); width across upper canines (C¹C¹); width across upper molars (M³M³); greatest mandible length excluding incisors (MAND).

Supplementary Data SD2.—Craniodental measurements (mm) of the known specimens of the new species Pseudoromicia principis sp. nov.