Kerivoula bicolor, Thomas, 1904

Sedlock, Jodi L., Heaney, Lawrence R., Balete, Danilo S. & Ruedi, Manuel, 2020, Philippine bats of the genus Kerivoula (Chiroptera: Vespertilionidae): Overview and assessment of variation in K. pellucida and K. whiteheadi, Zootaxa 4755 (3), pp. 454-490 : 464-467

The status of K. bicolor and K. pusilla

Two taxa originally named as distinct species, K. bicolor Thomas 1904 and K. pusilla Thomas 1894 , were considered by Hill (1965) to represent subspecies of K. whiteheadi , and have been provisionally dealt with in that fashion subsequently (e.g., Corbet & Hill 1992; Francis 2008). At the time, Hill (1965) had only the holotype of K. whiteheadi from Luzon, a badly damaged specimen, for comparison with K. bicolor , also represented only by the holotype, and with K. pusilla , represented by two specimens. He also noted two specimens referred to as K. pusilla by Sanborn (1952), which he did not examine. We know of no specimens of K. bicolor collected since the holotype, and only a few K. pusilla have been reported from Borneo ( Phillipps & Phillipps 2016), but we have obtained large numbers of K. whiteheadi from the Philippines that allow some morphological assessment of the distinctiveness of these three taxa.

As shown in Figure 3 , a plot of condylo-canine length vs. braincase height of specimens of K. whiteheadi from the Philippines form a cloud of points that represents the range of variation in the nominate form. The two specimens referred to K. pusilla by Sanborn (FMNH 56687 and 61077) have the shortest condylo-canine length of our Philippine sample, but they fall within the range of other specimens from Mindanao and within the cloud of Philippines points overall. Our measurements of the holotype of K. pusilla from Borneo show it to be similar in condylocanine length, but with a notably lower braincase height, well below that of any Philippine specimen ( Fig. 3 ; Tables 1 , 3 ). Additionally, our examination of the cranium of the holotype showed it to have unusually large cochlea, as described by Thomas (1894) and confirmed by Hill (1965). While the cochlea of K whiteheadi is large ( Fig. 6 ), the K. pusilla holotype’s cochlea (from a photograph we took) is proportionately substantially larger. Given the difference in braincase height and cochlea size in comparison to our large sample, we recommend that K. pusilla not be treated as a subspecies of K. whiteheadi .

We also examined the holotype (and only known specimen) of K. bicolor . The cranium is badly damaged, and we were unable to measure the condylo-canine length. Further, the cochlea were not present for comparison. However, the braincase height (4.74 mm) is well below the range shown by K. whiteheadi ( Fig. 3 ; Table 1 ). Given the comment by Hill (1965:531) that “both [ K. bicolor and K. pusilla ] have greatly expanded cochlea”, and that K. w. bicolor “closely resembles K. w. pusilla , differing … only in its white-tipped wings, white and not grayish underparts, slightly narrower rostrum and slightly less pronounced rostral sulcus … [which] seem only of subspecific importance”, we provisionally conclude that these should not be considered conspecific with K. whiteheadi , and can be treated as subspecies of a single species, K. p. pusilla and K. p. bicolor , until further specimens are available for more detailed study.

Biogeography and diversification

Our time-calibrated phylogenetic tree of Cyt b sequences obtained with BEAST has implications for the timing and geography of diversification of Kerivoula in the Philippines and adjacent areas ( Fig. 4 ). As noted above, support for many nodes is strong, but confidence limits (i.e. 95% HPD intervals) on the timing of most nodes are broad, so that caution is warranted.All mention of dates that follow are given as the average; these should be viewed as rough estimates, plus or minus about 30%, with the confidence limits shown in Fig. 4 .

The genus Kerivoula is here estimated to have diverged from Phoniscus , its sister-genus, about 27 million years ago (MYA; Fig. 4 ). The initial split between the K. pellucida species group (A–F) and all other species took place about 25 MYA, and the split between the K. papillosa species group (K–M) and the K. hardwickii species group (G–J) about 21 MYA. Thus, initial diversity within Kerivoula in Southeast Asia appears to have been well established by more than about 20 MYA.

Diversification within the species groups has been more recent. K. pellucida (A–B) and K. whiteheadi (C–F) last shared a common ancestor about 18 MYA, but diversification within these species (as currently defined) began about 3.7 MYA and 7.1 MYA, respectively. Members of the K. papillosa group (K–M) last shared a common ancestor about 15.3 MYA, but most species have diverged within the last 10 MYA ( Fig. 4 ). Members of the K. hardwickii species group (G–J) last shared a common ancestor about 17.2 MYA, with subsequent diversification.

We note that most diversification within the Philippines is relatively recent. Kerivoula pellucida in the Philippines diverged from their Sunda Shelf conspecifics about 3.7 MYA, but divergence within the Philippines (clade A) appears to have begun only about 2.0 MYA. Populations referred to K. whiteheadi last shared a common ancestor about 7.1 MYA, and have formed geographically cohesive clades (C, D, E, and F) subsequently, with specimens from northern Luzon, Camiguin Norte, Lubang, and Mindoro (C) sharing a common ancestor about 1.4 MYA. The oldest clade that includes K. papillosa in the Philippines diverged about 7.0 MYA, but this includes a specimen from Sarawak, which makes the interpretation of diversification within the Philippines ambiguous. Finally, Philippine populations of K. hardwickii (G’) diverged from their Sunda Shelf relatives (G”) by about 9.6 MYA, the oldest of the implied colonizations of the Philippines by Kerivoula .

These data imply that Kerivoula arrived in the Philippines by about 10 MYA. Kerivoula whiteheadi occurs throughout the entire Philippines ( Esselstyn et al. 2004; Heaney et al. 2010; Fig 2 ). However, K. pellucida occurs only on Greater Mindanao, Greater Negros-Panay, and Greater Palawan; K. hardwickii only occurs in Greater Mindanao and Greater Palawan; K. papillosa only occurs in Greater Luzon and Greater Mindanao ( Fig. 2 ). Thus, only one of the four species is widespread within the Philippine archipelago, and it ( K. whiteheadi ) exhibits only a moderate level of incipient speciation that began about 7 MYA.

However, our genetic data ( Figs. 1 , 4 ) indicate the presence of two distinct clades within the Philippine K. hardwickii lineage that are morphologically distinct, and our morphological data ( Fig. 3 ; Table 3 ) show that both occur on Mindanao ( Appendix A ). We take this as tentative evidence of speciation in Kerivoula within the archipelago that has progressed to the point of secondary sympatry. This requires further study, but suggests that although in places where more than one species of Kerivoula is present, each is usually the product of a separate colonization from other regions in Southeast Asia. Where divergence dates among sympatric taxa are deeper, in the case of the K. hardwickii group, some full diversification within the archipelago may have taken place, in this oldest of the Philippine lineages.

Finally, we note that K. whiteheadi from southern Luzon (clade D) are more closely related to specimens from the central Philippines (clade E) than they are to those from central and northern Luzon and adjacent islands (clade C; Fig. 4 ). Although it is initially surprising that all populations from Luzon are not monophyletic, the geological history of the Philippines is consistent with this. The estimated divergence date of 5.8 MYA for clades D and E from clade C reflects the fact that southern Luzon (the area represented by clade D) was a distinct island that first arose about 6.6 MYA. However, this island did not become connected to central and northern Luzon until roughly 0.5 MYA when the Bondoc Peninsula and adjacent areas were uplifted ( Aurelio et al. 1991; Heaney et al. 2016:49-50; Fig. 2 ). Thus, the sister-group relationship of clades D and E appears to reflect the prior history of southern Luzon and Greater Negros-Panay (including Cebu) as two separate but nearby islands at the time of this phylogenetic split, 5.8 MYA.

Figures

FIGURE 1. Phylogenetic relationships of East Asian Kerivoula and two outgroup species based on Cyt b haplotypes. This topology results from a Bayesian analysis. Numbers at each node correspond to the Bayesian posterior probabilities (left) and to the proportion of bootstrap support of a maximum likelihood reconstruction (right). If both analyses supported a node by 98% or more support, this node is symbolized with a plain circle. The letters to the right of each clade refer to the groups discussed in the text. The scale bar corresponds to a branch length of 0.05 substitution per site.

FIGURE 2. Map of the Philippines indicating the collecting localities of Kerivoula pellucida and K. whiteheadi. The land exposed during the Pleistocene Last Glacial Maximum is shown in light green.

FIGURE 3. Bivariate plot of condylo-canine length vs. height of braincase for specimens of Kerivoula summarized in Tables 1 and 3. Color indicates geographic region and symbol indicates taxon. Philippine specimens of each taxon are enclosed within polygons.

FIGURE 4. Time-calibrated phylogeny of Kerivoula species based on Cyt b haplotypes. Outgroups were omitted for ease of readability. The greyed boxes represent 95% highest posterior density (HPD) intervals of node ages derived from a BEAST analysis. Time is expressed in million years.

FIGURE 6. Scanning electron microscope images of K. pellucida cranium and mandible, FMNH 202782 from Bohol Island, Philippines.

APPENDIX A. List of all Kerivoula specimens examined. Specimens included in morphological analysis and comparisons are in bold. § Holotypes * Paratype.