Content uploaded by Heru Handika
Author content
All content in this area was uploaded by Heru Handika on Apr 26, 2024
Content may be subject to copyright.
Journal of Mammalogy, 2024, XX, 1–18
https://doi.org/10.1093/jmammal/gyad126
Advance access publication 25 January 2024
Research Article
© The Author(s) 2024. Published by Oxford University Press on behalf of the American Society of Mammalogists, www.mammalogy.org.
Submitted 31 August 2022; Accepted 11 December 2023
Research Article
Three new shrews (Soricidae: Crocidura) from West
Sumatra, Indonesia: elevational and morphological
divergence in syntopic sister taxa
Jonathan A. Nations1,5,*,, Heru Handika1, Ahmad Mursyid2, Ryski Darma Busta2, Apandi3, Anang S. Achmadi4, Jacob A. Esselstyn1
1Museum of Natural Science and Department of Biological Sciences, Louisiana State University, Baton Rouge, LA 70803, United States
2Department of Biology, Faculty of Mathematics and Natural Sciences, Andalas University, Padang, Sumatera Barat 25163, Indonesia
3Directorate of Scientic Collection Management, Deputy of Infrastructure Research and Innovation, National Research and Innovation Agency (BRIN), Cibinong,
Jawa Barat 16911, Indonesia
4Museum Zoologicum Bogoriense, Research Center for Ecology and Ethnobiology, National Research and Innovation Agency (BRIN), Cibinong, Jawa Barat 16911,
Indonesia
5Present addresses: Department of the Geophysical Sciences, University of Chicago, Chicago, IL 60637, United States Field Museum of Natural History, Chicago, IL
60505, United States, Florida Museum of Natural History, University of Florida, Gainesville, FL 32611, United States
*Corresponding author: Florida Museum of Natural History, University of Florida, Gainesville, FL 32611, United States. Email: jnations1@u.edu
Version of Record, rst published online January 25, 2024, with xed content and layout in compliance with Art. 8.1.3.2 ICZN.
Nomenclatural statement.—A Life Science Identier (LSID) number was obtained for this publication: urn:lsid:zoobank.org:pub: 451A0A0F-2AF1-4E1B-98F3-
D120049C23DD
Associate Editor was Melissa Hawkins
Abstract
We describe 3 new species of shrews (Eulipotyphla, Soricidae, Crocidura) from West Sumatra, Indonesia. Two of these taxa were found
above 1,800 m on Mt. Singgalang. The third taxon was found above 1,660 m on Mt. Talamau, 65 km northwest of Mt. Singgalang. We also
resurrect Crocidura aequicauda based on 2 specimens from Mts. Tujuh and Kerinci, which lie near the border between West Sumatra
and Jambi provinces. Several methodological approaches support our ndings: linear cranial morphometrics, landmark-based 2D
geometric morphometrics, and molecular phylogenetics using both mtDNA and 6 nuclear exons. A multilocus species-tree analysis
places the 3 new species and C. aequicauda in a clade with the Javan endemics C. monticola and C. umbra. Although the 2 taxa from
Mt. Singgalang are recovered as sister species, 1 is nearly twice the size of the other, and they are divergent in several other morpho-
logical characters, such as tail length, cranium size, and pelage color and texture. Recently diverged yet morphologically disparate
sister taxa living syntopically in an isolated habitat “island,” like the montane forests of Mt. Singgalang, is unusual in mammals but
documented in other Crocidura on neighboring Java and Borneo; these 2 new taxa represent the rst known case of this phenomenon
on Sumatra. Our results bring the number of Sumatran Crocidura to 10, 9 of which are endemic to the island. All 3 of the new species
appear to be endemic to a single mountain and were not detected in similar surveys of nearby mountains. If this local endemism
pattern is common, it would indicate that Sumatra’s mammal diversity may be severely underestimated, largely due to the paucity
of small-mammal surveys and museum specimens.
Key words: biodiversity, Crocidura, Indonesia, new species, Sunda.
Tiga cecurut semut baru (Soricidae: Crocidura) dari Sumatra Barat, Indonesia: perbedaan elevasi dan morfologi
dalam sintopik, taksa saudara
Abstrak
Kami mendeskripsikan tiga species baru cecurut (Eulipotyphla, Soricidae, Crocidura) dari Sumatera Barat, Indonesia. Dua dari tiga
taksa ini ditemukan pada ketinggian di atas 1,800 meter di Gunung Singgalang. Takson ketiga ditemukan di atas 1,660 m di Gunung
Talamau, 65 km di sebelah barat laut Gunung Singgalang. Kami juga membangkitkan kembali Crocidura aequicauda berdasarkan dua
specimen dari Gunung Tujuh dan Gunung Kerinci, yang terletak didekat perbatasan antara provinsi Sumatera Barat dan provinsi
Jambi. Beberapa pendekatan metodologis mendukung temuan kami: morfometrik linear tengkorak, morfometrik geometric 2D ber-
basis landmark, logenetika molekuler menggunakan mtDNA dan enam ekson inti. Analisa multilokus kekerabatan antar spesies
menempatkan tiga species baru ini dan C. aequicauda dalam satu klad dengan species endemik Jawa, C. monticola dan C. umbra.
Walaupun dua taksa dari Gunung Singgalang ditemukan sebagai spesies saudara, salah satunya memiliki ukuran dua kali lebih
besar dari yang lainnya, dan mereka berbeda dalam beberapa karakter morfologi lainnya, seperti panjang ekor, ukuran tengkorak,
dan warna dan tekstur rambut. Taksa bersaudara yang baru saja bercabang tetapi memiliki morfologi berbeda yang hidup terisolasi
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
2 | Nations et al.
di habitat “pulau,” seperti hutan pegunungan di Gunung Singgalang, merupakan hal yang tidak biasa pada mamalia, tetapi juga ter-
catat pada Crocidura lainnya di Jawa dan Kalimantan yang berdekatan; dua species baru ini merepresentasikan kasus pertama yang
diketahui tentang fenomena ini di Sumatera. Hasil kami menjadikan jumlah Crocidura Sumatera menjadi sepuluh species, sembilan
spesies endemik di pulau tersebut. Ketiga spesies baru tersebut tampaknya endemik di satu gunung saja dan tidak terdeteksi di
survei serupa di gunung-gunung terdekat. Jika pola endemisme lokal ini umum, ini menunjukkan bahwa keanekaragaman mama-
lia Sumatera mungkin masih jauh dari estimasi sebenarnya, sebagian besar karena kurangnya survei mamalia kecil dan specimen
museum dari kawasan tersebut.
Kata kunci: biodiversitas, Crocidura, Indonesia, species baru, Sunda.
The ongoing effort to document global biodiversity has revealed
that the number of living mammal species has been greatly
underestimated, particularly in diverse tropical regions (Reeder
et al. 2007; Heaney et al. 2016; Esselstyn et al. 2021; Parsons et al.
2022). Intense anthropogenic pressures exerted across the global
tropics today add urgency to the need to accurately document
biodiversity. Perhaps no group of mammals better exemplies
historic undercounting than the white-toothed shrew genus
Crocidura (Soricidae), a group of 219 currently described (Mammal
Diversity Database 2023), small-bodied (<50 g), nocturnal mam-
mals with a largely conserved morphology. Delimiting species
within Crocidura using traditional morphometrics has been dif-
cult, as many species look alike, and the paucity of specimens
has compounded this predicament. However, recent efforts are
revealing the existence of many cryptic species, and 44 new taxa
have been described since the end of the 20th century alone
(Mammal Diversity Database 2023). Nearly all these discover-
ies have resulted from combined morphological and molecular
approaches. The realization that many evolutionarily distinct
populations of shrews—often distantly related—may co-occur
in areas previously thought to support only 1 or 2 species has
sparked interests in elevational divergence, convergent evolution,
character displacement, and community ecospace occupation in
Crocidura (Esselstyn et al. 2013; Demos et al. 2016a, 2016b; Van de
Perre et al. 2020).
Recently improved resolution in vertebrate taxonomy more gen-
erally has also enhanced our understanding of the biogeography in
Sundaland, Wallacea, and the Philippines. Where island vicariance
was once thought to be a key driver of diversication, recent anal-
yses of animal diversity in these areas suggest a complex history
of colonization, and a prominent role of in situ diversication of
species (Oliveros and Moyle 2010; Esselstyn et al. 2013; Sheldon et
al. 2015; Demos et al. 2016a; Rowe et al. 2019; Shaney et al. 2020a,
2020b). Examples in shrews include elevational and morphological
divergences of sister species pairs (e.g. Crocidura monticola and C.
umbra; Demos et al. 2016b) and the endemic radiation of 20 Crocidura
species on Sulawesi, complete with multiple examples of morpho-
logical convergence (Esselstyn et al. 2021). Current knowledge con-
tends that nearly all species of Sunda, Wallacean, and Philippine
shrews are endemic to a single island.
The island of Sumatra is a biodiversity hotspot, famous for
housing some of the remaining mammalian megafauna of the
region (e.g. tigers, rhinoceros, orangutans). The Barisan Mountains
run nearly the entire length of the island, forming a long chain
of wet, tropical lowland forests at their base that quickly rises
to montane cloud forests—a habitat gradient that often houses
high biodiversity. The shrew fauna of Sumatra has been sparsely
sampled, and all but one of the currently recognized species
of Sumatran Crocidura were described from single specimens
(Ruedi 1995). The most recent taxonomic revision of Sumatran
Crocidura determined there to be 5 species present, along with C.
vosmaeri from nearby Bangka Island (Ruedi 1995), though recent
publications suggest the presence of additional taxa. A phyloge-
ographic study of Sundaland shrews points to the presence of an
undescribed Sumatran species that is sister to C. monticola and
C. umbra from Java (“C. sp. nov. 2”; Demos et al. 2016a). Another
Sumatran specimen “previously misidentied as C. paradoxura”
(Hinckley et al. 2022:493) was recovered as sister to the unnamed
taxon “C. sp. nov. 2” (Demos et al. 2016a; Hinckley et al. 2022). This
discovery has additional taxonomic relevance as the morphology
of the misidentied C. paradoxura was the basis for synonymizing
C. paradoxura and C. aequicauda (Robinson and Kloss 1918; Ruedi
1995).
In 2011 and 2018, we undertook small-mammal inventories
that specically targeted shrews on 3 mountains in West Sumatra.
Here we document 3 new species of Crocidura from West Sumatra,
Indonesia, and resurrect a fourth species, using cranial morpho-
metrics, external measurements, and multilocus sequence data
drawn from these new specimens.
Materials and methods
Fieldwork
We surveyed small mammals in 3 locations in West Sumatra,
Indonesia: Mount Tujuh (2011), Mount Singgalang (2011 and
2018), and Mount Talamau (2018; Fig. 1). The 2011 trip to Mt.
Singgalang surveyed the eastern side of the mountain, while
the 2018 trip focused on the northern side. All surveys used
a mix of Victor snap traps and pitfall lines, and virtually all
shrews were captured in pitfalls. Pitfall lines consisted of 6 to
10 large 20- to 30-l buckets connected by a drift fence made
of tarp, though occasionally we used smaller 1- to 3-l buckets
with and without a drift fence. On Mount Tujuh, we set pitfall
lines at 1,540 m (mix of large and small buckets), 1,600 m (mix
of large and small buckets), and 1,950 m (large buckets) for a
total of ca. 300 pitfall nights. On Mt. Singgalang in 2011, we set
pitfall lines (large buckets) at 1,700 and 2,000 m for a total ca.
120 pitfall nights. In 2018, we added large-bucket pitfall lines at
1,800, 1,880, and 2,100 m for a total of ca. 330 pitfall nights, and
a line of 10 small buckets for 1 night at 2,826 m. On Mt. Talamau
we set pitfall lines (large buckets) at 1,050, 1,600, 1,660, 1,800,
and 2,000 m for a total of ca. 515 pitfall nights. All specimens
were deposited at the Field Museum of Natural History, Chicago
(FMNH); Louisiana State University Museum of Natural Science,
Baton Rouge (LSUMZ); and Museum Zoologicum Bogoriense
(MZB), Bogor. Specimens were measured and weighed, and
after tissues harvested, bodies were preserved in formalin,
often with the skull removed, or prepared as dried museum
skins with dried and cleaned skeletons. Tissues were pre-
served in liquid nitrogen or 95% ethanol. All collecting meth-
ods followed the recommended guidelines of the American
Society of Mammalogists (Sikes et al. 2016). We incorporated
additional comparative material from University of California
Museum of Vertebrate Zoology, Berkeley (MVZ); Rijksmuseum
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 3
Natuulijke Historie, Leiden (RMNH); and National Museum of
Natural History, Smithsonian Institution, Washington, District
of Columbia (USNM).
Morphological analyses
External measurements including total length (TL), head–body
length (HB), tail length (Tail), hind-foot length (HF), and ear
length (Ear) were taken from eld notes. Cranial measurements
were taken with digital calipers to the nearest 0.01 mm follow-
ing Esselstyn et al. (2021): condyloincisive length (CIL), breadth
of braincase (BBC), interorbital width (IOW), rostral length (RL),
rostral width (RW), postpalatal depth (PPD), postpalatal length
(PPL), distance from occipital condyle to glenoid fossa (CGF),
upper toothrow length (UTL), distance from alveolar of P4 to M3
(P4–M3), width of the palate from M3 to M3 (PW), and labial width at
M2 (M2–M2). The skull from the type of C. aequicauda (cataloged as
both Federated Malay States Museum FMSM 448/14 and Natural
History Museum BMNH 1919.11.5.28), a junior synonym of C.
paradoxura, is badly damaged (Robinson and Kloss 1918), and we
used the 8 measurements from C. paradoxura MZB 16790 reported
in Ruedi (1995), as the author concluded that the specimens MZB
16790 and FMSM 448/14 are conspecic (Ruedi 1995). We per-
formed a principal components analysis (PCA) on the correlation
matrix of natural log-transformed cranial measurements and
visualized the results in R 4.2.0 (R Core Team 2022), and tidyverse
1.3.2 (Wickham et al. 2019). We made a variety of univariate and
bivariate plots of external and cranial measurements to visualize
potentially diagnostic features. Measured specimens are listed in
Supplementary Data SD1.
Linear cranial measurements are useful in differentiating
Crocidura, but often overlap among species. Size-corrected cra-
nial measurements identify proportional differences in cranial
dimensions (e.g. a wide braincase for a given skull length) that are
not easily identied with univariate linear measurements alone.
However, the correlation between cranial measurements and
skull size may not be consistent among cranial dimensions, and
therefore ratios calculated by dividing cranial measurements by
skull length may not accurately represent a size-corrected meas-
urement. To exploit the strengths of each approach, we quanti-
ed differences in skull proportions between Sumatran Crocidura
using multiple regression to estimate the mean distribution of
11 cranial measurements, conditioned on the length of the skull
(condyloincisive length), or a size-corrected estimate for each cra-
nial dimension. We then constructed 12 linear regression mod-
els using each cranial measurement as a response variable, and
species as a categorical, group-level predictor variable. Our esti-
mate of CIL simply used CIL as the response and species as the
predictor. The remaining 11 models (1 per measurement) used
the linear skull measurement as the response, and species + CIL
as predictors, generating a measurement estimate for each spe-
cies, conditioned on the effect of skull length (i.e. a size-corrected
estimate). We used Bayesian linear regression modeling in the R
package brms 2.18.8 (Bürkner 2018). Before running the models,
we mean-centered and scaled each measurement to a standard
deviation of 1.0. We used regularizing priors, and ran 4 chains for
each model, with each chain set to 2,000 iterations of warm-up
and 2,000 iterations of sampling. All data and scripts for these
analyses are archived on Zenodo as DOI: 10.5281/zenodo.7916697.
We quantied cranial shape variation in Sumatran Crocidura
using a 2D geometric morphometric landmarking approach. We
photographed the dorsal and ventral sides of the crania of 113
Sumatran shrews housed in the LSUMZ collection using a Fujilm
X-Pro2 Camera with a Fujilm 90 mm F2 lens and a 16 mm macro
extension tube. We placed 16 dorsal landmarks and 23 ventral land-
marks on each skull (Supplementary Data SD2) using the digitize2D()
function in the R package geomorph 4.0.3 (Baken et al. 2021). In geo-
morph, we performed a generalized Procrustes analysis (GPA) super-
imposition using gpagen(), and shape coordinates were subjected to
a PCA with gm.prcomp(). PC scores are reported in Supplementary
Data SD1. We visualized the shape morphospaces using ggplot2.
Genetic analyses
We extracted genomic DNA from tissue samples (liver, muscle)
using Qiagen DNEasy Blood & Tissue kits (Qiagen, Germantown,
Maryland) following the manufacturer’s instructions. We
amplied mitochondrial DNA (mtDNA) and nuclear DNA
(nDNA) regions using polymerase chain reaction. We amplied
the mitochondrial protein-coding gene cytochrome b (Cytb) for
33 specimens of new and previously described Crocidura spe-
cies collected on recent expeditions to Sumatra. Additionally,
we sequenced portions of 6 nuclear exons: apolipoprotein B
(Apob), breast cancer susceptibility 1 (Brca), brain-derived neu-
rotrophic factor (BDNF), growth hormone receptor exon 10
(Ghr), prostaglandin E4 receptor (Ptger4), and von Willebrand
factor exon 28 (vWf) from multiple individuals of the putative
species collected on Mount Singgalang and Mount Talamau.
We aligned the new mitochondrial and nuclear sequences to
previously published sequences, which included specimens
from the 2011 surveys of Mts. Tujuh and Singgalang (Demos
Fig. 1. Map of West Sumatra province showing the sampling locations,
Mt. Tujuh, Mt. Talamau, and Mt. Singgalang, from the 2011 and 2018
surveys, as well as Mt. Kerinci where the Crocidura aequicauda type
was collected. Inset shows the area in the context of SE Asia. Map was
generated in the R package tmap (Tennekes 2018).
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
4 | Nations et al.
et al. 2016a; Hinckley et al. 2022), using MEGA v.11 (Tamura et
al. 2021). Specimen catalog numbers and GenBank accessions
are available in Supplementary Data SD3, and the Cytb align-
ment is available in Supplementary Data SD4 and archived on
Zenodo as DOI: 10.5281/zenodo.7916697.
We estimated the Cytb mitochondrial gene tree using a suite
of tools available in IQTree 1.6.12 (Nguyen et al. 2015). We esti-
mated the optimal substitution model for the mitochondrial
data using ModelFinder (Kalyaanamoorthy et al. 2017), which
is integrated into the tree topology estimation in IQTree. We
used the UFBoot2 method (Hoang et al. 2018) to estimate top-
ological support. We calculated inter- and intraspecic Jukes–
Cantor mitochondrial distances in the R package Ape 5.6-2
(Paradis et al. 2004) using a Cytb alignment of all Sumatran and
Javan Crocidura species.
We concatenated alignments of our 6 nuclear exons and esti-
mated a phylogenetic tree with IQTree 1.6.12 (Nguyen et al. 2015).
We estimated the optimal substitution model for each gene using
ModelFinder and topological support with UFBoot2. Specimen
catalog numbers and GenBank accessions are available in
Supplementary Data SD3, and the concatenated nDNA alignment
is available in Supplementary Data SD5 and archived on Zenodo as
DOI: 10.5281/zenodo.7916697. We also estimated a species tree using
StarBEAST in BEAST 2.6.7 (Bouckaert et al. 2019) with the 6 individ-
ual nDNA alignments. We unlinked the clock, substitution, and tree
models for all genes. We used a relaxed log-normal clock model
with a uniform prior, and constant population size. We ran 4 chains
of 108 iterations and stored every 2,000th sample. Convergence
was assessed in Tracer 1.7.1 (Rambaut et al. 2018), using estimated
sample size (ESS) > 200 as an indication of chain convergence.
We removed the rst 20% of the trees as burn-in, combined the
gene trees and species trees from each chain using LogCombiner
(Bouckaert et al. 2019), then estimated maximum clade credibil-
ity gene trees and species tree using TreeAnnotator (Bouckaert et
al. 2019). Specimen catalog numbers and GenBank accessions are
available in Supplementary Data SD3, and the 6 nDNA alignments
are archived on Zenodo as DOI: 10.5281/zenodo.7916697.
Results
Fieldwork
We collected 8 species of Crocidura in total from the 3 Sumatran
mountains. On Mt. Tujuh, we collected C. beccarii, C. lepidura, C.
neglecta, and C. paradoxura. On Mt. Singgalang, we collected C. bec-
carii, C. hutanis, C. neglecta, and C. paradoxura along with 2 unde-
scribed taxa. The rst new taxon was documented in 2011 on Mt.
Singgalang at 2,000 m elevation (Demos et al. 2016a, 2016b). On
the second trip to Mt. Singgalang, the same taxon was collected
on the north side of the mountain at 2,100 m, and an additional
new taxon was collected at 1,880, 2,100, and 2,826 m. On Mt.
Talamau, we collected C. hutanis, C. lepidura, C. neglecta, and C. par-
adoxura as well as an additional species that was not collected on
Mt. Tujuh or Mt. Singgalang.
Morphological evidence
Morphological results support the presence of 3 new species
from Sumatra (Fig. 2). PC1 from the linear measurement analyses
demonstrates that the 2 smaller new species occupy an intermedi-
ate cranial size between the tiny C. neglecta and the medium-sized
C. beccarii (Fig. 2A). A bivariate plot of upper toothrow length and
tail length (Fig. 2B) clearly distinguishes the individual species,
and the 3 new species all have relatively longer tails than any
previously described species with a similar toothrow length. In
addition, MZB 16790 and the C. aequicauda type specimen, FMSM
448/14, occupy a distinct region of morphospace (Fig. 2B). The rst
2 PCs of the dorsal landmark analysis (Fig. 2C) show substantial
overlap between many species; however, the new species from
Talamau occupies its own region of morphospace. The rst 2 PCs
of the ventral landmark analysis (Fig. 2D) however clearly distin-
guish most Sumatran species, and the 3 new species each occupy
their own regions of morphospace and show little overlap with
previously described taxa.
Linear regression models of cranial measurement all demon-
strated proper convergence. The size-corrected estimates of lin-
ear cranial measurements display a distinctive suite of cranial
features for each Sumatran Crocidura species (Fig. 3). For example,
all 3 of the new species have narrow rostra for their skull size
(estimates are negative, or below the global mean of zero), but
only 2 have a dorsoventrally compressed skull (negative relative
postpalate depth values; Fig. 3). Additional comparisons are dis-
cussed below.
Genetic evidence
All new sequences were published as GenBank accession num-
bers OQ944004 to OQ944100 (Supplementary Data SD3). Our
mitochondrial alignment of Southeast Asian shrews contains
760 characters. ModelFinder selected the GTR + I (proportion
of invariable sites) + Γ (gamma-distributed rate heterogene-
ity) model, then used this substitution model to estimate the
topology. The resulting mitochondrial gene tree shows 3 mono-
phyletic, unnamed taxa from Sumatra (Fig. 4; Supplementary
Data SD6). These 3 species form a clade with the specimen MZB
16790 as well as C. monticola and C. umbra from Java, nested
within a larger, moderately supported clade that also contains
C. neglecta from Sumatra and Borneo and several species from
mainland Asia and Taiwan. Jukes–Cantor mitochondrial dis-
tances among Sumatran species (Table 1) range from 0.04 to
0.15. Each of the 3 new species is at least 0.08 distant from
other Sumatran species. Intraspecic mitochondrial diver-
gence means range from 0 to 0.02 (Table 1).
For our concatenated alignment of 3,243 characters from
6 nuclear exons, ModelFinder selected the HKY + I model for
BDNF and Ptger4, HKY + R (FreeRate model; Yang 1995) for Apob
and Brca, GTR+G for Ghr, and GTR + I + G for vWf. The resulting
nuclear tree recovers each of the 3 unnamed taxa from Sumatra
as monophyletic (Fig. 5; Supplementary Data SD7). Two species
from Singgalang form a clade with MZB 16790 that is sister to C.
monticola and C. umbra from Java. The third new taxon is sister to
this clade of 5 species.
All parameters in the StarBEAST analysis demonstrated evi-
dence of convergence (ESS > 200). The resulting species-tree topol-
ogy from 6 exons places the 2 new species from Mt. Singgalang
as sister to C. umbra and C. monticola from Java, and the new spe-
cies from Mt. Talamau as sister to these 4 taxa, all with moderate
to high nodal support (Fig. 6). These results are concordant with
the concatenated nuclear tree and previous species-tree analyses
using nuclear DNA (Demos et al. 2016a). Individual gene trees are
presented as Supplementary Data SD9.
The results of our mitochondrial gene tree estimation, concat-
enated nuclear tree estimation, and morphometric analyses sup-
port the presence of 3 previously undocumented Crocidura species
on Sumatra. We describe these species below. Comparisons
between the new and previously described taxa focus on similarly
sized species, as these are the most difcult to distinguish. When
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 5
describing dentition, we use the dental homology of Hutterer
(2005): I3 C1 P1 M3/i2 c1 p0 m3 = 28. All measurements referenced
in the descriptions are available in Table 2.
Systematics
Crocidura balingka, new species
Balingka Shrew
Crocidura sp. nov. 2: Demos et al. 2016a:5164
Crocidura sp. 1: Demos et al. 2016b:8
Holotype
LSUMZ 40239, an adult male collected 23 November 2018 by
Ahmad Mursyid and prepared as a cleaned skull and skeleton,
study skin, and frozen liver and lung tissues. External measure-
ments are total length = 108 mm, tail length = 48 mm, hind-foot
length = 12 mm, ear length = 8 mm, and mass = 5.6 g. The voucher
specimen and a tissue sample will be permanently curated at
MZB with the catalog number MZB 43651; duplicate tissue sam-
ples will be retained at LSUMZ.
Type locality
Indonesia, Sumatera Barat, Agam, Kecamatan IV Koto, Kenagarian
Balingka, Jorong Pahambatan, northern slope of Mt. Singgalang,
0.37389°S, 100.32799°E, 2,100 m elevation.
Paratypes
LSUMZ 40214, 40234, 40235, 40236, 40237, 40238, 40240, 40241,
40242.
Other examined material
FMNH 212952, 212964, 212965, 212967, 212971, 212972, 212973,
212974, 212976.
Etymology
Named in honor of the people of Balingka, a village in IV Koto,
Agam, West Sumatra, who, through their generosity and knowl-
edge, have provided critical support to Indonesian and foreign
mammalogists interested in the biodiversity of Mt. Singgalang.
We recommend the English common name “Balingka Shrew.”
Fig. 2. Inter- and intraspecic morphological variation in Sumatran Crocidura support the existence of 3 new species. (A) First 2 axes of a PCA from
11 linear cranium dimensions. The rst axis, representing 91% of the variation, separates most species by size. (B) Bivariate plot of the natural log of
tail length and condyloincisive length for the 8 Sumatran Crocidura. Lower plots show the rst 2 PCs of an analysis of 16 dorsal landmarks (C) and 23
ventral landmarks (D). All linear measurements and PC scores are reported in Supplementary Data SD1.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
6 | Nations et al.
Diagnosis
Crocidura balingka is a small- to medium-sized shrew (4.1 to 6.5 g),
dark gray-brown in color, with a thin, medium-length tail roughly
80% of head–body length (Supplementary Data SD9). Individual
hairs of the dorsum are paler brown at the base and darker brown
toward the tips. The ventral pelage is slightly paler (gray-brown)
than the dorsal pelage; however, there is no clear demarcation
between dorsal and ventral coloration. The tail is relatively long
for similar-sized Sumatran Crocidura, with very short, sparse
applied hairs along the entire length of the tail, giving a hairless
impression when viewed with the naked eye. However, a small
tuft of 1-mm-length hairs is visible at the tail tip. The proximal
portion of the tail is lightly bristled with 3-mm hairs extending
to approximately 1/2 of the total tail length. The skull is small
(16.9 to 17.9 mm CIL; Table 2) and has a poorly developed lamb-
doidal crest. The junction of the lambdoidal suture and the sag-
ittal suture is anteriorly positioned on the skull. The edges of the
braincase are not rounded but taper both anteriorly and posteri-
orly from the widest point, which juts out laterally. The braincase
is relatively narrow for a skull of this size. The rostrum tapers
strongly, but anterior to the I3 the taper wanes, resulting in a nar-
row rostrum (Fig. 3) with lateral surfaces that run parallel to the
midline of the skull. The palate is relatively small in both length
and breadth and the upper tooth row is relatively short, resulting
in a longer postpalatal length relative to the skull size (Fig. 3).
The coronoid processes are strongly tapered dorsally, and angular
processes are very long and thin.
Description
Dorsal pelage dark gray-brown (Supplementary Data SD9). Hairs
are medium gray near the base and darken to charcoal brown
toward the tips. The dorsal pelage is short, measuring 3 mm in
length at mid-dorsum. The ventral pelage is slightly lighter in
color, but there is no clear demarcation between dorsal and ven-
tral coloration. The dorsal portions of the hind feet are brown,
with sparse, short, medium-brown hairs, similar in color to the
dorsum, covering the entire surface. There are faint tufts of light
brown hairs at the distal ends of the digits. The digits are slightly
paler than the proximal regions of the foot. The forefeet show
a similar color and pilosity to the hind feet and have very faint
tufts of light brown hairs at the distal phalanges (1 to 4 hairs
per digit). The rhinarium is brown and the lips are unpigmented.
Most mystacial vibrissae are 12 to 13 mm in length, with the long-
est extending 16 to 18 mm from the face. The chin vibrissae are
4 mm long and very sparse. Hairs on the dorsal portion of the
hind feet are dark, similar in color to the hairs of the dorsum,
with detectible tufts near the distal ends of the digits. The hind
feet are narrow. The sole and palm are pale. The thenar pads are
Fig. 3. Size-corrected cranial measurement variation in Sumatran Crocidura. Mean-centered and scaled measurement estimates are on the y-axis, and
species are on the x-axis. The point represents the median measurement probability, the wider vertical line indicates the 60% probability interval, and
the thin vertical line represents the 89% probability interval. All measurement estimates (apart from condyloincisive length) are conditioned on the
condyloincisive length, reecting an estimate of each cranial character free from the inuence of skull length. The mean value for each measurement
is 0, shown with a dotted line. The greater uncertainty around C. aequicauda estimates stems from the small sample size (n = 1 to 2).
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 7
Fig. 4. Maximum likelihood Cytb gene tree inferred in IQTree. Black circles at nodes indicate a UFBoot2 bootstrap value of ≥95%. The remaining
UFBoot2 values are reported at the nodes. Sumatran species are highlighted. The locality of each specimen is noted in parentheses in the form of
island locality for insular species. Nodes are collapsed apart from the 3 newly described species, and the outgroup branch length to Suncus murinus is
truncated for a clearer presentation of results. The gure without the collapsed nodes is available as Supplementary Data SD6.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
8 | Nations et al.
medium brown, and the hypothenar and digital pads are pale
brown. Claws on the fore and hind feet are medium in length,
with an intermediate curvature. The tail is medium length and
slender, with sparse 3-mm bristles on the proximal half of the
tail. Applied hairs are short, giving the tail a hairless impression
when viewed with the naked eye. They are dark dorsally, gradu-
ally becoming a lighter gray-brown ventrally. The tip of the tail
has a tuft of 1-mm hairs that are slightly paler than the applied
hairs. The skull is small and narrow for a shrew of this size. The
lambdoidal crest is weakly formed. The interorbital constriction
is narrow for a Sumatran Crocidura of this size (Table 2; Fig. 3).
The upper toothrow is short and the palate is narrow for the skull
length, resulting in a long postpalatal region (Fig. 3). The rostrum
narrows sharply toward the distal portion of the maxilla (Fig. 7).
I1 is falciform and rounded and extends beyond the premaxilla. It
is larger than I2, and I2 is larger than I3. The upper canine is dis-
tinctly larger than I3. The P4 talonid is dorsoventrally compressed
with a strong convex posterior margin. M1 is slightly larger in
occlusal area and height than M2, and M3 is considerably smaller
than both.
Comparisons
Crocidura balingka is easily distinguished from the larger Sumatran
Crocidura (C. lepidura, and C. hutanis) based on the size of all exter-
nal and cranial measurements (Table 2). The new species is sim-
ilar in size to C. beccarii, C. paradoxura, and C. neglecta. Though
similar in body length and mass, C. paradoxura has a long tail
(Table 2), elongate skull (as evidenced by a long condyloincisive
length and narrow braincase; Fig. 3; Table 2), pale feet with lightly
pigmented skin, and distinct light gray coloration, none of which
are shared by C. balingka. Therefore, C. balingka will be compared
to C. beccarii and C. neglecta in turn.
Crocidura balingka overlaps with C. beccarii in all external meas-
urements (head–body length, tail length, hind-foot length, ear
length, and mass); however, C. balingka is less robust than C. bec-
carii and has a lower mean head–body length, hind-foot length,
and mass measurements (Table 2). The new species has a higher
mean relative tail length than C. beccarii (Table 2). Dorsally, the
feet are darker in C. balingka than in C. beccarii. The claws of C.
balingka are slightly shorter and more sharply curved than those
of C. beccarii. Consistent with external measurements being
smaller, the skull of C. balingka is more diminutive than that of C.
beccarii, as evidenced by smaller condyloincisive length, postpal-
atal depth, brain breadth, interorbital width, rostral length, and
rostral width (Table 2). The foramen magnum is proportionally
larger in C. balingka than in C. beccarii. Crocidura balingka also has
a similar palatal width but narrower M2–M2 width, indicating less
robust dentition than in C. beccarii (Table 2; Fig. 3). These cranial
differences manifest on the rst axis of the linear measurement
PCA, and the dorsal- and ventral-landmark PCAs (Fig. 2).
The smallest Crocidura found on Sumatra is C. neglecta. This
species is found on both Sumatra and Borneo, but our compar-
isons are limited to specimens from Sumatra. Though they over-
lap in some external and cranial measurements, C. balingka has
higher average measurement values for all external and cranial
characters (Table 2). Crocidura balingka also has a distinctly longer
tail than C. neglecta, both in absolute and relative measures. The
palate of C. balingka is narrower and longer than of C. neglecta, as
evidenced by a longer toothrow, longer P4–M3, and narrower M2–M2
width relative to skull length (Fig. 3). The feet of C. balingka are
slightly darker than the pale brown feet of C. neglecta. The base
of the tail of C. balingka is slender, unlike the proximally thick tail
base in C. neglecta that strongly tapers distally. The ventral portion
of the tail is also much darker in C. balingka than in C. neglecta.
The pinnae of C. balingka are much darker than the pale pinnae
of C. neglecta.
Distribution and ecology
Crocidura balingka appears to be a microendemic, montane spe-
cialist restricted to elevations above 2,000 m on Mt. Singgalang.
During 2 surveys of Mount Singgalang, C. balingka was frequently
caught in pitfall traps above 2,000 m. A brief survey (only 10 pit-
fall trap-nights) at 2,826 m near the summit of Mt. Singgalang
did not obtain any C. balingka specimens, though given this lim-
ited effort, we contend that its upper elevational limit remains
unclear. Substantial pitfall effort (around 300 pitfall nights total)
at 1,700, 1,800, and 1,880 m did not capture any C. balingka, sug-
gesting that there is a lower limit to its elevational distribution.
Table 1. Inter- and intraspecic Cytb distances between Sumatran Crocidura species. The means are reported, with the minimum and
maximum distances in parentheses. The diagonal shows intraspecic distances.
aequicauda balingka barapi beccarii dewi hutanis lepidura neglecta paradoxura
aequicauda -
balingka 0.09
(0.09, 0.09)
0 (0, 0)
barapi 0.10
(0.10, 0.11)
0.11
(0.10, 0.12)
0
(0.0, 0.01)
beccarii 0.14
(0.14, 0.15)
0.14
(0.13, 0.15)
0.13
(0.12, 0.14)
0.02
(0, 0.03)
dewi 0.08
(0.08, 0.08)
0.08
(0.08, 0.08)
0.09
(0.08, 0.09)
0.13
(0.13, 0.14)
0 (0,0)
hutanis 0.17
(0.15, 0.17)
0.15
(0.14, 0.16)
0.14
(0.13, 0.15)
0.06
(0.04, 0.07)
0.14
(0.13, 0.15)
0.02
(0, 0.04)
lepidura 0.15
(0.15, 0.16)
0.15
(0.14, 0.15)
0.13
(0.13, 0.14)
0.05
(0.05, 0.06)
0.14
(0.14, 0.14)
0.04
(0.03, 0.05)
0 (0,0)
neglecta 0.11
(0.09, 0.12)
0.15
(0.14, 0.16)
0.12
(0.12, 0.13)
0.13
(0.13, 0.14)
0.14
(0.13, 0.15)
0.14
(0.14, 0.15)
0.13
(0.13, 0.14)
0.02
(0, 0.04)
paradoxura 0.15
(0.12, 0.15)
0.14
(0.13, 0.16)
0.13
(0.13, 0.14)
0.08
(0.07, 0.09)
0.15
(0.15, 0.16)
0.08
(0.07, 0.10)
0.09
(0.09, 0.09)
0.14
(0.13, 0.15)
0.02
(0, 0.04)
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 9
Anthropogenic forest disturbance extends to roughly 1,600 m on
Mt. Singgalang, which may affect the distribution of C. balingka;
however, our pitfall lines at 1,800 and 1,880 m were in forest with
minimal human impact, and captured several other Crocidura,
further suggesting a restricted, montane distribution for this
species. Crocidura beccarii, C. paradoxura, and the Crocidura from
Singgalang described below were captured in the same pitfall
lines as C. balingka, suggesting syntopy among at least these 4
Fig. 5. Maximum likelihood phylogeny inferred in IQTree from a concatenated alignment of 6 nuclear genes detailing the phylogenetic placement of
Crocidura balingka sp. nov., C. dewi sp. nov., and C. barapi sp. nov. Black circles at nodes indicate a UFBoot2 bootstrap value of >0.95. Sumatran species
are highlighted. Nodes of non-Sumatran taxa are collapsed. The gure without the collapsed nodes is available in Supplementary Data SD7.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
10 | Nations et al.
species. Crocidura balingka was not detected in similar habitats at
similar elevations on recent surveys of Mt. Tujuh (185 km SE of
Mt. Singgalang) or Mt. Talamau (65 km NW of Singgalang).
Crocidura dewi, new species
Dewi Shrew
Holotype
LSUMZ 40245, an adult female, collected 4 December 2018 by
Jonathan Nations and prepared as a cleaned skull and skeleton,
study skin, frozen liver and lung tissue samples, and a gastroin-
testinal tract preserved in ethanol. External measurements: total
length = 163 mm, tail length = 73 mm, hind-foot length = 17 mm,
ear length = 12 mm, and mass = 10.7 g. The voucher specimen
and a tissue sample will be permanently curated at MZB with
the catalog number MZB 43652, with additional tissues retained
at LSUMZ.
Type locality
Indonesia, Sumatera Barat, Agam, Kecamatan IV Koto, Kenagarian
Balingka, Jorong Pahambatan, Mt. Singgalang, Talago Dewi crater
lake; 0.39489°S, 100.33360°E, 2,826 m elevation.
Paratypes
LSUMZ 40216, 40217, 40246, 40247, 40248.
Etymology
Noun of the Indonesian word for “goddess.” Named for the pic-
turesque crater lagoon Talago Dewi, near the summit of Mt.
Singgalang where 4 of the 6 specimens were captured. We recom-
mend the English common name “Dewi Shrew.”
Diagnosis
Crocidura dewi is a medium-sized shrew (8.2 to 10.7 g), gray-brown
in color, with thick, wooly fur that is 7 mm long at mid-dorsum
(Supplementary Data SD9). Individual hairs are dark gray at the
base, paler and browner at the distal tips, and overall, the pelage is
iridescent. The dorsal pelage is only slightly darker than the ven-
tral pelage, and there is no clear demarcation between the two.
The tail is relatively long, roughly 85% of the head–body length,
with very short applied hairs that are uniform in color dorsoven-
trally. The tail is lightly bristled with 4-mm hairs on the proximal
1/3 of the tail. The dorsal portion of the hind feet are chocolate
brown proximally, and paler distally, with distinct 1 to 2 mm pale
gray hair tufts on the tips of digits II and III. The claws are deeply
curved. The braincase is tall and compressed anteroposteriorly,
with a short condyle–glenoid and postpalatal length relative to
skull length (Fig. 3), giving the braincase a circular shape from
the dorsal view. A lambdoidal crest is present. The rostrum has a
strong taper that wanes anterior to the I3, resulting in a narrow,
straight rostrum with edges parallel to the skull midline (Fig. 7).
Dentition is small for a shrew of this size, evidenced by a wide
palate but a narrow M2–M2 distance relative to skull length (Fig. 3).
Description
Gray-brown in color, with thick and wooly fur that measures
7 mm at mid-dorsum (Supplementary Data SD9). The pelage is
dark gray at the base, becoming lighter and browner at the distal
tips, the opposite pattern of other Sumatran Crocidura. The fur has
a shiny, iridescent quality. The ventral pelage is very slightly paler
than the dorsal pelage. Pinnae hairs are short, brown, and lighter
at the tip. The rhinarium is brown and the lips are unpigmented.
Mystacial vibrissae are long, up to 19 mm. The mental vibrissae
are short and sparse. Dorsal portions of the hind feet are choc-
olate brown, transitioning to paler brown near the digits. There
are distinct tufts of pale brown and gray hairs at the tips of the
hind digits, especially on digits II and III. The forefeet are similar
in color and pilosity to the hind feet. The thenar and hypothenar
pads of the fore and hind feet are dark brown, and the digital
pads are medium brown. The claws on the fore and hind feet are
well-developed and deeply curved. The tail is medium length to
long (Table 2), with sparse, short 4-mm bristles on the proximal
third of the tail. Applied hairs on the tail are uniformly brown and
short, giving the tail a hairless impression. The skull is medium in
size. The interorbital constriction is wide for a Sumatran Crocidura
of this size (Fig. 3). The posterior edge of the occipital bone curves
gently outward then inward at the midline, then outward again
over the foramen magnum. The lambdoid crest is present but
not well-developed. The braincase is short anteroposteriorly,
with a width just below average for a skull of this size, result-
ing in a rounded braincase (Figs. 3 and 7). The widest portion of
the zygomatic process of the maxilla is posteriorly angled, giv-
ing the impression that the zygomatic process is pointing toward
the posterior of the skull rather than laterally. This point extends
to the posterior edge of M2. Anterior to the infraorbital foramen,
the angle of the rostrum moves sharply toward the midline of
the skull, and the rostrum becomes very narrow anteriorly. I1 is
falciform and rounded and extends beyond the premaxilla. I2
Fig. 6. Species tree estimated using StarBEAST detailing the
phylogenetic placement of Crocidura balingka sp. nov., C. dewi sp.
nov., and C. barapi sp. nov. Black circles at nodes indicate a posterior
probability (pp) > 0.9, and gray squares indicate a pp < 0.9 and >
0.7. Nodes with a pp < 0.7 have values shown. Sumatran species are
highlighted. Outgroup branches are truncated. The scale bar is in
coalescent units.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 11
Table 2. Cranial and external measurements from 8 species of Sumatran Crocidura. All measurements are in millimeters except Mass (g). Columns represent species, with the number
of specimens sampled in parentheses. Cells report mean value ± standard deviation, with the range of measurements in parentheses. Total = total length; HB = head and body length;
Tail = tail length; HF = hind-foot length with claw; Mass = body mass; CIL = condyloincisive length; BBC = breadth of braincase; IOW = interorbital width; RL = rostral length; RW =
rostral width; PPD = postpalatal depth; PPL = postpalatal length; CGF = distance from occipital condyle to glenoid fossa; UTL = upper toothrow length; P4–M3 = distance from alveolar of
P4 to M3; M2–M2 = labial width at M2. For C. aequicauda specimens, cells with single values from MZB 16790, and empty cells are missing measurements.
Measurement C. aequicauda
(n = 2)
C. balingka (n = 13) C. barapi (n = 9) C. beccarii (n = 17) C. dewi (n = 4) C. hutanis (n = 9) C. lepidura (n = 14) C. neglect (n = 17) C. paradoxura (n
= 16)
Total 154 ± 4.24
(151 to 157)
111.23 ± 5.21
(102 to 122)
115.22 ± 3.99
(110 to 121)
124.88 ± 5.77
(114 to 132)
155.25 ± 5.8
(149 to 163)
134.44 ± 7.6
(121 to 148)
166.57 ± 11.12
(147 to 192)
100.82 ± 5.23
(93 to 112)
183.94 ± 6.69
(171 to 193)
HB 73.5 ± 4.5
(70 to 77)
61 ± 4.51
(53 to 69)
62.22 ± 2.82
(59 to 66)
73.35 ± 4.17
(65 to 80)
84.75 ± 3.59
(82 to 90)
80 ± 6.46
(70 to 91)
95.79 ± 9.21
(80 to 117)
61.88 ± 3.53
(56 to 70)
77.19 ± 5.72
(65 to 85)
Tail 80.5 ± 0.71
(80 to 81)
50.23 ± 1.83
(48 to 53)
53 ± 1.73
(51 to 55)
51.53 ± 2.74
(46 to 55)
70.5 ± 2.52
(67 to 73)
54.44 ± 2.51
(51 to 58)
70.79 ± 4.08
(63 to 76)
38.94 ± 3.38
(33 to 44)
106.75 ± 4.23
(99 to 116)
HF 15.5 ± 0
(15.5)
11.46 ± 0.78
(10 to 13)
12.67 ± 0.5
(12 to 13)
13.12 ± 0.49
(12 to 14)
16.75 ± 0.5
(16 to 17)
14 ± 0.87
(13 to 16)
18.14 ± 1.41
(15 to 20)
10.88 ± 1.11
(7 to 12)
17 ± 0.89
(15 to 18)
Mass 5.22 ± 0.58
(4.1 to 6.5)
6.23 ± 0.54
(5.4 to 7)
7.1 ± 1.31
(5.2 to 10.2)
10.05 ± 1.23
(8.2 to 10.7)
9.3 ± 1.66
(7.6 to 11.4)
19.16 ± 3.99
(13.4 to 27.5)
4.48 ± 0.72
(3.5 to 6.2)
8.55 ± 1.08
(6.2 to 10)
CIL 21.7 17.33 ± 0.32
(16.92 to 17.96)
18.39 ± 0.46
(17.8 to 19.39)
19.86 ± 0.52
(18.93 to 20.67)
21.48 ± 0.26
(21.22 to 21.83)
20.74 ± 0.67
(20.05 to 21.82)
24.96 ± 0.82
(23.2 to 26.29)
16.54 ± 0.42
(15.7 to 17.09)
20.99 ± 0.44
(20.22 to 21.65)
BBC 9.4 7.92 ± 0.17
(7.65 to 8.27)
8.48 ± 0.23
(8.06 to 8.78)
9.02 ± 0.23
(8.54 to 9.37)
9.38 ± 0.27
(9 to 9.61)
9.29 ± 0.4
(8.81 to 10.08)
10.72 ± 0.37
(10.04 to 11.41)
7.73 ± 0.22
(7.31 to 8.06)
8.9 ± 0.35
(7.79 to 9.22)
IOW 4.7 3.8 ± 0.08
(3.67 to 4.01)
4.21 ± 0.13
(4.02 to 4.46)
4.28 ± 0.15
(3.98 to 4.59)
4.78 ± 0.08
(4.67 to 4.83)
4.37 ± 0.16
(4.14 to 4.61)
5.3 ± 0.34
(4.79 to 6.05)
3.84 ± 0.22
(3.62 to 4.51)
4.65 ± 0.23
(4.34 to 5.28)
RL 6.75 ± 0.17
(6.45 to 7.04)
7.47 ± 0.27
(7.24 to 8.13)
7.99 ± 0.31
(7.5 to 8.71)
8.84 ± 0.25
(8.52 to 9.13)
8.49 ± 0.45
(7.99 to 9.24)
10.45 ± 0.57
(9.18 to 11.27)
6.47 ± 0.22
(5.9 to 6.74)
8.31 ± 0.22
(7.9 to 8.72)
RW 2.3 1.78 ± 0.09
(1.62 to 1.92)
1.81 ± 0.14
(1.7 to 2.16)
2.32 ± 0.22
(1.99 to 2.72)
2.12 ± 0.05
(2.05 to 2.17)
2.39 ± 0.14
(2.18 to 2.58)
2.9 ± 0.5
(2.39 to 3.63)
1.94 ± 0.11
(1.76 to 2.16)
2.28 ± 0.19
(2.05 to 2.77)
PPD 3.9 3.17 ± 0.13
(2.95 to 3.36)
3.33 ± 0.18
(3.16 to 3.73)
3.66 ± 0.11
(3.49 to 3.82)
4.04 ± 0.19
(3.8 to 4.24)
3.84 ± 0.19
(3.49 to 4.04)
4.47 ± 0.16
(4.18 to 4.7)
3.22 ± 0.11
(3.06 to 3.39)
3.95 ± 0.17
(3.66 to 4.23)
PPL 11.8 8.07 ± 0.21
(7.67 to 8.4)
8.45 ± 0.43
(7.97 to 9.46)
8.93 ± 0.34
(8.24 to 9.55)
9.55 ± 0.34
(9.23 to 10)
9.44 ± 0.43
(8.93 to 9.95)
11.1 ± 0.39
(10.34 to 11.71)
7.78 ± 0.27
(7.32 to 8.31)
9.65 ± 0.23
(9.17 to 10.08)
CGF 8.5 6.96 ± 0.19
(6.71 to 7.33)
7.22 ± 0.3
(6.8 to 7.77)
7.5 ± 0.24
(6.97 to 7.78)
7.82 ± 0.23
(7.52 to 8.06)
7.84 ± 0.36
(7.43 to 8.35)
9.01 ± 0.27
(8.5 to 9.47)
6.57 ± 0.25
(6.16 to 6.93)
8.2 ± 0.26
(7.75 to 8.69)
UTL 9.35 ± 0.07
(9.3 to 9.4)
7.47 ± 0.16
(7.25 to 7.8)
8.15 ± 0.39
(7.82 to 9.08)
8.84 ± 0.24
(8.4 to 9.13)
9.68 ± 0.14
(9.56 to 9.87)
9.28 ± 0.3
(8.92 to 9.62)
11.23 ± 0.5
(10.02 to 11.9)
7.03 ± 0.18
(6.65 to 7.28)
9.15 ± 0.22
(8.7 to 9.43)
P4 –M35.6 4.32 ± 0.17
(4.04 to 4.61)
4.75 ± 0.26
(4.49 to 5.18)
5.06 ± 0.19
(4.54 to 5.27)
5.44 ± 0.3
(4.99 to 5.64)
5.24 ± 0.16
(5.05 to 5.53)
6.19 ± 0.37
(5.39 to 6.68)
3.91 ± 0.14
(3.62 to 4.12)
5.12 ± 0.21
(4.62 to 5.38)
M2–M26.15 ± 0.49
(5.8 to 6.5)
4.75 ± 0.12
(4.55 to 4.93)
5.16 ± 0.19
(5 to 5.66)
5.79 ± 0.18
(5.36 to 6.06)
6.12 ± 0.1
(6 to 6.23)
6.02 ± 0.25
(5.59 to 6.39)
7.33 ± 0.36
(6.48 to 7.93)
4.99 ± 0.13
(4.77 to 5.21)
5.92 ± 0.17
(5.54 to 6.14)
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
12 | Nations et al.
is much larger than I3. The upper canine is larger than I3. M1 is
slightly larger than M2, and M3 is smaller in comparison to the
other upper molars but large and well-developed for a Southeast
Asian Crocidura (Fig. 7). The mandible is gracile for a shrew of this
body size. The coronoid process is short and strongly tapers dor-
sally. The angular processes are long and thin. i1 is dorsoventrally
compressed.
Comparisons
Overall body size distinguishes C. dewi from the smaller Sumatran
Crocidura (C. balingka, C. beccarii, and C. neglecta; Table 2) and
shorter tail length, browner pelage, and darker feet easily distin-
guish C. dewi from C. paradoxura (Table 2; Supplementary Data
SD9). Detailed comparisons of C. dewi are focused on the larger
Sumatran species, C. hutanis and C. lepidura.
Crocidura dewi has a similar head–body length and tail length
as C. lepidura; however, C. lepidura has a greater mass (Table 2) and
is a much darker chocolate brown in color with shorter, 5 mm
dorsal pelage. The feet of C. dewi are paler than those of C. lepi-
dura, and also have more hair and pale-colored tufts at the distal
end of the rear digits II and III that are absent in C. lepidura. The
claws of C. dewi are distinctly more curved than those of C. lepi-
dura, which has only weakly curved claws that extend outward
from the digits. Crocidura dewi is smaller in nearly every cranial
measurement than C. lepidura (Table 2). The i1 of C. dewi is narrow
and gracile while the i1 of C. lepidura is robust.
Crocidura dewi can be distinguished from C. hutanis by its lighter
gray-brown coloration, longer dorsal pelage (7 mm vs. 4 mm in C.
hutanis), and much longer tail, both absolute and relative values.
The hind foot of C. dewi is longer on average than the hind foot
of C. hutanis. The claws of C. dewi are much more curved than the
lightly curved claws of C. hutanis. The overall appearance of the
cranium of C. dewi is less robust than the cranium of C. hutanis,
with smaller dentition, a weak lambdoidal crest, and a dorsoven-
trally narrower rostrum. Crocidura dewi has a much wider palate
but narrower M2–M2 width than C. hutanis, relative to skull length
(Fig. 3), indicating smaller cheek teeth. The dentaries of C. dewi
are less robust than C. hutanis.
Distribution and ecology
As with the sister taxon C. balingka, C. dewi appears to be a
microendemic, montane specialist, most abundant at high eleva-
tions on Mt. Singgalang. Crocidura dewi was captured at 1,880 m
(1 specimen), 2,100 m (1 specimen), and 2,826 m (4 specimens) on
Mt. Singgalang. Crocidura neglecta, C. beccarii, C. balingka, C. hutanis,
and C. paradoxura were all caught in pitfall lines with C. dewi. The
number and elevational distribution of the captures, considered
in conjunction with the limited nature of our sampling at 2,826
m (only 10 pitfall nights), suggest that C. dewi is less abundant at
the lower levels of its elevational range (1,800 to 2,000 m) than at
the highest elevations. The long, wooly fur of C. dewi is reminis-
cent of the fur of other shrews and rodents in tropical montane
habitats. The apparent lower elevational limit of C. dewi (1,880 m)
is below that of C. balingka and at the upper limit of C. hutanis and
C. neglecta captures on Mt. Singgalang. Claw curvature has been
used to describe locomotor or microhabitat afnities in small
mammals, with highly curved claws associated with climbing
(Tulli et al. 2016). The deeply curved claws of C. dewi, vastly dif-
ferent from the long, slightly curved claws of the similarly sized
C. lepidura and C. hutanis, suggest climbing or scrambling behav-
ior rather than burrowing, but currently nothing is known of its
habits.
Crocidura barapi, new species
Barapi Shrew
Fig. 7. Images showing dorsal, and ventral views of the skull and lateral views of the dentary of the holotypes of 3 new shrew species from Sumatra:
(A) Crocidura balingka, LSUMZ 40239; (B) Crocidura dewi, LSUMZ 40245; (C) Crocidura barapi, LSUMZ 39782 (image reversed horizontally).
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 13
Holotype
LSUMZ 39782, an adult male, collected 21 April 2018 by Heru
Handika and prepared as a cleaned skull and uid-preserved
specimen and frozen liver samples. External measurements: total
length = 113 mm, tail length = 51 mm, hind-foot length = 13 mm,
ear length = 9 mm, and mass = 6.4 g. The voucher specimen and
a tissue sample will be permanently curated at MZB with the
catalog number MZB 43653, with a duplicate tissue retained at
LSUMZ.
Type locality
Indonesia, Sumatera Barat, Pasaman Barat, Kecamatan Pasaman,
Nagari Aur Kuning, Jorong Lubuak Landua, Mt. Talamau,
0.08660°N, 99.96884°E, 2,004 m elevation.
Paratypes
LSUMZ 39748, 39752, 39759, 39768, 39772, 39773, 39776, 39777,
39785, 39792.
Etymology
Noun for “volcano” in the Minangkabau language of West
Sumatra, a reference to the Talamau volcano where the discov-
ery took place and an homage to the Minangkabau people of the
region who facilitated this research. We recommend the English
common name “Barapi Shrew.”
Diagnosis
Crocidura barapi is a small to medium shrew (5.4 to 7 g), dark gray-
brown in color, with slightly wooly fur that measures 5 mm at
mid-dorsum (Supplementary Data SD9). The pelage is light gray
at the base, becoming chocolate brown at the distal tips. The ven-
tral pelage is slightly paler than the dorsal pelage, but it is not
discretely bicolored. The tail is relatively long for similar-sized
Southeast Asian Crocidura, roughly 85% of the head–body length,
with very short, sparse applied hairs. The proximal portion of
the tail is lightly bristled with 6-mm hairs extending to approxi-
mately 1/3 of the total tail length. The skull is small for its body
size, and at in lateral view (postpalatal depth; Table 2; Fig. 3).
The braincase is wide and the interorbital region is average for a
skull of this size (Fig. 3). There is no lambdoidal crest. The palate
is narrow. The rostrum is narrow and long, and strongly tapers
toward the anterior portion of the skull; however, the taper wanes
anterior to I3 (Fig. 7). The dentition is weak. Toothrow length and
palatal width are average for the skull size, but the M2–M2 width
is small, indicating weak molars.
Description
Dorsal pelage is a dark gray-brown, with hairs being grayer
toward the base and darkening to chocolate brown at the tips
(Supplementary Data SD9). The dorsal pelage measures 5 mm in
length at the mid-dorsum. The ventral pelage is lighter in color,
but there is no discrete boundary between the dorsal and ven-
tral pelage. The dorsal portion of the hind feet is uniformly pale
brown, with sparse, short, medium-brown hairs, similar in color
to the dorsum, covering most of the feet, but with greater den-
sity on medial portions. There are no distal tufts of hair on the
digits, and the digits are similar in color to the proximal regions
of the foot. The soles of the hind feet are pale brown, and the the-
nar, hypothenar, and interdigital pads are paler than the plantar
surface. The forefeet share a similar pale brown color and uni-
form, sparse pilosity with the hind feet. The claws on the fore and
hind feet are long and slightly curved. The pinnae hairs are short,
sparse, and uniformly brown. The rhinarium is brown and the
lips are unpigmented. The mystacial vibrissae are 12 to 13 mm
in length, with the longest extending 15 mm from the face. The
mental vibrissae are short and sparse. Hairs on the dorsal por-
tion of the hind feet are chocolate brown. The tail is relatively
long, with sparse 6-mm hairs on the proximal third of the tail.
Applied hairs are short, giving the tail a hairless impression, and
are dark on the dorsal portion of the tail, gradually becoming a
paler gray-brown ventrally. The skull is relatively small and nar-
row for a shrew of this size, and short dorsoventrally. There is no
lambdoidal crest. The braincase is wide, but the postpalatal and
condyle–glenoid lengths are average for the skull length, resulting
in a somewhat rounded braincase (Figs. 3 and 7). The interorbital
constriction is slightly wide for a Sumatran Crocidura of this size
(Fig. 3). The rostrum is long and very narrow, and tapers sharply
toward the distal portion of the maxilla (Fig. 3). The palate is aver-
age in width. The dentition is weak for a shrew of this size; inci-
sors and canines are relatively small, and premolars and molars
are not robust (Fig. 7). I1 is falciform and rounded and extends
beyond the premaxilla. I1 is larger than I2, and I2 is larger than
I3. The upper canine is only slightly larger than I3. M1 is slightly
larger than M2, and M3 is small in comparison.
Comparisons
Overall body size easily distinguished C. barapi from the larger
Sumatran Crocidura (C. lepidura, C. dewi, and C. hutanis), and
shorter tail length and darker pelage and feet easily distinguishes
C. barapi from C. paradoxura. We will therefore limit the C. barapi
comparisons to similarly sized Sumatran shrews, C. beccarii, C. bal-
ingka, and C. neglecta.
The only similar size shrew to C. barapi detected on Mt. Talamau
is C. neglecta. Though largely overlapping in head–body length, C.
barapi is a more robust shrew than C. neglecta with, on average,
a higher mass and hind-foot length, and longer, woolier pelage
(5 mm vs. 2.5 mm). Crocidura barapi can easily be distinguished
from C. neglecta by its much longer tail (Table 2). The pinnae, fore-
feet, and hind feet of C. barapi are much darker than C. neglecta.
The skull of C. barapi is also longer and wider than the skull of C.
neglecta in raw values (Table 2), and the dentition of C. barapi is
larger with nonoverlapping values of upper toothrow and molar
toothrow lengths (Table 2). However, relative to skull length, C.
barapi has a narrow M2–M2 width while C. neglecta has a wide pal-
ate, short molar row, and short toothrow (Fig. 3). These differences
in cranial morphology result in a separation of C. barapi and C.
neglecta in all measures of cranial morphospace (Fig. 2).
Crocidura barapi is generally smaller than the allopatric C.
beccarii with, on average, smaller head–body length, mass, and
hind-foot length values, though C. barapi has a longer average
tail length than C. beccarii (Table 2). The skull of C. barapi is more
diminutive than C. beccarii as evidenced by a smaller condyloinci-
sive length, narrower rostrum, shorter upper toothrow (Table 2),
and relatively narrower braincase (Fig. 3). The incisors take up a
proportionally smaller portion of the toothrow in C. barapi than
C. beccarii (Fig. 3). The occipital bone of C. barapi is relatively at
along the posterior edge of the cranium, while C. beccarii has a
more rounded posterior edge of the skull. These cranial differ-
ences lead to a clear separation of C. barapi, C. beccarii, and C.
neglecta on the rst 2 components of the linear and geometric
PCAs (Fig. 2).
Crocidura barapi is larger in most external and cranial meas-
urements than the allopatric C. balingka. It has a longer dorsal
pelage than C. balingka, and the stiff bristles on its tail do not
extend as far on the tail in C. barapi as in C. balingka. The skull
of C. barapi is on average larger than the skull of C. balingka in
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
14 | Nations et al.
all measurements, and interorbital width, rostral length, upper
toothrow length, and M2–M2 width are nonoverlapping (Table 2).
The dentition of C. barapi is more robust than the dentition of C.
balingka (Fig. 7), evidenced by a longer molar length and upper
toothrow length relative to skull size (Fig. 3). C. barapi has a wider
interorbit than C. balingka relative to skull size (Fig. 3). The hind
feet of C. barapi are darker and wider than those of C. balingka.
Currently C. barapi is known from only Mount Talamau; it was
not detected on Mount Singgalang where C. balingka is abundant.
Distribution and ecology
Crocidura barapi was relatively abundant above 1,662 m on Mount
Talamau. As of now, this species has only been detected on this
mountain. Pitfall traps were not placed above 2,004 m on Mount
Talamau. Given the overall montane distribution of C. barapi,
we suspect that it occurs above the 2,000 m sampling limit we
achieved on this nearly 3,000 m volcano. Crocidura barapi and C.
neglecta appear to be parapatrically distributed along the eleva-
tional gradient. In our survey, C. neglecta was detected from 1,050
to 1,662 m, while C. barapi was found from 1,600 to 2,004 m, or the
upper limits of our pitfall efforts. Crocidura barapi was found on
the same pitfall lines as C. paradoxura and C. lepidura, and found
to overlap with C. hutanis and C. neglecta only at the lower eleva-
tions of its known range (~1,650 to 1,700 m). Surveys on Mount
Singgalang, 65 km southeast of Talamau, did not detect C. barapi.
Long-tailed shrew from Mt. Tujuh
In 1991, a single shrew with a relatively long tail was collected
at 2,200 m on the slopes of Mt. Tujuh in Jambi Province (Fig.
1), assigned to the species C. paradoxura, and cataloged as MZB
16790/IZEA 4503 (Supplementary Data SD9). In a subsequent revi-
sion of Malay Archipelago Crocidura, MZB 16790 was compared to
the type of C. aequicauda, collected on nearby Mt. Kerinci (Fig. 1;
Robinson and Kloss 1918; Ruedi 1995). Based on the tail length
relative to head–body length, 104% in the C. aequicauda type, 116%
in MZB 16790, and 159% in the C. paradoxura type (Dobson 1886),
along with the relatively pale feet and gray-brown fur color, Ruedi
(1995) referred these 2 specimens to C. paradoxura. The taxon C.
aequicauda has therefore been considered to be a junior synonym
of C. paradoxura (Ruedi 1995). However, Ruedi (1995) did not exam-
ine the holotype of C. paradoxura when making this decision and
there were few, if any, additional C. paradoxura specimens availa-
ble at the time of his revision.
Recent molecular phylogenetics work, leveraging both mito-
chondrial and nuclear DNA sequences, revealed that the Mt. Tujuh
specimen MZB 16790 is not nested within C. paradoxura. Instead,
this specimen, referred to as “Crocidura cf. paradoxura MZB16790,”
“Crocidura sp 1 aH 2020,” and “Crocidura sp 1” (see Supplementary
Data SD1: Fig. S1, Supplementary Data SD2: Fig. S2, and Table S1
in Hinckley et al. 2022), is sister to C. balingka (“Crocidura sp. 2 TCD
2016” in Hinckley et al. 2022: Fig. S1). Our phylogenetic analyses
suggest that MZB 16790 is sister to the clade that contains C. bal-
ingka and C. dewi (Figs. 4 and 6) or nested within the clade (Fig. 5).
Since 1995, numerous C. paradoxura specimens, including topo-
types, have been collected and archived, and our examination of
MZB 16790 reveals that it is not representative of the distinctive
C. paradoxura (Supplementary Data SD9). The morphological (Figs.
2 and 3; Supplementary Data SD9) and genetic differences (≥0.08
Jukes–Cantor Cytb distance from other Sumatran shrews; Table 1)
between MZB 16790 and the 3 newly described taxa warrant its
recognition as a species, supporting the conclusions of Hinckley
et al. (2022). Previous morphological analyses determined that
MZB 16790 and the C. aequicauda type represent the same species
(Ruedi 1995), a conclusion supported by the proximity of their col-
lecting localities (Fig. 1). Therefore, we recognize C. aequicauda as
a species. Crocidura aequicauda is currently represented by 2 spec-
imens: the type (Federated Malay States Museum 448/14, also
cataloged as British Museum of Natural History 19.11.5.28) and
MZB 16790. Based on our examination of MZB 16790, C. aequicauda
is a medium-sized shrew that can be distinguished from other
Sumatran shrews using fur color and body proportions. The long
(7 to 8 mm) slate-gray fur is lighter than all Sumatran shrews
other than C. paradoxura (Supplementary Data SD9). It is larger in
head–body length than C. neglecta, C. balingka, and C. barapi, and
smaller in head–body length than C. dewi and C. lepidura (Table 2).
The tail of C. aequicauda is perhaps the most distinguishing fea-
ture, as it is longer in both absolute (Table 2) and relative length
(110% of head–body length) than all other Sumatran shrews apart
from C. paradoxura, which has a much longer tail, often with a
white tip (Supplementary Data SD9).
Discussion
Recent collecting efforts and subsequent morphological and
molecular analyses are revealing the degree to which small-
mammal diversity in Southeast Asia has been underestimated.
Here we present morphological and molecular evidence of the
presence of 3 new species of Crocidura on Sumatra. These discov-
eries bring the total number of Sumatran Crocidura to 9, including
the Bangka Island endemic, C. vosmaeri. Apart from C. neglecta,
all Sumatran shrew species are endemic to the island, though
molecular evidence suggests that there is substantial inter-island
evolutionary divergence within C. neglecta (Demos et al. 2016a;
Hinckley et al. 2022; this study). Results from surveys of 3 moun-
tains in West Sumatra suggest that each of these new taxa
has a very limited distribution and may be endemic to a single
mountain. The 2 new species from Mt. Singgalang, C. dewi and C.
balingka, were not detected on Mt. Talamau despite similar ele-
vational sampling, which lies only 65 km to the northwest (Fig.
1). Similarly, C. barapi was only detected on Mt. Talamau. None
of these 3 species were detected on Mt. Tujuh, 190 km southeast
of Mt. Singgalang and 250 km southeast of Mt. Talamau (Fig. 1).
Similar patterns of localized, montane endemism have been
noted in other Sumatran mammals (e.g. squirrels; Hinckley et
al. 2020) and in Crocidura on the neighboring islands of Borneo
(Hinckley et al. 2022), Java (Esselstyn et al. 2013; Demos et al.
2016a), and Sulawesi (Esselstyn et al. 2021).
The newly described C. balingka and C. dewi are recovered
as recently diverged, reciprocally monophyletic sister taxa in
the mitochondrial gene tree (Fig. 4), the coalescent-based spe-
cies tree (Fig. 6), and in 5 of the 6 individual exon gene trees
(Supplementary Data SD8). Additionally, they strongly differ in
every quantitative and qualitative morphological metric used
in this study. This pattern is not unique to this species pair, but
rather represents the rst Sumatran example of a phenome-
non known from other Malay Archipelago Crocidura (Fig. 8). On
Mt. Gede in Java, C. monticola, a smaller shrew that is widespread
across Java, is abundant at mid-elevations while its larger-bodied
closest relative, C. umbra, is more abundant at higher elevations
(Fig. 8). On Mt. Kinabalu and neighboring Mt. Tambuyukon in
Borneo, as elevation increases, C. foetida sensu lato (Hinckley et al.
2022) gradually decreases in abundance, and its larger-bodied sis-
ter taxon C. baluensis (Fig. 8) increases in abundance (Hinckley et
al. 2022). Each of the 3 larger, high-elevation shrews are thought
to be endemic to a single massif. Notably, phenotypic and genetic
differences between C. dewi and C. balingka are much greater than
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 15
differences between the other sister pairs (Tables 1 and 2; Fig.
8; Demos et al. 2016b; Hinckley et al. 2022), and the elevational
overlap between these 2 species may be greater than in the other
examples, though sampling is limited. The causal mechanisms
leading to this pattern of elevational, morphological, and molecu-
lar divergences are unknown. Demos et al. (2016a, 2016b) consid-
ered ecological speciation along an elevational gradient followed
by subsequent niche separation as a possible mechanism for the
distribution of C. monticola and C. umbra on Mt. Gede—however,
testing this hypothesis or alternatives such as vicariance via
high-elevation refugia during the Pleistocene requires a better
understanding of the historical biogeography, demography, and
ecology of these Sunda shrews.
Though it is the sixth largest island in the world and has a
human population near 60 million, Sumatran small mammals
and Sumatran vertebrates in general (Arin et al. 2022) are
underrepresented in museum collections and, as a result, under-
studied. The last taxonomic revision of Sumatran Crocidura, using
all available material, was based on 42 specimens, 30 of which
the author collected a few years prior to the publication (Ruedi
1995). Since this time, 4 surveys of 3 mountains in West Sumatra
have uncovered 3 new species of Crocidura, suggesting that the
shrew diversity of Sumatra may be much greater than previously
thought. Recent Sumatran phylogeographic studies suggest that
reptile, amphibian, and teleost populations are structured along
a north–south gradient, following the Barisan mountains, and
that northern, central, and southern portions of the mountain
chain house their own endemic species (Lumbangtobing 2010;
Harvey et al. 2017; Sarker et al. 2019; Shaney et al. 2020a; Arin
et al. 2022). The northernmost region of the island is especially
rugged and mountainous, a geography that often houses high
levels of biodiversity and endemism (Lumbangtobing 2010;
Harvey et al. 2017; Sarker et al. 2019; Shaney et al. 2020b), yet
lacks comprehensive small mammal surveys. Equally, many iso-
lated mountains in the south of the island remain unsurveyed
for small mammals. Recent work on lizards in the subfamily
Draconinae suggests a pattern of 4 biogeographic regions across
Sumatra (Shaney et al. 2020b)—however, within these bioge-
ographic regions, individual mountains house, or are antici-
pated to house, their own endemic species (Shaney et al. 2020a).
Though only separated by a total of 250 km, we found that just
2 of the 8 described Sumatran Crocidura species—C. neglecta and
C. paradoxura—were found in all 3 montane sites discussed here,
and all 3 of the newly described species appear to be endemic to
1 mountain, supporting the pattern suggested by Shaney et al.
(2020a). Additionally, the C. paradoxura samples from Singgalang
and Tujuh form distinct clades (Supplementary Data SD6) and
have a 4% Cytb divergence (Table 1), suggesting little maternal
gene ow between these 2 mountain populations. While limited
in geographic scale, our results support the hypothesis of sub-
stantial endemism along a north–south gradient in Sumatra.
This nding points to the need of revisiting other taxonomic
issues, such as the status of C. weberi, known from 1 specimen
from Lake Singkarak in West Sumatra (Jentink 1890), as a jun-
ior synonym of C. beccarii (Ruedi 1995). Much of the lowland for-
ests, often up to 1,500 m, has and continues to undergo extreme
anthropogenic modication from logging and conversion to
oil palm plantations. Additional eld surveys and collection of
vouchered specimens are urgently needed to fully document the
mammal diversity on this large, tropical island.
Supplementary data
Supplementary data are available at Journal of Mammalogy online.
Supplementary Data SD1.—.csv table containing all the spec-
imens, museums, catalog numbers, external measurements (in
mm, except mass in g), linear cranial measurements (in mm), lin-
ear principal components scores, dorsal geometric morphometric
principal components scores, and ventral geometric morphomet-
ric principal components scores.
Fig. 8. Three examples of a low to mid-elevation, medium-sized shrew overlapping with a higher elevation, larger-bodied sister taxon. Elevation
is on the x-axis, and condyloincisive length is on the y-axis. On Mt. Kinabalu in Borneo, Crocidura foetida sensu lato is present at lower elevations
and Crocidura baluensis is found at higher elevations. On Mt. Gede in Java, C. monticola is found at lower elevations and C. umbra is found at higher
elevations. In Sumatra, we hypothesize the same pattern between C. balingka (lower elevation) and C. dewi (higher elevation), despite the elevational
overlap. Elevation and condyloincisive lengths for C. foetida sensu lato and C. baluensis were taken from the supporting information of Hinckley et al.
(2022).
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
16 | Nations et al.
Supplementary Data SD2.—Figure containing the 16 dorsal
and 23 ventral 2D geometric morphometric landmarks.
Supplementary Data SD3.—.csv table containing the GenBank
accession numbers for each sequence used in the molecular
analyses; 1 mitochondrial gene and 6 nuclear exons.
Supplementary Data SD4.—Alignment of mitochondrial Cytb
sequence data used in IQTree analysis, in nexus format.
Supplementary Data SD5.—Alignment of concatenated
nuclear DNA sequence data used in concatenated IQTree anal-
ysis, in nexus format. Concatenated genes are APO, BDNF, BRCA,
GHR, PTGER, and vWF. Partitions are listed in the “sets” block at
the end of the nexus le.
Supplementary Data SD6.—Figure of Cytb phylogenetic tree
generated in IQTree. The same tree is presented in Fig. 4 but with-
out the collapsed nodes. Node labels represent UFBoot2 values.
Supplementary Data SD7.—Figure of concatenated nDNA
phylogenetic tree generated in IQTree. The same tree is presented
in Fig. 5 but without the collapsed nodes. Node labels represent
UFBoot2 values.
Supplementary Data SD8.—Figure of APO, BDNF, BRCA, GHR,
PTGER, and vWF phylogenetic trees generated in the StarBEAST
species-tree analysis. Node labels represent posterior probabilities.
Supplementary Data SD9.—Figure of the dorsal photograph of
representatives of the newly described Sumatran taxa, Crocidura
aequicauda, and C. paradoxura. (A) C. balingka LSUMZ 40239, (B) C.
barapi LSUMZ 39748, (C) C. dewi LSUMZ 40248, (D) C. aequicauda
MZB 16790, (E) C. paradoxura MZB 41100.
Acknowledgments
We are grateful to the people of Balingka and Lubuk Landua in
Sumatera Barat, Indonesia, for their logistical support and warm
hospitality. The Indonesian Ministry of Research and Technology,
the Ministry of Environment and Forestry, and the local gov-
ernments of West Sumatra provided permits. We received eld
and logistical support from Tri Haryoko, Suparno, and Nanang
Supriatna at Museum Zoologicum Bogoriense, Tengku Lidra,
Andri Saputra, and fellow students at Andalas University, along
with Subir Shakya, Matthew Brady, and Oscar Johnson at LSUMZ.
The staff at the Field Museum of Natural History, Louisiana
State University Museum of Natural Science, and University of
California Museum of Vertebrate Zoology provided access to
specimens in their care. We thank Dr. Bob Timm and Dr. Arlo
Hinckley for their detailed reviews that greatly improved this
manuscript.
Author contributions
JAN and JAE conceived the study. ASA, Apandi, AM, HH, and RDB
organized eldwork and collection permits. JAN and JAE col-
lected skull measurements from museum specimens. HH took
photographs. JAN performed analyses and wrote the rst draft.
All authors contributed to eldwork, specimen collection, data
acquisition, and edited and approved the nal version of the
manuscript.
Funding
This work was funded by US National Science Foundation (OISE-
0965856, DEB-1343517, DEB-1441634, and DEB-1754393 to JAE), a
US National Science Foundation Graduate Research Fellowship,
the American Society of Mammalogists Fellowship, and the
American Society of Mammalogists Jim Patton Award to JAN, and
the Alfred L. Gardner and Mark S. Hafner Mammalogy Fund at
LSUMZ.
Conict of interest
None declared.
Data availability
All data and code used in this study are available in a GitHub
repository available at https://github.com/jonnations/
Sumatra_Shrew_1, and are archived on Zenodo as DOI: 10.5281/
zenodo.7916697.
References
Arin U, Smart U, Husemann M, Hertwig ST, Smith EN, Iskandar
DT, Haas A. 2022. Phylogeographic inference of Sumatran
ranids bearing gastromyzophorous tadpoles with regard to the
Pleistocene drainage systems of Sundaland. Scientic Reports
12(1):1–12. https://doi.org/10.1038/s41598-022-14722-9
Baken EK, Collyer ML, Kaliontzopoulou A, Adams DC. 2021. geo-
morph v4 0 and gmShiny: enhanced analytics and a new graph-
ical interface for a comprehensive morphometric experience.
Methods in Ecology and Evolution 12(12):2355–2363. https://doi.
org/10.1111/2041-210x.13723
Bouckaert R, Vaughan TG, Barido-Sottani J, Duchêne S, Fourment
M, Gavryushkina A, Heled J, Jones G, Kühnert D, De Maio
N, et al. 2019. BEAST 25: an advanced software platform
for Bayesian evolutionary analysis. PLoS Computational
Biology 15(4):e1006650. https://doi.org/10.1371/journal.
pcbi.1006650
Bürkner P-C. 2018. Advanced Bayesian multilevel modeling with
the R package brms. The R Journal 10(1):395–411. https://doi.
org/10.32614/rj-2018-017
Demos TC, Achmadi AS, Giarla TC, Handika H, Maharadatunkamsi,
Rowe KC, Esselstyn JA. 2016a. Local endemism and
within-island diversication of shrews illustrate the impor-
tance of speciation in building Sundaland mammal diversity.
Molecular Ecology 25(20):5158–5173. https://doi.org/10.1111/
mec.13820
Demos TC, Achmadi AS, Handika H, Maharadatunkamsi, Rowe KC,
Esselstyn JA. 2016b. A new species of shrew (Soricomorpha:
Crocidura) from Java, Indonesia: possible character displace-
ment despite interspecic gene ow. Journal of Mammalogy
98(1):183–193. https://doi.org/10.1093/jmammal/gyw183
Dobson GE. 1886. Description of new species of Soricidae in the col-
lection of the Genoa Civic Museum. Annali del Museo Civico di
Storia Naturale di Genova 4:564–567.
Esselstyn JA, Achmadi AS, Handika H, Swanson MT, Giarla TC,
Rowe KC. 2021. Fourteen new, endemic species of shrew (genus
Crocidura) from Sulawesi reveal a spectacular island radiation.
Bulletin of the American Museum of Natural History 454:1–108.
https://doi.org/10.1206/0003-0090.454.1.1.
Esselstyn JA, Maharadatunkamsi KCR, Achmadi AS, Siler CD, Evans
BJ. 2013. Carving out turf in a biodiversity hotspot: multiple,
previously unrecognized shrew species co‐occur on Java Island,
Indonesia. Molecular Ecology 22(19):4972–4987. https://doi.
org/10.1111/mec.12450
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
Journal of Mammalogy, 2024, Vol, XX, Issue XX | 17
Harvey MB, Shaney K, Sidik I, Kurniawan N, Smith EN. 2017. Endemic
dragons of Sumatra’s volcanoes: new species of Dendragama
(Squamata: Agamidae) and status of Salea rosaceum Thominot.
Herpetological Monographs 31(1):69–97. https://doi.org/10.1655/
herpmonographs-d-16-00012
Heaney LR, Balete DS, Duya MRM, Duya MV, Jansa SA, Steppan SJ,
Rickart EA. 2016. Doubling diversity: a cautionary tale of pre-
viously unsuspected mammalian diversity on a tropical oce-
anic island. Frontiers of Biogeography 8(2):1–19. https://doi.
org/10.21425/F5FBG29667
Hinckley A, Camacho-Sanchez M, Ruedi M, Hawkins MT, Mullon M,
Cornellas A, Tuh Yit Yuh F, Leonard JA. 2022. Evolutionary history
of Sundaland shrews (Eulipotyphla: Soricidae: Crocidura) with
a focus on Borneo. Zoological Journal of the Linnean Society
194(2):478–501. https://doi.org/10.1093/zoolinnean/zlab045
Hinckley A, Hawkins MT, Achmadi AS, Maldonado JE, Leonard JA.
2020. Ancient divergence driven by geographic isolation and
ecological adaptation in forest dependent Sundaland tree
squirrels. Frontiers in Ecology and Evolution 8:1–18. https://doi.
org/10.3389/fevo.2020.00208
Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. 2018.
UFBoot2: improving the ultrafast bootstrap approximation.
Molecular Biology and Evolution 35(2):518–522. https://doi.
org/10.1093/molbev/msx281
Hutterer R. 2005. Homology of unicuspids and tooth nomencla-
ture in shrews. In: Merritt JF, Churcheld S, Hutterer R, Sheftel
B, editors. Advances in the biology of shrews II. New York:
International Society of Shrew Biologists; p. 379–404.
Jentink FA. 1890. Mammalia from the Malay Archipelago II. Rodentia,
Insectivora, Chiroptera. In: Weber M, editor. Zoologishe
Ergebnise einer reise Niederländisch Ost-Indien. Leiden (The
Netherlands): E.J. Brill; p. 115–130.
Kalyaanamoorthy S, Minh BQ, Wong TK, von Haeseler A, Jermiin
LS. 2017. ModelFinder: fast model selection for accurate phy-
logenetic estimates. Nature Methods 14(6):587. https://doi.
org/10.1038/nmeth.4285
Lumbantobing D. 2010. Four new species of the Rasbora trifasciata-
group (Teleostei: Cyprinidae) from Northwestern Sumatra,
Indonesia. Copeia 2010(4):644–670. https://doi.org/10.1643/
CI-09-155
Mammal Diversity Database. 2023. Mammal Diversity Database.
Version 1.11. Zenodo. https://doi.org/10.5281/zenodo.4139722.
Accessed 1 May 2023.
Nguyen LT, Schmidt HA, von Haeseler A, Minh BQ. 2015. IQ-TREE:
a fast and effective stochastic algorithm for estimating
maximum-likelihood phylogenies. Molecular Biology and
Evolution 32(1):268–274. https://doi.org/10.1093/molbev/msu300
Oliveros CH, Moyle RG. 2010. Origin and diversication of Philippine
bulbuls. Molecular Phylogenetics and Evolution 54(3):822–832.
https://doi.org/10.1016/j.ympev.2009.12.001
Paradis E, Claude J, Strimmer K. 2004. APE: analyses of phylogenet-
ics and evolution in R language. Bioinformatics 20(2):289–290.
https://doi.org/10.1093/bioinformatics/btg412
Parsons DJ, Pelletier TA, Wieringa JG, Duckett DJ, Carstens BC.
2022. Analysis of biodiversity data suggests that mam-
mal species are hidden in predictable places. Proceedings
of the National Academy of Sciences of the United States
of America 119(14):e2103400119. https://doi.org/10.1073/
pnas.2103400119
R Core Team 2022. R: a language and environment for statisti-
cal computing. Vienna (Austria): R Foundation for Statistical
Computing. https://www.R-project.org/.
Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. 2018.
Posterior summarization in Bayesian phylogenetics using Tracer
17. Systematic Biology 67(5):901–904. https://doi.org/10.1093/
sysbio/syy032
Reeder DM, Helgen KM, Wilson DE 2007. Global trends and biases in
new mammal species discoveries. Occasional Papers, Museum
of Texas Tech University 269:1–36. https://doi.org/10.5962/BHL.
TITLE.156951.
Robinson HC, Kloss CB. 1918. Results of an expedition to Korinchi
Peak, Sumatra. Part II: mammals. Journal of the Federated
Malay States Museum 8:1–80.
Rowe KC, Achmadi AS, Fabre PH, Schenk JJ, Steppan SJ, Esselstyn
JA. 2019. Oceanic islands of Wallacea as a source for dispersal
and diversication of murine rodents. Journal of Biogeography
46(12):2752–2768. https://doi.org/10.1111/jbi.13720
Ruedi M. 1995. Taxonomic revision of shrews of the genus
Crocidura from the Sunda Shelf and Sulawesi with descrip-
tion of two new species (Mammalia: Soricidae). Zoological
Journal of the Linnean Society 115(3):211–265. https://doi.
org/10.1111/j.1096-3642.1995.tb02461.x
Sarker GC, Wostl E, Thammachoti P, Sidik I, Hamidy A, Kurniawan
N, Smith EN. 2019. New species, diversity, systematics, and con-
servation assessment of the puppet toads of Sumatra (Anura:
Bufonidae: Sigalegalephrynus). Zootaxa 4679(2):365–391. https://
doi.org/10.11646/zootaxa.4679.2.9
Shaney KJ, Harvey MB, Hamidy A, Kurniawan N, Smith EN. 2020a.
Phylogeny and biogeography of Sumatra’s cloud forest liz-
ards of the genus Dendragama and status of Acanthosaura
schneideri. ZooKeys 995:127–153. https://doi.org/10.3897/
zookeys.995.49355
Shaney KJ, Maldonado J, Smart U, Thammachoti P, Fujita M, Hamidy
A, Kurniawan N, Harvey MB, Smith EN. 2020b. Phylogeography of
montane dragons could shed light on the history of forests and
diversication processes on Sumatra. Molecular Phylogenetics
and Evolution 149:106840. https://doi.org/10.1016/j.
ympev.2020.106840
Sheldon FH, Lim HC, Moyle RG. 2015. Return to the Malay
Archipelago: the biogeography of Sundaic rainforest birds.
Journal of Ornithology 156(S1):91–113. https://doi.org/10.1007/
s10336-015-1188-3
Sikes RS, and the Animal Care and Use Committee of the American
Society of Mammalogists. 2016. 2016 Guidelines of the American
Society of Mammalogists for the use of wild mammals in
research and education. Journal of Mammalogy 97(3):663–688.
https://doi.org/10.1093/jmammal/gyw078
Tamura K, Stecher G, Kumar S. 2021. MEGA11: molecular evolu-
tionary genetics analysis version 11. Molecular Biology and
Evolution 38(7):3022–3027. https://doi.org/10.1093/molbev/
msab120
Tennekes M. 2018. tmap: thematic maps in R. Journal of Statistical
Software 84(6):1–39. https://doi.org/10.18637/jss.v084.i06
Tulli MJ, Carrizo LV, Samuels JX. 2016. Morphological variation of
the forelimb and claw in Neotropical sigmodontine rodents
(Rodentia: Cricetidae). Journal of Mammalian Evolution
23(1):81–91. https://doi.org/10.1007/s10914-015-9300-2
Van de Perre F, Willig MR, Presley SJ, Mukinzi IJ-C, Gambalemoke MS,
Leirs H, Verheyen E. 2020. Functional volumes, niche packing
and species richness: biogeographic legacies in the Congo Basin.
Royal Society Open Science 7(3):191582. https://doi.org/10.1098/
rsos.191582
Wickham H, Averick M, Bryan J, Chang W, McGowan LDA, François
R, Grolemund G, Hayes A, Henry L, Hester J, et al. 2019. Welcome
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024
18 | Nations et al.
to the Tidyverse. Journal of Open Source Software 4(43):1686.
https://doi.org/10.21105/joss.01686
Yang Z. 1995. A space-time process model for the evolution of DNA
sequences. Genetics 139(2):993–1005. https://doi.org/10.1093/
genetics/139.2.993
Appendix I
Specimens examined
Crocidura aequicauda: Mt. Tujuh: MZB 16970.
Crocidura balingka: Mt. Singgalang: FMNH 212964, 212965,
212967, 212971 to 212974, 212976, LSUMZ 40214, 40234 to
40242.
Crocidura barapi: Mt. Talamau: LSUMZ 39748, 39752, 39759,
39768, 39772, 39773, 39776, 39777, 39782, 39785, 39792.
Crocidura beccarii: Mt. Singgalang: FMNH 212953, LSUMZ 40203,
40207, 40208, 40211 to 40213, 40215, 40243, 40244; Mt. Talamau:
LSUMZ 39749; Mt. Tujuh: FMNH 212819, 212821, 212823, 212828,
212830, 212831, 212841, 212843, 212850, 212851, 212887, 213411,
213412, RMNH 38350, 38351.
Crocidura dewi: Mt. Singgalang: LSUMZ 40216, 40217, 40245 to
40248.
Crocidura hutanis: Mt. Leuser: RMNH 38410; Mt. Singgalang:
FMNH 212950, LSUMZ 40204 to 40206, 40209, 40210, 39786 to
39791, 39794, 39795.
Crocidura lepidura: Kateman River, East Sumatra: USNM 123140;
Sindingin: LSUMZ 39332; Mt. Talamau: LSUMZ 39796 to 39798,
39800, 39842, 39843, 39845; Mt. Tujuh: FMNH 212856, 212861,
212866, 212868, 212873 to 212875, 213413, 213414.
Crocidura neglecta: Bukit Lawang: MVZ 192177, 192178; Mt.
Singgalang: 40218 to 40222; Mt. Talamau: LSUMZ 39801, 39804 to
39828, 39802, 39803; Mt. Tujuh: FMNH 212877, 212878.
Crocidura paradoxura: Mt. Singgalang: FMNH 212954 to 212956,
212959, 212960, LSUMZ 40223 to 40233; Mt. Talamau: LSUMZ 39830,
39833, 39835, 39836, 39840; Mt. Tujuh: FMNH 212882, 212884, 213415.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyad126/7589606 by Louisiana State University user on 12 March 2024