M A S A R Y K O V A
U N I V E R Z I T A
PŘÍRODOVĚDECKÁ FAKULTA
T a x o n o m i c k á r e v i z e k l a d u
Crocidura dolichura:
i n t e g r a c e g e n e t i c k ý c h a
m o r f o l o g i c k ý c h d a t
D i p l o m o v á p r á c e
V E N D U L A M I K E Š O V Á
V e d o u c í p r á c e : M g r . A d a m K o n e č n ý , P h . D .
Ú s t a v B o t a n i k y a Z o o l o g i e
P r o g r a m Z o o l o g i e
Brno 2026
M U N I
S C I
Bibliografický záznam
Autor:
Název práce:
Studijní program:
Vedoucí práce:
Rok:
Počet stran:
Klíčová slova:
Vendula Mikešová
Přírodovědecká fakulta
Masarykova univerzita
Ústav Botaniky a Zoologie
Taxonomická revize kladu Crocidura dolichura: integrace
genetických a morfologických dat
Zoologie
Mgr. Adam Konečný, Ph.D.
2026
68 + 19
Fylogeneze, taxonomie, geometrická morfometrie, Crocidura,
systematika, Afrika
Bibliographic record
Author: Vendula Mikešová
Faculty of Science
Masaryk University
Department of Botany and Zoology
Title of Thesis: Revision of the Crocidura dolichura clade: integrating genetic and
morphological data
Degree Programme: Zoology
Supervisor: Mgr. Adam Konečný, Ph.D.
Year: 2026
Number of Pages: 68 + 19
Keywords: Phylogeny, taxonomy, geometric morphometries, Crocidura,
systematics, Africa
Abstrakt
Bělozubky rodu Crocidura patří mezi nejméně fylogenetický a taxonomicky vyřešené skupiny
savců. Navzdory významnému pokroku spojenému s rozvojem molekulárních metod jejich
vysoká druhová bohatost, kryptická diverzita, morfologická podobnost a nedostatek
genetických dat nadále komplikují porozumění jejich biodiverzitě. V rámci afrotropických
bělozubek existuje několik druhových komplexů s nejasně vyřešenými mezi- i
vnitrodruhovými vztahy. Jedním z nich je monofyletický klad obsahující druh Crocidura
dolichura a jemu příbuzné druhy. V současné literatuře chybí práce, která by komplexně
shrnovala genetické vztahy uvnitř tohoto kladu, a přestože byla zaznamenána zajímavá
variabilita uvnitř některých druhů, nebyl tento klad dosud detailně studován. Tato práce se
zaměřuje na revizi tohoto kladu a využívá dosud nejrozsáhlejší dostupný dataset
mitochondriálního markem cytochromu b, morfologická data zahrnující externí a
kraniodentální míry a geometrickou morfometrii a data o geografickém rozšíření jedinců.
Cílem bylo objasnit fylogenetické vztahy, zhodnotit míru divergence mezi liniemi a posoudit
jejich taxonomicky status pomocí integrativního přístupu, kombinujícího genetická,
morfologická a geografická data. Výsledky ukazují čtyři jasně oddělené druhy, z nichž jeden
představuje dosud nepopsaný taxon. Na nižší úrovni bylo identifikováno devět genetických
linií. Tyto genetické vztahy byly porovnány s morfologickými znaky, přičemž byly
identifikovány rozdíly jak v externích, tak v kraniodentálních znacích a tvaru lebky. Tyto
výsledky naznačují, že východní linie Crocidura dolichura může také představovat
samostatný druh. Tato studie přináší první ucelený pohled na komplex Crocidura dolichura a
představuje důležitý krok k jeho taxonomickému vyřešení. Zároveň poukazuje na potřebu
dalšího výzkumu, zejména s využitím genomických dat, lepší dostupnosti morfologického
materiálu, barcodingu typových exemplářů a širšího vzorkování k plnému objasnění evoluční
historie této skupiny.
Abstract
White-toothed shrews of the genus Crocidura are among the least phylogenetically and
taxonomically resolved groups of mammals. Despite significant advances in molecular
methods, their high species richness, cryptic diversity, morphological similarity, and lack of
genetic data continue to complicate our understanding of their biodiversity. Within Afrotropical
Crocidura, there are several species complexes with unresolved inter- and intraspecific
relationships. One of these is the Crocidura dolichura clade. In the current literature, there is
no one comprehensive study summarizing genetic relationships within this clade, and
although there is notable variability reported within some species, the clade has not yet been
studied in detail. This study focuses on a revision of the Crocidura dolichura clade and is
based on the most extensive available dataset of the mitochondrial marker cytochrome b,
morphological data including external and craniodental measurements, geometric
morphometries, and data on the geographic distribution of individuals. The aim was to clarify
phylogenetic relationships, assess the level of divergence among lineages, and evaluate their
taxonomic status using an integrative approach, which combines genetic, morphological and
geographical data. The results reveal four clearly distinct species, one of which represents a
previously undescribed taxon. At a finer scale, nine genetic lineages were identified. These
genetic relationships were compared with morphological characters, revealing differences in
external and craniodental traits as well as skull shape. These results suggest that the eastern
lineage of Crocidura dolichura may also represent a separate species. This study provides
the first comprehensive look into the Crocidura dolichura complex and represents an
important step toward its taxonomic resolution. It also highlights the need for further research,
particularly using genomic data, improved availability of morphological material, barcoding of
type specimens, and broader sampling to fully understand the evolutionary history of this
group.
U N I
S C I
MASARYKOVA U N I V E R Z I T A
P R Í R O D O V E D E C K Á FAKULTA
KOTLÁŘSKÁ 2 , 611 37 BRNO
I C : 0 0 2 1 6 2 2 4
QIC: C Z 0 0 2 1 6 2 2 4
Z A D A N Í
D I P L O M O V É P R Á C E
Akademický rok: 2025/2026
Ústav: Ústav botaniky a zoologie
Studentka: Be. Vendula Mikešová
Program: Zoologie
Specializace: Zoologie
Ředitel ústavu PřF MU Vám ve smyslu Studijního a zkušebního fádu MU určuje diplomovou práci s nazvem:
Název práce: Taxonomická revize kladu Crocidura dolichura: integrace genetických a morfologických
dat
Název práce anglicky: Revision of the Crocidura dolichura clade: integrating genetic and morphological data
Jazyk práce: angličtina
Oficiální zadání:
The Crocidura genus is the most species-rich among mammals, yet despite ongoing research, it remains one of the
least taxonomically resolved. Due to conserved morphology, high cryptic diversity, and limited genetic data, species
delimitation is challenging, leading to species complexes with uncertain relationships. A recent phylogenetic reconstruction
of Afrotropical Crocidura shrews highlighted several such complexes in need of taxonomie revision. One such
is the Crocidura dolichura clade.
This group, distributed across Central Africa, currently comprises three described species but includes several divergent
lineages that may represent undescribed taxa. Among these are C. dolichura, which appears to consist of three
geographically and genetically distinct lineages that may themselves represent new species; C. crenata; C. grassei;
and two deeply diverged sister lineages of unknown taxonomie status.
This thesis aims to perform a comprehensive taxonomie revision of the C. dolichura clade by integrating genetic, morphological,
and geographical data. The student will reconstruct the phylogeny of this complex, compiling cytb data and
tissue samples from collaborators and museums, and mapping the distribution of all specimens. The second part of
the thesis will analyze species' external and cranial measurements and will assess skull variation through geometric
morphometries, a method that statistically quantifies shape and size differences to help distinguish closely related
species. Resolving these discrepancies will help to clarify the phylogenetic relationships, taxonomy, and geographic
distribution of Crocidura in sub-Saharan Africa, moving toward a comprehensive understanding of their diversity.
S T R A N A 1 I 2
Literatura: Bannikova, AA; Zemlemerova, ED; Lebedev, VS; Lavrenchenko, LA (2021) The phylogenetic relationships
within the Eastern Afromontane clade of Crocidura based on mitochondrial and nuclear data. MAMMALIAN BIOLOGY
101: 1005-1018.
Dianat, M; Voet, I; Ortiz, D; de Bellocq, JG; Cuypers, LN; Kryštufek, B; Bureš, M; Čížková, D; Bryjová, A; Bryja, J;
Nicolas, V; Konečný, A (2023) Cryptic diversity of Crocidura shrews in the savannahs of Eastern and Southern Africa.
MOLECULAR PHYLOGENETICS AND EVOLUTION 180: 107708.
Igbokwe, J ; Nicolas, V; Oyeyiola, A; Obadare, A; Adesina, AS; Awodiran, MO; Van Houtte, N; Fichet-Calvet, E;
Verheyen, E; Olayemi, A (2019) Molecular taxonomy of Crocidura species (Eulipotyphla: Soricidae) in a key biogeographical
region for African shrews, Nigeria. COMPTES RENDUS BIOLOGIES 342: 108-117.
Jacquet, F; Denys, C; Verheyen, E; Bryja, J ; Hutterer, R; Peterhans, JCK; Stanley, WT; Goodman, SM; Couloux, A;
Colyn, M; Nicolas, V (2015) Phylogeography and evolutionary history of the Crocidura olivieri complex (Mammalia,
Soricomorpha): from a forest origin to broad ecological expansion across Africa. BMC EVOLUTIONARY BIOLOGY
15:71.
Jacquet, F; Nicolas, V; Colyn, M; Kadjo, B; Hutterer, R; Decher, J ; Akpatou, B; Cruaud, C; Denys, C (2014) Forest
refugia and riverine barriers promote diversification in the West African pygmy shrew (Crocidura obscurior complex,
Soricomorpha). ZOOLOGICA SCRIPTA 43: 131-148.
Nicolas, V; Jacquet, F; Hutterer, R; Konečný, A; Kouassi, SK; Durnez, L; Lalis, A; Colyn, M; Denys, C (2019) Multilocus
phylogeny of the Crocidura poensis species complex (Mammalia, Eulipotyphla): Influences of the palaeoclimate on
its diversification and evolution. JOURNAL OF BIOGEOGRAPHY 46: 871-883.
Voet, I; Denys, C; Colyn, M; Lalis, A; Konečný, A; Delapré, A; Nicolas, V; Cornette, R (2022) Incongruences between
morphology and molecular phylogeny provide an insight into the diversification of the Crocidura poensis species
complex. SCIENTIFIC REPORTS 12: 10531.
Vedoucí práce: Mgr. Adam Konečný, Ph.D.
Datum zadání práce: 28.10. 2024
V Brně dne: 16. 3. 2026
Zadání bylo schváleno prostřednictvím IS MU.
Bc. Vendula Mikešová, 7.11. 2025
Mgr. Adam Konečný, Ph.D., 7. 11. 2025
Mgr. Iveta Hodová, Ph.D., 7. 11. 2025
S T R A N A 2 Z 2
Declaration
I declare that I have prepared my thesis independently under the mentorship of the thesis
supervisor and using the information sources provided in the thesis. I used Al tools for
grammar correction and assistance with R scripts.
Acknowledgements
Firstly, I would like to thank my supervisor, Adam Konečný, for his encouragement throughout
the two years, for giving me the opportunity to travel for my internships, and for providing
sequence data for the analysis.
I am especially grateful to Violaine Nicolas and her team at the Museum National
d'Histoire Naturelle, Paris, who welcomed me into the collections and provided a substantial
amount of material for the analyses. This thesis would not have been possible without her
contribution and generosity. I would also like to thank her colleagues, particularly Raphael
Cornette, for helping me in my first steps in learning geometric morphometry.
Many thanks go to Adam Ferguson at the Field Museum in Chicago for making my visit
to the collections possible, as well as to all his colleagues. I would also like to thank Terrence
Demos and Julian Kerbis for welcoming me in Chicago, for taking the time to discuss my work
with me, and for the material used in this thesis.
For providing sequence data, I would also like to thank Herwig Leirs and his team
(University of Antwerp, Belgium) and Jake Esselstyn and his team (Louisiana State
University).
For financial support, I would like to thank the Vertebrate Research Group at the
Department of Botany and Zoology, Masaryk University, Brno, as well as the Freemovers
program.
Finally, I would like to thank Ondřej Mikula for lending me photography equipment and
assisting with the implementation of the branch-cutting algorithm, and Veronika Chalupová
for her constant help with laboratory work.
Table of contents
1 Introduction 11
1.1 Aims of the thesis 14
2 Introduction to the Crocidura dolichura clade 15
2.1 Species accounts 16
2.1.1 Crocidura dolichura W. C. H. Peters, 1876 16
2.1.2 Crocidura crenata Brosset, Dubost & Heim de Balsac, 1965 17
2.1.3 Crocidura grassei Brosset, Dubost & Heim de Balsac, 1965 18
3 Materials and methods 20
3.1 Genetics 20
3.2 Morphology 22
4 Results 26
4.1 Genetics 26
4.2 Morphology 28
4.2.1 External measurements 28
4.2.2 Craniodental measurements 30
4.2.3 Geometric morphometry 34
5 Discussion 41
5.1 The Crocidura dolichura clade 41
5.2 C.df.ludia 42
5.3 C. grassei 46
5.4 C. crenata 48
5.5 C. dolichura 53
5.6 Taxonomie implications 58
6 Conclusion 60
Bibliography 62
Appendices 69
1 Introduction
Biodiversity cannot be properly studied, monitored, or protected without accurate description
and delimitation of species. Taxonomy forms the basis of all biological sciences, as correct
identification and classification of organisms is essential for understanding biodiversity,
ensuring consistency in scientific research, and reaches the fields of medicine,
conservation, agriculture, and public health (Guerra-Garcia era/., 2008). In some groups,
however, this task is particularly challenging. Shrews (Soricidae) represent one such
taxonomically difficult group, notorious for their morphological similarity, cryptic diversity,
and high species richness. Despite their ecological importance, high diversity, and threat of
extinction, small mammals, including shrews, remain understudied, compared to larger,
more charismatic animals (Verde Arregoitia, 2016). This is true particularly in tropical
regions where biodiversity is highest. These problems are especially visible in the genus
Crocidura, the most species-rich among all mammalian genera, and even more so, in the
African representatives of the genus, where numerous attempts have been made to
separate them into clusters, often referred to as species complexes (Dianat ef a/., 2023;
Voet ef a/., 2022; Nicolas ef a/., 2019). While some of these classifications are based on
modern molecular data and combine different approaches to resolve relationships within
complexes, others are outdated and rely solely on morphological data. Given the growing
number of recognised species in this genus in recent years (Voet ef a/., 2026; Craig ef a/.,
2025; Nations etal., 2024; Esselstyn etal., 2021; Konečný etal., 2020), this group will likely
continue to grow as we dig deeper to try and resolve its taxonomie issues. It has become
clear that the most effective approach to resolving these chaotic relationships is to use the
methods of integrative taxonomy, which uses multiple different approaches and
perspectives, including phylogeography, comparative morphology, population genetics,
ecology, behaviour, and others, to delimit units of diversity (Dayrat, 2005).
A recent reconstruction of the whole phylogeny of Afrotropical Crocidura revealed a
number of species complexes or clades in need of taxonomie revision (Mikešová, 2024).
One of them is the C. dolichura clade (Fig. 1). This clade, as presented within this thesis,
has not yet been described, and as C. dolichura is the oldest described species within this
group, I have named the clade accordingly. This clade forms a well-supported and
consistently recovered monophyletic group, which was recovered across multiple analyses
with varying numbers of sequences. There is very limited genetic material from this clade
11
published, and the clade has not been formally recognized or described as a unit in
literature, thus the relationships within the clade remain poorly understood. Due to this lack
of integrative studies, there are unresolved questions regarding the diversity, relationships,
and species delimitation within the clade. There is already a large amount of collected
material from this clade stored at the Museum National d'Histoire Naturelle, Paris (MNHN),
and the Field Museum of Natural History, Chicago (FMNH), and this material has not yet
been thoroughly studied. This provided a great opportunity to explore it and combine the
newly acquired morphological and molecular data to better understand the structure of this
clade and the relationships within it. In particular, this thesis addresses the following
research questions: how many lineages are present within the clade, which of these are
divergent enough to be considered separate species, how is this divergence reflected in
morphological variation, and what steps are required next to fully resolve the taxonomic
relationships within the clade.
12
1.1 Aims of the thesis
In this thesis, I aim to contribute to the understanding of species diversity of
Afrotropical Crocidura shrews, by performing a comprehensive revision of the Crocidura
dolichura clade, integrating genetic and morphological data.
Specifically, the aims of the thesis are as follows:
- to reconstruct the phylogenetic relationships within the Crocidura dolichura clade
using mitochondrial cytochrome b (cytb) sequences obtained from our collections,
museum collections, and collaborators, and to map the geographic distribution of all
analysed specimens to evaluate distribution patterns within the group
- to analyse morphological variation among lineages by examining external and cranial
measurements and by using geometric morphometries to quantify differences in skull
shape and size
In the first part of this thesis, I introduce the Crocidura dolichura clade, summarizing the
current state of knowledge in the literature and providing an overview of the species included
within it. The following section describes the acquisition of material for both genetic and
morphological analyses. The described results allow me to delineate individual lineages of
the clade, highlight the morphological and genetic differences among them, and propose
potential new species within the genus. Resolving the currently unclear taxonomic status of
this clade will contribute to a more accurate understanding of the ever-growing genus
Crocidura and its diversity.
14
2 Introduction to the Crocidura dolichura clade
Dollman (1916) first defined the C. dolichura group in his series On the African shrews
belonging to the genus Crocidura, with the main defining character being the absence of
caudal bristle hairs. Within this group, he included C. dolichura, C. maurisca, C. niobe, C.
bottegi, C. monax, C. littoralis, C. ultima, and two shrews which are now synonyms of C.
mariquensis. Similarly, Dippenaar (1980) defines the monax-dolichura complex,
characterised by low pilosity of the tail, and includes C. monax, C. ultima, C. maurisca, C.
littoralis, C. lanosa, C. kivuana, C. niobe, C. dolichura, and C. latona. This grouping has,
however, been disproven by more recent molecular works, and the low pilosity character
appears to be highly plastic within Crocidura, as not all species grouped by low tail pilosity
cluster together genetically (Stanley etai, 2015).
Following the description of C. crenata and C. grasses (Brosset et a/., 1965a, 1965b),
both noted to be morphologically similar to C. dolichura, a different grouping began to
emerge. Suggestions of this grouping have been shown in several works. While describing
the rainforest community of Belinga-Makoukou, Gabon, Brosset (1988) calls C. dolichura,
C. crenata, and C. grasses twin-species, grouping them on account of the slender body type
compared to others.
When it comes to genetically inferred relationships, C. dolichura and C. crenata have
been shown as sister species based on 16s rRNA (Querouil etai, 2006). A study based on
both mitochondrial and nuclear genes revealed a sister relationship between C. crenata and
C. grassei (Dubey ef a/., 2008). All three species were included together for the first time in
Mikešová (2024), where the close genetic relationship of all three species is confirmed.
Within the framework of my thesis, following Mikešová (2024), Crocidura dolichura
clade includes three currently recognised species: Crocidura dolichura W. C. H. Peters,
1876, Crocidura crenata Brosset, Dubost & Heim de Balsac, 1965, and Crocidura grassei
Brosset, Dubost & Heim de Balsac, 1965.
The species of this clade are distributed in Central Africa (Fig. 2). All three species
occur in tropical rainforest habitats, primarily moist lowland rainforest, riparian and marshy
areas, with C. dolichura also occurring in montane forests.
It has been thought that these animals might be quite rare or difficult to catch (Hutterer
etai, 2024). The latter is more likely, as all three species have been described as very agile
and good at jumping and leaping (Brosset, 1988). As this adaptation helps them escape
15
predators, it similarly helps them escape the traps. However, more recent sources find these
species to be less rare than previously thought. C. dolichura has been described as common
in lowland rainforests (Lasso etal., 1996), C. grassei and C. crenata were discovered to be
locally common in a Monts Doudou survey (Goodman and Hutterer, 2004). Although there
are currently only 30 published cytb sequences of species within this clade, the large number
of specimens collected by Violaine Nicolas' team from MNHN alone suggests that they might
not be as rare after all.
2.1 Species accounts
Fig. 2 Distribution of C. dolichura, C. crenata, and C. grassei based on IUCN distribution
maps, and type localities.
2.1.1 Crocidura dolichura W. C. H. Peters, 1876
The original description of Crocidura dolichura (Peters, 1876) is based on a single female
specimen from Bonjongo, Cameroon, and provides detailed information on dentition and
external morphology. However, it does not provide a specific diagnosis for the species or a
comparison with other Crocidura species.
16
C. dolichura is described as a slender-bodied shrew, with a dark chocolate-brown
dorsal pelage and grayish-brown ventral pelage. The tail is longer than head-body length,
very sparsely haired with only a few longer hairs at the base. The ears are relatively large
and appear naked, and the snout is pointed.
Dentition is characterised by a shorter secondary cusp of the upper incisor and a
comparatively large but not taller first unicuspid. The anterior cusp of the upper premolar is
well developed. The lower incisor is described as having a blunt projection along its cutting
edge, meaning it has a single denticulation. The lower premolar is unicuspid.
The total length is 143 mm, with a head-body length of 65 mm, tail length of 80 mm,
and hind foot length of 14 mm. Additional measurements include a head length of 20 mm,
ear height of 9 mm and width of 10 mm, and a distance of 16 mm from the nostril to the ear
opening. The upper tooth row measures 7.9 mm and the lower tooth row 7.2 mm.
According to the IUCN (2026), C. dolichura is distributed across Central Africa, from
Nigeria and Cameroon in the west, eastwards along the northern side of the Congo River to
the Albertine Rift (Fig. 2). Geographic variation within C. dolichura has been discussed
multiple times in the literature. Goodman and Hutterer (2004) reported that populations from
the central and eastern Congo Basin tend to have longer tails, larger skulls with a higher
braincase, and relatively smaller upper incisors compared to the typical form. Similarly,
material from the Albertine Rift region examined by Kerbis et al. (2013) shows a longer tail
and a more greyish dorsal pelage compared to western populations.
2.1.2 Crocidura crenata Brosset, Dubost & Heim de Balsac, 1965
Crocidura crenata is described based on a type specimen from Belinga, Gabon, and a
cotype from Makokou, Gabon (Brosset et al., 1965a) as a small, slender-bodied shrew,
similar in size and appearance to C. dolichura, but differing in a more uniform and darker
coloration, more pigmented and more naked extremities, ears, and tail, and longer feet by
about one quarter. The dorsal pelage is dark chocolate-brown. The ventral pelage is the
same colour, only with a lighter tint. The tail is long and sparsely haired, lacking vibrissae
and without a terminal tuft.
The skull is generally similar in size and shape to that of C. dolichura but differs in
several details: the facial region is narrower, the braincase is less oval, and the anteorbital
foramen is positioned above the first molar rather than more anteriorly.
17
Dentition is characterised by smaller teeth than in C. dolichura and more developed
cingular ridges on the unicuspids. The most diagnostic feature is visible on the lower incisor,
which bears two distinct denticulations, a character typical for Sylvisorex.
The head and body length is 71 mm, with a tail length of 85 mm in the type specimen
(70 mm in the truncated cotype), and a hind foot length of 17-17.5 mm. Cranial
measurements of the type include a condylo-incisive length of 19 mm, braincase width of
7.7 mm, maxillary width of 5.1 mm, braincase height of 4.8 mm, and an interorbital width of
4 mm. The upper tooth row measures 8.8 mm and the lower tooth row 8 mm.
According to the IUCN (2026), C. crenata is distributed primarily in western Central
Africa, occurring in parts of Cameroon, Gabon, Equatorial Guinea, and the Republic of
Congo (Fig. 2). An outlying record is visible in the eastern Democratic Republic of the
Congo.
2.1.3 Crocidura grassei Brosset, Dubost & Heim de Balsac, 1965
Crocidura grassei is described based on a type and cotype from Belinga, Gabon (Brosset
et al., 1965b). It is described as morphologically very similar to C. dolichura, but larger
overall, with a skull that can be regarded as a scaled-up version of the C. dolichura skull,
with some minor differences.
The pelage is uniformly ash-grey dorsally, with a slightly paler ventral surface and no
clear line between the two. The extremities are distinctly depigmented, with white or whitish
hands, feet, and ventral tail surface. The tail is approximately equal in length to the headbody
length, sparsely haired, and largely devoid of vibrissae, with the skin visible through
short, stiff hairs.
The skull differs in dorsal view from C. dolichura by a slightly broader parietal region
and a more regular (less sinuous) occipito-parietal suture. In ventral view there is a
difference in the shape of the foramen magnum. In lateral view, C. grassei has a less
pronounced brain case.
Regarding dentition, the upper row is described as the dolichura type, differing in a
larger premolar. In the lower row, the incisor is raised and lacks any rudimentary
denticulation, meaning that C. grassei might be the most evolved out of the naked-tail group.
Head- body length ranges from 64-76 mm, tail length between 62-75 mm, the hind foot
measures 16-18 mm. Cranial measurements include a condylo-incisive length of 24.3 mm,
a maximum braincase width of 10.2 mm, maxillary width of 7 mm, and a braincase height of
18
6 mm. The interorbital constriction measures 5.2 mm. The upper tooth row ranges from 10.2
to 10.5 mm, and the lower tooth row from 9.8 to 10.3 mm.
According to the IUCN (2026), C. grassei has a similar distribution to C. crenata,
distributed only in western Central Africa, occurring in parts of Cameroon, Gabon, Equatorial
Guinea, and the Republic of Congo (Fig. 2).
19
3 Materials and methods
3.1 Genetics
The majority of sequences were obtained from the research team of Violaine Nicolas at the
Museum National d'Histoire Naturelle in Paris (MNHN). Additional sequences were provided
by the Field Museum in Chicago (FMNH), along with contributions from other collaborators
and members of our research group. I have also downloaded all available cytb sequences
from Genbank. Additional tissue samples brought from MNHN and FMNH were sequenced
by Veronika Chalupová in our lab. The complete cytb alignment consisted of 330 sequences
(20 from Genbank, 310 unpublished, Table 1).
DNA extractions were performed using GeneJET Genomic DNA Purification kit
(Thermo Fisher Scientific) or innuPREP Forensic Kit (1ST Innuscreen GmbH). The
mitochondrial gene cytb was amplified using the primers L14723 and H15915 (Lecompte et
al., 2002) and the Multiplex P C R kit (Qiagen). The following P C R cycle parameters were
used: an initial denaturation step of 2 min at 95°C, then 35 cycles of a 30s denaturation at
95°C, a 30s annealing at 54°C, and a 60s extension at 72°C, followed by a final 10 min
extension. P C R products were purified with Exonuclease and Quick CIP (New England
Biolabs) and sent for Sanger sequencing to Eurofins (Germany).
The sequences were aligned and edited using AliView (Larsson, 2014). The Maximum
Likelihood tree was created using IQ-TREE via the W-IQ-TREE web server (Trifinopoulos
et al., 2016), with a specific partition file and ultrafast bootstrap 1000. As outgroup, I chose
two sequences of C. crossei from the close sublade IV sensu (Dubey etal., 2008). The tree
was then edited using Figtree ver 1.4.4. (Rambaut, 2018) and Inkscape. Lineages were
identified based on the topology of the tree and geography. There were nine lineages
identified, C. grassei, C. cf. ludia, three within C. crenata, and three within C. dolichura.
These lineages were then compared and analysed both genetically and morphologically.
The branchcutting algorithm (Mikula, 2018) was used on the phylogenetic tree to test the
species delimitation. This algorithm assesses the importance of each branch by virtually
removing it (reducing its length to zero) and evaluating the impact on overall pairwise
distances among tips. Branches whose removal results in a significant change are
interpreted as representing interspecific divergence.
20
Exact geographic coordinates were available for 310 of the samples, 11 were estimated
based on the specified locality, and only 9 samples have missing coordinate data. Altitudes
were estimated based on the coordinate data.
The uncorrected genetic distances (p-distances) and K2P distances with standard
errors between selected lineages were calculated using MEGA12 (Kumar et al., 2024) with
1000 bootstrap replicates.
Fig. 3 Geographic distribution of all barcoded specimens within the Crocidura
dolichura clade.
Table 1 Sample sizes for each revealed lineage of the C. dolichura clade, used for cytb
sequencing, for GMA (mandible, dorsal, ventral, lateral view), external measurements and
craniodental measurements. As sample sizes vary among individual external and
craniodental measurements, mean sample sizes are presented here.
cytb mandible dorsal ventral lateral external craniodental
c. cf. ludia A 20 2 9 8 8 10 11
c. cf. ludia B 18 0 0 0 0 0 0
c. crenata A 73 32 39 38 34 62 12
c. crenata B 48 27 27 28 24 34 5
c. crenata C 9 0 0 0 0 0 0
21
cytb mandible dorsal ventral lateral external craniodental
C. dolichura \N 101 58 62 57 51 73 26
C. dolichura E 37 7 24 22 21 29 24
C. dolichura C 9 1 1 1 1 1 1
C. grassei 15 39 49 49 45 10 9
Total 330 166 211 203 184 219 88
3.2 Morphology
Specimens for the morphological analyses were provided by the MNHN and FMNH. All
specimens (skulls) were photographed by the same person under standardized conditions
with the same equipment, using an OM System OM-5 camera with a digital 60mm macro
lens mounted on a tripod under controlled lighting conditions. For available specimens,
photographs of bodies or skins were taken to assess colour variation between lineages
using a compact digital camera under comparable lighting conditions. Because of the high
similarity between all species excluding C. grassei, only genetically identified individuals of
C. dolichura, C. crenata, and C.clludia were included in the morphological analysis. C.
grassei is distinctly larger and thus distinguishable from the others. A subset of individuals
was barcoded, photographed, and identified as C. grassei, and the holotype and paratype
were also included in the geometric morphometric analysis (GMA). The remaining
specimens were assigned to C. grassei based on their morphological similarity to these
references and were included in the GMA. I had access to only one specimen from the
lineage C. dolichura C. Skulls or bodies of individuals from lineage C. cf. ludia B and C.
crenata C were not available for analysis. Juvenile individuals, showing an unclosed
bassiocipital suture, were excluded from the analyses.
External measurements (in millimetres) were taken from field notes and material
provided by the curators of the museums. They include total length (Total), tail length (Tail),
head and body length (HB), hindfoot length (HF), ear length (Ear), and body mass (Mass,
in grams). For the FMNH material, the HB length was calculated from the total length and
tail length. Altogether, the external measurements were taken from 219 individuals (Table
1) - specific numbers for each measurement are specified in the Results part.
Fifteen craniodental measurements were taken by callipers (Fig. 4). Based on
Esselstyn et al. (2021), Demos et al. (2017) and Lavrenchenko et al. (2016), I have chosen
22
these measurements: condylo-incisive length (CIL), breadth of braincase (BBC), interorbital
width (IOW), rostral length (RL), rostral width (RW), post-palatal depth (PPD), post-palatal
length (PPL), distance from occipital condyle to glenoid (CGL), upper toothrow length (UTL),
distance from P4 to M3 (P4-M3), labial width at M2 (M2W), palatal width (PW), lower
toothrow (LTR), mandibular length (ML) and coronoid height (CH). The average number of
specimens measured for each lineage is shown in Table 1.
Only barcoded individuals were included in the external and craniodental
measurements, lineages with only a single specimen were excluded.
Differences in external and craniodental measurements among species were compared
using Welch's one-way ANOVA, which is robust to unequal variances and unequal sample
sizes. Pairwise comparisons were performed using Welch's t-tests with Holm correction for
multiple comparisons.
For digitizing the landmarks for GMA, I used the tpsDIG2 ver. 2.32 software (Rohlf,
2006). For the dorsal view, 22 fixed landmarks and 40 sliding landmarks were used, 70 fixed
landmarks on ventral view, 26 fixed landmarks on lateral view and 16 fixed and 60 sliding
landmarks on mandible. The placement of landmarks was based on Dianat etal. (2023) and
is shown in Fig. 5. Although the lateral view is not usually used in analyses of this kind, I
wanted to test the higher braincase hypothesis and larger incisor in this way. The number
of specimens used for each view for each species is specified in Table 1. The number of
mandibles is considerably lower than for other views, as mandibles are often preserved
intact, and the lingual view is not accessible in that way. The effect of sex on shape variation
was tested using Procrustes ANOVA on the ventral view of the skull. The result showed no
significant effect of sex on shape variation and consequently sexes were pooled in all
subsequent analyses.
23
All subsequent analyses were performed in R ver.4.4.2 (R Core Team 2024), using the
Geomorph package (Adams and Otarola-Castillo, 2013). To remove the effects of
differences in position, rotation, and scale, a generalized Procrustes analysis (GPA) was
performed (Rohlf and Slice, 1990; Gower, 1975; Sneath, 1967). This step removes size
variation, so all subsequent analyses of shape were conducted on size-independent
Procrustes coordinates. Shape variation was visualized using principal component analysis
(PCA) on Procrustes-aligned coordinates. Centroid size, a measure of overall skull size,
was extracted from the G P A results and compared among lineages using one-way ANOVA
followed by Tukey's post hoc tests. Canonical variate analysis (CVA) was performed on all
four views, using principal components, which explained approximately 90% of variation.
Shape changes along the principal component and canonical variate axes were visualized
using thin-plate spline (TPS) deformation grids. Linear discriminant analysis (LDA) was
performed on principal components explaining 90% of the total shape variability to assess
classification among lineages, with accuracy evaluated using leave-one-out crossvalidation.
Shape differences among lineages were tested in all views using Procrustes
ANOVA with permutations, which is a multivariate analysis of Procrustes-aligned shape
variables (Cardini, 2024). Pairwise comparisons between lineages were then tested to
assess which lineages differed significantly. Mean shapes for each lineage were computed
and visualized using wireframes to illustrate differences in skull morphology.
25
4 Results
4.1 Genetics
Based on the cytb ML tree, I have identified nine major lineages of the C. dolichura clade
(Fig. 6). C. grassei forms a single lineage, C. crenata is separated into three lineages (A, B
and C), and C. dolichura is also separated into three lineages corresponding to western (W),
eastern (E), and central (C) populations. My data reveal an entirely new and separate
mitochondrial species-level branch within this group, of unknown identification, which I have
decided to name C. cf. ludia on account of the similarities discussed later. All lineages show
very high support (>90), except the node grouping C. dolichura E and C. dolichura W (79).
When it comes to species-level relationships, C. crenata forms a sister lineage to C. grassei.
Together, they are sister to C. dolichura. C. cf. ludia is then sister to all the remaining species
within the C. dolichura clade. All of these relationships have very high support. The whole
uncollapsed tree is shown in Appendix 5. The branchcutting algorithm delimited seven
lineages within the C. dolichura clade, grouping together lineage C. crenata C and B, and
C. cf. ludia A and B.
26
C. cf. ludia B
C. cf. ludia A
C. grassei
C. crenata C
C. crenata B
C. crenata A
C. dolichura W
Fig. 6 ML tree (IQtree, 1000 bootstrap), showing nine major lineages of the C.
dolichura clade. Bootstrap values are shown above branches.
The cytb p-distances between species and lineages are summarized in Table 2 and 3.
High intraspecific divergence is evident in several species, most notably within C. dolichura,
where the distance between C. dolichura C and W reaches 12%, between E and W11.7%,
and between E and C 10.5%. The next highest intraspecific distance occurs within C. cf.
ludia, with 8.6% divergence between the two lineages. Within C. crenata, the intraspecific
distances are lower but still notable, ranging from 5.6% between B and C to 8.2% between
A and B. K2P showed slightly higher values, but did not change relative relationships among
lineages. As they did not provide any additional information, only p-distances are presented.
K2P distances can be seen in Appendix 4.
27
Table 2 P-distances averaging over all sequence pairs between species. Standard error
estimates are shown above the diagonal.
C. crenata C. grassei C. dolichura C. cf. ludia
c. crenata 0.009 0.008 0.008
c. grassei 0.150 0.009 0.009
c. dolichura 0.148 0.139 0.008
c. cf. ludia 0.131 0.141 0.132
Table 3 P-distances averaging over all sequence pairs between lineages. Standard error
estimates are shown above the diagonal.
O LU S
< CD < m O 2 2 2
,
£ £ £ 5 2
o o
d d d d d d d d d
c. cf. ludia A 0.009 0.010 0.010 0.011 0.010 0.010 0.010 0.010
c. cf. ludia B 0.086 0.009 0.010 0.010 0.010 0.010 0.010 0.010
c. crenata A 0.126 0.128 0.008 0.008 0.010 0.010 0.010 0.010
c. crenata B 0.134 0.138 0.082 0.007 0.011 0.010 0.010 0.011
c. crenata C 0.128 0.140 0.079 0.056 0.011 0.011 0.011 0.011
c. dolichura C 0.125 0.128 0.136 0.159 0.155 0.009 0.009 0.010
c. dolichura E 0.132 0.129 0.144 0.151 0.146 0.104 0.009 0.010
c. dolichura W 0.133 0.135 0.142 0.158 0.154 0.121 0.117 0.009
c. grassei 0.143 0.138 0.148 0.153 0.144 0.145 0.148 0.136
4.2 Morphology
4.2.1 External measurements
Significant differences among the six lineages with sufficient sample size were detected for
all examined measurements. Differences among lineages in external measurements are
illustrated in Fig. 7 and Table 4. Complete pairwise comparisons are provided in Appendix
1.
Ear length showed the fewest significant differences among lineages. The most
pronounced difference was observed between C. crenata A and C. crenata B (p.adj =
0.0006).
28
Head-body length indicated that C. grassei is significantly larger than the remaining
taxa in this clade (e.g., vs. C. dolichura W: p.adj = 2.27 * 10"5
). In addition, significant
intraspecific differences were detected, with C. crenata A being larger than C. crenata B
(p.adj = 0.007), and C. dolichura E larger than C. dolichura W (p.adj = 0.0004).
A similar pattern was observed for hind foot length, where the intraspecific differences
between C. crenata A and B (p.adj = 0.000682) and C. dolichura E and W (p.adj =4.1 x
10~6
) were even more pronounced.
Although a comparable trend was visible in body mass, the differences between C.
crenata A and B (p.adj = 0.152), as well as between C. dolichura E and W (p.adj = 0.246),
were not statistically significant. As in the previous measurements, C. grassei differed
markedly from the remaining taxa (e.g., vs. C. dolichura W: p.adj = 8.09 x 10~7
).
Tail length showed that C. dolichura E has a significantly longer tail than the other
lineages (e.g., vs. C. dolichura W: p.adj = 1.51 x 10~10
), whereas C. cf. ludia A has a
significantly shorter tail than the rest (e.g., vs. C. crenata A: p.adj = 2.81 x 10~9
).
The tail-to-body ratio differed significantly among most lineages, except for the
comparisons between C. crenata A and B (p.adj = 0.308) and between C. cf. ludia A and C.
grassei (p.adj = 0.766), where no significant differences were detected.
Table 4 Mean external measurements (± SE) in mm for each lineage. Tail-to-body ratio is
in %.
Measurement cf. ludia A crenata A crenata B
Ear 9.78 ± 0.28 (rv=9) 10.32 ±0.19 (n=63) 8.61 ±0.33 (n=31)
HB 73.60 ±1.20 (rv=10) 74.87 ± 0.71 (n=62) 70.54 ±1.03 (n=35)
HF 13.96 ±0.28 (rv=10) 15.91 ±0.11 (n=63) 14.98±0.19(n=35)
Mass 6.00 ± 0.51 (n= 10) 7.90 ± 0.24 (n=59) 7.04 ± 0.29 (n=33)
Tail 65.60 ±1.26 (rv=10) 85.82 ±0.74 (n=61) 82.91 ± 0.97 (n=35)
Tail/Body ratio 89.28 ±1.95 (rv=10) 115.07 ±1.18 (n=61) 118.17 ±1.79 (n=35)
Table 4 continued Mean external measurements (± SE) for each lineage Tail-to-body
ratio is in %.
Measurement dolichura E dolichura W grassei
Ear 9.31 ± 0.27 (n=29) 9.21 ±0.18 (n=70) 10.80 ±0.36 (rv=10)
HB 71.64 ±0.80 (n=28) 66.99 ± 0.67 (n=75) 86.80 ±2.04 (rv=10)
HF 14.64 ±0.24 (n=29) 13.01 ±0.13(n=75) 16.00 ±0.30 (rv=10)
Mass 6.29 ±0.17 (n=29) 5.80 ±0.20 (n=71) 11.35 ±0.45 (rv=10)
Tail 96.00 ±1.28 (n=28) 82.68 ± 0.79 (n=74) 78.00 ±1.49 (rv=10)
29
Measurement dolichura E dolichura W grassei
Tail/Body ratio 134.32 ±2.03 (n=28) 124.49 ±1.80 (n=74) 90.32 ±2.82 (n=10)
Ear HB HF
2,5
0 cf. ludiaA 0 crenataA crenata B E5 dolichura E 0 dolichura W Ö grassei
Fig. 7 Boxplots showing variation in external measurements from six lineages of the C.
dolichura clade: Ear - ear length (mm), HB - head-body length (mm), HF - hind foot length
(mm), Mass - body mass (g), Tail - tail length (mm), and Tail/Body ratio - tail-to-body ratio
(%). Numbers above boxes indicate sample sizes.
4.2.2 Craniodental measurements
Variation in craniodental measurements among lineages is shown in Fig. 8 and Table 5.
Complete pairwise comparisons are provided in Appendix 2.
The skull of C. grassei is significantly larger than the remaining lineages in all
craniodental measurements (all p.adj < 0.01), and the following section therefore focuses
on comparisons among the other species and lineages.
Breadth across the braincase (BBC) did not differ significantly within species, nor did
the width across the upper second molars (M2W), maxillary toothrow length (P4-M3), post-
30
palatal depth (PPD), post-palatal length (PPL), palatal width (PW), and rostral width (RW)
(all p.adj > 0.05).
Within C. dolichura, lineages E and W differed significantly only in a few craniodental
measurements, namely condyle to glenoid length (CGL) (p.adj = 0.027), interorbital width
(IOW) (p.adj = 0.000179), and maxillary toothrow length (LTR) (p.adj = 0.029).
C. crenata A was significantly larger than C. crenata B in condylo-incisive length (CIL)
(p.adj = 0.000213), lower toothrow length (LTR) (p.adj = 0.033), mandible length (MDL)
(p.adj = 0.025), upper toothrow length (UTR) (p.adj = 0.006), coronoid height (CRH) (p.adj
= 0.014) and rostral length (RL) (p.adj = 0.044).
C. crenata A also generally showed larger values than the other lineages (excluding C.
grassei), particularly in CIL, LTR, MDL, and UTR, where differences were consistently
significant (all p.adj < 0.05).
C. cf. ludia A showed the greatest similarity to C. dolichura W, differing significantly only
in rostral width (RW) (p.adj = 0.05).
Table 5 Mean craniodental measurements (± SE) in mm for each lineage.
Measurement cf. ludia A crenata A crenata B
BBC 8.27 ± 0.06 (n = 10) 8.06 ± 0.07 (n = 12) 8.00 ± 0.02 (n = 5)
CGL 7.73 ± 0.09 (n = 10) 8.30 ±0.10 (n = 12) 8.09 ± 0.03 (n = 5)
CIL 19.33 ± 0.15 (n = 10) 20.44 ±0.14 (n = 12) 19.46 ±0.09 (n == 5)
CRH 4.23 ± 0.03 (n = 11) 4.25 ± 0.05 (n = 12) 3.90 ± 0.06 (n = 5)
IOW 4.17 ±0.02 (n = 11) 4.19 ±0.04 (n = 12) 4.08 ± 0.06 (n = 5)
LTR 7.61 ±0.05 (n = 11) 7.97 ± 0.08 (n = 12) 7.65 ± 0.07 (n = 5)
M2W 5.25 ± 0.05 (n = 11) 5.33 ± 0.05 (n = 12) 5.28 ± 0.07 (n = 5)
MDL 11.57 ±0.09 (n = 11) 12.05 ±0.07 (n = 12) 11.48 ±0.12 (n == 5)
P4-M3 4.55 ± 0.04 (n = 11) 4.59 ± 0.04 (n = 12) 4.52 ± 0.03 (n = 5)
PPD 3.59 ± 0.03 (n = 11) 3.61 ± 0.03 (n = 12) 3.57 ± 0.06 (n = 5)
PPL 8.76 ± 0.08 (n = 9) 9.66 ± 0.09 (n = 12) 9.36 ± 0.10 (n = 5)
PW 1.98 ±0.03 (n = 11) 2.27 ± 0.03 (n = 12) 2.18 ±0.05 (n = 5)
RL 6.71 ±0.11 (n = 11) 6.98 ± 0.07 (n = 12) 6.71 ± 0.05 (n = 5)
RW 2.48 ± 0.06 (n = 11) 2.28 ± 0.04 (n = 12) 2.25 ± 0.03 (n = 5)
UTR 8.18 ±0.06 (n = 11) 8.49 ± 0.05 (n = 12) 8.20 ± 0.05 (n = 5)
31
Table 5 continued Mean craniodental measurements (± SE) for each lineage.
Measurement dolichura E dolichura W grassei
BBC 8.18 ±0.05 (n = 24) 8.13 ±0.04 (n = 27) 9.76 ±0.12 (n = 9)
CGL 8.00 ±0.05 (n = 24) 7.77 ± 0.06 (n = 27) 8.92 ± 0.06 (n = 9)
CIL 19.20 ± 0.12 (n == 24) 19.19 ±0.11 (n == 27) 23.21 ±0.18 (n == 9)
CRH 4.12 ±0.04 (n = 27) 4.25 ± 0.03 (n = 28) 5.32 ± 0.08 (n = 9)
IOW 4.30 ± 0.03 (n = 25) 4.10 ±0.03 (n = 28) 4.72 ± 0.02 (n = 9)
LTR 7.27 ± 0.04 (n = 27) 7.46 ± 0.05 (n = 28) 9.47 ± 0.07 (n = 9)
M2W 5.27 ± 0.02 (n = 25) 5.22 ± 0.04 (n = 28) 6.59 ± 0.06 (n = 9)
MDL 11.16 ±0.08 (n == 27) 11.28 ±0.07'(n == 28) 14.33 ±0.11 (n == 9)
P4-M3 4.56 ± 0.03 (n = 25) 4.58 ±0.03 (n = 28) 5.64 ± 0.06 (n = 9)
PPD 3.74 ± 0.04 (n = 25) 3.67 ± 0.03 (n = 28) 4.08 ± 0.03 (n = 9)
PPL 9.10 ±0.07 (n = 24) 8.85 ± 0.06 (n = 26) 10.44 ±0.07 (n == 9)
PW 2.11 ±0.02 (n = 25) 2.05 ± 0.02 (n = 27) 2.62 ± 0.03 (n = 9)
RL 6.42 ± 0.05 (n = 25) 6.63 ± 0.06 (n = 27) 8.36 ±0.12 (n = 9)
RW 2.28 ± 0.02 (n = 25) 2.26 ± 0.02 (n = 28) 3.04 ± 0.06 (n = 9)
UTR 7.87 ± 0.05 (n = 25) 8.03 ±0.05 (n = 28) 10.11 ±0.07 (n == 9)
32
BBC
10
10 1 2
24 27
12
CGL CIL
6.0
5.5
5.0
4.5
4.0
7.0
6 5
6.0
5.5
5.0
'15
4.2
3.9
3.0
CRH IOW
27
12
-M
28
» •
5.00-
4.75- 25
28
4.50- 11 1 2 5
4 . 2 5 - 1 1 •
4.00- • ^
3.75
9
22
20
Hi
10
g
I—•—i
12
f = 1-
LTR
12
I t
11 X 5 27
28
M2W MDL P4-M3
---a----a-
1 1 1 2
5 25 28
*****
9
m
1
12
1
27
28
PPD PPL
25
11 12 5
28 in-
9
12
I J 24
9 •
3
RL RW
25 27
* r -
f
3.5-
3.0-
2.5
2.0
* + 4
28
6.0
5.5
5.0
4.5
2.75
2.50
2.25
2.00
1.75
11
10
0
11
12 25
28
PW
12 27
UTR
4»
11
12
25 28
cf. ludia A * crenata A $ crenata B 8 dolichura E * dolichura W 8 grassei
Fig. 8 Boxplots showing variation in craniodental measurements for the examined lineages
(in mm). Shown are breadth across the braincase BBC, condyle to glenoid length (CGL),
condylo-incisive length (CIL), coronoid height (CRH), interorbital width (IOW), lower
toothrow length (LTR), second upper molar width (M2W), mandible length (MDL), maxillary
33
toothrow length (P4-M3), post-palatal depth (PPD), post-palatal length (PPL), palatal width
(PW), rostral length (RL), rostral width (RW), upper toothrow length (UTR). Numbers above
boxplots indicate sample sizes.
4.2.3 Geometric morphometry
Size
Centroid size comparisons among lineages across all views revealed similar results. C.
grassei is significantly larger than all other lineages in all views (p.adj < 0.05). C. crenata A
is larger than the remaining lineages except for C. grassei (p.ajd < 0.05) in all views except
mandible, where there is not a significant difference compared to C. cf. ludia. However, this
result is influenced by the low number of photographed specimens for this lineage (two).
There were no significant differences between the other lineages (Fig. 9).
Dorsal Ventral
55
50
o
45
Lateral
39
35
33 • t
55
53
45
Mandible
32.5
30.0
27.5-
25.0
22.5
* •
4.
1
^ ^ • c f . l u d i a A c r e n a t a A I—•—I c r e n a t a B I—•—I d o l i c h u r a E d o l i c h u r a W I—•—I g r a s s e i
Fig. 9 A boxplot showing variation in centroid size among lineages.
34
Shape
The Procrustes ANOVA yielded different results across the four views. C. crenata, C.
dolichura, and C. grassei were consistently significantly different in all views (p < 0.05). C.
cf. ludia A, however, was not significantly different from C. dolichura W in all but the lateral
view (p = 0.013).
Regarding intraspecific differences, C. crenata A and B were significantly different in
the lateral and dorsal views, but not in the ventral view or mandible shape. In contrast, C.
dolichura E and Wwere significantly different in all views, although the p-value in the lateral
view was close to the threshold of significance (p = 0.041).
In dorsal view, Principal Component 1 (PC1, 65.7% of total shape variability) primarily
reflects variation in the relative proportions of the rostrum and braincase (Fig. 10). Negative
PC1 values are associated with a more elongated and slender rostrum combined with a
relatively shorter and narrower braincase. In contrast, positive PC1 values correspond to a
shorter rostrum and broader and longer braincase, and a wider interorbital region.
PC2 (21.4% of total shape variability) captures subtler variation in skull shape, primarily
related to changes in the relative length and curvature of the braincase. Negative PC2
values are associated with a broader and more elongated braincase, whereas positive
values correspond to a more compact and rounded braincase. Additional variation is visible
in the rostral region, where positive PC2 values show a slight anterior elongation.
The PCA of dorsal skull shape shows partial separation among lineages, with
considerable overlap in the central region of the morphospace. C. grassei forms the most
clearly separated cluster from the rest, primarily along PC1. C. dolichura E shifts towards
more negative values along PC2, while C. dolichura W is more broadly distributed. Within
C. crenata, lineage A shifts towards more negative values along PC1, with lineage B being
completely within the hull of A, but more central. C. cf. ludia A is completely within the hull
of C. dolichura W.
In the dorsal view, classification accuracy based on shape variables was 72.38%. C.
grassei showed the highest classification success (98.0%), followed by C. dolichura W
(83.9%) and C. crenata A (79.5%). In contrast, C. cf. ludia A was not correctly classified
(0%), while C. crenata B and C. dolichura E showed lower classification success (33.3%
and 50%, respectively).
35
-0025 0.0(10 0 025
PC1 [65.7«.)
• cfludiaA •* crenataA a crenataB o dolichuraE • dolichuraW o grassei
Fig. 10 Principal component analysis (PCA) of Procrustes-aligned coordinates of the skull
in dorsal view, with thin plate spline deformation grids illustrating shape changes along the
axes with a magnification factor of three.
In ventral view, PC1 (86% of total shape variability) also reflects variation in the relative
proportions of the rostrum and the posterior skull region (Fig. 11). Negative PC1 values are
associated with a relatively elongated and more slender rostral region, including the area of
the upper toothrow and palate, combined with a shorter braincase. In contrast, positive PC1
values correspond to a more compacted rostral region and an elongated braincase,
specifically glenoid to condyle region.
PC2 (9.2% of total shape variability) captures overall robustness of the skull and the
teeth. Negative PC2 values are associated with a broader rostrum and wider teeth, as well
as a broad wide braincase, whereas positive values correspond to a narrower and more
elongated rostral region and more slender teeth as well as a narrow braincase.
The P C A of ventral skull shape reveals partial overlap among most lineages, with the
exception of C. grassei, which forms a clearly distinct cluster separated from all other taxa
primarily along PC1. Although there is some overlap, C. crenata and C. dolichura tend to
occupy distinct regions along PC2, with C. crenata shifted toward higher values. No clear
differentiation is observed between the two C. crenata lineages. Lineages C. dolichura E
and W show substantial overlap, although C. dolichura E extends a bit more toward higher
36
PC1 values. C. cf. ludia falls entirely within the morphospace of C. dolichura W, showing no
clear separation from this lineage.
In the ventral view, classification accuracy based on shape variables was 70.3%. C.
grassei showed the highest classification success (100%), whereas cf. ludia A and C.
crenata B showed low accuracy (12.5% and 3.6% respectively). Both C. dolichura lineages
showed high classification success (>77%), and C. crenata A reached 79%.
Fig. 11 Principal component analysis (PCA) of Procrustes-aligned coordinates of the skull
in ventral view, with thin plate spline deformation grids illustrating shape changes along the
axes with a magnification factor of three.
In lateral view, PC1 (75.2% of total shape variability) primarily reflects variation in the
height of the braincase (Fig. 12). Negative values show higher braincase and a relatively
shorter rostral length, while positive values are associated with a flatter braincase and
elongated rostrum. PC2 (8.2% of total shape variability) shows subtle differences in the
relative size of rostral length and post-palatal length.
Although there is overlap among all the lineages, there is a clear gradient visible
between the species along PC1, where C. dolichura occupies the most negative values,
followed by C. crenata and C. cf. ludia A in the central region, while C. grassei is at the most
positive end of the axis.
37
In the lateral view, classification accuracy based on shape variables was 68.85%. C.
grassei showed the highest classification success (95.6%), followed by C. dolichura W
(76.5%). In contrast, C. cf. ludia A was not correctly classified (0%), while C. crenata A, C.
crenata B, and C. dolichura E showed moderate classification success (64.7%, 54.2%, and
42.9%, respectively).
-C.C2& 0.02S
PCI (75.2%)
• cf.ludiaA * crenataA A crenataB o dolichuraE • dolichuraW O grassei
Fig. 12 Principal component analysis (PCA) of Procrustes-aligned coordinates of the skull
in lateral view, with thin plate spline deformation grids illustrating shape changes along the
axes with a magnification factor of three.
In the mandible, there is a strong overlap among all the lineages (Fig. 13). PC1 (45%
of total shape variability) primarily reflects variation in the coronoid process orientation
compared to the mandibular body. Negative PC1 values are associated with a more
posteriorly oriented coronoid process. PC2 (25.8% of total shape variability) shows more
subtle, harder to pinpoint, differences. Positive values seem to be associated with a broader
ramus.
While there is some separation among lineages, with C. grassei forming a separate
cluster from C. crenata, there is still extensive overlap. C. dolichura W occupies a broad
region of the morphospace, overlapping with all other lineages.
38
In the mandible, classification accuracy based on shape variables was 65.45%. C.
grassei and C. dolichura W showed the highest classification success (82.1% and 81.0%,
respectively). In contrast, C. cf. ludia A was not correctly classified (0%), while C. crenata
A, C. crenata B, and C. dolichura E showed moderate to low classification success (50.0%,
40.7%, and 28.6%, respectively).
•0.050 -0.025 0000 0.025
PCI (45%]
• cf.ludiaA A crenataA A crenataB o dolichuraE • dolichursW O grassei
Fig. 13 Principal component analysis (PCA) of Procrustes-aligned coordinates of the
mandible, with thin plate spline deformation grids illustrating shape changes along the axes
with a magnification factor of three.
In the CVA plots, created on principal component scores explaining approximately 90% of
the total shape variability, C. grassei forms a separate cluster in all views, except mandible,
where there are some outliers overlapping with the other species. C. crenata is well
separated from C. dolichura in all views, with a few outliers. C. dolichura E and Ware slightly
overlapping, with C. dolichura E shifted towards more negative values in all views except
lateral, where they are completely overlapping. There is no clear distinction between C.
crenata A and B. The deformations show similar results as the PCA and I did not include
them.
39
• cf.ludiaA A crenataA crenataB dolichuraE • dolichuraW grassei
Fig. 14 Canonical variate analysis (CVA) of shape variation based on principal components
explaining 90% of total variability, (a) ventral view, (b) dorsal view, (c) mandible, and (d) lateral
view. Specimens are colored by lineage.
40
5 Discussion
5.1 The Crocidura dolichura clade
Here, for the first time, I describe the Crocidura dolichura clade, based on the most
comprehensive dataset of genetic and morphological data currently available. According to
my results, this clade includes three described species: C. dolichura, C. crenata, and C.
grassei, plus a newly identified species of unknown identity, here named C. cf. ludia. These
species are characterized as small to medium-sized shrews with a long, naked tail, a high
tail-to-body ratio, and an overlapping range in the moist tropical forests of Central Africa.
Altitudes range from 40 to 700 m, and only C. dolichura can be found at higher elevations,
reaching almost 3000 m on Mt Bigugu, Rwanda. From all the collected material, it is clear
that these shrews are not rare, but rather quite abundant.
The four species show cytb genetic divergences between 13 and 15%, which is well
above the commonly used threshold of interspecific divergence among Crocidura shrews,
discussed below. All four species overlap in their distribution, further confirming the
robustness of their separation. Their clear separation is supported by several wellrecognizable
differences in external morphology. C. grassei and C. cf. ludia are well
distinguishable from the other species, as C. grassei is notably larger, while C. cf. ludia has
a shorter tail than C. crenata and C. dolichura. The latter two species are more difficult to
distinguish from one another, however, C. dolichura has lighter ventral pelage colour
compared to the dorsal colour, whereas in C. crenata the ventral pelage is the same as the
dorsal. A photographic comparison of the bodies of available lineages is shown in Appendix
3.
These species show further subdivision into lineages at a lower level. In total, there are
nine recognized lineages, with C. dolichura diverging into three lineages, C. crenata into
three lineages, and C. cf. ludia into two lineages.
If we consider genetic distances alone, all nine lineages within this clade could be
considered separate species, as examples from the literature on Crocidura report
interspecific distances on cytb ranging from 6.4 to 11.8% (Bannikova et al., 2021), 5.5 to
23.5% (Jacquet et al., 2012) or even as low as 2.6% (Voet et al., 2026). Generally,
interspecific distances in mammals range from 2.5 to 19.23% (K2P) (Bradley and Baker,
2001). My results show the lowest divergence between C. crenata B and C, with a value of
41
5.6%. This is the only divergence that does not meet the interspecific criteria of all the
mentioned studies. The divergences among the remaining lineages range from 7.9% to
15.9% and could all be considered interspecific if supported by additional evidence.
However, genetic distances are highly variable, especially within Crocidura, and cannot
alone serve as a reliable measure for species delimitation, as there is no universal threshold
of genetic divergence that would allow consistent separation of species across taxa.
The following sections examine each species in detail, focusing on their diagnostic
characteristics and intraspecific variation.
5.2 C. cf. ludia
The following description of this species is based on the material stored at FMNH,
representing lineage C. cf. ludia A.
This branch represents a highly diverged species of unknown identity. In Genbank, the
sequences from this lineage were identified as C. ludia. After a careful evaluation of the
specimens stored in FMNH and comparison with the original description, I have concluded
that this identification is incorrect.
In the original description of C. ludia, this species is mentioned as being related to C.
dolichura, with a shorter tail and a smaller skull (Hollister et ai, 1916). It was included in C.
dolichura by Meester and Setzer (1977), but later regarded again as a separate species by
Hutterer and Dippenaar (1987). Comparison with the original description of C. ludia reveals
both similarities and notable differences. External measurements such as ear length (9 mm
in the description vs. 9.78 ± 0.28 mm) and hind foot length (14 mm vs. 13.96 ± 0.28 mm)
are comparable. However, specimens of C. cf. ludia A show a longer tail (65.60 ± 1.26 mm
compared to 53-57 mm in the type material) and a greater head-body length (73.60 ± 1.20
mm vs. 63-67 mm). Similarly, cranial measurements are larger in C. cf. ludia A, in
comparison to the type specimen, including condylo-incisive length (19.33 ±0.15 mm vs.
18.2 mm), interorbital width (4.17 ± 0.02 mm vs. 4.0 mm), and upper tooth row length (8.18
± 0.06 mm vs. 7.8 mm), while breadth across braincase is comparable (8.27 ± 0.06 mm vs.
8.2 mm). Overall, these results suggest that C. cf. ludia A is similar to C. ludia, but differs in
its larger size and relatively longer tail. The description also mentions that C. ludia differs
from C. dolichura by its brown coloration, which is puzzling, as C. dolichura is also brown.
The known distribution of C. ludia based on the IUCN (2026) Red List map shows
substantial overlap with the distribution of C. cf. ludia (Fig. 15).
42
After careful examination of all described non-barcoded Crocidura shrews, comparing
their measurements (Happold and Happold, 2013) and distribution with those of C. cf. ludia,
no species was found to fully match this combination of characters. The two most similar
species are C. congobelgica and C. ludia (Hollister et al., 1916). However, C. congobelgica
is known only from the Democratic Republic of the Congo and is characterized by a relatively
greater maxillary width compared to other members of the naked-tailed group, which
distinguishes it from the studied specimens, as the skull of C. cf. ludia is much the same as
that of C. dolichura W. Although clear differences, described above, are present between
C. cf. ludia and C. ludia, particularly in size and tail length, no other described species shows
a closer overall similarity. Thus, based on the morphological resemblance of the analysed
specimens to the holotype of C. ludia, as well as their overlapping distribution, this lineage
is here referred to as C. cf. ludia.
Genetics and distribution
C. cf. ludia forms a sister lineage to all other species within the clade. The p-distances to
the other species within the C. dolichura clade range from 13.1% (to C. crenata) to 14.1%
(to C. grassei), indicating a high level of genetic divergence and clear separation from other
members of the clade. C. cf. ludia forms two distinct lineages separated by a p-distance of
8.6%. While this value falls within the range of interspecific variation in Crocidura shrews
(Jacquet et al., 2012), I did not have the specimens of C. cf. ludia B available for
morphological examination to support the possibility of two separate species.
C. cf. ludia A is widely distributed across Central Africa, occurring in Cameroon, the
Central African Republic, and the Democratic Republic of the Congo, with Uganda
representing the easternmost limit of its range (Fig. 15). C. cf. ludia B has been recorded
only from two localities (Inkanamongo and Obenge) in the Democratic Republic of the
Congo. As observed in other species, the locality of Inkanamongo appears to represent an
area of interest, with genetically divergent specimens recorded. One possible explanation
for this pattern is the presence of a barrier, such as the Tshuapa River. The altitude ranges
from 300 to 700 m.
43
• C. cf. India A
O C. cf. ludia B
& C. ludia type locality
C. ludia IUCN
EBQ191J n kanamofigo.CODJudiaB
1» EB0321 inkanamonga COD ludiafi
« E BO409 Irikanamango CO •_ Iijd bB
EB0223_lnkariSmongo_CODjwJi^e
n EBO21S_lnkanamon90_C0D_ludiaB
ft EB04SJnkanamwigo_COD_l udiaB
EBO500Jnkanamongo_C• DJudiaB.
™ TLL44_Obenge_CO0_lLidiaB
m\ ' 1lX*3_Oi»ft9e_CODjiidiaB
~~i— Tli.4a~Oberi9e~COD~ludiae
i« TLL152jObenge_ CODJudiaB
TLL149_Obenge_C0DJudiaB
^ EBO52Jnkanamongo_G• DJudiaB
— EBO303_lrkanarnongQ_COD_liidiaE
I EBO156.1rikanarriongaJCODJudiaB
m EB06l_liikananiorigo_COO_ljdiaB
* EB035_lnkanamongo COD_ludiaB
EB063_inhanamcingo_CODJntigB
RT4590_Bodjodo_CMR_ludiaA
- MN H22239S_Tshuapa_CDD_I udiaA
- MN H2324M_Tshuaria_CDD_lLKiiaA
FMNH219676 TshuajM CODJudiaA
Crpc>o\ira_Tsrioapa_CObjw>aA
FMNH219570_Tshuapa_COD_lud iaA
FMNH219S71 _Ts«i i»pa_COD_ludBA
FMNH219673_Tshi»apa_COD_ludiaA
FMNH2226Q9_T&hijapa_CODJudiaA
EU426&97_ludiaA
R1 »T4~ N90llO_CAFJudiaA
' "_Ngotta_CAF_ludisA
BA0611 _Ngotto_CAF_l w(iaA
FH NH222607_T5huapa_CQDJudiaA
MCO051_Bortai_CAF_lud iaA
BDA114_KpQnyo_COD_ludiaA
Jdd0567_SerYiiiki_UGA ludiaA
MWD75624_Yangarribi_CODJudiaA
IMVVQ7561D^enggrnCi.COOjwJiaA
NfW0756?3_yang3mbi_COD_ludiaA
Fig. 15 ML tree and distribution of lineages C. cf. ludia A and C. cf. ludia B. The map
also shows the IUCN distribution of C. ludia and the type locality.
Morphology
C. cf. ludia is a small-medium shrew with a mean head-body length of 73.60 ± 1.20 mm and
a tail-body ratio of 89.3%. While similar in HB length to the other species within this clade
(excluding C. grassei), it is well recognised by a shorter tail. Similarly to the rest of the
species, the tail appears naked, with a few longer bristles at the base. Dorsal pelage is
reddish brown, similar to C. crenata and the western form of C. dolichura. Ventral palage is
lighter, grayish-brown.
In the morphospace, from all views, C. cf. ludia falls within the hull of C. dolichura W.
The Procrustes ANOVA indicates that it differs from C. dolichura Wonly in the lateral view,
which also shows the clearest separation in the PCA plot. In the wireframe overlay (Fig. 16),
there is a slightly higher braincase visible in C. dolichura W as well as a shorter upper incisor.
There are no clear differences between the shape of the skull in ventral view (Fig. 17). The
same can be seen in the craniodental measurements, where C. cf. ludia does not differ from
C. dolichura W in all but one measurement (RW, p.adj = 0.05). Interestingly, the skull of C.
cf. ludia is more similar to that of C. dolichura Wthan the skull of C. dolichura E. Based on
the two photographed mandibles, there is only one denticulation on the lower incisor.
44
Fig. 16 Overlay of mean wireframe shapes of C. cf. ludia and C. dolichura W in
lateral view, illustrating differences in cranial shape and landmark positions.
dolichura W — • — cf. ludia A
Fig. 17 Overlay of mean wireframe shapes of C. cf. ludia and C. dolichura W in
ventral view, illustrating differences in cranial shape and landmark positions.
45
5.3 C. grassei
Genetics and distribution
My data only includes 15 sequences of C. grassei, noticeably less than the other species,
therefore it is difficult to assess the diversity of lineages within this species. Nevertheless, a
deeper divergence is visible between the sample from Inkanamongo and the rest (Fig. 18),
a pattern similar in the other species as well. This sample also expands the known area of
distribution of C. grassei, shown in the IUCN (2026) Red List map (Fig. 18). Otherwise, the
occurrence of C. grassei overlaps with that of C. crenata in Gabon and the Republic of
Congo (Fig. 2).
Fig. 18 ML tree and distribution of C. grassei lineage.
Morphology
C. grassei is well distinguished and recognisable within the C. dolichura clade, being of
larger size than the rest. The mean head-body length of the species is 86.80 ± 2.04 mm,
while the other lineages do not exceed 75 mm. The tail-body ratio is 90.32%. The pelage
colour is dark grayish brown, most similar to that of C. dolichura E. The ventral colour is
lighter, pale beige.
Based on my results, the close comparison to C. dolichura in the original description
(Brosset et ai, 1965b), the skull even being described as a scaled-up version, might be
overstated. In all views of the skull, except the mandible, C. grassei is well distinguishable
from the rest based on shape alone. Main differences include a relatively shorter braincase,
46
forming a more rounded shape, and a rostral part that is relatively more elongated and wider
(Fig. 20). In lateral view, the braincase appears flatter, forming a nearly horizontal profile
from its highest point to the tip of the rostrum (Fig. 19). C. grassei shows no denticulation
on the lower incisor.
grassei
Fig. 20 Overlay of mean wireframe shapes of C. dolichura\N and C. grassei in dorsal
view, illustrating differences in cranial shape and landmark positions.
/
^ y V V l Fig.
19 Overlay of mean wireframe shapes of C. dolichura W and C. grassei in lateral
view, illustrating differences in cranial shape and landmark positions.
47
5.4 C. crenata
Genetics and distribution
Three lineages of C. crenata were identified in the analysis (Fig. 21). However, the branchcutting
algorithm delimited C. crenata B and C as a single OTU. This is consistent with the
p-distance values, which show a much smaller divergence between these two lineages
(5.64%) compared with the distances between A and B (8.22%) and between A and C
(7.95%).
The distribution data did not confirm the occurrence of C. crenata in the eastern
Democratic Republic of the Congo, as shown in IUCN (2026). However, this record could
be correct, as lineage C does appear in the central part, in the Congo Basin, which was not
previously known. Lineages A and B appear to be allopatric, with C. crenata A distributed in
the southern part of Gabon, extending into the Republic of the Congo, whereas C. crenata
B occurs further north, reaching southern Cameroon and the Central African Republic (Fig.
21).
48
Fig. 21 ML tree and distribution map of C. crenata A, C. crenata B and C. crenata C
lineages.
Morphology
My data confirms most of the differences of C. crenata compared to C. dolichura, which are
well-defined in the original description (Brosset etai, 1965a). C. crenata is unifrom in colour
both in ventral and dorsal pelage, whereas C. dolichura has a lighter ventral pelage. Both
lineages of C. crenata (A and B) have a slender rostrum and overall shape of the skull, with
a visibly narrower braincase and teeth, well visible in the wireframes of mean shape in Fig.
22.
49
One difference that is visible in the lateral view of the skull is a higher braincase in C.
dolichura compared to C. crenata (Fig. 23).
— • — dolichura W — • — crenata A
Fig. 22 Overlay of mean wireframe shapes of C. dolichura W a n d C. crenata A in ventral
view, illustrating differences in cranial shape and landmark positions.
dolichura W — • — crenata A
Fig. 23 Overlay of mean wireframe shapes of C. dolichura W and C. crenata A in lateral
view, illustrating differences in cranial shape and landmark positions.
However, when individual lineages are considered separately, some of the differences
described in the original description become less consistent. The hindfoot length of C.
dolichura E is not smaller but comparable to that of C. crenata B. The tail cannot be a
distinguishing feature between C. crenata A, B and C. dolichura W, however, all three have
a shorter tail compared to C. dolichura E. In terms of dorsal coloration, C. crenata differs
50
markedly from C. dolichura E: C. dolichura E has a dark brown pelage, whereas C. crenata
has a reddish-brown coloration that is indistinguishable from C. dolichura W.
The one main distinguishing character mentioned in the description of C. crenata, which
should separate it from all other Crocidura, the two denticulations on the lower incisor, is not
exclusive to this species. According to my data, both C. dolichura E and C. dolichura W
have the same character (Fig. 24). However, this character appears to be highly variable,
most likely due to differences in tooth wear, and therefore cannot serve as a clear
distinguishing feature.
Fig. 24 Comparison of the denticulations on the lower incisor in C. dolichura and C.
crenata lineages
51
There are also significant differences within the crenata species. The two lineages (A
and B) differ in ear, HB, and HF lengths, with lineage A showing consistently greater
measurements. C. crenata A also has a longer skull, as shown by the significantly higher
values measured in CIL, UTR, RL, LTR, and MDL. However, there are no clear differences
in shape uncovered by the GMA, nor are they apparent in the wireframe comparisons,
suggesting that the observed variation between the lineages is primarily driven by size rather
than shape. This is also well visible in the centroid size comparison (Fig. 9). This size
difference could be explained by several factors. In shrews, both Bergmann's rule and its
inverse have been documented (Farkova et al., 2024; Cranbrook and Piper, 2008). In this
case, however, the altitudinal ranges of the two lineages overlap, and no clear difference is
observed. The Central African rainforest is characterized by high environmental
heterogeneity, which has been shown to promote substantial genomic turnover even over
small spatial scales (Morgan et al., 2020). There is a visible geographical separation
between the two C. crenata lineages. A comparable pattern of geographically structured
morphological differentiation has been documented in Central African forest rodents, where
populations from Gabon and Cameroon form distinct clusters associated with different
habitat types and bioclimatic zones (Mboumba, 2024; Mboumba et al., 2023). These studies
suggest the presence of historical forest refugia in south-western Cameroon, southern
Gabon, and northern Gabon, which may have resulted in persistent barriers, such as the
Ogooue and/or Ivindo Rivers, limiting post-refugial expansion (Nicolas et al., 2011). In the
case of C. crenata, a potential barrier formed by the Ogooue and Ivindo Rivers appears to
separate lineages A and B. However, this interpretation requires further testing.
52
5.5 C. dolichura
Genetics and distribution
Three lineages of C. dolichura were discovered in the analysis, corresponding to three
geographical areas at the west, east and central part of the species' distribution: C. dolichura
W, C. dolichura E, and C. dolichura C (Fig. 25). The branchcutting algorithm confirmed the
separation into the three units.
The p-distances show high values between all pairwise comparisons, ranging from
10.5% to 12%. In the inferred tree topology (Fig. 25), C. dolichura E and W form sister
lineages, with C. dolichura C as their sister lineage.
There are three well-defined geographically separated clusters visible in the distribution
of C. dolichura (Fig. 25). The western cluster (C. dolichura W) occupies Gabon, the southern
part of Cameroon, part of the Republic of the Congo, reaching its westernmost point in the
southeastern part of the Central African Republic. The eastern cluster (C. dolichura E) is
distributed mainly in and around the Albertině Rift, in Uganda, Rwanda, and Burundi. The
occurrence reaches even further to the east with one specimen collected in Mpanga Forest,
Uganda, near Lake Victoria. The westernmost points of the eastern cluster reach the central
part of the Democratic Republic of the Congo. These two clusters not only occur in separate
geographical areas, but also in quite different ecological regions: the lowland rainforest of
the Congo Basin versus the montane system of the Albertině Rift, which exhibits high
altitudinal variation (Anthony et al., 2026; Plumptre et al., 2007; White, 1983). The most
unexpected is the central cluster (C. dolichura C), which is distributed south of the Congo
River in Inkanamongo and Baliko. The occurrence of C. dolichura south of the Congo River
was not previously known. Interestingly, in contrast to C. cf. ludia, the Tshuapa River does
not appear to act as a barrier for C. dolichura, as this lineage occurs on both sides of the
river. The occurrence data of the barcoded specimens show some differences from the
known distribution of C. dolichura demonstrated in the IUCN (2026) distribution map. First,
there are no records from Nigeria. Second, the distribution south of the Congo River is not
represented in the IUCN maps. Finally, the distribution of C. dolichura extends further east
than currently shown.
53
-A, uc
1
FM
^ j " C. dolichura type locality
# C. dolichura C
O C dolichura E
O C. dolichura W
I I C. dolichura IUCN
1 oao
3 Km
. -yNh?-S57'l R.= hkr: COD riolirhuraC
jj EBOit3_irihariaiHfinaa_COD_aaiitriuraC
— -tLEBQ26_lntanBmQngD_COD dolich.irsC
"ft EBQ101 lnk_tfamang_jCOD_d_lp_fiijf_C
L31
EBEMI_inlia™mQngD_COD_dDlictiuraC
jj*E601M_lnkaiiamortgo_COO_aollaiJrtC
EBOaejnLanamongD COD_dDli_h_raC
4s
— EB0323Jn»anamor90_COD_dai|i;lnjraC
L
EB04S_lnL3namongn_COD_dDliGhiuraC
I FMWh1&SD70_KaDooc_COu_doilchijraE
FMNh20_S__. -.r.'L. COD Cio_id_r_. dul _h_r_ E
.;: i M N H ^ / L T uon&we_COD dolldiuraE
FMNH20_e.3 IIG—
bwe_COD _dcilichuraE
<] FMNH;03?03_l!(wcihe_COD_flolicriLiaE
FMNH2036_S_KiYj_COD Craid_ra_doli_hLjrB,_E
' W F M N H . 2 7 ^ K,V.. co"n Ci-oui(lvra_!Joiiehkirai_E
FMNH227E47_Kivu_CGD_Cm cid ura_dnliahu ra_E
-i FMNH203702j»mD*eJ»DjJolleri_raE;
: FMN-ia :;t::)7 Kivu COD CIO: d_r_ _j| E
+ . FMNH._7$41 Ki_u COD Crwitlurci coliclmia F
TbM FMNH2C3G05 K>j LCD C_cid_ra_dolicri_ra_E
R S 4 [ FMNH?27543_Kiw_Ct>D_Crocidi)ra_dolicliLJn!_E
' a; FMNH227_»l2_Kiv_ COD_Croddura_ctoli_riuia_E
f FMNH2275J4 K'--J r-nn Cmridnrs dr.lirhura F
FMNH22r546_KivL_CaD_Crocidjia_dDlictiuia_E
,«i BDA245_Kpc.nyc._caD_dDl«duraE
. FHN-M*41T7 *r.rr-ir LJGA _jl
_'i_'_E
J. IJGAIflO? R*er,7on IJGA (lolichnraF
FMNH323&B4_l
.'j.inc;:i_LC;A_L:rL.Lidjra_dMiCI1ura_E
L FMNH223973. Kibela UGA ctolicriuraE
FMNH2_3559JJGA_d_lietlL_r_E
FMhH?230!iQ Kihale UGA jtokjbijrsE.
1
FMMH2?396D_Kil>dlB_UGA_*t(iCI1u(aE
*5 AWF3977 BarwnH;A r.C-G S„ Kinir_i EM dr-icb iraE
1 FMNrHl_7_,fch__Kitiir.i LiN » I C
FW N H137597_Kibira_BNI_dDliEhuraE
SAD 113_GA8_dollehuraW
' GArXIKJ GAB dafchuraW
Rl4iS7_D|S_CMR_dollchLraW
CM1G794& Korup CMRjiolicriiiiHW
F Ri4379_D|a_CMR_dQlicHuiaW
R14537_Dja_CMR_doIichu iaW
R16SaSNgono_C AFd MichuraW
» RIG200_NgDt<9 CAF_d dichuraW
" BA06J
4 NrjotlO CAF_d5|ichuraW
RVr-l'- i: CAF crj chLiaW
: ••:>!,.. ••l.-irvr: CAF dftittiuri^Y
|i R1B242JJgirtii_CAF_diili_huraW
1
R19MI Nnulli; CAF qln inhnrciuV
i EaLoj-i JolifiL-aW
.o_r_A0fl?1 R.aMi.ll CAF (JolidlUfaW
.. •' BAOESS Kuollo _A- (Jolicl-.u-ii'rtRl^lJlr
Nr,r:lc CAF dr: ch. ia:V
J l : 0 R1B0BE_EalL.Lri_CAF_dolic;hjraW
it R.vr:.i„ CAF »Li,?rt
M *• BA0&13_N30ltc CAF_dkllH*ijraW
RI823e NqutlO C A F delicliurpW
].-L BAD375_Bii(jLii LAI -_.iL.iivv"
M=; Rl344.« MqrUft CAF flftlifthuiflW
5C R1S310_Ng.Hy_CAF d. I.IIUIJV^
R19»r --1.;]!^ C A F dr: r.h-..i.ivV
_:M'L'E!4_ Kurj[j UMH d"> 'I |. I CAF dr: ir.h.ji.-=',V
.,, lil'finj ULIIJ.II CAI t3 ^'n.uivV
•R-32'.7 _.^:.>u- CAF dolichuralV
f° H i e314_Hgoilo_C AF_d tHlehuraW
' :• 'I i. CAF dc: ;l.i;::V
•• H.wii:v NSOHO C M aonou-ow
R-SJ7-1 'ii :-. CAF doichjiaW
BA04D8_Nggttg CAF_dQ|ichjr8W
*JBAOC53 Ba-Gjr CAF dc-ichuri'i'.'
™ R1M33_BaOL.i CAF eonctufa*
: : :
^"•:2_T Ex\J__ LI IXG JokhurLI.^
> R17!l_3 Cdraa CCC rlr.lir|-,ira'.-v
ftl 7574_Od__l__COG_dOIICTldraW
L M G^S03_MoL_va_GAB_dDlk_riuraW
• dfrGA1174_M_Lwva_GA£__oar. GAB.rtoiichu^
»J C AK27_Labayi_COG_dalichuraW
,L, CAK?S leoa-i COr. dolichuraW
• CAK"C_ _.jr-j..i _'GG dolicl'jri.'A
1
IJD
CAK
CAK-fw Aimomr.fl,-,rto COG dri itturwA'
GAOISB_Kili GAB_dDlicriuraW
R161.if: Anrllf- OAF, rloNrj-.irV.'.
j GA2B79_MijLk_laba_ GAE_drjlichuraW
CA;iS:; Monfe^a CAR cfll-II.I.T.V
GA23SB_MOLHva_GAB_*|iCrnjiaW
GAT._2_ Kli CAR tiolichLiaW
R1 €it>64_AnSill___GAB_fldlitriLljW
CJAI712_Mod_va_GAB_iJii)i_riuraW
VW1LW4_Lek»di SAB_d_li_hufaW
i •• Ea-nhiam^ COO dr rauiaW
Fig. 25 ML tree and distribution map of C. dolichura W, C. dolichura C and C.
dolichura E lineages.
54
Morphology
The external and craniodental measurements reported in the original description of C.
dolichura correspond well to my data measured in C. dolichura W. The type specimen is
described as having a head-body length of 65 mm and a tail length of 80 mm, which are
comparable to the mean values observed in C. dolichura W (66.99 ± 0.67 mm and 82.68 ±
0.79 mm, respectively). Tail-to-body ratio in the type (approximately 123%) is also very
similar to that of C. dolichura W (124.49 ± 1.80). Ear length in this lineage (9.21 ±0.18 mm)
closely matches the value reported in the original description (9 mm), while hind foot length
(13.01 ± 0.13 mm) is slightly smaller than the reported 14 mm. Craniodental measurements
show similar results as well, with upper tooth row length 8.03 ± 0.05 mm compared to 7.9
mm in the type, and lower tooth row length 7.46 ± 0.05 mm compared to 7.2 mm.
The same cannot be said for C. dolichura E, which shows some differences in external
proportions compared to the holotype. The type specimen has a head-body length of 65 mm
and a tail length of 80 mm, whereas individuals of C. dolichura E are larger on average, with
a mean head-body length of 71.64 ± 0.80 mm and a substantially longer tail of 96.00 ± 1.28
mm. This is reflected in the tail-to-body ratio (134.32 ± 2.03) compared to approximately
123% in the type specimen. Hind foot length in C. dolichura E (14.64 ± 0.24 mm) is slightly
larger than 14 mm in the type, while ear length (9.31 ± 0.27 mm) remains comparable to
that of the type. The craniodental measurements are comparable to the type.
This, along with the distribution of C. dolichura W corresponding to the type locality of
C. dolichura, suggests that C. dolichura W is the true form of C. dolichura. C. dolichura E
has a larger HB length, longer tail, and larger tail-to-body ratio.
In contrast to the clear differences observed in external measurements, craniodental
measurements show more limited differences. Significant differences were detected in
condyloglenoid length (CGL), interorbital width (IOW), and lower tooth row length (LTR). C.
dolichura E shows higher mean values in CGL (8.00 ± 0.05 vs. 7.77 ± 0.06) and IOW (4.30
± 0.03 vs. 4.10 ± 0.03). This is well visible in the wireframes of the two lineages (Fig. 26). C.
dolichura W has a slightly greater lower tooth row length (7.46 ± 0.05 vs. 7.27 ± 0.04). As
the overall length of the skull does not differ between the two lineages, these differences
suggest that C. dolichura E has a longer posterior part of the skull (from glenoids to
condyles) and a shorter anterior part. In contrast, C. dolichura W has a longer anterior part,
reflected in the longer toothrow, and a shorter braincase. All remaining cranial
measurements did not differ significantly between the two lineages.
55
dolichura W dolichura E
Fig. 26 Overlay of mean wireframe shapes of C. dolichura Wand C. dolichura E in ventral
view, illustrating differences in cranial shape and landmark positions.
The P C A plots show similar patterns across views. In dorsal view, C. dolichura E is
shifted towards negative values along PC2, associated with a relatively elongated
braincase. In the ventral view, C. dolichura E occupies the positive values along PC1,
corresponding to a relatively elongated braincase and shorter rostral region. In the lateral
view, C. dolichura E is slightly shifted towards the negative values along PC1, associated
with a higher braincase. This difference is minimal and not clearly visible in the wireframe
overlay (Fig. 27).
In both the lateral (Fig. 27) and ventral (Fig. 26) wireframe overlays, a smaller incisor in
C. dolichura E is visible, as reported by Goodman and Hutterer (2004). However, this
difference is not reflected in the UTR measurements, where, although C. dolichura E shows
a slightly smaller mean value (7.87 ± 0.05 compared to 8.03 ± 0.05 in C. dolichura W), the
difference is not statistically significant. The higher braincase and larger skull in eastern
lineages of C. dolichura reported in the same study were not confirmed in my results.
56
— • — dolichura W — • — dolichura E
Fig. 27 Overlay of mean wireframe shapes of C. dolichura VM and C. dolichura E in lateral
view, illustrating differences in cranial shape and landmark positions.
In addition to the quantitative measurements, there were also observable differences in
pelage colouration. C. dolichura E has a dark brown pelage, while C. dolichura W has a
cinnamon brown pelage colour that is visibly lighter and more reddish. Neither lineage,
however, would I describe as grey or greyish, as reported by Hollister et al. (1916), or later
by Hutterer and Dippenaar (1987), in comparisons with C. ludia, where this trait is even
considered a distinguishing characteristic. This is even more puzzling, considering the
original description mentions chocolate brown colouration, and only the dorsal side as
greyish brown (Peters, 1876).
Fig. 28 Close up comparison of the dorsal pelage of C. dolichura E (top), and C.
dolichura W (bottom).
57
Fig. 29 Comparison of of the dorsal pelage as well as tail length of C. dolichura E (top
two), and C. dolichura W (bottom two).
5.6 Taxonomic implications
Within the Crocidura dolichura clade, I identified a minimum of two potential new species.
One of them is C. cf. ludia, similar in morphology to C. ludia, distinguished by larger body
dimensions. As no described Crocidura species matches both the morphology and
distribution of these specimens, C. cf. ludia likely represents an undescribed species.
The eastern lineage of C. dolichura differs from the typical western form in several
external characters, particularly tail length and hindfoot length, as well as in coloration.
These differences are supported by high genetic divergence, indicating a long history of
independent evolution. In addition, the two forms occupy distinct ecological regions, with C.
dolichura W occurring in the lowland forests of the Congo Basin and C. dolichura E in the
montane systems of the Albertine Rift. Such environmental contrasts have been shown to
promote diversification in African mammals and may contribute to speciation (Bowie et al.,
2023; Couvreur et al., 2021). Although craniodental differences are relatively subtle, the
combined evidence supports the recognition of C. dolichura E as a distinct species.
Although a difference in size is observed between C. crenata A and B, no clear
differences in cranial shape were detected, and the level of genetic divergence is lower than
in the other cases. These findings suggest that the observed variation might be considered
intraspecific. While the Ogooue and Ivindo Rivers could represent a potential barrier, the
58
two lineages may be in an early stage of divergence, where genetic and morphological
differentiation remains incomplete.
59
6 Conclusion
In this thesis, I introduce for the first time the Crocidura dolichura clade as a well-supported
phylogenetic unit. Using 310 newly acquired unpublished cytb sequences, I reconstructed
the phylogeny of this clade and assessed the relationships within. Owing to the large number
of specimens examined, this work provides more robust and representative species
measurements compared to previously published data, which are often based on limited
sample sizes. The coordinates of genetically confirmed specimens also provide a reliable
record of their distribution.
The results show that this clade includes four distinct species: C. dolichura, C. crenata,
C. grassei, and C. cf. ludia. The latter represents a previously unrecognized species,
different in characteristics from all other described genetically unconfirmed Crocidura
species, and thus may represent a new species to be described formally. Morphologically,
it is most similar in external and craniodental measurements to the western form of C.
dolichura, but is recognizable by a comparatively shorter tail. These four species are further
separated into nine genetic lineages.
This thesis also provides the first integrative assessment of variation within C. dolichura.
A substantial genetic divergence was observed between the eastern and western lineages,
primarily reflected in external morphology. Most notably, C. dolichura E has a longer body,
hindfoot, and tail, and a darker dorsal pelage compared to C. dolichura W. There are some
differences in the morphology of the skull as well. All these differences together support the
recognition of these lineages as separate species. There is an additional central lineage
discovered in my analysis, with substantial genetic divergence from the two others. The
morphological differences, however, could not be assessed. The distribution south of the
Congo River broadens the known distribution of this species.
While variation within C. dolichura has been discussed in the literature before, the
differences observed between the two lineages of C. crenata were unexpected. C. crenata
A shows consistently higher values in both craniodental and external measurements,
reflecting a generally larger skull and body size. Here, however, the p-distance is not as
great as within C. dolichura and most likely reflects only intraspecific variation.
Despite these findings, several aspects of the C. dolichura clade remain unresolved. A
major limitation of this study is the absence of morphological data for three lineages (C.
crenata C, C. dolichura C, and C. cf. ludia B), due to the lack of available specimens. Once
60
these specimens become available, it will be possible to develop a more comprehensive
understanding of each species, including how the lineages C. crenata C, C. dolichura C,
and C. cf. ludia B differ from other lineages within the species and what processes have
driven their diversification. Further research should also incorporate genomic data (nuclear
markers), which would allow for the use of different species delimitation methods, to
compare with the delimitation presented in this thesis.
Most importantly, unless the holotype museum specimens for cases such as C. ludia
are barcoded, there is no reliable way to move forward in identifying new lineages. All work
being done in the taxonomy of Crocidura could be undermined once these types are
genetically identified and matched to already barcoded specimens. By focusing on this issue
and addressing it first, all subsequent work will be made much easier.
61
Bibliography
Adams, D.C, Otárola-Castillo, E., 2013. geomorph: an r package for the collection and
analysis of geometric morphometric shape data. Methods in Ecology and Evolution 4,
393-399. doi: 10.1111 /2041 -21 OX. 12035
Anthony, N.M., Boupoya, C.A., Freedman, A., Hardy, O., Inogwabini, B.-l., Lewis, S.,
Lowenstein, J.H., Lezine, A.-M., Saatchi, S., Smith, T., Texier, N., Walters, G., 2026.
The Evolution of Congo Basin Ecosystem Diversity. Resilience and Sustainability in
the Congo Basin. Springer, Cham, pp. 1-40. doi: 10.1007/978-3-032-02023-9_4-1
Bannikova, A.A., Zemlemerova, E.D., Lebedev, V.S., Lavrenchenko, L.A., 2021. The
phylogenetic relationships within the Eastern Afromontane clade of Crocidura based
on mitochondrial and nuclear data. Mammalian Biology 101, 1005-1018.
doi: 10.1007/S42991 -021-00120-7
Bowie, R.C.K., Monahan, W.B., Fjeldsá, J., 2023. Climate Cycles, Habitat Stability, and
Lineage Diversification in an African Biodiversity Hotspot. Diversity 15(3), 394.
doi: 10.3390/d15030394
Bradley, R.D., Baker, R.J., 2001. A Test of the Genetic Species Concept: Cytochrome-b
Sequences and Mammals. Journal of Mammalogy 82, 960-973. doi: 10.1644/1545-
1542(2001 )082%3C0960:ATOTGS%3E2.0.CO;2
Brosset, A., 1988. Le peuplement de Mammiféres Insectivores des foréts du Nord-Est du
Gabon. Revue d'Écologie (La Terre et La Vie) 43, 2 3 ^ 6 .
doi: 10.3406/revec. 1988.5411
Brosset, A., Dubost, G., de Balsac, H., 1965a. Une nouvelle espece de Crocidura du
gabon. Mammalia29, 268-274. doi: 10.1515/mamm. 1965.29.2.268
Brosset, A., Dubost, G., de Balsac, H., 1965b. Mammiferes inedits recoltes au Gabon.
Biológia Gabonica 148-174.
Cardini, A., 2024. A practical, step-by-step, guide to taxonomic comparisons using
Procrustes geometric morphometries and user-friendly software (part A): introduction
62
and preliminary analyses. European Journal of Taxonomy 934, 1-92.
doi: 10.5852/ejt.2024.934.2527
Couvreur, T.L.P., Dauby, G., Blach-Overgaard, A., Deblauwe, V., Dessein, S., Droissart,
V., Hardy, O.J., Harris, D.J., Janssens, S.B., Ley, A.C., Mackinder, B.A., Sonké, B.,
Sosef, M.S.M., Stévart, T., Svenning, J., Wieringa, J.J., Faye, A., Missoup, A.D.,
Tolley, K.A., Nicolas, V., Ntie, S., Fluteau, F., Robin, C , Guillocheau, F., Barboni, D.,
Sepulchre, P., 2021. Tectonics, climate and the diversification of the tropical African
terrestrial flora and fauna. Biological Reviews of the Cambridge Philosophical Society
96, 16-51. doi:10.1111/brv. 12644
Craig, E.W., Bryjová, A., Bryja, J., Meheretu, Y., Lavrenchenko, L.A., Peterhans, J.C.K.,
2025. Integrative taxonomie revision of endemic dwarf shrews from the Ethiopian
Highlands. Journal of Vertebrate Biology 74, 25031.1-18. doi:10.25225/jvb.25031
Cranbrook, E.O., Piper, P.J., 2008. Post-Pleistocene evolution of Bornean shrews
Crocidura foetida (Mammalia, Soricidae). Biological Journal of the Linnean Society 94,
413-419. doi: 10.1111 /j. 1095-8312.2008.00965.x
Dayrat, B., 2005. Towards integrative taxonomy. Biological Journal of the Linnean Society
85, 407-417. doi:10.1111/j.1095-8312.2005.00503.x
Demos, T.C., Achmadi, A.S., Handika, H., Maharadatunkamsi, Rowe, K.C., Esselstyn,
J.A., 2017. A new species of shrew (Soricomorpha: Crocidura) from Java, Indonesia:
possible character displacement despite interspecific gene flow. Journal of
Mammalogy 98, 183-193. doi:10.1093/jmammal/gyw183
Dianat, M., Voet, I., Ortiz, D., Goůy de Bellocq, J., Cuypers, L.N., Kryštufek, B., Bureš, M.,
Čížková, D., Bryjová, A., Bryja, J., Nicolas, V., Konečný, A., 2023. Cryptic diversity of
Crocidura shrews in the savannahs of Eastern and Southern Africa. Molecular
Phylogenetics and Evolution 180, 107708. doi: 10.1016/j.ympev.2023.107708
Dippenaar, N.J., 1980. New species of Crocidura from Ethiopia and northern Tanzania
(Mammalia: Soricidae). Annals of the Transvaal Museum 32(6): 125-154.
doi: 10.10520/AJA00411752 181
63
Dollman J.G. 1916 XVII.—On the African shrews belonging to the genus Crocidura.—VII.
The Annals and Magazine of Natural History 8(17): 188-209.
https://doi.Org/10.1080/00222931608693767
Dubey, S., Salamin, N., Ruedi, M., Barriere, P., Colyn, M., Vogel, P., 2008. Biogeographic
origin and radiation of the Old World crocidurine shrews (Mammalia: Soricidae)
inferred from mitochondrial and nuclear genes. Molecular Phylogenetics and Evolution
48, 953-963. doi:10.1016/j.ympev.2008.07.002
Esselstyn, J.A., Achmadi, A.S., Handika, H., Swanson, M.T., Giarla, T.C., Rowe, K.C.,
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. doi:10.1206/0003-0090.454.1.1
Farková, L, Lázaro, J., Kryštufek, B., Vohralík, V., 2024. Geographical patterns in
mandible variation of the common shrew (Sorex araneus) in Europe. Biological Journal
of the Linnean Society 142, 113-120. doi:10.1093/biolinnean/blad118
Goodman, S., Hutterer, R., 2004. A Report on the Shrews (Mammalia: Soricidae) of Monts
Doudou, Gabon: Elevational Distribution and Preliminary Insights into Their Ecology.
California Academy of Sciences Memoir 28, 93-105.
Gower, J.C., 1975. Generalized Procrustes Analysis. Psychometrika 40, 33-51.
doi:10.1007/BF02291478
Guerra-Garcia, J., Espinosa, F., García-Gómez, J., 2008. Trends in taxonomy today.
Zoologica Baetica 19,15-49.
Happold, M., Happold, D., 2013. Mammals of Africa Volume IV - Hedgehogs, Shrews and
Bats. Bloomsbury Publishing, London UK.
Hollister, N., Chapin, J.P., Lang, H., 1916. Shrews collected by the Congo Expedition of
the American Museum. Bulletin of the American Museum of Natural History 35: 663-
680.
Hutterer R., Dippenaar N.J. 1987. A new species of Crocidura Wagler, 1832 (Soricidae)
from Zambia. Bonner zoologische Beiträge 38: 1-7.
64
Hutterer, R., Jocque, M., Mertens, J., Mpongo, D.M., Mbende, M., 2024. Three small
species of shrews (Soricidae) from the Salonga National Park, Democratic Republic of
the Congo, representing two new taxa and a new record. Mammalia 88, 601-607.
doi: 10.1515/mammalia-2024-0036
IUCN, 2026. . IUCN Red List of Threatened Species. Retrieved on 19 April 2026 from
https://www. iucnredlist.org/en
Jacquet, F., Nicolas, V., Bonillo, C , Cruaud, C , Denys, C , 2012. Barcoding, molecular
taxonomy, and exploration of the diversity of shrews (Soricomorpha: Soricidae) on
Mount Nimba (Guinea). Zoological Journal of the Linnean Society 166, 672-687.
doi: 10.1111 /j. 1096-3642.2012.00856.x
Kerbis, J., Huhndorf, M., Plumptre, A., Hutterer, R., Kaleme, P., Ndara, B., 2013.
Mammals, other than bats, from the Misotshi-Kabogo highlands (eastern Democratic
Republic of Congo), with the description of two new species (Mammalia: Soricidae).
Bonn Zoological Bulletin 62, 203-219.
Konečný, A., Hutterer, R., Meheretu, Y., Bryja, J., 2020. Two new species of Crocidura
(Mammalia: Soricidae) from Ethiopia and updates on the Ethiopian shrew fauna.
Journal of Vertebrate Biology 69, 20064.1. doi: 10.25225/jvb.20064
Kumar, S., Stecher, G., Suleski, M., Sanderford, M., Sharma, S., Tamura, K, 2024.
MEGA12: Molecular Evolutionary Genetic Analysis Version 12 for Adaptive and Green
Computing. Molecular Biology and Evolution 41, msae263.
doi: 10.1093/molbev/msae263
Larsson, A., 2014. AliView: a fast and lightweight alignment viewer and editor for large
datasets. Bioinformatics 30, 3276-3278. doi:10.1093/bioinformatics/btu531
Lasso, C , Hutterer, R., Rial, A., 1996. Records of shrews (Soricidae) from Equatorial
Guinea, especially from Monte Alen National Park. Mammalia 60: 69-76.
doi: 10.1515/mamm. 1996.60.1.69
Lavrenchenko L.A., Voyta L.L. & Hutterer R. 2016. Diversity of shrews in Ethiopia, with the
description of two new species of Crocidura (Mammalia: Lipotyphla: Soricidae).
Zootaxa 4196: 38-60. doi:10.11646/zootaxa.4196.1.2
65
Lecompte, É., Granjon, L, Peterhans, J.K., Denys, C , 2002. Cytochrome b-based
phylogeny of the Praomys group (Rodentia, Murinae): a new African radiation?
Comptes Rendus Biologies 325, 827-840. doi: 10.1016/S1631-0691 (02)01488-9
Mboumba, J.F., 2024. Morphometric analysis within populations of Hylomyscus
walterverheyeni (Rodentia: Muridae) confirms complex relationships between small
rodent populations of West Central African forests. International Journal of Frontier
Research in Science 2, 001-008. doi:10.56355/ijfrr.2024.2.1.0029
Mboumba, J.F., Ntie, S., Mickala, P., 2023. Morphometric variability of Praomys misonnei
(Rodentia: Muridae) in west central African forests reveals complex relationships
between populations. World Journal of Advanced Science and Technology 3, 032-
040. doi: 10.53346/wjast.2023.3.1.0054
Heim de Balsac H. & Meester J. (1977) Order Insectivora. Part 1. Pp. 1-29. In: Meester J.
& Setzer H.W. (eds) The Mammals of Africa: An Identification Manual. Smithsonian
Institution Press, Washington.
Mikešová, V., 2024. Diverzita, distribuce a taxonomie afrotropických bělozubek (Crocidura,
Soricidae). Bachelor thesis, Masaryk University, Brno, https://is.muni.cz/auth/th/vu7g0/
Mikula, O., 2018. Cutting tree branches to pick OTUs: a novel method of provisional
species delimitation. bioRxiv. 419887. doi: 10.1101/419887
Morgan, K., Mboumba, J.-F., Ntie, S., Mickala, P., Miller, C.A., Zhen, Y., Harrigan, R.J., Le
Underwood, V., Ruegg, K., Fokam, E.B., Tasse Taboue, G.C., Sesink Clee, P.R.,
Fuller, T., Smith, T.B., Anthony, N.M., 2020. Precipitation and vegetation shape
patterns of genomic and craniometric variation in the central African rodent Praomys
misonnei. Proceedings of the Royal Society B: Biological Sciences 287, 20200449.
doi: 10.1098/rspb.2020.0449
Nations, J.A., Handika, H., Mursyid, A., Darmá Busta, R., Apandi, Achmadi, A.S.,
Esselstyn, J.A., 2024. Three new shrews (Soricidae: Crocidura) from West Sumatra,
Indonesia: elevational and morphological divergence in syntopic sister taxa. Journal of
Mammalogy 105, 372-389. doi:10.1093/jmammal/gyad126
66
Nicolas, V., Jacquet, F., Hutterer, R., Konečný, A., Kan Kouassi, S., Durnez, L, Lalis, A.,
Colyn, M., Denys, C , 2019. Multilocus phylogeny of the Crocidura poensis species
complex (Mammalia, Eulipotyphla): Influences of the palaeoclimate on its
diversification and evolution. Journal of Biogeography 46, 871-883.
doi: 10.1111 /jbi. 13534
Nicolas, V., Missoup, A.D., Denys, C , Kerbis Peterhans, J., Katuala, P., Couloux, A.,
Colyn, M., 2011. The roles of rivers and Pleistocene refugia in shaping genetic
diversity in Praomys misonnei in tropical Africa. Journal of Biogeography 38, 191-207.
doi: 10.1111 /j. 1365-2699.2010.02399.x
Peters, W.C.H., 1876. Über die von dem verstorbenen Professor Dr. Reinhold Buchbolz in
Westafrika gesammelten Säugethiere. Monatsberichte der Königlichen Preussischen
Akademie der Wissenschaften zu Berlin 1876, 469-485.
doi: 10.5281 /zenodo. 18026505
Plumptre, A.J., Davenport, T.R.B., Behangana, M., Kityo, R., Eilu, G., Ssegawa, P.,
Ewango, C , Meirte, D., Kahindo, C , Herremans, M., Peterhans, J.K, Pilgrim, J.D.,
Wilson, M., Languy, M., Moyer, D., 2007. The biodiversity of the Albertině Rift.
Biological Conservation 134, 178-194. doi:10.1016/j.biocon.2006.08.021
Querouil S., Barriere P., Colyn M., Hutterer R., Verheyen E., 2006. A molecular insight into
the systematics of African Crocidura Crocidurinae, Soricidae using 16SrRNA
sequences. In: Merritt J, Churchfield S., Hutterer R., Sheftel B. (eds.) Biology of the
Soricidae II: pp. 99-113.
R: The R Project for Statistical Computing, n.d. Retrieved on 16 March 2026 from
https://www.r-project.org/
Rambaut, A., 2018. Figtree ver 1.4.4. Institute of Evolutionary Biology, University of
Edinburgh, Edinburgh. Retrieved on 6 April 2026 from
https://www.scirp.org/reference/referencespapers?referenceid=3470267
Rohlf, F.J., 2006. tpsDig, Digitize Landmarks and Outlines, Version 2.05. Stony Brook, NY
Department of Ecology and Evolution, State University of New York. Retrieved on 16
67
March 2026 from
https://www.scirp.org/reference/referencespapers?referenceid=1491321
Rohlf, F.J., Slice, D., 1990. Extensions of the Procrustes Method for the Optimal
Superimposition of Landmarks. Systematic Biology 39, 40-59. doi: 10.2307/2992207
Sneath, P.H.A., 1967. Trend-surface analysis of transformation grids. Journal of Zoology
151, 65-122. doi:10.1111/j.1469-7998.1967.tb02866.x
Stanley, W.T., Hutterer, R., Giarla, T.C., Esselstyn, J.A., 2015. Phylogeny,
phylogeography and geographical variation in the Crocidura monax (Soricidae)
species complex from the montane islands of Tanzania, with descriptions of three new
species. Zoological Journal of the Linnean Society 174, 185-215.
doi: 10.1111 /zoj. 12230
Trifinopoulos, J., Nguyen, L.-T, von Haeseler, A., Minh, B.Q., 2016. W-IQ-TREE: a fast
online phylogenetic tool for maximum likelihood analysis. Nucleic Acids Research 44,
W232-W235. doi: 10.1093/nar/gkw256
Verde Arregoitia, L.D., 2016. Biases, gaps, and opportunities in mammalian extinction risk
research. Mammal Review 46, 17-29. doi: 10.1111/mam. 12049
Voet, I., Delapre, A., Denys, C , Lalis, A., Cornette, R., Nicolas, V., 2026. Integrative
taxonomy resolves a new West African Crocidura (Mammalia: Soricidae) species,
cryptic to Crocidura grandiceps in color, but distinct in size. European Journal of
Taxonomy 1039, 297-322. doi: 10.5852/ejt.2026.1039.3197
Voet, I., Denys, C , Colyn, M., Lalis, A., Konečný, A., Delapré, A., Nicolas, V , Cornette, R.,
2022. Incongruences between morphology and molecular phylogeny provide an insight
into the diversification of the Crocidura poensis species complex. Scientific Reports 12,
10531. doi: 10.1038/s41598-022-12615-5
White F. 1983. The Vegetation of Africa: A descriptive memoir to accompany the
UNESCO/AETFAT/UNSO vegetation map of Africa. UNESCO, Paris.
68
Appendices
Appendix 1
Pairwise comparisons of external measurements among lineages, showing sample sizes
and p-values from Welch's t-tests with Holm correction.
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
Ear cf. ludia A crenata A 9 63 0.125 0.714 ns
Ear cf. ludia A crenata B 9 31 0.011 0.1 ns
Ear cf. ludia A dolichura E 9 29 0.238 0.714 ns
Ear cf. ludia A dolichura W 9 70 0.102 0.714 ns
Ear cf. ludia A grassei 9 10 0.038 0.307 ns
Ear crenata A crenata B 63 31 4.31 e-05 0.000603 ***
Ear crenata A dolichura E 63 29 0.003 0.036
Ear crenata A dolichura W 63 70 3.00E-05 0.00045 ***
Ear crenata A grassei 63 10 0.254 0.714 ns
Ear crenata B dolichura E 31 29 0.105 0.714 ns
Ear crenata B dolichura W 31 70 0.117 0.714 ns
Ear crenata B grassei 31 10 0.000137 0.002 **
Ear dolichura E dolichura W 29 70 0.749 0.749 ns
Ear dolichura E grassei 29 10 0.003 0.036 *
Ear dolichura W grassei 70 10 0.001 0.017 *
HB cf. ludia A crenata A 10 62 0.376 0.752 ns
HB cf. ludia A crenata B 10 35 0.066 0.263 ns
HB cf. ludia A dolichura E 10 28 0.193 0.579 ns
HB cf. ludia A dolichura W 10 75 0.000223 0.002 **
HB cf. ludia A grassei 10 10 5.98e-05 0.000598 ***
HB crenata A crenata B 62 35 0.00094 0.007 **
HB crenata A dolichura E 62 28 0.004 0.021 *
HB crenata A dolichura W 62 75 3.84e-13 5.76e-12 ****
HB crenata A grassei 62 10 0.000166 0.001 **
HB crenata B dolichura E 35 28 0.402 0.752 ns
HB crenata B dolichura W 35 75 0.005 0.026 *
HB crenata B grassei 35 10 5.47e-06 7.11 e-05 ****
HB dolichura E dolichura W 28 75 3.35e-05 0.000368 ***
HB dolichura E grassei 28 10 1.72e-05 0.000206 ***
HB dolichura W grassei 75 10 1.62e-06 2.27e-05 ****
69
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
HF cf. ludia A crenata A 10 63 2.76e-05 0.000276 ***
HF cf. ludia A crenata B 10 35 0.007 0.04 *
HF cf. ludia A dolichura E 10 29 0.077 0.23 ns
HF cf. ludia A dolichura W 10 75 0.008 0.042
*
HF cf. ludia A grassei 10 10 9.29e-05 0.000743 ***
HF crenata A crenata B 63 35 7.58e-05 0.000682 ***
HF crenata A dolichura E 63 29 2.03e-05 0.000223 ***
HF crenata A dolichura W 63 75 2.9e-34 4.35e-33 ****
HF crenata A grassei 63 10 0.774 0.774 ns
HF crenata B dolichura E 35 29 0.257 0.514 ns
HF crenata B dolichura W 35 75 1.26e-12 1.76e-11 ****
HF crenata B grassei 35 10 0.01 0.042
*
HF dolichura E dolichura W 29 75 3.15e-07 4.1e-06 ****
HF dolichura E grassei 29 10 0.002 0.012 *
HF dolichura W grassei 75 10 5.32e-07 6.38e-06 ****
Mass cf. ludia A crenata A 10 59 0.005 0.035
*
Mass cf. ludia A crenata B 10 33 0.098 0.295 ns
Mass cf. ludia A dolichura E 10 29 0.602 1 ns
Mass cf. ludia A dolichura W 10 71 0.719 1 ns
Mass cf. ludia A grassei 10 10 3.76e-07 3.76e-06 ****
Mass crenata A crenata B 59 33 0.025 0.152 ns
Mass crenata A dolichura E 59 29 2.85e-07 3.7e-06 ****
Mass crenata A dolichura W 59 71 3.92e-10 5.88e-09 ****
Mass crenata A grassei 59 10 8.24e-06 7.42e-05 ****
Mass crenata B dolichura E 33 29 0.031 0.156 ns
Mass crenata B dolichura W 33 71 0.000871 0.007
**
Mass crenata B grassei 33 10 3.4e-07 3.74e-06 ****
Mass dolichura E dolichura W 29 71 0.062 0.246 ns
Mass dolichura E grassei 29 10 3.06e-07 3.7e-06 ****
Mass dolichura W grassei 71 10 5.78e-08 8.09e-07 ****
Tail dolichura W grassei 74 10 0.015 0.052 ns
Tail dolichura E dolichura W 28 74 1.08e-11 1.51e-10 ****
Tail dolichura E grassei 28 10 3.87e-09 3.48e-08 ****
Tail crenata B dolichura E 35 28 6.73e-11 8.75e-10 ****
Tail crenata B dolichura W 35 74 0.849 0.849 ns
Tail crenata B grassei 35 10 0.013 0.052 ns
Tail crenata A crenata B 61 35 0.02 0.052 ns
Tail crenata A dolichura E 61 28 1.5e-08 1.2e-07 ****
Tail crenata A dolichura W 61 74 0.004 0.022 *
70
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
Tail crenata A grassei 61 10 0.000351 0.002
**
Tail cf. ludia A crenata A 10 61 2.34e-10 2.81 e-09 ****
Tail cf. ludia A crenata B 10 35 4.5e-10 4.95e-09 ****
Tail cf. ludia A dolichura E 10 28 4.42e-16 6.63e-15 ****
Tail cf. ludia A dolichura W 10 74 1.69e-09 1.69e-08 ****
Tail cf. ludia A grassei 10 10 6.23e-06 4.36e-05 ****
Tail/Body ratio cf. ludia A crenata A 10 61 3.37e-09 3.37e-08 ****
Tail/Body ratio cf. ludia A crenata B 10 35 3.65e-11 4.75e-10 ****
Tail/Body ratio cf. ludia A dolichura E 10 28 1.16e-15 1.74e-14 ****
Tail/Body ratio cf. ludia A dolichura W 10 74 1.08e-13 1.51e-12 ****
Tail/Body ratio cf. ludia A grassei 10 10 0.766 0.766 ns
Tail/Body ratio crenata A crenata B 61 35 0.154 0.308 ns
Tail/Body ratio crenata A dolichura E 61 28 1.54e-10 1.69e-09 ****
Tail/Body ratio crenata A dolichura W 61 74 2.57e-05 0.000128
***
Tail/Body ratio crenata A grassei 61 10 2.69e-06 1.61e-05 ****
Tail/Body ratio crenata B dolichura E 35 28 1.59e-07 1.27e-06 ****
Tail/Body ratio crenata B dolichura W 35 74 0.015 0.044 *
Tail/Body ratio crenata B grassei 35 10 2.05e-07 1.44e-06 ****
Tail/Body ratio dolichura E dolichura W 28 74 0.00054 0.002 **
Tail/Body ratio dolichura E grassei 28 10 9.91 e-11 1.19e-09 ****
Tail/Body ratio dolichura W grassei 74 10 8.65e-09 7.78e-08 ****
71
Appendix 2
Pairwise comparisons of craniodental measurements among lineages, showing sample
sizes and p-values from Welch's t-tests with Holm correction.
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
B B C cf. ludia A crenata A 10 12 0.035 0.248 ns
B B C cf. ludia A crenata B 10 5 0.001 0.012 *
B B C cf. ludia A dolichura E 10 24 0.231 1 ns
B B C cf. ludia A dolichura W 10 27 0.072 0.434 ns
B B C cf. ludia A grassei 10 9 9.23e-08 1.2e-06 ****
B B C crenata A crenata B 12 5 0.388 1 ns
B B C crenata A dolichura E 12 24 0.207 1 ns
B B C crenata A dolichura W 12 27 0.412 1 ns
B B C crenata A grassei 12 9 8.7e-09 1.3e-07 ****
B B C crenata B dolichura E 5 24 0.003 0.023 *
B B C crenata B dolichura W 5 27 0.007 0.056 ns
B B C crenata B grassei 5 9 2.53e-07 2.78e-06 ****
B B C dolichura E dolichura W 24 27 0.511 1 ns
B B C dolichura E grassei 24 9 7.5e-08 1.05e-06 ****
B B C dolichura W grassei 27 9 1.08e-07 1.3e-06 ****
C G L cf. ludia A crenata A 10 12 0.000351 0.003 **
C G L cf. ludia A crenata B 10 5 0.003 0.021 *
C G L cf. ludia A dolichura E 10 24 0.02 0.082 ns
C G L cf. ludia A dolichura W 10 27 0.719 0.719 ns
C G L cf. ludia A grassei 10 9 8.77e-09 1.14e-07 ****
C G L crenata A crenata B 12 5 0.06 0.18 ns
C G L crenata A dolichura E 12 24 0.014 0.069 ns
C G L crenata A dolichura W 12 27 0.000163 0.001 **
C G L crenata A grassei 12 9 3.56e-05 0.000356 ***
C G L crenata B dolichura E 5 24 0.138 0.276 ns
C G L crenata B dolichura W 5 27 2.98e-05 0.000328 ***
C G L crenata B grassei 5 9 6.4e-08 7.68e-07 ****
C G L dolichura E dolichura W 24 27 0.004 0.027 *
C G L dolichura E grassei 24 9 1.49e-10 2.09e-09 ****
C G L dolichura W grassei 27 9 8.56e-13 1.28e-11 ****
CIL cf. ludia A crenata A 10 12 2.46e-05 0.000197 ***
CIL cf. ludia A crenata B 10 5 0.459 1 ns
CIL cf. ludia A dolichura E 10 24 0.523 1 ns
72
Measurement groupl group2 n1 n2 p padj p.adj.signif
CIL cf. ludia A dolichura W 10 27 0.468 1 ns
CIL cf. ludia A grassei 10 9 1.27e-11 1.65e-10 ****
CIL crenata A crenata B 12 5 3.05e-05 0.000213 ***
CIL crenata A dolichura E 12 24 4.1e-07 3.69e-06 ****
CIL crenata A dolichura W 12 27 2.74e-07 2.74e-06 ****
CIL crenata A grassei 12 9 1.08e-09 1.19e-08 ****
CIL crenata B dolichura E 5 24 0.104 0.52 ns
CIL crenata B dolichura W 5 27 0.076 0.454 ns
CIL crenata B grassei 5 9 8.09e-10 9.71 e-09 ****
CIL dolichura E dolichura W 24 27 0.935 1 ns
CIL dolichura E grassei 24 9 3.93e-12 5.89e-11 ****
CIL dolichura W grassei 27 9 8.69e-12 1.22e-10 ****
C R H cf. ludia A crenata A 11 12 0.783 1 ns
C R H cf. ludia A crenata B 11 5 0.003 0.023 *
C R H cf. ludia A dolichura E 11 27 0.038 0.188 ns
C R H cf. ludia A dolichura W 11 28 0.616 1 ns
C R H cf. ludia A grassei 11 9 2.4e-07 2.88e-06 ****
C R H crenata A crenata B 12 5 0.001 0.014 *
C R H crenata A dolichura E 12 27 0.064 0.256 ns
C R H crenata A dolichura W 12 28 0.94 1 ns
C R H crenata A grassei 12 9 3.78e-08 4.91 e-07 ****
C R H crenata B dolichura E 5 27 0.015 0.095 ns
C R H crenata B dolichura W 5 28 0.002 0.019 *
C R H crenata B grassei 5 9 9.91 e-09 1.49e-07 ****
C R H dolichura E dolichura W 27 28 0.014 0.095 ns
C R H dolichura E grassei 27 9 1.7e-08 2.38e-07 ****
C R H dolichura W grassei 28 9 2.73e-07 3.00E-06 ****
IOW cf. ludia A crenata A 11 12 0.692 1 ns
IOW cf. ludia A crenata B 11 5 0.211 0.656 ns
IOW cf. ludia A dolichura E 11 25 0.003 0.029 *
IOW cf. ludia A dolichura W 11 28 0.047 0.281 ns
IOW cf. ludia A grassei 11 9 1.82e-12 2.55e-11 ****
IOW crenata A crenata B 12 5 0.164 0.656 ns
IOW crenata A dolichura E 12 25 0.034 0.236 ns
IOW crenata A dolichura W 12 28 0.059 0.294 ns
IOW crenata A grassei 12 9 4.96e-10 5.95e-09 ****
IOW crenata B dolichura E 5 25 0.017 0.139 ns
IOW crenata B dolichura W 5 28 0.743 1 ns
IOW crenata B grassei 5 9 0.000176 0.002 **
73
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
IOW dolichura E dolichura W 25 28 1.63e-05 0.000179 ***
IOW dolichura E grassei 25 9 7.05e-12 9.16e-11 ****
IOW dolichura W grassei 28 9 3.97e-17 5.96e-16 ****
LTR cf. ludia A crenata A 11 12 0.001 0.01 **
LTR cf. ludia A crenata B 11 5 0.701 0.701 ns
LTR cf. ludia A dolichura E 11 27 5.23e-05 0.000418 ***
LTR cf. ludia A dolichura W 11 28 0.041 0.124 ns
LTR cf. ludia A grassei 11 9 1.39e-13 1.81e-12 ****
LTR crenata A crenata B 12 5 0.008 0.033 *
LTR crenata A dolichura E 12 27 4.07e-07 4.07e-06 ****
LTR crenata A dolichura W 12 28 2.21 e-05 0.000199 ***
LTR crenata A grassei 12 9 8.81e-12 1.06e-10 ****
LTR crenata B dolichura E 5 27 0.002 0.011 *
LTR crenata B dolichura W 5 28 0.05 0.124 ns
LTR crenata B grassei 5 9 1.7e-09 1.87e-08 ****
LTR dolichura E dolichura W 27 28 0.006 0.029 *
LTR dolichura E grassei 27 9 1.09e-14 1.53e-13 ****
LTR dolichura W grassei 28 9 5.00E-15 7.5e-14 ****
M2W cf. ludia A crenata A 11 12 0.275 1 ns
M2W cf. ludia A crenata B 11 5 0.726 1 ns
M2W cf. ludia A dolichura E 11 25 0.728 1 ns
M2W cf. ludia A dolichura W 11 28 0.689 1 ns
M2W cf. ludia A grassei 11 9 1.41e-11 1.97e-10 ****
M2W crenata A crenata B 12 5 0.591 1 ns
M2W crenata A dolichura E 12 25 0.309 1 ns
M2W crenata A dolichura W 12 28 0.132 1 ns
M2W crenata A grassei 12 9 1.51e-11 1.97e-10 ****
M2W crenata B dolichura E 5 25 0.873 1 ns
M2W crenata B dolichura W 5 28 0.515 1 ns
M2W crenata B grassei 5 9 9.79e-08 1.08e-06 ****
M2W dolichura E dolichura W 25 28 0.349 1 ns
M2W dolichura E grassei 25 9 1.69e-09 2.03e-08 ****
M2W dolichura W grassei 28 9 8.48e-12 1.27e-10 ****
MDL cf. ludia A crenata A 11 12 0.000817 0.007 **
MDL cf. ludia A crenata B 11 5 0.574 0.574 ns
MDL cf. ludia A dolichura E 11 27 0.002 0.016 *
MDL cf. ludia A dolichura W 11 28 0.023 0.116 ns
MDL cf. ludia A grassei 11 9 5.87e-13 7.63e-12 ****
MDL crenata A crenata B 12 5 0.004 0.025 *
74
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
MDL crenata A dolichura E 12 27 1.88e-09 2.07e-08 ****
MDL crenata A dolichura W 12 28 3.16e-08 2.84e-07 ****
MDL crenata A grassei 12 9 2.37e-11 2.84e-10 ****
MDL crenata B dolichura E 5 27 0.048 0.19 ns
MDL crenata B dolichura W 5 28 0.184 0.552 ns
MDL crenata B grassei 5 9 6.29e-09 6.29e-08 ****
MDL dolichura E dolichura W 27 28 0.235 0.552 ns
MDL dolichura E grassei 27 9 1.79e-14 2.68e-13 ****
MDL dolichura W grassei 28 9 9.9e-14 1.39e-12 ****
P4-M3 cf. ludia A crenata A 11 12 0.501 1 ns
P4-M3 cf. ludia A crenata B 11 5 0.507 1 ns
P4-M3 cf. ludia A dolichura E 11 25 0.925 1 ns
P4-M3 cf. ludia A dolichura W 11 28 0.621 1 ns
P4-M3 cf. ludia A grassei 11 9 6.49e-10 9.09e-09 ****
P4-M3 crenata A crenata B 12 5 0.194 1 ns
P4-M3 crenata A dolichura E 12 25 0.505 1 ns
P4-M3 crenata A dolichura W 12 28 0.809 1 ns
P4-M3 crenata A grassei 12 9 5.36e-10 8.04e-09 ****
P4-M3 crenata B dolichura E 5 25 0.393 1 ns
P4-M3 crenata B dolichura W 5 28 0.23 1 ns
P4-M3 crenata B grassei 5 9 3.06e-09 3.67e-08 ****
P4-M3 dolichura E dolichura W 25 28 0.636 1 ns
P4-M3 dolichura E grassei 25 9 3.14e-09 3.67e-08 ****
P4-M3 dolichura W grassei 28 9 1.01e-09 1.31e-08 ****
PPD cf. ludia A crenata A 11 12 0.554 1 ns
PPD cf. ludia A crenata B 11 5 0.742 1 ns
PPD cf. ludia A dolichura E 11 25 0.002 0.022 *
PPD cf. ludia A dolichura W 11 28 0.052 0.395 ns
PPD cf. ludia A grassei 11 9 4.42e-09 6.63e-08 ****
PPD crenata A crenata B 12 5 0.527 1 ns
PPD crenata A dolichura E 12 25 0.007 0.065 ns
PPD crenata A dolichura W 12 28 0.156 0.78 ns
PPD crenata A grassei 12 9 7.3e-09 1.02e-07 ****
PPD crenata B dolichura E 5 25 0.049 0.395 ns
PPD crenata B dolichura W 5 28 0.203 0.812 ns
PPD crenata B grassei 5 9 0.000289 0.003 **
PPD dolichura E dolichura W 25 28 0.101 0.606 ns
PPD dolichura E grassei 25 9 3.18e-07 3.82e-06 ****
PPD dolichura W grassei 28 9 1.31e-08 1.7e-07 ****
75
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
P P L cf. ludia A crenata A 9 12 5.54e-07 6.09e-06 ****
P P L cf. ludia A crenata B 9 5 0.000964 0.007 **
P P L cf. ludia A dolichura E 9 24 0.005 0.024 *
P P L cf. ludia A dolichura W 9 26 0.374 0.374 ns
P P L cf. ludia A grassei 9 9 5.69e-11 7.4e-10 ****
P P L crenata A crenata B 12 5 0.048 0.145 ns
P P L crenata A dolichura E 12 24 7.59e-05 0.000607 ***
P P L crenata A dolichura W 12 26 2.81 e-07 3.37e-06 ****
P P L crenata A grassei 12 9 1.94e-06 1.94e-05 ****
P P L crenata B dolichura E 5 24 0.063 0.145 ns
P P L crenata B dolichura W 5 26 0.002 0.015 *
P P L crenata B grassei 5 9 1.68e-05 0.000151 ***
P P L dolichura E dolichura W 24 26 0.013 0.052 ns
P P L dolichura E grassei 24 9 1.48e-12 2.07e-11 ****
P P L dolichura W grassei 26 9 1.01e-13 1.52e-12 ****
PW cf. ludia A crenata A 11 12 3.31 e-07 3.64e-06 ****
PW cf. ludia A crenata B 11 5 0.017 0.101 ns
PW cf. ludia A dolichura E 11 25 0.001 0.01 *
PW cf. ludia A dolichura W 11 27 0.068 0.338 ns
PW cf. ludia A grassei 11 9 9.66e-12 1.35e-10 ****
PW crenata A crenata B 12 5 0.185 0.37 ns
PW crenata A dolichura E 12 25 0.000143 0.001 **
PW crenata A dolichura W 12 27 2.6e-06 2.6e-05 ****
PW crenata A grassei 12 9 7.46e-08 8.95e-07 ****
PW crenata B dolichura E 5 25 0.278 0.37 ns
PW crenata B dolichura W 5 27 0.075 0.338 ns
PW crenata B grassei 5 9 0.000239 0.002 **
PW dolichura E dolichura W 25 27 0.075 0.338 ns
PW dolichura E grassei 25 9 8.8e-11 1.14e-09 ****
PW dolichura W grassei 27 9 6.84e-12 1.03e-10 ****
RL cf. ludia A crenata A 11 12 0.044 0.176 ns
RL cf. ludia A crenata B 11 5 0.988 1 ns
RL cf. ludia A dolichura E 11 25 0.026 0.132 ns
RL cf. ludia A dolichura W 11 27 0.516 1 ns
RL cf. ludia A grassei 11 9 9.24e-09 1.39e-07 ****
RL crenata A crenata B 12 5 0.006 0.044 *
RL crenata A dolichura E 12 25 1.15e-06 1.15e-05 ****
RL crenata A dolichura W 12 27 0.000604 0.005 **
RL crenata A grassei 12 9 1.66e-07 1.83e-06 ****
76
Measurement groupl group2 n1 n2 P p.adj p.adj.signif
RL crenata B dolichura E 5 25 0.001 0.009 **
RL eren ata B dolichura W 5 27 0.304 0.912 ns
RL crenata B grassei 5 9 1.12e-07 1.34e-06 ****
RL dolichura E dolichura W 25 27 0.009 0.053 ns
RL dolichura E grassei 25 9 1.82e-08 2.37e-07 ****
RL dolichura W grassei 27 9 1.6e-08 2.24e-07 ****
RW cf. ludia A crenata A 11 12 0.014 0.096 ns
RW cf. ludia A crenata B 11 5 0.005 0.05 *
RW cf. ludia A dolichura E 11 25 0.008 0.062 ns
RW cf. ludia A dolichura W 11 28 0.005 0.05 *
RW cf. ludia A grassei 11 9 5.22e-06 5.74e-05 ****
RW crenata A crenata B 12 5 0.503 1 ns
RW crenata A dolichura E 12 25 0.892 1 ns
RW crenata A dolichura W 12 28 0.671 1 ns
RW crenata A grassei 12 9 1.18e-07 1.77e-06 ****
RW crenata B dolichura E 5 25 0.492 1 ns
RW crenata B dolichura W 5 28 0.678 1 ns
RW crenata B grassei 5 9 2.35e-07 3.29e-06 ****
RW dolichura E dolichura W 25 28 0.668 1 ns
RW dolichura E grassei 25 9 6.12e-07 7.96e-06 ****
RW dolichura W grassei 28 9 7.02e-07 8.42e-06 ****
UTR cf. ludia A crenata A 11 12 0.000889 0.005 **
UTR cf. ludia A crenata B 11 5 0.851 0.851 ns
UTR cf. ludia A dolichura E 11 25 0.000539 0.004 **
UTR cf. ludia A dolichura W 11 28 0.078 0.156 ns
UTR cf. ludia A grassei 11 9 9.37e-14 1.22e-12 ****
UTR crenata A crenata B 12 5 0.001 0.006 **
UTR crenata A dolichura E 12 25 4.17e-10 4.17e-09 ****
UTR crenata A dolichura W 12 28 3.7e-07 3.33e-06 ****
UTR crenata A grassei 12 9 2.8e-12 3.36e-11 ****
UTR crenata B dolichura E 5 25 0.000339 0.003 **
UTR crenata B dolichura W 5 28 0.038 0.115 ns
UTR crenata B grassei 5 9 3.16e-11 3.48e-10 ****
UTR dolichura E dolichura W 25 28 0.027 0.108 ns
UTR dolichura E grassei 25 9 2.84e-15 3.98e-14 ****
UTR dolichura W grassei 28 9 8.04e-16 1.21e-14 ****
77
Appendix 3
Photographed specimens of lineages a) C. dolichura W b) C. dolichura E c) C. crenata A d)
C. cf. ludia A e) C. grassei.
78
Appendix 4
K2P distances averaging over all sequence pairs between lineages. Standard error
estimates are shown above the diagonal.
C.cf./ucr/'aA
C.cf./ucŕ/'aB
C.crenataA
C.crenataB
C.crenataC
C.dolichuraC
C.dolichuraE
C.dolichuraW
C.grassei
C. cf. ludia A 0.0096 0.0124 0.0130 0.0138 0.0134 0.0146 0.0131 0.0138
C. cf. ludia B 0.0938 0.0122 0.0123 0.0143 0.0125 0.0131 0.0124 0.0122
C. crenata A 0.1415 0.1443 0.0089 0.0097 0.0133 0.0143 0.0127 0.0135
C. crenata B 0.1509 0.1569 0.0889 0.0078 0.0146 0.0143 0.0140 0.0135
C. crenata C 0.1442 0.1597 0.0859 0.0596 0.0159 0.0165 0.0153 0.0146
C. dolichura C 0.1422 0.1449 0.1545 0.1852 0.1800 0.0117 0.0113 0.0141
C. dolichura E 0.1513 0.1465 0.1657 0.1749 0.1689 0.1162 0.0120 0.0146
C. dolichura W 0.1519 0.1544 0.1623 0.1842 0.1795 0.1370 0.1321 0.0118
C. grassei 0.1633 0.1574 0.1705 0.1769 0.1652 0.1674 0.1698 0.1540
79
Appendix 5
Uncollapsed whole ML tree of C. dolichura clade.
n.ilryoLp EKPKi4 E-:p-nn.i \iriA .-rass^i
outgroiip_BE11ft4_Pobs_BEN_(
EBO191_l nkandmongo.CODJudiaB
100 EB0321_lnhanamongo_COD_ludiaB
jg EBQ4Q9_lnkanamongo COD JudiaB
g . EB0223_lnkananwf>go_COD_ludiaB
gg EB021B_lnkanamongD_COD_ludiaB
79 EB045_lnkanarriongo_COD_ludiaB
EBO500_lr>kan3mong0i_CODJudiaB
100 TLL44_Obenge_COD_ludiaB
TLL43_ObengB CODJudiaB
TLL48_Gt>enge_COajudiaB
00 TLL152_Obenge_COD_ludiaB
TLL149_Obenge_COD_ludiaB
EB052jrtkaramoiw_CODjudiaB
EBO308_lnkanarT>ongei_COD ludiaB
7 5 EB0156_lnkanamongo_COD_ludiaB
35 EBO5l_lr>kanamoiig0i_CODJudiaB
56 EB035_lnkanarnorgo_COD_ludJaB
EBQ63_lnka namongo CODJudiaB
R14690_Bodjouo_CMR_ludiaA
FMNH222398_Tshuapa_COD_ludiaA
FMNH2224tw_Tshuapa_CODJudiaA
FMNH219676_Tshuapa_COD_ludiaA
Crocidura Tshuapa COD luciaA
FMNH2195T0_Tsliu8pa_COD_ludJaA
yw FMNH219571_T5huapa_COD_lLidlaA
FMNH219673 Tshuapa_COD_ludiaA
FMMH222609_Tshuapa_COD_lud»aA
EU426997JudiaA
i r R19374_NgottoCAF_ludiaA
TP0BA0586 _NgtHtO_CAF_1udiaA
" BAD611 Ngotto_CAF_ludLaA
- FMNH222607 Tshuapa_COD_ludJaA
NC0051_Bohou_CAF_lud«aA
BDA114_Kponyo_COD_ludiaA
• Jck10567_Semliki_UGA_ludJaA
MW075624 Yangambi_CODJudiaA
EJ7 MW07561D_Yar>gambi_COa_ludiaA
MW075623_Yangambi_COD_ludiaA
100
EBOl05_lnkanamon9o_COD_gfassei
100 FMNHl62140_Woleu-Mlem_GAB_Crotidura_gra:
R22693 Odzala COG grassei
FMNH167717 Daodou_GAB_Cro[:ldura_grasafli
0Q gi FMNH1S7710_Moukalaha_GAB_grassei
86 FMNH1677ia_Doudou_GAB_Cracidura_grassei
6# GA2578_Moueva_GAB_gra&sel
EF524759_Moiieva_GA6_graasei
7 9
FMNH167716 Doudou_GAB_gfassei
&1
G7 FMNH1677D1_Doijdou_GAB_CrQcidura_grasBai
71 CAK14B_Barnbama CDG_grassei
g 6 FMNH1G7719_Monls_GAB_gras£&i
go CAK19_Kissiki_CQG grassei
100 FMNH167663_Doudou_GAB_grassei
FMN H16770B_Mou kalaba_GAB_graasai
i MN597604_COD_crenataC
t LlK169_COD_crenalaC
' BDA332 Kponyo_COD_crenataC
1
MN597609_COD_crenataC
,/ MWD75605 Yangambi_COD_crenataC
i COB1245 Yan9ambi_COD_cferiataC
j£— EO427032_creiaiaC
W C77_Yangambi_COD_crenataC
MWQ75607_Yangambi_COD_crenalaC
GA0117 GAB_crenataB
100 R14581_Dja^CMR_crenataB
R14585_Dja_CMR_crenataB
i R14737J3ja_CMR_crenataB
FMNH162154_Woleu-NLern_GAB_crenataB
R14694_Dja_CMR_crenataB
10Q R14542 D]a_CMR_crenalaB
59 R14586_Dja CMR_crenataB
97 TCB77_TchabglMbal»_CMR_crenataB
85 CMi07943_Kofup_CMR_crenataB
gg GA0029_Mvoum_GAB_crenalaB
100 GA0Q83_GAB_crenataB
7 4
GA0OB8_GAB_crenataB
R19422 Nqotlo CAF crenataB
100
R16120_NgottO_CAF_crenataB
R18122JMgotlo_CAF_crenataB
R1927 7 _Ngotio_CAF_crenalaB
R18333_Ngollo_CAF_cf«nalaB
R18123_Ngotlo_CAF_crenalaB
R19487 Ngotto CAF_crenalaB
80
R19487_Ngotto_CAF_crer>alaB
g ? BA056*_Ngolto_CAF_dBnatBB
R1 -8317_Ngotto_CAF_crwat aB
l w ) R194aa NgďK>_CAF_crenataB
B 8 Ria493_Ngolto_CAF_crenřitaB
R193fl2_Ngo«o_CAF wenataB
R19144_Ngotto_CAF_crenataB
R19904_Ngotto_CAF_crenata B
100 R 1 8 9 3 2
- N
9 ° n o
- C A F
- c
* e n a l a B
R19399_Ngotio_CAF «enalaB
BA06GDMgotto_CAF_crenat aB
R19127_NgoNo_CAF_crena1aB
BADBB9_Ngolto_CAF cranalaB
li PO0D6_hlimtja_GIN_crenataB
g 2 eA08B5_NgcWo_CAF_crenalaB
100R19294_Mgotto_CAF_creťiíi|aB
BA1031 NcjúIiů CAF cranataB
R18742_Ngotto_CAF_crenalaB
g o R1936Q_Ngoíto_CAF_crenalaB
BA0559_N9Otta_CAF crenataB
R1957 3_NgQttc._CAF_wenalaB
100 BA0628_Ngotlo_CAF_crenataB
IDO R18740_hgOMí>_CAF_CWiaiaB
^ R1S326 Ngalla_CAF crena!a B
I R19D38_Ngotto_CAF_crenalaB
R19589_Ngotto_CAF_crenalaB
R19220JMgoltn_CAF crenalaB
R19015_Ngotto_CAF_crenalaB
R19263_Ngotto_CAF_ciena1aB
OAOi 97_Maloutiga_GAB_Crei>ataA
GAO 163_ Malounga_GAB_cranalaA
GAG125_Maluurioa_GAB_crenalaA
s 1
GAŮi2e_M0lQunga_GAB_crenalaA
FMNH167702 Moukalaba GAB creťiaraA
GA2S72_Moueva GAB_crenataA
GAD645_MouBva GABcrenataA
GA291G Moukalatid GAB crenataA
FMNH1677Q4_Moukalaba_GAB_ctenalaA
GA2174_Mw«va_GAB_cfenalaA
GA3099_Moueva^GAB_CíenalíiA
GA0625 Mouova GAB c/anaiaA
GA2695_Moueva GAB_crenalaA
EF524762_MouflYa_GAB_creTiBtaA
GAGĚŮ7 Moueva GAB cienalaA
GA• 6B6_Maueva_GAB_cranataA
4f3 GA2755_Moueva_GAB_cienal3A
GA0628_Moi*ewa_GAB_creoaiaA
GA2087_Mouava_GAB a enalaA
Xj CAKia_KíssÍkÍ_COG_cranataA
' CAK96_Loungou COG_crenataA
R22B59 Odzala COG_«analaA
^tjO R17184_Oifce.la_COG_c(enalaA
R1724Q_Odiala_CQG_c>enalaA
=í22aC5_Ubamo_COG_iuerataA
9)4 R17ia6_Odzala COG crůnataA
R17284_Otírala COG_cranataA
R16910_Wboroo COGcrenalaA
R227B1 Ockala COG crenataA
CAK92_Loungou_COG_cfenataA
R17229_Od;ala_COG_crenaleA
R168l6_Odz3la_COG_cranaiaA
R17ia5_Odzala COG.wanaTaA
GA259t_Moii-ava GAB_crenataA
GA120& Mnueifa_E
FMNH2O3701 llombwft COD doliůhuraE
FMI\IH203699_ltomb*ů COD_dolichuraE
FMMH203T03 Uambwa_COD_dkjlicJiur8E
9Jn FMNH203698 Ki*j_CQD_Croddura dolrehura E
K-1
FMNH227545_Kivu_COD_Crocidura_dolichjra_E
' FMNHZZ7547_Kivií_COD_Crocidure_iW'Chijr3_E
FMNH203702_ltombwe_COO_dollCrH.f9E
FMNH203597 Klvu COD Crocidur3_dol FMNHZ27542_Kivu_COD_CFDCidura_dolicriiira_E
Šfi- FMNH227^_K(vu_CCO_Ox<ádura_dolichura_E
^ FMNH227M6_Ktvu_COD_CTOcklufa_dolcjciduia_ddlchuraJ=
FMNH223973_Kibalo_LIGA_iíolJctiurBE
FMNH223569_UGA_dolicriuraE
FMNH223S59^Kibate_UGA_dollchuraE
FMNH223960_Kibala UGA_dolichuraE
^ AWF3977_Bangupa_COD_dolicr>jraE
awf:bHiigiipü c;cju iiuiidiurai
W MWö75S17_Yangambi_CaD_do(lchuFHE
^ MW07»Z9_Yangomrj._COD_dti|itliLiraE
iB FWNH1J7594_Klo.ľa_BNi_doiichijťaE
UC FMNH160127 Kign/I UGA Crúcidura dolichura E
KF110762_Bwindl_UGA_dolichuraE
F VI'.K : ; NJyiingwe KVVA dni -i ir íl
FMNH207284 Nyungwu UGACrocIdura dolíĽtiura E
FM N H207282_Nyungwo_FiWA_dol icliuraE
T
FMMH13T69Ď_Kibira_BNI_doiichurHE
FMNH137693_Klbira_BNI_d0ilchur9E
FMNH137597 Kihira SNI dolichuraE
GA0113 GAB dc*churaW
GAD084 GAB_dolpChuraW
R14357_Dja_CMR_dolictmmW
CM107948_Kurjp_CMR_dolichiJíaW
R14379_Dja_CW RjtolichuraW
R1453 7_D|a_CW RjJolichuraW
Ria3fl5_NgGHö CAF ílňlifinjraW
R182O0_Ngotto_CAF dolicauraW
BA0614 Ngotto_CAF_dollchurBW
R1B149_Ngollo_CAFdoltchuraW
100 R19366_Ngotto_CAF_doiichijraW
7j4 Fii8242_higotto_CAF_dolichuraW
R19361_Mgotto_CAF_dollehuraW
R19302_Batouri_CAF doilchuiaW
BA0821 _Balou n_CAF_dolictiuraW
BA0SSÍ_N90iio_CAF_doltchurítW
R19446_Ngoi1o_CAF doltchuraW
R19058_BatouriCAFdollchuf•W
3BADe44_Baloun_CAF_doHchuraW
. 94 BA0613 Ngolla CAF dulicriuraW
H1fl22S_Mgt>tto_CAF_dQlichuraW
IM "
1 0 0 B
82
OD r
10C
i FMN
L™'
7:
R1fl226_Ngotto_CAF_dalichijřaW
,n BAOD75_BaloiJri CAF dolichiiraW
90 R1S449 Ngollo CAF dolKhuraW
R1931 t>_NgDl!o_CAF_dolicliijreW
=í 19931 _Ngot!o_CAF_*)|ichiJ raW
CM107S42_Karup_CMR_dollchijiaW
BAŮ576 NgaHQ_CAF_dalictuir_W
Sfe R19732_NgoMo_CAF_(tolichuraW
« RieidSr^tloCAFdokhuraW
^ R19725_N9ono_CAF dofcchuraW
1 f i BA0a49_Ngc4tc_CAF_dolichur_W
r 19044_Ngotto_CAF_dolichuraW
R18727_Ngotlo_CAF doltehuraW
^ BAQ690_Ngotto_CAF_dc4ichuraW
R16296_NgottD_CAF_dolichuroW
R19$03_EeWji-j_CAF_ijolichufaW
3_R19247 BalQLii CAF dolicfiuraW
10D R18314_NgoKo_CAF_dcl icbjraW
6
R1879^ NgoluraW
R19033 Baiouh_CAF_dolichiJFaW
FMNH202829 Bioko GNQ Crociduru doli-hura W
FMNH202B20_Moka_GNQ_dollchuraW
FMNHzrjZ819_eník0i_GNO_CrocidLira_doJichurB_W
FMNH202S27_BiokD_GhKl_Crocfdijra_(IOltCliUfa_W
FMNH202824_Moka GNQ dollchuraW
KPD51968_Moka_GNQ_dollchiJr9W
FMNH202ej1_Biciko GNQ Croadura_dDllchurB_W
3 FMNH202B25 Moka GNQ _dolichuiaW
FMNH2Q2&30_MokB_GNQ_dolichiiiaW
MVZ i g6209_SOTgrnbengl\V\ rnlni.nVv
GA0256_KIII_GAB dolichjraW
GAD2B9 Klli GAB dolitriiiraW
GA2272_Moue«i_GAB_dolicriLiraW
S GA3wg_Mwe*a_GAB_dollcriuraW
GA2665_Moueva GAB doltehuraVV
GA28B5_Moukalabe_GAB_dolicriuraW
GA283i_Mtani_GAB_dolicf>-raW
GA1638_Moue«_GAB_doliciiuraA
GA2895 Moukalaba GAB dolicíiuraW
GAa463_Mousva GAB dolichuraW
GA2253_ Moueva GAB _dol IctiuraW
O GA223B Mouava GAB dollchuraW
GA2B28_M-Bni_GAB_d-li-h_r_W
1 SDCAKZ7_Letiayi_COG_dolichu>BW
, ' CAK2B_Lebay._COG_dollchkJ.aW
19 CAK105 Ltmngou COG dolichoraW
W CAK183_Sirrwniboíic,
o COG_dollchuraW
00 CAK154_SimomborKlo_COG_dolPchinaW
CAK169 Simombofido_COG doltehur-W
GAO198_Kili_GAB_dolichuraW
— R1613S_Abeille5_GAB_dol*hUtBW
I^GA2879_Mc*iMlaba_GAB_(k)licr.uraW
V GA2153 Moueva GAB úeitehuraW
T GA2388_Moueva GAB BoltchuraW
13 GA0222_ Klll GAB dollchuraW
í R16064 Aballlas GAB dollchuraW
iL GA1712_Mweva_GAB_dcJichuraW
80 VNi8CM_Leli«li_GAB_dDlidiiirBW
CAK139_Bambama_COG_d0llehuraW
100 G
i 1
G
100
83
Appendix 6
R script used for GMA.
library(geomorph)
library(MASS)
library(ggplot2)
library(dplyr)
tpsjile <- "FILE.tps"
infojile <- "INFO_FILE.txt"
landmarks <- readland.tps(tps_file, specID = NULL, readcurves = TRUE)
info <- read.table(info_file, header = FALSE, stringsAsFactors = FALSE, fill = TRUE)
meta <- data.frame(
species = factor(info[,1]),
specimen_id = info[,2],
sex = factor(trimws(info[,3]))
)
meta$sex[meta$sex %in% c("", "I", "i", "NA", "na")] <- NA
gpa <- gpagen(landmarks)
meta$Csize <- gpa$Csize
# Sex effect
known <- !is.na(meta$sex)
fit_sex <- procD.lm(
gpa$coords[,,known] ~ species + log(Csize) + sex,
data = droplevels(meta[known,]),
iter = 999
)
summary(fit_sex)
# PCA
pea <- gm.prcomp(gpa$coords)
pvar <- round(100 * pca$dA
2 / sum(pca$dA
2), 1)
pca_df <- as.data.frame(pca$x)
pca_df$species <- meta$species
pca_df$specimen_id <- meta$specimen_id
84
p_pca <- ggplot(pca_df, aes(Comp1, Comp2, fill = species)) +
geom_point(shape = 21, size = 3, color = "black") +
coord_equal() +
theme_minimal() +
labs(
x = pasteO("PC1 (", pvar[1],"%)"),
y = pasteO("PC2 (", pvar[2], "%)"),
fill = NULL
)
print(p_pca)
# PCA deformation grids
mean_shape <- mshape(gpa$coords)
par(mfrow = c(2,2), mar= c(1,1,1.8,1))
plotRefToTarget(mean_shape, pca$shapes$shapes.comp1$min,
method = "TPS", mag = 3, main = "PC1 min")
plotRefToTarget(mean_shape, pca$shapes$shapes.comp1$max,
method = "TPS", mag = 3, main = "PC1 max")
plotRefToTarget(rnean_shape, pca$shapes$shapes.comp2$min,
method = "TPS", mag = 3, main = "PC2 min")
plotRefToTarget(mean_shape, pca$shapes$shapes.comp2$max,
method = "TPS", mag = 3, main = "PC2 max")
# Centroid size
size_aov <- aov(Csize ~ species, data = meta)
summary(size_aov)
TukeyHSD(size_aov)
p_size <- ggplot(meta, aes(species, Csize, fill = species)) +
geom_boxplot() +
theme_bw(base_size =15) +
theme(
axis.text.x = element_blank(),
axis.ticks.x = element_blank(),
legend, position = "bottom"
) +
labs(x = NULL, y = "Centroid size", fill = NULL)
print(p_size)
# CVA / LDA on PCA scores
pc_use <- 1:min(27, ncol(pca$x), nrow(pca$x) - length(levels(meta$species)))
lda_res <- lda(pca$x[, pc_use], grouping = meta$species)
85
cva_df <- as.data.frame(predict(lda_res)$x)
cva_df$species <- meta$species
cva_df$specimen_id <- meta$specimen_id
eig <- lda_res$svdA
2
cv_var <- round(100 * eig / sum(eig), 1)
p_cva <- ggplot(cva_df, aes(LD1, LD2, fill = species)) +
geom_point(shape = 21, size = 3, color = "black") +
coord_equal() +
theme_minimal() +
labs(
x = pasteO("CV1 (", cv_var[1],"%)"),
y = pasteO("CV2 (", cv_var[2], "%)"),
fill = NULL
)
print(p_cva)
# CVA deformation grids
Id1_min <- gpa$coords[,,which.min(cva_df$LD1)]
Id1_max <- gpa$coords[,,which.max(cva_df$LD1)]
Id2_min <- gpa$coords[,,which.min(cva_df$LD2)]
Id2_max <- gpa$coords[,.which.max(cva_df$LD2)]
par(mfrow = c(2,2), mar= c(1,1,1.8,1))
plotRefToTarget(mean_shape, Id1_min, method = "TPS", mag = 3, main = "CV1 min")
plotRefToTarget(mean_shape, Id1_max, method = "TPS", mag = 3, main = "CV1 max")
plotRefToTarget(rnean_shape, Id2_min, method = "TPS", mag = 3, main = "CV2 min")
plotRefToTarget(rnean_shape, Id2_max, method = "TPS", mag = 3, main = "CV2 max")
# Leave-one-out classification
explained <- pca$dA
2 / sum(pca$dA
2)
n_pc <- which(cumsum(explained) >= 0.90)[1]
pc_data <- as.data.frame(pca$x[, 1:n_pc])
pc_data$species <- meta$species
lda_cv <- lda(species ~ ., data = pc_data, CV = TRUE)
tab <- table(True = meta$species, Predicted = lda_cv$class)
print(tab)
accuracy <- mean(lda_cv$class == meta$species) * 100
cat("Overall accuracy:", round(accuracy, 2), "%\n")
accuracy_df <- data.frame(
species = rownames(tab),
86
correct = diag(tab),
total = rowSums(tab)
)
accuracy_df$accuracy_percent <- round(accuracy_df$correct / accuracy_df$total * 100, 2)
# Mean shapes and wireframe
# Fill in landmark links:
wireframe <- rbind(
c(1,2),
c(2,3),
c(3,4)
)
speciesjneans <- lapply(levels(meta$species), function(sp) {
mshape(gpa$coords[,,meta$species == sp])
})
names(species_means) <- levels(meta$species)
plot(NULL,
xlim = range(sapply(species_means, function(x) x[,1])),
ylim = range(sapply(species_means, function(x) x[,2])),
asp= 1,xlab = "", ylab = "")
for (sp in names(species_means)) {
shp <- species_means[[sp]]
for (i in 1:nrow(wireframe)) {
segments(
shp[wireframe[i,1],1], shp[wireframe[i,1],2],
shp[wirefrarne[i,2],1], shp[wireframe[i,2],2]
)
}
}
# Procrustes ANOVA and pairwise comparison
fit_shape <- procD.lm(gpa$coords ~ species, data = meta, iter = 999)
summary(fit_shape)
pairwise_shape <- pairwise(fit_shape, groups = meta$species)
summary(pairwise_shape)
87