Functional Skeletal Morphology and Its Implications for
Locomotory Behavior Among Three Genera of
Myosoricine Shrews (Mammalia: Eulipotyphla:
Soricidae)
Neal Woodman1* and Frank A. Stabile2
1Department of Vertebrate Zoology, USGS Patuxent Wildlife Research Center, National Museum of Natural History,
Smithsonian Institution, Washington, DC 20013-7012
2Department of Biology, The College of New Jersey, Ewing, New Jersey 08628
ABSTRACT Myosoricinae is a small clade of shrews
(Mammalia, Eulipotyphla, Soricidae) that is currently
restricted to the African continent. Individual species
have limited distributions that are often associated
with higher elevations. Although the majority of spe-
cies in the subfamily are considered ambulatory in
their locomotory behavior, species of the myosoricine
genus Surdisorex are known to be semifossorial. To bet-
ter characterize variation in locomotory behaviors
among myosoricines, we calculated 32 morphological
indices from skeletal measurements from nine species
representing all three genera that comprise the sub-
family (i.e., Congosorex, Myosorex, Surdisorex) and
compared them to indices calculated for two species
with well-documented locomotory behaviors: the ambu-
latory talpid Uropsilus soricipes and the semifossorial
talpid Neurotrichus gibbsii. We summarized the 22
most complete morphological variables by 1) calculating
a mean percentile rank for each species and 2) using
the first principal component from principal component
analysis of the indices. The two methods yielded simi-
lar results and indicate grades of adaptations reflecting
a range of potential locomotory behaviors from ambula-
tory to semifossorial that exceeds the range repre-
sented by the two talpids. Morphological variation
reflecting grades of increased semifossoriality among
myosoricine shrews is similar in many respects to that
seen for soricines, but some features are unique to the
Myosoricinae. J. Morphol. 276:550–563, 2015. VC 2015
Wiley Periodicals, Inc.
KEY WORDS: anatomy; digging; fossorial; Insectivora;
Soricomorpha; substrate use; terrestrial
INTRODUCTION
Shrews (Mammalia: Eulipotyphla: Soricidae)
exhibit a limited range of locomotory behaviors
and substrate use. Most soricids are classified as
terrestrial, ambulatory predators that forage pre-
dominantly on epigeal invertebrates. Only about
10% of shrews are considered semifossorial, and a
few other species are scansorial or semiaquatic
(Hutterer, 1985; Churchfield, 1990). Despite the
predominance of ambulatory behavior in the fam-
ily, semifossorial species occur in all three soricid
subfamilies and are geographically widespread
(Hutterer, 1985; Churchfield, 1990). Paradoxically,
even ambulatory soricids possess a number of
external characters that are common to semifosso-
rial and fossorial mammals. These include a fusi-
form body with short, dense fur; short, stout
limbs; and small pinnae and eyes (Shimer, 1903;
Eisenberg, 1981; Churchfield, 1990; Stein, 2000).
Variation in soricid skeletal characteristics that
are typically tied to locomotory behavior suggests
that substrate use is more nuanced and diverse
than is recognized by standard stereotypical cate-
gories (Woodman and Gaffney, 2014).
Myosoricinae is a clade (Querouil et al., 2001;
Willows-Munro and Matthee, 2011) of small- to
medium-bodied African shrews that comprises at
least 19 species of Mouse Shrews and Forest
Shrews (genus Myosorex Gray, 1838), three species
of Mole Shrews (Surdisorex Thomas, 1906), and
three species of Congo Shrews (Congosorex Heim
de Balsac and Lamotte, 1956; Meester, 1953;
Additional Supporting Information may be found in the online
version of this article.
Contract grant sponsor: National Science Foundation [through
the Natural History Research Experiences Program of the United
States National Museum of Natural History (USNM) (to F.A.S)].
*Correspondence to: Neal Woodman, USGS Patuxent Wildlife
Research Center, National Museum of Natural History, MRC 111,
Smithsonian Institution, PO Box 37012, Washington, DC 20013-
7012. E-mail: woodmann@si.edu
Frank A. Stabile is currently at Department of Ecology & Evolu-
tionary Biology, Yale University, New Haven, CT 06511.
Received 4 August 2014; Revised 9 December 2014;
Accepted 19 December 2014.
Published online 10 February 2015 in
Wiley Online Library (wileyonlinelibrary.com).
DOI 10.1002/jmor.20365
VC 2015 WILEY PERIODICALS, INC.
JOURNAL OF MORPHOLOGY 276:550–563 (2015)
Hutterer, 2005; Stanley et al., 2005; Kerbis
Peterhans et al., 2008, 2009, 2010, 2013; Taylor
et al., 2013). Most myosoricines have limited geo-
graphical distributions in regions that are diffi-
cult to access, and their ecology, substrate use,
and locomotory behavior remain poorly docu-
mented. Species of Myosorex are generally con-
sidered to be ambulatory, whereas species of
Surdisorex are known to be semifossorial (Dun-
can and Wrangham, 1971; Coe and Foster, 1972;
Hutterer, 1985). Our early observations indicated
that locomotory differences between these two
genera are reflected to some extent in external
morphological characteristics, including the rela-
tive dimensions of the palm, digits, and claws of
the forefeet (Woodman and Stabile, 2015). Subse-
quent inspection of postcranial skeletons of Myo-
sorex and Surdisorex revealed abundant
variation (Fig. 1). External and skeletal differen-
ces just within the genus Myosorex suggested the
potential for a broader diversity of locomotory
behaviors than is currently recognized.
Fig. 1. Anterior aspect of left humeri of nine myosoricine shrews and two moles. Shrews: (A) M. cafer (FM 165585); (B) M. kihaulei
(FM 204860); (C) M. geata (FM 197673); (D) C. phillipsorum (FM 177689); (E) M. varius (FM 165592); (F) M. blarina (FM 144209);
(G) M. zinki (FM 174117); (H) S. polulus (USNM 589820); (I) S. norae (USNM 589817). Moles: (J) U. soricipes (USNM 574297); and
(K) N. gibbsii (USNM 273085).
551MYOSORICINE LOCOMOTORY ADAPTATIONS
Journal of Morphology
In the absence of direct behavioral observations
for most myosoricines, we investigated the poten-
tial for using skeletal morphology to assess loco-
motory mode. We studied and measured skeletal
elements from nine myosoricine species represent-
ing all three genera and calculated 32 morphologi-
cal indices from them. Most of these indices derive
from previous studies of a variety of other mam-
mals and were designed to aid in understanding
locomotory adaptations (Price, 1993; Lemelin,
1999; Sargis, 2002; Weisbecker and Warton, 2006;
Weisbecker and Schmid, 2007; Kirk et al., 2008;
Samuels and Van Valkenburgh, 2008; Hopkins and
Davis, 2009; Elissamburu and De Santis, 2011;
Woodman and Gaffney, 2014). We statistically
summarized these indices and constructed a rela-
tive scale to provide a predictive measure of loco-
motory behavior from ambulatory to semifossorial.
Herein, we evaluate the relative locomotory behav-
ior of each species based on this scale and summa-
rize skeletal changes that accompany increased
semifossoriality in the myosoricine clade.
MATERIALS AND METHODS
Because of their small size, individual bones of soricids are
difficult to manipulate and measure accurately or precisely
with hand-held calipers. Instead, the scapula, humerus, ulna,
radius, femur, and tibiofibula were digitally photographed. The
resulting images were imported into Adobe Photoshop CS3
Extended 10.0.1 (Adobe Systems, San Jose, CA), and 21 varia-
bles (Fig. 2) were measured using the Custom Measuring Scale
in the Analysis menu following Woodman and Gaffney (2014).
Following the procedures outlined in Woodman and Morgan
(2005; see also Woodman and Stephens, 2010; Sargis et al.,
2013a, 2013b), we obtained digital x-ray images of the bones of
the manus and the pes by x-raying feet of dried skins using a
Kevex X-Ray Source 4.1.3 (Kevex, Palo Alto, CA) with Varian
Image Viewing and Acquisition 2.0 software (VIVA, Waltham,
MA) in the Division of Fishes, National Museum of Natural
History, Washington, DC. We imported the resulting images
into Adobe Photoshop CS3 Extended and measured lengths and
widths from each of the bones and claws of the five rays of the
manus and from ray III of the pes (Fig. 3). “Digit” herein refers
to the bones and tissues associated with the phalanges, and
“ray” refers to those associated with the phalanges and meta-
carpal. As a proxy for body size, we calculated head-and-body
length (HB) by subtracting tail length from total length as
recorded by the original collectors. All measurements are
reported in mm and are summarized in Supporting Information
Supplement 1.
We measured 87 individuals representing three genera and
nine species of myosoricine shrews. Specimens used in this
study (Appendix) are deposited in the following institutions:
American Museum of Natural History, New York, NY; Field
Museum of Natural History, Chicago, IL; National Museum of
Natural History, Washington, DC. Sample sizes were limited by
the availability of skeletons (Bell and Mead, 2014). To better
understand the differences in skeletal characters distinguishing
terrestrial and semifossorial species of shrews, we measured 10
specimens of the ambulatory Chinese Shrew-mole, Uropsilus
soricipes Milne-Edwards, 1871 and 26 specimens of the semifos-
sorial Shrew-mole, Neurotrichus gibbsii (Baird, 1857). Moles
(Talpidae) are a likely sister-group to shrews (Meredith et al.,
2011), and their adaptations for semifossorial and fossorial loco-
motion and substrate use are well documented (e.g., Edwards,
1937; Sanchez-Villagra et al., 2004).
To assess locomotory function, we calculated 32 osteological
indices (Table 1), most of which have previously been used to
characterize locomotory adaptation and substrate use among
soricids (Woodman and Gaffney, 2014), rodents (Price, 1993;
Weisbecker and Schmid, 2007; Samuels and Van Valkenburgh,
2008; Elissamburu and De Santis, 2011) and other taxa (Leme-
lin, 1999; Sargis, 2002; Weisbecker and Warton, 2006; Kirk
et al., 2008; Hopkins and Davis, 2009). We compared tabled
indices for myosoricines with those for 1) the talpids Uropsilus
Fig. 2. Measurements of the long bones of the postcranial skel-
eton. (A) Posterior aspect of left scapula: SL, greatest length of
scapula. (B) Anterior aspect of left humerus: HL, length of
humerus; HDW, distal width of humerus (epicondylar breadth);
HDPC, length of deltopectoral crest. (C) Anterior aspect of left
humerus: HAR, axis of rotation of the humerus; HLD, least
mediolateral diameter of humerus; HTT, length from head of
humerus to distal edge of teres tubercle; HTTR, teres tubercle
input lever for rotation (measured at a right angle to HAR). (D)
Anterior aspect of left radius: RL, length of radius; RDW, distal
width of radius. (E) Lateral aspect of left ulna: UL, total length;
UFL, functional length (output lever arm); UOP, length of olecra-
non process (input lever arm). (F) Posterior aspect of left ulna:
UPC, width of proximal crest; ULD, least mediolateral diameter.
(G) Posterior aspect of left femur: FL, length; FDW, distal width
(epicondylar breadth); FLD, least mediolateral diameter. (H)
Anterior aspect of left tibiofibula: TL, length; TDA, width of dis-
tal articular surface; TDW, distal width.
552 N. WOODMAN AND F.A. STABILE
Journal of Morphology
and Neurotrichus, 2) two ambulatory (Cryptotis parvus, C. mer-
riami) and one semifossorial (C. lacertosus) species of soricine
small-eared shrews from Woodman and Gaffney (2014), and 3)
summary statistics for terrestrial, semifossorial, and fossorial
rodents reported by Samuels and Van Valkenburgh (2008). We
also determined 1) the length of each element of the forelimb
relative to the total length of the forelimb (calculated as
HL1RL1ML1PPL1MPL1DPL), 2) the length of each bone
of ray III of the manus relative to the total length of the ray
(ML1PPL1MPL1CL), 3) the length of each element of the
hind limb relative to the total length of the hind limb
(FL1TL1hML1hPPL1hMPL1hDPL), and 4) the length of
each bone of ray III of the pes relative to the total length of the
ray (hML1hPPL1hMPL1hCL). Where data permitted, indi-
ces were calculated for each individual and summarized statis-
tically for each species (Supporting Information Supplement 2).
To account for missing data, we also calculated indices from
mean values of variables (Table 2). The latter values are pre-
sented in the tables, unless otherwise stated. All indices are
expressed as whole number percentages.
Of the 32 indices we investigate in our study, 10 could be cal-
culated for only three species of myosoricines, primarily
because of the lack of complete ulnae and tibiofibulae. We used
two different analyses to combine the remaining 22 indices to
calculate a summary score for each species. The summary
scores provide an overview of interspecific variation and permit
us to define relative grades of ambulatory and semifossorial
adaptation. In the first analysis, we computed the percentile
rank of each species for each of the 22 morphological indices
(Tables 2 and 3). We then averaged the 22 percentile ranks to
obtain a mean percentile rank for each species that represents
its relative adaptation for ambulatory versus semifossorial loco-
motion on a possible scale from 0 (most ambulatory) to 100
(most semifossorial). To estimate the variance in the mean per-
centile ranks, we carried out a simple jackknife resampling pro-
cedure. As a further test of the stability of the scale, we
recalculated mean percentile ranks using all 32 indices despite
missing data (Supporting Information Supplement 3). In our
second analysis, we used principal component analysis (PCA) of
a covariance matrix of the same 22 indices to examine variation
among myosoricines in multivariate space. We used a covari-
ance matrix for this analysis because all indices were measured
on the same 0–100 scale, so no standardization of variable
scales was necessary. With the exception of the two species of
Surdisorex, whose semifossorial behaviors are documented, our
interpretations of myosoricine locomotory modes are based on
these two summary scales. To ease comparisons in tables, spe-
cies of myosoricines are listed in increasing order by our a pos-
teriori mean percentile rank. Univariate statistics were
calculated in Excel 97–2003 (Microsoft Corp., Redmond, WA)
and multivariate statistics in Systat 11.00.01 (Systat Software,
Chicago, IL).
RESULTS
Mean Percentile Ranks
Within a possible range from 0 (most ambula-
tory) to 100 (most semifossorial), the mean percen-
tile ranks for myosoricines range from 18 to 83
(Table 3; Fig. 4). Mean percentile ranks for the
ambulatory talpid Uropsilus (33) and the semifos-
sorial talpid Neurotrichus (75) provide a general
guide for this scale. Standard deviations for mean
percentile ranks, calculated from jackknife resam-
pling, range from 0.4 to 1.1 among the myosori-
cines, but are higher for the two moles. The
relative order of myosoricine species remained the
same for each iteration of the jackknife procedure.
When we calculated percentile ranks for all 32
indices, the order of these species also remained
the same (Supporting Information Supplement 3),
indicating some degree of stability in our sample.
The addition or subtraction of taxa used in the cal-
culations could change the mean rankings sub-
stantially, however, so they are only meaningful
within the context of our study.
Based on the distribution of mean percentile
rankings, the most ambulatory myosoricine among
the nine species we examined is M. cafer, and the
most semifossorial is S. norae. The average differ-
ence in rankings between adjacent myosoricine
species on this scale is eight percentile points
(Table 3). Differences >8 percentile points between
adjacent species occur in only two places. At the
lower end of the scale, M. cafer is separated from
M. geata by 12 percentile points, and on the
higher end of the scale, M. zinki is separated from
S polulus by 16 percentile points. As a compari-
son, the mean percentile ranking for Uropsilus
places the ambulatory mole between M. kihaulei
and C. phillipsorum. This suggests that M. cafer,
M. geata, and M. kihaulei are more suited to
ambulatory locomotion than other myosoricines.
The semifossorial Neurotrichus ranks above M.
zinki and below S. polulus and S. norae, which
suggests that the two species of Surdisorex are
more adapted for digging than Neurotrichus.
Fig. 3. Dorsal aspect of ray III of the manus illustrating varia-
bles measured on the manus and pes: ML, length of metacarpal
or metatarsal; PPL, length of proximal phalanx; MPL, length of
middle phalanx; DPL, length of distal phalanx; MW, width of
metacarpal; PPW, width of proximal phalanx; MPW, width of
middle phalanx; DPW, width of distal phalanx. An “h” preceding
the abbreviation indicates the measurement derives from the
pes (e.g., hMPL5 length of middle phalanx of the pes).
553MYOSORICINE LOCOMOTORY ADAPTATIONS
Journal of Morphology
TABLE 1. List of indices used in this study, with their abbreviations in parentheses
1. Intermembral index (IM5 [HL1RL]/[FL1TL]) compares the lengths of the forelimbs and hind limbs (Sargis, 2002). IM typi-
cally increases with increasing fossoriality in rodents (Samuels and Van Valkenburgh, 2008).
2. Humerofemoral index (HFI5HL/FL) represents the length of the humerus as a proportion of the length of the femur (Sargis, 2002).
3. Metapodial index (FOOT5ML/hML) indicates the relative sizes of the forefeet and hind feet by comparing the length of meta-
carpal III to that of metatarsal III.
4. Distal phalanx length index (CLAW5DPL/hDPL) compares the relative size of distal phalanx III of the manus to distal phalanx
III of the pes. It increases with increasing fossoriality in rodents (Samuels and Van Valkenburgh, 2008).
5. Claw length index (CLI5CL/hCL) gauges the relative size of claw III of the manus to claw III of the pes.
6. Scapulohumeral index (SHI5SL/HL) indicates relative lengths of the scapula and humerus. This index is typically greater for
more semifossorial eulipotyphlans (Woodman and Gaffney, 2014).
7. Brachial index (BI5RL/HL) shows the relative proportions of the proximal (humerus) and distal (radius) elements of the fore-
limb. The index decreases with increasing fossoriality among rodents (Samuels and Van Valkenburgh, 2008).
8. Shoulder moment index (SMI5HDPC/HL) gauges the size and mechanical advantage of the deltoid and pectoral muscle groups,
which are important in the movement, rotation, and counter-rotation of the humerus (Reed, 1951), by comparing the length of the
deltopectoral crest, on which these muscles insert, to the length of the humerus. This is the same as the delto-pectoral crest length
index (Sargis, 2002). The index increases with increasing fossoriality among rodents (Samuels and Van Valkenburgh, 2008).
9. Humeral robustness index (HRI5HLD/HL) indicates the robustness of the humerus and its ability to resist bending and shear-
ing stresses. The index increases with increasing fossoriality among rodents (Samuels and Van Valkenburgh, 2008).
10. Humeral rotation lever index (HTI5HTTR/HAR) shows the relative length of the teres tubercle measured at right angles to
the longitudinal axis of rotation of the humerus. The teres tubercle is an elongate process on the eulipotyphlan humerus that
serves as the insertion for the large latissimus dorsi and teres major muscles and as a lever for rotating the humerus (Reed,
1951). HTI increases with increased semifossoriality among species of Cryptotis (Woodman and Gaffney, 2014).
11. Teres tubercle position index (TTP5HTT/HAR) represents the relative position of the teres tubercle along the axis of rotation
of the humerus (HAR). In more robust humeri with larger surfaces for muscle attachment, the teres tubercle is often more dis-
tally positioned. Hence, the index should increase with greater semifossoriality.
12. Humeral epicondylar index (HEB5HDW/HL) measures the width of the distal humerus relative to the length of the humerus and
represents the area available for the origins of muscles involved in flexing, pronating, and supinating the forearm. It typically
increases in mammals with increasing semifossoriality and fossoriality (Hildebrand, 1985b; Samuels and Van Valkenburgh, 2008).
13. Radial distal width index (RDW5RDW/RL) measures the relative width of the distal end of the radius, providing a gauge of
its robustness and its resistance to the stresses associated with digging.
14. Olecranon length index (OLI5UOP/UFL) is one of many variations on the index of fossorial ability of Hildebrand (1985a). The
ulna acts as a lever that pivots at the trochlear notch, and OLI is used to gauge the amount of force exerted by the triceps bra-
chii muscle on the olecranon process that is transmitted to the functional arm of the ulna. More semifossorial and fossorial
mammals generally have a longer olecranon process to accommodate a larger triceps brachii, resulting in larger OLI (Reed,
1951; Vizcaino and Milne, 2002; Samuels and Van Valkenburgh, 2008; Woodman and Gaffney, 2014).
15. Triceps metacarpal outforce index (TMO5UOP/[UFL1ML]) gives the length of the olecranon process as a proportion of the
functional arm provided by the ulna and metacarpal III. A variation of OLI, this index measures the amount of force input on
the olecranon process that is transmitted to the tip of the metacarpal of ray III (Price, 1993).
16. Triceps claw outforce index (TCO5UOP/[UFL1ML1PPL1MPL1CL]) expresses the length of the olecranon process relative
to the combined functional lengths of the ulna and ray III. An extension of Hildebrand’s (1985a) index of fossorial ability and
Price’s (1993) triceps metacarpal outforce index, TCO represents the proportion of force input on the olecranon process by the
triceps muscle that is transmitted to the tip of the claw of ray III, which is the initial point of contact with the soil.
17. Olecranon crest index (OCI5UPC/UFL) is a measure of the relative length of the olecranon crest, a prominent structure on
the olecranon process of many eulipotyphlans, which is the insertion for much of the triceps brachii. OCI is an approximate
gauge of muscle size, and, therefore, another measure of the relative input force on the ulna (Woodman and Gaffney, 2014).
18. Ulnar robustness index (URI5ULD/UFL) measures the robustness of the ulna and its ability to resist bending and shearing
stresses. URI increases with semifossoriality and fossoriality (Samuels and Van Valkenburgh, 2008).
19. Relative length of the manual distal phalanx [%DPL5DPL/(ML1PPL1MPL)] is the length of distal phalanx III of the manus
relative to the combined length of the proximal three bones of ray III.
20. Relative length of the manual claw [%CL5CL/(ML1PPL1MPL)] is the length of claw III of the manus relative to the com-
bined length of the proximal three bones of ray III.
21. Relative support for the claw (%CLS5DPL/CL) represents the proportion of claw III of the manus that is supported by distal
phalanx III.
22. Metacarpal width index (MW35MW/ML) measures the robustness of metacarpal III of the manus in relation to its length.
23. Phalangeal index (PI5 (PPL1MPL)/ML) shows the lengths of the proximal and middle phalanges relative to the metacarpal. The
index varies considerably among rays of an individual, so ray III is typically used for comparisons among species (Lemelin, 1999).
24. Manus proportions index (MANUS5PPL/ML) measures the size of the proximal phalanx relative to the metacarpal (Kirk
et al., 2008; Samuels and Van Valkenburgh, 2008). This is the same as Kirk et al.’s (2008) proximal phalangeal index.
25. Crural index (CI5TL/FL) measures the relative lengths of proximal (femur) and distal (tibiofibula) long bones of the hind
limb. Like the brachial index, this decreases with increasing fossoriality among rodents (Samuels and Van Valkenburgh, 2008).
26. Pes length index (PES5hML/FL) represents the length of metatarsal III relative to femur length and has been used to indi-
cate the relative size of the hind foot (Samuels and Van Valkenburgh, 2008).
27. Femoral robustness index (FRI5FLD/FL) quantifies the robustness of the femur and its ability to resist bending and shearing stresses.
28. Femoral epicondylar index (FEB5FDW/FL) approximates the area available for the origins of the gastrocnemius and soleus
muscles used in extension of the knee and plantar-flexion of the pes in rodents (Samuels and Van Valkenburgh, 2008). In shrews
and talpids, the same region is the origin for the plantaris (toe flexor), gastrocnemius (pes extensor), extensor digitorum longus
(extensor and adductor of digits; dorsoflexor of foot), and insertion for the caudofemoralis (femur retractor) and adductor longis
(femur adductor; Reed, 1951).
29. Distal tibiofibular articulation index (DTA5TDA/TDW) measures the width of the articular region for the astragalus between
the lateral and medial malleolus relative to the distal width of the tibia.
30. Relative length of the pedal distal phalanx [%hDPL5hDPL/(hML1hPPL1hMPL)] is the length of the distal phalanx of ray
III of the pes relative to the combined length of the proximal three bones of that ray.
31. Relative length of the pedal claw [%hCL5hCL/(hML1hPPL1hMPL)] is the length of the claw of ray III of the pes relative to
the combined length of the proximal three bones of that ray.
32. Relative support for the pedal claw (%hCLS5hDPL/hCL) is the proportion of the claw of ray III of the pes supported by the
distal phalanx.
All indices are expressed as whole number percentages.
TABLE 2. Mean indices calculated for myosoricines
IM HFI FOOT CLAW CLI SHI BI SMI HRI HTI TTP
M. cafer — 88 71 100 114 92 — 46 10 15 36
M. geata — 91 72 93 116 94 — 47 9 16 39
M. kihaulei — 90 70 97 117 94 — 46 9 16 40
C. phillipsorum — 90 76 113 107 104 — 50 11 20 39
M. varius 70 86 69 109 126 99 103 48 10 18 43
M. blarina 72 89 77 107 128 100 96 50 9 19 41
M. zinki — 82 75 120 152 108 — 47 13 18 42
S. polulus — 84 79 162 174 113 — 62 17 39 55
S. norae 70 81 71 143 166 120 103 62 17 35 51
Soricines
C. parvus 71 86 66 97 103 100 103 42 9 17 40
C. merriami 73 92 68 99 96 95 93 44 9 17 38
C. lacertosus 75 92 67 125 131 123 102 44 15 33 51
Talpids
U. soricipes 70 88 54 81 88 125 128 44 10 16 39
N. gibbsii 66 82 41 123 123 174 119 62 22 38 50
Rodents
Terrestrial 74 — — 79 — — 100 42 9 — —
Semifossorial 76 — — 116 — — 92 46 10 — —
Fossorial 85 — — 159 — — 91 54 11 — —
HEB RDW OLI TMO TCO OCI URI %DPL %CL %CLS MW3
M. cafer 32 — — — — — — 19 36 51 11
M. geata 35 — — — — — — 21 42 50 13
M. kihaulei 35 — — — — — — 23 46 49 12
C. phillipsorum 42 — — — — — — 23 39 59 13
M. varius 35 13 19 14 10 30 5 27 52 52 14
M. blarina 39 15 24 18 12 31 8 28 57 49 15
M. zinki 47 — — — — — — 29 61 48 16
S. polulus 58 — — — — — — 45 76 59 20
S. norae 60 17 31 23 14 40 9 46 78 60 20
Soricines
C. parvus 36 13 18 13 9 24 6 16 35 46 10
C. merriami 35 12 20 15 11 27 7 15 29 52 10
C. lacertosus 58 17 28 21 14 40 9 36 63 58 20
Talpids
U. soricipes 44 9 16 12 8 17 4 22 40 55 12
N. gibbsii 54 19 26 21 13 36 7 68 104 65 41
Rodents
Terrestrial 25 — 16 — — — 4 — — — —
Semifossorial 29 — 22 — — — 6 — — — —
Fossorial 37 — 33 — — — 7 — — — —
PI MANUS CI PES FRI FEB DTA %HDPL %HCL %HCLS
M. cafer 91 55 — 48 8 21 — 15 26 58
M. geata 90 54 — 46 10 23 — 19 31 62
M. kihaulei 88 53 — 47 10 23 — 19 32 59
C. phillipsorum 87 55 — 46 10 24 — 18 32 56
M. varius 88 52 149 44 9 22 45 20 34 60
M. blarina 84 50 142 40 10 24 46 23 39 59
M. zinki 86 51 — 42 11 25 — 21 34 60
S. polulus 83 49 — 42 10 24 — 25 39 64
S. norae 84 50 135 42 11 25 44 27 38 69
Soricines
C. parvus 96 58 145 46 8 23 41 13 26 49
C. merriami 98 62 143 44 10 22 43 12 23 50
C. lacertosus 97 58 147 41 10 28 48 23 38 60
Talpids
U. soricipes 104 66 186 60 9 26 50 18 30 60
N. gibbsii 135 81 169 50 10 27 46 30 46 65
Rodents
Terrestrial — 62 119 50 8 18 — — — —
Semifossorial — 58 109 41 9 21 — — — —
Fossorial — 52 105 36 10 25 — — — —
Comparable values for soricines and rodents are from Woodman and Gaffney (2014) and Samuels and Van Valkenburgh (2008),
respectively. Abbreviations explained in Table 1.
555MYOSORICINE LOCOMOTORY ADAPTATIONS
Principal Component Analysis
The first four principal components (PCs) from
our PCA all have eigenvalues >1,000 (Table 4).
Most indices have high positive loadings on PC1,
which explains >59% of the variation in the
model. One exception is the index FOOT, which
has a low negative loading on this component.
PC2 is mostly a contrast between FOOT and the
three indices PES, MANUS, and PI. It explains
>19% of the variation and distinguishes M. blar-
ina (low negative score) and the two talpids (high
positive scores) from the other eight species based
on the relative lengths of their metapodials and
phalanges. A plot of factor scores on these two
components is shown in Figure 5. PC3 (not shown)
is a contrast between FOOT and %CLS. It
TABLE 3. Percentile ranks for the 22 indices used to calculate the mean percentile rank (mean rank) for each taxon
HFI FOOT CLAW CLI SHI SMI HRI HTI TTP HEB % DPL %CL % CLS
M. cafer 45 55 32 23 5 18 41 5 5 5 5 5 41
M. geata 5 41 14 32 18 36 14 23 23 23 14 32 32
M. kihaulei 18 68 23 41 18 18 14 23 41 23 36 41 18
C. phillipsorum 18 23 59 14 50 64 59 68 23 50 36 14 72
M. varius 59 77 50 59 32 50 41 45 68 23 50 50 50
M. blarina 32 14 41 68 41 64 14 59 50 41 59 59 18
M. zinki 81 32 68 77 59 36 68 45 59 68 68 68 5
S. polulus 68 5 95 95 68 86 82 95 95 86 77 77 72
S. norae 95 55 86 86 77 86 82 77 86 95 86 86 86
U. soricipes 45 86 5 5 86 5 41 23 23 59 23 23 59
N. gibbsii 81 95 77 50 95 86 95 86 77 77 95 95 95
MW3 PI MANUS PES FRI FEB %HDPL %HCL %HCLS Sum Mean Rank SD
M. cafer 5 23 27 23 5 5 5 5 14 397 18 60.8
M. geata 36 32 41 45 54 27 36 23 68 669 30 60.7
M. kihaulei 18 45 50 32 54 27 36 36 27 707 32 60.7
C. phillipsorum 36 59 27 45 54 50 18 36 5 880 40 60.9
M. varius 50 45 59 59 18 14 50 55 50 1054 48 60.7
M. blarina 59 81 82 95 54 50 68 81 27 1157 53 61.1
M. zinki 68 68 68 77 91 73 59 55 50 1343 61 60.9
S. polulus 82 95 95 77 54 50 77 82 77 1690 77 61.0
S. norae 82 81 82 77 91 73 86 68 95 1818 83 60.4
U. soricipes 18 14 14 5 18 86 18 14 50 720 33 61.3
N. gibbsii 95 5 5 14 54 95 95 95 86 1648 75 61.4
Abbreviations of indices are explained in Table 1. Standard deviation (SD) of mean percentile rank is from jackknife resampling.
See also Supporting Information Supplement 3.
Fig. 4. Boxplots of mean percentile ranks for each species. Mean (crosses), SD (boxes), and
range (whiskers) are from jackknifed resampling (Table 3).
556 N. WOODMAN AND F.A. STABILE
Journal of Morphology
explains >6% of the variation and effectively sepa-
rates C. phillipsorum (low negative score) from
the other 10 species based on its relatively short
claws and long distal phalanges. PC4 (not shown)
explains >5% of the variation and is most influ-
enced by FRI and FEB. This component separates
Uropsilus and M. zinki (low negative scores) and
M. cafer and M. varius (high positive scores) from
the remaining seven species based on the relative
robustness of their femurs.
PC1 incorporates 21 of the 22 indices and effec-
tively describes a locomotory gradient from more
ambulatory (more negative) to more semifossorial
(more positive). The individual scores on PC1
(Table 5) provide a convenient means of ranking
the locomotory ability of each species in the analy-
sis. This ranking presents the same order of spe-
cies as that provided by the mean percentile
rankings (Table 3). The average difference in
scores between adjacent myosoricine species on
this scale is 0.362 (Table 5). Differences in adja-
cent scores greater than that value occur at three
places, two of which are the same as the large
gaps in the scale of mean percentile rankings: M.
cafer is separated from M. geata by 0.595, M.
kihaulei is separated from C. phillipsorum by
0.421, and M. zinki is separated from S. polulus
by 0.769. The talpid Neurotrichus scores between
M. zinki and S. polulus, which is the same relative
position it had for the mean percentile rankings.
The score for the ambulatory mole Uropsilus, how-
ever, places it between M. cafer and M. geata,
which is lower on the scale compared to its posi-
tion based on mean percentile rankings. The order
of species on the PC1 scale suggests that only M.
cafer is more suited to ambulatory locomotion
than Uropsilus.
DISCUSSION
Morphological Variation Related to
Locomotory Mode
We used the scale from our mean percentile
ranks analysis as our primary model of relative
locomotory mode in myosoricines. By understand-
ing variation in the relative proportions of ele-
ments of the limbs (Tables 6 and 7) and the
performance of each of the 32 indices (Table 2) in
the context of this model, we reconstructed general
patterns of variation in the myosoricine postcra-
nial skeleton as they relate to locomotory function.
In the following discussions, more ambulatory spe-
cies are considered to be those that had lower
Fig. 5. Plot of factor scores on the first two PCs from PCA of
22 indices from nine myosoricines and two talpids. PC1 accounts
for >59% of the variance; PC2 for >19% (see Tables 4 and 5).
TABLE 5. Rounded factor scores on PC1 (see Fig. 5)
Species PC1 Score
M. cafer 21.46
M. geata 20.87
M. kihaulei 20.84
C. phillipsorum 20.42
M. varius 20.13
M. blarina 0.20
M. zinki 0.54
S. polulus 1.32
S. norae 1.42
U. soricipes 20.90
N. gibbsii 1.13
TABLE 4. Component loadings and eigenvalues from PCA of
covariance matrix of 22 morphological indices (see Fig. 5)
Variable PC1 PC2 PC3 PC4
MW3 28.995 1.291 1.794 1.223
DPL 28.868 2.504 2.897 0.154
HDPL 27.848 0.165 8.18 2.002
CL 27.797 1.614 9.949 0.178
TTP 27.65 21.652 5.917 5.013
HTI 26.729 1.031 210.025 3.459
HCL 26.677 22.136 2.655 4.142
CLAW 26.654 22.142 27.462 6.209
HEB 25.737 7.152 25.578 29.791
SMI 25.552 21.559 28.39 8.866
CLI 24.639 213.789 7.745 2.582
HFI 22.631 7.8 5.768 0.842
HRI 22.15 12.59 28.077 2.19
HCLS 21.076 8.986 10.591 0.722
SHI 20.644 17.098 24.121 210.049
FRI 18.135 27.118 0.94 213.356
FEB 17.323 15.982 21.945 217.054
PES 15.771 223.769 20.081 22.215
PI 15.755 222.774 26.911 23.596
MANUS 15.291 223.778 3.439 20.559
CLS 13.875 18.455 212.458 10.399
FOOT 25.774 23.522 14.393 4.789
Eigenvalues 11,504.461 3,708.648 1,191.546 1,004.96
Total variance
explained (%)
59.667 19.235 6.18 5.212
Abbreviations of variables are explained in Figures 2 and 3.
557MYOSORICINE LOCOMOTORY ADAPTATIONS
Journal of Morphology
TABLE 6. Proportional contribution of the bones of the forelimb and hind limbs relative to their respective total lengths, calculated
from mean values
Forelimb length
Mean (mm) % HB Humerus Radius Metacarpal
Proximal
phalanx
Middle
phalanx
Distal
phalanx Claw
Myosorex varius 27.3 32 35 36 12 6 4 6 12
Myosorex blarina 27.3 37 36 35 12 6 4 6 13
Surdisorex norae 30.6 35 33 34 12 6 4 10 17
Cryptotis parvus 17.5 30 35 36 12 7 5 4 8
C. merriami 20.7 30 37 35 12 7 4 4 7
C. lacertosus 24.9 30 34 36 12 7 4 8 13
U. soricipes 25.4 36 31 40 11 7 4 5 9
N. gibbsii 21.5 31 31 37 8 6 4 13 19
Hind limb length
Mean (mm) % HB Femur Tibia Metatarsal
Proximal
phalanx
Middle
phalanx
Distal
phalanx Claw
Myosorex varius 36.9 44 30 45 13 5 3 4 7
Myosorex blarina 35.6 48 31 44 12 5 3 5 8
Surdisorex norae 40.0 46 31 42 13 5 3 6 8
Cryptotis parvus 24.0 41 30 44 14 6 3 3 6
C. merriami 27.8 40 31 44 13 6 3 3 5
C. lacertosus 31.7 39 30 44 12 5 3 5 8
U. soricipes 36.3 51 25 46 15 7 3 4 7
N. gibbsii 31.8 45 26 44 13 7 3 7 11
Comparable values for Cryptotis are from Woodman and Gaffney (2014). Except for mean forelimb length and mean hind limb
length (mm), all numbers are percentages. %HB is limb length as a proportion of the head-and-body length.
TABLE 7. Mean lengths of individual elements of ray III of the manus and pes relative to the total length of the ray including the
claw
Metacarpal/
Metatarsal
Proximal
phalanx
Middle
phalanx
Distal
phalanx Claw
Manus
M. cafer 38 21 14 14 27
M. geata 37 20 13 15 30
M. kihaulei 36 19 13 15 31
C. phillipsorum 38 21 12 17 28
M. varius 35 18 13 18 34
M. blarina 35 17 12 18 36
M. zinki 34 17 12 18 38
S. polulus 31 15 11 26 43
S. norae 31 15 10 26 44
Cryptotis parvus 38 22 14 12 26
C. merriami 37 23 13 11 21
C. lacertosus 31 18 12 22 39
U. soricipes 35 23 13 16 29
N. gibbsii 21 17 11 33 51
Pes
M. cafer 48 20 12 12 21
M. geata 47 18 11 15 24
M. kihaulei 47 18 11 14 24
C. phillipsorum 46 19 10 14 24
M. varius 47 18 10 15 25
M. blarina 44 17 11 17 28
M. zinki 46 18 11 15 25
S. polulus 44 18 10 18 28
S. norae 43 18 10 20 29
C. parvus 47 20 11 10 20
C. merriami 47 22 12 9 19
C. lacertosus 44 18 11 17 27
U. soricipes 42 22 11 15 26
N. gibbsii 38 21 10 21 32
Values for Cryptotis are from Woodman and Gaffney (2014). All numbers are percentages.
558 N. WOODMAN AND F.A. STABILE
Journal of Morphology
scores on the scale of mean percentile ranks; more
semifossorial species had higher scores.
Forelimb versus hind limb. Adaptations for
fossoriality in mammals can include reductions in
the overall lengths of the limbs, with the hind
limbs shortening at a greater rate than the fore-
limbs (Shimer, 1903; Samuels and Van Valken-
burgh, 2008). This does not appear to be the
general pattern in mysoricines, which show no
clear trend (Table 2: IM, HFI, FOOT; Table 6).
Moreover, the myosoricine humerus typically
becomes shorter relative to the femur (HFI) in
more semifossorial species, a pattern that is also
seen in talpids (Reed, 1951).
Scratch-digging fossorial mammals typically
have measurably enlarged foreclaws and underly-
ing distal phalanges (Hildebrand, 1985b). The
hind claws also are generally larger but not to the
same extent as the foreclaws (Shimer, 1903; Reed,
1951). This pattern holds for myosoricines as well
(Table 2: CLAW, CLI).
Forelimb. Unlike rodents (Samuels and Van
Valkenburgh, 2008) and talpids (Table 6), there is
no clear pattern of forelimb reduction among the
few myosoricines for which we could obtain com-
plete measurements. Although the humerus and
radius may shorten slightly, the proportional
lengths of many of the bones contributing to the
forelimb are remarkably stable both in myosori-
cines and soricines (Table 6). The most variable
elements are the claw and the distal phalanx,
which, as noted previously, increase in length with
increasing semifossoriality.
Scapula. Reed (1951) described the elongation
of the scapula among semifossorial and fossorial
talpids. More semifossorial species of myosoricines
also have longer scapulae than more ambulatory
species (Table 2: SHI). One exception is C. phillip-
sorum, which has a longer scapula than would be
expected based on its mean percentile ranking.
Humerus. The radius becomes shorter relative
to the humerus in more fossorial rodents and tal-
pids (Samuels and Van Valkenburgh, 2008). No
such pattern emerged among myosoricines or sori-
cines (Table 2: BI; Table 6).
The length of the deltopectoral crest of the
humerus increases with increasing semifossorial-
ity among rodents and talpids (Table 2: SMI) as
the size and power of the deltoid and pectoral
muscles increase (Samuels and Van Valkenburgh,
2008). With the exception of C. phillipsorum,
which has a long crest for a short-clawed shrew,
myosoricines exhibit a similar pattern of lengthen-
ing deltopectoral crest with increasing
semifossoriality.
Among mammals, it is common for more semi-
fossorial species to have more robust limb bones,
including humeri (Hildebrand, 1985b; Samuels
and Van Valkenburgh, 2008; Woodman and Gaff-
ney, 2014). With the exception of C. phillipsorum,
which has a more robust humerus than expected
based on its mean percentile rank, the humerus
increases in breadth with increasing semifossorial-
ity among myosoricines (Table 2: HRI).
In soricines, the teres tubercle is visibly longer
(Table 2: HTI) and more distally positioned along
the humerus (TTP) with increasing semifossor-
iality (Fig. 1; Woodman and Gaffney, 2014). With
the exception of C. phillipsorum, which has an
unexpectedly long tubercle, both indices increase
with increasing semifossoriality among
myosoricines.
The relative breadth of the distal end of the
humerus typically increases with increasing semi-
fossoriality among mammals (Hildebrand, 1985b),
and this is the pattern among myosoricines (Table
2: HEB), with the exception of C. phillipsorum,
which has a broader distal end of the humerus
than expected.
Radius. The mammalian radius is generally
more robust in more semifossorial species, and
there is a similar pattern among the few myosori-
cines for which radii were available to measure
(Table 2: RDW).
Ulna. Semifossorial and fossorial mammals
generally have a longer olecranon process and a
shorter functional arm of the ulna, resulting in
greater force being transmitted. This increase in
force is indicated by larger OLI, TMO, TCO, and
OCI (Reed, 1951; Price, 1993; Samuels and Van
Valkenburgh, 2008; Woodman and Gaffney, 2014).
We had only three species of myosoricines with
complete ulnae, but they exhibited an increase in
all four indices with increasing semifossoriality
(Table 2). Surdisorex norae is notable because it
has a higher mean OLI than those calculated for
soricine shrews, semifossorial rodents, and the
semifossorial mole Neurotrichus, suggesting that it
can bring more power to bear when digging than
any of those other mammals.
As with other bones of the forelimb, the mam-
malian ulna increases in robustness with increas-
ing semifossoriality and fossoriality, and this
pattern holds for myosoricines (Table 2: URI).
Manus. Among myosoricines, distal phalanges
(%DPL) and claws of the manus (%CL; Tables 2
and 7; Fig. 6) increase in length with increased
semifossoriality. The distal phalanx lengthens at a
faster rate than the claw (%CLS), which provides
more semifossorial species with greater underlying
support for the claw (Woodman and Gaffney, 2014;
Woodman and Stabile, 2015). An exception is C.
phillipsorum, which has a relatively short claw,
but a long distal phalanx, resulting in a higher
measure of support than would be expected based
on the claws and the distal phalanges of other
mysoricines.
In contrast to the distal phalanges and claws,
the metacarpals and proximal and middle pha-
langes shorten with increasing semifossoriality
559MYOSORICINE LOCOMOTORY ADAPTATIONS
Journal of Morphology
(Table 7; Fig. 6A). Moreover, the proximal and
middle phalanges shorten at a higher rate than
the metacarpal, resulting in decreases in the
indices PI and MANUS (Table 2). This is oppo-
site of the trend in talpids, in which the meta-
carpal shortens much more than the proximal
and middle phalanges with increasing semifos-
soriality (Table 2; Reed, 1951). As in other
groups of mammals (Hildebrand, 1985b), the
breadths of the bones of the manus (MW3)
increase with increasing semifossoriality in
myosoricines.
Hind limb. Like the forelimb, there is no
clear pattern of hind limb reduction or increase
with locomotory mode among the few myosori-
cines for which we could obtain complete meas-
urements (Table 6). The proportional lengths of
most individual elements of the hind limb remain
relatively stable, although the tibiofibula (CI;
Table 2) and metatarsal III (PES) exhibit some
shortening relative to the femur with increased
semifossoriality, a pattern also seen in rodents
(Samuels and Van Valkenburgh, 2008). The most
variable elements are the claw and the distal
phalanx, which increase in length with increasing
semifossoriality, but the increases are much lower
in magnitude than for the same elements of the
forelimb.
Femur. As for other groups of mammals
(Samuels and Van Valkenburgh, 2008; Woodman
and Gaffney, 2014; Table 2), the breadths of the
diaphysis (Table 2: FRI) and the distal end of the
femur (FEB) generally increase with increasing
semifossoriality among myosoricines.
Tibiofibula. The breadth of the distal tibiofib-
ular articulation with the astragalus shows no
clear pattern among myosoricines (Table 2: DTA),
unlike in talpids and soricines (Woodman and
Gaffney, 2014).
Pes. In general, morphological changes in
the bones of the myosoricine pes mirror those
for the manus but at lower magnitude. The
metatarsals and proximal and middle phalanges
generally shorten with increasing semifossorial-
ity (Fig. 6B), whereas distal phalanges
(%hDPL) and claws (%hCL) lengthen (Table 7).
The distal phalanx increases in length at a
faster rate than the claw (%hCLS), thereby pro-
viding greater support for the claw in more
semifossorial species. An exception is M. zinki,
which has a longer metatarsal and shorter dis-
tal phalanx and claw than expected based on
the relative lengths of these elements on the
forefoot (Fig. 6B).
General Morphological Trends
Morphological variation associated with locomo-
tory behavior in myosoricines takes the form of a
graded series in which individual elements of the
skeleton lengthen or shorten and become more
robust with greater adaptation for semifossoriality.
In summarizing this variation, we treated the var-
iation as a continuum and addressed changes in
morphology as a unidirectional transition from
ambulatory to semifossorial. No complete phylog-
eny for the Myosoricinae currently exists, so it is
difficult to understand locomotory adaptations
within a true evolutionary framework. Regardless,
it is important to keep in mind that rather than
representing a single lineage sampled at discrete
intervals, the variation we describe is present at
the extant tips of genetically divergent lineages,
each of which has some period of independent evo-
lutionary history (Querouil et al., 2001; Willows-
Munro and Matthee, 2009, 2011). The discrete
morphological gradation we document suggests
that adaptation overrides phylogeny in our
Fig. 6. Plots of (A) mean proportional lengths of the metacar-
pal (mc), proximal phalanx (pp), middle phalanx (mp), distal
phalanx (dp), and protruding claw (cl) of ray III of the manus;
and (B) mean proportional lengths of the metatarsal (mt), proxi-
mal phalanx (pp), middle phalanx (mp), distal phalanx (dp), and
protruding claw (cl) of ray III of the pes. Lengths of individual
bones and claws are relative to the complete length of the ray
including the claw. The protruding claw is that part of the claw
that extends beyond the tip of the distal phalanx.
560 N. WOODMAN AND F.A. STABILE
Journal of Morphology
dataset. If this proves to be so, it would indicate
that the appearance of gradation is illusory.
Because selection for ambulatory behavior is as
likely as selection for semifossorial behavior, it is
also important to consider that evolution may
occur in either direction.
In evolving toward a more fossorial mode of
life, myosoricines exhibit a number of morphologi-
cal changes that are typical adaptations for
scratch-digging mammals (Shimer, 1903; Reed,
1951; Hildebrand, 1985b; Samuels and Van Val-
kenburgh, 2008; Woodman and Gaffney, 2014).
The claws and underlying distal phalanges of the
manus and pes increase in length and breadth,
while the metacarpals, metatarsals, and proximal
and middle phalanges shorten and broaden.
These changes are more obvious on the forefoot
than on the hind foot. The diaphyses of the
humerus, radius, ulna, and femur become more
robust, and regions of the bones involved in the
origin and insertion of forelimb extensor and
flexor muscles increase in size, resulting in
enlarged bony processes. The olecranon process of
the ulna increases in length relative to the distal
functional arm.
Myosoricines also show morphological changes
that are unique to eulipotyphlans, soricids, and
their own subfamilial clade. Adaptations for
semifossoriality and fossoriality often include
reductions in the lengths of the limbs (Shimer,
1903), as occurs in rodents and moles (Reed,
1951; Samuels and Van Valkenburgh, 2008). A
shorter forelimb is a consequence of shortening
of the antebrachium to deliver greater leverage
in digging, while a shorter hind limb results
from shortening of the tibiofibula and hind foot.
Shortening of the hind limb is typically greater
than shortening of the forelimb (Samuels and
Van Valkenburgh, 2008). Yet, there is no evi-
dence for limb shortening in myosoricines or
other soricids. This absence does not, however,
translate as complete stasis in pectoral girdle.
The myosoricine scapula generally increases in
length with increasing semifossoriality (as in
other soricids and talpids), and there are modest
decreases in the lengths of humerus and radius
(as in talpids). Like other groups of mammals,
increased fossoriality in myosoricines is tied to
noticeable changes in the morphology of the
humerus; it becomes more robust, the deltopec-
toral crest elongates, and the medial and lateral
epicondyles enlarge. In contrast to soricines, in
which the deltopectoral crest is less pronounced
proximally and is offset toward the lesser tuber-
osity (Reed, 1951; Woodman and Gaffney, 2014),
in myosoricines the crest forms a continuous
ridge that extends proximally to the greater
tuberosity. Although these differences were not
quantified, the myosoricine humerus also has a
more prominent trochlea and a longer medial
epicondyle than in soricines (compare to Reed,
1951; Woodman and Gaffney, 2014), two charac-
teristics also present in some lower Miocene
myosoricines (Klietmann et al., 2014:Fig. 6). A
teres tubercle is a feature of the humerus unique
to talpids, soricids, tachyglossids, and a few
early mammals (Hildebrand, 1985b; Martin,
2005; Woodman and Gaffney, 2014). As in other
soricids, this process increases in size and
becomes more distally positioned with increasing
semifossoriality, and it suggests that, like tal-
pids, some degree of humeral rotation is involved
in the digging stroke (Reed, 1951; Woodman and
Gaffney, 2014).
The unique and variable morphologies of the
myosoricine skeleton allow for predictions about
digging adaptations and point toward an ecological
radiation in behavior and substrate use. How the
sometimes subtle variation in morphology may be
affected by soil composition, prey availability, pre-
dation risk, and other ecological factors remains to
be determined.
Congosorex phillipsorum
Congosorex phillipsorum is distinctive among
modern Myosoricinae because it possesses a
seemingly contradictory combination of ambula-
tory and semifossorial characteristics. The spe-
cies has comparatively short claws, the bones of
the manus are not particularly broad, and it
ranks relatively low on both the mean percentile
ranks and PC1 scales, suggesting that it is a
more ambulatory species. Yet, a number of indi-
vidual characters point toward a more semifosso-
rial mode. It possesses a longer scapula than
would be expected for an ambulatory species and
the processes of the humerus are more devel-
oped. These latter include a long (but not more
distally positioned) teres tubercle, broad lateral
and medial epicondyles, and long deltopectoral
crest (Fig. 1). The diaphysis of the humerus (but
not the diaphysis of the femur) is also more
robust than would be expected for an ambula-
tory species. Although it bears short claws on its
forefeet, C. phillipsorum has moderately long
distal phalanges, resulting in a higher measure
of support for the claws than would be expected
based on the claws and distal phalanges of other
myosoricines. In contrast, the hind feet of C.
phillipsorum have moderately long claws and
short distal phalanges, resulting in a low level of
support for the claws. Overall, the skeletal char-
acteristics of C. phillipsorum suggest a unique,
but enigmatic, functional mode for this species.
The robust humerus has relatively strong mus-
cle attachment regions, but it is not obvious how
the relatively slender bones of manus would
bear the force generated by those muscles
(Woodman and Stabile, 2015). The claws are
561MYOSORICINE LOCOMOTORY ADAPTATIONS
Journal of Morphology
well-supported but short, suggesting either that
C. phillipsorum excavates in a novel substrate
or that its primary activity is not typical
scratch-digging. Unfortunately, a number of
potentially relevant bones, particularly the ulna,
are incomplete, making comprehensive study of
limb function impossible.
Recent molecular phylogenetic studies (Wil-
lows-Munro and Matthee, 2009; Taylor et al.,
2013) indicate that the genus Myosorex is para-
phyletic with respect to the genus Congosorex
and suggest a higher level of taxonomic diver-
sity within Myosorex than is currently recog-
nized. While this arrangement implies that
some of the morphological differences we docu-
ment in Myosorex may relate to older divergen-
ces than indicated by current taxonomy, it also
makes it more difficult to understand the
unique skeletal proportions of C. phillipsorum
as a consequence of phylogeny. The phylogenetic
positions of Congosorex and M. zinki as succes-
sive outgroups to a South African clade of Myo-
sorex (Taylor et al., 2013) suggest a complex
history of locomotory evolution within a Congo-
sorex-Myosorex clade.
ACKNOWLEDGMENTS
The authors thank R. Hutterer, who drew
attention to the interesting skeletal morphology
of Surdisorex and motivated us to begin our
study of the myosoricine postcranial skeleton.
The authors are grateful to W. T. Stanley and J.
C. Kerbis Peterhans for graciously permitting us
to examine recently-collected specimens of poorly
documented species that resulted from their field
work. The following curators and collections
managers permitted us to study specimens under
their care: N. B. Simmons, R. S. Voss, and E.
Westwig (AMNH); L. R. Heaney, B. D. Patterson,
and W. T. Stanley (FMNH). A. L. Gardner and
two anonymous reviewers provided valuable
comments on previous versions of this manu-
script. Any use of trade, product, or firm names
is for descriptive purposes only and does not
imply endorsement by the United States
government.
APPENDIX : SPECIMENS EXAMINED
Specimens examined for this study are deposited
in the following institutions: American Museum of
Natural History, New York (AMNH); Field
Museum of Natural History, Chicago (FMNH);
National Museum of Natural History, Washington
United States National Museum of Natural His-
tory (USNM). Symbols: *, no skeleton available; †,
manus and pes measurements could not be
obtained.
Soricidae:
Congosorex phillipsorum (n57). Tanzania: Iringa District
(FMNH 177683–177689).
Myosorex blarina (n53). Uganda: Kasese District (FMNH
144209, 144211). Zaire: Ruwenzori (FMNH 26285*).
Myosorex cafer (n5 2). South Africa: Kwazulu Natal Prov.
(FMNH 165585, 165587).
Myosorex geata (n5 12). Tanzania: Kilosa District (FMNH
166767, 166775, 166777, 197667, 197670–197673); Morogoro
District (FMNH 158299–158302); Mpwapwa District (FMNH
166767).
Myosorex kihaulei (n5 12). Tanzania: Kilombero District
(FMNH 155620–155622); Makete District (FMNH 204685,
204856, 204858, 204860, 204862); Rungwe District (FMNH
163552, 163554, 163558, 163559).
Myosorex varius (n5 5). South Africa: Kwazulu Natal Prov.
(FMNH 165588†, 165589–165592).
Myosorex zinki (n5 2). Tanzania: Kilimanjaro Region (FMNH
174117, 174119).
Surdisorex norae (n521). Kenya: Aberdare Mountains
(AMNH 187262; FMNH 190622–190624, 190625†, 190626;
USNM 182581*–182586*, 589811–589813, 589814†, 589815,
589816†, 589817, 589818†, 589819*).
Surdisorex polulus (n5 23). Kenya: Mount Kenya (USNM
163975*, 163976*, 163979*, 163981*, 163982*, 163984*,
163987*, 163989*–163991*, 163993*, 163996*–164000*,
164002*–164007*, 589820).
Talpidae:
Neurotrichus gibbsii gibbsii (n5 26). USA: Oregon (USNM
13410*, 65707*, 79788*, 80217*, 80437*–80441*, 89023*,
204484*–204487*, 264887*, 557433†–557435†,557437†,
557451†, 557460†–557463†, 560096†). Washington (USNM
273085†).
Uropsilus soricipes (n5 10). China: Sichuan (USNM 175142*,
256119, 260743*, 260751*, 574297, 574298*–574301*,
574302†).
LITERATURE CITED
Bell CJ, Mead JI. 2014. Not enough skeletons in the closet:
Collections-based anatomical research in an age of conserva-
tion conscience. Anat Rec 297:344–348.
Churchfield S. 1990. The Natural History of Shrews. Ithaca,
NY: Comstock Publishing Associates.
Coe MJ, Foster JB. 1972. The mammals of the northern slopes
of Mt. Kenya. J East Afr Nat Hist Soc Natl Mus 131:1–18.
Duncan P, Wrangham RW. 1971. On the ecology and distribu-
tion of subterranean insectivores in Kenya. J Zool 164:149–
163.
Edwards LF. 1937. Morphology of the forelimb of the mole (Sca-
lops aquaticus, L.) in relation to its fossorial habits. Ohio J
Sci 37:20–41.
Eisenberg JF. 1981. The Mammalian Radiations. Chicago, IL:
The University of Chicago Press.
Elissamburu A, De Santis L. 2011. Forelimb proportions and
fossorial adaptations in the scratch-digging rodent Ctenomys
(Caviomorpha). J Mammal 92:683–689.
Hildebrand M. 1985a. Walking and running. In: Hildebrand M,
Bramble DM, Liem KF, Wake DB, editors. Functional Verte-
brate Morphology. Cambridge, MA: Belknap Press. pp 38–57.
Hildebrand M. 1985b. Digging of quadrupeds. In: Hildebrand
M, Bramble DM, Liem KF, Wake DB, editors. Functional Ver-
tebrate Morphology. Cambridge, MA: Belknap Press. pp 89–
109.
Hopkins SSB, Davis EB. 2009. Quantitative morphological
proxies for fossoriality in small mammals. J Mammal 90:
1449–1460.
Hutterer R. 1985. Anatomical adaptations of shrews. Mamm
Rev 15:43–55.
562 N. WOODMAN AND F.A. STABILE
Journal of Morphology
Hutterer R. 2005. Order Soricomorpha. In: Wilson DE, Reeder
DM, editors. Mammal Species of the World: A Taxonomic and
Geographic Reference, 3rd ed. Baltimore, MD: The Johns
Hopkins University Press. pp 220–311.
Kerbis Peterhans JC, Hutterer R, Kaliba P, Mazibuko L. 2008.
First record of Myosorex (Mammalia: Soricidae) from Malawi
with description as a new species, Myosorex gnoskei. J East
Afr Nat Hist 97:19–32.
Kerbis Peterhans JC, Stanley WT, Hutterer R, Demos TC,
Agwanda B. 2009. A new species of Surdisorex Thomas, 1906
(Mammalia, Soricidae) from western Kenya. Bonn Zool Beit
56:175–183.
Kerbis Peterhans JC, Hutterer R, Mwanga J, Ndara B,
Davenport L, Karhagomba I, Udelhoven J. 2010. African
shrews endemic to the Albertine Rift: Two new species of
Myosorex (Mammalia: Soricidae) from Burundi and the Dem-
ocratic Republic of Congo. J East Afr Nat Hist 99:103–128.
Kerbis Peterhans JC, Huhndorf MH, Plumptre AJ, 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 (Mamma-
lia: Soricidae). Bonn Zool Bull 62:203–219.
Kirk EC, Lemelin P, Hamrick MW, Boyer DM, Bloch JI. 2008.
Intrinsic hand proportions of euarchontans and other mam-
mals: Implications for the locomotor behavior of plesiadapi-
forms. J Hum Evol 55:278–99.
Klietmann J, Nagel D, Rummel M, van den Hoek Ostende LW.
2014. Heterosorex and Soricidae (Eulipotyphla, Mammalia) of
the fissure Petersbuch 28; micro-evolution as indicator of
temporal mixing? C R Palevol 13:157–181.
Lemelin P. 1999. Morphological correlates of substrate use in
didelphid marsupials: Implications for primate origins. J Zool
247:165–175.
Martin, T. 2005. Postcranial anatomy of Haldanodon exspecta-
tus (Mammalia, Docodonta) from the Late Jurassic (Kimmer-
idgian) of Portugal and its bearing for mammalian evolution.
Zool J Linn Soc 145:219–248.
Meester J. 1953. The genera of African shrews. Ann Transvaal
Mus 22:205–214.
Meredith RW, Janecˇka JE, Gatesy J, Ryder OA, Fisher CA,
Teeling EC, Goodbla A, Eizirik E, Simao TL, Stadler T,
Rabosky DL, Honeycutt RL, Flynn JJ, Ingram CM, Steiner
C, Williams TL, Robinson TJ, Burk-Herrick A, Westerman M,
Ayoub NA, Springer MS, Murphy WJ. 2011. Impacts of the
Cretaceous terrestrial revolution and KPg extinction on mam-
mal diversification. Science 334:521–524.
Price MV. 1993. A functional–morphometric analysis of fore-
limbs in bipedal and quadrupedal heteromyid rodents. Biol J
Linn Soc 50:339–360.
Querouil S, Hutterer R, Barrie`re P, Colyn M, Kerbis Peterhans
JC, Verheyen E. 2001. Phylogeny and evolution of African
shrews (Mammalia: Soricidae) inferred from 16s rRNA
sequences. Mol Phylogenet Evol 20:185–195.
Reed CA. 1951. Locomotion and appendicular anatomy in three
soricoid insectivores. Am Midl Nat 45:513–670.
Samuels JX, Van Valkenburgh B. 2008. Skeletal indicators of
locomotor adaptations in living and extinct rodents.
J Morphol 269:1387–1411.
Sanchez-Villagra MR, Menke PR, Geisler JH. 2004. Patterns of
evolutionary transformation in the humerus of moles (Talpi-
dae, Mammalia): A character analysis. Mamm Study 29:163–
170.
Sargis EJ. 2002. Functional morphology of the forelimb of
tupaiids (Mammalia, Scandentia) and its phylogenetic impli-
cations. J Morphol 253:10–42.
Sargis EJ, Woodman N, Reese AT, Olson LE. 2013a. Using
hand proportions to test taxonomic boundaries within the
Tupaia glis species complex (Scandentia, Tupaiidae).
J Mammal 94:183–201.
Sargis EJ, Woodman N, Morningstar NC, Reese AT, Olson LE.
2013b. Morphological distinctiveness of Javan Tupaia hypo-
chrysa (Scandentia, Tupaiidae). J Mammal 94:938–947.
Shimer HW. 1903. Adaptations to aquatic, arboreal, fossorial
and cursorial habits in mammals. III. Fossoriality. Am Nat
37:819–825.
Stanley WT, Rogers MA, Hutterer R. 2005. A new species of
Congosorex from the Eastern Arc Mountains, Tanzania,
with significant biogeographical implications. J Zool 265:
269–280.
Stein BA. 2000. Morphology of subterranean rodents. In: Lacey
EA, Patton JL, Cameron GN, editors. Life Underground. The
Biology of Subterranean Rodents. Chicago, IL: The University
of Chicago Press. pp 19–61.
Taylor PJ, Kearney TC, Kerbis Peterhans JC, Baxter RM,
Willows-Munro S. 2013. Cryptic diversity in forest shrews of
the genus Myosorex from southern Africa, with the descrip-
tion of a new species and comments on Myosorex tenuis. Zool
J Linn Soc 169:881–902.
Vizcaino SF, Milne N. 2002. Structure and function in armadillo
limbs (Mammalia: Xenarthra: Dasypodidae). J Zool 257:117–
127.
Weisbecker V, Schmid S. 2007. Autopodial skeletal diversity in
hystricognath rodents: Functional and phylogenetic aspects.
Mamm Biol 72:27–44.
Weisbecker V, Warton DI. 2006. Evidence at hand: Diversity,
functional implications, and locomotor prediction in intrinsic
hand proportions of diprotodontian marsupials. J Morphol
267:1469–1485.
Willows-Munro S, Matthee CA. 2009. The evolution of the
southern African members of the shrew genus Myosorex:
Understanding the origin and diversity of a morphologically
cryptic group. Mol Phylogenet Evol 51:394–398.
Willows-Munro S, Matthee CA. 2011. Exploring the diversity
and molecular evolution of shrews (family Soricidae) using
mtDNA cytochrome b data. Afr Zool 46:246–262.
Woodman N, Gaffney SA. 2014. Can they dig it? Functional
morphology and degrees of semifossoriality among some
American shrews (Mammalia, Soricidae). J Morphol 275:745–
759.
Woodman N, Morgan JJP. 2005. Skeletal morphology of the
fore foot in shrews (Mammalia: Soricidae) of the genus
Cryptotis, as revealed by digital x-rays. J Morphol 266:60–
73.
Woodman N, Stabile FA. 2015. Variation in the myosoricine
hand skeleton and its implications for locomotory behavior
(Eulipotyphla: Soricidae). J Mamm 96.
Woodman N, Stephens RB. 2010. At the foot of the shrew:
Manus morphology distinguishes closely-related Cryptotis
goodwini and Cryptotis griseoventris (Mammalia, Soricidae)
in Central America. Biol J Linn Soc 99:118–134.
563MYOSORICINE LOCOMOTORY ADAPTATIONS
Journal of Morphology