Geometric morphometrics and cladistics: Testing evolutionary relationships in mega- and microbats

Abstract

Traditionally, morphometric data have consisted of distances, angles, or ratios, and have been considered inappropriate for cladistic analyses. Recently, geometric morphometrics, based on homologous landmark point-coordinates, has provided a number of advantages over traditional morphometric data and methods, including the possibility that phylogenetically informative characters and character-states may be extracted and used in cladistic analyses. Using two data sets of 3-dimensional point coordinates collected from skulls of bats, we empirically evaluate this possibility. Partial warps were extracted from the point-coordinate matrix, and these were then re-coded by gap-coding, for use in the cladistic analyses. In the case of samples from Eidolon helvum populations (two mainland localities and four islands in the Gulf of Guinea), analyzing males and females separately, our analyses based on these data were unable to detect consistent phylogeographic patterns among the populations. In the case of samples from plecotine bat species, these analyses produced a consensus cladogram showing considerable concordance with an earlier cladistic analysis by us of this group. In both cases, our results reflect those of earlier studies (based on both morphologic and genetic data), suggesting that the data and analytic techniques described herein may have interesting utility in cladistic analyses.

Full text

INTRODUCTION
It has been argued that continuous-state
characters should be excluded from cladis-
tic analysis mainly due to two reasons: (1)
these data are inappropriate (not phyloge-
netically informative), and (2) the methods
for their conversion into codes are arbitrary.
Pimentel and Riggins (1987) stated that
only character states derived from charac-
ters showing discrete variation can provide
phylogenetic information, whereas continu-
ous characters are transformational homolo-
gies not subject to test and, therefore invalid
for phylogenetic analysis. On the other
hand, more recently Rae (1998: 221) con-
cluded that “metric data [...] fulfill the sole
criterion for inclusion in phylogenetic
analysis, the presence of homologous char-
acter states, and thus cannot be excluded as
a class of data.” This position is also sup-
ported by Thiele (1993), who noted that
data scored for cladistic analyses may be
quantitative or qualitative, continuous or
discrete, and show overlapping or non-over-
lapping values between taxa.
Traditionally, morphometric research has
relied on statistical analyses of distances,
Acta Chiropterologica, 7(1): 39–49, 2005
PL ISSN 1508-1109 © Museum and Institute of Zoology PAS
Geometric morphometrics and cladistics: testing evolutionary
relationships in mega- and microbats
WIESºAW BOGDANOWICZ1, JAVIER JUSTE2, ROBERT D. OWEN3,and ANNA SZTENCEL1
1Museum and Institute of Zoology, Polish Academy of Sciences, 00-679 Warszawa, Poland
E-mail: wieslawb@miiz.waw.pl
2Estación Biológica de Doñana (CSIC), Avda. Mª Luisa s/n, Sevilla 41013, Spain
3Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409, USA
Traditionally, morphometric data have consisted of distances, angles, or ratios, and have been considered
inappropriate for cladistic analyses. Recently, geometric morphometrics, based on homologous landmark point-
coordinates, has provided a number of advantages over traditional morphometric data and methods, including
the possibility that phylogenetically informative characters and character-states may be extracted and used in
cladistic analyses. Using two data sets of 3-dimensional point coordinates collected from skulls of bats, we
empirically evaluate this possibility. Partial warps were extracted from the point-coordinate matrix, and these
were then re-coded by gap-coding, for use in the cladistic analyses. In the case of samples from Eidolon helvum
populations (two mainland localities and four islands in the Gulf of Guinea), analyzing males and females
separately, our analyses based on these data were unable to detect consistent phylogeographic patterns among
the populations. In the case of samples from plecotine bat species, these analyses produced a consensus
cladogram showing considerable concordance with an earlier cladistic analysis by us of this group. In both
cases, our results reflect those of earlier studies (based on both morphologic and genetic data), suggesting that
the data and analytic techniques described herein may have interesting utility in cladistic analyses.
Key words: geometric morphometrics, partial warps, gap-coding, phylogeny, Microchiroptera, Megachiroptera

angles or ratios to study quantitative vari-
ation among taxa. Recently, geometric
techniques have been developed for the di-
rect analysis of coordinates of homologous
landmarks. This geometric morphometrics
methodology has several advantages over
the traditional morphometrics (sensu Mar-
cus, 1993), e.g., by partitioning size from
shape for separate analyses and by provid-
ing a graphic way of locating and compar-
ing variability in different components of
shape among studied groups (e.g., Rohlf
and Marcus, 1993). Recently a number of
arguments have been made in favor of re-
garding certain morphometric variables as
putatively homologous characters and in-
cluding them, sometimes along with other
non-morphometric variables, in parsimony-
based cladistic analyses. Zelditch et al.
(1995; see also Fink and Zelditch, 1995 and
Swiderski et al., 1998; but cf. Adams and
Rosenberg, 1995; Naylor, 1996; and Mon-
teiro, 2000) have proposed the use of partial
warp-based traits in phylogenetic analysis.
These traits are derived from the shapes of
objects under study, as defined by the se-
lected number of x, y(and zin 3-dimension-
al analysis) coordinates of homologous
landmarks. The shapes are fitted to the ref-
erences by stretching/compressing and
shearing until complete identity of their
landmark configurations is achieved. Eigen-
vectors of the resulting bending-energy ma-
trix are defined as new shape variables,
principal warps which yield another shape
space with the origin defined by the refer-
ence. Projections of the shapes being com-
pared onto principal warps yield partial
warps, which warps are analogous to factor
analysis projections, with the eigenvectors
(principal warps) derived from a description
of non-uniform differences between ob-
served forms and some reference form
(O’Higgins, 2000). Principal warps, togeth-
er with the uniform component, also supply
a basis for the space in which we compute
relative warps, which are the same as prin-
cipal components (in particular, they are
mutually orthogonal). Both partial and rela-
tive warps can be used in many multivariate
statistical analyses as quantitative shape
variables. MacLeod (2002) illustrated some
of the weaknesses of partial warps using
empirical and simulated examples. Unfortu-
nately, the empirical trilobite example used
by this author is not a good test case be-
cause, as noted by Wagner (2000), trilobite
matrices are characterized by poor resolu-
tion of states.
The aim of our study was to re-evaluate
the usefulness of methods applying partial
warp analysis based on 3-dimensional in-
formation in recovering phylogenetical-
ly informative characters. This was tested
by producing trees derived from classical
(linear) parsimony of re-coded data, and
squared-change parsimony and continu-
ous maximum likelihood of continuous
characters, and comparing them with
a presumably known and well-established
phylogeny of two groups of bats belong-
ing to Megachiroptera and Microchiro-
ptera.
MATERIALS AND METHODS
Variables
For testing interspecific relationships among fruit
bats four taxa occurring in western Africa and with
well-understood phylogenetic relationships were se-
lected: Eidolon helvum, Rousettus aegyptiacus, Myo-
nycteris torquata and M. brachycephala. In addition,
the first species was also used in evaluation of evolu-
tionary ties among its populations in the islands of the
Gulf of Guinea. Eidolon helvum is the second largest
fruit bat in Africa, with unique morphological (An-
dersen, 1912), ecological (Thomas, 1983), and repro-
ductive (Bernard and Cumming, 1997) characteris-
tics. Juste et al. (2000) reported that of four island
populations examined (plus two mainland popula-
tions), the population from Annobón, the smallest and
farthest, shows remarkable morphological and genet-
ic differentiation, whereas the rest are similar phenet-
ically and with low genetic distances among them.
40 W. Bogdanowicz, J. Juste, R. D. Owen, and A. Sztencel

A set of 20 homologous cranial landmarks (see
Bookstein, 1991) was defined on bone sutures, foram-
ina, and inflection points along the edges of cranial
structures (Fig. 1) on the ventral view of skulls of the
four species of fruit bats (Appendix). Dental land-
marks were set on the bone (at the edge of the alveoli)
to avoid variation due to differential tooth-wear. To
facilitate repeatability, each landmark was gently
marked in pencil on the surface of the bone, under
20x magnification, before being recorded. Three-di-
mensional coordinates of landmarks were digitized
using a 3-D Reflex Microscope (Reflex Measurement
Ltd., Butleigh, Somerset BA6 8SP, UK). This is a
highly precise, non-contact instrument that uses a
small light spot to digitize coordinates in any position
within a magnified field. The microscope was period-
ically re-calibrated to ensure a linear scale error of
less than 30 µm over 100 mm in the x-axis. Land-
marks were collected under a 20× magnification lens
from four different aspects, thus the skull was rotated
three times to attain the complete data collection.
Amount of error due to the new digitizing of the four
reference landmarks was evaluated after each shift,
and the whole transformation was discarded, and the
digitizing process restarted, if the greatest error of any
of the four reference points was larger than 0.1 mm in
any direction. All these landmarks were recorded by
a single individual, and were taken without reference
to prior values.
The same instrument was applied to gather 19
three-dimensional coordinates from the dorsal side of
the skulls of microbats, including members of the
tribe Plecotini sensu stricto (genera Corynorhinus,
Plecotus, Barbastella, Euderma, Idionycteris) and
some other taxa (Otonycteris hemprichi, Antrozous
pallidus, and Eptesicus fuscus) within the subfamily
Vespertilioninae, plus Myotis lucifigus of the subfam-
ily Myotinae (sensu Hoofer and Van Den Bussche,
2003), which was used as an outgroup (see Appen-
dix). The plecotine bats represent the only supra-
generic group within Chiroptera that is Holarctic in
distribution, and there is considerable morphologic
and karyotypic evidence supporting their monophyly
(e.g., Frost and Timm, 1992; Tumlison and Douglas,
1992; Bogdanowicz et al., 1998; see also Juste et al.,
2004). Also in this case all landmarks were recorded
by a single individual.
Analyses
Geometric coordinates were checked and visual-
ized using Morphologika (O’Higgins and Jones,
1998). The average location of each landmark (if
necessary) was obtained through a non-documented
Geometric morphometrics and cladistics in mega- and microbats 41
FIG. 1. Location of landmarks on the skulls of Megachiroptera (left) and Microchiroptera (right)

option in the program allowing to save the coordi-
nates of the warped mean (P. O’Higgins, in litt.). The
NTSYS-pc ver. 2.11T (Rohlf, 2000) package was
used to calculate partial warps. The uniform compo-
nent was estimated by sweeping the partial warps
from the projections of the aligned coordinates into
the tangent space and then using SVD to extract the
non-singular dimensions (Rohlf and Bookstein,
2003). Partial warps (plus uniform components
scores) were re-coded using the gap weighting
method of Thiele (1993). A new character state Xnew
can be calculated with the following formula:
Xnew = n* [(x - min) / (max - min)],
where max and min are the maximum and minimum
mean value of the character across all species, x is the
mean value of the current taxon and nis the number
of allowed character states.
Phylogenetic analyses were performed with
PAST (Hammer et al., 2001) and PHYLIP (Felsen-
stein, 2004b). The first taxon was always (but inten-
tionally) treated as the outgroup. In the case of fruit
bats the branch-and-bound algorithm was applied.
For the Plecotini project we used heuristic search,
with the subtree pruning and regrafting option. This
algorithm is similar to the nearest neighbour inter-
change but with a more elaborate branch swapping
scheme. The character optimization was based on the
Wagner criterion, assuming that characters are re-
versible and ordered, meaning that 0- > 2 costs more
than 0- > 1, but has the same cost as 2- > 0.
RESULTS
Fruit Bats
In the cladogram (Fig. 2A) representing
interspecific relationships among these
megabats Rousettus aegyptiacus figures as
next most basal to the outgroup, followed
by the Myonycteris clade, which contains
M. brachycephala and M. torquata.
In the cladograms representing intraspe-
cific relationships of the E. helvum popu-
lations, Rousettus is seen as basal, with
a clade of the two Myonycteris species
the next most basal (Fig. 2B and 2C). The
relationships among the Eidolon popu-
lations are figured differently, depend-
ing on whether based on the male (tree of
936 steps, ensemble consistency index, CI
= 0.76) or female shortest tree (1,002 steps,
CI = 0.79) data. In the cladogram based on
42 W. Bogdanowicz, J. Juste, R. D. Owen, and A. Sztencel
FIG. 2. Interspecific relationships (A) among four
species of fruit bats derived from the maximum
parsimony of re-coded partial warp scores (strict
consensus tree). Interpopulation affinities on the basis
of two equally long cladograms in E. helvum are also
shown (B, XX; C, YY). Numbers below branches
indicate the bootstrap support values (percentage) for
the same nodes selected in the topologies obtained
under the evolutionary model of maximum
parsimony (after 100 iterations and 500 reorderings).
Please note that in each case only one possible
orientation to the reference is taken into account

female shape characteristics (Fig. 2B), pop-
ulations from Annobón and Principe are
sister taxa, forming a sister group to the
remaining four populations studied. Within
these four, Río Muni is basal, followed
by Bioko, which is sister to the Cameroon/
São Tome clade. In the male-based clado-
gram (Fig. 2C), Principe is again in the
most basal group, but is sister to Río Muni.
Together these two populations are sister
to the remaining four, among which São
Tome is most basal, followed by Bioko,
which is sister to Cameroon and Anno-
bón. Thus, there is little or no concordance
to be found between the interpopula-
tional cladograms based on the female
and male shape characteristics, and neither
represents a recognizable geographic pat-
tern.
Microbats
The three shortest trees were 1,110 steps
long and had an ensemble CI of 0.70. The
strict consensus (Fig. 3A) reflects a high de-
gree of concordance among the three clado-
grams, which are in agreement concerning
the placement of the non-plecotine taxa
Geometric morphometrics and cladistics in mega- and microbats 43
FIG. 3. Relationships among bats belonging to the tribe Plecotini: A — consensus of three most parsimonious
trees from the analysis of re-coded partial warps; B — the most parsimonious tree based on morphological and
karyotypic evidence (Bogdanowicz et al., 1998); C — extracted super-tree of Plecotini (Jones et al., 2002);
D — cladogram for Plecotini based on ca. 2.6 kb pairs of mitochondrial DNA sequence (after Hoofer and Van
Den Bussche (2003), but with only relationships supported strongly by either or both Bayesian and Parsimony
analyses being depicted

44 W. Bogdanowicz, J. Juste, R. D. Owen, and A. Sztencel
Otonycteris,Eptesicus, and Antrozous (out-
groups treated analytically as ingroups).
The ‘correct’ arrangement of these taxa sug-
gests that our characters: (a) carry phyloge-
netic information at this taxonomic level,
and (b) are correctly polarized.
Within the plecotine taxa, the genus Ple-
cotus is sister to all other taxa, with P. aus-
triacus and P. kolombatovici sisters, with
P. auritus a sister to that clade. The remain-
ing taxa are figured as an unresolved poly-
tomy among the genera Euderma, Idiony-
cteris, Corynorhinus (2 species), and Bar-
bastella (2 species). (The majority consen-
sus of the three trees agrees with this, but
shows Euderma and Idionycteris as sister
species, and an unresolved trichotomy
among this clade, Corynorhinus, and Bar-
bastella).
DISCUSSION
Interspecific and Intrapopulation Rela-
tionships of Fruit Bats
The resulting phylogenetic hypothesis at
the species level was consistent with the
previously established arrangements of
fruitbats based jointly on sequences derived
from the mitochondrial cytochrome b and
16S rRNA genes of a wide representation of
Megachiroptera (Álvarez et al., 1999; Juste
et al., 1999; see also Juste et al., 1997), de-
spite outstanding differences in the evolu-
tionary relationships among fruitbats sug-
gested by molecular and ’classical‘ morpho-
logical data (Springer et al., 1995; Kirsch
and Lapointe 1997). On the other hand, it is
not surprising because only a four-taxon hy-
pothesis was tested using re-coded partial
warps, including two taxa (M. brachycepha-
la and M. torquata) belonging to one genus
(Romagnoli and Springer, 2000).
The situation is much more complicated
in the case of evolutionary affinities of
a single species — Eidolon helvum — on
the islands of the Gulf of Guinea (Central
Africa). An Eidolon ancestor probably
reached mainland Africa independently
from other fruitbats, and this colonization
likely took place far earlier than the Late-
Pliocene date (3 Myr) of its only fossil
(Howell and Coppens, 1974) and maybe
even earlier than the other African coloniza-
tions (Juste et al., 1999). A recent phyloge-
netic study suggests that Eidolon’s origin
may be closer to the typically Asian Ptero-
pus group than to any extant African fruit
bat (Juste et al., 1999).
In ‘classical’ multivariate morphology,
the populations of E. helvum from the is-
lands of Bioko, Príncipe, and São Tomé
do not show significant phenetic differenti-
ation, although a trend towards a reduction
of size is found in the latter two islands (Fig.
4). In terms of allozyme variation, the low
genetic distances among these populations,
as well as their values of Wright’s fixation
indexes, suggest that gene flow has ham-
pered differentiation on these islands (Juste
et al., 1999). Only the fourth insular pop-
ulation, from Annobón, was characterized
by such remarkable morphological and ge-
netic differentiation that it has been accord-
ed the status of a separate subspecies:
E. helvum annobonensis. Although we
might expect better or more consistent reso-
lution among these island populations based
on re-coded partial warps, it appears that
resolution of our approach mirrors that of
more traditional approaches, and is unable
to discern phylogeographic patterns among
the Eidolon populations. Still, the partial
warp analysis is able to tell apart the mor-
phological differences specific of the fe-
males of the population from Annobón.
Interspecific Relationships within Plecotini
A comparison of our strict consensus
tree with the pertinent portion of the
’supertree‘ of Jones et al. (2002) is of
considerable interest (see Fig. 3A and 3C),
in that if the rooting is disregarded, the

Geometric morphometrics and cladistics in mega- and microbats 45
FIG. 4. The geographic situation in the Gulf of Guinea, Central West Africa (based on Juste et al., 2000). The
neighbour-joining topologies based on selected morphological characters, by sex, built on Mahalanobis D2
distances between populations of E. helvum from the Gulf of Guinea, according to Juste et al. (2000), are also
presented
topologies are concordant (albeit with dif-
fering resolution). Considering only the ple-
cotine taxa, and configuring the topology of
Jones et al. (2002) to reflect a primary di-
chotomy of Plecotus as sister to all other
plecotine taxa (as in our consensus tree),
their tree would figure Corynhorhinus as
sister to a clade containing Barbastella and
Euderma/Idionycteris. Although our con-
sensus tree figures these last three clades as
an unresolved trichotomy, the topologies
are concordant. Thus, the difference be-
tween our phylogenetic hypothesis for the
plecotines, and that of Jones et al. (2002), is
one of polarization, or more precisely, the
ordering of character-state transformation
series. In fact our consensus tree is more
likely correct than Jones et al. (2002), be-
cause it is very difficult to postulate that
the Euderma/Idionycteris clade is basal to
a clade containing both Plecotus (Old
World) and Corynorhinus (New World). If
we assume that the Plecotini originated in
the Old World, their phylogeny requires ei-
ther: (a) two dispersals from Old to New
World, or (b) one dispersal from Old to New
World, followed by a dispersal from New to
Old World; whereas our phylogeny could be
resolved to require only one dispersal from
Old to New World (if we allow Euderma
and Idionycteris to be sisters, as in our strict
consensus tree, and further resolve the tri-
chotomy to allow the Euderma/Idionycteris
clade to be sister to Corynorhinus), which is
reasonable on biogeographic grounds.
We also note that the pertinent portion of
the tree presented by Hoofer and Van Den
Bussche (2003) is concordant with our con-
sensus tree, though their data were unable to
provide much resolution at this level, or
even demonstrate that the Plecotini are
monophyletic. They did agree, however, on

the sister-group relationship between Eu-
derma and Idionycteris.
Geometric Morphometrics, Partial Warps
and Phylogeny
In the present study, in the case of
samples from plecotine bat species, these
analyses produced a reasonable consensus
cladogram showing considerable concor-
dance with an earlier cladistic analysis by us
of this group, and our results reflect those of
earlier studies (based on both morphologic
and genetic data; e.g., Frost and Timm,
1992, and Bogdanowicz et al., 1998). The
most fundamental problem in using tradi-
tional morphometric shape variables in a
standard cladistic analysis is the fact that it
requires the use of Manhattan distances (it
is implied by the use of linear parsimony).
Geometric morphometrics yields variables
corresponding to an arbitrary rotation of
shape space (depending in part on the refer-
ence configuration orientation). This arbi-
trariness does not matter in morphometrics
because the multivariate methods used in
morphometrics are invariant to rotation.
Manhattan distances are not and thus one
must select a particular rotation as being es-
pecially meaningful. Gap coding applied
one variable at a time is also a problem be-
cause the results depend on the arbitrary ro-
tation of the space (e.g., a different orienta-
tion of the reference configuration will yield
different shape variables that will be coded
differently by gap coding) (reviewed in
MacLeod and Forey, 2002; see also Felsen-
stein, 2004a). The squared change parsimo-
ny does not have this problem so that would
be quite compatible with morphometric data
since the solution is invariant to rotation of
the data. Continuous maximum likelihood
methods are also compatible with morpho-
metric data (F. J. Rohlf, in litt.). Neverthe-
less, at least as far as complex morphologi-
cal structures are concerned, such models
have probably no biological meaning. Mo-
reover, despite forcing continuous data into
integer codes only the classical, linear parsi-
mony of re-coded partials warps revealed
considerable logical phylogenetic configu-
ration for a 13-taxon example of Plecotini,
contrary to the methods utilizing continuous
data. It appears that a period of active ex-
perimentation with these methods is now
needed to further explore their appropriate-
ness and compatibility. An approach worth
investigating in this context is presented by
Bookstein (2000, 2002).
ACKNOWLEDGEMENTS
The following institutions are acknowledged for
allowing us to examine specimens under their care
and their curators and/or collection managers for fa-
cilitating the loan or examination at each collection:
The American Museum of Natural History, New
York, USA (Nancy Simmons); The Carnegie Museum
of Natural History, Pittsburgh, USA (Suzanne
McLaren); The Estación Biológica de Doñana, Se-
villa, Spain (José Cabot); The Croatian Natural His-
tory Museum, Zagreb, Croatia (Igor Pavliniƒ); The
Harrison Institute, Sevenoaks, United Kingdom (Paul
J. J. Bates); Hungarian Museum of Natural History,
Budapest, Hungary (Gábor Csorba); Mammal Re-
search Institute, Polish Academy of Science, Bia»o-
wieóa, Poland (Elwira Szuma); The Museum of
Southwestern Biology, Albuquerque, New Mexico,
USA (William Gannon); The Museum of Texas Tech
University, Lubbock, USA(Robert Baker); The Royal
Ontario Museum, Toronto, Canada (Judith Eger); and
The Zoologisches Forschungsinstitut und Museum
Alexander Koenig, Bonn, Germany (Reiner Hutte-
rer). The senior author was supported by the State
Committee for Scientific Research (Komitet Bada½
Naukowych) – grant no. 6 P04C 07718. The Reflex
Microscope used to measure the fruitbats was kindly
loaned to RDO by Richard Cifelli, University of
Oklahoma, USA. Dr. F. J. Rohlf made a substantial
contribution to this paper by providing detailed com-
ments. In fairness to Rohlf, it should be noted that he
does not agree with the primary conclusions of this
paper.
LITERATURE CITED
ADAMS, D. C., and M. C. ROSENBERG. 1995. Partial
warps, ontogeny, and phylogeny: a comment on
46 W. Bogdanowicz, J. Juste, R. D. Owen, and A. Sztencel

Zelditch and Fink. Systematic Biology, 47:
168–173.
ÁLVAREZ, Y., JUSTE, J., E. TABARÉS, A. GARRIDO-
PERTIERRA, C. IBÁÑEZ, and and J. M. BAUTISTA.
1999. Molecular phylogeny and morphological
homoplasy in fruitbats. Molecular Biology and
Evolution, 16: 1061–1067.
ANDERSEN, K. 1912. Catalog of the Chiroptera of
the British Museum. I. Megachiroptera. British
Museum (Natural History), London, 854 pp.
BERNARD, R. T. F., and G. S. CUMMING. 1997. Af-
rican fruitbats: evolution of reproductive patterns
and delays. Quarterly Review of Biology, 72:
253–274.
BOGDANOWICZ, W., S. KASPER, and R. D. OWEN.
1998. Phylogeny of plecotine bats: reevaluation
of morphological and chromosomal data. Journal
of Mammalogy, 79: 78–90.
BOOKSTEIN, F. L. 1991. Morphometric tools for land-
mark data. Geometry and biology. Cambridge
University Press, New York, 435 pp.
BOOKSTEIN, F. L. 2000. Creases as local features of
deformation grids. Medical Image Analysis, 4:
93–110.
BOOKSTEIN F. L. 2002. Creases as morphometric char-
acters. Pp. 139–174, in Morphometrics, shape,
and phylogeny (N. MACLEOD and P. FOREY, eds.).
Taylor and Francis, London, 308 pp.
FINK, W. L., and M. L. ZELDITCH. 1995. Phylogenetic
analysis of ontogenetic shape transformations: a
reassessment of the Piranha genus Pygocentrus.
Systematic Biology, 44: 343–360.
FELSENSTEIN, J. 2004a. Inferring phylogenies. Sinauer
Associates, Sunderland, Massachusetts.
FELSENSTEIN, J. 2004b. PHYLIP — Phylogeny Infer-
ence Package, version 3.6. University of Wash-
ington, Seattle (program).
FROST, D. R., and R. M. TIMM. 1992. Phylogeny of
plecotine bats (Chiroptera: ‘Vespertilionidae’):
summary of the evidence and proposal of a logi-
cally consistent taxonomy. American Museum
Novitates, 3034: 1–16.
HAMMER, Ø., D. A. T. HARPER, and P. D. RYAN , 2001.
PAST: Paleontological statistics software pack-
age for education and data analysis. Palaeonto-
logia Electronica, 4(1).
HOOFER, S., and R. A. VAN DEN BUSSCHE. 2003. Mo-
lecular phylogenetics of the chiropteran family
Vespertilionidae. Acta Chiropterologica, 5 (supp-
lement): 1–63.
HOWELL, F. C., and Y. COPPENS. 1974. Eidolon sp.
from the late Pliocene of Ethiopia. Comptes
Rendus de l’Académie des Sciences Paris, 278:
2421–2427.
JONES, K. E., A. PURVIS, A. MACLARNON, O. R. P.
BININDA-EMONDS, and N. B. SIMMONS. 2002. A
phylogenetic supertree of the bats (Mammalia:
Chiroptera). Biological Reviews, 77: 223–259.
JUSTE, B., J., C. IBÁÑEZ, and A. MACHORDÓN. 1997.
Evolutionary relationships among the African
fruit bats: Rousettus egyptiacus,R. angolensis,
and Myonycteris sp. Journal of Mammalogy, 78:
766–774.
JUSTE, J., C. IBÁÑEZ, and A. MACHORDOM. 2000.
Morphological and allozyme variation of Eidolon
helvum (Mammalia: Megachiroptera) in the is-
lands of the Gulf of Guinea. Biological Journal of
the Linnean Society, 71: 359–378.
JUSTE, J., Y. ÁLVAREZ, E. TABARÉS, A. GARRIDO-
PERTIERRA, C. IBÁÑEZ, and and J. M. BAUTISTA.
1999. Phylogeography of African fruitbats (Me-
gachiroptera). Molecular Phylogenetics and Evo-
lution, 13: 596–604.
JUSTE, J., C. IBÁÑEZ, D. TRUJILLO, J. MUÑOZ, J., P.
BENDA, and M. RUEDI. 2004. Mitochondrial phy-
logeography of the long-eared bats (Plecotus) in
the Mediterranean Palaearctic and Atlantic Is-
lands. Molecular Phylogenetics and Evolution,
31: 1114–1126.
KIRSCH, J. A., and F. J. LAPOINTE. 1997. You aren’t
(always) what you eat: evolution of nectar-feed-
ing among old world fruitbats (Megachiroptera:
Pteropodidae). Pp. 313–330, in Molecular evolu-
tion and adaptive radiation (T. J. GIVNISH and K.
J. SYTSMA, eds.). Cambridge University Press,
New York, 621 pp.
MACLEOD, N. 2002. Phylogenetic signals in mor-
phometrics data. Pp. 100–138, in Morphomet-
rics, shape, and phylogeny (N. MACLEOD and
P. FOREY, eds.). Taylor and Francis, London,
308 pp.
MACLEOD, N., and P. FOREY (eds.). 2002. Morpho-
metrics, shape, and phylogeny. Taylor and Fran-
cis, London, 308 pp.
MARCUS, L. F. 1993. Some aspects of multivariate sta-
tistics for morphometrics. Pp. 96–130, in Con-
tributions to morphometrics, Vol. 8 (L. F. MAR-
CUS, E. BELLO, and A. GARCIA-VALDECASAS,
eds.). Museo Nacional de Ciencias Naturales
(CSIC), Madrid, 264 pp.
MONTEIRO, L. R. 2000. Why morphometrics is spe-
cial: the problem with using partial warps as
characters for phylogenetic inference. Systematic
Biology, 49: 796–800.
NAYLOR, G. P. 1996. Can partial warps scores be used
as cladistic characters? Pp. 519–530, in Advances
in morphometrics (L. F. MARCUS, M. CORTI, A.
LOY, G. J. P. NAYLOR, and D. SLICE, eds.). NATO
ASI Series A, Vol. 284. Plenum Press, New York,
587 pp.
Geometric morphometrics and cladistics in mega- and microbats 47

O’HIGGINS, P. 2000. The study of morphological vari-
ation in the hominid fossil record: biology, land-
marks and geometry. Journal of Anatomy, 197:
103–120.
O’HIGGINS, P., and N. JONES. 1998. Morphologica
Software. University College London (program).
PIMENTEL, R. A., and R. RIGGINS. 1987. The nature of
cladistic data. Cladistics, 3: 201–209.
RAE, T. C. 1998. The logical basis for the use of
continuous characters in phylogenetic systemat-
ics. Cladistics, 14: 221–228.
ROHLF, F. J. 2000. NTSYS-pc numerical taxonomy
and multivariate analysis system, version 2.11T.
Exeter Software, Setauket.
ROHLF, F. J., and F. L. BOOKSTEIN. 2003. Computing
the uniform component of shape variation. Sys-
tematic Biology, 52: 66–69.
ROHLF, F. J., and L. MARCUS. 1993. A revolution in
morphometrics. Trends in Ecology and Evolu-
tion, 8: 129–132.
ROMAGNOLI, M. L., and M. S. SPRINGER. 2000.
Evolutionary relationships among Old World
fruitbats (Megachiroptera: Pteropodidae) based
on 12S rRNA, tRNA valine, and 16S rRNA gene
sequences. Journal of Mammalian Evolution, 7:
259–284.
SPRINGER, M. S., L. J. HOLLAR, and J. A. KIRSCH.
1995. Phylogeny, molecules versus morphology,
and rates of character evolution among fruitbats
(Chiroptera: Megachiroptera). Australian Journal
of Zoology, 43: 557–582.
SWIDERSKI, D. L., M. L. ZELDITCH, and W. L. FINK.
1998. Why morphometrics is not special: coding
quantitative data for phylogenetic analysis. Sys-
tematic Biology, 47: 508–519.
THIELE, K. 1993. The Holy Grail of the perfect char-
acter: the cladistic treatment of morphometric da-
ta. Cladistics, 9: 275–304.
THOMAS, D. W. 1983. The annual migrations of three
species of West African fruit bats (Chiroptera:
Pteropodidae). Canadian Journal of Zoology, 61:
2266–2272.
WAGNER, P. J. 2000b. Exhaustion of cladistic
character states among fossil taxa. Evolution, 54:
365–386.
TUMLISON, R., and M. E. DOUGLAS. 1992. Parsimony
analysis and the phylogeny of the plecotine bats
(Chiroptera: Vespertilionidae). Journal of Mam-
malogy, 73: 276–285.
ZELDITCH, M. L., W. L. FINK, and D. L SWIDERSKI.
1995. Morphometrics, homology, and phyloge-
netics: quantified characters as synapomorphies.
Systematic Biology, 44: 179–189.
48 W. Bogdanowicz, J. Juste, R. D. Owen, and A. Sztencel
APPENDIX
Specimens used in the present study. Acronyms: AMNH: American Museum of Natural History, New York,
USA; Acronyms: CM — Croatian Natural History Museum, Zagreb, Croatia; CMNH – Carnegie Museum of
Natural History, Pittsburgh, USA; EBD – Estación Biológica de Doñana, Spain; HNHM – Hungarian Museum
of Natural History, Budapest, Hungary; HZM – Harrison Institute, Sevenoaks, United Kingdom; MRI –
Mammal Research Institute, Polish Academy of Science, Bia»owieóa, Poland; MSB – Museum of Southwestern
Biology, Albuquerque, New Mexico, USA; ROM – Royal Ontario Museum, Toronto, Canada; TTU – Texas
Tech University, Lubbock, USA; ZFMK – Zoologisches Forschungsinstitut und Museum Alexander Koenig,
Bonn, Germany. Letters after species number refer to geographic location in the Gulf of Guinea: a – Annobón;
b – Bioko; ca – Cameroon; p – Principe; rm – Río Muni; st – São Tome
Received 18 June 2004, accepted 22 December 2004
Megachiroptera:
Eidolon helvum, XX(n = 44): CMNH40990 – ca;
EBD17386 – st; EBD17387 – st; EBD17388 – st;
EBD17520 – st; EBD17540 – st; EBD17598 – a;
EBD17600 – a; EBD17603 – a; EBD18488 – a;
EBD18825 – b; EBD18826 – b; EBD18827 – b;
EBD18840 – b; EBD18874 – a; EBD18875 – a;
EBD18876 – a; EBD18897 – rm; EBD18898 – rm;
EBD18899 – b; EBD18901 – b; EBD18907 – a;
EBD18913 – st; EBD19044 – b; EBD19064 – b;
EBD19217 – rm; zBD19218 – rm; EBD19224 – rm;
EBD19225 – rm; EBD19226 – rm; EBD19227 – rm;
EBD19238 – rm; EBD20348 – st; EBD20358 – st;
EBD20362 – st; EBD20448 – st; EBD8818 – rm;
EBD8820 – rm; ROM50865 – ca ROM54940 – ca;
ROM55626 – ca; ROM55627 – ca; ZFMK73368 –

Geometric morphometrics and cladistics in mega- and microbats 49
ca; ZFMK74302 – ca; YY (n= 68): EBD17382 – p;
EBD17389 – st; EBD17471 – st; EBD17519 – p;
EBD17538 – p; EBD17539 – p; EBD17542 – st;
EBD17595 – rm; EBD17597 – a; EBD17599 – a;
EBD17601 – a; EBD17602 – a; EBD18199 – p;
EBD18204 – st; EBD18212 – st; EBD18487 – a;
EBD18489 – a; EBD18492 – a; EBD18494 – a;
EBD18828 – b; EBD18829 – b; EBD18833 – b;
EBD18834 – p; EBD18836 – p; EBD18837 – p;
EBD18844 – rm; EBD18869 – p; EBD18879 – b;
EBD18900 – b; EBD18906 – a; EBD18915 – st;
EBD18917 – p; EBD18958 – a; EBD19020 – rm;
EBD19037 – b; EBD19038 – rm; EBD19039 – rm;
EBD19053 – b; EBD19054 – b; EBD19065 – b;
EBD19214 – rm; EBD19215 – rm; EBD19216 – rm;
EBD19220 – rm; EBD19221 – rm; EBD19222 – rm;
EBD19239 – rm; EBD19247 – rm; EBD20344 – st;
EBD20345 – p; EBD20346 – p; EBD20347 – p;
EBD20349 – st; EBD20350 – p; EBD20351 – p;
EBD20352 – p; ; EBD20355 – st; EBD20357 – p; ;
EBD20359 – p; EBD20364 – p; EBD8819 – rm;
EBD8821 – rm; HZM21.4835 – ca; ROM39044 – ca;
ROM39058 – ca; ZFMK64328 – b; ZFMK64331 – b;
ZFMK64333 – b.
Myonycteris brachycephala, XX (n= 9): EBD
18934, 18935, 17410, 17524, 17470, 18904, 17477,
17468, 18936; YY (n= 8): EBD17413, 22284,
17469, 17411, 17526, 18937, 19066, 37413.
Myonycteris torquata, XX (n = 42): AMNH
236237, 236239, 236240, 236246, 236247, 236249,
236254, 236255, 236256, 240999, 241000, 241001,
241002, 241003, 241004; CMNH107996, 40951,
40957; EBD13767, 15046, 15057, 15061, 15065,
15066, 15110, 19058, 22486, 22488, 22489, 22493,
22501, 22503; ROM39393, 43356, 57148, 69001;
TTM17954; TTU17952, 3938, 3939; USNM241112;
ZFMK61621; YY (n = 53): AMNH236236, 236242,
236243, 236245, 236250, 236251, 236252, 236253;
BM139122; CMNH40949, 40950, 40952, 40953,
40954, 40955, 40956, 58253; EBD13888, 15011,
15012, 15013, 15054, 15055, 15056, 15058, 15059,
15060, 15062, 15063, 15727, 17715, 19057, 20487,
20488, 20489, 22487, 22490, 22495, 22497, 22498,
22499, 22504, 22505, 22506, 22507, 22508, 22509,
22510; SNH511901; TTU3937, 3940; ZFMK61623,
69609.
Rousettus aegyptiacus, XX (n= 96): AMNH
240988, 240990; CMNH58254; EBD13869, 15152,
15191, 15680, 17352, 17396, 17405, 17406, 17496,
17497, 17533, 17534, 17535, 18203, 18207, 18241,
18250, 18525, 18550, 18557, 18558, 18559, 18841,
18856, 18857, 18858, 18859, 18860, 18861, 18863,
18864, 18865, 18866, 18873, 18918, 18938, 18939,
18952, 19021, 19029, 19030, 19031, 19032, 19033,
19049, 19050, 19052, 19056, 19059, 19267, 19268,
19270, 19276, 19277, 19278, 20178, 20179, 20181,
20183, 20185, 20188, 20190, 20191, 20193, 20196,
20197, 22293, 22307, 22313, 223311; ROM43348,
46756, 55663, 55698, 55699, 55700, 55701, 55702,
55704, 55707, 55732, 55733, 55735, 55737, 56227;
TTM17963; TTU17956, 17962; ZFMK444084,
64312, 64313, 64315, 64319; YY (n = 97): AMNH
240977, 240989, 240996, 318299; CMNH3928; EBD
13679, 15192, 15462, 15463, 17397, 17398, 17521,
17531, 17536, 17597, 17597, 17599, 17601, 17602,
17673, 18208, 18209, 18213, 18242, 18252, 18487,
18489, 18492, 18494, 18830, 18832, 18850, 18851,
18854, 18855, 18865, 18867, 18868, 18870, 18880,
18881, 18891, 18902, 18902, 18906, 18921, 18953,
18954, 18958, 18976, 18978, 18979, 19006, 19007,
19034, 19051, 19060, 19061, 19074, 19269, 19271,
19272, 19273, 19279, 19280, 20180, 20184, 20186,
20195, 20663, 22294, 22308, 22309, 22310, 22312;
HZM504098; ROM46755, 55693, 55694, 55885,
56207, 56226, 56250, 56251, 56255, 58309, 58340;
TTM17958; TTU3923, 3924, 17961; ZFMK64309,
64310, 64314, 64316, 64317, 64318.
Microchiroptera:
Antrozous pallidus, XX (n= 3): ROM67106,
67107, 67111; YY (n= 3): ROM67110, 67121,
78631.
Barbastellus barbastellus, XX (n= 3): HNHM
53.29.1, 58.65.1, 2000.43.4; YY (n= 4): HNHM
57.56.1., 57.110.1, 70.18.1; MRI85469.
Barbastella leucomelas, YY (n= 1): HNHM
2743.3.
Corynorhinus mexicanus, YY (n= 2): AMNH
203933, 203934.
Corynorhinus rafinesquii, XX (n= 3): AMNH
142010, 142003, 166892; YY (n= 3): AMNH
142005, 142011, 142006.
Eptesicus fuscus, XX (n= 3): ROM41705,
43176, 46037; YY (n= 3): ROM19388, 24448,
54179.
Euderma maculatum, XX (n= 1): MSB96066.
Idionycteris phyllotis, XX (n= 1): AMNH
178893 ; YY (n= 1) AMNH185341.
Myotis lucifugus, XX (n = 3): ROM43605,
88947, 88951; YY (n= 3): ROM40122, 43595,
43594.
Otonycteris hemprichi, XX (n= 1): HZM1122.
Plecotus auritus, XX (n= 3): MRI10638, 49960,
57090; YY (n= 3): MRI85094, 91305, 91306.
Plecotus austriacus, XX (n= 3): MRI130/91303,
7868, 96981; YY (n= 3) MRI12410, 38248, 96982.
Plecotus kolombatovici, XX (n= 3): CM3004,
3006, 3008; YY (n= 2): CM2152, 3054.