Figures
Figure 1: Definition of three-dimensional fixed landmarks (red) and
semi-landmarks (yellow) set on the cranium and humerus of crocodile
newts (genera Tylototriton and Echinotriton ) for the
geometric morphometrics analysis of SD. Shown are the cranium of SMF1134
(male T. verrucosus ) and humerus of ZSM0830-2012 (female T.
himalayanus) derived from mCT data.
Figure 2: Morphospace of humerus shape of crocodile newts
(Tylototriton spp. and Echinotriton andersoni ) build by
the first PC axes of a PCA of 56 GPA-aligned 3D landmarks. Shape changes
at the minimum and maximum of each axis are presented as TPS-deformed
grids from the mean shape. Colour code correspond to setting in
subsequent figures.
Figure 3: Morphospace of cranial shape of crocodile newts
(Tylototriton spp. and Echinotriton andersoni ) build by
the first PC axes of a PCA of 45 GPA-aligned 3D landmarks. Shape changes
at the minimum and maximum of each axis are presented as TPS-deformed
grids from the mean shape. For colour coding see Figure 2.
Figure 4: Unique allometry in humerus (a) and cranium (b) shape of nine
species of crocodile newts (Tylototriton spp. andEchinotriton andersoni ) estimated by multivariate regression.
Shape changes to the mean shape are presented as TPS-deformation grids
for the largest (upper) and smallest (lower) fitted value. For colour
coding see Figure 2.
Figure 5: Common allometry in humerus (a) and cranium (b) shape of
crocodile newts (Tylototriton spp. and Echinotriton
andersoni ) of females and males estimated by multivariate regression.
Shape changes to the mean shape are presented as TPS-deformation grids
for the largest (upper) and smallest (lower) fitted value. The shape
changes for the humerus are magnified by the factor of three. For colour
coding see Figure 2.
Figure 6: TPS-deformation grids from the mean shape (reference) to the
different mating modes (target: circle dance, amplexus) of crocodile
newts (Tylototriton spp. and Echinotriton andersoni ) of
the humerus (upper rows) and cranial shape (lower rows). The shape
changes are magnified by the factor of three.
Figure 7: Trajectory analysis of SD in humerus shape of crocodile newt
(Tylototriton spp. and Echinotriton andersoni ) for whole
data set (upper left) and for mean shape predictions for each sex and
species (upper right). TPS-deformation grids of two exemplary species
with different trajectories are illustrated in the lower rows. Those
trajectories are marked in the upper graphs by an arrow indicating the
direction from male to female. Shape changes for those are shown from
male (reference) to females (target). The shape changes are magnified by
the factor of two. For colour coding see Figure 2.
Figure 8: Trajectory analysis of SD in cranial shape of crocodile newts
(Tylototriton spp. and Echinotriton andersoni ) for whole
data set (upper left) and for mean shape predictions for each sex and
species (upper right). In the lower row TPS-deformation grids of two
exemplary species with different trajectories are illustrated. Those
trajectories are marked in the upper graphs by an arrow indicating their
directions. Shape changes for those are shown from male (reference) to
females (target). The shape changes are magnified by the factor of
three. For colour coding see Figure 2.
Figure 9: TPS-deformation grids from the mean shape (reference) to male
and female shapes (target) of Tylototriton himalayanus (amplexus)
and T. kweichowensis (circle dancer) for the humerus (upper rows)
and cranium (lower rows). The selected species represent different
sexual dimorphism-trajectories and different mating modes. The shape
changes are magnified by the factor of two for the humerus and by the
factor of three for the cranium.
Data accessibility: Raw-data (3D geometric morphometric
landmarks) is accessible under
https://www.doi.org/10.6084/m9.figshare.14381315 [upon
acceptance].
Competing interests: The authors have no conflicts of interest
to declare.
Author contributions: PP – study design, specimen loans,
CT-scanning, data collection, data analysis, data interpretation,
drafting manuscript. MZ – CT-scanning, data collection. TB –
CT-scanning. AK – study design, specimen loans, supervision. All
authors edited and approved the manuscript draft.
Acknowledgments : Collection-based research of PP was partly
funded by the Wilhelm-Peters-Fonds of the Deutsche Gesellschaft für
Herpetologie und Terrarienkunde e.V. (DGHT), funding number WP-01/2017.
Further, PP received funding for travel costs by the PROMOS program of
the DAAD and the Reinhold-und-Maria-Teufel-Stiftung, Tuttlingen. We
thank all curators and collection managers of natural history museums
for granting access to salamanders in their care: Raffael Ernst and
Markus Auer (MTKD, Dresden), Gunter Köhler and Linda Acker (SMF,
Frankfurt a. Main), Mark-Oliver Rödel and Frank Tillak (ZMB, Berlin),
Frank Glaw and Michael Franzen (ZSM, Munich), Dennis Rödder and Morris
Flecks (ZFMK, Bonn), Silke Schweiger and Georg Gassner (NHMW, Vienna),
Annemarie Ohler (MNHN, Paris), Lauren Scheinberg (CAS, Berkeley, CA) and
Kevin de Queiroz and Esther Langan (USNM, Suitland, MD). The
geomorph-community is thanked for much help with the R script. Further
thanks go to Katharina Foerster and James Nebelsick being always a
source of good scientific advice. This work represents a contribution to
obtain the PhD degree of PP at the University of Tuebingen.