Cranial ontogenetic variation in early saurischians and the role of heterochrony in the diversification of predatory dinosaurs Christian Foth1,2,3, Brandon P. Hedrick4, Martín D. Ezcurra2,5,6 1 SNBS, Bayerische Staatssammlung für Paläontologie und Geologie, Richard Wagner-Str. 10, D-80333 München
2 Department of Earth and Environmental Sciences, Ludwig-Maximilians-Universität, Richard-Wagner-Str. 10, D-80333 München, Germany
3 Department of Geosciences, University of Fribourg/Freiburg, Chemin du Musée 6, 1700 Fribourg, Switzerland
4 Department of Earth and Environmental Science, University of Pennsylvania, 251 Hayden Hall, 240 S 33rd Street, Philadelphia, PA 19104, USA
Table S1 List of landmarks and semi-landmark description. Semi-landmarks (semi-LMs) are marked with a ‘S’.
Most anterior point of the premaxilla along the tooth row (This point is reconstructed for Alioramus and Zupaysaurus).
Contact between premaxilla and maxilla along the tooth row.
Contact between the maxilla and jugal along the ventral margin of the skull
Contact between the nasal process of the premaxilla and the anterodorsal process of the nasal along the dorsal margin of the external naris
Most anterior point of the antorbital fossa (in those taxa without an antorbital fossa, the most anterior point of the antorbital fenestra was marked).
Most anterior point of the anterior process of the lacrimal along the dorsal margin of the antorbital fenestra.
Most ventral point of the ventral process of the lacrimal along the posteroventral margin of the antorbital fenestra.
Most posterior point of the jugal process of the maxilla along the ventral margin of the antorbital fenestra.
Most anterior point of the jugal (Depending on the configuration of maxilla, lacrimal and jugal, the landmarks 7, 8 and 9 can be similarly located).
Contact between the ventral process of the lacrimal and the jugal along the anteroventral margin of the orbit.
Contact between the jugal and the postorbital along the posterior margin of the orbit.
Most dorsal point of the anterior process of the postorbital along the posterodorsal margin of the orbit.
Contact between the jugal and the postorbital along the anterior margin of the lateral temporal fenestra.
Ventral contact between postorbital and squamosal along the dorsal margin of the lateral temporal fenestra.
Contact between the descending process of the squamosal with the quadratojugal along the posterior margin of the lateral temporal fenestra (For those taxa, where these bones do not contact to each other, e.g. the juvenile Coelophysis, the most anteroventral point of the descending process of the squamosal was marked.
Contact between jugal and quadratojugal along the ventral margin of the lateral temporal fenestra.
Dorsal contact between postorbital and squamosal.
Contact between the descending process of the squamosal with the quadratojugal along the posterior margin of the skull (For those taxa, where these bones do not contact to each other, e.g. the juvenile Coelophysis, the most posteroventral point of the descending process of the squamosal was marked.
Posteroventral corner of the quadratojugal.
Three semi-LMs on the ventral margin of the maxilla from LM 2 to LM 3.
Five semi-LMs along the narial margin starting and ending in LM 4.
Three semi-LMs along the anterior margin of the antorbital fenestra from LM 6 to LM 8.
One semi-LM along the posterior margin of the antorbital fenestra from LM 6 to LM 7.
One semi-LM along the ventral margin of the orbit from LM 10 to LM 11.
Four semi-LMs along the anterodorsal margin of the orbit from LM 10 to LM 12.
Three semi-LMs along the posterodorsal margin of the orbit from LM 12 to LM 11.
One semi-LM along the anterodorsal margin of the lateral temporal fenestra from LM 13 to LM 14.
Two semi-LMs along the posterodorsal margin of the lateral temporal fenestra from LM 14 to LM 15.
Two semi-LMs along the posteroventral margin of the lateral temporal fenestra from LM 15 to LM 16.
Two semi-LMs along the anteroventral margin of the lateral temporal fenestra from LM 16 to LM 13.
Two semi-LMs along the ventral margin of the skull from LM 3 to LM 19.
Two semi-LMs along the posterior margin of the quadratojugal from LM 19 to LM 18.
Most posterior point of the descending process of the maxilla contacting the nasal and/or the lacrimal.
Five semi-LMs along the anterodorsal margin of the maxilla from LM 2 to LM 51.
Fifteen semi-LMs along the dorsal margin of the skull from LM 1 to LM 18.
Figure S1Illustration of the landmarks and semi-landmarks positions on the skull of Sciurumimus albersdoerferi. Landmarks are shown as red dots, while semi-landmarks are marked with a ‘S’ and are shown as yellow dots.
2. Landmark error after Singleton (2002)
The percent error for digitizing landmarks and semi-landmarks by hand was estimated for the skull reconstruction of the juvenile Coelophysis (with n = 10 repetitions) using the method described by Singleton (2002). On the basis of the Procrustes coordinates the mean Procrustes distances to the respective consensus coordinates of each landmark were calculated. Then the relation of these distances to the mean distance of the consensus landmarks to the centroid of the consensus shape was calculated as a percentage of the former from the latter. Landmark and semi-landmark error varies between 0.117 percent (LM 51) and 0.738 percent (LM 3) with a mean of 0.283 percent, having no significant effect on shape analyses.
Table S2 Percent error for each landmark for the skull of the juvenile specimens of Coelophysis with n = 10.
3. Taxon sampling
Table S3List of taxa used in the present analyses with data of occurrences (in million of years, Myr) and sources of images. Bold taxa mark ontogenetic series used for the study.
Hurum & Sabbath, 2003; modified after Tsuihiji et al., 2011
Carr & Williamson, 2004
Brusatte et al., 2009
4. Comments on sliding semi-landmarks
To test the influence of sliding semi-landmarks during general Procrustes analysis (GPA) semi-landmarks were slid in tpsRelW (Rohlf, 2003) to minimize the bending energy and the Procrustes distance, both with a maximal iteration of ten. Afterwards, the resulting Procrustes shapes were compared with that of the regular GPA. The example at hand shows the results for the alignment for the skull of the juvenile Massospondylus. The minimization of the bending energy results on artificial elongation of the external naris and the formation of a frontal “crest”-like structure in front of the orbit, while the minimization of the Procrustes distance leads to deformation of the narial shape, the shape of the orbit and the shape of the skull roof in the postorbital region. As these artificial shape alignments affected several taxa within the data set (e.g. Carnotaurus, Compsognathus, Mamenchisaurus, Shunosaurus, Syntarsus), we decided to treat the semi-landmarks as landmarks. However, to appraise the influence of the semi-landmarks in the original data set, we repeated the shape analyses with a second data set containing only landmark information and compared it with the original results.