Figure 1.
Undeformed cast of NHML ZD.1961.776 in (a) anterior and (b) inferior views.
Figure 2.
Blue landmarks indicate those eliminated for analyses using the artificially deformed crania, red landmarks indicate those eliminated for analyses of Paradolichopithecus, and black landmarks indicate those eliminated in all analyses.
Table 1.
List of landmarks used in these analyses.
Table 2.
List of semilandmark curves used in these analyses.
Table 3.
Number of trials per user used to generate the range of variability around intraobserver error.
Table 4.
Comparative sample of papionins used in these analyses.
Figure 3.
ISER VGr/345 in (a) lateral, (b) superior, (c) basal and (d) anterior views.
(a) in approximate Frankfurt horizontal; (b-d) occlusal plane horizontal.
Figure 4.
Comparison of the original cranium (left column), deformed cranium 1 (second column), reflected & averaged cranium 1 (third column) and algorithmically symmetrized cranium 1 (right column) in anterior (top), lateral (middle) and basal (bottom) views.
Reflected & averaged specimens do not appear perfectly symmetrical as only bilateral landmark points were used in this computation, rather than semilandmark curves or patches.
Figure 5.
Histogram illustrating the distribution of pairwise Procrustes distances within each group of Papio.
The dashed line represents the mean within-group pairwise distance for all groups.
Figure 6.
Histogram illustrating the distribution of all pairwise Procrustes distances within each cercopithecid group.
Dashed line represents the mean intraspecific pairwise Procrustes distance for all cercopithecids.
Figure 7.
Histogram illustrating the distribution of pairwise Procrustes distances in a study of intraobserver error.
Table 5.
Procrustes distances between the original undeformed cranium and each of the five manual deformations (original to deformed column), as well as their modifications that were reflected & averaged bilaterally (original to reflected & averaged column), and algorithmically symmetrized model (original-retrodeformed).
Figure 8.
Comparison of the original cranium (left column), deformed cranium 2 (second column), reflected & averaged cranium 2 (third column) and algorithmically symmetrized cranium 2 (right column) in anterior (top), lateral (middle) and basal (bottom) views.
Reflected & averaged specimens do not appear perfectly symmetrical as only bilateral landmark points were used in this computation, rather than semilandmark curves or patches.
Figure 9.
Comparison of the original cranium (left column), deformed cranium 3 (second column), reflected & averaged cranium 3 (third column) and algorithmically symmetrized cranium 3 (right column) in anterior (top), lateral (middle) and basal (bottom) views.
Reflected & averaged specimens do not appear perfectly symmetrical as only bilateral landmark points were used in this computation, rather than semilandmark curves or patches.
Figure 10.
Comparison of the original cranium (left column), deformed cranium 4 (second column), reflected & averaged cranium 3 (third column) and algorithmically symmetrized cranium 4 (right column) in anterior (top), lateral (middle) and basal (bottom) views.
Reflected & averaged specimens do not appear perfectly symmetrical as only bilateral landmark points were used in this computation, rather than semilandmark curves or patches.
Figure 11.
Comparison of the original cranium (left column), deformed cranium 5 (second column), reflected & averaged cranium 5 (third column) and algorithmically symmetrized cranium 5 (right column) in anterior (top), lateral (middle) and basal (bottom) views.
Reflected & averaged specimens do not appear perfectly symmetrical as only bilateral landmark points were used in this computation, rather than semilandmark curves or patches.
Figure 12.
PCA of the Procrustes aligned coordinates for the original (star), deformed (squares), reflected & averaged (circles) and algorithmically symmetrized (triangles) crania.
Arrows connect the deformed to the reflected & averaged model, and the reflected & averaged model to the algorithmically symmetrized cranium. These arrows are for aid in visualization and do not represent real data. (a) PCA including both semilandmark curves and type I, II and III landmarks. PC1 accounts for 49% and PC 2 18% of the variance within this sample. (b) PCA of the Procrustes aligned coordinates including only types I-III landmarks. PC 1 accounts for 52% and PC 2 20% of the variance within this sample.
Figure 13.
PCA of the algorithmically symmetrized specimens with the full sample of Papio.
Landmarks 1–3, 13–15, and 24–27 were eliminated from the original dataset to accommodate the retrodeformed specimens. Specimens are labeled in the graph as per the key. Lines represent convex hulls surrounding each genus.
Figure 14.
Deformed and algorithmically symmetrized scans of ISER VGr/345 (Paradolichopithecus arvernensis) in (a) anterior, (b) lateral, (c) superior and (d) basal views.
Figure 15.
PCA of all extant cercopithecid taxa, Paradolichopithecus, and algorithmically symmetrized Paradolichopithecus.
Specimens are labeled in the graph as per the key. Lines represent convex hulls surrounding each genus.
Table 6.
Procrustes distances between the retrodeformed Paradolichopithecus specimens, the other Paradolichopithecus individuals and the extant taxon means; details about extant species in Table 4.