Figure 1.
Three-dimensional digital reconstruction of skeleton of Beelzebufo ampinga highlighting sources of material for reconstruction.
A, dorsal view; and B, right lateral view (with left limbs removed for visual clarity). Beelzebufo specimens used in model in dark blue. Light grey cranial and vertebral materials inferred from known morphology of Beelzebufo specimens, primarily through mirror-imaging. Dark grey jaws and postcranial elements modelled on large female specimen of Ceratophrys aurita (LACM 163430). See Supporting Information S1 for detailed description of model.
Figure 2.
Three-dimensional digital reconstruction of skeleton of Beelzebufo ampinga.
A, dorsal view; and B, right lateral view (with left limbs removed for visual clarity). See Figure 1 for sources of material for reconstruction, and Supporting Information S1 for detailed description of model.
Figure 3.
Three-dimensional digital reconstruction of skull of Beelzebufo ampinga.
A, anterior; B, posterior; C, dorsal; D, ventral; and E, right lateral views. Parts of posterior region of skull lack complete symmetry because respective sides use different combinations of specimens. See Supporting Information S1 for detailed description of model.
Figure 4.
Skull reconstructions of Beelzebufo ampinga.
A, dorsal view, as in Fig. 3C, with areas of digital model representing actual (non-mirrored) specimens in dark blue; B, dorsal view, illustrated reconstruction based on Fig. 3C, but with right side of 3C mirrored for symmetry and with missing regions silhouetted in grey. Shape of orbital, narial, and temporal fenestrae based on bone extrapolation from edges, facets, and other anatomical features. See Supporting Information S1 for detailed description of model.
Figure 5.
Three-dimensional digital reconstruction of skeleton of Beelzebufo ampinga highlighting specimen FMNH PR 2512.
Elements of skeleton of FMNH PR 2512, the most complete specimen discovered to date, highlighted in dark blue. Other Beelzebufo specimens in light blue. Light grey cranial and vertebral materials inferred from known morphology of Beelzebufo, created primarily through mirror-imaging. Dark grey postcranial elements and jaws modelled on large female specimen of Ceratophrys aurita (LACM 163430). See Supporting Information S1 for detailed description of model.
Figure 6.
Map of Mahajanga Basin study areas and stratigraphy (modified from Rogers et al., 2013: fig. 1).
The majority of specimens of Beelzebufo have been discovered in the Anembalemba Member of the Maevarano Formation in the Berivotra Study Area, but the taxon has also been recovered from the Masorobe Member in the Berivotra Study Area and the Lac Kinkony Member in the Lac Kinkony Study Area.
Figure 7.
Mahajanga Basin Project locality MAD98-25.
The image, taken in July 2007, shows a member of the field crew surface collecting at the locality from which the most complete specimen of Beelzebufo ampinga, FMNH PR 2512, was recovered during 11 expeditions between 1998 and 2011.
Figure 8.
FMNH PR 2512 cranial bone morphology.
A, detail of external exostosis on right squamosal; B, slices through posterior part of frontoparietal-braincase region showing details of thick laminar dermal bone overlying spongy endochondral bone; C, as B, but through occipital pillar.
Figure 9.
Stereophotographs of braincase and frontoparietal region of FMNH PR 2512 with labeled line drawings.
A, dorsal; B, ventral; C, anterior; D, posterior; E, left lateral; and F, right lateral views.
Figure 10.
Stereophotographs of right squamosal of FMNH PR 2512 with labeled line drawings.
A, dorsolateral; and B, ventromedial views.
Figure 11.
Stereophotographs of right quadratojugal-quadrate and pterygoid of FMNH PR 2512 with labeled line drawings.
A, dorsolateral; and B, ventromedial views.
Figure 12.
A, distal (lateral); B, dorsal; C, labial; D, lingual; and E, mesial (medial) views of left premaxilla, UA 9622. F, lingual view of right premaxilla, FMNH PR 1960.
Figure 13.
Reconstructions of right maxilla in labial view.
A, illustrated reconstruction based on composite digital model; B, outline reconstruction showing main specimens (FMNH PR 2507, FMNH PR 2510 [reversed], FMNH PR 2512 [reversed], UA 9635) in combined digital model. Additional data taken from neighbouring elements and positional information in skull reconstruction. See Figs 14, 15 for detailed views of individual specimens.
Figure 14.
A, labial; and B, lingual views of left maxilla, FMNH PR 2510. C, labial; and D, lingual views of right maxilla, UA 9635. E, labial; F, lingual; and G, dorsolingual views of right maxilla, FMNH PR 2499.
Figure 15.
A, labial; B, lingual; C, oblique lingual; D, dorsolingual; and E, posterolingual views of partial right maxilla that includes part of the pars facialis, FMNH PR 2507. F, labial; G, lingual; and H, dorsal views of fragmentary pars facialis of left maxilla, FMNH PR 2512.
Figure 16.
Anterior portion of left nasal.
A, dorsal; B, ventral; C, anterior; D, medial; and E, lateral views, UA 9951; F, ventral view, UA 9958.
Figure 17.
Posterior portion of right nasal (or anterior portion of left frontoparietal), UA 9615.
A, dorsal; B, ventral; C, lateral; and D, oblique ventral views. Specimen positioned as posterior portion of right nasal for cranial reconstruction in Figs. 1–5.
Figure 18.
A, dorsal; and B, ventromedial views of otic plate of left squamosal, FMNH PR 2512. C, ventromedial; D, ventral; and E, anteroventral views of right squamosal in articulation with lateral shelf of frontoparietal, FMNH PR 2512. F, dorsolateral; and G, ventromedial views of postquadrate flange formed from fused right squamosal and quadratojugal, FMNH PR 2536. H, lateral; I, ventromedial; and J, ventral views of part of small right squamosal, UA 9625.
Figure 19.
Quadratojugal/quadrate, FMNH PR 2512.
A, lateral; B, posterior; and C, medial views of left quadrate with part of conjoined quadratojugal. D, anterior; E, dorsolateral; and F, posteroventromedial views of right quadrate, quadratojugal, and pterygoid.
Figure 20.
A, dorsolateral; and B, lateral views of left quadratojugal, FMNH PR 2512. C, lateral; and D, medial views of right quadratojugal, FMNH PR 1959. E, lateral; F, medial; and G, dorsal views of left quadratojugal, UA 9639. H, lateral; I, medial; J, dorsolateral; K, anteroventromedial; and L, anteromedial views of right quadratojugal, UA 9621.
Figure 21.
Frontoparietal and braincase, FMNH PR 2512.
A, dorsal; and B, ventral views.
Figure 22.
Frontoparietal and braincase, FMNH PR 2512.
A, anterior; B, anteroventral; C, posterior; D, posteroventral; and E, oblique right posterolateral views.
Figure 23.
Frontoparietal and braincase, FMNH PR 2512.
A, left lateral; B, left anterolateral; and C, right lateral views.
Figure 24.
Approximately anterodorsal-posteroventral progression of μCT slices through braincase, FMNH PR 2512.
A, slice YZ 386; B, slice YZ 637; C, slice YZ 719; and D, slice YZ 992. Scan slices in YZ plane of reconstructed volume. Note that anteroposterior and dorsoventral biological axes deviate approximately 45° from scan reconstruction XZ and YZ axes.
Figure 25.
Approximately anteroventral-posterodorsal progression of μCT slices through braincase, FMNH PR 2512.
A, slice XZ 290; B, slice XZ 315; C, slice XZ 370; and D, slice XZ 425. Note that anteroposterior and dorsoventral biological axes deviate approximately 45° from scan reconstruction XZ and YZ axes.
Figure 26.
Mediolateral progression of μCT slices through right ear region, FMNH PR 2512.
A, slice XY 425; and B, slice XY 441, through lateral semicircular canal and otic chamber. C, slice XY 500, close to boundary between lateral semicircular canal and otic chamber. D, slice XY 640; and E, slice XY 665, including occipital canal. Scan slices in XY plane of reconstructed volume.
Figure 27.
Left frontoparietal and otoccipital, UA 9675.
A, posterior; B, anterior; C, medial; D, lateral; and E, ventrolateral views.
Figure 28.
Internal morphology of left frontoparietal and otoccipital, UA 9675.
A, medial; B, dorsolateral; and C, posterior views of digital segmentation of μCT dataset. Small opaque images at left for orientation; larger semi-transparent images at right, including occipital canal rendered in red and inner ear structures in blue.
Figure 29.
A, posterior; B, anterior; C, medial; and D, oblique medial views.
Figure 30.
Right angulosplenial, UA 8677.
A, dorsal; B, dorsolateral; and C, ventral views.
Figure 31.
Details of maxillary dentition.
A, lingual view of partial right maxilla showing tooth tip, FMNH PR 2506. B, lingual; and C, anterolingual views of partial right maxilla, UA 9945.
Figure 32.
Three-dimensional digital reconstruction of axial skeleton of Beelzebufo ampinga.
A, dorsal; B, ventral; and C, right lateral views of axial column. As in Fig. 1, with material of Beelzebufo ampinga in dark blue. Mirrored left portion of neural arch of fifth presacral vertebra in model (FMNH PR 2512 Vertebra B) and centrum and transverse process of sacral vertebra (FMNH PR 2003) are mirrored in light grey. Dark grey postcranial elements modelled on large female specimen of Ceratophrys aurita (LACM 163430). See Supporting Information S1 for detailed description of model.
Figure 33.
Atlas and second presacral vertebrae.
A, anterior; B, anterodorsal; C, posterior; D, dorsal; E, ventral; F, left lateral; and G, right lateral views, UA 9600 (holotype). H, anterior; and I, anterodorsal views of atlas vertebra, FMNH PR 2512. Note scanning artifacts on UA 9600 most easily traceable as horizontal lines in D and E, and as vertical lines in F and G.
Figure 34.
A, anterior; B, posterior; C, dorsal; D, ventral; E, left lateral; and F, right lateral views. UA 9947 interpreted as possible third presacral vertebra and placed in that position for digital reconstruction in Figs 1, 2, 5, and 32.
Figure 35.
Presacral Vertebra A, FMNH PR 2512.
A, anterior; B, posterior; C, dorsal; D, ventral; E, left lateral; and F, right lateral views. Vertebra A interpreted as possible fourth presacral vertebra and placed in that position for digital reconstruction in Figs 1, 2, 5, and 32.
Figure 36.
Presacral Vertebra B, FMNH PR 2512.
A, anterior; B, posterior; C, dorsal; and D, right lateral views. Vertebra B interpreted as possible fifth or sixth presacral vertebra and placed in fifth presacral position for digital reconstruction in Figs 1, 2, 5, and 32.
Figure 37.
A, anterior; B, posterior; C, dorsal; D, ventral; E, left lateral; and F, right lateral views. UA 9948 interpreted as sixth or seventh presacral vertebra.
Figure 38.
Sacral vertebra, FMNH PR 2300.
A, anterior; B, posterior; C, dorsal; and D, ventral views.
Figure 39.
A, dorsal; B, ventral; C, anterior; D, posterior; E, left lateral; and F, right lateral views, UA 9636. G, anterior view of less complete urostyle fragment, FMNH PR 2512, representing only the right cotyle.
Figure 40.
A, external; and B, internal views, UA 9620. C, external; and D, internal views, FMNH PR 2512.
Figure 41.
A, anterior; B, posterior; C, lateral; D, medial; and E, distal views.
Figure 42.
Right tibiale–fibulare, UA 9957.
A, anterior; B, posterior; C, proximal; D, distal; E, lateral; F, medial; and G, oblique distal views.
Figure 43.
Morphology–only (maximum parsimony) strict consensus of 63,108 most parsimonious trees using full matrix, rooted on Alytes obstetricans.
Numbers at nodes represent jackknife GC/Bremer values.
Figure 44.
Morphology–only Bayesian inference tree, rooted on Alytes obstetricans.
Numbers at nodes are posterior probabilities.
Figure 45.
Combined evidence (maximum parsimony) strict consensus of four most parsimonious trees, rooted on Ascaphus montanus.
Numbers at nodes represent jackknife GC/Bremer values.
Figure 46.
Combined evidence Bayesian Inference tree, rooted on Ascaphus montanus.
Note that as the Ascaphus species lay on a very long branch at the base of the tree, they have been omitted to reduce figure size. Numbers at nodes are posterior probabilities.
Figure 47.
Intraspecific size range of Beelzebufo ampinga.
Left squamosals of A, UA 9629; B, FMNH PR 2512 (reversed for comparison); and C, UA 9614, all in dorsal view. Skull silhouettes based on Fig. 4B and scaled by variation in size range of selected squamosals. Assuming isometric growth trajectory, individual represented by UA 9629 would have been about 20 percent larger than FMNH PR 2512, and that represented by UA 9614 about half the size of FMNH PR 2512.
Figure 48.
Intraspecific differences in pattern of bone growth.
A, dorsal; B, lateral; and C, ventrolateral comparisons of digital volume of relatively robust quadratojugal fragment UA 9639 (blue, at right) with that of quadrate-quadratojugal of FMNH PR 2512 (grey, mirror-imaged, at left). Integrated volumes (centre) show relatively greater medial and lateral development of bone growth in UA 9639, particularly in quadratojugal buttress.
Figure 49.
Combined evidence (maximum parsimony) tree showing placement of fossil genera discussed in text.
Arariphrynus (Early Cretaceous); Baurubatrachus, Uberabatrachus and Beelzebufo (Late Cretaceous); and Wawelia (Miocene) all fall within Nobleobatrachia. Alternative positions (as shown) were obtained for Early Cretaceous Cratia and Eurycephalella. All except Beelzebufo are from South America.