The holotype of
Crassigyrinus scoticus
Watson from the Viséan (Lower Carboniferous) of Edinburgh shows the side of the skull of a very primitive amphibian with fish-like proportions, an osteolepiform fish configuration of bones round the nostril and a preopercular bone on the cheek. ‘
Macromerium scoticum
’ Lydekker from the same locality and horizon proves to be a
Crassigyrinus
mandibular ramus. This is corroborated by discovery of a skull and anterior skeleton of Crassigyrinus from the Namurian (basal Upper Carboniferous) of Cowdenbeath, Fife. The skull of
Crassigyrinus
is also shown to have a loosely articulated basioccipital which did not form a finished occipital condyle and a mandible with coronoid teeth. However, it shares a number of derived (synapomorph) characters with the anthracosauroid amphibia of the Carboniferous and early Permian, notably the characteristic tabular horn, the probable absence of posttemporal fossae, the nature of the dermal ornament, the histology of the teeth and a true basipterygoid articulation. The last character may also indicate relations to loxommatid and seymouriamorph amphibia and amniotes. The pattern of bones of the
Crassigyrinus
skull table, however, is the primitive tetrapod (‘temnospondyl’) one. The postcranial skeleton is both primitive and degenerate. The vertebrae each have a single crescent-shaped centrum (‘intercentrum ’) and neural arches as poorly ossified, unfused bilateral halves. Prezygapophyses are unbuttressed facets and postzygapophyses totally lacking. There is room for a virtually unconstricted notochord. The diameter of the centra increases posteriorly from the small (partly reconstructed) atlas-axis complex. Ribs are long, well-ossified and cylindrical, but lack well-ossified rib-heads. The fore-limb is minute, with a typical primitive tetrapod humerus, which, however, retains some foramina otherwise seen only in
Ichthyostega
and fishes. The elongate ventral scales are unlike those of any ‘labyrinthodont’ amphibia. It is suggested that the apparent ‘otic notches’ of
Crassigyrinus
may mark the position of persistent spiracles, while the stapes, not preserved in any specimen, may have been like that known in the Coal Measure anthracosaurs and in the primitive temnospondyl
Greererpeton
. Combined with an air-filled spiracular cleft the stapes could have been tuned to underwater rather than aerial hearing.
Crassigyrinus
appears to have been a large
Amphiuma
-like underwater predator. A case is made for the ‘sister-group’ relation of
Crassigyrinus
to the anthracosauroids and a cladogram presented of the subgroups involved. It is, however, difficult to make a case for the close relationship of
Crassigyrinus
and the Seymouriamorpha and the closeness of relationship of the latter to anthracosauroids is questioned.
Crassigyrinus
shares several primitive characters with
Ichthyostega
, but they are only distantly related, so that the loss of those characters in all other tetrapods must have been polyphyletic. There are other characters in which one or the other is clearly the more primitive, but the polarity of a number of alternative character states in the two genera is equivocal. The cladistic use of out-group comparison is impotent to solve the problem because rival sister-groups for the Tetrapoda have been proposed using,
inter alia
, the disputed characters.
A morphometric analysis of Tobleria bicuspis, a Voltziales seed cone from the early Permian Jambi palaeoflora, Sumatra (Indonesia)
