M. cornalianus possibly differs from the other species of the
genus in the development of this character: Maisch &
Matzke (1998) found that the dental groove was present in
the entire upper jaw in at least one specimen that they
prepared. This contradicts the descriptions given by Re-
possi (1902) and Besmer (1947), which Maisch & Matzke
(1997B) found unsubstantiated. However, the specimen
examined by Maisch & Matzke (1997B) also differed from
those of the previous authors in that there was no clear
heterodonty. It is possible that these characters are polymorphic
within the species, or that more than one species
is involved in the debate. Additional specimens need to be
prepared before definite conclusions can be drawn.
The second condition is exemplified by Shonisaurus,
which is unique among known ichthyopterygians for having
deep dental sockets. The bone of attachment is restricted
to the bottom of the socket, unlike in ankylosed theco-
donty. This condition, tentatively named ichthyosaurian
thecodonty by Motani (1997A), is solely based on a figure
given by Camp (1980: fig. 23). However, there is a high
possibility that this figure is not based on the original
Shonisaurus specimens from Nevada. McGowan & Motani
(1999) failed to locate any material resembling this figure
in the original specimens, whereas there is a specimen at
UCMP that closely resembles Camp’s (1980: fig. 23) illustration
(RM, pers. obs.). This Mexican specimen (UCMP
27141), also collected by Camp, was identified as Shastasau-
rus altispinus by Callaway & Massare (1989A) but was
later reassigned to Shonisaurus sp. by Motani (1999B). It is
possible that Camp (1980) was aware of the similarity between
the Nevada and Mexican forms, but did not have
time to incorporate this knowledge fully in the manuscript
for his 1980 monograph because of his deteriorating health.
Cymbospondylus has been said to have a similar tooth implantation
as described for Shonisaurus (Merriam, 1908),
but examination of the original specimens revealed that it
is impossible to determine tooth implantation without further
preparation. True thecodonty, in which there is no
bone of attachment (Romer, 1956; Edmund, 1960, 1969), is
not known for ichthyopterygians.
The infolding of dental roots, also referred to as plici-
dentine, is usually present in ichthyopterygians. This
structure has yet to be confirmed for Utatsusaurus, but was
already present in Grippidia (Mazin, 1981 A ) and in isolated
teeth from the Lower Triassic of Spitsbergen (Wiman,
1910; Mazin, 1981B). Mixosaurus cornalianus has been reported
to lack this structure (Besmer, 1947), but it is necessary
to scrutinize this description by examining the bottom
of the roots. It was once believed that plicidentine appeared
in ichthyopterygians along with aulacodonty to
compensate for the lack of firm attachment (Peyer, 1968).
However, the presence of plicidentine in the forms from
the Lower Triassic of Spitsbergen (W iman, 1910) falsifies
this hypothesis.
Tooth replacement in ichthyosaurs is known only for
two genera. Edmund (1960) showed for Ichthyosaurus that
replacement teeth entered the pulp cavities of their predecessors
at an early stage, as in iguanid lizards. In contrast,
Utatsusaurus seems to have had a varanid type replacement,
where replacement teeth never entered the pulp
cavities of their predecessors (Motani, 1996). An additional
specimen of Utatsusaurus described by Motani et al.
(1998) lends further support to this view: no resorption
cavity is present in numerous dissociated teeth (only three
of which were left in situ in the specimen). It is not known
when the transition from the varanid to the iguanid type of
tooth replacement occurred.
Pectoral girdle
(Fig. 70)
The endochondral pectoral girdle is similar in the basal
and mixosaurian plans: both the scapula and coracoid
have a fan-shaped expansion distally and a short and
narrow stem proximally. The proximal stems of the two
elements, which meet to form a glenoid, are thicker than
the distal expansions. Coracoids may be distinguished
from the scapulae in that the distal expansion is extended
anteriorly, resulting in a highly asymmetrical shape of the
entire element. This asymmetry of the coracoid was retained
in the stem and parvipelvian body plans; when
there is only one notch in the coracoid, it usually occurs
anteriorly. Cymbospondylus is the only exception to this
pattern, as discussed in the next paragraph. The articulation
between the coracoid and scapula is weak in the basal
and mixosaurian plans, but is more robust in the stem and
parvipelvian plans.
The scapula and coracoid of Cymbospondylus are unique.
In contrast to the other taxa, the coracoid is extended
posteriorly, whereas the anterior expansion is almost absent.
Furthermore, there is a coracoid foramen that is otherwise
unknown for ichthyosaurs. The scapula is unique in
having an anterodorsal emargination, which makes the
element triradiate. Otherwise, the shape of the scapula is
essentially similar to the basal pattern: one could derive
the shape of the Cymbospondylus-type scapula by carving
off the anterodorsal margin of a Mixosaurus-type scapula.
Such an emargination of the scapula is not rare among
extant squamates. Therefore, although the scapular blade
appears narrow, this narrowness is probably not homologous
with that seen in parvipelvian scapulae, which seems
to have evolved from the Californosaurus-type scapula. The
articulation between the scapula and coracoid is weak in
Cymbospondylus.
Shastasaurine coracoids are unique in that they are
elongated lateromedially. The plesiomorphic stem of the
bone had been elongated and widened in shastasaurines,
and thus the anterior and posterior margin of the bone
appears concave.
Grippia/ Chaohusaurus
Utatsusaurus
Mixosaurus Cymbospondylus Besanosaurus
Presacral count « 40 « 4 0 1 5 0 « 6 5 «60
Fig. 70. Comparison of postcranial parts among selected ichthyosaurs. Row 1: cartilaginous pectoral girdle; Row 2; pelvic girdle; Row
3: pelvic fin; Row 4: pectoral fin; Row 5: presacral count.