The evolution of the dermal pectoral girdle is poorly
understood. Utatsusaurus had a somewhat cruciform or
spade-shaped interclavicle, with a long posterior and short
anterior processes. The interclavicle of the holotype figured
by Shekama et al. (1978) lacks the posterior process
and thus appears triangular. From this basic design, posterior
and anterior processes became reduced to various
extents. For example, parvipelvian interclavicles have a
reduced anterior process and thus appears T-shaped,
whereas both anterior and posterior processes are reduced
in Mixosaurus, resulting in triangular interclavicles (Merri-
am, 1908). The evolution of the clavicle is very poorly
known, and requires further study. The cleithrum is unknown
for any ichthyosaur.
Forefin
(Fig. 70)
The forefin is pentadactyl in basal forms, with a slight tendency
toward an increase in the number of elements anteriorly
(phalangeal formula of [2 or 3]-[4 or 5]-5-[4 or 5]-2).
Carpals are clearly distinguishable from metacarpals, which
retain short but complete shafts. The shafts of metacarpals
and phalanges are retained in Mixosaurus, and probably in
Cymbospondylus (Sander, 1989: fig. 7), but they disappeared
in the main stem-plan. The earliest shaft-less phalanges
and metacarpals are known from the Lower Saurian
Level of Spitsbergen (Wiman, 1910), indicating the rapid
evolution of this feature in the Early Triassic. These elements
are disarticulated but there are reasons to believe
that many of them are not carpals: among the numerous
elements of various sizes reported, there are none with
shafts. Also, femora closely resembling those of Besanosau-
rus co-occur with the elements, and this genus is known to
have phalanges and metapodials without shafts (Dal Sas-
so & Pinna, 1996).
M any p a rv ip e lv ian forefins h a v e stru c tu re s called
“n otches”, which are emarginations of the leading-edge
elements (they also o ccur along the trailing edge in the
basal euichthy osaurs Toretocnemus and Californosaurus).
These notches m ay o r m ay not be homolog ous to the
plesiomorphic shafts of phalanges and metacarpals (Caldwell,
1 997; Motani, 1999B). E xtensive histological and
phylogenetic studies are required to resolve this issue.
Hyperdactyly (presence of additional digits relative to
the basal amniote condition of five digits), characteristic of
parvipelvian ichthyosaurs, happened in several ways.
Most common was the posterior addition of digits, resulting
in the presence of digits VI, VII, and even VIII in some
forms. The digital count also increased from splitting of
existing digits, which occurred anterior to the primary axis
in Ichthyosaurus, and posterior to the axis in Stenopterygius
and Suevoleviathan (Motani, 1999B). The most exceptional
was the anterior addition of digits, limited to the Ophthal-
mosauria. Digit I was absent in most merriamosaurian
ichthyosaurs (Motani, 1999), and the additional anterior
digits cannot be considered homologous to the plesiomorphic
digit I. There were up to three additional digits anteriorly
in Platypterygius and Caypullisaurus, which had the
highest digital counts of no less than 10 digits per forefin
in at least some individuals (W ade, 1984; F ernAndez,
1997).
Hyperphalangy (presence of extra phalanges in addition
to the basal amniote phalangeal formula of 2-3-4-5-3)
was commonly present in all ichthyopterygians. (Note,
however, that complete forefins are unknown for many
Triassic ichthyosaurian taxa.) The phalangeal count exceeds
15 in some parvipelvians, including Platypterygius
and Ichthyosaurus.
Pelvic girdle
(Fig. 70)
The ischio-pubic plate was large in the basal, stem, and
mixosaurian plans, but was reduced in the parvipelvian
plan (hence the name of the latter). Basal parvipelvians
had a somewhat large ischium, but it became reduced in
the more derived forms and eventually fused to the pubis,
which is also reduced. Openings in the ischio-pubic plate
also went through extensive modifications. In the basal
forms, the median thyroid fenestra was not well developed,
and the obturator foramen was enclosed within the
pubis (e.g., Utatsusaurus). The thyroid fenestra became
enlarged in the stem and mixosaurian plans, whereas the
obturator foramen gradually opened posteriorly, and eventually
became connected to the thyroid fenestra in the stem
group (e.g., Califomosaurus). The thyroid fenestra was reduced
again in the parvipelvian plan, although it reappeared
occasionally, as in Temnodontosaurus. The obturator
foramen was located between the pubis and ischium in
parvipelvians, and therefore the structure should be called
the obturator fenestra (Romer, 1956).
In the basal, stem, and mixosaurian plans, the ilium has
a thick proximal end that participates in the acetabulum,
and a thinner distal expansion. This expansion may not be
visible, depending on the preservational angle (Fig. 70, il
vs. il’). The distal expansion is reduced to various degrees
in parvipelvians, and the whole element eventually becomes
rod-like (e.g., Stenopterygius). Even in these forms,
the proximal end that participates in the acetabulum is
never thin. There is a controversy regarding the orientation
of the ilium in Suevoleviathan (Maisch, 1998; Motani,
1999B). Although the readers should consult Maisch (1998)
for fair comparisons, the evidence for the orientation given
in Fig. 70 is summarized here: (1) the acetabular end of the
ilium is always thick in ichthyosaurs, which is functionally
reasonable; (2) the ilium nev er contacts the entire pro ximal
margin of the ischio-pubic plate in ichthyosaurs, except
possibly in some ophthalmosaurians, but only near
the acetabulum; (3) distal expansion of the ilium is known
in the more basal ichthyosaurs, so its retention in Suevoleviathan
is not surprising; (4) the three bones do not m eet to
form the acetabulum in Maisch’s (1998) reconstruction,
which is v e ry unlikely.
Hindfin
(Fig- 70)
The hindfin shares many characteristics with the forefin: it
is pentadactyl in basal forms, and digit I is lost in merria-
mosaurians. Hyperphalangy and hyperdactyly are much
less pronounced in hindfins than in forefins. Basal ichthyosaurs
have two proximal tarsals, the astragalus and calca-
neum. The number of elements in the proximal tarsal row
increases to three in euichthyosaurs, but the identity of the
third element has yet to be firmly established. This element
occurs anteriorly to the other two. In some specimens of
Stenopterygius, even the zeugopodial row contains an extra
element, in addition to the tibia and fibula.
Vertebral column
(Fig. 68)
Ichthyosaurs were axial swimmers, so the vertebral structure
reflects their swimming styles to some extent. All
ichthyosaurs have 40 or more presacral vertebrae (Motani
et al., 1996), far greater in number than in terrestrial reptiles
with limbs (Hoffstetter & Gasc, 1969). Basal ichthyosaurs
have about 40 presacral vertebrae, whereas stem
ichthyosaurs and Mixosaurus have 50 or more. The highest
presacral count occurs in Guanlingsaurus liangae, which has
about 80 vertebrae anterior to the pelvic girdle (Yin et al.,
2000). The vertebral count is reduced to about 45 in parvipelvians,
with the exceptions of Leptonectes solei and Aegi-
rosaurus leptospondylus.
The caudal vertebral columns of all ichthyosaurs are
curved in the middle, forming a peak. This caudal peak is
usually associated with anticlination of the neural spines.
The vertebral .count to the caudal peak varies from about
60 in basal forms to over 120 in Guanlingsaurus. The caudal
peak of parvipelvians, and possibly of their sister taxa, is
usually referred to as the tailbend. Among parvipelvians,
the lowest count to the tailbend is found in Ichthyosaurus
breviceps (about 70) and the highest in Temnodontosaurus
(up to 100).
The shape and size of vertebral centra is variable within
a single individual, and many taxa share one basic
pattern. The cervical centra are pentagonal in anteroposterior
view, with the peak pointing ventrally. The shape
becomes almost circular toward the mid dorsal region, and
then hexagonal toward the anterior caudal region. The
comers of the hexagons may be rounded. The hexagons
tend to be “bottom heavy” (i.e., the bottom half is expanded
compared to its upper counterpart), and are as high as
wide in the anterior caudal region, but higher than wide
posteriorly. In the forms with tailbends, the hexagons become
wider than high anterior to the bend, and then suddenly
becomes higher than wide after the bend. In lateral
view, height/length ratio of the centra changes.
The largest centra are the ones in the anterior caudal
region, from whence the size decreases both anteriorly and
posteriorly. The first two or three centra may be fused to
form the atlas-axis complex, but such a fusion has not been
established for mixosaurs or basal forms. Cymbospondylus
clearly had separated atlantal and axial centra (Merriam,
1908). RM was unable to locate the critical portion of
UCMP 9950 figured by Merriam (1908), but this condition
can be confirmed in UCMP 9913, which was also figured
by that author. Utatsusaurus also seems to have shared this
separation of the atlas and axis: the atlantal centra occur
near the occiput in the holotype and another specimen as
an isolated disk (Motani et al., 1998), much resembling
that of UCMP 9913 (Cymbospondylus). The axis is unknown
for Utatsusaurus, or at least has not been properly identified.
With the lack of definitive information from other
forms, it seems most likely that the separation of the atlas
and axis is plesiomorphic for ichthyosaurs.