Our preferred hypothesis is that the ancestral balistoids had
small scales like the triacanthoids and some of the more
primitive tetraodontoids, simply because we have no particular
reason to believe that a large-scaled balistoid ancestral line is
more probable than the proposed small-scaled ancestry. If one
accepts small scales as ancestral to balistoids, then the
enlargement of scales in Eospinus relates it to either ostraciids
or spinacanthids (Protobalistum) in the accepted phylogeny or
indicates, contrary to other evidence, that spinacanthids and
ostraciids are sister groups to which Eospinus has an
unresolved relationship. None of these alternatives are refuted
by the possibility that the ancestral balistoids had large scales,
in which case the large scales of Eospinus are phyletically
uninformative within the balistoids.
In Protobalistum the partial carapace has a ventrolateral
ridge, similar to the ridges in this position in most ostraciids
and in Eospinus. Just as with scale enlargement, we postulate
that the carapace ridge in Protobalistum is an independent
acquisition from that in Eospinus and ostraciids. Thus, ridges
are interpreted as a common functional aspect of carapace
construction irrespective of the phylogenetic origin of the
carapace.
CARAPACE SPINES.—Eospinus has a large median carapace
spine directed forward from the snout and another fixed spine
directed posteriorly from the middle of each side of the body.
No such spines are present in spinacanthids and balistids
(among the latter neither the flexible quill-like spines found on
the side of the body of males of the monacanthin Amanses
scopas (Cuvier) nor the enlarged tubercles and spiny processes
borne on the scales of the caudal peduncle in many balistins and
monacanthins seem homologous to the fixed carapace spines of
Eospinus). Among ostraciids only the Eocene Eolactoria has a
median, but much shorter, carapace spine on the snout, in the
same position as that of Eospinus (the Recent Ostracion
rhinorhynchus Bleeker has a blunt, rounded enlargement on the
snout, but not a spine). Bilateral carapace spines are present in
many ostraciids, either on the supraorbital ridge or along the
dorsolateral and ventrolateral ridges of the carapace, but none
has a large spine projecting posteriorly from the middle of each
side of the body.
While the median snout spines of Eospinus and Eolactoria
could be considered a synapomorphy, we discount the
significance of this evidence because of the great degree of
homoplasy in carapace spines in ostraciids. For example,
seemingly identical large carapace spines bilaterally in front of
the eyes have been acquired independently by genera in two
different phyletic lines of Recent ostraciins, and this is also
homoplastic to the occurrence of such preorbital spines in the
Oligocene Oligolactoria (Tyler and Gregorova, 1991). Moreover,
the numerous spiny scale processes of pelagic larval
molids, which apparently act as flotation and defensive devices,
include a long anteriorly projecting median snout spine and one
projecting posteriorly from each side of the middle of the body
(especially in Masturus). These are similar in size and
placement to those of Eospinus but they are not associated with
a carapace. As molids are one of the most derived groups of
tetraodontoids and more closely related to tetraodontids and
diodontids than to ostraciids, it is reasonable to assume that the
molid snout and body spines have been acquired independently
from those of ostraciids. Because there is no other evidence of
a close relationship between Eospinus and Eolactoria, we
postulate that the snout spines in these two genera (and in larval
molids) are homoplastic, and that the presence of carapace
spines of any kind in Eospinus is not evidence of relationship
with ostraciids.