out proof, that it is mainly a case of the latter, but it 

 could well be some combination of both possibilities. 



As in the evolution of the triacanthoids and balistoids, 

 the major features in the diversification of the aracanids 

 from a prebalistidlike group and of ostraciids from 

 aracanids are largely reductive. 



It is not suggested here that the aracanids evolved 

 from a group with a level of organization like that of the 

 Recent or few known fossil balistids, but rather that 

 aracanids and balistids share a common ancestry at a 

 lower level of organization, around that of the 

 triacanthids. However, this joint ancestral line of the 

 aracanids and balistids must have been more generalized 

 than any of the Recent triacanthids, and probably split 

 off from the triacanthids at a level of organization 

 somewhat like that exemplified by Protacanthodes, the 

 Eocene ancestral triacanthid that probably evolved from 

 hoUardiinlike triacanthodids. That is, the triacanthid 

 line leading to the balistids and aracanids probably did 

 not yet have such specialized features as found in the Re- 

 cent triacanthids as: 1) the especially large notched in- 

 cisors; 2) the forward extension of the prefrontal suturing 

 to the vomer; 3) a fully elongate soft dorsal fin with many 

 more rays than the anal fin; 4) the uppermost pectoral 

 fin ray much reduced in size and one half smaller than 

 the other; 5) the second and third basal pterygiophores of 

 the spiny dorsal fin greatly reduced in size; 6) the ab- 

 sence of a free parhypural and autogenous haemal spine 

 of the penultimate vertebra; 7) an elongate tapering 

 caudal peduncle with a deeply forked caudal fin. The 

 triacanthid line leading to the balistids and aracanids 

 would have had to retain from its triacanthodid ancestry 

 more generalized conditions of these features, much as 

 did Protacanthodes, the basal triacanthid. Whether the 

 balistid-aracanid ancestral line diverged from the early 

 triacanthids at a level of organization slightly more 

 generalized than that represented by Protacanthodes or 

 at a level between that oi Protacanthodes and the Recent 

 triacanthids is difficult to ascertain, for too many critical 

 features of Protacanthodes remain unknown, but it was 

 probably the former, as discussed under the Balistidae. 



This ancestral pretriacanthid line is seen as diverging 

 into two radiations, one line leading with little change to 

 the Eocene Protacanthodes and the other triacanthids on 

 the one hand, and the other line leading through greater 

 changes to the balistids and aracanids, with the balistids 

 (as discussed under that family) remaining more 

 generalized and thus anatomically closer to their pre- 

 triacanthid ancestors than did the aracanids, whose 

 specializations are built around the defensive shell that 

 encases their bodies. While the balistid line continued to 

 lengthen the soft dorsal and anal fin bases and to develop 

 a locking mechanism between the first two dorsal spines 

 and an elaborate flexing mechanism of rudimentary rays 

 at the end of the pelvis, the aracanid line lost the spiny 

 dorsal fin and pelvic apparatus completely while 

 shortening the bases and decreasing the number of rays 

 in both the soft dorsal and anal fins. 



When the aracanid line split off from that of the pre- 

 balistids it rapidly lost, still in the Eocene, the spiny dor- 



Figure 147.— Typical body form in the Recent 

 Ostraciidae: Ostracion lentiginoaum. 



sal and pelvic fins as the developing carapace took on the 

 major line of defensive protection. The aracanids became 

 less flexible and less swift in sustained swimming as the 

 soft dorsal and anal fins were reduced in number of rays 

 and basal pterygiophores. While the pelvis was com- 

 pletely lost, there remains some trace of either the now 

 absent spiny dorsal fin supports or of those of the now ab- 

 sent anterior part of the originally longer based soft dor- 

 sal fin. The long supraneural element of aracanids that 

 extends out from the anterior end of the dorsal fin is un- 

 doubtedly a modified basal pterygiophore, but whether 

 from the spiny dorsal fin or the anterior region of the soft 

 dorsal fin is impossible to say with assurity, because as 

 the spiny dorsal was being lost so was some of the soft 

 dorsal fin, probably from anteriorly to posteriorly judging 

 from the position of that which remains. 



In balistoids the tendency has been for the spiny dor- 

 sal fin and its pterygiophores to migrate anteriorly as the 

 fin is reduced in size and number of elements from 

 balistids to monacanthids, and the latter retain no rem- 

 nants of the third spine, second pterygiophore, and 

 supraneural element of balistids. If the spiny dorsal fin of 

 aracanids migrated anteriorly as it became rudimen- 

 tary, it may have left free one of the more posterior pte- 

 rygiophores (corresponding in balistids to either the 

 second pterygiophore or to the supraneural strut, the lat- 

 ter itself being a modified third pterygiophore) which 

 then became enlarged to support the carapace. Con- 

 versely, in the only group of plectognaths other than 

 aracanids to lose the spiny dorsal fin, the gymnodonts, 

 the loss of the fin, on the evidence of Triodon, has been 

 by the posterior migration of the spines and pterygio- 

 phores as they became rudimentary. If the loss in 

 aracanids was by posterior migration, then the supra- 

 neural still represents an enlarged pterygiophore, or an 

 enlarged consolidation of two or more of them. Equally 

 possible is that the aracanid spiny dorsal fin, regardless 

 of whether it migrated anteriorly or posteriorly, lost not 

 only the spines but also all the pterygiophores, and that 

 the supraneural is actually an enlarged pterygiophore, or 

 an enlarged consolidation of two or more of them, from 

 the soft dorsal fin that became available for carapace 

 support as the soft dorsal was shortened from anteriorly 

 to posteriorly. 



