velopment of the balistid spiny dorsal fin apparatus. 



The above hypothesis in no way implies that Eoplectus 

 is directly ancestral to the balistids, for the balistids 

 clearly are derived from triacanthids. However, fishes 

 like the Recent triacanthids are too specialized to be con- 

 sidered as the ancestral group to the balistids, and it is 

 simply suggested that balistids could have evolved from 

 an early line of Protacanthodes-like triacanthids in the 

 Eocene still having an Eoplectus-hke arrangement of the 

 first three basal pterygiophores of the spiny dorsal fin. 



The Spinacanthinae appear to me to represent an ex- 

 tinct lineage of triacanthodids not ancestral to any of the 

 Recent groups of plectognaths, although with our present 

 negligible knowledge of the internal anatomy of the two 

 species of this subfamily, almost any phylogenetic state- 

 ment about them is mostly speculation. If the 

 Spinacanthinae were evolutionary dead ends, this sub- 

 family of triacanthodids joins the Cryptobalistinae sub- 

 family of triacanthids as the only two groups of fossil 

 plectognaths not clearly on evolutionary lines leading to 

 Recent groups of plectognaths. The reasons for con- 

 sidering the Cryptobalistinae as an evolutionary dead 

 end are discussed at length by Tyler (1968:243-249) and 

 for the Spinacanthinae below. 



The Spinacanthinae seem to have been an unsuccess- 

 ful experiment in specialization from a more normal 

 basal triacanthodid configuration. These specializa- 

 tions include: 1) the enormous elongation of the dorsal 

 fin spines and increased length of the fin base; 2) the 

 reduction in number of fin rays in the soft dorsal and 

 anal fins and the shortening of the fin bases; 3) the ex- 

 tremely high placement of the small eye in the head; 

 4) the forward migration of the spiny dorsal fin origin to 

 above the eye; 5) the probably reduced size of the pelvic 

 fin spine, if present at all; 6) the steep profile of the 

 snout; 7) the immense size of the body of one of the 

 species; 8) the development of enormous more or less 

 hexagonal scale plates forming a complete carapace over 

 the anterior part of the body behind the head in one of 

 the species. 



The trend for the elaboration of the dorsal fin spines 

 and lengthening of its base concomitant with a reduction 

 in the number of soft dorsal and anal fin rays and a 

 shortening of their bases is not compatible with a possi- 

 ble ancestry of the triacanthids or balistids, both of 

 which have reduced the spiny dorsal fin and elaborated 

 the soft dorsal and anal fins. The position of the eye and 

 the origin of the spiny dorsal fin in spinacanthins are, 

 respectively, higher and further forward than in any 

 triacanthids or balistids, and the snout is much steeper 

 than in those two families. Additionally, there is good 

 evidence that the hollardiin triacanthodids gave rise to 

 the triacanthids through a Protacanthodes-\ike line, and 

 that the balistids were derived from the early triacanthids 

 before they became as specialized as the Recent species. 



All of this would seem to me to exclude the 

 spinacanthins from further consideration as possible 

 ancestors of the triacanthids and balistids, and, even 

 neglecting the role of the eoplectins as the ancestors of 

 the gymnodonts, there is nothing compelling in the com- 



position of spinacanthins to think of them as possibly 

 related to the gymnodonts, other than that the pelvic fin 

 may have been reduced in size or absent and that the soft 

 dorsal and anal fins are reduced in size. 



This leaves but one group of plectognaths to be con- 

 sidered as hypothetically derivable from spinacan- 

 thins: the ostracioids. Ostracioids, like spinacanthins, 

 have short-based soft dorsal and anal fins and have com- 

 pletely lost the pelvis and pelvic fin, while in spinacan- 

 thins the pelvic fin may have been of reduced size or ab- 

 sent, and the condition of the pelvis is unknown. The 

 scales in one of the two genera of spinacanthins, Proto- 

 balistum, are large thick hexagonal plates which form a 

 continuous cuirass around the anterior half of the body 

 behind the head, the plates interdigitating along their 

 edges of contact and having surface granulations. In os- 

 tracioids the plates of the carapace are remarkably 

 similar to those of Protobalistum, although they cover 

 the head as well as most of the body. It is easy to envision 

 a transition from the mostly isolated and approximately 

 circular scale plates of Spinacanthus, with surface 

 granulations and a higher central spine, to the hexagonal 

 carapace plates in the partial body cuirass of 

 Protobalistum, with surface granulations and interdigi- 

 tated edges, to the fuller carapace covering of ostracioids. 

 While the teeth of Protobalistum are too large and 

 heavy to be conveniently ancestral to those of os- 

 tracioids, the smaller more conical teeth of Spinacanthus 

 could easily be ancestral to those of ostracioids. The 

 profile of the snout of some ostracioids is as equally steep 

 as in spinacanthins, and the eye of ostracioids, although 

 somewhat larger than in the two species of 

 spinacanthins, is located high in the head. However, it is 

 difficult to compare eye position between two groups in 

 which one has a spiny dorsal fin and the other does not. 

 Weighing against any phylogenetic significance to 

 these similarities between spinacanthins and ostracioids 

 are three factors: 1) the presence of an enormously well- 

 developed spiny dorsal fin in spinacanthins, which one 

 would expect to be at least no larger than in other 

 triacanthodids if not of greatly reduced size in a line 

 leading to the ostracioids; 2) enlarged scale plates, 

 whether or not with granulations and hexagonal pat- 

 tern, are not unique to Protobalistum and ostracioids 

 among the plectognaths, but are found in some 

 tetraodontids and molids as well, and the scales of 

 balistids can easily be envisioned as ancestral to those of 

 ostracioids; 3) the osteological evidence from Recent 

 species indicates that the ostracioids are derived from 

 the same ancestral line as that which gave rise to the 

 balistids, while it is discussed above why the spinacan- 

 thins are unlikely to be ancestral to the triacanthids and 

 balistids. 



The low number of fin rays and short base of the soft 

 dorsal fin in spinacanthins is probably correlated with 

 the trend for the enormous enlargement of the spiny dor- 

 sal fin, while for hydrodynamic reasons the apposed anal 

 fin would similarly become reduced in size to match its 

 equivalent above it. Thus, the short-based soft dorsal 

 and anal fins in spinacanthins could have no phy- 



