by attempting to disarticulate the bones after long 

 maceration. When a bone is said to be cartilage filled 

 along a particular edge, this refers to the presence of car- 

 tilaginous material extending into the otherwise ossified 

 substance of the bone so that an upper and a lower layer 

 of bone, separated by cartilage, are distinguishable at the 

 edge of the bone. Such a condition, of course, is only pos- 

 sible in endochondral bones. As a general rule, the 

 amount of interdigitation between bones increases with 

 increasing specimen size, while at the same time the 

 amount of cartilage at the edges of endochondral bones 

 decreases. 



The suffix oideo is adopted for the names of infraor- 

 dinal categories. 



Throughout the text, vernacular versions of the 

 classically based and more formal superfamilial, 



familial, and subfamilial names, ending respectively in 

 oidea, idae, and inae, are employed, with the last two 

 vowels dropped for adjectival usage and with these 

 replaced by an "s" for nominative usage, so that super- 

 families end in -oid or -oids (triacanthoid, balistoid, os- 

 tracioid, triodontoid, tetraodontoid, moloid), families in 

 -id or -ids (triacanthodid, triacanthid, balistid, mona- 

 canthid, aracanid, ostraciid, triodontid, tetraodontid, 

 diodontid, molid), and subfamilies in -in or -ins (spina- 

 canthin, eoplectin, hollardiin, triacanthodin, protacan- 

 thodin, cryptobalistin, triacanthin, ostraciin, lactophry- 

 sin, tetraodontin, canthigasterin). The names of two of 

 the familial categories, the Triacanthodidae and Triacan- 

 thidae, unfortunately are rather similar, and the distinc- 

 tion between such terms as triacanthoid, triacanthodid, 

 and triacanthid is a necessary mental excercise. 



PROLOGUE 



Synopsis of the Phylogeny of the Plectognathi 



The anatomical data on the fossil and Recent Plectog- 

 nathi presented in this monograph indicate that the 

 family Triacanthodidae contains the most generalized 

 fossil and Recent species of the order, and that these 

 basal plectognaths gave rise on the one hand to the line 

 leading to the triacanthids, their nearest relatives, and 

 through them to the balistoids and ostracioids, while, on 

 the other hand, and with even greater modification away 

 from the ancestral level of organization, the triacan- 

 thodids gave rise to the triodontids, and through them to 

 the tetraodontoids and molids. The triacanthodids and 

 their anatomically and phylogenetically closest deriva- 

 tive groups, the triacanthids, balistoids, and ostracioids, 

 are here considered to represent the suborder Scleroder- 

 mi or Balistoidei, while the other major line of plectog- 

 nath radiation, that has diverged even further from the 

 ancestral triacanthodid type, comprising the triodon- 

 tids, tetraodontoids and molids, is here considered to 

 represent the suborder Gymnodontes or Tetraodon- 

 toidei. 



Two fossil forms of the basal triacanthoids from the 

 Eocene are thought to be especially pertinent to the phy- 

 logeny of the plectognaths, for Protacanthodes forms a 

 strong link between the triacanthodids and the triacan- 

 thids and thus to the other derived scleroderms 

 (balistoids and ostracioids), while Eoplectus is nearly a 

 perfect intermediary between the triacanthodids and the 

 triodontids, the most generalized of the gymnodonts, and 

 thus to the derived tetraodontoids and molids. 



More particularly, of the four subfamilies of Triacan- 

 thodidae recognized here, it is obvious that the Triacan- 

 thidae are derived from the HoUardiinae rather than 

 from the Triacanthodinae or the fossil Spinacanthinae 

 and Eoplectinae because of similarities in the shapes and 

 positions of the bones in the rear of the skull and of the 

 shape of the pelvis and because there is every reason to 

 believe that the Spinacanthinae were evolutionary dead 



ends without issue surviving today and that the Eoplec- 

 tinae were ancestral to the triodontids and the other 

 gymnodonts. 



The evolution of the Triacanthidae from the Triacan- 

 thodidae was probably mediated through a form like 

 Protacanthodes, the Eocene representative of the most 

 generalized subfamily of triacanthids. Of the other two 

 subfamilies of triacanthids, the triacanthins include all 

 of the Recent species and obviously evolved from such 

 Oligocene forms as Acanthopleurus, while the aberrant 

 cryptobalistins of the Oligocene were, like the 

 spinacanthin triacanthodids, failed evolutionary ex- 

 periments that did not give rise to forms living today. 

 The balistoids and ostracioids are reasoned to share a 

 common ancestral line to the triacanthids in the Eocene, 

 and it is unquestioned that the balistids gave rise to the 

 monacanthids and that the aracanids are ancestral to the 

 ostraciids. Two subfamilial lines of evolution are evident 

 in the ostraciids. 



The recent discovery of the eoplectin triacanthodids, 

 with a well-developed spiny dorsal fin, the best 

 developed pelvic fin among plectognaths, and a generally 

 triacanthodid appearance, but with a fully developed 

 gymnodontlike crushing beak, tends to confirm that the 

 triacanthodids gave rise to a line of triodontidlike forms 

 which in turn were ancestral to the two main radiations 

 of the gymnodonts, the molids on the one hand and the 

 tetraodontids and diodontids on the other hand. The 

 tetraodontids show two subfamilial phyletic lines, the 

 tetraodontins for the vast majority of diversified forms, 

 related through Carinotetraodon to the canthigasterins. 



In summary, it seems clear, especially in the light of 

 our present knowledge of Protacanthodes and Eoplec- 

 tus, as discussed by Tyler (1973b), that the basic diver- 

 sification of the plectognaths was a relatively rapid event 

 of the lower Eocene. In the upper portion of the lower 

 Eocene of Monte Bolca, Italy, alone, there are represen- 



