fused with the main body of the plate, but even in the 

 12.3 mm specimen this thinner parhypural region is con- 

 tinuous with the main body of the plate. The rest of the 

 plate is composed of the last vertebral centrum fused 

 with all of the hypural elements. In about the anterior 

 two-thirds of its length the caudal plate possesses a lateral 

 flange whose depth increases anteriorly. The anterior end 

 of the lateral flange articulates by fibrous tissue with the 

 posterolateral process from the haemal spine of the 

 penultimate vertebra. The posterior edge of the caudal 

 plate supports the caudal fin rays. 



Caudal fin rays. — Nine in number; the uppermost 

 ray and the lowermost ray unbranched, the other rays 

 becoming increasingly branched toward the middle rays, 

 which are branched in triple dichotomies. The bifid 

 bases of the rays articulate by fibrous tissue with the pos- 

 terior edge of the caudal plate. 



DORSAL A^^D ANAL FINS. 



Fin rays and pterygiophores. — Fifteen fin rays 

 usually present, the first two rays and the last ray un- 

 branched, the others branched in single or double dicho- 

 tomies. Distal pterygiophores either absent or unos- 

 sified. The fin rays are supported through fibrous tissue 

 at their bifid bases by 12 basal pterygiophores. The 

 pterygiophores are cartilage filled at both ends, and all of 

 them, with the exception of the last, are stout rods which 

 decrease in width posteriorly in the series. The last 

 pterygiophore is rodlike anteriorly, but posteriorly it is 

 expanded into a flat plate. The ventral edge of the last 

 pterygiophore articulates by fibrous tissue with the bony 

 roofs over the neural canal of the fourth and fifth caudal 

 vertebrae. The other pterygiophores articulate ventrally 

 by fibrous tissue with the irregular dorsal edges of the 

 neural arch plates of the 12th abdominal to 4th caudal 

 vertebrae. The pterygiophores articulate by fibrous tis- 

 sue dorsally with one another and with the dorsal fin 

 rays. 



Anal Fin. 



Fin rays and pterygiophores. — Thirteen or fourteen 

 fin rays usually present; the first two rays and the last 

 ray unbranched, the others branched in single or double 

 dichotomies. Distal pterygiophores either absent or unos- 

 sified. The fin rays are supported through fibrous tissue 

 at their bifid bases by nine basal pterygiophores. The 

 basal pterygiophores decrease slightly in length pos- 

 teriorly in the series and are all, with the exception of the 

 last pterygiophore, stout rods. The last pterygiophore is 

 rodlike anteriorly, but slightly expanded posteriorly into 

 a flat plate. The amount of expansion of the last anal fin 

 pterygiophore is not equal to that of the last dorsal fin 

 pterygiophore, but in large specimens the difference is 

 not great. The pterygiophores are cartilage filled at both 

 ends. The dorsal ends of the pterygiophores articulate by 

 fibrous tissue with the irregular ventral edges of the 



haemal processes of the first to fifth caudal vertebrae. At 

 their distal ends the pterygiophores articulate with one 

 another by fibrous tissue. 



Anatomical diversity.— The few genera of Recent 

 diodontids form an anatomically compact family of 

 gymnodonts that has probably changed little in overall 

 configuration since the Eocene, with the exception of an 

 increase in the size of the scales and in certain rearrange- 

 ments of the dentition and jaw supporting structures. 



The few relatively complete fossil diodontids are from 

 the Eocene of Monte Bolca, Italy, and are referable to 

 Diodon tenuispinus Agassiz and D. erinaceus Agassiz. 

 Agassiz (1844b) simply said that D. erinaceus was larger 

 and rounder, and had shorter and heavier spines, than in 

 the more fully described D. tenuispinus. The types of D. 

 tenuispinus are in Paris, and Le Danois (1955, 1959) 

 proposed the new generic name Prodiodon for 

 tenuispinus on the basis of the supposedly anteriorly 

 elongate sphenotics broadly entering the edge of the or- 

 bit above and of the supposedly sutured (rather than ful- 

 ly fused) premaxillaries which would distinguish 

 tenuispinus from all Recent species. 



The three type-specimens of P. tenuispinus (see 

 Material Examined) have been examined for this work. 

 None of the three show much detail of the jaws, skull 

 structure, vertebral column, and armament. A dorsal 

 view illustration of the only one of these three specimens 

 that is in counterpart, and considered at the Paris 

 museum to be the holotype, is presented by Le Danois 

 (1959:215, fig. 184). The illustration, in my opinion, is 

 wildly inaccurate and impressionistic, being rendered 

 with heights of artistic license. I find it difficult to 

 decipher any pertinent details in the dorsoventrally com- 

 pressed and probably fractured cranium, but Le Danois 

 showed the shapes and sizes of most of the bones there, 

 with the sphenotic being quite unlike that of any Recent 

 species, lacking a lateral prong and being prolonged an- 

 teriorly to form much of the lateral edge of the upper or- 

 bit, much as in the tetraodontid Colomesus. I seriously 

 doubt that the degree of detail shown in the illustration 

 by Le Danois is actually present. 



Le Danois described and showed the upper jaws with 

 the premaxillaries sutured as in tetraodontids, with in- 

 terlocking emarginations, while saying that the dentaries 

 are either fused to one another or at least closely at- 

 tached. To my eyes the jaws that are exposed in the three 

 specimens seem to be fully fused medially without any 

 evidence of sutured right and left halves. Several trans- 

 versely elongate trituration teeth are present to either 

 side of the midline on what is apparently the inner sur- 

 face of the lower jaw in the specimen in counterpart. Le 

 Danois showed the pectoral and dorsal fins far more 

 clearly than they exist on the specimen. What vague in- 

 dication remains of the dorsal fin in the fossil is far 

 shorter based than illustrated by Le Danois, with the 

 true fin corresponding only to the posterior half of the il- 

 lustrated fin. The illustration clearly shows 23 verte- 

 brae, but I would estimate that the much vaguer out- 

 lines represent about 18 to 20 vertebrae. 



351 



