separated off first, followed by the Molidae. The Diodon- 

 tidae form the sister group of the Tetraodontidae plus 

 Canthigasteridae." The classification based on this phy- 

 logeny divides the plectognaths into two suborders in a 

 rather different way than ever proposed before, revised as 

 follows (Winterbottom 1974:99): 



Order Tetraodontiformes 

 Suborder Triacanthoidei 

 Family Triacanthodidae 



Subfamily Triacanthodinae 

 tSubfamily Cryptobalistinae 

 Subfamily HoUardiinae 

 Family Triacanthidae 



Subfamily Triacanthinae 

 tSubfamily Protacanthinae 

 Suborder Tetraodontoidei 

 Superfamily Balistoidea 

 tFamily Spinacanthidae 

 Family Balistidae 



Subfamily Balistinae 

 Subfamily Monacanthinae 

 Family Ostraciidae 



Subfamily Aracaninae 

 Subfamily Ostraciinae 

 Superfamily Tetraodontoidea 

 tFamily Eoplectidae 

 tFamily Zignoichthyidae 

 Family Triodontidae 

 Family Tetraodontidae 



Subfamily Tetraodontinae 

 Subfamily Canthigasterinae 

 Family Diodontidae 

 Family Molidae. 



The phylogenetic conclusions reached by Winter- 

 bottom on the basis of the myology of the plectognaths 

 and those presented here on the basis of their osteology 

 are in general agreement, although the hypotheses of this 

 monograph emphasize the role on the one hand of a 

 hollardiinlike Eocene stock of triacanthodids as an- 

 cestral to triacanthids through a protacanthodinlike line, 

 with the basal triacanthids giving rise to the balistoids 

 and ostracioids, and on the other hand of an eoplectin- 

 like Eocene stock of triacanthodids as ancestral to the 

 triodontids and hence to the tetraodontoids and molids. 

 Even though the phylogenies based on Winterbottom's 

 myological analysis and that of the osteology given here 

 are not far apart, and differ mainly in details of the rela- 

 tionships of fossil forms for which there is no myological 

 evidence, the classifications adopted are widely diver- 

 gent. 



While finding Winterbottom's philosophy of classifica- 

 tion highly worthwhile, eminently logical and entirely 

 defensible, the preference followed here is to weigh and 

 balance the value of both the generalized and specialized 

 features of the fossil and Recent species, rather than to 

 rely on the specialized alone, and the classification sug- 

 gested here reflects that bias. 



Winterbottom and I plan to prepare a joint paper after 

 the publication of this monograph discussing the merits 

 of alternatives to our respective modes of phylogenetic 

 interpretation and classification of various plectognath 

 linages and subgroups. Until then, the reader is left to 

 his own inclination. 



Historical Review of the Nonclassificatory Anatomical Work on the Plectognathi, 

 with notes on certain anatomical systems of special historical interest. 



Comments are given in the preceding section on the 

 contributions of such anatomically oriented workers as 

 Cuvier, Dareste, Hollard, Winther, Regan, Kaschkaroff, 

 Rosen, Gregory, Fraser-Brunner, and Le Danois, all of 

 whom were using an osteological framework for their 

 classifications of the plectognath fishes. A number of 

 other workers have described the general osteology of one 

 or several species of Plectognathi from a more or less 

 purely anatomical point of view, without concerning 

 themselves with questions of phylogeny and classifica- 

 tion. Specific details of these works are mentioned in the 

 subsequent analysis, but some of these shorter contribu- 

 tions are of sufficient importance to merit brief general 

 comment. 



Wellenbergh (1840) was the first person to treat an in- 

 dividual plectognath species in a manner similar to that 

 of Geoffroy Saint-Hilaire's (1827) anatomy of the Nile 

 puffer. Wellenbergh's description of the bones of the 

 skull of Mola leaves much to be desired by con- 

 temporary standards, but the branchial and hyoid arches 

 were described rather accurately. After Wellenbergh's 

 work there came a veritable flood of literature on the 

 anatomy of molids. Goodsir (1841) described with some 

 detail the histology of the shagreenlike skin and the 

 gelatinous subdermal tissue of Mola, as well as its 

 general myology. His work is of importance here because 

 of his description of the abortive nature of the posterior 

 end of the molid vertebral column. Cleland (1862) like- 



