PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 267 

 Serial arrangement of the species of Ostracion. 



I. Form triagoual, spineless <^>- ^queter. 



11. Form triagonal, with two ventral spines 0. irigonus, 0. hicamlalis. 



III. a. Form triagonal, with two ventral and two horizontal frontal spines, the 



posterior extension of whose base gives in young individuals a semblance to 



a tetragonal, in adults to a pentagonal shape 0. qnairicornis. 



h. Lik-e the last in every respect except that there is a median dorsal spine. 

 (This is frequently seen in the young of 0. qmdricornis, disappearing at 

 different stages of growth in different individuals, hut in others persist- 

 gjj|\ 0. quadricornis, subspecies notacanthus. 



IV. a. Form subtriagonal, approaching to pentagonal, the posterior extension of the 



orbital crest being more pronounced than in III. Frontal spines smaU, 

 vertical, frequently double, two small spines upon the dorsal ridge and two 

 on each ventral keel, all the spines obsolescent with age. . . . 0. concatenatus. 

 b. (Aside-shoot from a.) Like the last, but with all characters exaggerated 

 and more persistent, the dorsal spine single and high, the spines on each 

 lateral keel four in number. 



0. turrltus (by GUnther considered to bo probably identical with 0. con- 

 catenatus). 



V. a. Form subtetragonal, approaching pentagonal, but with a dorsal surface 



clearly defined, though the affinity to the triagoual forms is indicated by a 

 pronounced elevation of the dorsum, surmounted by a high spine. Frontal 

 spines horizontal, stronger. The fullness of the anterior part of the body 

 observed in the forms already studied is suggested by a bulging of the 



ventral surface 0. Fornastni, 0. cornutus, Linn. =0. dlaplianus, Schn. 



In 0. diaphamis the dorsal surface is flatter than in 0. Foniasini, but there 

 are small spines on the dorsal and ventral keels, obsolescent in age, which 

 suggest the preceding form. The two forms together, or an average between 

 them, form a needed link in the series. 

 h. (A side-shoot with great development of frontal and ventral spines.) Forms 

 similar to the last but approximating still more closely to the tetragonal, 

 particularly in adult age ; without dorsal spine, though with a trace of its 

 presence in an elevated dorsal ridge. Horizontal spines very prominent. 



0. arciis, Schn. = 0. cornutus, Bloch. 



VI. Form tetragonal, spineless, similar to the last, but with squarer angles. "A 



more or less sharp protuberance in front of the dorsal tin, from which several 

 pointed lines radiate." Horizontal spines absent. This is the transition 

 from the subpeutagonal and subtetragonal to the truly tetragonal forms. 



0. ornatus. 

 VII. a. Form tetragonal, spineless, similar to the last, but with lower though still 



very distinct dorsal ridge, A trace of rostral prominence 0. rmsus. 



h. (Side-shoot from a.) Similar to a, but with prominent rostral hump. 



0. rhinorhynchus. 

 c. Form truly tetragonal, back convex, not ridged, dorsal and ventral keels 



blunt . . .' r. 0. cuMcus, 0. punctatiis, 0. sehcp. 



VIII. Form tetragonal, spineless, back flat, ridges sharp 0. Benardl, O. solorensis. 



Such is the continuity of the gradation in this series that it is ahnost 

 impossible to distribute the species into subgenera, though the extreme 

 forms wouhl be considered by many writers as belonging to well-marked 

 genera were the hitermediate forms not known. The transition is pe;r- 

 fect, without a break from 0. triqueter to 0. Eenardi and 0. solorensis. 

 Even the size, abundance, tfnd distribution of the spines are seen to be 

 correlated to the shape of the body, for these are to be regarded, as 



