26G PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 



lateral walls of the carapace. It is evident, from the study of tliese 

 transitions, that the form of the carapace cannot furnish a certain basis 

 for the establishment of distinct genera. 



"The character of the spines has, however, a greater value than that 

 of the form of the carapace. Although it be true that spines do not 

 occur exclusively uj)on this or that form of carapace, since there are 

 triangular trunk-fishes without spines, others armed with frontal and 

 anal spines, and others with anal spines alone, while there are also 

 quadrangular ones, spineless, or armed on the forehead and beneath the 

 tail, still there may be observed a certain consistency in their arrange- 

 ment as regards their position, their form, their luimber, and their di- 

 rection. But this constancy does not extend to their persistency since 

 some spines, or indeed all of them, are absorbed and disai^pear entirely 

 in adult individuals of certain species. In this manner all the spines 

 disappear with age in Ostraeion concatenatus, and if one were disiiosed 

 to see generic characters in its arming-, three genera might be founded 

 upon Ostraeion stellifer, BL, Schu. (in which the forehead, the dorsal 

 keel, and the ventral ridge are spinous), Ostraeion hicuspis, Blum., figured 

 by A. Smith (which has onl^' dorsal and ventral spines), and Ostraeion 

 concatenatus, Bl. (which has the carajiace entirely spineless). In real- 

 ity these species are merely nominal ; Ostraeion stellatus and Ostraeion 

 bicuspis being young individuals of the species of which 0. eoneatenatus 

 is the adult. In one other species, Ostraeion cornutus, Linn, (not Bloch), 

 the spines in the middle of the lateral dorsal ridge, and those on the 

 ventral ridge, decrease with age, and in the adult finally disappear. 

 In other species the spines are much more constant, but their propor- 

 tions, very different in accordance with the age of the individual, render 

 it sufficiently evident that they afford a character of very doubtful 

 value. I should, however, note the fact that there is no known example 

 of an Ostraeion with horizontal frontal and anal sjiines in which these 

 spines disappear in adult age." 



As has already been stated, the subgenera adopted by Bleeker are 

 founded solely upon the number and position of the spines. In Tetro- 

 somiis he places one pentagonal species, but in Acanthostraeion and 

 Ostraeion he includes triagonal, tetragonal, and i)entagonal forms with- 

 out discrimination. Notwithstanding the strong grounds taken by him 

 in regard to the importance of the shajje of the carapace it seems to 

 afford the most reliable guide in an arrangement of the species of this 

 genus. An arrangement with reference to the position of the spines 

 produces some incongruous results, while the other plan harmonizes to 

 a great extent with all structural features as well as with the geograph- 

 ical distribution of the group. Hollard remarked that the serial grada- 

 tion of the species was of great interest, but he did not Avork it out 

 with the care which might have been expected. I have endeavored to 

 indicate what seems to me to be a natural series, from the triagonal 

 spineless form through the ]>eiitagoiial form, provided with many spines, 

 to the tetragonal spineless form at the other extreme. 



