Xlll 



is the mind generally to impressions early received or which have become 

 current, that insensibly the premises in dispute are assumed and results 

 viewed with preconceptions reflected from the assumed premises.. 



But at the same time, caution must be exercised lest too great impor- 

 tance is attached to the minor modifications. For example, the great 

 frontal bone in the Gadinae and near related . subfamilies is single, as in 

 many other fishes, but in the subfamily Lotinae and in the family Mer- 

 luciidae, two entirely separate bones exist instead. Again, the inferior 

 pharyngeal bones are generally distinct in the Teleocephali, but in several 

 families they are united more or less early, and, in the extreme forms, 

 very soon, losing all trace of suture, and the eminent Johannes Midler 

 was led to separate the forms so distinguished from other fishes as a 

 distinct order (Pharyngognathi) ; that such a combination, however, was 

 somewhat hasty is demonstrable, independently of hypothetical considera- 

 tions as to the values of characters by certain facts. First, the combination 

 thus formed was a heterogeneous one, definable by no other internal or 

 external common characters, and composed of forms which respectively 

 agreed in structure, in all other respects, in the closest manner with other 

 widely separated types, and thus the character became tainted with suspi- 

 cion. Second, in another form (Haploidonotus) agreeing (generically) in 

 almost all details — and very characteristic ones moreover — with forms 

 (Sciaenidae) possessed generally of entirely separated bones, the pharyngeal 

 bones were found united as entirely as, and even more so than, in typical 

 Pharyngognathi of Midler, and it thus became evident that per se a com- 

 bination based on such a character would violently divorce forms from their 

 natural allies, and it was equally evident that the character itself was one 

 liable to recur in very dissimilar groups, and not even having the advan- 

 tage of being a technical expression of a natural group. 



With these remarks, examination may be made of the various orders of 

 fishes that have been adopted, commencing with those forms that appear 

 to be the most generalized or least removed from the Ganoids: the sequence 

 herein adopted is the most convenient for present purposes, and is also 

 believed to be a tolerably close exponent of nature. 



But as it will be necessary to make use of some elements concerning 

 which much difference of opinion prevails among anatomists, the author 

 deems it advisable to digress in order to examine into the merits of the 

 questions in dispute, and present his reasons for the nomenclature subse- 

 quently adopted. 



EXCURSUS OX THE SHOULDER GIRDLE OF FISHES. 



Few problems involving the homologies of bones in the vertebrate 

 branch have been in so unsatisfactory a condition as that respecting the 



