388 ri?HEs. 



subdivisions, we see him placing the herrings between the mullets 

 and the amphacanthes (buro), the cods near the gasterestes, the 

 sword-fisli near the genus anarrhicus, and leaving the rhinchobdelles 

 and the liog mares, in the same family with the eels. 



It was only after nearly forty years of study on fishfes, not from 

 authors, but on the fishes themselves, on their skeletons, on their vis- 

 cera, after having dissected several hundred species, that I became 

 convinced of the necessity of never mixing any acanthopterygians 

 with fishes of other families; that 1 learned that the acanthoptery- 

 gians, which form three-fourths of all the known fishes, are also the 

 type with which nature has taken most pains, and which it has main- 

 tained in the greatest resemblance Avilh itself, through all the varia- 

 tions of detail which it has caused it to undergo. 



xA.ll the other characters would only be made use of as inferior to 

 this one, and could not be allowed to supersede it ; but the extreme 

 constancy of the general, and the predominating influence of the re- 

 gulating character, has rendered it very diificult to give to fishes, in 

 Avhich it exists, precise and sensible applications of subordinate cha- 

 racters : thus, the different families of the acanthopterygians pass so 

 insensibly into each other, that we cannot determine where one 

 begins or another ends. 



The family of pearches, for instance, which is essentially distin- 

 guished bv its palatine teeth from that of the sciences, comprises a 

 considerable group, perfectly natural in all other respects, one part 

 of which possesses those teeth, whilst another is deprived of them. 



The same thing occurs in the family otherwise Avell characterized, 

 of the cuirass-jawed : the greatest part of these genera is allied to 

 the pearches ; the other to the scisjnas as far as regards the palatine 

 teeth. 



There are some passages which are sensible from a part of the 

 genera of the family of the scienes, to those of the chetodons, by the 

 scales which cover, more or less, their vertical fins; and notwithstand- 

 ing we are on the other hand obliged to bring together the family of 

 the sj)ares of several genera of sciaenes, which have no trace of those 

 same scales. 



Transitions no less marked, unite certain genera of spares, such as 

 the picarelsand the gerres, with other genera, such as the equalus, 

 which cannot be widely separated from the zeus, which, in their 

 turn, lead to the family of the scombres, and this last passes by 

 shades so fine to those fishes in form of ribands, called tsenoides, 

 that it is almost impossible to say where the limit should be 

 placed to separate them from each other. 



It is therefore the duty of naturalists desirous of making known 

 beings according to their true relations, to acknowledge that the 

 acanthopteiygian fishes, which form the old genera of pearches, the 

 cepola, down to and comj-rising the sciaenas, sparus, chcetodons, 

 zeus, and scorabcrs, and other fishes in the form of ribbands, 

 and which form, notwithstanding the innumerable quantity of 

 their species, com])ose but one natural family, in which we 

 can distinguish shades, perceive commencements of groups, and 



