Fishes of the Family Stromateidse. 119 



It will be seen that there are great differences between the 

 genera, but in all cases the following generalizations apply. 

 A greater or less number of the anterior prgecaudals have no 

 transverse processes, and are followed by vertebrai with 

 downwardly directed transverse processes, successively in- 

 creasing in length, and in the posterior praicaudal region 

 uniting ventrally to form a closed ha3mal arch. The two 

 first prsecaudals have no ribs and are followed by vertebrae 

 with sessile ribs, which are inserted lower down on each 

 succeeding one, becoming attached ventrally to the hasmal 

 arch when the transverse processes unite. 



Throughout this family considerable changes take place 

 during growth, and the non-recognition of this fact has been 

 the cause of multiplication of nominal species. In some 

 genera, e. g. Nomeus^ Cuhiceps^ the shape of the body 

 changes from short and deep in the young to elongate in the 

 adult ; but this does not occur throughout the family. The 

 vertical fins always undergo considerable changes, in most 

 cases being relatively of much greater height in the young, 

 and often during growth they become much more densely 

 scaly, and some of the spines are overgrown and imbedded. 

 The ventral fins are always much longer in the young ; the 

 pectorals, on the contrary, are subsymmetrical and rounded 

 in the young, and during growth, owing to the greater elon- 

 gation of the rays of the upper halt of the fin, become asym- 

 metrical and relatively of greater length. Thus the length of 

 the ventrals and pectorals or the height of the dorsal and anal 

 are of little value for systematic purposes, unless the size of 

 the specimen described is given. 



Another cause of the multiplication of species on the part 

 of some systematists has been their ignorance of the fact that 

 certain characters regarded as specific are really common to 

 all the species of a genus. Thus, in all the forms with 

 deciduous scales, grooves or furrows, corresponding to the 

 intermuscular septa, are seen on the body ; and these have 

 given rise to tho specific names of sulcatus, bilineatus, &c. 

 Also the presence or absence of pores on the body has been 

 taken as a character by which to distinguish between species, 

 their supposed absence being in many cases due to the fact 

 that they cannot easily be detected in all preserved specimens. 

 Generally the first two or three articulated rays of the dorsal 

 are unbvanched, and so have been reckoned as spines by 

 some authors. 



The genera of this family have in not a few instances been 

 iusufEcicntly diagnosed, and the addition of several which 



