332 H. H. NEWMAN. 



cheeks top of head, etc., heavily shaded with a bluish-black colora- 

 tion, reminding one of the blue glint of the male F. heteroclitus 

 instead of- bronze yellow as described. The black spot on the 

 posterior rays of the dorsal may, as in the illustration (Plate 

 XXVII., 4), be composed of a series of spots arranged in a sort 

 of circle. This is more apt to be the case in very large specimens. 

 The color pattern of both sexes is gradually modified during 

 the lifetime of the individual and all stages in the production of 

 the complete adult pattern are readily found. The young of 

 both sexes are always cross barred somewhat like the adult male, 

 but the bars are less numerous, numbering seven to ten as com- 

 pared with from fourteen to twenty in adults. The increase in 

 number of bars takes place by means of a longitudinal splitting 

 of individual bars and by the appearance of small new bars 

 between the old ones (see Text Plate V., Figs. 8 and 9). In the 

 former case the two bars produced by the division of one simply 

 spread apart and at the same time broaden out ; in the latter case 

 the small alternate new bars merely increase in size until they 

 become nearly as large as the original bars, although it is usually 

 easy to distinguish the latter by their greater length. In one 

 unusually large male I observed fourteen well developed bars 

 nearly all of which had begun to split at the ends as though pre- 

 paring to double the number of bars once more. As a rule the 

 most anterior bars are the first to show signs of splitting, the ten- 

 dency proceeding antero-posteriorly. It seems to be a very general 

 rule that meristic changes of this sort proceed in this direction. 

 The color pattern of the females is at first similar to that of the 

 young male, but during the second season, probably, a marked 

 change begins to take place. The eight or ten bars that exist at 

 that time show decided irregularities in outline, each bar, begin- 

 ning with the most anterior, sending out at two places forward 

 and backward processes, which, on examination, prove to be ar- 

 ranged in two longitudinal lines, the upper one on a line with the 

 eye and the lower one on a line with the angle of the operculum 

 (Text Plate V., Fig. 2). The processes especially those of the 

 upper bar, continue to elongate anteriorly and posteriorly until 

 those of adjacent bars fuse together into a continuous longitudinal 

 stripe, the remaining portion of the bars becoming attenuated and 



