90 PKELIMINARY EXPERIMENTS ON 



development had proceeded normally (as in my experiments D, F, G, 

 and H), the black chromatophores were kept constantly contracted, the 

 resulting colour being a bright greenish yellow. This condition was 

 retained until the end of the third week, when the black chromato- 

 phores began to be partially relaxed, producing a distinct green-finned 

 stage, which gradually passed into the ultimate dark-finned condition, 

 the change being more or less synchronous with the abandonment of 

 pelagic habits. 



The supraorbital tentacles are larger than in the specimen 15 mm. 

 long described above, and are pigmented. They can be seen in the 

 larger specimens without the aid of a lens. 



The lower jaw does not protrude beyond the upper in any of these 

 specimens. In other structural details the fry do not materially differ 

 from the smaller specimen already described. The eyes in all specimens 

 are set equally far forward. 



The dimensions of the body and number of fin-rays in all five fishes 

 are given in the subjoined table. To complete the series I have added 

 the figures already given for the specimen killed on November 1st : — 



Maxiniinii 



In explanation of the figures I should add that the length of body 

 includes the caudal fin ; but the depth has been determined from the 

 base of the dorsal fin to the ventral surface, the exact line being drawn 

 across the base of the pectoral fins. The number of fin-rays was in 

 each case determined with the microscope. 



It will be noticed that the number of dorsal and anal fin-rays is 

 remarkably constant, since it is identical in five out of six specimens. 

 The abnormality which occurs in the smallest fish is apparently not a 

 variation due to difference of size, but a true individual variation affect- 

 ing the upper and lower fin alike. The caudal fin is subject to greater 

 variation, partly due to growth. 



On comparing these numbers with those assigned by Holt to the later 

 larva figured in his paper,* a considerable difference is noticeable, since 

 Holt's larva, having a body length of 18 mm., possessed twenty-seven 

 dorsal (12+ 15) and nineteen anal rays, thus exceeding by one dorsal 

 and one anal ray the numbers exhibited by the abnormal specimen in 



* Ann. Mus. Hist. Nat. Marseille, 1899, p. 46, pi. 6, fig 64. 



