12 EEPEODUCTION AND DEVELOPMENT OF TELEOSTEAN FISHES 



as follows : — The ova fertilized on April 27tli were placed on tlie 

 following day in a glass jar provided with, a constant circulation of 

 water, the jar having a laj^er of gravel at the bottom in which the 

 lower end of a glass cylinder rested, the outflow passing through a 

 siphon which took the water from the inside of this cylinder. The 

 water supplied on this day was taken near the shore on the flood 

 tide three hours before high water, and had a density of 1"025. 

 The ova sank in this. On April 30th I procured some buckets of 

 water from near the mouth of the Sound, at the Duke Eock, and 

 this had a density of 1"026. As it replaced the other water in the 

 circulation jar the ova rose to the surface. The specific gravity of 

 the ova of this species is therefore about 1*0255. The temperature 

 was 9'2° to 8-0° on May 1st, 8-9° on May 2nd, in the water passing 

 through the jar. On May 4th the water brought in had a density 

 of 1'024, and on account of small tides and continuous rain I could 

 not get water any denser. The ova consequently sank to the 

 bottom, and by May 7th all of them were dead. This case seems 

 to show conclusively that death was due solely to the too low 

 density of the water used. The circulation was nearly constant, and 

 therefore the ova must have been sufficiently supplied with oxygen ; 

 and the temperature was very little higher than that of the open 

 sea. However, the ova lived nine days. Of the ova brought in on 

 May 13th, I placed the T. gurnardus in an apparatus like that 

 described above, and left the T. cuculus in a jar of still water taken 

 from the shore, having added common salt to it to cause the ova to 

 float. On May 15th I went to sea on board the trawler " Lola." 

 After my return, on May 19th, I found a single newly-hatched larva 

 of T. cuculus in the jar, the rest of the ova being dead. This is some- 

 what surprising, namely, that an egg should live six days and finally 

 hatch, in a small quantity of unchanged water to which common salt 

 had been added. The temperatures of course had not been noted. 



Develvpment and structure. — I have given a figure of the ovum of 

 Trigla gurnardus to show the early stage, not having drawn one of 

 T. cuculus till the stage shown in fig. 4. The dimensions of the 

 ovum in the two species are exactly the same ; the diameter of the 

 ovum is 1*45 mm., of the oil-glouble '8 mm. There is but a single 

 large oil-globule, which is as usual movable (see under Scomber) in 

 the early stages. The stage shown in fig. 4 in T. cuculus was 

 reached in five days three hours. Chromatophores of two colours 

 were present at this stage, some black, the others orange ; both 

 kinds are present on the sides of the embryo, only the orange on the 

 yolk-sac. The characters of the larva are shown in fig. 6, but the 

 figure is not coloured. The black dendritic chromatophores are here 

 as usual specially developed along the edges of the median fin-fold, 



