142 BULLETIN OF THE UNITED STATES FISH COMMISSION. 



edges of the germinal disk in rei)tiles, birds, and mammals, and wbicli 

 eventually incloses the embryos of these forms, and in which they 

 undergo the largest part of their development. A similar spacious 

 cavity api)ears to exist in the egg-cases of the oviparous sharks, and 

 rays, filled with water or a thin serous fluid, in which the embryo de- 

 velops as it absorbs the contents of the yolk sack. The same remark 

 applies to the eggs of the Amphibia and AmpMoxtis; so it appears that 

 all of the vertebrates below the reptiles have an egg-membrane, or 

 what answers to it, and a water or respiratory space in which the germ 

 or vitellus is included, and in which it undergoes a more or less com- 

 plete development. 



The vitellus or germ included by the egg-membrane of the mackerel 

 is globular and very nearly fills up the cavity, bounded by the mem- 

 brane, so that the water space or breathing chamber is small. The 

 vitellus is made up of three principal portions : a thin outer germinal 

 layer, as shown in Fig. 2, which incloses a globular yolk-mass, in which 

 it is ordinarily difficult to distinguish the contour of the yolk spheres 

 of which it is composed. Imbedded in this yolk-mass there is always 

 a single oil sphere perfectly globular, highly refringent, and measuring 

 about one one-hundredth of an inch in diameter, and which always oc- 

 cupies an eccentric jjosition in the upper hemisphere of the living egg. 

 It is the presence of this oil sphere which causes the egg to float; the 

 drop of oil is always in such a jiosition as to keep the developing em- 

 bryo inverted or turned upon its back. This is probably due to a purely 

 physical cause ; the oil sphere, being the lightest part of the egg, will 

 always be found in its upper hemisphere, while the germinal disk or 

 embryo appears to be the heaviest part, and in consequence is always 

 found in the lowermost hemisphere looking back downwards. It is 

 difiicult to trace any protoplasmic filaments passing ofl' from the ger- 

 minal layer of the mackerel's ovum down amongst the yolk spheres, in 

 consequence of which it is difiicult to demonstrate the latter. It has 

 likewise not been my good fortune to trace and learn what is the fate 

 of the germinative vesicle of the mackerel egg, but it will suffice to say 

 that when the egg is mature it can no longer be distinguished ; nothing 

 whatever remains to indicate its former position, and the whole egg is 

 now more transparent and presents the appearance shown in Fig. 2. 

 The germinal layer or pellicle, however, is found to include a great 

 number of very minute refringent corpuscles scattered through its sub- 

 stance ; these disappear as the germinal disk is formed by the aggrega- 

 tion of the protoplasm of the germinal layer at one pole of the vitellus 

 to form the germinal disk. I have frequently seen them apparently 

 dissolve and disappear while I was observing them through the micro- 

 scope. Whether these represent the remains of the more fluid and 

 refringent germinative vesicles I am unable to say, but I am inclined to 

 doubt it from the fact that if the germinal pellicle is removed and 

 stained with haematoxylon these corpuscles do not tinge, while the pro- 



