FORMATION OF THE GERMINAL LAYERS IN TELEOSTEI. 205 



egg through this channel. It would seem, however, on a priori grounds, that 

 a micropyle can be of little service in aiding fertilisation in such an egg as that 

 of the herring. Around the egg envelope at the time the egg leaves the ovi- 

 duct there is a comparatively thick layer of a viscous substance which hardens 

 in sea water. So long as it remains semifluid it collects at the lower pole, and 

 if two or more eggs are in contact, the viscous substance around each forms a 

 thick ring around the point of contact, and thus the whole mass is cemented 

 together. It follows, therefore, that the moment an egg touches any object in 

 its descent through the water it becomes attached to that object, and does not 

 again change its position during development. It seems probable that the 

 micropyle cannot have a fixed position in relation to the axis of the ovum, as 

 is the case in pelagic ova. If this supposition is justified, it follows that the 

 micropyle must frequently be covered by the thicker portions of the hardened 

 viscous layer. In any case the opening of the micropyle must be filled with 

 the viscous layer before this hardens, as Kupffer has already pointed out. 

 Thus spermatozoa would find as much difficulty in penetrating the ovum at the 

 micropyle as at any other point. It has been proved by repeated experiment 

 that herring ova may be fertilised from twelve to twenty-four hours after they 

 have been placed in sea water, and by this time the viscous layer is so hard that 

 the ova are not easily displaced from their point of attachment. In any case, 

 therefore, spermatozoa are able to penetrate this hardened layer, which indeed 

 offers more resistance than the egg membrane itself. It has been generally 

 admitted that in fish ova the spermatozoa enter by the micropyle and by this 

 only. The presence of a viscous covering of the egg envelope in such ova as 

 that of the herring may modify the use of the micropyle. If the viscous 

 covering of the eggs of the herring and allied forms is phylogenetically of recent 

 origin, it may be that the changed conditions have rendered the micropyle 

 useless. These remarks are, however, simply offered as suggestions, a 

 thorough investigation of the whole subject being very desirable. 



The behaviour of the germinal protoplasm during the time that a ripe 

 unfertilised ovum remains in sea water cannot easily be observed in the living 

 egg. In order to investigate this point, I have cut sections of a large number 

 of eggs which have remained unfertilised in sea water for a time varying from 

 one to forty-eight hours, and I have also examined sections of ripe ova taken 

 from females which had been dead some hours. As I have previously described, 

 the germinal protoplasm at first forms a comparatively even network between 

 the yolk spheres. There is, it is true, a little more protoplasm at the surface of 

 the yolk than towards the centre, but it must be remembered that the yolk 

 spheres are smaller towards the surface, and there is thus more room for the 

 protoplasm. This exact relationship is not long maintained in any case. There 

 is, however, a difference in its behaviour in the fertilised and in the unfertilised 



