FORMATION OF THE GERMINAL LAYERS IN TELEOSTEI. 217 



impregnation. He does not, however, state the temperature at which the 

 observations were made, but I gather from remarks in another portion of the 

 paper that this was probably between 60° -8 and 64° 4 F. 



The influence of temperature on the rate of development of fish ova is 

 very great, and has been already studied by Meyer in the case of the herring. 

 The sensibility of the ova of the herring to a change of temperature is almost 

 as marked as that of pelagic fish ova. There can be no doubt, however, that 

 under natural conditions the eggs of the herring are not so liable to sudden 

 changes of temperature as are those which float at or near the surface of the 

 sea. The range of temperature at which the eggs of the herring will develop 

 normally is much wider than is the case for the eggs of the Salmonidse. 



On the completion of the four-cell stage — that is, after the formation of the 

 third furrow — the segmented disc measured in one case 9407 mm. in diameter, 

 and *3198 mm. in thickness ; in another egg the measurements were *9595 

 mm. and "2822 mm. respectively. After the furrows are completed, the 

 individual cells separate more or less from one another. This is accomplished 

 in the following manner : — The adjacent cells begin to separate at the outer 

 limit of the furrow between them, and at the same time a similar process 

 commences at the inner extremity of the same furrow. Thus, at the point 

 where the two furrows cross one another, there is a space formed, owing to 

 the protoplasm receding somewhat from the former point of contact. In this 

 respect the segmenting disc of the herring ovum presents a very different 

 appearance from that which is seen in many other fish ova, for instance those 

 of the Salmonidae. Such an arrangement is of frequent occurrence amongst 

 the Invertebrata, and is there connected with the formation of the segmenta- 

 tion cavity. In the herring the segmentation cavity arises at a much later 

 stage, and has no connection with this structure, which, indeed, is here 

 only a temporary one. As the cells become more completely separated, the 

 central cavity is lost. This mode of separation of the cells in the early 

 segmentation stages is not confined to the herring amongst fishes. A similar 

 phase is found in the ova of Perca, Leuciscus, and other forms. In the species 

 which show this partial separation of the early segmentation spheres, the 

 cells at a later stage are always loosely aggregated together. On the other 

 hand, in the group of which Salmo may be taken as the type, the cells are 

 never so completely separated from one another, and there is almost an 

 entire absence of those large spaces between the cells during the segmentation 

 stage, which are of such common occurrence in the herring ovum. In this 

 respect Salmo approaches more nearly to the Elasmobranch type, and the 

 difference is probably connected with the distribution of food yolk. 



Under normal conditions the separation of the four cells is never complete 

 in the herring ovum. They present the appearance of four conical mounds 



