ORIGIN OP HYPOBLAST IN PELAGIC TELEOSTEAN OYA. 31 
not clear, and the lower margin of the blastodisc becomes 
quite flat again before the marginal periblast is pushed under 
it. The gradual development of the periblast (there called the 
intermediary layer) in Trachinus has been shown in my 
paper already referred to, and it is not necessary to repeat it 
again here. About the time free cells are formed in it, a 
section has the appearance shown in fig. 1. The epidermal 
layer of the epiblast is already differentiated, and there is no 
segmentation cavity; nor, so far as I can make out, does the 
periblast extend quite under the disc. As early as the sixteen- 
cell stage I have noticed that the central cells of the disc do 
not lie on the yolk, so that there is a shallow cavity between 
the disc and the yolk in its central area ; but this becomes 
filled up as segmentation goes on, and does not represent the 
true segmentation cavity in which the hypoblast is formed. 
As to whether the nuclei in the periblast arise by true free 
cell-formation or are derived from a subdivision of the lower 
part of the first cleavage spindle I cannot say, but in Tra- 
chinus, at least, there appears to be no doubt that they do 
not come from the margin of the germinal disc, as Whitman 
asserts is the case in Ctenolabrus. If, as Hoffmann asserts, 
the first cleavage spindle formed of the male and female pro- 
nuclei really does divide at right angles to the axis of the egg 
so as to form at the outset two layers, his archiblast and para- 
blast (non Klein), it would appear more probable that the 
nuclei found later should be the result of the subdivision of 
the original nucleus of the layer than that they should arise 
independently. In Trachinus and Motella the ring of peri- 
blast gradually becomes more granular before cells appear, the 
granules cluster together in groups, and it has certainly 
appeared to me sometimes as if free cell-formation really did 
take place. Nothing short of a careful investigation of this 
stage in a large number of different forms may be expected to 
settle this question definitely. 
Next as to the origin of the hypoblast, Hoffmann and Messrs. 
Agassiz and Whitman, though differing in their ideas as to the 
origin of the periblast, are all equally confident that this layer 
