A STUDY IN MORPHOLOGY. 71 



different planes would not intersect the same number in all cases, and there may have 

 been two more in fig. 15 than in fig. 13. 



I think, then, that the facts indicate that c of fig. 11 is a yolk-pyramid, rather than 

 a primary mesoblast, and that after it divides into two, as in fig. 13, each part gives 

 rise to a central portion c, and a peripheral endoderm cell. 



If we accept this view and regard the cell c as a yolk-pyramid, two views as to the 

 relationship between the egg of Lucifer and an ordinary Crustacean egg at once 

 suggest themselves. 



We may hold that Lucifer presents the primitive or ancestral form of segmentation, 

 of which centrolycethal segmentation is a secondary modification. In this case we may 

 suppose that as the supply of food-material gradually increased, new food-bearing cells 

 or yolk-pyramids were added until all the cells were included, and the segmentation 

 cavity was entirely filled and obliterated by them. 



According to the other view, we may hold that the segmentation of the Lucifer egg 

 is a secondary modification, which has been brought about by the gradual reduction of 

 the amount of food-material, and its restriction, at last, to a single one of the cells of 

 the segmenting egg. 



There does not seem to be much difficulty in deciding which of these views is most 

 satisfactory and probable. Lucifer is undoubtedly a very primitive Malacostracan, but 

 it can hardly be regarded as a primitive Crustacean ; and the occurrence of perfectly 

 centrolycethic segmentation in the Copopods, Phyllopods, Amphipods, and Isopods, as 

 well as in the Decapods forms below as well as forms above Lucifer forbids us to 

 believe that the egg of Lucifer is ancestral, or the unmodified descendant of an ances- 

 tral type of egg ; and we must therefore believe that the egg of Lucifer has been 

 simplified by the loss of the greater part of its food-yolk. 



A change of this kind is not without a parallel, and I have shown (" The 

 Acquisition and Loss of a Food- Yolk in Molluscan Eggs," ' Studies from the 

 Biological Laboratory of the Johns Hopkins University/ vol. i., part iv.) that the 

 resemblance between the segmenting egg of the Oyster and a Molluscan egg with a 

 food-yolk can only be explained by the supposition that the Lamellibranchs have 

 inherited a rudimentary food-yolk which was functional at some past time, and that 

 the assumption gives an explanation of all the peculiarities of oyster segmentation. 



If we accept this view, and regard the egg of Lucifer as simplified by secondary 

 change, it is extremely instructive to note that the loss of a food-yolk has brought it 

 back to a type of segmentation which is directly comparable with that of ordinary 

 Metazoan eggs, and we must therefore believe that a segmentation cavity is poten- 

 tially present in all centrolycethic eggs, or else that the segmentation cavity of the 

 egg of Lucifer is not homologous with that of ordinary eggs. 



