MUSEUM OF COMPARATIVE ZOOLOGY. 101 



the micropyle. I believe it will be seen from his results that there is 

 still very good reason for maintaining that the tubular or columnar 

 structures of the capsular membrane, which have been the objects of 

 so much study, are derived from the granulosa cells, one from each cell, 

 and that the process by which the capsular membrane is formed is 

 neither simply a cell secretion nor exclusively a cell metamorphosis. 

 Although Eigenmann has not succeeded in getting stages which show 

 clearly all the steps in the formation of the capsule in Perca, he has 

 shown that there exist conditions in the later stages of the development 

 of the egg in Esox (Eigenmann, '90, Plate III. Fig. 37) which seem to 

 me of considerable importance in interpreting the conditions in Perca. 

 In Esox the cells become elongated, and the central (axial) portion 

 retains the granular and stainable properties of unmodified cell proto- 

 plasm. This axial portion is not cylindrical, but conical ; its base is 

 directed outward and contains the nucleus. The peripheral portion 

 — which is more and more abundant as one approaches the zona — 

 is more homogeneous than the axial part, and reacts with dyes in a 

 different way. The cell boundaries have been previously lost. The 

 boundary between these two constituents of the cell is not at first 

 sharp, so that this phase of the process may perhaps be regarded as 

 one of metamorphosis rather than of secretion. 



I believe that Perca must pass through some such stage as this during 

 the earlier part of the process which produces the capsular membrane. 

 I imagine that the distinction between the axial and the peripheral por- 

 tions of the cell becomes more and more sharply defined as the thicken- 

 ing of the capsule goes on. Meanwhile the axial portion does not long 

 retain the indifferent condition, but is metamorphosed, especially at its 

 periphery, into a highly refracting substance, so that there is reason for 

 regarding the structure as tubular. This metamorphosis advances till it 

 has practically obliterated the cell, even though a nucleus with a small 

 amount of enveloping protoplasm may still be made out at its distal 

 end in very late stages of ovarian growth. At any time before this, and 

 after the distinction between a funnel part and a tubular part has arisen, 

 the less modified distal portion of the cell may doubtless be easily sepa- 

 rated from the secreted gelatinous substance and also from the meta- 

 morphosed cell process. Such at least is the view which I have formed, 

 after a comparison of the granulosa in Perca and in Esox. 



It would certainly be remarkable if the perch were the only represent- 

 ative of bony fishes in which such a process took place. I believe that 

 there are a few cases already known which may prove upon renewed 



