547 



parts, but the egg passed into a marked resting- stage before the furrows 

 were completely formed. At the next cleavage the egg divided into 16 

 spheres, and in succeeding stages into 32, 64 etc. wich great regularity. 



Besides these modes of development one very aberrant form was 

 observed. In the earliest stage observed the egg consisted of three 

 large spheres and four much smaller ones. These spheres then divided 

 in somewhat irregular succession, the inequality in size remaining for a 

 long time very marked. The spheres ultimately became, as usual, 

 nearly equal , the larvae acquired cilia and developed as well as the 

 others. 



Besides the above-described forms 1 have observed eggs divided 

 into four equal and sharply defined spheres, but unfortunately did not 

 prove them capable of further development. In view, however, of the 

 great variations which certainly exist, there is little reason to doubt 

 that these eggs were normal and capable of full development. 



In all the foregoing cases the first cleavage was definite and clearly 

 defined. It sometimes happened, liowever, that the first cleavage was 

 only imperfectly carried out and the egg passed into a resting stage be- 

 fore the spheres were fully formed. The egg was in several instances 

 observed to divide imperfectly and irregularly into 8 spheres, but a well 

 defined cleavage did not occur until the next active stage, when 16 

 distinct and regular spheres were formed. 



In a larger number of cases the egg underwent considerable chan- 

 ges of form, but without the occurrence of any cleavage, some time 

 before the first actual division of the vitellus. This appears to repre- 

 sent a kind of attempt at cleavage which does not go further than the 

 division of the nuclei within the egg. 



I have made many series of sections through the egg before clea- 

 vage and have been able fully to establish the fact that the nucleus 

 divides regularly, as in some Crustacean eggs, during the period which 

 precedes the cleavage of the vitellus, so that when the first cleavage 

 does occur each sphere contains one of these nuclei. 



The variations in the segmentation appear therefore to depend in 

 part upon the period at which division of the vitellus follows that of 

 the nuclei. If at an early period the vitellus may divide into two, four 

 or eight spheres. If at a later period cleavage may not take place until 

 16 or 32 nuclei have been formed. 



In some cases the vitellus divides completely (so far as external 

 features go) with the nuclei, in other eases only incompletely, and in 

 still other simply changes its form without the occurrence of any actual 

 division. 



These various forms of development will be fully described and 



