374 bulletin: museum of comparative zoology. 



it represents a mitosis which, for some reason or other, is never com- 

 pleted. Further than this we can say little, except that, as Woltereck 

 has shown, it can have nothing to do with reduction, because in the 

 parthenogenetic eggs of Cypris the chromosomes appear in the full 

 somatic number in the first polar spindle. 



d. The Chromosomes of the Cleavage Spindles. 



One of the most interesting features which have appeared from this 

 study is the occurrence in Gonionemus of a reduced number of chromo- 

 somes in the early cleavage spindles. I have already (p. 360) called 

 attention to the possibility that this condition may perhaps not be typi- 

 cal of this animal, but may be in some way abnormal. To settle this, 

 however, requires renewed examination of fresh material of this same 

 form. Till such material is available I prefer to regard the process 

 as the normal one, the more so since there is nothing in the preparations 

 to suggest that it is in any way abnormal. 



A similar case has been described in Cyclops by Hacker ('95), who 

 discovered that the early cleavage spindles show only half the somatic 

 number of chromosomes. Each of these chromosomes, however, is shown 

 by its form to consist of two ordinary chromosomes joined end to end, 

 and is therefore bivalent. The reduced number of chromosomes is per- 

 manently retained by the primordial germ cells, which are differentiated 

 as early as the eight-cell stage. But the ordinary somatic cells after this 

 stage acquire the full somatic number. This alteration, as E. B. Wilson 

 (:00, p. 274) remarks, "must consist in the dividing of each bivalent 

 rod into its two elements." Although no exact parallel to this condition 

 seems ever to have been described, vora Bath ('93, p. 106, footnote) 

 states that he has found nuclei with only half the somatic number of 

 chromosomes in blood cells and cells of the primitive kidney and midgut 

 of larval salamanders. 



Although the conditions in Gonionemus closely resemble those in Cy- 

 clops, an apparent difference is found in the fact that in Cyclops there 

 are only half the somatic number of chromosomes in the first cleavage, 

 while in Gonionemus there are the full number in this division. But 

 this divergence is more apparent than real, since it is caused merely 

 by a variation in the time when the pairing of the chromosomes takes 

 place. Hacker's ('95) figures show very clearly the exact method of this 

 pairing in Cyclops. In his Figure 45 — a polar view of the early meta- 

 phase of the first cleavage immediately after the disappearance of the 



