DAVIS: SPERMATOGENESIS., 127 



pole, as described by the Schreiners, but believe this an accidental 

 arrangement, which is more common near the pole, since in this 

 region the threads are crowded more closely together. Moreover it 

 seems hardly probable that the chromosomes should conjugate granule 

 by granule. 



Among the species in which the Schreiners have described the side 

 bv side union of the chromosomes is Salamandra, but both INIontffom- 

 ery (:03, :04) and Moore and Embleton (:06) have found an end to 

 end vmion in closely related urodeles. 



According to the Schreiners each of the polar loops later becomes 

 converted into a tetrad in the following way: The two conjugants 

 become widely separated, remaining connected only at one or both 

 ends, to form loop- or ring-shaped elements, and at the same time a 

 longitudinal split appears in each conjugant. This necessitates a 

 very rapid shortening and thickening of the chromatic threads and the 

 simultaneous appearance of a longitudinal split, but on both these 

 points their figures are far from convincing. 



It will be seen that the end result is the same as in Orthoptera, i. e. 

 that the longitudinally split arms of the loops, or the opposite sides of 

 the rings, represent the imivalent components. 



The tA'pes of bivalent chromosomes described by these authors 

 in the prophase of the first division strikingly resemble those found in 

 the Orthoptera. 



4. The Maturation Divisions. ' 



The extensive literature on this important stage in the development 

 of the germ cells has been so often, and so thoroughly reviewed that 

 it will be unnecessary for me to attempt an extended discussion here. 

 The entire problem of maturation is at present in a very unsatisfactory 

 state and there seems to be no prospect, for some time at least, of 

 reducing the widely diverse accounts to a common basis, although 

 the striking similarities in the shape and behavior of the bivalent ele- 

 ments in widely separated forms leads to the hope that in time this 

 may be accomplished. 



In any discussion of the subject the method of formation of the 

 bivalent chromosomes must be taken into account and this has, of 

 course, been done only within very recent years. For this reason the 

 older accounts of the maturation division — such as those of Boveri 

 ('87), Flemming ('88), Hertwig ('90), Brauer ('93), Meves ('96), von 

 Lenhossek ('98), McGregor ('99), and Kingsbury (:02), in which 



