OETHOPTERAN SPERMATOGENESIS 693 



but it would appear from several statements that this must 

 be due to union of chromosom.es, and not to a unique cross- 

 splitting of a continuous spireme. On page 527 this statement 

 occurs: "It does not seem to be a matter of any consequence how 

 the bivalent arrangement is produced, since there is so much 

 variability in the process, but the temporary union in pairs of 

 somatic chromosomes is the really important feature." (2) For 

 Periplaneta the analysis of the tetrads into their univalent parts 

 is not made clearly, but the idea of variability of association 

 and dissociation is held. How much this means for these authors 

 may be gathered from the description of the different methods 

 by which the com.mon X-shaped chromosome is produced in Os- 

 munda (p. 522). At least three variations in constitution are 

 mentioned. In Periplaneta the prophase chromosomes are in 

 the form of loops with their free ends polarized at one side of the 

 nucleus. A chromatin nucleolus is present. (3) The metaphase 

 chromosomes are in the form of rings and are produced by approxi- 

 mation of the free ends of the loops. (4) Upon the spindle the 

 rings are so arranged that they lie in its axis. They are there- 

 fore of the Stenobothrus type. (5) In the metaphase the annular 

 chromosomes are separated into half rings. (6) Fiber attach- 

 ment is median as in the spermatogonia. (7) The anaphase 

 chromosomes are irregular half rings, or V's, with the original 

 longitudinal cleft of the chromatin thread dividing them into 

 double elements. (8) As a result of the first spermatocyte divi- 

 sion the reduction, or segregation, of the biparental chromosome 

 groups is effected. (9) After a very nearly complete rest stage 

 the reduced number of chromosomes reappears in the second 

 spermatocyte metaphase and there they divide "in the ordinary 

 premaiotic manner." 



Stevens ('05) describes the maturation phenomena of (Blat- 

 tella) Blatta germanica with these results: (1) The 23 chromo- 

 somes of the spermatogonium become converted by telosynapsis 

 into 11 tetrads, leaving the unpaired accessory chrom.osome. 

 (2) Rings, crosses, U's and rods occur in the first spermatocyte 

 prophase but they are of essentially the same type of construc- 

 tion. (3) Variation disappears in the metaphase and all the 



