ORTHOPTERAN SPERMATOGENESIS 711 



shows as a darker spot at both high and low focus. Usually these 

 two extensions fall within the width of the ring's wall and both 

 are in the same vertical axis. Sometimes the annular body is more 

 pointed at this inner side and the extensions may fall beyond the 

 outer limits of the ring proper. Under such circumstances the 

 enclosed space becomes drawn out in this direction also and passes 

 somewhat into the extended wall of the ring (figs. 21 to 30, 129 to 

 133) . Should the diameter of the curved rods be large in proportion 

 to their length, the enclosed space is small, or may even disappear, 

 and a cleft shows in the outer margin where non-synaptic ends are 

 in contact (figs. 2, 3, 11). Viewed laterally or en face the annular 

 chromosomes are with difficulty distinguishable from the V's and 

 crosses (figs. 8, 9, 10, 25, 111), but oblique views are very clear 

 and convincing (figs. 83, 133). (5) In division, during the sper- 

 matogonial mitoses, chromatids lying superimposed, are sepa- 

 rated from each other, begining at the inner end, where the fiber 

 attaches, and moving in the same vertical plane to opposite 

 poles of the spindle. Mecostethus shows exactly the same con- 

 ditions in the first spermatocyte except for the fusion in pairs 

 of homologous chromosomes. Ring-shaped chromosomes of 

 the Hippiscus type maintain the same spatial relations in the 

 cell and undergo similar movements of the chromatids. 



(6) Of much importance is the fiber attachment for determin- 

 ing fixed points on the chromosomes. As has been shown, par- 

 ticularly in Stenobothrus, the point at which the fiber attaches 

 remains the same from one generation of cells to the other. 

 Where all the chromosomes are straight rods in the spermato- 

 gonium this point is always at the inner ends, and, in the first 

 spermatocyte, the chromosomes occupy a similar position, as 

 in Mecostethus, with the single exception that they are now 

 joined in pairs. Where rings appear these relations are again 

 preserved, only fusion has occurred at both ends of the univalent 

 chromosomes. Terminal fiber attachment and rings of the 

 Plippiscus type always occur together. (7) On separation all the 

 anaphase chromosomes are simple V's and the former elements 

 of the rings are no longer distinguishable from the other deriva- 

 tives (figs. 5, 6, 7, 35). 



