382 EDMUND B. WILSON 



In some cases the two arms actually meet, uniting to form a 

 closed ring, of the type first made known by Paulmier ('98) and 

 often observed since, both in insects and in other animals, such as 

 Tomopteris, or the grasshoppers; but this type is much rarer in 

 Oncopeltus and Lygaeus than in some other species. (3) The 

 third type is that of the tetrad-rod, which consists of a straight rod, 

 which shows both a longitudinal split and a transverse median 

 suture. These forms are readily deducible from the double cross 

 by reduction and final disappearance of the lateral arms, the posi- 

 tion of which is now indicated by the transverse suture. As will 

 be shown hereafter (especially in the case of Protenor) the double 

 crosses undergo in their later stages precisely this change ; but the 

 evidence indicates that some of the tetrad-rods never pass through 

 the double cross stage. (4) The fourth type is the double- F, 

 best described as a F-shaped figure that is longitudinally split in 

 the plane of the two branches, from the apex of the V towards the 

 free ends, accompanied by a greater or less degree of separation 

 of the two halves thus produced. Figures of this type are espe- 

 cially common in the earlier stages (fig. 108, photos. 17, 18) and may 

 be recognized soon after the beginning of Stage hin a much more 

 elongate form, as shown in fig. 107, photos. 17 (from a smear- 

 preparation). In the final prophases (Stage j) all the bivalents 

 finally condense to form dumb-bell figures, though the double 

 crosses (now much condensed, and often more or less opened out 

 in a ring form) may sometimes still be distinguished in the early 

 metaphases. In the course of this process the lateral arms of the 

 crosses sooner or later disappear, and a cross constriction appears 

 at the points where they have been. These conditions will be 

 more fully considered later in the case of Protenor. 



Owing to the uncertainty regarding synapsis and the impossibil- 

 ity of tracing the bivalents individually through the confused 

 period, it is not possible to offer more than a somewhat conjectural 

 interpretation of the origin and relationships one to another of 

 these various forms. Paulmier, McClung and other earlier stu- 

 dents of the insects assumed the primary type to be a tetrad-rod, 

 representing two univalent chromosomes, united end to end, and 

 longitudinally split. From this type the double cross was assumed 



