414 EDMUND B. WILSON 



A particular interest attaches to the striking and constant con- 

 trast between the X-chromosome and the large bivalent; I do 

 not here refer to their conspicuous difference in texture and stain- 

 ing capacity in the earlier stages but to a characteristic difference 

 in morphological composition that persists up to the verymeta- 

 phase of the first division. The JT-chromosome is always bipar- 

 tite a simple rod, longitudinally split. Sometimes it is curved, 

 sometimes slightly constricted at the middle point, sometimes 

 irregular in form; but many of these variations are evidently 

 quite accidental. It never shows the least approach to the double 

 cross form nor does a transverse suture at the middle point make 

 its appearance. In the final prophase it enters the spindle at 

 right angles to. the latter, and undergoes a longitudinal division 

 (figs. 132 to 134, cf. Montgomery, '01, '06, Wilson, '11 a). The 

 large bivalent, on the other hand, is always, sooner or later, a 

 quadripartite body. In most cases it forms a fine double cross, 

 of the same type as that already described in case of Oncopeltus, 

 and showing the same variations. All gradations are found, in 

 different nuclei, on the same slide, between forms in which the 

 arms of the cross are equal (figs. 120, 122, 123, photos. 40, 41, 49) 

 and those in which one pair (the 'lateral 'arms) are but just 

 perceptible (figs. 121, 124, 125). In some cases, comparatively 

 rare, the lateral arms are quite wanting, and the bivalent appears 

 as a longitudinally split rod (figs, 126, 129, 130) but in these forms 

 a distinct transverse cleft or suture is often seen at the middle 

 point; and in a few cases the rod is sharply bent at this point to 

 form a F-shaped figure (fig. 127). Sometimes the transverse 

 cleft is not seen in the earlier stages : but always in the later pro- 

 phases 'it becomes evident (figs. 129, 130, 131, photos. 48, 51). 

 In these stages, as in Oncopeltus, the lateral arms of the double 

 crosses sooner or later disappear, being apparently progressively 

 drawn in to the axial arms, and in their place appears first a 

 transverse suture and later a constriction across which the first 

 division takes place. In the early metaphases the large bivalent 

 shows two extreme types, connected by all intermediate transi- 

 tions. At one extreme are ring-like tetrads (fig. 133) which are 

 evidently derived from crosses by shortening of the arms and per- 



