STUDIES ON CHROMOSOMES 415 



sisteiice of the central opening. At the other extreme are short 

 tetrad-rods (fig. 132) which clearly show two division-planes at 

 right angles to each other. These forms may be derived either 

 from the more elongate rod-like forms of earlier stages or from 

 crosses by disappearance of the lateral arms. In both types the 

 quadripartite structure is unmistakable. The large bivalent 

 ultimately assumes a dumb-bell form, with its longer axis parallel 

 to that of the spindle-axis, and undergoes a 'transverse' division; 

 while, as already stated, the X-chromosome always enters the 

 spindle with its long axis transverse to that of the spindle, and 

 undergoes longitudinal division (figs. 132, 134; see also Wilson, 

 '11, fig. 9, Montgomery, '01, figs. 136, 138). 



No observer who studies these nuclei attentively can fail to be 

 struck by the remarkable difference between the large bivalent 

 and the large X-univalent. Its explanation is obvious ; the former 

 is preparing to divide twice, the latter once, in the course of the 

 two maturation-divisions. But this does not yet touch the root 

 of the matter. We have still to ask why two chromosomes of 

 equal size in the same nucleus differ so widely in respect to their 

 mode of division. The reply is again obvious. It is because one 

 of them has double the chromosomic value (or valence) of the 

 other, the bivalent representing two chromosomes of the original 

 diploid groups, while the univalent represents but one. This 

 conclusion, which is hardly more than a statement of fact, is 

 confirmed by a very interesting anomaly shown in fig. 128, and in 

 photo. 43. This nucleus contains nine chromosomes, and the 

 large bivalent is absent as such, while in its place appear two sepa- 

 rate and equal chromosomes of half its size (B, B in the figure), 

 while the four smaller bivalents are plainly recognizable {h-h) . The 

 explanation obviously is that through an abnormality of synapsis 

 the two members of this particular pair have failed to unite and 

 have therefore remained univalent. Both of these chromosomes 

 have the form of simple, longitudinally divided rods, without trace of a 

 quadripartite structure, while the four smaller bivalents all show the 

 cross-form (though the lateral arms are but slightly developed in 

 one of them) . It may be surmised, I think, that if the later history 

 these separate univalents could be followed out, they would be 



