THE REDUCTION OF THE CHROMOSOMES 227 



form, at that time a purely theoretical postulate, he assumed to be of 

 such a character that each daughter-nucleus should receive only half the 

 number of ancestral germ-plasms possessed by the mother-nucleus. 

 This he termed a reducing division (Reduktionstheilung), and suggested 

 that this might be effected either by a transverse division of the chromo- 

 somes, or by the elimination of entire chromosomes without division. By 

 either method the number of 'ids' would be reduced; and Weismann 

 argued that such reducing divisions must be involved in the formation 

 of the polar bodies, and in the parallel phenomena of spermatogenesis.' 1 



Reduction in Weismann's sense, then, is a reduction of the number of 

 kinds of germ-plasm or ancestral hereditary qualities present, this 

 reduction being brought about by means of a redistribution, half of the 

 qualities to one daughter nucleus and the remainder to the other daughter 

 nucleus. The change in the number of chromosomes is a consequence of 

 the manner in which this redistribution is accomplished, as we shall see. 



Interpretations Based on Weismann's Theory. — As would be ex- 

 pected, there were announced certain interpretations of chromosome be- 

 havior based on Weismann's idea. Several cytologists thought that 

 they found the chromsomes actually dividing transversely at one or the 

 other of the two maturation mitoses. This interpretation, however, 

 proved to be incorrect. Much light w 7 as thrown on the problem when 

 Henking (1891), Rtickert (1891, etc.), Haecker (1890-9), vom Rath 

 (1892-3), and others showed that the double chromosomes appearing in 

 the reduced number on the spindle at the first maturation mitosis are not 

 split chromosomes like those seen in somatic divisions, but are pair- oi 

 chromosomes, or hivalent chromosomes, each arising b} r an end-to-end 

 conjugation (synapsis) of two somatic chromosomes. The two parts oi 

 each bivalent then separate at the first or second maturation division, 

 the entire chromosomes thus being segregated into two groups, each 

 with the reduced number. Thus it appeared unnecessary thai a single 

 chromosome, representing a linear series of different qualities, should be 

 transversely divided in order for Weismannian reduction to occur: it 

 was only necessary to assume that the whole chromosomes differ qualita- 

 tively from one another, so that when the two members of a bivalent 

 pair separate there would be a segregation of different qualities. It is in 

 the light of this bivalent chromosome conception thai we are to interpret 

 the many early reports of a transverse division of the chromosome during 

 maturation. What was called a transverse division was merely tin- 

 separation of two entire chromsomes placed end-to-end. 



A number of workers soon found that in many cases there Is nothing 

 even simulating a transverse division, either of single chromosomes or <>t 

 bivalent pairs, but that both maturation divisions are apparently longi- 

 tudinal (Flemming, Brauer 1S93, Moore L896, Meves L896, Gregoire, 

 etc.). How t , then, is there any Weismannian reduction if there is neither 



