J. W. H. Harrlson and L. Doncastbr 245 



the only possible arrangenient ; I give it merely as an illustration of 

 the way in which I regard the hirtarki chnunosomes as perhaps being 

 composed of a variable number of units each corresponding with one 

 zoniirid chromosome. It is of course probable, if the hirtaria chromo- 

 somes are compound, that those oi zonaria are so also, and that the true 

 " unit," whatever that may be, is smaller than either. Meek' has shown 

 that the chromosome lengths in various species of insects and other 

 animals are terms of an arithmetical series, as if they were composed of 

 units of standard length, and the comparison nf hirtiiria with zonariit 

 leads by a different route tn a somewhat similar conclusion. 



Another aspect of the individuality question is touched by the 

 phenomena of the maturation divisions of the hybrids. I have shown 

 that although the majority of the chromosomes fail to pair in synapsis, 

 it is certain that some of th(un do so. The number of those which pair 

 is sometimes not more than ten (five pairs), giving a maximum number 

 of 65 in first sjDermatocyte equatorial plates in the cross zonaria % x 

 hirtaria ^ ; in other cases, especially in the converse cross, the number 

 of pairs appears sometimes at least to be larger. If I am right in 

 suggesting that about three of the small hirtaria chromosomes corre- 

 spond with single zonaria units, we might expect these to pair normally 

 together, and the indications of occasional unequal pairing suggest that 

 some of the compound hirftiria chromosomes may pair also, perhaps le.ss 

 regularly, with zonaria units. This would lead to the conclusion, not 

 that chromosomes are individuals in the sense of being indivisible units, 

 but that they are composed of units, and that pairing in synapsis i.s due 

 to some affinity between chromosomes made up of similar components. 

 This is of cour.se the hypothesis which has been widely accepted on 

 evidence of a different kind, in connexion with the hypothesis that the 

 chromosomes are the bearers of Mendelian unit characters. 



If there is any truth in this view, it may give some clue to the 

 baffling problem of interspecific sterility. The hirtaria — zonaria hybrids 

 are completely sterile. If the haploid set of 56 chromosomes of zonaria 

 consists on the whole of the same units as the 14 haploid chromosomes 

 of hirtaria, combined in different ways, the sterility between the two 

 nearly allied species may have arisen simply from a difference of 

 grouping. Since the units are gi'ouped differently, the chromosomes 

 cannot pair properly in synapsis, and this may be the cause of the 



' C. F. U. Meek, "A metrical analysis of chromosome complexes, etc.," Phil. Trans. 

 Roy. Hoc. Vol. B. 203, p. 1, 1912. 



Jouru. of Gen. in 17 



