Duplication 395 



genetic ratios. If the duplicated segment lies next to the iden- 

 tical segment as at the Bar locus (that is, if the duplication is a 

 repeat), the position effect will sometimes result. 



McClintock has reported an interesting situation in which 

 one allele for brown midrib, Bm, when present as a duplication, 

 produced nonbrown midrib tissue in an otherwise homozygous 

 6m bm (brown midrib) plant. By X-rays two deficiencies were 

 produced in chromosome V in maize. Each deficiency was a 

 small, intercalary one that became a small ring, and each carried 

 the Bm gene. Strangely, in each deletion, the original centro- 

 mere had broken in half so that both the small ring deletion and 

 the rod-shaped remainder of the chromosome had half a centro- 

 mere, and each half centromere was functional. When a plant 

 possessed two normal chromosomes V, each of which had the 

 bm gene, and one of the rings with the Bm gene, it was non- 

 brown midrib because one Bm gene was dominant over two bm 

 genes. The small ring-shaped chromosomes, however, behaved 

 abnormally in somatic mitoses. They were lost from the nuclei 

 or they changed in size. In these Bm duplications the plants 

 were green because of the Bm gene but had streaks of brown 

 midrib tissue. Cytological examination showed that most of 

 these streaks of brown midrib tissue had cells in which the ring 

 chromosome bearing the Bm gene became lost at some previous 

 somatic division. The ring chromosome was lost during any 

 stage of development. If it was lost very early, the entire plant 

 was brown midrib; if it was lost later, there were wide sectors 

 of brown midrib tissue; and if lost very late in ontogeny, there 

 were merely small patches of bm, cells. Thus by the peculiar 

 behavior of these chromosomes during somatic development, 

 many types of variegated brown midrib plants were produced. 



The way these ring chromosomes change in size is very inter- 

 esting. Let us assume a ring whose segments are numbered 1 

 to 8 (Fig. 103). After it has divided, in some mitotic prophases 

 the two sister halves form one continuous ring with two centro- 

 meres instead of two freely separating rings with one centromere 

 each. Whether this occurs by a somatic crossing over or in some 

 other way has not been determined. At metaphase, this small 

 dicentric ring opens out on the equator with each centromere 

 orientated towards one pole. Especially if the original ring is 

 very small, the new dicentric ring usually remains at the equator 



