THE CYTOGENETICS OF AUTOHETEROPLOID PLANTS 355 



plant was placed on the silks of an rr plant, the ear produced 13 colored 

 and 27 colorless kernels (expectation, 1:2). Pollen from an RRr 

 plant on silks of an rr plant gave 646 colored and 355 colorless kernels 

 (expectation, 2:1). Pollen from a normal rr plant on the silks of an 

 RRr plant gave 819 colored and 213 colorless kernels (expectation, about 

 4:1). A certain amount of deviation from expected ratios in such 

 cases is due to the lagging and loss of the extra chromosome in meiosis, 

 so that the spores and gametes with n chromosomes actually outnumber 

 those with n -\- 1. This results in a deficiency of dominant plants among 

 the offspring. 



In the foregoing example one dominant factor (R) is sufficient to 

 overcome the effect of two recessives (rr). Other genes are known of 

 which this is not true, the appearance of the triploid tissue depending 

 rather upon which type of factor outnumbers the other. ^^ In such cases, 

 of course, a different set of ratios is expected. 



5 ^ « i <» M H S 



Fig. 202. — Multivalent chromosomes in Prunus laurocerasus. {After Meurman, 19296.) 



Other Heteroploids. — The cytological and genetical features of plants 

 with five or more homologous chromosomes or chromosome sets are not so 

 well known as those of triploids and tetraploids. The chromosome 

 configurations in meiosis are very variable. In the pentaploid Tulipa 

 Clusiana (60 chromosomes) there are univalents, bivalents, trivalents, 

 quadrivalents, and quinquevalents; but among these the quinquevalents 

 are least frequent and the bivalents most frequent. In the octoploid 

 Dahlia variabilis (64 chromosomes) there are bivalents, quadrivalents, 

 sexavalents, and octovalents, with the last named the least frequent. 

 In Dahlia Merckii with 36 chromosomes (4n + 4) there are quadrivalents 

 and two sexavalents, suggesting that the plant is a tetraploid with two 

 of the eight chromosomes of the set still further reduplicated. In 

 the 22/1 Prunus laurocerasus there are many multivalent associations, 

 with trivalents most. common (Fig. 202). Such variations in plants 

 not known to be truly autoheteroploid may be due in part to incomplete 

 homology. The genetic ratios observed in such plants will obviously 

 depend upon the strength of dominance, the manner in which the multi- 

 valents are distributed in meiosis, and the viability and fertility of the 

 various heteroploid gametes and zygotes. ^^ 



1^ East and Hayes (1915) on Zea endosperm, von Wettstein (1923) on triploid 

 moss gametophytes. 



1^ Newton and Darlington (1929) on Tulipa, Lawrence (1929) on Dahlia, Meurman 

 (19296) on Prunus. Haldane (19306) gives expected genetic ratios for An, 6n, 8n, 

 lOn, 12/i, and 16« autopolyploid types. 



