ularity to be due to segmental interchange. Otherwise 
the genome constitution of 4.barbata, which was repre- 
sented as AA B’B’ by Nishiyama (1986), seems identical 
to that of A. abyssinica. 
Hewxaploid <X Hexaploid—N ishiyama (1929) examined 
chromosome pairing in the F; of four hexaploid species 
combinations and found 21 normal bivalents in the ma- 
jority of cells. Sometimes, however, 1 or 2 bivalents failed 
to form properly, resulting in univalents, a trivalent or 
a quadrivalent. Nishiyama suggested these irregularities 
were probably caused by mating between semi-homolo- 
gous chromosomes which are not normal partners. Nishi- 
yama expressed the degree of affinity between the gen- 
omes of the various species by the ratio of the number of 
bivalents found in the F; to the number expected. Full 
affinity between the parental genomes was given the 
value of unity, while no affinity was equal to zero. The 
values obtained by Nishiyama are: 
A, fatua. . . . .0,.998. . . . . A. sativa 
A, fatua. . . » 10.992... . . A shertlts 
A.sativa. . . . .0.983. . . . . A. byzantina 
A. sterilis . . . . 0.986. . . . . A. byzantina 
Spier (1984) reported complete bivalent formation in 
the F, of A. sterilisX A. sativa. A few cells damaged by 
pressure showed two univalents. 
Joshi (1951) studied meiotic behavior in the Fi’s of 
crosses involving A. sativa, A. sterilis and A. fatwa and 
found 6-13% of the pollen mother cells showed irregu- 
larities, the nature of which was not stated. The paren- 
tal species showed less than 2% irregularity, while the 
I, of spring X winter varieties of A. sativa showed 14— 
31% of the cells having irregularities. In hybrids of 4. 
byzantina with A. sterilis, A. sativa and A. fatua 59-88 Yo 
of the cells showed irregularities. ‘These results are prob- 
[ 288 ] 
