ably comparable to those of Nishiyama (1929) who, in 
hybrids involving A. byzantina as one parent, found uni- 
valents in 29-83% of the cells examined, while in hy- 
brids involving only A. sativa, A. sterilis and A. fatua, 
univalents were found in 5-16% of the cells. 
From the above evidence it is clear that a high degree 
of homology exists between the genomes of the various 
hexaploid species. Nishiyama (1989) gave AABBCC as 
the genome constitution of 4. fatua and it is probable 
that this applies to all the hexaploid species. 
Tetraploid X Diploid—Meiosis in the triploid Fy of 
A. barbataX A. strigosa was investigated by Nishiyama 
(1929) who found the close genome affinity of 1.041 be- 
tween the two species. Usually 7 bivalents inclusive of 
trivalents were formed, but sometimes 8-9 bivalents in- 
clusive of trivalents were found. The number of triva- 
lents was 0-8 and rarely a quadrivalent was seen. How- 
ever, at least 8 lagging univalents were found at meta- 
phase I of all cells. Spier (1984) made the same cross 
and reported similar results, except that at least 4 uni- 
valents were found in all cells. 
Nishiyama (1936) made a further study of the genome 
constitution of A. barbata by crossing it with a synthetic 
autotetraploid of A. strigosa having the genome consti- 
tution of AAAA,. Inthe Fy; hybrid 12 bivalents were 
frequently found and sometimes 18. The number of uni- 
valents found varied from 0 to 5, but in most cases was 
1 or 2. In an extreme case all chromosomes entered into 
pairing relations, forming 11 bivalents and 2 trivalents. 
Thus it is evident that one genome of A. barbata is com- 
pletely homologous with genome A of A. strigosa, while 
the second is partially homologous to it. Nishiyama 
represented the genome constitution of A. barbata as 
AAB’B’. 
[ 284 ] 
