than in the triploid A. longiglumisX A. abyssinica. Thus 
it is possible that one genome of A. sativa is homologous 
to the genome of A. longiglumis, but we do not know 
whether the genome of A. longiglumis is A or B’. 
Hewaploid XX Tetraploid—Nishiyama (1929) studied 
pairing behavior in the F; pentaploid hybrids of 4. bar- 
bataX A. fatua and A. barbataX A. sterilis. Chromo- 
some pairing was generally found to be loose with the 
bivalents formed often mated only at one end. On the 
basis of full homology 14 bivalents would be expected, 
but in A. barbataX A. fatua 2-11 bivalents inclusive of 
1—4 trivalents were found, while in 4. barbataX A. steri- 
lis 7-18 bivalents inclusive of 0-4 trivalents were found. 
The genome affinity between A. barbata and A. fatua 
was 0.456 and between A. barbata and A. sterilis it was 
0.675. These figures are even more significant when it 
is recalled that the chromosomes of A. barbata can pair 
among themselves as can the chromosomes of the hexa- 
ploids. 
Emme (1982) reported that meiosis in the F; of the 
two pentaploid hybrids A. sativa A. abyssinica and A. 
sativaX A. barbata showed 7 to 9 bivalents. Similarly 
Spier (1984) reported that meiosis in the F; of A. abys- 
sinicaX A. sterilis showed 5 to 11 bivalents with 28 to 18 
univalents per cell. ‘Trivalents were sometimes found 
but not quadrivalents. These results agree with those of 
Nishiyama in indicating a very low homology between 
the genomes of the tetraploid and hexaploid species 
studied. 
The lack of homology between the genomes of the 
tetraploids and hexaploids was further demonstrated by 
Lesik (1948) who obtained synthetic amphidiploids of 
A. sativaX A. abyssinica using colchicine. These plants 
had 85 pairs of chromosomes and 100% fertility. Ex- 
[ 286 ] 
