amination of meiosis showed 35 bivalents were usually 
formed, but sometimes 2-8 univalents were observed. 
Conclusions 
The evidence from cytology indicates that speciation 
on the diploid level in Avena has involved structural 
changes in the chromosomes, but these changes have not 
been extensive in the species so far investigated. Our 
ignorance of the genome constitution of many of the 
diploid species is complete. 
The evidence indicates that the tetraploid Avena spe- 
cies have two partially homologous genomes which are 
also homologous to the genomes of the diploid species 
investigated. Whether the tetraploids are auto- or allo- 
polyploids is not clear. Evidence on this point could 
readily be obtained by synthesizing tetraploids involving 
the various diploid combinations. The tetraploid species 
may have arisen independently or they may be derived 
from a common tetraploid ancestor. 
The origin of the hexaploid species is obscure. While 
it is probable that at least one diploid Avena genome is 
present in the hexaploids, it is certain that 4. barbata 
and A. abyssinica played no part. The identity of the 
tetraploid which did contribute two genomes to the hex- 
aploids is a complete mystery. As with the tetraploids, 
the evidence from cytology does not indicate whether 
the hexaploids arose independently by allopolyploidy or 
whether the different species are derived from a common 
hexaploid ancestor or whether a combination of these 
processes has occurred. 
Part III. THe EvipeEnceE FROM GENETICS 
Studies on the inheritance of taxonomic characters in 
the genus Avena have been largely restricted to the man- 
ner of attachment of the florets to the rachilla. The culti- 
[ 287 | 
