576 Transactions. ^ 
vation altogether. The filaments were found to be shrivelled, and the 
pollen-grains were fluid. The stage of development of the anthers appeared 
to be of importance, because it was usually in the central floret of the 
spikelet that the injury had occurred. In the remaining 84 per cent. of 
the florets the anthers were apparently uninjured. 
The other series of temperatures which did not lead to the killing of 
anthers and also left the ovaries uninjured were as follows :— 
(а.) Reduce from 39? to 29° in six hours, hold at 29? for two and three- 
minimum temperature to the open-air temperature at 60° Е. Result—all 
plants killed, anthers and ovaries; and in the majority of cases the 
glumes as well. 
It was as a result of the failure in these three cases—(a), (b), and (c)— 
that the final and fairly successful series of temperatures recorded in para- 
graph (3) above was selected. 
_ In this final case the florets in which the anthers had been killed were 
It may be that artificial crossing of wheats can be more rapidly carried 
out by killing the anthers by freezing than by the usual method, but this 
would need suitable cooling-apparatus. Where this is available the plan 
will produce plants showing the dominant, and will be kept, while those 
that are selfed will not show the dominant character, and will be pulled 
out and rejected. 
