[THOMPSON] EARLINESS AND LATENESS IN WHEAT 145 



maximum number possible as determined by direct crosses between the 

 earliest and the latest. 



As in previous years the greatest care was taken to eliminate 

 environmental errors as much as possible. All seeds were sown on 

 two successive days. The plot of ground on which the plants were 

 grown was chosen with particular reference to its flatness and uni- 

 formity. The plants were spaced in regular rows six inches apart and 

 three inches apart in the row. At harvest time two plants at both 

 ends of each row were discarded as were those on either side of a 

 blank resulting from the failure of a seed to germinate or from the 

 death of a plant. The latter fact accounts for the differences in the 

 numbers recorded for the different families. 



The season was unfavourable for the investigation for two reasons. 

 In the first place, a cold spring in which almost no growth was made 

 was followed by a very hot dry summer, with the result that the varie- 

 ties ripened much more nearly at the same time than is usually the 

 case. In the second place a severe epidemic of rust broke out in spite 

 of the dryness and the seed of the later varieties and hybrids became 

 shrunken and hardened much sooner than the normal ripening time. 

 Both these influences tended to lessen the real hereditary differences 

 between the different races, and to make interpretation of the results 

 more difficult. 



On account of the very large number of plants to be grown and 

 recorded in order to secure a reasonably large number of each family, 

 it was found to be impossible to continue the work with all the crosses 

 of the previous year. Record taking was therefore either abandoned 

 entirely, or of a very general nature with some of the least important 

 crosses. One very important cross, the ene involving the widest 

 difference, could not be investigated completely because of the partial 

 sterility of many of the F2 plants, resulting from the remoteness of 

 the relationship of the parents {Triticum vulgare and T. durum). In 

 most cases of wide crosses in wheat the Fi plants are very vigorous 

 and productive, but the F2 plants show various degrees of sterility. 

 Consequently, in this case, not enough viable seed could be obtained 

 from many F2 plants to raise sufficiently large F3 families. 



The only contribution to the literature of the subject since the pre- 

 vious article (2) was written, has been made by Caporn, (l) who worked 

 with oats. He crossed a single pair of parental varieties whose ripen- 

 ing times did not overlap. Unfortunately, the F2 data are not avail- 

 able, but the F3 families were for the most part intermediate and show- 

 ed greater variation than either parent. Many families were dis- 

 tinctly earlier than others. The parental types were not recovered. 



