440 Trials of New Varieties of Cereals. [Aug., 



be drilled allowing for this space and for trimming the ends 

 will be about 580 ft. in the case of a 13-coulter drill with 12 

 coulters working. 



Where, as is frequently the case, it is found impossible to 

 avoid placing the trial on land which has been differently 

 cropped or manured in previous years, the strips should, if 

 possible, cross the lines marking the different soil conditions 

 rather than run parallel with them. If possible, the longi- 

 tudinal direction of the strips should be north and south, or 

 as nearly so as the shape of the field admits. 



It will be noted that one acre is required for each race when 

 only one control is used, and two acres for each race when two 

 controls are used, because in the case of two control races the 

 experiment is duplicated in every respect by the addition of the 

 second control race. 



Drilling will be found to be much less complicated than 

 would appear from the above directions. Once the drill has 

 been adjusted this goes forward as rapidly as with ordinary 

 drilling. It is quite possible to drill 6 or 8 acres, viz., 6 or 8 

 separate yield trials in one day if the drill is made ready the 

 day before and if the superintendent is familiar with 

 the method, and has two or three intelligent helpers, 

 one of whom must be an expert drill sman. A good deal of 

 time is occupied in cleaning out the drill (or half the drill 

 where the same control is used for several new races), but 

 no more than when single half-acre plots of each race are 

 drilled. The cleaning out is much more easily done with a 

 drill of the Massey-Harris forced-feed type than with a cup- 

 drill, but a " steerage " is very necessary. 



Cutting. — The method to be followed in cutting and in the 

 subsequent operations will depend partly on the state of the 

 crops at harvest time and partly on the degree of accuracy 

 which is aimed at. 



A source of systematic error is introduced in all strip methods 

 of comparison if there is " interference " of one race with 

 another along the lines of separation. 



In the Warminster trials of 1920 and 1921 the two races 

 compared were of very similar habit of growth — so much so 

 as to be almost indistinguishable at all stages, and it was 

 obvious that there was no interference of one race with the 

 other. The whole of the area drilled, excluding the first and 

 last strin, was therefore harvested and weighed. 



In other cases, however, interference will often arise either 



