1922.] 



Trials of New Varieties of Cereals. 



machine. The balance should be fitted with a cradle on which 

 the sheaves are placed. The weighing of 500 pairs of sheaves 

 may be performed in about three hours by one assistant reading 

 the weights; one entering them in a prepared book; a man 

 lifting the sheaves on and off the scale; and two men carrying 

 the balance. This operation is therefore feasible for a series of 

 variety trials. 



A comparatively simple attachment has been devised for a 

 self-binder in order to deliver sheaves corresponding to equal 

 areas, instead of sheaves of approximately equal weights which 

 is the present arrangement in all self-binders. 



If the strips have to be cut by hand, each half-drill-strip 

 may, obviously, with little difficulty be divided into a number 

 oi equal plots, and the sheaves on each plot weighed. 



The only practical method of collecting the produce is to 

 bulk all the " C " sheaves into one small stack and all the 



A " sheaves into another; to thresh the stacks and weigh 

 and record the grain and straw threshed from each stack. 



These figures for the total weights of grain and straw on 

 half an acre of each race obviously give no indication of the 

 probable error which attaches to them in the absence of any 

 weighings of the produce of smaller areas, but a very close 

 approximation to the probable error of these weights can be 

 arrived at by a statistical treatment of the sheaf-weights. 



It has been found in repeated experiments that the ratio of 

 grain to straw is constant within very narrow limits for the 

 toe rare when grown under the conditions above described. 

 Tt may therefore be safely assumed that the probable error 

 of the total grain-weights of each race is not appreciably 

 greater than that of the average total produce on a large 

 number of small areas of each race. 



Prom the tabulation of results in a very large number of 

 similar cases the writer finds that the probable error of the 

 weight of grain is, in fact, less than that of the corresponding 

 weights of grain plus straw. 



The obrect of the half-drill-strip method is to minimise the 

 effect of divergencies in the conditions external to the plant, 

 and there is no doubt whatever that this result is obtained. 

 Tf the sheaf -weights are determined either for small areas like 

 1 "00 acre bv weighing pairs of sheaves, or even only for 1/40 

 nicre half-drill-strips of each race, then the results can be stated 

 in terms of (1) weight per acre of grain, (2) weight per acre 

 of straw, (8) weight per acre of total produce, and (4) probable 



