442 



Trials of New Varieties of Cereals. 



[Aug., 



added to the nearest sheaf before the machine comes round 

 again. It will generally be best to cut only one way of the 

 field for a reason which will appear later. 



Binding. — Whether the plots are cut by hand or by a reaper 

 or by a self-binder, the sheaves of the two races must be kept 

 rigidly distinct. The best plan is to use string of two different 

 colours. If a binder is used, a loop of coloured string should 

 be attached to each sheaf of one of the races, before the next 

 turn of the binder. 



Field Weighing.— In 1920 and 1921 at Warminster the half- 

 drill-strips of the two races under comparison were cut with 

 a side-delivery reaper — each half-drill-strip separately — and in 

 order to leave approximately the same amount of stubble on 

 all the plots, so that the straw-weights as well as the grain- 

 weights might be comparable, the strips were only cut in one 

 direction. 



Machines of this type (now generally superseded by self- 

 binders) have an advantage for this particular purpose : — They 

 throw off two sheaves for each revolution of the rakes, and as 

 the rakes are driven by the travelling wheel each two follow- 

 ing sheaves represent the produce of equal areas. In 1921 the 

 area corresponding to each two sheaves was within a negligible 

 fraction equal to 1/500 acre. In this case, therefore, the 

 effect was to split each of the two half-acres into about 250 

 plots of equal size, and, obviously, it would be possible in such 

 a case to repeat Hall nnd Mercer's Rothamsted experiment and 

 obtain not only the total weight of both grain and straw on the 

 two half-acres, but also it would be possible to obtain the 

 weight of grain and straw on each 1/500 acre and by this means 

 l ed nee the probable error of the comparison in respect of both 

 grain and straw to probably a fraction of 1 per cent. This 

 would, however, entail so much labour and supervision as to be 

 practically impossible for a series of plots, and, moreover, 

 would be a " work of supererogation." What was done at 

 Warminster and what is practical was as follows : — As soon as 

 the sheaves had been tied with red and plain string respec- 

 tively each sheaf was weighed. This was done on a Salter's 

 spring balance graduated to single ounces up to 20 pounds. A 

 balance reading to tenths of pounds would be preferable in 

 order to Give a more simple record for statistical purposes. The 

 balance may be suspended from a pole about 6 ft. long carried 

 on the shoulders of two men, and the sheaves weighed two at 

 a time in the order in which they are thrown off by the 



