.1920.] Report on the Treatment of Seeds. 973 



Classes of Treatment — Broadly speaking, the various methods 

 of treating crops to increase production may be divided into 

 three classes : — 



1. Those which are nearly always successful, such as the 



application of sulphate of ammonia and nitrate of 

 soda to corn or to grass laid in for hay ; of super- 

 phosphate to swedes ; of salt and of nitrate of soda 

 to mangolds, etc. 



2. Those which apparently succeed in some cases, but fail 



in others. 



3. Those which fail altogether to give crop increases. 



The use of artificial fertilisers belongs to the first category. 

 Methods of field trials have been devised by which an experi- 

 menter can say with comparative certainty whether or not a 

 fertiliser or a mixture of fertilisers would yield an increase in 

 crop, given a favourable season. He cannot say this with 

 absolute certainty, but the odds are 25 or 30 to i against his 

 being wrong. \Vhile, therefore, he may make a mistake in any 

 particular case he will not make many mistakes in advising, 

 say, 100 farmers. 



It is comparatively easy in a short test to find whether any 

 given process belongs to the first or second category, but it is 

 more difficult to discover whether it belongs to the second or 

 the third. 



Broadly speaking, the results of the recent tests made at the 

 colleges and experimental stations go to show that the elec- 

 trolytic treatment of seed does not belong to the first category. 

 In the majority of the trials the treatment has had no effect ; 

 in some there have been gains, in others losses. On the whole 

 there has been nothing to indicate with certainty any increase 

 in crop. It does not, however, follow that the process 

 necessarily belongs to category 3 — the worthless class ; it may 

 still belong to category 2. A single positive result in 100 

 failures would put it into this class, but obviously this would 

 require a close examination of all the alleged successes, and, 

 what is equally important, of all the failures, before a definite 

 decision could be given. 



Pot Experiments. — Experimental tests with treated seed 

 were made at Rothamsted in 191 8 and in 191 9; the experi- 

 ments were all made in pots, this being the most convenient 

 method for rapid work. 



To avoid misapprehension, it should be clearly understood 

 that the vessels used are not iiower pots, and that the process 



