igi2.] Summary of Agricultural Experiments. 863 



variety Miss Dudgeon was anxious to have for trial, she was only 

 able to obtain the third-grade size, which in many cases were little 

 larger than an ordinary marble. 



From the commencement of the application of the discharge a 

 distinct difference was observed between the two plots, the crops 

 under the electrified wires being, from the commencement of growth 

 up to the time of ripening, distinctly in advance of those outside the 

 influence of the charged wires. The tubers were ready for lifting 

 quite a week earlier ; also the greater height of the haulms and 

 luxuriance of leaf-growth were very marked. 



The following table gives the weights per acre of the crops lifted 

 in the electrified and control plots : — 



Not 



electrified. Electrified. Increase, 



tons cwt. • tons cwt. tons cwt. 



• Ringleader 5 17 ... 8 1 ... 2 4 



Windsor Castle ... 9 18 ... 11 15 ... 1 17 



Golden Wonder ... 8 2 ... 8 15 ... 13 



Great Scot 10 6 ... 11 16 ... 1 10 



The cost of applying the electric discharge was £5 19s. 6<2., 

 comprising petrol, £1 6s. ; lubricating oil, 6s. ; and depreciation at 

 10 per cent, on apparatus (costing £175) for three months, ^4 js. 6d. 

 The same expense would, however, have covered the cost of electrifying 

 an area of 15 acres instead of the 8 acres treated in the experiment. 



Utilisation of Atmospheric Electricity in Plant Cultivation (Jour. 

 d'Agric. Pratique, September 29th, 19 10). — An apparatus which has 

 been used during a number of years at Angers for fixing atmospheric 

 electricity with very good results is described in this publication, and 

 the results obtained with a number of crops are given. 



Experimental Error in Field Trials (Jour, of Agric. Science, Vol. IV., 

 Part 2, October, 191 1). — The question of the experimental error in 

 agricultural investigations continues to attract attention. It was 

 discussed at some length by some of the leading research workers in 

 the Supplement to the December issue of this Journal, and the matter 

 is carried to a higher stage of refinement in an article by Messrs. Hall 

 and Mercer in the above periodical. An analysis of the sources of 

 variation in field experiments shows that they may be divided into 

 two categories : — (a) Variation in the climatic or seasonal conditions 

 under which the experiments are made; (b) the natural fluctuation of 

 the plant under observation under the influences of casual environ- 

 mental conditions, such as small differences of soil, situation, or 

 cultural treatment. The first varies from year to year, while the 

 second may affect crops growing in the same field under the same 

 seasonal conditions. The authors lay down that the limits of the 

 error due to the first class cannot be evaluated on the assumption that 

 they obey any law, but they show, on a consideration of experiments 

 carried out at Rothamsted, that errors of the second class obey a 

 definite law, and, consequently, can be predicted. They are of opinion 

 that errors of the latter class can be best evaded by scattering the plots 

 receiving the same treatment about the area under experiment, and 

 as a suitable system they recommend that each unit of comparison 



