216 Experiments on the Artificial Fecundation of Wheat. 



reduced to English standards of quantity), in one experiment, 

 from 34 to 4G, and in another from 3G to 39 bushels of wheat- 

 grain per acre ; and, what is more extraordinary, as the result 

 of a process applied so late in the period of growth as that of 

 flowering, the recorded increase in the produce of straw was in 

 the one case as much as from 1848 to 2728 lbs., and in the other 

 from 2059 to 2240 lbs, per acre. A process which involves no 

 outlay for manure, is so simple and easy of application, and the 

 results of which were vouched for by such high authority, seemed, 

 at any rate, worth a careful trial, and I therefore submitted it to 

 one in the course of last summer. 



Before, however, describing the experiments and their results, 

 it may be well to make a few remarks on the conditions essential 

 to the attainment of successful and reliable results in field experi- 

 ments generally, and upon the state of the plots selected for the 

 tiials with M. Hooibrenk's process. 



If the object of an experiment be to compare the cfTecls of 

 different manures, or of any other means of increasing the crop, 

 the general condition of the soil, its evenness, and the characters 

 of the season in which the experiment is made, must each and 

 all be taken into account. The land may be in too high or in 

 too low a condition for the pur])ose ; the plots may appear to be 

 j)retty uniform both in quality and condition, and ^et if they 

 were all treated in the same manner, instead of differently, they 

 would probably give very different amounts of produce over 

 equal areas ; or the season may be much more favourable, or 

 unfavourable, for the effects of one of the modes of treatment that 

 is to be compared with others, than would be the case taking the 

 average of a number of years. 



It is seldom that the opportunity occurs of selecting duplicate 

 plots for a comparative experiment, the history of which for 

 many years past is accurately recorded, and the evenness of 

 which or otherwise can therefore be accuratclv ascertained. Such 

 an opportunity was, however, at command for the purposes of 

 the trial of Mr. Hooibrenk's process. Plots were selected from 

 among those on whit h wheat had been grown at Kothamsted 

 for many years in succession by the use of the same manure 

 applied year after year on the same land. By reference to the 

 report on the experiments in question, published in the last two 

 Numbers of this Journal, it will be found that, in most cases, 

 duplicate experiments were made with the same manure, the 

 duplicate plots being designated "a" and "i" respectively. 

 Three pairs of these plots were selected, namely, 12a and h, 13a 

 and b, and 14a and b ; and in order to show how far the duplicate 

 plots were in equal condition, and, therefore, properly comparable 

 with one another, the average produce of wheat per at re per 



