A[/ricuUural Experiments in the Field. 515 



perhaps intervene before the pLant has a fair chance of making a start. 

 Three weeks tell very much on crops which, like onr root crops, have 

 but a short time for gi-owth. But as it is impossible to cultivate a 

 whole field precisely alike, to secure the same conditions throughout 

 a large area, it is, I believe, on the whole, best to confine field trials on 

 arable land to plots amounting to the twentieth part of an acre. 



The next circumstance which greatly influences the success of field 

 trials is the kind of soil selected. The field itseK should, if possible, 

 be level ; it should be imiform in quality, and not too shallow ; it 

 should be perfectly drained, and in good physical condition. For 

 general experimental purposes, perhaps it is best to have soil which is 

 neither very light nor very heavy ; no great preponderance, whether of 

 clay, or of organic matter, such as peat, or of sand should occur. 

 Experiments for particular soils constitute in themselves a particular 

 class of trials ; but if we wish to ascertain what is the general efiect of 

 certain manuring constituents, such as ammonia, or phosphate of lime, 

 or potash, it is not desirable to have a soil in which one or other of 

 the chief constituents greatly preponderates. Under all circumstances, 

 the physical condition of the soil, the nature of the subsoil and its 

 depth, as well as the depth of the sm-face soil, should be carefully 

 recorded, and the chemical condition of both should also be ascertained. 

 Moreover, the agricultural condition of the experimental field should 

 be perfectly well known. The neglect of this point produces a gi-eat 

 deal of confusion, and renders the account of many of the recorded 

 field experiments altogether unintelligible. We should ascertain how 

 the field has been cultivated in previous years, when it was last 

 manured, what was the weight of corn, or roots, or hay, which it 

 yielded during the whole of the last rotation, also what is the average 

 yield in good seasons as well as in bad ones. Fiu-ther, we should 

 know whether natural maniu'e, such as farmyard manure, lime, and 

 compost manure, or whether artificial manures, such as guano, super- 

 phosphate, or bone dust, have produced a particularly marked efiect. 

 Information on all these points often throws much light on the 

 agi'icultural as well as on the chemical condition of the land, and 

 affords useful indications as to the mode in which field trials should 

 be arranged by the intelligent experimentalist. 



In order to illustrate how the neglect of these agricultm-al conditions 

 often entirely spoils the experiments, and tends to waste of time and 

 waste of money in the purchase of artificial or natural manures, I shall 

 mention some experiments which I had tried in 1864 on some clover 

 land. The experimental field was divided into six portions. Two 

 parts were left unmanured. The four others were manm-ed as follows : — 

 The first plot received nitrate of soda alone at the rate of 3 cwt. 

 per acre ; the second It cwt. of nitrate of soda, and 4 cwt. of super- 

 phosphate of lime, of known composition ; the third 6 cwt, of salt per 

 acre ; and the fourth 3 cwt. Now, the two unmanured jjlots yielded 

 the following per acre : — The first, 2 tons 11 cwt. 801bs. of green 

 hay ; the second, 2 tons 11 cwt. Slbs. The plot manured with nitrate 

 of soda alone, at the rate of 3 cwt. per acre, yielded 2 tons 14 cwt. 

 321bs. ; the next plot, maniu-ed with nitrate of soda and superphosphate, 



