230 ANNUAL OF SCIENTIFIC DISCOVERY. 



reckoned by per cents nor by tenths of per cents, but by the minutest 

 fractions. 



A heavy crop of thirty-seven bushels of -wheat, grain and straw 

 included, removes from an acre of land but three hundred pounds 

 total of mineral matters ; and the annual removal of the heaviest 

 crop of wheat from a soil for one hundred years diminishes its min- 

 eral matters by less than 0.4 per cent. If, then, in the selection of a 

 sample, the average composition is departed from to the amount of 

 four parts in one thousand, the analysis may represent the soil by the 

 value of three thousand seven hundred bushels of wheat per acre, or 

 by what represents, so far as mineral ingredients can, the fertility of 

 a century. 



AVhat freaks and accidents is not the soil analyst the sport of? A 

 bird, squirrel, or dog, relieving nature at the spot where he collects 

 his sample, innocently magnifies the phosphoric acid or alkalies of the 

 surrounding acres a hundred fold. The soil gathered toward the end 

 of a long rain, whereby its soluble matters are carried deep into the 

 subsoil, is declared poor, by analysis, whereas, if taken after a fort- 

 night of drought, it might appear extraordinarily fertile. Boussingault 

 found in his rich garden-soil in June, during wet weather, 0-00034 

 per cent, of nitric acid. In the following September, after a period 

 ofdryness, it contained 0.0093 per cent., or twenty-seven times as 

 much as in June. Tin's ingredient is indeed more liable to fluctua- 

 tion in amount than any other, both because it is formed in the soil, 

 and because it is not subject to the absorbent action which the soil 

 exercises over most other of its soluble constituents ; but the same 

 variation occurs among the other ingredients according to the direc- 

 tion of the capillary movement of the soil-water, though in less de- 

 gree. 



Independently, however, of all considerations and calculations 

 like the above, we have proof evidence at least that supports these 

 considerations, and has never been publicly refuted that it is prac- 

 tically impossible to obtain average specimens of the soil. I refer to 

 investigations made as long ago as 18469, under the direction of the 

 Prussian Landes Oekonornie Collegium, and reported by the distin- 

 guished Magnus. The account of these experiments is given in de- 

 tail in Erdman's Journal fur Praktisclie Chemie, vol. xlviii. 



Landes Oekonomie Collegium at that time carried on systematic 

 experiments in agriculture at fourteen distinct stations scattered 

 through the Prussian domain. The trials which we now speak of 

 were made for the ostensible purpose of studying the exhaustion of 

 the soil by cropping. The plan was to analyze the fourteen soils, 

 the history of which for years previous was accurately known, then 

 crop them with rape until " exhausted," then compare together the 

 original composition of the soils with their composition after exhaus- 

 tion, taking into account as well the composition of the crops re- 

 moved. The research began with collecting and analyzing the soils. 

 In order to meet, as far as possible, the difficulties of securing aver- 

 age specimens, equal portions of the soil of each field were taken 

 with the spade at ten or twelve different points, and thoroughly 

 intermixed ; of each sample, three separate portions were analyzed, 

 in most cases by different operators, who in many instances were the 



