343 



of soil, Wohltinann usocl zinc boxes, 85 cm. loiii; and about 47 cm. wide 

 (surface area 0,4 square meter) and I meter deep, /. e. about .'U inches 

 lonji', 18^ wide, aud oD deep. The ziuc was protected by asphalt var- 

 nish. To secure natural temperature, the boxes were sunk in the earth 

 so as to have the soil within and without at the same level, and ]>lants 

 of the sanu? species as those grown in the boxes were also grown in tlie 

 soil between them. To facilitate j)roper regulation of water supply from 

 below, drainage, and collection of drainage water, a ]K'rforated zinc tube, 

 which was coated inside and outside with as])Iialt varnish, was laid on 

 the bottom of each box and covered by semicylindrical drain tile. This 

 tube passed through the end of the box at tin; bottom. To the end of 

 the tube outside the box was connected a pei'])endi(;ular glass tube so 

 that water could be introduced at will and the height of the water table 

 maintained at any desired level and conveniently observed. 



A faucet was attached to the tube at a height of UO cm. from 

 the bottom. By this the drainage water was collected in a graduated 

 o-liter Hask which stood underneath. The boxes were placed on a 

 brick foundation. To provide for getting at the drainage tubes, pas- 

 sages a little more than 1 meter deep were made between the rows of 

 boxes. The boxes were tilled to a depth of 2o cm. from the bot- 

 tom with pure qiuirtz sand (which thus served as subsoil), and then to 

 within 2 cm. of the top with the soil selected for the experiment. 

 This was a sandy loam which contained but little humus and had been 

 cro])ped for 3 years without manure. The soil was taken li;om the 

 tield to a dei)tli of 'So cm., i. e. so as to inchule only surface soil, 

 and was sifted and tlnuonghly mixed, so that it should be of uniform 

 character for all the boxes. Investigations Avere made of fineness (elu- 

 triation), water-holding capacity, and chemical composition of both sand 

 and soil. 



Seven kinds of ])lants Avere grown, barley, wheat, oats, yellow lupines, 

 blue lupines, ])eas, and beans. The seed was chosen with a view to its 

 special fitness for the experiment and was tested in germination trials. 

 The individual seeds were carefully selected, so as to have those of each 

 plant as nearly alike as practicable in api)earance, weight, aud specific 

 gravity. They were sown at distances apart corresi)onding- to those in 

 the best field practice. 



The fertilizing materials were furnished in nitrate of soda, sulphate of 

 potash, and plain super]diosphate, each singly and all three together. 

 To compare the efiects of phosphoric acid in different combinations, 

 "double superphosphate" was used in some of the earlier trials, l)ut 

 rei)lace<l by a mixture of plain superphosphate and potash salt after 

 trials of two successive seasons had shown that phosphoric acid had 

 but little effect. The quantities of fertilizers Avere made equivalent to 

 the largest ordinarily emi»loyed, so as to insure generous sui)ply. 



The boxes containing plants were arranged in 7 roAvs running east 

 and west aud 11 rows running north and soutli. In addition to these 



