177 



«0il extracts are not equally suitable for inoculating certain plants ; but 

 the most systematic and conclusive experiments on this part of the 

 subject are those of Nobbe, who concluded that there is only one micro- 

 organism concerned (BeijerincVs Bacillus radicicola), but that it becomes 

 so altered in the symbiosis with certain plants as to be rendered useless 

 for the purpose of infecting others. (^) This is, however, not always 

 the case, since the bacteria from pease and vetches, for instance, seem to 

 be, mutually available. On the other hand, the bacterium as modified 

 by the pea (or vetch) is useless for serradella, robinia, and for red and 

 other varieties of clover Nobbe thinks this may account for the diver- 

 gent results obtained by different observers, some of whom found 

 lupins, others serradella, &c., to be the greatest nitrogen collector ; the 

 diiference being in reality, most probably due to the inoculation having 

 been more or less suitable or unsuitable. The importance of this ques- 

 tion will be better realised in discussing the results of the field experi- 

 ments. 



FIELD EXPERIMENTS ON SOIL INOCULATION. 



The greater portion of the inoculation experiments described by Sal- 

 feld were made on peaty soils (Moorland,) and it was at the then new 

 Experiment Station of the Central Moor-Kommission on the Great 

 Bourtanger Moor that the first trials were made in 1887 (i.e. immediately 

 ^after the announcement of Hellriegel's discovery) when the usual me- 

 thods of reclaiming the land seemed almost hopeless. It had generally 

 been found possible on old cultivated, peaty soils, which had been man- 

 ured with dung, to obtain satisfactory crops of horse beans and peas, 

 after the application of quicklime or marl, with kainite and phophatea, 

 but on the new land these plants did not thrive unless supplied with 

 nitrogenous manure. The attempt to commence with buckwheat, as a 

 peaty-soil plant, instead of horse-beans, did not meet with success in 

 presence of large quantities of lime (32 cwt. per acre), and it was not 

 advisable to apply less lime on account of the clover. 



The first experiments made in 1887, in which horse-beans, grey. 

 Prussian peas, field peas, garden peas, and Ervum monanthos were in- 

 oculated by applying soil to the field, afforded clear indications of a 

 favourable character, but as no numerical results are given for the 

 different plots, we will pass on to the second series of experiments, 

 /made in 1888. 



The land selected had an area of 1 hectare (about 2^ acres). The 

 northern half of the field had been twice burnt, and in 1886 received 

 32 cwt., of lime per acre. Kainite, containing 140 lb., of potash, basic 

 slag containing 107 lb., of phosphoric acid, and ammonium sulphate and 

 sodium nitrate, containing together 54 lb., of nitrogen per acre, were 

 applied, and a yield of 18 . 74 cwt., per acre of winter-rye grain obtained 

 in 1887. In the 1888 experiments the land was divided into eighteen 

 plots, about five yards wide, and separated from each other by a path 

 and a ditch alternately. The plants employed were : (1) Peas, four 

 plots ; (2) horse beans mixed with peas, eight plots ; (3) horse beans 

 and lentils (Ervum), four plots; (4) Pisum arvens, two plots. The 



9 Landw. Versuchs-Stat. 1891. Bd. 39, and 1893, Bd.42; also Hannor. Land u 

 .Forstwirtschaftl. Zeit. 1894, p. 79. 



