402 THE FOOD OF PLANTS 



nitrogenous soil or one containing a measured quantity of saltpetre. On estimating 

 the amount of nitrogen present in the original seeds and in the plants when harvested 

 no perceptible gain from the free nitrogen of the air could be detected. The slight 

 gain of nitrogen observed in the first researches in ordinary air was due to the 

 absorption of ammonia, &c. from the latter 1 . These results were obtained not 

 only with Lepidium and with oats, but also with beans and lupins, probably 

 because the plants were for the most part grown upon previously ignited pumice- 

 stone and watered with a solution of the necessary ash-constituents, so that no 

 root-tubercles were formed. Lawes, Gilbert, and Pugh* obtained similar results, 

 using for the most part a soil composed of burnt clay. The increase in the 

 amount of nitrogen observed in all cases by Ville 3 , even in cereals, was due 

 either to faulty experimentation 4 or to a fixation of nitrogen by organisms in 

 the soil. 



Practical experience has left no doubt that the Leguminosae obtain more 

 nitrogen than is present in the soil, and it was Hellriegel who showed in 1888 

 that these plants assimilate this gas in the free condition, but only when provided 

 with root-tubercles. The further progress of our knowledge has already been 

 indicated 6 , as well as the reasons which led certain authors to conclude that all 

 plants can assimilate free nitrogen. Berthelot (1885) was the first to prove that 

 certain bacteria have this power, while Winogradsky isolated one of these bacteria 

 and studied its special properties in great detail. 



Hellriegel and most other workers, including Boussingault , estimated the 

 increased percentage of nitrogen in the crop from sand, humus ', or watery 

 nutrient solutions. Hellriegel placed almost non-nitrogenous sand in glass vessels 

 of 4-8 litres capacity, and brought the cultures into the open during fine weather, 

 so as to attain the maximal possible growth. The complete sterilization of the 

 soil and the use of distilled water for watering sufficed to prevent any formation of 

 root-nodules. Similarly no perceptible development of nitrogen-forming organisms 

 occurred in sand covered with pieces of quartz on which a layer of cotton wool 

 was superposed. The seeds were sterilized by means of i to 2 per cent, solution of 

 formic aldehyde, or by washing with alcohol, ether, or a dilute solution of mercuric 

 chloride. Beyerinck, Prazmowski, &c., by using special vessels and adopting certain 

 precautions, ensured the absence of all micro-organisms, but even then, as both 



1 Boussingault, Ann. d. chim. et d. phys., 1838, ii. se'r., T. LXVII. 



8 Lawes, Gilbert and Pugh, Phil. Trans., 1862, Vol. CLI, p. 431. The following authors obtained 

 similar results, although their experiments were hardly performed with the same care : Mene 

 (Compt. reud., 1851, T. xxxn, p. 180) ; Harting (ibid., 1855, T. XLI, p. 942) ; Cloez et Gratiolet 

 (Ann. d. chim. et d. phys., 1851, Hi. ser., T. xxxn, p. 41); Bretschneider (Jahresb. d. Agr.- 

 Chem., 1861-2, p. 123). 



3 Ville, CompL rend., 1852, T. xxxv", p. 464; 1854, T - xxxvm, pp. 703, 723; also in Rech. 

 exp., 1853 and 1857. 

 I * Compt. rend., 1855, T. XLI, p. 757. 



5 MacDougal gave in 1894 a summary of all the works which had appeared up to that time 

 (Minnesota Bot. Studies, 1894, Vol. 9, p. 186). 



On the effect of heating the soil, cf. Hellriegel, Ber. d. Bot. Ges., 1889, p. 131 ; Richter, 

 Versuchsst., 1896, Bd. XLVII, p. 269. 



