BIOLOGY OF SOILS 143 



is not diminished but increased by their growth. This applies 

 more or less to all plants belonging to the order leguminosse. 



HeUriegeVs Discovery. No satisfactory explanation of 

 these remarkable facts was forthcoming, however, until 

 the year 1886, when Hellriegel and Wilfarth showed that, 

 under certain conditions, leguminous plants can assimi- 

 late the free nitrogen of the air and grow in soil devoid 

 of compounds of nitrogen, from which plants usually obtain 

 their supplies of that element. 



This peculiar property they traced to the development of 

 warts or nodules on the roots of the plants (Fig. 17), and 

 showed that these are produced by bacteria which enter the 

 plant and are the cause of the fixation of nitrogen. When 

 grown in sterilised soil the warts or nodules are not formed, 

 and in the absence of compounds of nitrogen development 

 of the plants is arrested. 



Sand Cultures. The fact can be demonstrated by sand 

 culture experiments similar to those previously described 

 (p. 7). When the sand has been prepared and mixed 

 with the necessary mineral ingredients of plant food com- 

 pounds of nitrogen being strictly excluded it is put into 

 pots and sterilised by heat. When cold the sand is satu- 

 rated with sterilised water and the seed of any leguminous 

 plant peas, clover, vetches, etc. deposited in it. The pots 

 are then covered with bell glasses and kept at a suitable 

 temperature. In due time the seeds germinate, but the plants 

 soon languish for lack of nitrogen and die. But, if the sand 

 in one of the pots be impregnated 1 with soil bacteria and 

 a small quantity of nitrate of soda is added 2 to another, a 

 very different result will be produced (Figs. 18 and 19). 



1 This may be done by sprinkling a few grains of moist earth from 

 an ordinary soil on the surface of the sand, or by shaking up a handful 

 of ordinary soil with water, and, when it has settled, decanting off the 

 liquid and adding it to the sand under experiment. 



' 2 Goldim?. 



