199 



out artificial inoculation will cause many persons to be skep- 

 tical regarding the value of inoculation. The fact that usu- 

 ally wild and cultivated legumes are naturally inoculated does 

 not indicate that inoculation of certain rarely grown plants is 

 unnecessary under all conditions. It would be just as logical 

 to argue against plowing as a preparation for hay grasses on 

 the ground that the grasses grow luxuriantly in their wild 

 state without any preparation of the land. 



Natural methods of inoculating legumes, or of bringing 

 the appropriate root-tubercle bacteria in contact with the roots 

 of young legumes are as follows : 



(1) Decay of tubercles, on old legume roots, thus free- 

 ing thousands of bacteria in the soil where the seed will be 

 dropped and where the next generation of legumes will grow. 



(2) Transportation of germs thus freed by means of 

 winds, flowing water, etc. 



(3) Inoculation of seeds before they fall by means of 

 germ- laden soil settling upon them or spattering upon them 

 during rains. 



(4) Changes in the nature or food habits of the root- 

 nodule bacteria by which it is claimed that these germs may 

 in time so adapt themselves as to cause tubercles on any 

 legumegrown continuously on the same field for several years. 



The writer is not in possession of very direct evidence 

 on this latter point, made by European writers, but there is 

 certainly some indirect evidence in its favor. 



Wherever any of these agencies are active, inoculation is 

 never absolutely necessary, and often superfiuous. 



When clover follows clover on the same land for several 

 years in succession, we have an example of the first mentioned 

 of these natural agencies. Of course in such a case artificial 

 inoculation is unnecessary. 



The case is similar when vetch is sown on land where 

 closely related wild plants have previously grown, a class very 

 common in uncultivated places, wood-lands, etc. In or near 

 garden spots and around the residence vetch is often inde- 

 pendent of artificial inoculation. 



When clover is sown in a region where clovers are exten- 



