82 MISCELLANEOUS PAPERS. 



a hole in the paraffined paper. After a month's time the roots of the 

 plants were examined, with the following results: 



Nodules on plant No. — ! 





Treatment. 



1. 



2- 



3. j 



4. 



5. 



6. 



Total. 



-.\ era ted 



4 

 



20 

 1 



1 



25 1 



2 j 



■ 8 

 



26 

 



4 

 



87 



Nonaerated . 



3 









See Plate VIII, figure 2, noting the difference in the plants in favor 

 of the noduled (aerated) plant. 



ASSOCLA.TIVE ACTION OF BACTERIA. 



Undoubtedh' the pure cultures of nodule-forming bacteria are sub- 

 ject to contamination when planters, using convenient rather than bac- 

 teriological methods, prepare quantities sufficient to inoculate then- 

 seed. Yet the purity of a culture immediatelv before it is applied to 

 seed or introduced into the soil of a cultivated field is unimportant, 

 provided there is present a sufficient number of virile nodule-forming 

 bacteria. 



Preliminary experiments at Washington some 3'ears ago indicated 

 that in the nitrogen-poor solution generally used the nodule-forming 

 bacteria would develop much more rapidly than other bacteria or con- 

 taminating 3^easts and molds. It might be expected that different 

 localities would have different contaminating forms, and during the past 

 ye,aY samples of cultures have been obtained from farmers in various 

 parts of the United States immediateh' before the seed or soil was 

 treated. Investigation of contaminating forms isolated from these 

 cultures shows that there are certain forms which inhibit the growth 

 of the nodule organism when both are gro^vn for some time in the 

 usual nitrogen-poor solution/' Using an extract of a favorable soil as 

 a culture medium, results very nearly parallel for the competing bac- 

 teria were secured, thouo-h one or onanism which resembles Bacillus coli 

 checked the growth of Pseudomonas radicieola in the synthetic medium 

 and not in the soil extract. - The relative virility of the competing cul- 

 tures at the time they begin their struggle is an exceedingly important 

 factor in determining which species of bacteria survives. 



Extended experiments on the interaction of groups of bacteria in 

 various media must be carried on before it will be possible to deter- 



« In view of this fact, the reason for our present method of distribution is apparent. 

 A relatively large amount of pure hquid culture used as a "starter" cuts down the 

 chances of ruining the farmer's culture through the competition of contaminating 

 forms; the time required to fill the solution with the organisms is shortened, and, as 

 shown subsequently, the quicker the bacteria can be introduced into the soil and 

 make use of the soil solution as a culture medium the less becomes the danger from 

 competition. 



100— VIII. 



