no BOTANICAL GAZETTE [augusi 



almost converted to the opinion that this particular rhizobium 

 was either not an autonomous organism, or that it had lost the 

 ability to grow outside of the host-plant. I have had no 

 difficulty in securing cultures of the rhizobium found in bean- 

 root tubercles, for example, but for reasons which will be given 

 in a subsequent paper I was desirous of obtaining cultures of the 

 comparatively large rod-shaped to Y-shaped form found in sweet 

 clover. For some time after beginning the present series of 

 experiments it appeared that my efforts would again be without 

 positive results ; but my former experiences induced me to pro- 

 ceed more cautiously and more carefully, keeping in mind the 

 probable physiological peculiarities of this particular rhizobium 

 as compared with bacteria in general. Now I feel certain that 

 in 1893 I had several small pure cultures of this particular 

 rhizobium, but failed to recognize them, and they were destroyed 

 without being investigated, and, unfortunately, I was further- 

 more obliged to discontinue the investigations at a time when 

 I was becoming somewhat familiar with my subject. 



In the following experiments young seed-grown plants or 

 Melilotiis alba were selected, which were about four inches high, 

 each plant having from six to fourteen fairly well developed, 

 mostly single, non-branching, more or less irregularly oval to 

 flattened tubercles. The roots and tubercles were thoroughly 

 and repeatedly washed in hydrant water, and dried with blotting 

 paper. A tubercle was cut into with a small sterilized knife, and 

 the knife-blade was lifted, thus breaking the tubercle across. 

 This prevented the blade from dragging exterior bacteria over 

 the interior of the tubercle. A short platinum needle was pushed 

 into the central portion of the tubercle and streak (line) cul- 

 tures made in Petri dishes w^ith Melilotus alba extract solidified 

 by means of agar. The inoculated Petri dishes were kept in the 

 dark at the spring and early summer temperature of the labora- 

 tory. ■ Most of the dishes showed white growths in a day or 

 two. Examination proved these to consist of several motile 

 bacteria, including a large motile bacillus resembling B. anthracis 

 or B. siibtilis. In the course of about five days from the time or 



