ROOT-NODULES OF LEGUMINOSAE. 1 15 



The membrane of the root-hair is not completely dissolved by this process, but becomes swollen 

 and thus penetrable by the bacteria. 



The above results lead to the conclusion that the nodule bacteria of different species do not 

 entirely agree, at least in their physiological peculiarities, and hence to the question whether they 

 comprise one or more species. 



Hiltner obtained nodules on the roots of the East Indian Acacia lophanta, grown in sterilized 

 nitrogen-free sand, by inoculating them with bacteria from peas and locust. This shows that at least 

 not every genus of Leguminosae requires a special species of bacteria for the formation of nodules. 

 On the other hand, Kirchner observed that, among about ioo species of Leguminosae grown yearly 

 in the garden at Hohenheim, only the soy-beans were free from nodules. When these were inoculated 

 with earth from Japan in which soy-beans had grown, large active nodules appeared. Hiltner 

 at Tharandt in experiments covering several years also found Soja kispida always nodule free. 

 Although morphologically the bacteria of these nodules could not be distinguished from native nodule 

 bacteria, Kirchner considered the physiological and biological differences sufficient to warrant their 

 recognition as a new species, Bacterium {Rkizobacterium) japonicum. This conclusion is not shared 

 by Cohn or Naudin, who found nodules on soy-beans which had not been inoculated. 



Hiltner states that in his experiments in recent years no nodules were formed on his uninocu- 

 lated plants, except in some of the water cultures which are extremely difficult to keep pure, and that 

 hence, when inoculations produced nodules, the bacteria used must have been the nodule-forming 

 bacteria. 



He found the bacteria of the most various Leguminosae, even of the Mimosae and Caesalpineae, 

 extraordinarily alike morphologically. He could find no constant difference, nor could he regularly 

 demonstrate the two groups of Beyerinck, much less the differences described by Gonnermann. On 

 the other hand he thinks that the artificial media generally used for the culture of nodule bacteria, 

 and on which bacteria from peas, clover, vetch, and Robinia flourish, is extremely unsuitable for 

 serradella and lupin bacteria. vSuch medium consists of gelatin weakly acidified with malic acid, 

 and containing an extract of Leguminosae and asparagin and grape-sugar. Concerning other dif- 

 ferences Hiltner says: 



"The investigations at Tharandt, however, have plainly shown a great difference, both bio- 

 logical and physiological, in the bacteria isolated from nodules of different leguminous plants. It 

 was evident, from the most carefully conducted inoculation experiments that the bacteria from 

 nodules of a definite species of leguminous plant, were most active upon this same species, both as 

 regards rapidity in the formation of nodules, and their nitrogen-gathering activity. On other species 

 of the same genus, or of the same group, however, they generally showed a materially decreased 

 activity, and finally, as a rule, such activity disappeared completely in genera which were farther 

 removed systematically." 



Within the Trifoliaceae a mutual transfer of bacteria was less successful than between Pisum 

 and Vicia. Very active bacteria from Trifolium pralense failed to produce nodules on Medicago 

 sativa grown in a sterilized mixture of sand and earth. 



The lupins showed to a marked degree this difference between species of the same genus. Even 

 the bacteria from Lupinus lutcus and L. anguslijolius behaved very differently and Kirchner found 

 among 14 species of lupin grown for several years in close proximity, and even mingled, that while 12 

 formed nodules from the start, Lupinus hirsutus and L. subcarnosus were completely free from them. 

 For the most part Robinia pseudacacia appeared to be very exclusive. It required much time to 

 produce nodules even when bacteria from most closely related genera, e. g., Caragana, were employed. 



In contrast to Robinia the genus Phascolus was easily infected by bacteria from the nodules of 

 other legumes. When inoculated with bacteria from Pisum and Robinia it produced nodules, which, 

 however, were inactive or nearly so as regards nitrogen fixation. Bacteria, from Phaseolus on the 

 other hand, formed weakly active nodules on Pisum, but had no effect on Robinia. 



All these results lead to the probable conclusion that the different nodule-bacteria of the Legu- 

 mino?ae are really only adaptation forms of the same species. Laurent's experiments in which he 

 obtained nodules on peas by needle-prick inoculations with bacteria from 30 different (unnamed) 

 species of Leguminosae, seems to support this conclusion, although he had no check plants, and 

 Zinsser's experiments are contrary. 



A study of the bacteriods from the nodules of different species gave the same result. Although 

 sometimes very different in form, the transformations which took place permitted the conclusion 

 that the nodule bacteria of the Leguminosae represent a single, extraordinarily polymorphic species, 

 a conclusion which agrees with Beyerinck's observations on the bacteroids in an old culture from 

 Phaseolus. 



Hiltner states also that he and Nobbe succeeded in transforming pea-bacteria into forms iden- 

 tical with bean-bacteria in their activity on peas and beans, as well as in the form of their bacteroids. 



