130 BACTERIA IN RELATION TO PLANT DISEASES. 
bacteria formed colonies with watery exudations and made no growth at all in higher concentrations 
(5 per cent). On the other hand, the addition of 0.3 per cent gelatin prevented the growth of Soja 
bacteria (in aba bouillon) while it did not influence the growth of pea bacteria. 
: The very definite difference between the two groups shown by their behavior on gelatin is also 
a ground for the establishment of two species. 
These results explain why it has been so difficult to obtain cultures from lupin or serradella on 
gelatin. When luxuriant colonies have appeared they were certainly formed not by Rh. Beyerinckii 
but by Rhizobium radicicola which had entered the nodule by chance. The infection of peas with 
bacteria obtained from lupin nodules in 1890 at Tharandt, therefore, gives evidence that pea bacteria 
may be contained in lupin nodules, for the true lupin bacteria are not able to cause nodules on peas. 
Experiments in pots were made with soy-bean plants using for inoculation pure cultures and 
also the contents taken directly from nodules. ‘The latter gave the better results, though improve- 
ment in growth began 3 days earlier with pure cultures. In field experiments with the same inocu- 
lating materials the nodules being rubbed up and added to pure water until it was twice as cloudy as 
the pure culture suspension, the pure cultures produced very materially better results than the 
nodule contents. 
In an experiment already referred to in which, after soil inoculation with soy-bean bacteria, 
oats were cultivated the first year and soja the second year, the soy-bean plants were almost free 
from nodules. In other tests with soy-bean in which the inoculations were made with bacteria mixed 
with either quartz sand, forest humus, compost earth, Dahlem earth, or water, only the quartz sand 
inoculation gave satisfactory results. Since it is not likely that the sandincreased the multiplication 
of the bacteria the various soils must have reduced the number of bacteria within a very short time. 
Plants inoculated with compost earth formed on the two plots an average of 0.32 and 1.89 nodules 
per plant, those with Dahlem soil 0.12 and 1.00 nodules per piant, while those with quartz sand formed 
an average of 7.07 and 15.33 nodules per plant. Sterilizing the earths did not improve them. When 
Dahlem soil was mixed with quicklime (1 kg. earth and 2 gr. lime intimately mixed) and then with 
bacteria for inoculating material, no nodule formation occurred. These results do not hold for all 
limed earth, however, only in the case of quicklime. This was still able to act injuriously after lying 
in moistearth 24 hoursbefore the bacteria were added. Had a longer time elapsed before the addition 
ot the bacteria probably they would not have been injured. From this it is evident that unless one 
is dealing with acid soil, quicklime should not be used immediately before inoculation. Experiments 
in pots made at the same time and under the same conditions gave the same results, thus confirming 
Salfeld’s conclusion that caustic lime exerts an injurious influence on the nodule bacteria. 
Long before Hartleb brought out his specific nutrient fluid, the relative advantages of solid and 
liquid media were tested by Nobbe and Hiltner who found that the liquid medium offered most advan- 
tage because the bacteria remained much longer alive and capable of causing infection. Although 
these researches were not published, Hiltner has taken this fact into account by keeping all his 
cultures, which are to be further cultivated, in liquid media. Over against this undeniable advantage 
must be set the fact that a liquid culture medium does not permit the necessary control of the purity 
of cultures. Hartleb’s claimthat only bacteroids in liquid media were virulent while the bacteria on 
solid media were completely inactive, shows that he was not sufficiently familiar with the methods 
of cultivation on solid media. Certainly his statement is not of general application. 
The previously noted experiments of Hiltner with the nodule contents in contrast to pure 
cultures gave evidence that the inoculating material loses in value as the bacteroids develop further 
from the original bacterial form. 
The addition of saltpeter to grape-sugar solutions in which pure cultures were grown strongly 
suppressed the nodule producing ability of the bacteria. While pure cultures in 1 per cent grape- 
sugar solution produced when used for inoculation an average of 19.2 nodules per plant such a 
solution with 0.1 per cent saltpeter produced an average of 8.5 nodules. 
The word virulence, says Hiltner, is a term used to express the degree of ability of the nodule 
bacteria to penetrate into the root tissues of the plants and to multiply therein. It is almost self- 
evident that the conditions of nutrition influence materially the degree of virulence. 
In a series of experiments upon the spread of nodule bacteria in forest soil it was found that 
locusts grown in the middle of an old beech forest formed when uninoculated only a few small nodules 
and developed very weakly. When inoculated the growth from the beginning was luxuriant andthe 
nodules were numerous. The results on the uninoculated plants led Hiltner to conclude that nodule 
bacteria may sometimes exist an extraordinarily long time in the soil without leguminous plants and 
that to all appearances, they are not dependent on symbiosis. It was noticeable that the nodules of 
inoculated plants were more active than those arising from bacteria already in the soil. This is 
referable rather to a difference in virulence than to a difference in the quantity of bacteria concerned. 
This is not proved to a certainty, however, by these experiments. More convincing were the 
