ROOT-NODULES OF LEGUMINOSAE. 123 
The following table gives the harvested crops in kilograms per plot, omitting fractions less than 
o.5 and adding those more than 0.5: 
Uninocu- Seed Soil Soaked | Natural soil 
lated. inoculated. | inoculated. | seeds. | inoculation. 
Wella FUpil os Gas asind cefeiee sane 112 89 385 2135 |) S138 
BMG MMII: aos ots lenis cie eis 40 37 150 93 162 
CrinISOR CLOVER Sh a3 380s sis fe) 244 112 14 17 
merradella hisses sieve caee spews | 41 54 41 160 | 42 
| | 
In summing up the 46 experiments of this year (1901) at Bremen it may be said that 54 per cent 
were favorable, some extraordinarily so, in spite of the many unfavorable conditions. This result 
certainly refutes the claim that inoculations in the open give recognizable results only in quite isolated 
cases. It is also plain that results are the more certain the less often the land used has borne the 
species of legume in question. The method of inoculation is of the utmost importance and must be 
selected with reference to the soil and the species of legume. 
In the experiments at Dahlem, 7. e., under Hiltner’s direct supervision, Soja hispida was used 
exclusively, since it is not there subject to spontaneous infection, but forms, when inoculated, large 
easily counted nodules in which, at a temperature suited to the Soja, active nitrogen assimilation 
takes place. In 1901 experiments were made to determine how long nodule bacteria remain active 
in the soil. Plots were used which had borne soy-beans the previous year, some uninoculated, others 
inoculated at that time. 
Both yellow and brown seeded varieties formed an average of 75 nodules on all plants grown in 
previously inoculated soil. Out of 356 plants examined none had less than 50 nodules. From this 
it is evident that the nodule bacteria which had wintered in the soil must have retained a high degree 
of virulence. Hence, when nodule bacteria prove virulent one year, they are also virulent for the 
same crop the next year, and even increasingly effective. This conclusion agrees with the results 
of growers who have found that one inoculation suffices for a series of years where the same crop is 
cultivated, and that further inoculation in such cases is useless. 
On soil which had been inoculated with Soja earth and cultivated in oats in 1900, the Soja-beans 
of 1901 were sparingly infected. These nodules were due, however, to an infection from neighboring 
plots rather than to bacteria which had persisted in the soil, as was shown by comparison of the two 
varieties in the four oat plots of 1900. The bacteria, therefore, can live from one year to the next 
in the Dahlemsoil only when the legume to which they belong is at hand so that nodules are formed. 
They are able to draw nourishment from the decaying roots, and, when these are gone, to hold their 
own with the other soil organisms for a limited time only. 
Another experiment in 1901 was made with soy-beans on land that had in 1900 borne various 
crops. In the uninoculated plots the average number of nodules varied from o to 1.2. Most of the 
plants were free. On the contrary in the plots where Phaseolus had grown in 1900, 41 per cent of the 
plants bore nodules. The results when inoculation took place 6 weeks before sowing were very 
moderate (3 to 14 nodules), and no difference could be observed between the plots which in 1900 
had borne legumes and those which had not. A 30 times larger amount of inoculating material 
caused a larger number of nodules, but even then the number did not equal those formed when inocu- 
lation took place at the time of sowing. What caused this failure, when 6 weeks elasped between 
the inoculation of the soil and the sowing of the seed, is not certain. At any rate it was not due to 
lack of moisture, as moisture conditions were excellent. 
Experiments with soy-bean to determine the best method of inoculation again showed the action 
of an injurious substance in the seed-coat. Inoculation made by strewing Dahlem sand which had 
been moistened with a pure culture produced plants which were absolutely nodule free. the most 
favorable results were obtained by germinating the seed and then inoculating just before sowing. 
These facts indicate that seeds which have been soaked exert no longer an injurious influence: This 
decreases as the germination progresses. The inoculation with pure cultures in all the experiments 
showed itself equal to natural earth inoculations and in most cases was superior to the latter. 
The method of inoculation by strewing earth previously infected with pure cultures proved 
beneficial only on cultivated moor soil, or in moist weather, or with seeds which germinate very rapidly. 
The method by seed-inoculation may entirely miscarry owing to destruction of the bacteria by 
poisonous substances extruded from the seed-coat during germination. This danger is much greater 
with large seeds, such as lupins, peas, and soy-beans, than with small seeds, but may occur with the 
latter, e. g., clover seed, if the soil is very dry and germination slow. 
