The Sources of th6 Nitrogen of our Leguminous Crops. 677 
doubtless, be some fine root-fibre not removed at the conclusion, 
so that where there is loss it is to be supposed that some of the 
original nitrogen of the sand has contributed to the growth. In 
the case of the garden-soil, with its high percentage of nitrogen, 
it is, of course, not impossible that there may have been some 
I loss by the evolution of free nitrogen. 
That there is at any rate no material gain of nitrogen in the 
sand or soil would seem to be confirmatory of the conclusion 
indicated by other evidence, that the fixation is not effected by 
the organisms within the soil, independently of the symbiotic 
growth of the nodules and their contents, and of the higher 
I plant to which they are attached, and to whose nitrogenous 
j supply they seem to contribute. Indeed, if the fixation had 
; taken place under the influence of microbes within the soil, 
independently of connection with the higher plant, we should 
expect some accumulation within it, or we should have to con- 
clude that the plant had availed itself of exactly the whole of 
the nitrogen so brought into combination — a supposition for 
which there would seem no reasonable justification. 
Turningr to the middle division of the Table, which shows 
the nitrogen in the seed sown, in the total vegetable matter 
grown, and the gain, and disregarding the changes in the soils, 
which it has been seen may well be done, it will be observed 
that, in the case of pots 1, 2, and 3, with sand, the gain in the 
plants is so large as to be very far beyond the limit of any possi- 
ble experimental eiTor. This certainly cannot be said of some 
of the experiments which have been conducted on other lines 
in regard to the question of the fixation of free nitrogen. 
The gain of nitrogen in these initiative experiments is, how- 
ever, much less than in many of those of Hellriegel and Wilfarth, 
and also much less than in subsequent experiments at Eotham- 
sted. This is not to be wondered at when the late period of 
the season, and the consequent character of the growth, are 
borne in mind. 
To refer to the figures, it is seen that, whilst the nitrogen 
supplied in the seed was only 0-03 gram or less, that in the 
products of growth was 0-2822 gram in pot 1, 0-5361 in pot 2, 
0-4857 in pot 3, and 0-6600 in pot 4 ; and the gains v/ere more 
than ^ of a gram in pot 1, more than \ a gram in pot 2, about 
|ths of a gram in pot 3, and more than fths of a gram in 
pot 4. 
The third division of the Table shows the total nitroefen at 
the commencement (in soil and seed together), at the conclusion 
(in soil and total vegetable matter grown), and the gain;!. 
But the significance of the results is more clearly seen in 
