556 
PROFESSOR H. MARSHALL WARD OH THE 
close of their period of vegetation, and especially as the seeds ripen : in accordance 
with this, their root-system is abundant and far-reaching in the soil poor in nitrogen. 
The nitrogenous materials not employed at the time are stored up in the root- 
tubercles, which are emptied as the seeds begin to ripen, the nitrogenous contents 
being absorbed for the good of the latter. 
As I have shown, this hypothesis is utterly untenable with regard to the tubercles 
on the roots of Vida Fab a, and everything points to its being equally so for the 
other Leguminosse : not only is Tschirch’s attempt to explain away the hyphse and 
the gemmules a failure, but his drawings indicate that he is not in possession of the 
histological facts necessary to constitute him an authority on the subject of the 
development and physiology of the root tubercles of the Leguminosee, whence his 
repeated assumptions lose in value. 
It now remains to gi ve some account of the experimental cultures which have had so 
much importance in leading to the discovery of the infecting hypha, and to the general 
results given above. I found some time ago that Beans grown in a soil which had 
been burnt did not develope the tubercles, and that the same results followed as a 
rule when the cultivations were made in solutions of nutritive materials. I have since 
observed that if cultures in burnt soil are watered with washings of common garden 
soil, or of compost heaps, the tubercles are almost sure to be developed in the course 
of a few weeks ; even river or pond water is dangerous—at least, in one case I traced 
the infection to the water with which I had watered the plants. Moreover, as I have 
shown, no experiment can be trusted if the seedling has been in contact with ordinary 
sand, turf, or leaf-mould. 
I find that a very convenient process for obtaining the tubercles is to allow 7 the 
seed to germinate for a week or ten days in common garden soil, and then carefully 
lift the seedling, wash its root very gently, and place it in a split cork so that its 
root goes on developing in a nutritive solution. In the majority of cases the young 
tubercles are quite evident some time during the fourth week following, and a little 
experience enables one to detect them at least a w r eek sooner than that. 
Now, seedlings carefully germinated in the same garden soil, previously sterilised 
by being burnt, and subsequently treated similarly, do not yield tubercles at all.* 
This fact is in itself strong' evidence of the infection coming from the soil. 
It may be well to state the composition of the nutritive solutions I have employed, 
and the mode of treating the plants grown in them. 
In all the cases referred to the solution employed was that given by Sachs in his 
‘ Vorlesungen liber Pflanzenphysiologie,’! care being taken that all vessels and utensils 
were clean, and the water and reagents pure. 
* Or only sporadically—a fact explained by cbance infection during the culture, 
t P. 342. English edition, p. 284. 
