30 HOW CROPS FEED. 
Inaccuracy of Ville’s Resulis.—In comparing the in- 
vestigations of Boussingault and Ville as detailed in their 
own words, the critical reader cannot fail to be struck 
with the greater simplicity of the apparatus used by the 
former, and his more exhaustive study of the possible 
sources of error incidental to the investigation—facts which 
are greatly in favor of the conclusions of this skillful and 
experienced philosopher. Furthermore Cloéz, who was 
employed by a Commission of the French Academy to 
oversee the repetition of Ville’s experiments, found that a 
considerable quantity of ammonia was either generated 
within or introduced into the apparatus of Ville during 
the period of the trials, which of course vitiated all his 
results. 
Any further doubts with regard to this important sub- 
ject have been effectually disposed of by another most 
elaborate investigation. 
Research of Lawes, Gilbert, and Pugh,—In 1857 and 
58, the late Dr. Pugh, afterward President of the Penn- 
sylvania Agricultural College, associated himself with 
Messrs. Lawes and Gilbert, of Rothamstead, England, 
for the purpose of investigating all those points con- 
nected with the subject, which the spirited discussion of 
the researches of Boussingault. and Ville had suggested as 
possibly accounting for the diversity of their results. 
Lawes, Gilbert, and Pugh, conducted 27 experiments on 
graminaceous and leguminous plants, and on buckwheat. 
The plants vogetated within large glass bells. They were 
cut off from the external air by he bells dipping into 
mercury. They were supplied with renewed portions of 
purified air mixed with carbonic acid, which, being forced 
into the bells instead of being drawn dhrough them, ef 
fectually prevented any ordinary air from getting access 
to the plants. 
To give an idea of the mode in which these delicate investigations are 
conducted, we give here a figure and concise description of the appara- 
