ATMOSPHERIC AIR AS THE FOOD OF PLANTS. 27 
their command, besides that contained in the seed itself, 
should be the free nitrogen of the atmosphere. For this 
purpose he prepared a soil consisting of pumice-stone and 
the ashes of stable-manure, which was perfectly freed from 
all compounds of nitrogen by treatment with acids and in- 
tense heat. In nine of his earlier experiments the soil thus 
prepared was placed at the bottom of a large glass globe, . 
B, fig. 4, of 15 to 20 gallons’ capacity. Seeds of cress, 
dwarf beans, or lupins, were deposited in the soil, and a 
proper supply of water, purified for the purpose, was add- 
ed. After germination of the seeds, a glass globe, D, of 
about one-tenth the capacity of the larger vessel, was filled 
with carbonic acid (to supply carbon), and was secured air- 
tight to the mouth 
of the latter, com- 
munication being 
had between them 
by the open neck at 
C. The apparatus 
was then disposed in 
a suitably lighted 
place in a garden, 
_ and left to itself for 
a period which va- 
ried in the different 
experiments from 14 
to 5 months, At 
a the conclusion of the 
} trial the plants were 
py ‘ lifted out, and, to- 
Fig. 4. gether with the soil 
from which their roots could not be entirely separated, 
were subjected to chemical analysis, to determine the 
amount of nitrogen which they had assimilated during 
growth. 
The details of these trials are contained in the subjoined 
