E 54 3 
almoft inoffenfive in this refpect, 
which I was the lefs furprized at, as 
their fubftance is but little different 
from the ftalks. The roots of a muf~ 
tard plant gave in the fun-ihine a 
good deal of air, when kept under 
water; which air was worfe than com- 
mon air, and extinguifhed a candle. 
Thefe roots even corrupted good air 
in the middle of the fun-fhine. 
A plant, with roots and all, placed — 
in ajar full of water, did, however, 
yield dephlogifticated air; fo that 
the bad effect of the roots was over- 
balanced by the plant itfelf, 
5 E C- 
