38 
by vegetation ; and the longer the plants continued 
in it, the worse it became. As the general result of 
all his experiments, however, Dr Priestley thought 
it probable, that the vegetation of healthy plants, 
growing in situations natural to them, had a salutary 
influence on the air -in which they grew, since one 
clear instance of the melioration of air in these cir- 
cumstances, should, says he, weigh against a hun- 
dred, where the" air is made worse by it *. 
252. Such is the series' of experiments from which 
Dr Priestley endeavoured to shew that plants, by 
their vegetation, purify corrupted air : and such are 
the terms in which his opinion is finally expressed. 
The probability only which he attached to this opi- 
nion, has, by others, been converted into certainty ; 
and time hitherto has appeared rather to confirm, 
than to correct the error. The contrariety, how- 
ever, that pervades all his experiments, and the 
" fearful odds" against which his general conclusion 
is drawn, are, of themselves, sufficient to beget doubt 
and distrust ; and if, to these circumstances, we add 
the numerous sources of fallacy which have now 
been stated, and the want of precision in the methods 
of analysis employed fj we must be convinced that 
* Obs. on Air, abridged, vol. iii. 265. 273. 275. 
j- Dr Priestley did not discover the use of nitrous gas, as aneu- 
diometrical test, till the close of the year 1 772, when his experi- 
ments on plants in that year were nearly brought to a conclusion. 
Previous to this period, he employed the burning of a candle, and 
the respiration of mice, as tests of the purity of air*; and it is 
worthy of remark, that his experiments in 1778, when his me- 
thods of analysis were improved, totally contradict those of an 
earlier period, when they were confessedly imperfecU 
