$2 General Refleciions ow Heat. 
have endeavoured to deduce all the phenomena from reasou- 
ing on the nature of that agent; the modern chemists, pursu- 
ing the reverse order, end at heat, satisfied that they have ac- 
counted for each effect, when they have ascertained its inva- 
riable antecedent. 
We may assert in general, then, that a fact in chemistry is 
adequately accounted for, when we have ascertained its in- 
variable antecedent ; but since the process would be tedious 
to do this for every chemical change, by an analytical exam- 
ination of the result, we refer each particular change to a 
class of phenomena which it resembles, of which a sufficient 
number have been examined, to show their invariable con- 
nexion with their antecedent. Thus, when we hold melted 
lead over the fire in a ladle, we account for the dross that 
forms on its surface, by saying that the lead has combined 
with oxygen; not because we have, in ‘this instance, ascer- 
tained the fact, but because its external appearance is such 
as to justify usin classing it among similar phenomena, where 
the fact has been ascertained by actual experiment. This is 
another fact to show, that our explanations of phenomena are 
little else than mere classifications. The same remark holds 
true in every department of physics : we account for a phy- 
sical fact by assigning its proximate cause or antecedent ; but 
we ascertain this, not sofrequently by actual experiment, as 
by reducing the fact in question to a class of phenomena, 
whose cause has been ascertained by actual experiment. 
With these principles in view, let us now proceed to in- 
quire how far Lavoisier accounted for the phenomena of 
combustion. 
It appears to me, that Lavoisier must have been 
considered as adequately accounting for combustion, pro- 
vided that the combination of oxygen with a base had always 
proved to be, as it appeared to him, an invariable antecedent 
to that process. In the progress of our science, however, 
a number of examples of combustion have been discovered, 
in which no oxygen is present. Thus certain substances 
burn in an atmosphere of sulphuretted hydrogen, or chlorine, 
or iodine ; and even certain solids into which no oxygen 
enters, as sulphur and iron filings, or sulphur and copper 
filings, exhibit similar phenomena in their action on each 
other. Oxygen therefore has now lost its character of inva- 
riable antecedent, and its combination with ‘oflammable 
bodies can no longer be pronounced the cause of combustion. 
