( 38o ) 



caufe a mucli ftronger detonation ? But 

 the proof that admits of no reply, or rather 

 the abfolute demonftration of the infuffi- 

 ciency of this hypothefis, is the following : 

 When thefe tumors burft by the force of 

 the imprifoned fluid at the time it difen- 

 gages itfelf, were this fiuid hydrogenous 

 gas, it would take fire at the fame moment, 

 and cover with flame the furface of the 

 lava ; but it is certain that not the lead ap- 

 pearance of the feebleil flame is difcoverable 

 in the lava in any eruption ; as I can af- 

 firm from the various obfervations I have 

 made, at a very fhort diftance, of the moft 

 minute accidents that take place within the 

 crater. 



Compelled, ^ therefore, to abandon this 

 explanation, it occurred to me whether at- 

 mofpheric air might not pofllbly produce 

 the phenomena in queftion. I confidered, 

 however, that, before this hypothecs was 

 admitted, it would be necefTary to fhow 

 that the atmofpheric air had free accefs 

 from without the mountain to the volcano ; 



that 



