100 ESSAY V. 



of the flame. If this is really the case, it would follow that, 

 by repeating the experiment, some of these particles would 

 always separate if a sufficiently strong flame were applied, 

 and it would be impossible to expel the red colour afterwards. 

 I dare not, however, advance this conjecture, though it is 

 grounded on some experiments, as a matter of certainty. 

 Accidents may sometimes make a thing appear in a quite 

 different light from what it really is, especially in experi- 

 ments upon such a small scale, where one cannot be careful 

 enough in forming conclusions. 



More experiments would elucidate this point ; but it 

 would perhaps hardly be worth while to make them, since 

 they seem to be more curious than useful. 



Quite different is the case with regard to the destruction 

 and restoration of the colour of manganese itself by means 

 of the simple flame ; for the several experiments I made all 

 agreed in the same result. In order to confirm them the 

 more, I have made experiments lately with manganese from 

 Upton Pine, near Exeter, in England, and have observed no 

 difference. 



The brown flame of the candle has probably more 

 phlogistic particles than the blue one. 



The presence of phlogiston seems in most cases to 

 communicate colour. Mineral substances, which contain 

 little phlogiston, are often with great difficulty deprived of 

 it ; a violent and long-continued heat is frequently required 

 for this purpose ; sometimes, however, we succeed on adding 

 more phlogiston, and by means of certain management with 

 regard to the fire itself. 



