89

temperature of the flame. This tempera-ture, however, evidently presents the limitto experiments of this kind, for bodies ex-posed to flame can never be hotter thanflame itself; whereas in the voltaic appara-tus there seems to be no limit to the heat,except the volatilization of the conduc-tors.

The temperatures of flames are proba-bly very different. Where, in chemicalchanges, there is no change of volume, asin the instance of the mutual action ofchlorine and hydrogene, prussic gas (cya-nogen) and oxygene, approximations totheir temperatures may be gained from theexpansion in explosion.

I have made some experiments of thiskind by detonating the gases by the elec-trical spark in a curved tube containingmercury or water ; and I judged of the ex-pansion from the quantity of fluid thrownout of the tube: the resistance opposed bymercury, and its great cooling powers, ren-dered the results very unsatisfactory in thecases in which it was used but with water,cyanogen and oxygene being employed,they were more conclusive. Cyanogen andoxygene, in the proportion of ope to two,