69

appeared cherry red, the expansion was notto more than 2; 5, and a part of this mightperhaps have been apparent only, owing tothe collapsing of the glass tube before itactually melted. It may be supposed thatthe oxidation of the fusible metal may havehad some effect in making the expansionappear less ; but in the first experiment theair was gradually brought back to its ori-ginal temperature of boiling water, whenthe absorption was scarcely sensible. If M.Gay Lussac’s conclusions be taken as theground work of calculation, and it be sup-posed that air expands equally for equal in-crements of temperature, it would appearthat the temperature of air capable of ren-dering glass luminous must be 1035° Fah-renheit.*

M. de Grotthus describes an experi-ment in which atmospheric air and hydro-gene, expanded to four times their bulk over

* The mode of ascertaining temperatures as high asthe point of fusion of glass by the expansion of air, seemsmore unexceptionable than any other. It gives for thepoint of visible ignition nearly the same degree as thatdeduced by Newton from the times of the cooling ofignited metal in the atmosphere.