44'4< M. G. Bontemps on some modifications in 



green opake, is cut to different depths. These effects are 

 produced by inequalities of cooling, as we have seen for man- 

 ganese and copper. 



5. Gold. 



Oxide of gold gives to the glass a pink tint, which by an 

 increase of quantity may attain a 'purplish-red. For this pur- 

 pose a small proportion of precipitated purple of Cassius is 

 added to the mixture of flint-glass ; but by the first melting 

 this mixture gives only a colourless transparent glass, which 

 must be heated again to show the pink colour. If, for instance, 

 a small solid cylinder has been formed with this first melted 

 glass, when cold it is quite white; but if this cylinder is 

 afterwards exposed to the heat of the working-hole of the fur- 

 nace, we see it acquire the red colour gradually as it is pene- 

 trated by heat ; and this colour remains fixed wlien the cylin- 

 der is gradually cooled again in the annealing kiln. 



I have remarked also, that by varying the degrees of heating 

 apiece of this glass of some length at a high temperature, and 

 re-cooling it several times, a great number of tints, varying 

 from blue to pink, red, opake yellow and green, may be pro- 

 duced. But I am not certain that this effect might not be 

 attributed to some fractions of silver mixed with the gold used; 

 and the only point that remains quite positive, is the fact of 

 the pink colour showing itself by a second Jiring in the glass 

 into the composition of which gold enters. 



To these results of colouring by metallic oxides, I shall add 

 an effect produced in the colouring of glass by charcoal, which 

 effect is of the same nature as those mentioned in the colour- 

 ing by copper and gold. 



An excess of charcoal in the mixture of a silico-alkaline 

 glass gives a yellow colour, which is not so bright as the yel- 

 low from silver, but good enough to be used in church win- 

 dows ; and sometimes, according to the nature of the wood 

 from which the charcoal has been made and the time at which 

 it has been cut, this yellow colour may be turned to dark red by 

 a second Jire. 



I doubt, indeed, whether all the results which I have men- 

 tioned can be explained only by various degrees of oxidation 

 of the metals. This multiplicity of colours, greater than the 

 number of oxides described for each metal, must lead us to 

 consider whether those pha^nomena are not the consequence of 

 ■physical laws. It is the peculiar character of our time, and the 

 result of the immense progress accomplished in chemistry and 

 natural philosophy, to bring their study to some united views, 

 which render the connexion of these two sciences indissoluble. 



