CHEMISTRY. 199 



of an inch from both surfaces, and the four edges of a duplicate 

 exposed specimen, which, after repolishing, exhibited the same 

 color. The glasses exposed were what are called colorless window- 



f lasses, though varying in hue from the whitest French plate to the 

 arkest green English sheet-glass. Long exposure of really col- 

 ored glasses caused no change, except in the purple becoming 

 darker; but it is possible that time would effect a change even in 

 these, a fact which he hoped to ascertain by examination of some 

 of the old stained glass of European churches. Familiar exam- 

 ples of these colored window-glasses, of a pinkish hue, may 

 be seen in many houses in Boston fronting the Common ; in 

 many windows several colors may be seen, according as each 

 pane has been exposed for a longer or shorter time. Under our 

 sunny skies, the change is much more rapid than in the more 

 humid and less clear atmosphere of England. 



He finds his first results confirmed by subsequent experiments, 

 that almost all kinds of the glass above mentioned undergo this 

 change from the influence of sunlight. The cause of this change 

 has been referred to the presence of oxide of manganese, the ox- 

 ide of iron, sulphur, etc. Some think there are not facts enough 

 accumulated to decide the question of cause. It is most likely 

 due to the presence of the oxide of manganese, which is used to 

 give glass a white color; if the materials were perfect this sub- 

 stance would be unnecessary, and the change would not occur ; 

 thus we find that the glass made from the very pure Berkshire 

 SMid very seldom changes. No change is observed in lead glass, 

 unless manganese is also present. This change must not be con- 

 founded with what is called " rust" in glass, which is a mere me- 

 chanical disintegration of the surface, from the washing out by 

 the rain of the soda in the glass. 



The phenomena illustrated by these experiments are very in- 

 teresting; and he suggested that the examination by spectrum 

 analysis might solve the problem as to the cause of the change 

 of color. 



In the " Comptes Rendus " for Jan. 14, 1867, M. Pelouze (who 

 must have been acquainted with Mr. Gaffield's account, published 

 years before) makes a communication on glass, in which, after 

 stating his opinion that there is no glass, in commerce, which does 

 not change its color on exposure to the sun, goes on to say that 

 pure glass, free from alkaline sulphate and oxide of iron, is not 

 colored by the solar rays ; that, for equal weights, the sesquiox- 

 ide of iron colors it less than the protoxide, and that the yellow 

 color manifested is much more intense than would have been pro- 

 duced by the same amount of iron in the state of peroxide ; that 

 a mere trace, almost imponderable, of sulphuret will color glass 

 yellow. In glass which grows yellow on exposure to the sun 

 there is protoxide of iron and sulphate of soda; light causes chem- 

 ical reaction, whence results peroxide of iron and sulphuret of 

 sodium; heat causes an opposite reaction, reproducing the sul- 

 phate of soda and the protoxide of iron, with a return in the glass 

 of the original color. Analysis shows in the colored glass a very 



