128 Dr. Thomson's Experiments to determine the [Aug. 



I have repeated this experiment several times, and it leaves no 

 doubt in my mind that an atom of mercury weighs 25, and an 

 atom of peroxide 27. Protoxide of course must weigh 26. 



I made repeated attempts to convert mercury into a peroxide 

 without any loss, but was not able to succeed. It may, perhaps, 

 be worth while to relate a single experiment of this kind to 

 enable the reader to judge of the degree of accuracy which I was 

 able to reach by this method ; because it may serve in some 

 measure to account for the deviations from each other which we 

 observe in the different experiments which have been published 

 by different chemists to determine the composition of the oxides 

 of mercury. 



I put into a flask previously accurately weighed, and the 

 weight marked on it by a diamond, 25 grains of mercury. A 

 quantity of pure nitric acid capable of dissolving it was poured 

 ■over it. After the solution was completed, the flask was placed 

 on a sand-bath, and left in a moderate heat till the whole con- 

 tents of it were evaporated to dryness. The flask Avas then 

 exposed to a temperature of about 600°, and kept in it till all 

 nitrous acid fumes ceased to exhale. The red oxide prepared in 

 this way not being quite free from white particles, it was obvious 

 that the nitrate had not been completely decomposed. I, there- 

 fore, heated a hessian crucible to redness, and introduced the 

 flask into it, holding it at a little distance from the sides and 

 bottom of the crucible. This occasioned the discharge of some 

 more nitrous acid fumes, and the peroxide at the bottom of the 

 flask had assumed a very fine red colour. I considered it in 

 consequence as pure ; but there had sublimed upon the sides of 

 the flask traces of a yellow matter, which was tasteless, and 

 dissolved in nitric acid without effervescence. It certainly did 

 not amount to 1-1 00th part of a grain. I was, therefore, unable 

 to determine its nature, though it was probably nothing but per- 

 oxide of mercury. I had covered the mouth of the flask with a 

 piece of glass before I introduced it into the crucible. On this 

 piece of glass a *mall portion of metallic mercury had sublimed. 

 I found this portion to weigh 0-2 gr. The weight of the perox- 

 ide of mercury was 26*7 grs. Had the 0*2 gr. of sublimed mer- 

 cury been peroxidized, it would have weighed 0*216 gr. Hence 

 the weight obtained was : 



Peroxide in flask , 26*7 



Ditto sublimed 0*216 



26-916 

 Loss 0-084 



27-000 



Here the loss amounted to 0-084 gr. or about -j4- part of the 

 •whole. I have no doubt that this loss was owing to a small 



