. 
218 Proceedings of Philosophical Societies. [Marcu, 
Vapour of naphtha. ..........eceeeeeee0 177'870 
oil of turpentine. .............. 177:870 
nitric acid (sp. gr. 1:494) ....... 531°990 
‘ammonia (sp. gr. 0°978) ........ 837-280 
vinegar (sp. gr. 1:007). ........ 875°000 
Dr. Ure terminates his paper by a very ingenious speculation 
on the connexion existing between the latent heat, elastic force, 
and specific gravity of gases or vapours. He conceives that 
when their tension is the same, the product of their densities 
into their latent heat will also be the same ; or, in other words, 
that the elasticity is always as thespecific gravity multiplied into 
the latent heat. I have no doubt that we might make consider- 
able progress in the generalization of the properties of elastic 
fluids by the application of mathematical reasoning; but it 
would be requisite in the first place to be possessed of a very 
accurate set of experiments on their expansion, latent heats, 
specific gravities, &c. Till these are furnished, mathematical 
reasoning, however ingenious, will serve only to lead us astray. 
Mr. Dalton in the first volume of his Chemistry, and M. Biot in 
his late work on Physics, have afforded us some striking exam- 
ples of the little advantage which results from the application of 
mathematical reasoning to loose or inaccurate data. 
(To be continued.) 
ARTICLE XI. 
Proceedings of Philosophical Societies. 
ROYAL SOCIETY. 
Jan. 21.—A paper, by Dr. T. Young, was read, entitled 
‘* Remarks on the Advantage of Multiplied Observations in the 
Physical Sciences, and on the Density of the Earth.” After 
some observations upon the application of the doctrine of 
chances to the physical sciences, the author showed that the 
combination of many different causes of error, each lable to 
change, has a tendency to diminish the aggregate variation of 
their joint effect. From calculation he then inferred, that the 
original conditions of the probability of different errors do not 
considerably modify the conclusions respecting the accuracy of 
the mean result, because their effect is included in the magnitude 
of the mean error from which these conclusions are deduced. He 
also showed, that the error of the mean arising from this limita- 
tion is never likely to be greater than Sths of the mean of all the 
errors divided by the square root of the number of observations. 
The author then proceeded to the application of the doctrine of 
chances to literary and historical subjects, particularly with 
, 9 
