162 Dr. Thomson on the [Sept. 



numerous and intricate corrections, which Biot describes at such 

 length in his Traite de Physique. It renders an absolute vacuum, 

 or even a very good vacuum, unnecessary, and enables us to 

 obtain a more correct result by a single experiment of five 

 minutes' duration than Berzelius was able to accomplish by 

 almost innumerable repetitions, when he endeavoured to deter- 

 mine the specific gravity of sulphurous acid gas. 



It is proper to mention before going further that the weighings 

 and the exhaustions of the flask were all conducted by my 

 assistant Mr. Harvey, who indeed performed the whole of the 

 experiments, and to whose accuracy and care 1 have been very 

 often indebted in other chemical investigations. 



Gay-Lussac first proved by a series of very satisfactory induc- 

 tions that the gases unite with each other in certain definite pro- 

 portions which may be denoted by very simple numbers. One 

 volume of one gas either unites with one volume, or two volumes, 

 or three volumes, or 1^ volume of another gas. He showed too 

 that a determinate change always takes place in the volume of 

 the new formed gas. The volume sometimes remains unaltered. 

 Thus one volume of chlorine gas and one volume of hydrogen 

 gas, when united together, constitute two volumes of muriatic 

 acid gas. More generally the volume is reduced to one half. 

 Thus a volume of oxygen gas united to a volume of vapour of 

 sulphur constitutes only one volume of sulphurous acid gas. In 

 one particular case, the volume is reduced to one-fourth part of 

 the original volume of the two constituents. Thus when one 

 volume of hydrogen gas and one volume of vapour of carbon 

 unite together so as to constitute olefiant gas, the bulk of the 

 defiant gas is not two volumes, nor even one volume, but only 

 half a volume. 



It is this remarkable general fact which g'lves such importance 

 to the accurate determination of the specific gravity of the 

 gases. We see from it that a simple relation exists between the 

 specific gravity of the gases and the weights of their atoms. 

 This relation is such that if we know the one we can very easily 

 deduce the other. I have shown this relation in a paper which 

 I published in a preceding volume of the Annals of Fhilosophy. 

 Indeed I was led by my knowledge of it to conclude that many 

 of the specific gravities of gases already determined by preceding 

 experimenters, were inaccurate. I even went the length of 

 adopting a set of hypothetical specific gravities of gases as the 

 true ones in the last edition of my System of Chemistry. My 

 object in the present paper is to give the result of a set of expe- 

 riments made on purpose to ascertain how far these hypothetical 

 numbers w ere borne out by actual trial. 



There tuj i'uur simple or undecompounded gases, the specific 

 gravities of which it is important to determine in the first place; 

 because it is iVom them we are enabled to deduce the verification 



