104 KEPOET— 1891. 



manner in which the molecular volume of a compound, taken 

 as a whole, and considered as one of its characteristic physical 

 properties, may be correlated with other such properties of 

 the same compound. The property which is obviously most 

 likely to have a close connection with the volume at the boiling- 

 point is the temperature of ebullition itself, reckoned, of course, 

 on the absolute scale ; and, as this temperature varies with the 

 pressure, that pressure, or (what is the same) the vapour ten- 

 sion of the liquid at its boiling-point, must also be taken into 

 account. 



The few records that we yet possess of the variation of the 

 boiling-point, pressure, and molecular volume of a liquid from 

 low pressures up to the critical point, are of great interest and 

 value, but the data have been as yet too few to lead to any 

 very certain general results, such as theoretical considerations 

 point to. 



I was lately led to a study of the connection of the molecular 

 volumes and boiling-points with one another and with the 

 chemical characters of liquids by observing a certain regularity 

 in the behaviour of the ethides of the elements. Having deter- 

 mined the molecular volumes of the three closely-related com- 

 pounds, ethyl chloride, bromide, and iodide, I was struck with 

 the fact that they were directly proportional to the absolute 



boiling-temperatures. An examination of all the available 



Y 

 recorded data proved that this equality of the ratio — always 



characterizes the fluorine, chlorine, bromine, and iodine com- 

 pounds of the same organic radicle. The actual deviation from 

 perfect equality exceeds 2 per cent, of the total value in only 

 two cases out of the eleven groups of substances examined ; but 

 a small variation in each group, after allowance is made for 



experimental error, does seem to occur in such a way that — 



is greatest for the fluoride and least for the bromide. Prac- 

 tically this variation may be neglected. 



The question whether — is equal for the compounds of the 



same organic radicle with any other natural group of elements 

 was one which could not, and cannot yet, be answered by an 

 appeal to du-ect experiment, as there is a want of data with regard 

 to the molecular volumes. But by using Thorpe's values for the 

 average atomic volumes of the elements it was possible to calcu- 

 late the molecular volumes with sufiicient approximation to the 

 truth, and then, making use of known boiling-points, to calculate 



the values of — . This was done, and the result left no doubt 

 at all that the law observed to hold good for compounds of the 



