76 Williamson On (he Atomicity of Aluminium. [Feb. 9, 



drogen, or to doubt any other of the results of our comparison of their 

 compounds ? or has it led chemists to diffusion experiments with its vapour, 

 proving it to contain uncombined HC1 and NH 3 , each occupying its own 

 natural volume ? Has it not been proved that at the temperature at which 

 sal-ammoniac vapour was measured, its constituents mix either without 

 evolving heat (that invariable function of chemical action), or, according to 

 another experimentalist, with evolution of far less heat than of the whole 

 quantity of hydrochloric acid and ammonia combined, on coming together 

 at that high temperature ? 



Again, SO 1 H 2 is known to represent the formula of one molecule of 

 hydric sulphate, yet the vapour formed from it occupies nearly the bulk of 

 two molecules. Has this fact cast any doubt on the atomic weights of the 

 elements S, O, or H ? Or has it led to the discovery of peculiarities in the 

 constitution of the vapour which would probably have escaped notice had 

 they not been anticipated by theory, peculiarities which go a long way 

 towards bringing the apparent anomalies within the law ? 



Nitric peroxide, N 2 O 4 , was considered, from our knowledge of other vo- 

 latile compounds of nitrogen, to be anomalous in its vapour-volume being 

 N 2 O l =4 vols. ; and we have been shown by the experiment of Messrs. 

 Playfair and Wanklyn, that the anomaly almost disappears when the com- 

 pound is evaporated by the aid of a permanent gas at a temperature consi- 

 derably below its boiling-point, as its theoretical molecule N 2 O 4 is then 

 found to occupy the two volumes which every undecomposed molecule 

 occupies. This explanation seems to me to be the more entitled to grave 

 consideration on the part of the discoverers of the new aluminium com- 

 pounds, from the fact that the evidence in favour of it has been admitted 

 to be conclusive by Dr. Odling, who classes nitric peroxide by the formula 

 N 2 O 4 = 2 vols. among compounds with normal vapour-densities, in virtue 

 of the fact that at low temperatures it can be obtained with that density, 

 though having half that density at higher temperatures. 



The arguments for admitting that the low vapour- densities of the alumi- 

 nium compounds are anomalous are even stronger than those which are ad- 

 mitted in the case of nitric peroxide ; for it did require very severe super- 

 heating to get the aluminium compounds to near four volumes, whereas it 

 required very ingenious devices to get nitric peroxide out of the four-volume 

 state. 



Such guiding principles as we have acquired in chemistry are the 

 noblest fruits of the accumulated labours of numberless patient experi- 

 mentalists and thinkers ; and when any new or old fact appears to be at 

 variance with those principles, we either add to our knowledge by discover- 

 ing new facts which remove the apparent inconsistency, or we put the case 

 by for a while and frankly say that we do not understand it. 



The decision of the atomic weight of aluminium has involved greater 

 difficulty than was encountered in the case of most other metals, owing to 

 the fact of our knowing only one oxide of the metal, and salts correspond- 



