4 o8 SCIENCE PROGRESS 



chemistry, whilst in books on mathematics the subject affords 

 examples of methods of integration and is distributed according 

 to the methods required for the different cases. We have only 

 to look at the published work 'of many well-known chemists to 

 realise how much we are in need of sound teaching in connection 

 with this subject of mass action and to see what extraordinary 

 ideas exist as to what constitutes a constant and as to the 

 importance that can be attached to obtaining constants. To the 

 most credulous, it must seem ridiculous that the attempt should 

 be made to base any serious deductions on the results obtained 

 by applying the law of mass action to the effects produced by 

 the action of a poisonous solution on a frog's leg, for example. 



In the first place, the student should have it impressed on 

 him that the mass action law is a generalisation of an axiomatic 

 nature, never apparently obeyed exactly and incapable therefore of 

 absolute proof; that even if the correct assumptions are made 

 with regard to the number and nature of the interacting molecules 

 there are many other disturbing factors, as a rule, that cannot be 

 taken into account. Then with regard to the deduction of the 

 mathematical formulae for the investigation of the various cases, 

 not only should all possible simple cases be dealt with but the 

 merits and defects of alternative formulae should be pointed out. 

 Writers of books on Physical Chemistry, for instance, James 

 Walker and Senter, who presumably know the relative merits 

 of the two formulae 



^ = t [ °s ^r~x and k = w, lQ g 



a — x. 



for the investigation of a unimolecular change, for some reason 

 or other give only the former without even hinting to the poor 

 student that they have passed off on to him a second-rate and 

 untrustworthy article ; that his values of K will all be affected 

 by an error in the one value a ; that near the beginning of the 

 experiment when a is determined the chance of error due to the 

 temperature not having become constant is greatest ; and that 

 the use of this formula does not allow of a strict comparison of 

 different stages of the interaction. B}- using the second formula, 

 which is immediately deducible from the first, the possibility of 

 an error in all the values of K due to an error in the one value a 

 is obviated ; if the values of K show no tendency to rise or fall 

 but remain constant within the limits of experimental error this 



