Reducing the Results of Experiments, 95 



not quarrel with me for giving accurate tables merely because 

 the special application made of them did not require the full 

 degree of accuracy attained. 



2. Mr. Lupton is then " startled " to find that " in the 

 case of experiments, the whole object of which is extreme 

 accuracy/' there is a difference of yoqo on eacn s ^ e °f *he 

 mean in the eight analyses made of the stock acid* — analyses, 

 by the way, which were not my own, and the results of which 

 I did not take. But will he tell me how the accuracy of my 

 results would have been affected by a constant error of even 

 ToVo * n ^ ne strength of all the solutions used ? Or how my 

 conclusions would have been affected if, for instance, I had 

 found 58'08 instead of 58'14 per cent, as the position of a 

 change, when, for other reasons, this position could not, as a 

 rule, be correctly ascertained within 0*5 or 1*0 per cent. ? 



3. Mr. Lupton complains (p. 420) that I have given no 

 details as to my method of drawing curves. I had thought 

 that the details on pp. 68, 69, and 184, he. cit., would have 

 been sufficient, and the only important omission I appear to 

 have made was that I did not mention that the flexible 

 ruler was held at its extremities, the middle portion only 

 being used for the drawing • but this omission cannot account 

 for Mr. Lupton's extraordinary conclusion, p. 421, that "hence 

 the hydrates of sulphuric acid are apparently made to depend 

 upon the flexibility of the steel lath used/'' Let Mr. Lupton 

 convince himself of the fallacy of this statement by a little 

 experimentation on the curves depicted above with laths of 

 different flexibility. 



4. Mr. Lupton refers to Prof. Lodge's " word of warning " 

 to experimentalists on the over-pressure of formulas (' Nature,' 

 July 18, 1889 ; see also my answer, August 8, 1889). It 

 is certainly remarkable that Mr. Lupton should consider that 

 this warning applies to me, who used not a single formula, 

 and not to himself, who is relying entirely on empirical 

 formuke. 



5. In the same way Mr. Lupton refers to Prof. Arrhenius's 

 premature criticism of my work, mentioning with approval 

 that Arrhenius pointed out " that my equations for sulphuric 

 acid required the introduction of nearly sixty constants.'" 

 Mr. Lupton is apparently in ignorance of my answer to 

 Arrhenius (Phil. Mag. xxix. p. 433), and of the fact that since 

 the time when this answer and the details of my results were 

 published Arrhenius has never made good his criticism. I 



* Why does Mr. Lupton make only a partial quotation from my table 

 of analyses (loc. cit. p. 72) ? And why does he not calculate the probable 

 error of the mean of these analyses ?' It is only iM Lq, 



