of Sulphuric-Acid Solutions. 135 



solutions at the points indicated, and strong evidence that 

 their occurrence had an intimate connexion with the existence 

 of a corresponding hydrate. Wishing, however, to make 

 assurance doubly sure I carried out, after my first paper icas 

 icritten, a series of experiments on the freezing-points. The 

 isolation of the tetrahydrate, whose percentage corresponds to 

 one of the most feebly marked of the changes which I had 

 previously discovered, amply justified my view of their con- 

 nexion with hydrates ; but the entire agreement of the minor 

 changes, marked on the diagram, with those previously 

 found was perhaps even more satisfactory. Of the thirteen 

 changes previously found to exist between 2 and 98 per cent, 

 nine were confirmed, while as regards the others, which were 

 situated in regions of very low freezing-point, data were either 

 insufficient or entirely lacking. 



Remembering that the results set out above were obtained 

 by the application of the same method and the same form of 

 curves to figures which in the majority of cases differed 

 entirely from one another in general form, the question arises 

 as to how such concordance could have been obtained. If we 

 refuse to conclude that the results are due to some property 

 possessed by all the figures in common of exhibiting changes 

 of some sort at these points, we are driven, I think, to explain 

 it in one of the three following ways : — (1) That I knowingly 

 " cooked " my results — an explanation which has not yet 

 been offered ; (2) that I did so unconsciously ; which would 

 seem to be quite impossible, owing to the properties studied 

 being entirely different, and to my having obtained the 

 same results with every different form of plotting and on 

 every different scale*; or (3) (which is obviously absurd) 



* I may mention a very strong proof of how independent my results are 

 of any unconscious " cooking," and of the particular form in which the 

 data are presented for examination. I received a short time ago from my 

 friend Mr. Hayes a series of numbers which I took to he what I was at 

 the time expecting from him — namely, a set of imaginary experiments 

 constructed from equations, with a view to seeing whether I would find out 

 the points at which the equation had been changed. After sending him 

 the results of my examination, I learnt that the values sent were my own 

 experimental results for the densities at 18° from 95 to 19 per cent., 

 metamorphosed beyond recognition, and I also learnt that 1 had cor- 

 rectly located the breaks which I had mentioned in my published work. 

 The results were 



92, 85-6, 79-7, { SOm f. breaks ' but } 49-6, and 30-4 per cent, ; as 

 ' '\ position uncertain J r 



against 93-6,84-5, 78*0, 72*8, 58, 51'0, and 29-5 per cent, for- 



merly given. 



The values which Mr. Hayes sent to me were .r=1250 (100-/0, and 



^ = 10 6 .— . _ - (p = percentage, and s = density). 



S" CIJ) 



The data for finding the first of these breaks were more meagre than 

 they should have been, and a note was made by mo that its position might 

 be at a percentage higher than 92. 



