Reducing the Results of Experiments. 97 



fitting of the stopper offers other capillary openings besides 

 the legitimate one through the stopper; and this, from its 

 very nature, is a defect which may be absent either through 

 accident or through careful manufacture in a particular bottle. 



8. Mr. Lupton then complains that 1 do u not mention any 

 precaution taken to reduce the flask and weights to their true 

 weights in vacuo.''' The complaint is just, and my only 

 excuse for the omission must be that every text-book on 

 physics contains an account of the necessary " precaution." 



9. The next complaint is that I mentioned no precautions 

 taken to remove the air dissolved in the solutions, or that 

 retained between the liquid and the glass. Again Mr. Lupton 

 is right ; I mentioned none, and, what is more, I took none, 

 because I did not see how they could be taken without alter- 

 ing the composition of the solution in the one case, or altering 

 the capacity of the bottle by heating it to redness in the other. 

 I did mention, however, that I had satisfied myself that the 

 air contained in the water used was insufficient to affect its 

 density (p. 73), and a fortiori it would be insufficient to affect 

 the density of the solutions made with it ; while the numerous 

 determinations of the water-contents of the bottle, both after 

 it had been left dry in air for a long time, and also after it had 

 been in continuous use throughout the day, were sufficient to 

 assure me that no appreciable error was introduced by any 

 air-film adhering to the glass. 



Mr. Lupton seems to imagine that 25 cubic centim. of a 

 liquid in which '0007 gram of air is dissolved will weigh "0007 

 gram more per 25 cubic centim. Matters are not quite so 

 simple as this, for the volume, as well as the weight, is altered 

 by the gas ; and if the effect of the air is similar to that of, 

 say, hydrochloric acid, the weight of the 25 cubic centim. 

 would be increased by only '0001 gram. 



10. Mr. Lupton then attacks me for applying the term 

 differentiation to the method which I adopted in analysing 

 curves, on the ground that a true differential (as I pointed 

 out myself on p. 67) is of infinitesimal magnitude. As, how- 

 ever, I explained fully the nature of the process which I 

 adopted (for which there is no proper appellation), the name 

 given to it was a very secondary consideration. 



11. Mr. Lupton then says that "the smoothing process 

 applied to my first differential diagram ought, if accurately 

 performed, to get rid of those slight changes in first differences 

 which in the second differences mark changes of curvature, and 

 therefore changes of hydration/' But why " ought " ? Surely 

 only on the strength of a foregone conclusion that there are no 

 changes of curvature or of hydration. I must ask Mr. Lupton, 



Phil. Mag. S. 5. Vol. 32. No. 194. July 1891. H 



