412 M. J. A. Groshans on the Specific Gravity of certain 



For d 20°/4° and d 20°/ 20° the densities (in Table G) have 

 been calculated three times from the first and fifth experiments. 

 Using the densities 20°/ 20°, we have 



v = 7-30208 logv=0'8634464, 



A.=12'5467 



Sum... 19-8487 « + £ = 1-0447. 



These figures are almost the same as those which were 

 obtained from Gerlach's two experiments. 



Table G shows that the interpolation-formula represents, 

 within a near approximation, the densities 20°/ 20° — that is to 

 say, the densities t°/t°; but not so well the densities 20°/ 4°, 

 or d t°/£°. 



It would seem as if the density t°/t° were the expression 

 most suited to the nature of solution ; the conversion of d t°/t° 

 into dt°/4° might resemble somewhat an arbitrary change, 

 altering the nature of the results. At the same time, the den- 

 sities t°)t° have the advantage of making it easier to compare 

 the results of different observers. This would still more be 

 possible if all investigators were to make use of the same fixed 

 temperature t° when the so-called ordinary temperature is in 

 question. 



We could also wish that it were always clearly stated if the 

 temperature in question be t°/t° or £°/4°; one is often in doubt 

 about this important point, especially when obliged to cite 

 from second-hand sources. But we must proceed. 



We have shown the method we have employed in Kohl- 

 rausch's experiments on the solution of KF, which are de- 

 scribed in the important work of G. Wiedemann on Electricity, 

 vol. i. p. 593. In all, we have chosen three experiments. 



Kohlrausch's experiments on a solution of KF. 



p. aq. A. c/ 18°/ 18°. tfl8°/4°? 



5 19 Gig 1-0424 1-041 



10 9 29 1-0854 1-084 



40 1-5 4-833 1-3798 1-378 



We have commenced by changing the given densities (those 

 -with three decimals), which we suppose to be 18°/4°, into 

 18°/18°. We have then calculated the two constants, taking 

 the mean of the first and third experiments ; and have thus 

 obtained the following mean result: — 



v = 2-69133 log »/ = 0-4299663 



X = 2-25270 



Sum ...4-94403 « + £= 1-5343 



