382 



Dr. W. W. J. Nicol on the Mutual 



Table VIII. (continued). 

 KN0 3 in Solutions of JSTaN0 3 . 



Mols. NaN0 3 . 



NaN0 3 . 



M.S. 



Mean M.S. 



KNO3. 



Mols. KNO,. ' 



1 







~:~ 



3111 



1 





1 







3111 



31-11 



31-11 



5-539 



1 



4-722 



35064 



I 











35 096 



1 35-080 



30-358 



5-405 



2 



9-444 



39560 













39-445 



39-502 



30-058 



5-352 



3 



14-167 



44-114 













44-229 



44-172 



30-005 



5-342 



6 



28-334 



59-614 













59-541 



59-578 



31-244 



5-563 



9 



42-501 



75-399 













75-063 



75-231 



32-730 



5-827 



12 



56-668 



91-092 













91-225 



, 91-159 



34-491 



6141 



15 



70-835 



107-050 













107-590 



j 107-320 



36-485 



6-496 



18 



85-002 



123-424 













123-410 



123-417 



1 



38-415 



6-840 



dissolved of the first salt. The last column gives the molecules 

 of the first salt present per 100 molecules of water. The 

 molecular weights used throughout are : — 



NaCl=58*5, KC1=74'6, NaN0 3 = 85, KNO 3 =101-l. 



With regard to the accuracy of these determinations one or 

 two words only need be said. In the 27 duplicate experi- 

 ments the sum of the differences between each pair amounts 

 to 3*057 parts of salt in the 100 parts of water, giving a 

 mean error of 0*113 ; but of the whole 27 only four much 

 exceed this mean error, leaving 23 experiments with a total 

 error of 1*359, and a mean error of 0*059, or +0*03 from the 

 mean. And it must not be forgotten that the results are 

 parts per 100 of water and not percentages, and the error 

 thus appears twice as great as it would have been had the 

 other mode of calculation been employed. 



The next point to be considered is the ratio of the two salts 

 in the saturated solutions formed when water is treated with 

 excess of both salts. This problem may now be attacked from 

 two sides. 



(1) That of direct experiment, and better even, since that 

 is beset with difficulties, (2) that of indirect experiment, for 

 it is evident that if the data derived from any pair of salts in 

 the duplicate experiments be plotted the curves thus obtained 

 will, if prolonged, intersect at the saturation-point of the salts 

 when each is present in excess. The saturation-point thus 

 obtained indirectly will depend not on one or two experiments 



