ON ELECTROLYSIS IN ITS PHYSICAL AND CHEMICAL BEARINGS. 385 



cases ; on the other hand, he had six times as many alternations of currents per 

 second, so the self-induction of his rheostat must be greater than that of mine (§ 3). 

 On quite other grounds, Kohlrausch's method is preferable. I may also say that 

 Kohlrausch used a galvanometer for measuring higher resistances. 



•' 2. Secondly, concerning your remark on subtracting the conductivity of the 

 water in order to find that of the dissolved salt. This mode of calculation I use, 

 not only because it corresponds with later established " laws," but also on true d 

 posterioi-i grounds (which indeed the reader of your critique cannot guess ; cf. Pt. I. 

 § 9). It would, however, to my mind, be little logical not to show that the rule 

 corresponds also with the general laws, according to my apprehension of them. So 

 I cannot understand your objection to this paragraph. The law of divided circuit 

 is, so far as I know, really true when two electrolytes are mixed. According to 

 my hypothesis, the conductivity of a solution is equal to the sum of the number of 

 ions, each divided by the friction it experiences against the fluid. This view leads 

 necessarily to the law in question, and it is also a necessary consequence of the 

 Williamson-Clausius hypothesis. I regret I cannot enter fully into this here. I 

 have, moreover, applied this rule throughout my memoir on conductivity of mixed 

 solution with, so far as I see, complete success. You say quit« simply, " which is 

 untrue." I wish you would tell me of any weU-eatablished fact against it. I am of 

 opinion that there exists no basis for the opinion that the law of divided circuits 

 does not apply to dilute solutions.' 



' 3. Thii'dly, " that the conductivity of water itself remains unaffected by the 

 presence of a foreign body ; which is improbable." 



' It appears as if you meant that the conductivity which Kohlrausch and I have 

 proved to exist in our distilled water is really the conductivity of water. That my 

 experiments contradict this is indubitable, and Kohlrausch also attributes it to 

 traces of saline impurity. Reaction-velocities give a pretty safe judgment on this 

 subject. I have thus pretty clearly shown that there are salts which cause the 

 greatest part of this conductivity (Part I. § 11). The probable effect of the 

 foreign body is to take much of the water from the said salts (cf. my memoir on 

 conductivity of mixtures). The salts, however, change their molecular conductivity 

 very little with great dilution ; and so the apparent conductivity of water does 

 remain very nearly unaffected by the presence of a foreign body in very dilute solu- 

 tion. I could say much more on this if time permitted. 



' 4. Fourth, " The first approximation to the relation between k and m is that 

 made by Kohlrausch, viz., that the two are proportional. This is roughly true for 

 very dilute salt-solutions." 



* For example, copper acetate (see Part I. p. 40) has the following : — 



1111 



276' 200' 1254' 6586' 



-=320, 512, 671, 740, times 10~'. 

 m 



' Again, copper sulphate has, according to Kohlrausch (p. 196 of his last memoir), 

 the following : — 



m= -5 l -05 -01 -001 -0001 -00001; 



- = 288 424 479 675 950 1062 1086. 

 m 



' You can find still better examples among mercury salts ! 



* These numbers, according to Kohlrausch's formula, should be the same ; so 

 the correspondence between formula and fact seems to me not " roughly true," but 

 altogether futile. (The dilution-exponent is in every case much more constant ; 

 though I build nothing upon the fact.) 



' That for extremest dilutions, under 15550, the molecular conductivity is fairly 

 constant, although even then probably not for mercury salts, I also have shown ; 

 since in these cases the dilution exponent rises to 2. 



' See footnote on pp. 358, 359. If Dr. Arrhenius did not mean that the law was 

 deductively proved by him (in § 15) for all cases, my criticism does not apply ; ex- 

 cept, indeed, to his mode of statement. O. L. 



1886. c 



