548 
Proceedings of the Royal Society 
does not obviate the two difficulties already cited; it may also intro- 
duce a new one, in the form of what du JBois-Reymond calls internal 
( innere ) polarisation.* Nor is the second addition to our descrip- 
tion more destructive of the accuracy of our criticism than the first. 
Professor Beetz thinks that, by measuring the same liquid in tubes 
of different lengths, one may take the difference of their resist- 
ances as the resistance of a column of liquid whose length is the 
difference of their lengths. But this can only be the case on the 
supposition that the state of matters at the junction of the liquids 
is the same for both determinations. Now diffusion does not cease 
at the end of each measurement and wait until the next begins. 
Nature is not so convenient. Every moment adds to the mixture 
of the solutions and changes their resistance. If, moreover, there 
be polarisation at the surface of contact, or (clay vessels being 
used) if there be also internal polarisation, it must begin at 
zero and increase from the first moment of contact up to the time 
of observation (supposing that to occur before the maximum is 
reached). In order that this condition may be the same for both 
measurements, the observations must be made after the same lapse 
of time from the first moment of contact, an occurrence which is 
manifestly improbable, and, if it should happen, impossible to 
know. It is true that a judicious choice of electrolytes may 
remove one or more, though never all, of these sources of error ; 
but such a possibility cannot be taken into consideration in dis- 
cussing Paalzow’s as a general method; while, at its best, as a 
special method, it has always the defect arising from the mixture 
of the liquids. How great or how small the error arising from 
mixture and polarisation may be, it is difficult to say. That can 
only be decided by future experiments. But it is clear that the 
error remains, and that the method, as described most minutely, is 
subject to the same criticism as in its simpler form. 
Passing from Paalzow, Professor Beetz proceeds to prove that 
we did not understand Kohlrausch and Nippoldt’s f work any better 
* Monatsberichte der Berliner Akademie, Aug. 4, 1856, p. 15, and Jan. 31, 
1859, p. 1. 
t Gottinger Nachrichten, Nov. 18, 1868, p. 415. Jahresbericht des phys. 
Vereins zu Frankf. 1867-68, p. 71. Pogg. Ann. cxxxviii. 1869, pp. 280 and 
370. 
