558 Proceedings of the Royal Society 
Paalzow’s agreement with Kolilrausoli as evidence for his own 
accuracy, and his <c agreement of results obtained by three observers 
in three quite different ways ” becomes the agreement of results 
obtained by two observers in the same way. We must now inquire 
what even this agreement amounts to. Kohlrausch and Nippoldt 
made two comparable observations. The first, however,* is ren- 
dered worthless by the fact that they assume the resistance of two 
vessels of the electrolyte, apparently without having made any 
accurate determination. The secondf agrees very well with Pro- 
fessor Beetz’ corresponding determination; but in order to conclude 
from this single agreement to general agreement, the unwarrant- 
able assumption must be made that the error found is not less than 
the average error.j: Kohlrausch and G-rotrian, whom Professor 
Beetz also cites, make two comparable observations, but both are 
questionable from the fact of their being unable to state accurately 
the constitution of the solutions whose resistance they measured. 
The authority with which Professor Beetz condemns our results 
as inaccurate on the ground of non-agreement with the agreeing 
results of various quite different methods may now be judged by 
the reader for himself. 
That the results of a new method applied by young experimenters 
should be even approximately accurate is, perhaps, hardly to be 
expected, and it will probably be found that our numbers are not 
quite exact. If Professor Beetz’ conclusion were well grounded 
they would need to be corrected only to the extent to which 
* Pogg. Ann. cxxxviii. p. 373. t Ibid. p. 376. 
t The “ perfectly satisfactory agreement ” between Kohlrausch and 
Nippoldt and Paalzow is based also upon comparison of a single pair of 
observations, the same unwarrantable supposition being made. With regard 
to this agreement it is interesting to notice the fact that Kohlrausch has 
lately corrected his first published numbers to the extent of 4 per cent. 
Paalzow’s observed conductivity instead of being a little more than 2 per 
cent, less than Kohlrausch’s, becomes a little less than 2 per cent, greater. 
If Paalzow were next to make the same discovery there would still be the 
same agreement, and Professor Beetz’ argument would be untouched. Even 
if such corrections should proceed alternately, ad infinitum, his argument 
would hold at all stages of the process as well as at the present! It would 
still be true that Professor Beetz agreed with Kohlrausch and Kohlrausch 
with Paalzow, and therefore Professor Beetz would be proved to be authori- 
tative. So long as Kohlrausch and Paalzow agree, — it matters not whether 
in accuracy or in error, — they nevertheless prove Professor Beetz accurate ! 
