For great values of C, the critical current density is therefore, 
independent of the value of the solubility product, i.e. the value 
is the same for all halogens, as D, is here about equal. 
It is different when C, has a small value, one that is comparable 
with VL. 
nee | i 
The critical current density is =0, when D, C, = D, ae Ord 
1 
and D, differing little, when C,=WL. For silver chloride, for 
which ZL = 10-10, the critical density will therefore be = 0 for 
C‚=105. Already at the smallest possible current density more 
AgCl will here be deposited in the liquid than on the anode. If on 
the other hand we work with an iodide, practically all the silver 
iodide will be deposited on the anode for C, = 102 as L4,; = 10-1 
and the critical current density is not = 0 until C, = 10. 
By the aid of the above considerations it is now possible to 
indicate in what way the electro-analytic determination of the halogens 
can take place most rationally, as will be set forth in the following paper. 
Chemical Laboratory of the University. 
Amsterdam, June 1916. 
Mathematics. — “Skew Frequency Curves.’ By M. J. van Uven. 
(Communicated by Prof. J. C. Kapreyn). 
(Communicated in the meeting of October 28, 1916). 
The skewness of a frequency-curve appertaining to some observed 
quantity « may be explained, as Prof. J. C. Kaprnyn’) has shown, 
without dropping the normal Gaussian law of error, namely by 
supposing that, instead of the observed quantity , a certain function 
of «: Z=F(«), is spread according to the normal law. 
Denoting the mean value of Z by M and the modulus of precision 
by A, the quantity 
z == h(Z—M)=— h F(a) - M= fiz) 
will be distributed round the mean value zero with the modulus of 
precision unity, so that the probability that z is found between z, and 
z, is represented by 
w?= 
«1 op 
1) J. CG. Kapreyn: Skew Frequency Curves in Biology and Statistics; Groningen, 
1903, Noordhoff. 
