130 
MR. B. T). STKEI.B ON THE MEASURE.MENT OF TONTC VEEOCITIES 
do those for potassium chloride. The explanation of this behaviour is rendered 
more difficult Ipy the fact that the transport number for magnesium sulphate has 
been determined by Hittorf’s method, using a gelatine partition to separate the 
anode and cathode portions of the solution. Two experiments made in this manner 
gave 0‘732 and 0744 for the anion transport number, numbers which agree well with 
the older determinations and with the figures in column S. It is worth pointing 
out that whilst p for these salts, when determined in water, shows a considerable 
change with concentration, in gelatine r)n the other hand p is approximately constant. 
Table IV. 
Group I. 
Group II. 
n = 0'5. 
/i = 1-0. , /( = 2-0. 
1 
n = 0-5. 
n = 1-0. 
n = 2-0. 
Li. 
0-716 
0-7.51 — 
Mg .... 
0-705 
0-722 
0-740 
Na. 
0-.597 
0-591 0-590 
Ca. 
0-681 
0-697 
0-715 
K. 
0-490 
0-488 0-489 
Sr. 
0-625 
0-665 
0-709 
Bu. 
0-576 
0-619 
0-633 
A jDoint that is shown with great clearness by a consideration of the figures in the 
fourth column is the periodicity of the transport number of the same anion with 
different cations belonging to the same group. Table IV. shows the periodicity in 
question for the chlorides of the elements of Grou})S I. and II. as far as these have 
been studied. Tables III. and IV. bring out very clearly the difference, referred to 
in the earlier part of this paper, between salts of the type of potassium chloride and 
those of the type of barium chloride and magnesium sulphate. For the former, p is 
practically independent of the concentration ; tor the latter, p changes with the 
concentration. 
A comparison of the figures in the fourth and sixth cohmms shows that foi 
potassium and sodium chlorides the values obtained by the two methods are both 
nearly constant, although not coincident with one another. For the other .salts a 
general agreement is seen to liold lietween the two sets ot figures ; the variation in p 
is also in the .same sense and of approximately the .same magnitude. 
In referring to the individual experiments it may he mentioned that for the 
measurement of potassium hydroxide only one gelatine partition was used. 
The experiment with silver nitrate is probably the lea.st accurate of all those 
tabulated, as the potential fall employed for the ani(in margin was very near the 
inferior limit for a good boundary, in order to ])revent melting of the jelly on entrance 
of the NOo ion. 
Mauv unsuccessful attem})ts have been made to measure solutions of sodium 
