IN AQUEOUS SOLUTIOX, AND THE EXISTENCE OF COMPLEX IONS. 
133 
The correctness of the assumption that tlie velocity of the l)onndary is in reality 
that of the preceding ion is capable of l)eing tested by the calculation of the 
conductivity of the intermediate solution from the observed velocity. For tins 
purpose it is necessary to first reduce the latter to that Avhich would be produced lyy 
a potential fall of 1 volt per centimetre. From the sum of the velocities thus 
obtained the conductivity is obtained by multiplication by e, where e is the (|uantity 
of electricity carried by 1 ion = 96,500 coulombs. Table VI. contains a comparison 
of the conductivities so calculated with those given by Koiilrausch. In the last 
four columns are given the values of the velocities, calculated from KohlRx\usch’.s 
tables, at the concentrations indicated in the second column, and those reduced from 
the author’s measurements. 
Since the sum of the velocities is pro 2 :)ortional to the conductivities, the sum of 
U and V must show the same agreement with the sum of U and V from 
Kohlrausch’s figures, as is shown in columns 3 and 4 ; that U and Y singly do not 
show such an agreement is due to the fiict that the velocity U = -- «, and hence if 
the Hittorfian p is used in the calculation, a difierent value for U will Ije obtained 
than if the author’s p is used. In columns 5 and 7 the p that is used is Hittorf's ; 
in columns 6 and 8 that obtained during the present research, and hence the want of 
agreement. The fact that l)y the use of the latter the figures in question would be 
brought into much greater concordance, speaks strongly in favour of their greater 
accuracy. 
The Existence of Complex Ions. 
With the three electrolytic measurements that have been considered, it has been 
found that it is only in the case of a few salts of the most simple type that 
experiment and theory are in agreement. In the case of all other salts, in the first 
place the transjjort numbei', Avhether measured by the older metliod of Hittorf or 
by the direct method, described in the present paper, is not independent of the 
concentration ; secondly, from the measurements of the conductivity it is not possilde 
to assign any specific ionic velocity to sucli ions as Mg, which is constant in different 
salts of the same cation; and finally, the current as measured l)y the galvanometer 
is not tlie same as that calculated from the observed A'elocity of tlie margin. 
For the change in transport number two explanations only seem possible, since in 
the majority of instances the infiuence of hydrolysis is })ractically negligible. 
The first of these is that with change in concentration there occurs a variation in 
the specific velocity occasioned l)y the electrolytic friction, and that the influence of 
the latter is greater for s(mie ions than for others, and usually ginater for tlie cation 
than for the anion. 
If we assume that at a given total concentration the specific mobilities of the ions are 
