1892.] Ionic Velocities. 289 



salt was made, and their sum compared with the value deduced from 

 the conductivity. 



^The first salt used was cobalt chloride, the alcoholic solution of 

 which is of a deep blue colour. The velocity of the chlorine ion was 

 measured by setting up the cobalt chloride with cobalt nitrate, the 

 colour of which is red, and that of the cobalt by a cobalt chloride and 

 calcium chloride pair, the latter salt being colourless. 



Some little difficulty was encountered in getting solutions of con- 

 venient strength. If very weak, the colours were not easily seen, 

 while, if the strength approached 0*1 gram- equivalent per litre, 

 irregularities of behaviour appeared. Finally, solutions of O05 gram- 

 equivalent were used, but even here the effects of too great concentra- 

 tion were still appreciable. 



Chlorine , . Vl = 0*000026 



Cobalt , . . 0-000022 



The sum is U = 0*000048 



This can be calculated from the conductivity (2'86 X 10~ 13 ), and 

 comes out 



TJ = 0-000060. 



Cobalt nitrate was then investigated. Its conductivity is greater 

 than that of the chloride (3 '80 x I0~ 13 ), which leads us to expect that 

 its behaviour will be normal at concentrations greater than those 

 at which the chloride becomes irregular. Such was found to be the 

 case. We should also suppose that its agreement with theory will 

 be closer. The pairs used were cobalt nitrate cobalt chloride and 

 cobalt nitrate calcium nitrate. 



Nitrate group (N"0 3 ) t>i = 0-000035 



Cobalt Vi = 0-000044 



The sum of these is 



U = 0-000079. 



The value calculated from the conductivity is 

 U = 0-000079. 



The explanation of the irregularities observed in strong solutions 

 is most easily found by the supposition that complex ions are 

 formed as the strength increases. The further discussion of this 

 point is postponed till some more experiments I am making are 

 ready for publication. 



A table of results is appended : 



