78 Proceedings of the Royal Irish Academy. 



guard-ring, and connected to an electrometer. The gauze is charged to 

 potentials + V and - V alternately, the alternating field being therefore of 

 what is sometimes called " square " wave-form. This is eifected by a rocking 

 device driven by clockwork, which, by means of a needle dipping into a 

 mercury cup, makes and breaks a contact. While the contact is broken, the 

 lower battery in the diagram has its positive pole earthed through the 

 high non-inductive resistance It. The negative pole being connected to 

 the metal gauze, the latter is therefore at the potential - V. While the 

 contact is made, the point P is at potential -r 2 F, and therefore the plate is 

 at potential -f V. The times of contact can be determined by means of the 

 Hipp chronoscope (C) to the yoVo °f ^ second. 



If the periodic time is varied, and if the charge received by the electro- 

 meter plate for one alternation of the field is considered, it is clear that an 

 ion of mobility v. will contribute nothing until a periodic time is reached 



such that ut = ^, where cl is the distance between the plates, and t is half 



the periodic time, the positive and negative fields being supposed to be of 

 equal duration. If the quantity of electricity received per alternation be 

 plotted against the voltage, the curve will show upward bends at points 

 corresponding to the mobilities of the different classes of ions. The product 

 of the current recorded by the electrometer and t is proportional to the 

 quantity of electricity per alternation, and it is this product that has been 

 plotted. In these experiments V was 10 volts, d was 3'05 cms., and the 

 values for periodic time varied from ••! to -8 seconds. The electrometer there- 

 fore did not so much record a steady current as a series of impulses. The 

 damping of the needle was very high and the motion fairly smooth. 



The only source of trouble was the rocking device used as contact maker 

 and breaker. It was easy to adjust this to give for any one periodic time 

 equal intervals of make and break, that is, equal durations of + F and - F, 

 or, as was generally done, to make the times slightly unequal so as to secure 

 that ions which did not reach the electrometer plate on the downward 

 journey were discharged at the end of their upward joiuTiey. For any one 

 adjustment the contact maker would give the times quite steadily in a 

 definite ratio ; but if the periodic time was altered, the ratio of the time of 

 make and break altered slightly also. This was not of any direct consequence, 

 as the time of the downward field (the field of the same sign as the ions 

 under examination), which alone enters into the calculation of mobility, was 

 in each case acciu-ately measured. But the indirect consequences were, as 

 will appear later, rather disturbing. 



