154 Messrs. Holborn and Austin on Cathode Disintegration 
In this way we obtain Table II., in which 7 represents 
the current-strength in scale-divisions, V the difference of 
potential between the cathode and the neighbouring tube- 
wall, and y the loss in weight of the cathode for a current of 
20 scale-divisions 2°46 x 10-° amp. flowing for 30 minutes. 
The observed values of y are compared with values calculated 
on the assumption that the loss of weight of different metals 
is proportional to their chemical equivalents. With the 
exception of a few large deviations where the observed values 
are too small, the differences (observed—calculated) fall 
within the limits of the errors of observation. 
The values for y lie in a straight line which cuts the axis 
of abscissas at the pomt V=350. The value 0-00060, which 
when multiplied by the chemical equivalent gives the tri- 
gonometric tangent of the angle between the axis of abscissas 
and the given line, would correspond to 0°00146 for a current- 
strength of 0°6x10-? amp. and 0:01 mg. as the unit of 
weight. From these observations it is impossible to discover 
any direct influence of gas-pressure on the disintegration. 
4, Observations in Tube B. 
The fall of potential at the cathode depends not only on 
the gas-pressure and the current-density, but also on the 
situation of the cathode. A cathode surrounded by a narrow 
tube shows in general a higher cathode-fall for the same 
pressure and current-density than one in a wider tube. We 
have attempted to discover how these conditions affect the 
disintegration by observations on platinum, gold, and silver 
in tube B (fig. 1). 
Table III. contains the results for platinum, silver, and 
gold. In the case of platinum the cathodes B and C were of 
the form already described, while D was of similar form but 
of twice the diameter. This difference in size made no 
appreciable difference in the disintegration. As in the former 
experiments, the current-strength was 0°6 x 10-° amp., with 
the exception of the two cases marked with an asterisk, where 
the observed loss of weight was doubled since the current- 
strength was only 0°3x10-? amp. The table shows that for 
platinum the loss of weight is the same as in tube A. 
Gold and silver behave differently. The loss of weight, 
indeed, would be represented graphically by straight lines, 
which when extrapolated cut the axis of abscissas at the point 
V=610 and rise much more steeply than in the former cases. 
For the same potential, the equivalent losses of weight stand 
in the relation 300: 267=1°12, which is the same ratio 
as exists between the two groups of silver observations. 
