172 Mr Richa?'dson ; On an attempt to detect radiation , etc. 
the diameter of the tube i and the lines of force were perpen- 
dicular to the excited wire (i. e. they were along the dotted line H 
of the figure). It was found necessary, to prevent the wire from 
glowing, to put one end (a) of it to earth ; this in no way 
interfered with the intensity of the discharge, as was shown by 
producing the spark gap at 0. 
With this apparatus experiments were made with copper, 
brass, soft iron and aluminium wires of about *7 mm. diameter 
and also with the steel wire 09 mm. diameter. The length of 
exposure varied up to three hours. The other conditions were 
changed as much as possible, e.g. the spark length was varied from 
1 to 4 cms. and exposures were taken with and without the 
magnetic field on. To increase the frequency of the discharges, 
and therefore the intensity of the skin currents, the Leyden jars 
were replaced by two small paraffin condensers of less than one 
centimetre capacity. 
The following measurements (for the case of the thin steel 
wire) indicate the maximum value of the current density attained 
at the surface. The capacity of the whole system including the 
wires was taken to be 10 cms. The self-induction was approxi- 
mately equal to 4 x 10 4 electromagnetic units. The period of the 
oscillations was therefore = 14 x 10 -8 seconds approximately. For 
the exponential factor e v ^ ' giving the diminution of the 
current density with the depth x from the surface, g was taken to 
be = 10 3 and a = 10 4 . If A is the current density at the surface 
that at depth x is .’. = He -6 * 103 *. The voltage for a spark length 
of 2 5 cms. is 75,000 or 250 electrostatic units of potential. The 
quantity of electricity set in motion at each discharge is therefore 
2500 electrostatic units giving an average total current through 
the wire of 2 x 10 10 units on the same system. Hence the current 
density at the surface of the wire during a discharge is given 
approximately by the relation 
ro 
2 7 tA X -0045 6 -6xio3* dx= 2 x 10 10 , 
J -0045 
i.e. 4 = 4 x 10 15 electrostatic units or 13 x 10 6 amperes per 
sq. cm. 
Since the plates were unaffected in the above experiments it 
was concluded that no photoelectric radiation was given off from 
the wires. It was still possible that radiation might be given off 
which had no action in a photographic plate, or whose action was 
feeble compared with its power of ionising the air through which 
it passed. Further experiments were therefore made to test 
whether the air in the neighbourhood of the wire possessed con- 
ducting properties. 
