INTENSITY RELATIONS IN THE SPECTRUM Of HELIUM. 143 



to the focal lengths of the lenses. With this arrangement, an exposure was made 

 for a definite time on the plate. 



A perforated metal plate was then introduced between the two lenses, and another 

 exposure for the same period of time was made on an adjacent portion of the same 

 plate. This perforated plate consisted of a thin sheet of metal drilled with small 

 holes at regular intervals of about a millimetre. By taking the mean of a number of 

 micrometric determinations of the diameters of these holes and of the distances 

 between them, the effective " density " of the metal plate could be calculated. The 

 difference in the lengths of corresponding lines in the two spectra thus denotes the 

 density step due to the plate, which is equal to the " density " of the perforated 

 metal plate, from which the density step per millimetre length of wedge was 

 calculated. The values thus found by the use of all the stronger Helium lines were 

 plotted on squared paper against the wave-lengths, and a curve was drawn through 

 these points. This curve was quite regular, and of the same type as that shown in 

 the previous paper for another wedge, though obtained now by a different and in 

 some respects better method. Actual values of the density gradient may be found 

 in the tables given in later sections of this communication. 



(IV.) Experimental. 



In the present investigation we have examined the radiation in front of a flat 

 aluminium cathode about 1 inch in diameter, which fitted closely into a cylindrical 

 tube, as in fig. 1. The tubes were highly exhausted by means of a Gaede mercury 





Fig. 1. 



pump, and after continuous sparking, connection with the pump was cut off and 

 Helium was introduced by heating a quantity of powdered Thorianite contained in a 

 fused silica bulb, which was connected with the vacuum tubes through a tube 

 containing pieces of caustic potash and a U-tube containing charcoal cooled with 

 liquid air. After sparking for some time, the tubes were sealed off, and were found 

 to contain, in addition to the Helium, a certain amount of Hydrogen and also of 

 Mercury vapour. A great part of the latter disappeared on further sparking, and 

 finally the Mercury spectrum settled down to a constant intensity. The pressure in 

 the tubes was such that the thickness of the dark space was about 1 mm. With 

 electrodes of these dimensions, the tubes could be run with a moderate current for an 



