1898.] on Some Neiv Studies in Cathode and Uonigen Eadiations. 589 



surface is from all ronud the edge of the latter, so that the atoms may 

 very possibly be all shot off again from the cathode in the form of a 

 hollow cone before they get further than a certain distance towards 

 the centre. Further, as the vacuum increases we know from our 

 experiments with our radiometer tube that the velocity of the positive 

 stream also increases very considerably, so that under the conditions 

 of a higher vacuum the atoms approaching the cathode have more 

 momentum and consequently get nearer to the centre before they 

 obtain a negative charge and are repelled in the cathode stream, thus 

 making the stream and the rings smaller in diameter. Of course, 

 once we start with a hollow convergent cone it is easy to understand 

 that the divergent cone will also be hollow, seeing that the atoms fiy 

 more or less rectilinearly crossing one another's paths at the focus. 

 How to explain the bright spots in the centres of the rings, which 

 ap (tears to indicate a central negative stream down the axis of the 

 hollow cones, is more difficult, but possibly the heterogeneous nature 



./Q~' 



Fig. S. 

 Apparatus for showiDg the cathude ray spectrum. 



of the cathode stream, due very j^robably to the varying veh^cities of 

 the negatively charged atoms, may bo sufficient to account for this. 



Crookes observed many years ago that cathode rays were deflected 

 by a magnet. Lenard was the first to show that the rays are not 

 homogeneous, but some are m( re easily deflected than others. Birke- 

 land went one step further than this, and showed that if a thin cathode 

 beam was deflected by a suitable magnetic field it was split up into 

 bundles of rays, and if allowed to fall upon the glass walls of the 

 tube, it gave fluorescent bands of alternate brightness and darkness, 

 which he termed the magnetic spectrum. 



Fig. 8 represents an ajiparatus for showing this effect. The 

 cathode rays proceeding from a flat aluminium disc are caused to 

 ])ass through a narrow slit in a piece of platinum which serves as 

 the anode. After passing through the slit, the rays impinge upon 

 the bulb, and if otherwise unaffected, produce a luirrow baud of 

 intense luminescence upon the glass. At each side of the bulb is 



