of view of the Kinetic Theory of Matter. 31 



The internal diameter of the tube is 3*5 cm and the two spheri- 

 cal electrodes are 7'5 cm apart. 



At comparatively high pressures, and with the ordinary con- 

 tact maker in the primary, two discharge streams can be dis- 

 tinguished, of almost equal intensity. But if contact be made 

 and broken very slowly, it is seen that these two streams are 

 not simultaneous phenomena, one being produced by the po- 

 tential-difference at make, and the other at break of the pri- 

 mary circuit. By decreasing the pressure in the tube continu- 

 ously, one of these streams is seen gradually to grow smaller, 

 and finally entirely to vanish. 



On further decreasing the pressure, until the discharge 

 stream from the cathode covers but a small portion of the 

 total area of the cathode a secondary stream from the anode is 

 found to exist, whether we allow the coil to work continu- 

 ously, or examine the discharge at each make and break of the 

 primary. At make, no discharge stream is seen from either 

 electrode, but at break, the two streams (one from the 

 cathode, the other from the anode) can easily be separated by 

 a magnet, as shown in fig. 5. In some of the early tubes made 

 for the study of these secondary streams, the spherical anode 

 was turned through a small angle, so that the discharge stream 

 from it ordinarily fell on the glass wall of the tube. In this 

 way the two discharge streams can easily be seen, without it 

 being necessary to separate them by a magnet. These experi- 

 ments have also been performed with the rectified alternating 

 potential-difference, described in £i; the same evidence of 

 secondary streams as just noted has, in every case, been ob- 

 served. This experiment will be further discussed iu part II 

 of the present paper. 



Fig. 6 illustrates a tube that shows these secondary discharge 

 streams under conditions somewhat different from those in the 

 tube just described. The tube has two spherical aluminum 

 electrodes each of 2'5 cm radius of curvature ; the electrode in 

 the main tube being 2 cm , that in the branch tube l-7 cm in diam- 

 eter. The diameter of the main tube is 2*5 cm , that of the 

 branch tube at right angles, l*8 cm . The distance between the 

 electrode in the main tube and the opposite glass wall is about 



ir. 



At a pressure considerably lower than that at which a dis- 

 charge stream is sent through the tube on the make of the 

 primary, there is seen to spring from the spherical formed end 

 of the glass tube lying opposite the cathode, a secondary 

 stream, (&,) in fig. 6. This secondary stream could be deflected 

 by a magnet, caused fluorescence of the glass walls where it 

 struck, and in all other respects acted as a discharge stream 

 would have acted, coming from a cathode located at the end of 

 the glass tube. 



