266 
this was caused by two branches of the discharge, or by a change of direc- 
tion of the discharge during the exposure, but the clear interference bands 
ahout the stream indicate that the former is correct. Another is the 
peculiar deflection of the negative stream of No. 6 F. Many photographs 
were taken and many observations were made with the ground glass in an 
attempt to secure a duplicate of this, but with no success. 
In the previous work the negative stream was not deflected by a mag- 
netic field of 1,500 gausses, but in this the deflection was well shown where 
the stream is clear enough. The negative stream is in very few cases as 
long or as well defined as the positive. Also in the previous work, it was 
found that if the knobs of the electrostatic machine were placed sufficiently 
close together, a spurk passed between them, while between the points 
there was a violet stream, which was not shown on the plate or per- 
ceptibly deflected by the magnetic field. An attempt to deflect this stream 
with a stronger field was not successful. 
In repeating the work of Precht’. particular attention was given to his 
observation with the point cathode and the blunt wire anode, that the spark 
changed to brush and the porential rese when the magnet was excited. 
The writer found that this change occurred in a great majority of the 
observations made, but it was found to occur also in as great a per cent. 
of the obseryations, whether the discharge passed between the points, | 
point anode and blunt wire cathode, or point cathode and blunt wire anode, 
whatever the sense of the magnetism with reference to the current. In a 
few cases a brush would break into a rich spark, but all attempts to deter- 
mine the conditions which caused the changes were unsuccessful. In the 
previous report it was suggested that in case of the discharge between two 
points the change in type of discharge might be explained as a result of a 
change of the spark length, but after repeating the experiment it was 
concluded, as was suggested by Precht®, that although the length of spark 
path might be partly the cause, it was not the whole cause. No attempt 
was made to reproduce the exact condition of Precht’s experiment either 
in the form or size of the point, but no doubt if these had been fulfilled the 
atmospherie and other conditions would have entered which would have 
made the results variable because with no part of the apparatus altered 
in any way entirely opposite transformations were found to exist on 
different days. 
5 J. Precht, Wied. Annalen (66-4, pp. 676-697), 1898. 
