G04 
MESSRS. W. SPOTTISWOODE AND J. FLETCHER MOULTON 
this will be in a positive unipolar discharge and it will give relief-phosphorescence and 
all the other phenomena of positive intermittence. 
The phenomenon of relief-phosphorescence supplies us with a very convenient method 
of determining experimentally the laws that govern the production of these molecular 
streams and the direction in which they travel. By varying the shape and size of 
the patch of tinfoil on the tube we vary the negative terminal from which the streams 
proceed, and by varying the length of the air-spark we vary the violence of the dis¬ 
charge. The opposite side of the tube acts as a very conveniently placed screen on 
which the effects are shown, so that we are able with a minimum of labour to examine 
the various phenomena to which these molecular streams give rise. We shall now 
give the results of this examination so far as we have had time to carry it. 
A very cursory examination suffices to show that it is impossible to accept the 
hypothesis that presents itself to the mind most naturally, viz. : that these molecular 
streams move in straight lines, or nearly so, starting in a direction normal to the 
surface of the terminal. Were this the case, then the shape of the phosphorescent 
patch due to a piece of tinfoil on a tube when employed to produce relief-phospho¬ 
rescence would be of a shape similar to that of the tinfoil. Experiment, however, 
shows it to be quite otherwise, and indeed it is at first sight extremely difficult to 
arrive at precise conclusions as to the law of the distortion observable in this patch, 
which we may term the phosphorescent image. There are certainly two causes at 
work in producing this distortion which are of great importance, but it is probable 
that at least one other cause is present, the existence of which, however, we cannot 
demonstrate at tins stage. 
The first cause which is undoubtedly at work to prevent the molecular streams 
forming an exact image of the tinfoil patch is that they do not all leave normally. It 
has been assumed rather than proved that such is their natural tendency, but we very 
much doubt whether even with the continuous discharge this is the case except to the 
extent that it is approximately true for those parts of the terminal that are not very 
near its edge. And it is certain that when we come to the intermittent discharge the 
direction in which the molecular streams which cause relief-phosphorescence can be 
made to go is inclined at such an angle to the normal that they must possess consider¬ 
able initial obliquity. A ring of tinfoil connected to earth was placed round a tube in 
which a current with positive air-spark was passing, and which gave the usual phos¬ 
phorescent phenomena. A small piece of glass lying loose within the tube was then 
shaken to such a position in the tube that the line from the piece of glass to the edge 
of the tinfoil opposite to the place where the glass lay made an angle of about half a 
right angle with the axis of the tube. A fine sharp shadow of the piece of glass was 
seen, and its direction was (so far as could be judged by the eye) just what it would 
have been had it been made by streams proceeding directly to the piece of glass from 
the tinfoil on the opposite side of the tube. Now had the streams started in a direction 
more nearly normal to the surface of the tinfoil, and subsequently become deviated to 
