608 
MESSRS. W. SPOTTISWOODE AND J. FLETCHER MOULTON 
as spreading out in directions outwards from the patch, as they would be if they were 
separated from the rest of the patch and formed a narrow slip of the same form as its 
edge. There is, however, the further complication (when the patch is of finite size) 
that all the normals at the various points of its surface pass through the axis of the 
tube, and would, if uninterfered with, form a reversed image on the opposite side of 
the tube. In so doing, they doubtless interfere with one another to a greater or less 
degree. It is not necessary, however, to examine carefully the results of this fresh 
element of complexity; speaking generally, the consequence is that the phosphor¬ 
escence assumes the shape of a central bright patch surrounded by a border of smaller 
luminosity. 
The second cause that operates to prevent the streams of molecules from making a 
perfect image of the patch of tinfoil which excites them is the property that such 
streams possess of interfering with one another during their passage through the gas. 
It is no doubt difficult to separate this cause from the last, for the streams only make 
themselves manifest at their extremities where they strike the glass, so that it is well- 
nigh impossible directly to distinguish between an initial obliquity and an obliquity that 
has been acquired during flight in consequence of the interference of other molecular 
streams. But there is abundant evidence of the independent existence of this latter 
cause of the obliquity of molecular streams. If we take a piece of tinfoil on a tube 
of not too large diameter and connect it to earth (the air-spark being, of course, in the 
positive) we shall, as we have said, produce a patch of green light on the opposite side 
of the tube. Let the outlines of this patch be carefully observed, and then let another 
patch be placed on the same section of the tube, but distant, say, a quadrant from the 
other. If this be also connected to earth, the former patch will be found to have 
altered in shape, and of course a fresh patch will have appeared corresponding to the 
second piece of tinfoil. If, now, the first piece of tinfoil be disconnected from earth, 
it will be found that the second patch has altered in shape. Thus the streams from 
the two loci of discharge must have interfered with each other, and as there was no 
community of origin this interference must have taken place during their passage 
through the gas.'" 
There is one form of this interference which is so marked, and which is so unmistak¬ 
ably a matter of interference, that it deserves special mention. We refer to the case 
in which the two pieces of tinfoil referred to in the last experiment are diametrically 
opposite, so that each is placed where the phosphorescent image of the other would 
naturally fall. In such a case the streams proceeding from each appear to beat back 
* The existence of this interference has also been demonstrated in the following way. A helix of tinfoil 
upon the tube connected to earth in the usual way was taken, the pitch being half a right angle. It 
gave the sharp phosphorescent helix of which we have already spoken. The tube was then touched at a 
point midway between two threads of the helix of tinfoil, so that the molecular streams from the finger 
cut normally through the surface formed by the molecular streams from the tinfoil. This was found to 
cause a distinct shifting of the corresponding pai’t of the phosphorescent helix. 
