ON THE SENSITIVE STATE OF VACUUM DISCHARGES. 
569 
But it may be objected that this explanation seems to assume that the sum total 
of the discharges due to the interfering system is equal to the total of the continuous 
discharge, an assumption manifestly improbable, seeing that the interfering system is 
produced by a smaller machine, and, in addition, has an air-spark interposed in its 
circuit, the effect of which is to decrease the quantity, although it greatly increases 
the tension of the resulting discharge. The answer to this is, that although consider¬ 
able change in the length of the air-spark is possible without destroying the typical 
form of the positive effect, yet in extreme cases we do find indications of just such an 
imperfect action of the interfering system as would be expected from considerations such 
as these. The truncated positive column and the hollow cone separated from it by a 
dark space are seen faintly amid the ordinary positive luminosity of the continuous 
discharge in the tube, giving the appearance of a superposition of two discharges, the 
positive luminosity of the one being continuous, and that of the other interrupted in 
the way above described. But even in the cases where the action is perfect and gives 
to us the pure typical form of positive effect, it must not be supposed that theory 
requires an equality between the sum total of the inductive discharges in the tube, 
and the quantity in the continuous discharge. If the relieving system is sufficient to 
create the structure we have described, and to maintain it for a portion of the interval 
between two discharges, it may well be that during the remainder of the time the 
discharge, finding this structure ready to hand, might use it during the remainder of 
the interval, seeing that it is of precisely the type that vacuum discharges find most 
suitable for their propagation. For it must be remembered that the experience we 
have of stria spaces, and above all of the dark space round the negative terminal, 
shows us that the normal state of discharge under fixed conditions is an invisible 
discharge through a space of definite length (depending primarily on the degree of 
exhaust) lying longitudinally along the path of the discharge between a sharp bright 
luminosity on the side from which the positive comes, and a hazy luminosity on the 
side from which the negative comes. All these conditions are present in the structure 
we have described,* and it may well be that in spite of its being due to foreign 
influences the continuous discharge makes use of it in the same way as it would 
make use of a stria of its own creation. 
Be this as it may, there can hardly be any doubt of the interpretation of the 
observed appearances so far as the discharge is concerned. But, in corroboration of it, 
we may add the following facts gathered from an examination of the discharge in 
other parts of the tube. If the interference be not too violent, it will be found that 
the discharge in the portion of the tube lying between the tinfoil and the negative 
terminal of the tube is sensitive, and gives all the signs of positive intermittence, 
while that between the tinfoil aud the positive terminal is either wholly or almost 
non-sensitive. This shows that the discharge in the former portion of the tube is 
* We stall presently find that this view derives strong corroboration from the behaviour of the 
continuous discharge when subjected to interference by negative impulses. 
