ON ELECTRICAL DISCHARGES THROUGH RAREFIED GASES. 
173 
intervals between the luminous discharges, during which the tube is dark/" If the 
ordinary non-sensitive discharge be observed in a similar way there are no such dark 
bands, and no available speed of the mirror suffices to show any break in the 
uniformity of the luminous image of the slit. The occurrence of these dark bands 
shows conclusively that the discharge in which they appear is intermittent and 
discontinuous, and it may be stated as a general proposition that a revolving mirror 
has thus invariably shown intermittence in the discharge whenever a sensitive column 
has been examined by it through a slit of proper dimensions, whatever may have been 
the particular means by which the sensitiveness has been produced. 
We have said that the telephone, as well as the revolving mirror, may be made 
to show that there is intermittence in the discharge in vacuum tubes so long as there 
is sensitiveness. The evidence it gives to us is as follows : When a telephone is 
in circuit with a non-interrupted, non-sensitive discharge, a rushing sound is frequently 
although not always heard. In a very large number of cases there is absolute silence. 
But as soon as the discharge becomes sensitive the silence is broken by a shrill sound, 
or if the rushing sound of which we have spoken previously existed, there is a sudden 
change in the character of the sound, which usually becomes musical. The pitch of 
the note is always high, and naturally varies with the circumstances of the discharge. 
When the air-spark is again abolished the note ceases, or gives way to the rushing 
sound mentioned above.! Occasionally, in the sensitive discharge, the rushing sound 
and the musical note are heard simultaneously. This may arise from rapidly succeeding 
periods of interruptedness and non-interruptedness; but it is not a phenomenon of any 
particular importance at the present stage of our inquiry4 
It will have been noticed that at the outset we defined the intermittence capable of 
causing sensitiveness to be such as would permit only so small a quantity of electricity 
to pass at each individual pulsation as not to interfere with the instantaneous character 
of the discharge. This is necessary in order to exclude such cases of intermittence or 
fluctuation of current as are observable within a single-coil discharge when a compara- 
* This experiment is more fully described in § 10. 
f It may fairly be objected that this test shows only the intermittence of the current in the circuit 
external to the tube, and does not directly tell us anything of the discharge through the tube. We shall, 
however, show later on how the telephone may be made to give us more direct evidence of the inter¬ 
mittence of the discharge through the tube. But to render clear the nature of this evidence would 
require that the theory of the sensitive discharge should be more fully developed than is possible at this 
early stage. 
J When the air interval is increased beyond striking distance, a silent discharge takes place across the 
interval; and the discharge through the tube is the same in character as that produced without an air- 
interval ; it is also non-sensitive. And instead of the shrill whistling which is heard with an air-spark 
proper, nothing is in this case audible but a soft rushing sound. A similar result is obtained when an arc 
is formed across the air-spark interval, which is found to be the case under certain circumstances which 
will be referred to later on. 
