474 RIESS ON THE INCANDESCENCE OF 
charge of a certain strength cannot be propagated continuously 
through a wire of certain definite dimensions ; it must be once or 
more than once intermitted. The points at which this inter- 
mittence occurs, and which are hypothetically pointed out by 
the bends, appear to be arbitrary, and only occasioned by casual 
inequalities in the substance of the wire. If for an experiment 
upon incandescence a wire is chosen which at some point has 
been squeezed or indented, in spite of the most careful straight- 
ening and smoothing, the first bend is sure to occur at that 
point. It is easy to produce a distinct bend at any part of the 
wire previously fixed upon. I have often indented a wire at a 
certain point with the nail of my finger, and always found that 
the first bend caused by the discharge took place at that point, 
whilst all the other phenomena in the connecting circuit re- 
mained unaltered. 
The Electric Spark. 
The present inquiry is connected with the question so re- 
peatedly raised concerning the nature of the electric spark, and 
in conclusion I will offer a few observations respecting it. If 
a sufficiently strong charge is passed through a good conductor 
and an imperfectly conducting liquid or gas, the discharge be- 
comes intermittent in the latter medium, and a luminous appear- 
ance is produced, an electrical spark, or some modification of it. 
Concerning the nature of this spark two essentially different 
opinions are entertained. According to the one, the medium 
itself (generally the air) through which the discharge passes be- 
comes luminous by a primary or secondary action of electricity. 
To this view an analogy is found in the above experiments, where 
it has been shown that a metallic wire becomes incandescent by 
an intermittent discharge, its substance therefore becoming lu- 
minous. The other opinion supposes the spark to consist, not 
of luminous particles of the medium itself, which it traverses, 
but of incandescent particles of the good conductors which border 
that medium, and which therefore generally consist of red-hot 
metallic particles. This view is also not contradicted by the 
above experiments. It is easily shown that the discontinuous 
discharge does not only take place in the imperfect conductor, 
but begins at a slight depth in the good conductor. If a spark 
be taken from a metallic conductor, the intermittent discharge is 
indicated by the pricking sensation in the finger; or if a piece 
