556 



Mr. A. Schuster. 



in Goldstein's original experiments), I have been able to supplement 

 his observations in several details. At present I only wish to state 

 that the edge of the dark space is, under these circumstances, far 

 from being an equipotential surface ; so we must look for some other 

 explanation. We can assure ourselves in another way that it is not 

 a certain minimum kinetic energy which determines the boundary of 

 the dark space. In my previous Bakerian Lecture I followed others 

 in the statement that an increase of current diminishes the thickness 

 of the dark space ; but this I find is not correct. If the dark space 

 is carefully watched while the current is diminished or increased by 

 altering the resistance of the circuit, it is seen to contract or expand 

 slightly, always being widest when the current is strongest. Such 

 an increase of current is accompanied by an increased fall of potential ; 

 the difference of the potential at which the dark space ends can 

 therefore be altered at will. 



I can at present only think of one way of accounting for the facts, 

 but wish for the present to express my views on this point with due 

 caution. 



From the magnetic experiments it appears that the velocities of 

 the molecules are reduced quickly in the luminous glow, but not at 

 any rate to the same exteut in the dark space. If that is the case, 

 there must be some change in the law of impacts, as we pass from the 

 dark space into the glow ; and the simplest supposition to make seems 

 to me to be that the strength of the electric field is an important 

 factor in the transformation of energy which takes place in the colli- 

 sions. We may imagine that if the electric field is sufficiently strong, 

 the ion will not lose much of its energy during impact, but that in a 

 weak field the velocities are reduced at a much quicker rate. It is 

 clear that if the molecular forces are strong, two molecules must 

 approach much more closely together before their mutual action 

 comes into play, and therefore what must be considered an impact 

 must happen more seldom in a strong than in a weak field. ISTear the 

 edge of the dark space the electric forces are found to diminish very 

 rapidly, and it is a question worth investigating whether the edge of 

 the dark space is a surface at which the electric force has some con- 

 stant critical value. I know of no facts which are against this view, 

 and my experiments have hitherto all been consistent with it. The 

 slio-ht widening of the dark space by increase of current would at 

 once be explained if my hypothesis is correct. In the above-men- 

 tioned case of two parallel kathodes, there is always a luminous layer 

 in the equidistant plane between them, even when the width of the 

 dark space due to one kathode alone extends sufficiently far to include 

 the other. In this plane the fall of potential is easily seen to be 

 small, and the shape which the dark space assumes seems to me to 

 agree very well with the supposition that there is a critical rate of 



