40G Prof. Thomson, On some experiments on the [May 10, 



electrode. At very low pressures there are luminous rays which 

 start from the negative electrode in a direction nearly at right 

 angles to it, and travel on in this direction until they meet the 

 glass tube which fluoresces brilliantly at the places where the 

 rays strike the glass. 



Let us now consider from our point of view what happens in 

 the tube. The gas near the negative electrode gets decomposed, 

 and the explosion which occurs when the decomposition takes 

 place drives the dissociated atoms about, these after going for 

 some distance recombine, giving out as is usual when gases 

 combine light and heat. The dark space corresponds to the 

 region when the dissociated atoms are moving about, the lumi- 

 nous glow to the space within which they are recombining. The 

 distance through which the molecules must travel before they 

 recombine must evidently depend upon the density of the gas, 

 for the smaller the density the fewer atoms will a given atom 

 meet with in a given distance, and so the fewer chances it has of 

 entering again into combination. Again, the number of collisions 

 which on an average an atom must make before it recombines will 

 depend on the nature of the gas — if the atoms of the gas have 

 a strong affinity for each other fewer collisions will be needed 

 than if the affinity is weak, so that the dimensions of the dark 

 space will be a smaller multiple of the free path in the first case 

 than in the second. Crookes found that the dimensions of the 

 dark space in carbonic acid gas was a smaller multiple of the free 

 path than that in hydrogen. We must however remember that 

 the free path we are considering is not the free path of the 

 molecules but that of the dissociated atoms, and about this we 

 have no direct evidence ; again, the disturbance produced by the 

 discharge will no doubt influence the free paths, so that we cannot 

 expect to be able to calculate with any accuracy the dimensions of 

 the dark space from the length of the free path. 



The electromotive intensity outside the negative glow is much 

 smaller than that inside, and the molecules are not so easily 

 decomposed as they are near the negative electrode, so that in 

 a region outside the negative glow we have no decomposition of 

 the molecules at all; this region would correspond to the dark 

 interval. 



Let us now consider what takes place at the positive electrode: 

 the molecules there are according to our view not so easily de- 

 composed by the electric field as those near the negative electrode, 

 and it may be that, as in our experiments with the parallel plates, 

 we can arrange matters so that practically all the discharge takes 

 place at the negative electrode ; in general however the conditions 

 will not be such that the phenomena at the two electrodes are 

 contrasted so much, there will be more molecules decomposed at 



