ON CHEMICAL COMBINATION AND ELECTKIC DISCHARGE. 485 



mucli smaller than that which it withstood previously without giving way, 

 is sufficient to send the discharge through it. In fact, when once the 

 discharge has been started through the dry gas it is almost as easy to send 

 the consecutive discharges through the dry gas as through the damp. If 

 the dry bulb is put away and left alone, it will, after some time, recover 

 its original strength. It takes, however, a considerable time to do this, 

 and it is very often several hours before the bulb is restored to its original 

 condition. The effect of depriving the gas of aqueous vapour is thus to 

 inci-ease the difficulty of getting the first spark to pass through the gas. 

 After once the spark has passed, there is very much less difference between 

 the dry and the wet gas. If, instead of these alternating currents, we use 

 the dii'ect current from a large battery of cells, we can easily measure the 

 change in the potential difference required for the first and consecutive 

 sparks ; all we have to do is to connect the electrodes of a discharge tube 

 with an electrostatic-voltmeter. For damp gases the change in the poten- 

 tial difference between the first and following sparks is not very great. For 

 very dry gases, however, the potential difference required to start the 

 first spark may be quite indefinitely large ; but for the second and follow- 

 ing sparks it will be almost identical with that for the damp gas. 



Another striking effect produced by aqueous vapour is the change it 

 produces in a phosphorescent glow which some gases emit after an electric 

 discharge has passed through them. When the discharge passes through 

 oxygen it is followed by a greenish yellow glow diffused through the bulb. 

 This glow is \'ery bright and lasts for a considerable time. It gives a con- 

 tinuous spectrum crossed by a few bright lines. When the discharge 

 passes through cyanogen it is succeeded by a white glow which is very 

 persistent, lasting in some cases for fifteen or twenty minutes. All tLr 

 gases in which I have observed this glow have the power of forming poly- 

 meric modifications. Aqueous vapour produces a very marked effect on 

 this glow. If an oxygen bulb is sealed up with some phosphorus pent- 

 oxide, then for a short time after sealing off the gas will glow brightly 

 after a discharge has passed, but as the gas gets drier the glow gets fainter, 

 and after a few days is hardly visible. On the other hand, air, which only 

 shows a very faint glow when moist, gives quite a bright glow when care- 

 fully dried. This glow is of the same character as that in damp oxygen. 



The way in which the presence of water vapour facilitates the discharge 

 and affects its appearance may be compared with its effect on chemical 

 combination. Dixon, Pringsheim, and Baker have shown that certain 

 typical combinations do not take place at all unless water vapour is pre- 

 sent. If we take the view — which is, I think, proved by the phenomena 

 accompanying electric discharges through gases — that the first discharge 

 through a gas in its normal state is accompanied by the splitting up of 

 some of the molecules of the gas, we can see why a cause which increases 

 the facility with which the first discharge passes through the gas should also 

 increase the tendency of the gas to enter into chemical combination. The 

 forces holding together the atoms in a molecule are so great that if we 

 were to take a single molecule of a gas by itself the electric field required 

 to pull the atoms in the molecule asunder would far exceed in intensity any 

 hitherto applied to a gas. For consider two atoms in a molecule each 

 charged with the quantity of electricity which the electrolysis of liquids 

 shows is carried by each atom of an electrolyte, and which from the results 

 of experiments of the electrolysis of steam we may infer is carried by an 

 atom of a gas. This charge, which we shall call e, is of the order 10~" in 



