128 Prof. H. Edlund on the Behaviour of 
If we consider first the phenomena attending the charge 
and discharge of the tinfoil coatings, we observe that when 
the density of the air in the tube is relatively considerable, 
e.g. 86 millim. or more, no perceptible induction-currents 
result ; on the other hand, when the density amounts to less 
than 1 millim. down to 0:00036 millim., induction-currents 
are formed and manifest their presence by brilliant luminosity. 
The constancy of the induced electromotive force is altogether 
dependent upon the improvement which takes place in the 
conducting-power of the air in the same proportion as the 
rarefaction increases up to the limit attainable by means of 
the mercury air-pump. If,on the other hand, we observe the 
form assumed by the phenomenon when the current of the 
induction-coil is passed into the tube through the platinum 
electrodes, we find that the current easily leaps over the in- 
terval between the electrodes when the pressure amounts to 
one atmosphere or less down to 0:004 millim.—a point at which 
the resistance becomes large enough to sensibly weaken the 
current, and from which onward this resistance increases, if 
the exhaustion is continued; so that at a pressure of 0:00036 
millim. the current can only occasionally leap over the interval 
between the electrodes, although this is only 3 millim. The 
resistance to the current at a pressure of only 0:00036 millim. 
is therefore greater than at a higher pressure. The two 
series therefore give altogether opposite results. I can only 
explain this by the presence of a hindrance to the passage of 
electricity from the electrodes into the air, whilst, on the con- 
trary, the resistance peculiar to the air becomes less as its 
exhaustion increases. This experiment leads therefore to the 
same result as my former experiments, in which I came to the 
conclusion that there exists an opposing electromotive force 
at the surface bounding the electrodes and the gas, which 
increases continuously with the exhaustion of the gas. 
Second Haperument.—The same tube was used, but the tin- 
foil coatings were made considerably broader than for the 
first experiment. 
Pressure of air. 
millim. 
50. Visible sparks constantly between the electrodes a and 
6, but no induced light at all between the tinfoil 
coatings ¢ and d. 
3. Sparks betweenaandb. The induction-light between 
c and d begins to show. : 
0°05. The spark between a and b extends considerably 
towards the other end of the tube. Strong induction- 
light between ¢ and d. 
