108 : Dr. E. Weintraub on the Are in 
value, and then remains stationary, or even slightly diminishes. 
To be perfectly correct we ought, therefore, to distinguish 
between three kinds of mereury vapour in the arc-stream : 
one ionized and conductive, the other non-conductive but 
hght emitting, and third, non-conductive and non-luminous 
ordinary mereury vapour. 
In the construction of the tube which contains the are care 
must be taken to avoid the presence of that third kind of 
mercury vapour as much as possible. For this purpose the 
tube is provided with a condensing-chamber, lying outside of 
the path of the are, in which most t of the superfluous mercury 
vapour condenses before it gets into the arc. A good form 
of a vertical tube is shown in fig. 2, where the cylindrical 
tube AC, placed above the auxiliary anode A, and parallel to 
the main tube BD, serves as a condensing-chamber. (See 
also fig. 3.) If such a condensing-chamber i is not provided, 
the potential drop across the lamp is much higher than under 
ordinary conditions; and if this voltage approaches too closely 
the impressed one the are goes out. 
In § 2 the influence of non-conducting mercury vapour in 
the are has been compared, as to its influence on the starting 
of the are, with the influence of any other non-conducting 
foreign gas. The analogy goes further, since the presence 
of non-conductive mercury vapour in the are has the same 
influence on the resistance and stability of the are as the 
presence of a foreign gas. Again, from a quantitative point 
of view, the action is much weaker in the case of mercury 
vapour. The presence of even small traces of air or hydrogen 
in the tube increases the resistance of the arc-path enor- 
mously, and renders the arc unstable. In view of the influence 
of foreign gases and of superfluous mereury vapour on the 
resistance of the mercury are, the approximate constancy of 
the potential drop across the lamp is strictly valid, and the 
numerical data given as to the voltage drop in the lamps, &c., 
are correct only when the vacuum is perfect, and care has 
been taken to provide the tube with a condensing-chamber. 
This being always presupposed, the conditions of stability 
of the mercury are are still rather complicated. First, as in 
any other arc, there must always be a certain amount of 
_ steadying resistance in series with the lamp. Second, again 
in common with all other ares, for any given impressed 
voltage there is a certain lower limit of current below which 
the are is not stable. With 100 volts applied to a tube con- 
suming about 80 this low limit is in the neighbourhood of 
3 amperes. If, however, 250 volts are impressed, the lamp 
will run steadily with a much lower current. The voltage 
