116 Dr. E. Weintraub on the Are in 
across the are is of the same magnitude as in the mercury 
are, the sum of polarization voltages at the anode and cathode 
being here about 8 volts, independent of what the material 
of the anode may be. In this there is a difference from the 
mercury arc, as between a mercury anode and graphite 
anode there existed a difference of 74 volts. This additional 
voltage is probably connected with the low boiling-point of 
mercury. The potential drop per inch arc length is nearly 
the same, about 2 volts, as in the mercury are, varying 
under the influence of the same conditions, as in the case of 
the arc through mercury vapours. 
The amount of light given off is very slight. The lamp 
has the appearance of a very dim, reddish-vellow flame, 
filling a glass tube. The difference in the efficiency as to 
light emission between a mercury are and an arc in alkali 
metal vapours is partly to be explained by the known fact 
that much more energy is stored up in red light than in 
green, for the production of the same physiological impres- 
sion ; partly by the fact that much less vapour is present in 
the arc than in the case of a mercury are. The condition of 
a potassium or sodium are surrounded by air is nearly the 
same as that of a mercury arc cooled by a water-jacket. I 
believe, however, from a series of experiments the descrip- 
tions of which would lead us too far, that these two causes. 
are not sufficient to explain the enormous difference in the 
efficiency of the alkali metal vapour arc and the mercury arc. 
In addition to the two causes mentioned, one must assume 
that different vapours have a different power for light emission 
when submitted to electrical excitation. 
The long are through the vapours of alkali metals that can 
be easily realized by following the description given above, 
will certainly prove useful for the spectral analysis of the 
light emitted by these vapours, as the arc takes place in 
vacuum in absence of any material that could introduce lines 
of its own into that spectrum. 
A large number of experiments have been carried out on: 
the behaviour of amalgams, beginning with very dilute ones 
and going up to such containing only a few per cent. of 
mercury. In case of dilute amalgams (up to about 10 per 
cent. potassium or sodium) the arc is at low current prac- 
tically that of mercury, while at higher currents the light of 
the alkali metal appears. Results of especial theoretical 
interest have not been obtained, and the very large number 
of experiments made may, therefore, be only very briefly 
summarized. With, for instance,a 10 per cent. potassium 
amalgam, at 24 amperes, the arc is that of mercury, while at 
