90 
LUMMER. 
trum reaching from wave-length 0.4/* up to 0.8/*. (See 
Fig. 1.) The small progress made by the Nernst lamp, the 
osmium, and similar new incandescent lamps is, in my opin- 
ion, due entirely to the raising of the temperature. Indeed, 
the light intensity is increased in the proportion 
1 : 2 12 : 3 12 = 1 : 4000 : 600000 
when we raise the temperature in the ratio 1 :2 :3. Thus, 
our sun emits per unit of surface 600,000 times more light 
energy than the filament of an Edison lamp, and increasing 
the temperature of this lamp by only 100° doubles the bril- 
liancy of its filament. A single Edison lamp heated up to 
3000°, at which temperature the filament is destroyed, lights 
a large lecture room like an arc light of 2000 candle power. 
Only when we are able to excite directly the electrons 
within the atom (or, to use my analogy, the bees in a bee- 
hive or the bells in a belfry), that is, when we have to deal 
with luminescence , do we get a relatively great economy in 
light production, much greater than can be obtained by heat- 
ing a body like carbon or platinum or the black body. Radi- 
ating gases and vapors, as, for example, in Geissler tubes or in 
the mercury arc lamp of Arons (introduced into the trade by 
Hewitt), are, therefore, much more economical light radi- 
ators than any of the other practical light sources. These 
are still sources of heat more than of light, and from a 
theoretical standpoint are not very far in advance of the 
lamps of the ancients. 
When Kirchiioff, starting out from his law and his 
famous experiments on the reversal of spectrum lines, drew 
the conclusion that the sun’s central parts must have a 
very high temperature, he was right, but he was not 
right in drawing this conclusion from his hypotheses. In- 
deed only when the reversing body is a black body and when 
the radiation from the llame, the spectrum lines of which 
become inverted, is due to the temperature only, can we 
conclude that the reversing light source has a higher tem- 
perature than the flame. But this is not the case, since 
Pringsheim showed that the incandescence of gases, in so 
