532 
Proceedings of the Poyed Society 
Spektralanalyse^ von Dr H. Kayser (Berlin, 1883), tliougli historical 
details are freely given, the name of Stewart does not occur even 
once! There are in the same work other instances of historical 
error nearly as grave. Thus the physical analogy, by which Stokes 
in 1852 first explained the basis of spectrum analysis, is given m 
Ur Kayser’s work ; but it is introduced by the very peculiar phrase 
wollen wir versuchen, eine meclianische Erkldrung der 
Erscheinungen zu geben, welche auf unsere Anschauungen fiber das 
Leuchten begrfindet ist . . . . , ; ” and the name of Stokes is not 
even mentioned in connection with it ! 
The second part of the paper deals with the question of the limits 
of accuracy of the reasoning which led Stewart, and those who have 
followed him, to results of such vast importance. Ur Kayser, 
indeed, announces his intention “ in aller Strenge niathematisch zu 
beweisen ” the equality of emissive and absorptive powers. But 
the mere fact that phosphorescent bodies, such as luminous paint, 
give out visible radiations while at ordinary temperatures, shows at 
once that there are grave exceptions even to the fundamental state- 
ment that the utmost radiation, both as to quantity and as to quality, 
at any one temperature, is that of a black body ; — and very simple 
considerations show that all the reasoning which has been applied 
to the subject is ultimately based on the Second Laio of Thermody- 
namics (or Carnot’s principle), and is therefore true only in the sense 
in which that law is true, ^.e, in the statistical sense. The assumed 
ultimate uniformity of temperature in an enclosure, which is 
practically the basis of every demonstration of the extended law of 
exchanges, is merely an expression for the average of irregularities 
which are in the majority of cases too regularly spread, and on a 
scale too minute, to be detected by our senses, even when these are 
aided by the most delicate instruments. The kinetic theory of 
gases here furnishes us with something much closer than a mere 
analogy. For the very essence of what appears to us uniform 
temperature in a gas is the regularity of distribution of the irregu- 
larities of speed of the various particles. And, just as in every 
mass of gas there are a few particles moving with speed far greater 
than that of mean square, so it is at least probable that a black body 
at ordinary temperatures emits (though, of course, excessively 
feebly) radiations of wave-lengths corresponding to those of visible 
