164 Profs. Hagen and Rubens on some Relations 
proportion for all the metals, so that it is of no consequence 
at what part of this spectral region the comparison is made. 
Table II. contains the values of 100—R at A=4p, 8p, 
and 12 for all the pure metals examined and for five alloys, 
the electric conductivity of which was accurately known. 
The numbers given for (100—R) are always the average 
values obtained for the different modifications of the same 
material. The only results we have herefrom excluded are 
those obtained by aid of the mirrors produced by cathodic 
dissipation. Their reflecting-power was found to be a little 
smaller than that of the other mirrors. For silver, this 
difference is scarcely perceptible; for. gold it amounts to 
1 per cent., for platinum to 2-3 per cent. We cannot decide 
whether these differences are due to a deviating molecular 
structure of the dissipated layers, or whether they are caused 
by a very slight oxidation of the metals. 
In addition to the value (100—R), Table II.* contains the 
electric conductivity « f, its square root, and finally the pro- 
duct (100—R) . ,/« for X=4, 8, and 12y. It is evident 
that this product has approximately the same value with all 
the metals for X=12 uw; 7. e¢. the intensities entering the 
metals are—in the region of long waves—in inverse proportion 
to the square root of the electrical conductivity, 
(100—R),./«<=const. ... 7 ae 
But also for X=8 pw, and even for \=4 pw, this equation is ap- 
proximately verified. At %»=4 pw the products (100—R) . V« 
vary with an average deviation of 21 per cent. from the 
number Cy=19'4. At ~X=8y the average value of the pro- 
ducts is Cs=13°0; the average deviation is 14°5 per cent. 
At X¥=12 p, lastly, the mean value is Oj,=11:0, and the 
mean deviation is not more than 9°6 per cent. 
The values registered in Table II. for bismuth are bracketed, 
and not considered in the calculation of the average values of 
C. The first reason is, because they do not possess the éxact- 
ness of the other numbers. This is accounted for by the 
deficiency of the curvature and polish of our bismuth mirrors. 
The cutting and polishing of cast bismuth is very difficult, 
because of the crystalline structure of this material and the 
irregular consistency of its surface. The mirror of cast 
bismuth could therefore only serve for the investigation of 
the change of the reflecting-power with increasing wave- 
length ; the absolute values had to be determined with the 
aid of mirrors made by cathodic dissipation. Although these 
*-R is given in per cents. of the incident radiation. 
+ W. Jager and H. Diesselhorst, Wissenschaftliche Abhandlungen der 
Physikal. Techn. Reichanstalt, iii. p. 269 (1900). 

