Geology and Mineralogy. 419 
9. Double Refraction of Electric Waves.—Prtser Lepepew has 
succeeded in reducing the size of Hertz’s apparatus for exhibit- 
ing the reflection-polarization and refraction of electric waves, so 
that he could work with waves only a fraction of a centimeter in 
length (A=0°6™) and with prisms scarcely more than a centi- 
meter in height. The arrangement of the mirrors and the prisms 
is the same as that of the observing telescopes and prism in 
the case of the optical spectrometer. A prism of sulphur 1°8°™ 
high, 13°" broad with a refracting angle of 25° was used. The 
values of the indices of refraction obtained were N.=2:2 N.=2:0 
for the two directions in which Boltzman had measured the 
dielectric constants of sulphur. The square roots of the dielectric 
constants are, 2°18 and 1°95, and Maxwell’s law therefore holds 
within the limits of error of observation. Instead of a spark 
micrometer, Lebedew employed a thermal junction of iron and 
constantan. His paper contains illustrations and descriptions of 
his apparatus.—Ann. der Physik und Chemie, No. 9, 1895, pp. 
lel Peso 
Il. GroLtocy AND MINERALOGY. 
1. On the Temperature Variation of the Thermal Conduc- 
tivity of Rocks; by Lorp Ketvin and J. R. Erskine Murray. 
(1.) The experiments described in this communication were 
undertaken for the purpose of finding temperature variation of 
thermal conductivity of some of the more important rocks of the 
earth’s crust. 
(2.) The method which we adopted was to measure, by aid of 
thermoelectric junctions, the temperatures at different points of a 
flux line in a solid, kept unequally heated by sources (positive 
and negative) applied to its surface, and maintained uniform for 
a sufficiently long time to cause the temperature to be as nearly _ 
constant at every point as we could arrange for. The shapes of 
the solid and the thermal sources were arranged to cause the flux 
lines to be, as nearly as possible, parallel straight lines; so that, 
according to Fourier’s elementary theory and definition of thermal 
conductivity, we should have 
k(M, B)_[v(M)—»(T)] + MT 
kK(T, M) [v(B)— »(M)] + BM’ 
where T, M, B denote three points in a stream line (respectively 
next to the top, at the middle, and next to the bottom in the 
slabs and columns which we used); »(T), v(M), v(B) denote the 
steady temperatures at these points; and 4(T, M), 4(M, B), the 
mean conductivities between TIT’ and M, and between M and B 
respectively. 
(3.) The rock experimented on in each case consisted of two 
equal and similar rectangular pieces, pressed with similar faces 
together. In one of these faces three straight parallel groves are 
cut, just deep enough to allow the thermoelectric wires and junc- 
