[WHEELER] THERMAL EXPANSION OF ROCK 23 
experiments on rocks made by Dr. F. D. Adams of the Department of 
Geology and Mineralogy. The method used was similar to that de- 
scribed below in connection with the present investigation. Two heatings 
were made. First a granite cylinder about 20 cm. in length was 
heated within a hollow cylinder of the nickel-steel. The expansion 
of both cylinders was measured with micrometer microscopes. The 
average coefficient of expansion, as obtained for the granite was 
0 .0000106, and for the nickel, 0 .0000099 per degree centigrade, for a tem- 
perature range of 17° to 782°C. A similar experiment was made with 
diabase in place of the granite, the temperature range being 18° to 793° C., 
the coefficient of expansion obtained for the diabase being 0.0000107, 
and for the nickel-steel, 0.0000102. Unfortunately a trial heating of 
the granite was made before any systematic readings were taken. This 
would affect to some extent, the results obtained for the granite. The 
permanent elongation for this second heating of the granite was given as 
0.272 mm.; that for the heating of the diabase as 0.29 mm. 
For the present investigation the apparatus and method used in 
the preliminary work just mentioned, were somewhat improved, and 
were found to be quite satisfactory. A cylinder of rock about 20 cms. 
long and 2.4 ems. in diameter was placed inside a specially constructed 
electric furnace with a nickel heating coil. The specimen was carefully 
protected from the air without by a thick packing of asbestos, in order 
that its temperature might be practically uniform throughout. The 
temperatures were got by the use of a platinum and platinum-rhodium 
thermo-electric couple, connected to a potentiometer bridge designed 
by Dr. Tory. The thermo-couple and bridge were first calibrated for 
water at 15.2° C.,for steam, for sulphur vapor and for the freezing points 
of aluminum and silver, and a curve was drawn through these points on 
a large scale, so that degrees of temperature which were plotted against 
the potentiometer bridge readings as well as the bridge readings them- 
selves, were represented by millimetre spaces on a large sheet of squared 
paper. Thus it was possible, having obtained the balance point on the 
bridge, to read the temperature directly from the curve with a good 
degree of accuracy. The temperatures (except in the case of the marble) 
were read at intervals of about 100° to 200° C. up to approximately 
1000° C. By means of two micrometer microscopes placed in front of 
mica covered openings in the furnace, it was possible to observe the 
ends of the stone cylinder, and thus determine the amount of its expan- 
sion between the different temperatures observed. When a reading 
was to be taken, sufficient resistance was introduced into the heating 
circuit to render the temperature nearly stationary, or only slowly 
rising,—a 110 volt direct current being sent through a resistance which 
was varied from about 4.5 to 20 ohms. 
