[WHEELER] THERMAL EXPANSION OF ROCK 25 
The furnace (See Figure I.) consisted of a hollow cylinder G H 
of the special nickel steel, the thermal expansion of which had been 
compared with that of the granite and diabase. This cylinder, within 
which was placed the stone A B to be tested, was about 27.5 cms. long 
and slightly larger in inside diameter than the cylinders of stone used. 
In one side of the nickel steel cylinder, at short distances from the ends, 
openings were made about 20 cms. apart, to make possible the focusing 
of the two micrometer microscopes M & N on the ends of the specimen 
of rock being examined. A hand-feed arc lamp placed between the mi- 
. croscopes was used to illuminate the ends of the stone cylinder. To in- 
sulate the heating coil (See dots. Fig. I.), which was of nickel wire of 
about 1.37 ohms total resistance, from the metal cylinder, the latter 
was covered with thin sheet asbestos, upon which the nickel wire was 
wound spirally, together with a coarse asbestos string, serving to in- 
sulate one turn of the spiral from the next. This heating coil extended 
about 3 ems. beyond the ends of the cylinder of rock placed within. 
Mica coverings C & D were placed over the openings at the ends of the 
stone and the furnace was wound with several layers of sheet asbestos 
EF, so that with the stone at 1000° C.the outside would not become very 
hot. For the present work the furnace was lengthened by winding 
with more sheet asbestos to make possible packing with asbestos wool 
for a greater distance from the ends of the stone. 
The hot junction of the thermo-couple was placed within the 
nickle steel cylinder of the furnace, usually near the middle of a groove 
along one side of the cylinder of stone, and insulated from the stone by 
means of a thin strip of mica and from the metal cylinder of the furnace 
by means of a strip of sheet asbestos. The cold ends passed into mer- 
cury at the bottom of two test tubes R &S suspended in a jar of melting 
ice to keep them at a uniform temperature of 0° C. From the mercury 
contained in the test tubes, copper leads V & W consisting of ordinary 
electric lighting twin wire, perhaps 50 ft. in length (to make it possible 
to perform the sulphur-vapour calibration in a hood at some distance 
from the potentiometer) led to the proper terminals of the potentio- 
meter. 
The potentiometer bridge used (See Fig. II) consists of a hard 
rubber cylinder A B about 34 inches in diameter on which is cut spiral 
groove of about one centimeter pitch running the whole length of the 
cylinder. The cylinder is divided into 30 equal divisions for each turn 
of the groove, making about 1300 equal divisions in all. The contact 
piece P is so attached that it moves along directly over the bridge wire 
as the cylinder is turned. The whole is attached permanently to a 
base-board. Immediately behind the cylinder is a resistance box made 
of 41 coils, 1 of 300 ohms, 10 of 100 ohms, 10 of 10 ohms, 10 of 1 ohm, 
