1132 
MR. G. F. ROHWELL ON THE EFFECTS OF HEAT ON CERTAIN 
silver) gives '00001921 ; and more recently Matthiessen (Phil. Trans., 1866) found 
it to be '000019436. 
Rods of other metals gave the following results :— 
Expansion-apparatus, 
Iron. -000011025 
Copper. -0000174433 
Lead . .. *0000302121 
Zinc. -0000288761 
Various observers. 
r -00001182 
l -00001194 
f-00001666 
-00001678 
[-00001722 
•00002924 
•00002918 
It is thus evident that the apparatus is capable of considerable accuracy, especially 
when we remember that we are dealing with rods of substance only 6 inches in length. 
I venture to think that such an apparatus would prove a useful adjunct to a Physical 
Laboratory; for it would not only afford a means of determining coefficients of ex¬ 
pansion with rapidity and accuracy, but it might be used for determining thicknesses 
in thousandths and ten-thousandths of an inch. For if the substance were introduced 
between the agate face of the micrometer-screw and the lever B, and if the index 
were then driven round to a certain point (say 20° of arc), and so with the other 
substances tested, an exact uniformity of pressure of the face of the screw on the 
substance would be secured, and minute thicknesses might thus be measured with 
accuracy. 
In constructing such an apparatus, extreme rigidity of the fixed parts is a main 
necessity; the base should be of thick marble, and the micrometer-screw support 
should be deeply let into it and firmly secured. Massive supports, apparently fixed 
with the utmost firmness, sometimes yield with surprising readiness to the extent of 
yJ-Q-th of an inch. The elasticity of the metal often causes great inconveniences. 
The levers must be firm and strong, but not unnecessarily heavy ; they should work 
very smoothly. Agate should be let into the brass at the points of contact. Care 
should be taken to avoid any transference of heat from the hot trough to other parts 
of the apparatus. By the use of a paraffine of high boiling-point, the temperature of 
a bar submitted to heat in the trough (F) can be raised to 340° C. Under such 
circumstances great care is requisite to prevent heat from being radiated or otherwise 
communicated to other parts of the apparatus. 
Means employed to determine the expansion in passing from the solid to the 
liquid condition. 
The determination of the amount of expansion which the various bodies examined 
undergo in passing from the solid to the liquid condition was effected in the following 
manner :—A conical tube of platinum weighing 44'844 grms. was carefully filled with 
mercury at a known temperature. It contained 105'43 grms. of mercury. It was 
