406 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



interrupter were admitted. Into the neck of the bell-jar, by means of a 

 screw-connector, was fitted a funnel-tube closed above by a stop-cock 

 and communicating with a horizontal glass capillary to be used as an 

 index. The bell-jar, funnel-tube, and a part of the capillary were filled 

 with a light transformer oil. The expansion of the oil, read on a scale 

 attached to the capillary, served as a measure of the heat developed 

 in the jar. Within the oil, around the interrupter, black paper was 

 loosely wrapped to prevent radiation. 



For comparison with calorimeter A, two calorimeters " B " and " C ' 

 of different resistances were constructed of the form shown in Figure X. 

 The resistance consisted of a straight manganine wire connected by 

 spriugs of copper to thick pieces of platinum sealed into the ends of a 

 glass tube 8 cm. in diameter. This tube was also filled with oil and 

 provided with a capillary index similar to that of calorimeter A. The 

 manganine wire was supported axially within the tube by circular discs 

 of glass. About these discs of glass, within the oil, half-way between 

 the wire and the walls of the tube, a cylinder of black paper was wound 

 to prevent radiation. The resistance of the manganine wires of calori- 

 meters B and C was measured on a Wheatstone bridge. Computed by 

 Rayleigh's formula, the correction for surface travel when these wires 

 should be used with the oscillatory current, was found to be negligibly 

 small on account of the high specific resistance of the material of the 

 wires. The calorimeters B and C had the following constants : 



To determine what amount of heat corresponded to 1 cm. of expansion 

 of the oil, the three calorimeters were calibrated by a direct dynamo cur- 

 rent through the bulb of the mercury interrupter and the two manganine 

 wires in series. The current was started, as in the Aaron's lamp, by 

 tipping the bulb so that the mercury of the two electrodes came momen- 

 tarily together. The energy expended in the three calorimeters by the 

 calibrating current could now be calculated from the current, the voltage 

 about the interrupter, the resistances of the two manganine wires, and 

 the time. 



