136 Prof. C. I. Burton and Mr. W. Marshall. [June 18, 



order to prevent liquids like chloroform or bisulphide of carbon 

 coming in contact with the india-rubber, the tube was filled up to 

 the neck F with clean dry mercury. Twelve tubes were provided, all 

 of which fitted to the same stopper, in order to prevent loss of time 

 in changing substances. The mercury was poured away and a fresh 

 rubber cap wired on for each experiment. 



The galvanometer used was of the dead-beat pattern, by White, of 

 Glasgow, of 642 turns and T242 ohms resistance. It was placed on 

 a slate shelf about 20 feet distant from the compression apparatus, 

 and it was not at all affected by the movements of the pump handle 

 or by any other part of the mechanism. The scale of the galvano- 

 meter was graduated in arbitrary divisions of nearly 2 cm. in length, 

 these being divided into tenths ; these small divisions could again be 

 read to quarters. As one large division on the scale corresponded 

 very nearly to 1 degree centigrade, the temperature readings are 

 very accurate. In order to have all substances as nearly as possible 

 .at the same temperature, a large covered vessel was sunk in the 

 water-tub which contained also the cold junctions, and in this vessel 

 the bottles of substances were placed the day before, along with mer- 

 cury and clean tubes ready for use. The suction tube of the pump 

 also drew its supply from the same source. 



After the gun had been filled with water taken from the tub, a tube 

 tilled with clean mercury to the neck E,and with the substance to be 

 compressed np to C, was then brought up to the stopper and well 

 pressed home, care being taken that no air bubble should be left in 

 inserting the stopper. The wires were arranged on opposite sides of 

 the tube, which was then pressed down on the spiral spring formed 

 by the coiled wire inside the gun, and the ram inserted. The dis- 

 charge valve at fig. 1 was opened, the ram pressed home and secured 

 by its key. Pressure was usually applied and blown off two or three 

 times to prove that everything was in good condition. As little or 

 no air space was left in any part of the apparatus, the pressure rose 

 very rapidly, two strokes of the pump usually sufficing. 



When the galvanometer was steady, its zero point was noted and 

 pressure immediately applied, an operation which took about two 

 seconds. As the apparatus could never be made perfectly tight, it, 

 was always necessary to continue pumping slowly, watching the 

 gauge so as to maintain the pressure as nearly as possible constant till 

 the galvanometer became steady, when the reading was taken and 

 pressure immediately relaxed. 



As soon as the galvanometer was again steady a new zero was 

 taken, pressure applied again, and so on, usually ten times. If the 

 results appeared to agree satisfactorily the ram was removed from 

 the gun, and the tube drawn up and carefully examined to see that 

 no leakage had taken place. It was then set aside, a clean tube 



