14 Mr. E. H. Griffiths and Miss Marshall on the 



withdrawn and for the first time inverted — the air-bubble 

 rising to the closed end. The case was opened for a moment 

 and the dropper at once lowered into place by means of the 

 previously attached thread : this operation only occupied a 

 couple of seconds. As the dropper left the case the latter 

 was instantly closed by a second operator to prevent the 

 escape of any benzene vapour left within it. Thus from the 

 time of inversion no change took place in the temperature, 

 and therefore in the volume of the air-bubble ; also no 

 appreciable change in the temperature of the calorimeter (0J 

 was caused by the introduction of the dropper and contents. 

 However, the mere act of opening the tube leading down to 

 the calorimeter caused a slight lowering of 0„ and in order to 

 re-establish the equality between X and O before commencing 

 an experiment, it was necessary to switch on, for a second or 

 two, the current from the exterior coil in the tank to the 

 calorimeter coil. When the observer at the differential- 

 thermometer galvanometer announced that 6 — l was small 

 and steady, the time for commencing the experiment had 

 arrived. 



10. In our later experiments, when we had become more 

 expert at the various operations, the time from the insertion 

 of the dropper to the commencement of an experiment was 

 from 3 to 5 minutes. In our earlier observations at 30° the 

 time was, however, much longer — from about 8 to 15 minutes. 

 During this interval there was generally apparent a slight 

 lowering of 1? which made us fear that the evaporation 

 through the capillary was appreciable, and it was not until 

 our tenth experiment that a means of meeting this difficulty 

 suggested itself. It was evidently necessary to wait until #, 

 became steady before commencing an experiment, and it was 

 difficult to see how to shorten the time required to establish 

 this condition. 



From Experiment X. onwards the procedure was as follows. 

 A glass rod whose upper end passed through a cork fitting 

 the opening of the entrance-tube was lowered into the 

 calorimeter. Round the lower three inches of this rod was 

 strapped (by fine platinum wire) a thin roll of cotton-wool of 

 which the upper two inches were saturated with benzene 

 while the lower end was left dry, so that there was no danger 

 of drops of benzene falling from it into the silver evaporating- 

 flask. 



The rod was withdrawn occasionally and more benzene 

 added if the upper portion of the roll had become dry. Obser- 

 vation of the galvanometer showed when the cooling effect 



