46 Dr. J. A. Harker on 



the steam tubes leading from the boiler, together with the 

 cylinder, were wrapped with the lead pipe described pre- 

 viously, and covered by several layers of green baize. A 

 side tube in the wall of the cylinder was arranged so as to 

 be nearly opposite the end of the steam-delivery tube, 

 and a small metal vessel was placed so as to collect any 

 drops from it. 



After steam from a subsidiary boiler had passed through 

 the steam jacket for about half an hour, and steam from 

 the electrical boiler along the horizontal tube for some 

 minutes, the three-way tap was turned so that the steam 

 was sent downwards, and the point of the tube B was 

 watched for about ten minutes. During this time two or 

 three large drops ('05 — '2 cc.) of condensed water formed 

 and fell off into the vessel. The experiment was 

 repeated under varying conditions and with steam from 

 different sources, but always with the same result. The 

 average total weight of water condensed in the two small 

 forms of glass apparatus is only about 10 grms., and this 

 amount of uncertainty is quite sufficient to make the results- 

 untrustworthy. 



Before attempting to make a new apparatus, I wished 

 to ascertain if it were possible to efficiently stir a larger 

 calorimeter in order that by increasing the water equiva- 

 lent the amount of water condensed might be increased 

 without necessitating a greater rise of temperature than 

 had previously been obtained. The up-and-down stirrer, 

 which has the objection of being always more or less 

 inclined to shake the calorimeter, has the further dis- 

 advantage that unless the blades and supporting rod are 

 very thick, they require a cage of guide-rods on which to 

 slide, which occupies a considerable part of the space in 

 the water. On the other hand, in a round calorimeter, a 

 rotating stirrer is not efficient, unless the blades are large 

 in proportion to the vessel, and it requires also to be placed 



