44 Dr. J. A. Harker on 



passing through the jacket) and the Berthelot form, a con- 

 siderable quantity of water is volatilised and condensed in 

 the lower part of the tube before the water really boils. 

 Berthelot makes no mention of this, and it is difficult to see 

 how it can be allowed for in the calculation of his results. 

 By the provision of the small steam-trap described, any 

 condensed water can be drawn off when the boiling has 

 commenced, and even if this is not done it is prevented 

 from lodging in the pipe and afterwards being carried on 

 into the condenser. 



Many experiments were made with this apparatus, using 

 the same calorimeter as in the earlier experiments, which 

 held about 1,700 grms. water, stirred by an up-and-down 

 stirrer. The condenser was of thin sheet copper with 

 spiral of narrow copper tube. It was joined to the vertical 

 tube projecting downwards by a piece of ebonite into which 

 it was screwed, and it was always arranged that the water 

 level of the calorimeter reached to the middle of this 

 piece of ebonite. By this means the otherwise very 

 considerable evaporation of the calorimeter water from 

 the surface of the vertical steam-pipe was entirely 

 obviated, conduction of heat downwards during the 

 initial and final periods of the experiments very much 

 lessened, and at the same time an efficient detachable 

 joint formed. It was some time before the cause of 

 the anomalous results obtained with this first condensing- 

 worm was explained. It lay in the fact ,that during 

 the passing of steam into the condenser, minute cracks 

 in the copper tube, caused probably by the strain of 

 the bending into its spiral form, allowed a small quantity 

 of water from the calorimeter to leak inwards to the con- 

 denser, even though when cold all remained quite tight for 

 hours. This was obviated by the use of solid drawn 

 copper tube in all subsequent experiments, instead of the 

 usual kind with brazed seam. 



