1880.] Lowering of the Freezing-Point of Water, §c. 537 



on for a long time, as the ice melted with considerable rapidity, from 

 the heating of the mercury by compression. 



Series VII. 



As in several cases the junction that was placed inside the bottle, 

 and which had been several times subjected to high pressures, appeared 

 to be affected by the compression and worked somewhat irregularly, it 

 was thought advisable to subject both the junctions to the same com- 

 pression, so as to have them both under the same conditions. For this 

 purpose both junctions were passed through the brass flange (fig. 1), 

 being well insulated by marine glue, the one remaining outside of the 

 test-tube in which the other was frozen. 



Fig. 1. 



The results with ice were exactly the same as when one junction 

 tv as placed outside of the bottle. Pressure still appeared to have an 

 effect upon the junctions, as after one or two series of compression 

 experiments they could be no longer relied upon, and worked very 

 irregularly. The only way to obviate this difficulty was to prevent 

 the junctions being submitted to pressure at all, and this was effected 

 by soldering a stout iron tube (fig. 2), about \ inch internal diameter, 

 into the brass flange, which reached to about the centre of the bottle. 

 The tube was closed at the bottom and contained a few cubic centi- 

 metres of alcohol, into which the junction was lowered from the 

 outside. The part of the tube which went into the bottle was then 

 frozen into a mass of ice, and the iron tube sustained all the interior 

 pressure. The alteration of temperature was conveyed slowly through 

 the iron to the junction. 



