84 Proceedings of Boyal Society of Edinburgh. [sess. 
The chief difficulty with this alloy was its expansion after solidify- 
ing, which invariably shatters any glass vessel in which it is allowed 
to cool down. This difficulty was avoided by using, instead of a 
glass tube, a thin indiarubber one, which was filled with the alloy. 
One end of this tube was bent up at right angles, and into the 
fusible metal at that end of the tube a thin strip of iron (tin-plate) 
was inserted. This junction was placed, beside the bulb of a 
thermometer, in asbestos wool contained in copper cylinders sepa- 
rated from each other by layers of asbestos, and heated by a small 
spirit-lamp underneath. The other ends of the iron strip and the 
fusible metal were joined to copper wires leading through a com- 
mutator to a sensitive mirror galvanometer with a scale at 12 feet 
distance, and the junctions, well varnished, were immersed in a large 
can of cold water. During the sets of observations (which lasted 
