Some Experiments on the Latent Heat oj Steam. 43 



circuiting took place. In addition, platinoid, manganin, 

 and pure nickel wires were all strongly attacked at ioo°, 

 when a current large enough for supplying the necessary 

 amount of heat is sent through them for any length of 

 time. Particularly with manganin, corrosion was found 

 to have taken place near the positive pole, and in this case 

 the liquid left in the boiler after an experiment had a con- 

 siderable quantity of a flocculent greenish yellow precipitate 

 in suspension, which was found to contain manganese. 



The experiment showed, however, that the supply of the 

 necessary heat from an internal source had several advan- 

 tages over the flame, and since in this case the whole of 

 the boiler and tubes above the calorimeter could be 

 wrapped with felt or other non-conducting material, the 

 correction for the heat gained by the calorimeter from 

 external sources was greatly diminished. It was, therefore,, 

 thought worth while to make a new frame wound with a 

 coil of platinum wire, which would not, at all events, suffer 

 from corrosion. A satisfactory form of frame was at last 

 constructed of stout brass wires, on which were slipped 

 corrugated pillars of " red fibre," a fairly good insulating 

 substance, which does not soften at ioo° to the same 

 extent as ebonite. These pillars were held in position by 

 rings of the same fibre, through which the brass wires 

 carrying the current to and from the coil passed. The 

 leading-in wires were so arranged that the ends of the 

 platinum coil were as far apart as possible to reduce any 

 possible conduction through the water to . a minimum. 

 The upper ends passed out to insulated terminals on the 

 brass lid. Two thin platinum wires in parallel were found 

 to be more satisfactory than a single thicker one. The 

 resistance of the coil which gave the most satisfactory 

 results was about 5 ohms, and the current employed varied 

 from 5^ to 7^ amperes, according to the rate of boiling 

 desired. Both with this apparatus (when no steam is 



