20 
MESSRS. C. T. HEYCOCK AND F. H. NEVILLE ON 
heate 1 to higher and higher temperatures, until finally it broke. The temj^erature 
was called the breaking-j^omt. 
Tlie arrangement rvas as follows :—A porcelain tube, open at both ends, was fitted 
at one end with a cork or plug of soft graphite. (We are much indebted to Messrs. 
Ali.bright and Wilson for their kindness in giving us a suitable sample of graphite.) 
The porcelain tube had a diameter of rather less than a centimetre, being identical 
with the tubes containing the pyrometers. In the figure, G is the plug of graphite 
closing one end of the tulie T. A saw-cut, about 2 millims. wide and a centimetre 
deep, was made longitudinally in the plug, and at right angles to the plane of the 
saw-cut a circular hole of about a millimetre in diameter was also drilled througii the 
plug. A lath of graphite just fitting into the saw-cut, and with a corresponding hole 
drilled in it, passed along the axis of the porcelain tube. A little rod of the metal 
or alloy to be examined was fitted through the holes in the lath and the plug, as it 
were riveting them together. The lath and plug with the alloy were then fitted 
into the tube, and the tube introduced into a tubular electric furnace, and secured in 
that position. The outer end of the lath was connected with a string passing over a 
pulley and stretched by a 50-gramme weight, a transverse stress of definite amount 
being thus maintained on the rod of metal. A pyrometer P was pushed into the 
otlier end of the furnace, so that the end containing the coil was in contact with the 
outer end of the graphite plug, the end of the p 3 U'ometer and the plug being as 
nearly as possible in the middle of the furnace tube. The heating current was then 
started, and very slowly and automatically increased. By this arrangement the 
temperature of the rod of alloy was known at each moment and could be raised witli 
any desired degree of slowness. Finally, the rod liroke ipiite suddenly, the weight 
fell and started an electric alarm so that the breakiDg temperature might be noted. 
The time occupied in heating the alloy up to its breaking temperature was three oi' 
four hours, or even longer. The graphite a})peared to protect the metal perfectly 
against oxidation, and yet it liurnt away so slowly that the same plug and lath were 
servicealfie for several experiments. 
In order to test the method we began l)y determining the lireakiiig-points of pure 
metals. ’ The following table gives the results. Column I. states the freeziDg-point, 
as determined by our pyrometers, and Column II. gives the breaking-point, as 
determined by the above method. 
