290 Messrs. Roberts and Wrigbtson on the 



filled with distilled water at a known temperature. Tbe 

 necessary data were thus afforded for accurately determining 

 its capacity at tbe temperature of tbe air. Molten silver was 

 then poured into it, the temperature at tbe time of pouring 

 being ascertained by tbe calorimetric method. Tbe precau- 

 tions, as regards filling, pointed out by Mr. Mallet, were 

 adopted ; and as soon as the metal was quite cold, the cone 

 with its contents was again weighed. 



Experiments were also made on the density of fluid bismuth ; 

 and two distinct determinations gave the following results : — 



10-072 } Mean 10 ' 039 - 



The invention of the oncosimeter* appeared to afford an 

 opportunity for resuming the investigation on a new basis, 

 more especially as the delicacy of the instrument had already 

 been proved by experiments on a considerable scale for de- 

 termining the density of fluid cast iron. The following is 

 the principle on which this instrument acts. 



If a spherical ball of any metal be plunged below the surface 

 of a molten bath of the same or another metal, the cold ball 

 will displace its own volume of molten metal. If the densities 

 of the cold and molten metal be the same, there will be equili- 

 brium, and no floating or sinking effect will be exhibited. If 

 the density of the cold be greater than of the molten metal, 

 there w T ill be a sinking effect, and if less a floating effect, when 

 first immersed. As the temperature of the submerged ball 

 rises, the volume of the displaced liquid will increase or 

 decrease according as the ball expands or contracts. In order 

 to register these changes the ball is hung on a spiral spring, 

 and the slightest change in buoyancy causes an elongation or 

 contraction of this spring which can be read off on a scale of 

 ounces, and is recorded by a pencil on a revolving drum. A 

 diagram is thus traced out the ordinates of which represent 

 increments of volume, or, in other words, of weight of fluid 

 displaced — the zero-line, or line corresponding to a ball in a 

 liquid of equal density, being previously traced out by revol- 

 ving the drum without attaching the ball of metal itself to 

 the spring, but with all other auxiliary attachments. By a 

 simple adjustment the ball is kept constantly depressed to the 

 same extent below the surface of the liquid ; and the ordinate 

 of this pencil-line, measuring from the line of equilibrium, thus 

 gives an exact measure of the floating or sinking effect at every 

 stage of temperature, from the cold solid to the state when the 

 ball begins to melt. 



* Wrightson, Journ. Iron and Steel Inst. No. II. 1879, p. 418. 



