Porcelain with Rise of Temperature* 97 



The coefficient of linear expansion at t° C. is given by 



~ =34-25 x 10" 7 + 21-4x 10-n. 

 l Q at 



Since the coefficient varies with the temperature, an uncer- 

 tainty is introduced due to the gradient of temperature 

 from the portion of the tube heated in the furnace to the 

 marks ; but this uncertainty must be very small. According 

 to a rough estimate made by the aid of the auxiliary tem- 

 perature readings in column 5, Table II., the correction for 

 the ends came out to be *6 per cent, at mean temperature 

 800°, falling to '4 per cent, at 600°. 



The error introduced into the temperature measurements, 

 if the part of the tube occupied by the wire did not precisely 

 coincide with the part between the diamond marks, must 

 also have been very small. Thus, for a temperature gradient 

 so steep as 100° per centim. in the neighbourhood of the 

 diamond marks, an error of 1° in the temperature would 

 necessitate a displacement of the wire by nearly 1 centim. 

 The chief difficulty in experiments of this kind seems to be 

 in insuring that the temperature shall be uniform all along 

 the tube. With an ordinary gas-furnace it is impossible to 

 be certain that this condition is satisfied. An accident to 

 the tube after experiment 10 prevented any further measure- 

 ments, and so no experiments were made while the tem- 

 perature was rising. 



As may be seen from Table III., there seem to have been 

 permanent changes in the length of the tube during the 

 course of the experiments, but these changes are very small 

 and apparently irregular. At the end of experiment 7 it 

 was observed that the tube had become very slightly bent ; 

 the irregularities in the length at 0° may be due to this. 

 The total increase in length at 0° from March 13 to April 8 

 was '02 per cent. 



For cubical expansions we get 



v t =v (l + 102-75 x 10- 7 * + 32-4 x lO" 10 ^) ; 

 and for the coefficient at t , 



-~ = 102-75 x lO" 7 + 64-8 x 10" 10 *. 

 v at 



In conclusion the writer must express his indebtedness for 

 the use of the Sidney College Chemical Laboratory, and also 

 for help which he has received from Mr. C. F, Green, B.A., 

 from time to time. 



Phil. Maa. S. 5. Vol. 49. No. 296. Jan. 1900. H 



