232 



Profs. W. E. Ayrton and John Perry. [Mar. 21, 



We were unable to obtain galvanometer and electrometer tests at 

 the same temperature, since the galvanometer did not give deflections 

 at low temperatures ; and as we had not at this time designed the time 

 roller described above, and as no auxiliary condenser was used with 

 the electrometer, the loss of charge was too rapid at high temperatures 

 to make accurate time observations. 



The oil-bath used in November, being small and otherwise incon- 

 venient, we proceeded in May of this year with the investigation, 

 using the apparatus described above. At low temperatures time ob- 

 servations of loss of charge were taken on a fixed scale. Near the 

 melting point time observations were taken on the roller in motion, 

 and at higher temperatures, one terminal of an auxiliary condenser of 

 one-third microfarad capacity and excellent insulation was attached 

 to the plate A, the other terminal of the condenser being connected 

 with B. Four Daniell's cells were used to charge A. 



At 20° * 9 C, the temperature of the room, A experienced no loss of 

 charge in ien minutes whether or not the condenser was attached. 



A number of preliminary experiments showed as before that there 

 was a great increase of conductivity at melting. For instance, on 

 May 15th, when the temperature was rising, the paraffin appeared to 

 be solid at 46° C, and it appeared to be liquid at 52° 0., and between 

 these two temperatures we obtained the following results, t being 

 measured in minutes always. 



Temperature. 



1 i E x 



I l0g E 3 



X 



in megohms. 



46° -5 C. 



-00114 



34190 X 10 6 



48° -3 



"00168 



26480 x 10 6 



50° -0 



-00225 



17320 xlO 6 



50° -4 



-00607 



6421 x 10 6 



50° '8 



-01600 



2436 x 10 6 



51° -15 



-03970 



998-2 xlO 6 



51° -2 



-07000 



556-8 xlO 6 



51° -3 



-30000 



129-9 xlO 6 



51° '35 



1 -26000 

 Afterwards when cooling. 



30 -93 x 10 8 



48° '4 



-0689 



565 '7 x 10 6 



48° -2 



-0308 



1266 x 10 6 



45° I 



-0133 



2930 x 10 6 



These results cannot be compared directly with those previously 

 obtained with the galvanometer, because in the former case the paraffin 

 was cooling from a high temperature, 90° C, whereas in this case 

 the temperature was only raised to 51°, and it is well known that 

 when paraffin is cooled from a high temperature, it will remain liquid 



