1878.] 



Contact Theory of Voltaic Action. 



233 



at temperatures at which it would be solid were it being gradually 

 heated from a low temperature. 



Curve AB, May 16th, gives the relative conductivities when the 

 paraffin was rising from 54° C. to 56° C. At 54° the paraffin seemed to 

 be solid, and it did not appear to be melted at its upper surface, even 

 for some time afterwards, when the thermometers indicated 60° 0. But 

 at 56° C. it was covered with white eruptions. With the condenser, 

 May 16th, we obtained the following tests. As we did not employ the 

 roller the conductivities deduced from detached readings are often in- 

 consistent, and yet each of these conductivities was deduced from a 

 number of observations, in some cases from twenty. This shows the 

 difficulty of making eye observations, for when using the time roller 

 and tracing on it with a pencil the path of the image of the cross 

 fibres, no such inconsistences were met with. 



Temperature. 



1 i Ei 



X 



in megohms. 



60° -2 C. 



1175 



4 -482 x 10 



The paraff 



m was quite liquid over the upper plate. 



62° 



-0205 



2 -569 x 10 6 



64° -5 



-0222 



2 -372 xlO 6 



68° -5 



-0261 



2 -018 x 10 6 



72° -5 



-0312 



1 -689 x 10 6 



76° -6 



-0373 



1 -412 xlO 6 



77° -8 



-0389 



1 -354 x 10 6 



77° -4 



-0334 



1-577 xlO 6 



76° -4 



-0316 



1 -666 x 10 6 



75° -3 



-0310 



1 '699 x10 s 



74° 



-0268 



1 -736 xlO 6 



72° -4 



-0246 



2 -141 x 10 6 



70° -7 



-0242 



2 177 xlO 6 



69° -6 



-0240 



2 -195 x 10 6 



These results may be compared with the galvanometer results, 

 because in both cases the temperature was cooling from a point much 

 above melting. Considering the great difficulty of measuring the 

 temperature, the agreement of these results with the galvanometer 

 results may be regarded as very satisfactory. 



May 20th. Time roller used for the first time, quick motion only 

 employed, that is, one revolution in two minutes. The battery was 

 kept connected with the plates for intervals of time varying from three 

 to ten minutes before plate A was insulated. The roller curves are in 

 themselves extremely interesting, but we think that to give them in 

 the present paper would be superfluous. The curve EF gives the 

 1 E 



values of - log _i without condenser from 52° '6 C. to 58° C. ; the co- 



