Heat of Granite and Basalt at High Temperatures. 65 



couple was unfortunately broken, so that in the first set of 

 determinations the average temperature was obtained from 

 the reading of a couple placed in the narrow space between 

 the cylinder and th« inside of the copper tube, a suitable 

 correction being added for the radial temperature gradient. 

 In this and the later experiments tests were always made for 

 leakage from the central wire to the couples. This was 

 always inappreciable except at temperatures about 600° C. 

 Tests were also made with the furnace at a steady high 

 temperature and no central current, when A was found to 

 be zero as it should be. One, or occasionally two determi- 

 nations were made per day. As the actual readings con- 

 stituting a day's work were very similar to those obtained in 

 the case of the granite determination, a sample of which is 

 given later, there is no need to insert them here. 



The daily mean results are shown in Table I. in chrono- 

 logical order, the symbols having the same meaning as 

 before. 



Table I. (Basalt.) 



e. 



K. 



W. 



9. 



K 



W. 



9. 



K. 



W. 



125° C. 



5-22 



245 



68° 0. 



5-10 



5-86 



180° C. 



4-82 



5-38 



40 



5-00 



2-32 



68 



491 



3-60 



295 



4-61 



5-27 



99 



503 



2-34 



69 



4-88 



2-85 



265 



4-85 



6-92 



52 



5-08 



3-43 



70 



5-01 



4-56 



146 



4-88 



7-07 



155 



513 



3-41 



69 



5*15 



578 



133 



499 



860 



180 



517 



2-83 



133 



5-19 



542 



104 



5-07 



8-98 



319 



4-93 



2-52 



131 



5-61 



2-85 



222 



4 75 



7-75 



233 



5-20 



256 



143 



4-88 



5*55 



339 



4-60 



7-46 i 



278 



4-81 



4-14 









378 



4-53 



7-27 



62 



4-98 



420 









431 



443 



721 



298 



495 



4-01 









500 



91 



190 



4-32 



4-87 

 4-70 



6-79 

 7-73 I 

 7-33 















204 



4-70 



7-87 i 



The values of K must be multiplied by 10 



The results in the first three columns are the first ones 

 obtained ; this set was brought to a conclusion by the 

 breakage of one of the wires of the multiple couple. These 

 results, which are not very consistent, are in marked con- 

 trast to those previously obtained for granite, as they show 

 an apparent rise of conductivity with temperature up to 

 about 200° C , above which a fall is shown. 



The cylinders were taken to pieces and the couples remade, 

 the single couple being placed in its proper groove, and the 

 cylinders cemented together again and replaced in the furnace. 



Phil Mag. S. (>. Vol. 27. No. 1 157. Jan. 1914. F 



