fasten only one end to tin- I run. and with tin- lOQSC end 

 tuik-h the Alloy at a. What du we mm -et ? .jS on the 



galvr. instead of 42! This si-.,,\\ a tint the tup I run ha- taken 

 up a great deal of heat, and brings in a counter F. M F. 

 or drag, and thus prevents the full working uf the /.A 

 as indicated by tin- galvanometer. 



\\ V have s en that these experiments do nut bear out the 

 old law vi/.. : "That the greater diffVrenre in temperature 

 en the t\\o ends of a couple, the greater V.. M. F. will 

 IK- produced " F< >r \\ e have shown Fig. 12, that couples 

 3# high. | ;4 on the galr. ; (Fig. 13) i% inches high, 



42 galr. ; and (Fig. 14) I ^ inches high, 49-50. Hut we find 

 that if we make the wire iMii/.e into Thermo-couple 

 inches high. \vc shall have a still /ess difference in tempera- 

 ture between the two ends of the two couples, but we 

 increase our K. M. F. considerably; the same heat 

 easily showing 60 galr.. Two couples with heat 45OF. 

 can be made- to indicate So galr. At this temperature the 

 thenno properties of the Iron are of course much impr 

 by the- Thomson effect. Indeed, rightly applied, IOO 

 difference in temperature gives better results than 2OO 

 applied wrongly. 



The true law of thermo-electricity is 



THE KLKCTRO-MOTIVE-FORCE is PROPORTIONAL TO 



THE KATE < >1 M'KI.h AT \\IIK II HEAT I'AS.sES THE FWO 

 JUNCTIONS. 



Turning to Figs. 16 and 17 for a moment, let us notice that 

 the peas ii/i<.'ays fall off the ZA first. This shows, as we 

 intimated before, that they receive the heat sooner, and arc- 

 soon hotter than the Iron molecules, and that the ZA 

 molecules (Fig. 19) strike in frictional stress against the Iron. 

 But it shows more ; for if the Iron is allowed to pen 

 into the body of the 7. A, a counter K. M V.. <T ( ; . 

 up. and the greatest F. M. I*", is produced by the first thrust 

 of the heat of the /A molecules against Iron. F very thing 



