28 PROCEEDINGS OP THE AMERICAN ACADEMY. 



object of this arrangement is to give an indication of the temperature at 

 the point where the inner junction lies against the guard-ring bar G 3 . 

 For a like purpose the junction ^is placed on the bar G 8 , 90o around the 

 guard-ring from G 3 . 



The line W W indicates the boundary of the cotton wrapping of the 

 rods. 



E, F, G and H, I, J, in Figure 3, are copper-german-silver junctions, 

 which have, with respect to certain points on a and /3, a function like 

 that which the junctions Z and Z' shown in Figure 4 have with respect 

 to points on the guard-ring bars. Thus E, with its corresponding ex- 

 ternal junction placed in ice-water or in steam, indicates the temperature 

 of a point midway between the junctions K and K'. H does the same 

 for the junctions iVand N', etc. 



Figure 6 shows a vertical longitudinal section of the apparatus on 

 the scale 1 to 6. G and C are the conductors carrying the main cur- 

 rent into and out of the apparatus. This current, about 25 amperes, 

 is measured by an ammeter not here shown. 



Figure 7 shows the arrangements on the top of the board B. It is to 

 be noted that each couple on a is joined in opposition to the corresponding 

 couple on /?. For example, the wires coming up from the junctions K 

 and N (Figure 3) are soldered to one and the same wire, which leads 

 away from B and may run to the galvanometer, though it is usually dis- 

 connected at its outer end. The wires from junctions A"' and N', re- 

 spectively, on the other hand, connect above B with two separate wires, 

 which lead off toward the galvanometer. The three wires thus running 

 from the couples KK' and N W make it possible to put into the galva- 

 nometer circuit either of these couples alone or both in opposition to 

 each other. The other pairs of couples are similarly connected, L L' 

 with ! , and MM' with PP. 



It is unnecessary to depict here the other apparatus used, concerning 

 which some details will be given later. 



The general order of procedure on October 22, 1904, a fair sample day, 

 was substantially as follows : 



Time. 



a.m., 8 : 05 Water ice-cold in left box, boiling in the other. 

 9 : 00 Main current put ou in condition F. 

 10 : 35 Measurement of sensitiveness of galvanometer. 



" " disturbing thermo-electric effects. 



" u main current. 



