42 PROCEEDINGS OP THE AMERICAN ACADEMY. 



after which the distance between the junctions of each couple was tested 

 by means of calipers ; and one or two junctions were moved a trifle in the 

 still yielding varnish in order to make this distance, as nearly as might 

 be, 1 cm. The windings were then recoated with asphaltum and baked 

 at 100° for some hours, then cooled and re-examined, then coated once 

 more at the points where the copper wires leave the rods, then finally 

 baked for six hours at 200° C. 



No grooves were made to receive the supplementary thermo-electric 

 junctions (see F and I, Figure 4), which were put on to give the tem- 

 peratures of the midway points of the various couples now fixed, accord- 

 ing to the method just given, on the bars ; for it was not necessary to 

 place these supplementary junctions with any great exactness. They 

 were at first put on with a thick layer of asphaltum, which was allowed 

 to dry for some ten days before baking ; but when baking was attempted 

 the heat so softened this coating that it was necessary to start anew. 

 Success in fastening these junctions was finally attained by keeping the 

 new coating of varnish in which they were imbedded at 50° for several 

 hours before the final baking. The worst placed of these junctions at 

 the last was the middle one on bar a, which was about half a millimeter 

 too far toward the K K' end of the bar. Such a displacement must have 

 produced certain temporary dissymmetries in the observations, but, with 

 the various reversals which were practised, cannot have had any serious 

 effect on the final result. 



Calibration of Thermo-electric Couples. 



With the same kinds of wire, and the same methods of annealing and 

 soldering, Mr. Churchill made, for calibration use, four more thermo- 

 electric couples, A, B, C, and D. The german-silver wire in each of 

 these was about 1.6 m. long. Near each end it ran through a glass tube 

 about 40 cm. long, to keep it from accidental contact with the copper 

 wire, which was lashed to the outside of the tube. For a calibration test 

 each junction, with a considerable part of the glass tube above it, was 

 placed in a mixture of ice and water or in water near the temperature of 

 the room, or in the vapor of some boiling liquid, as the bulb and stem of 

 a thermometer under test would be placed. The temperatures, except 

 those of melting ice and boiling water, were determined by means of one 

 or the other of a pair of Baudin thermometers, which Mr. Churchill had 

 carefully compared with a Tonnelot corrected to the hydrogen scale by 

 the Bureau International des Poids et Mesures. 



