on the Surfaces of Solid Bodies. 243 



tube : the air could pass also through a third communication 

 direct to the tube gg without being either dried or saturated 

 with moisture. In order to retain a constant temperature, the 

 chloride-of-calcium tube, and the tube containing fragments of 

 glass moistened with water, and also the tube forming the direct 

 communication, were immersed in water contained in a large 

 vessel. 



In order that the air blown in might have accurately the same 

 temperature as the pile, it was at the beginning of an experi- 

 ment led through a long spirally bent tube of copper, which was 

 placed in a large vessel filled with water having the temperature 

 of the pile. This was not easily attainable, and it turned out to 

 be requisite to have the whole room in which the experiments 

 were carried on as nearly as possible at the same temperature as 

 the pile. At the time of the year when it is necessary to heat a 

 room, the heating takes place in the evening ; and the morning, 

 when the temperature had equalized itself, was employed for the 

 experiments. It was then a superfluous precaution to place the 

 chloride-of-calcium tube and the tube containing fragments of 

 glass in the same vessel of water. 



The thermo-pile, which by preference was taken for this inves- 

 tigation, contained fifty-six pairs of antimony and bismuth plates, 

 which were so thin that a section of the pile did not amount to 

 more than 13 millims. square. In addition to this pile, another 

 was used consisting of twenty-eight pairs in which each bar was 

 3 millims. in diameter. It displayed effects altogether similar 

 to the more delicate one. Finally, it should be mentioned that 

 the ends of both piles were not pointed, but flat, and lay pretty 

 nearly in the same plane, so that the plates, if they were not in 

 contact with all, were still in contact with the greater number of 

 the bars. 



For the observation of the rise of temperature or of the cur- 

 rent induced by it, I availed myself of the galvanometer, which 

 is described in the paper " On the Diathermancy of Dry and 

 Moist Air," with two mirrors connected after the manner of 

 astatic needles*. A very sensitive astatic needle-galvanometer 

 which I had employed in my former researches was, however, 

 sometimes used. In isolated cases both galvanometers, one 

 behind the other, were introduced into the circuit. Then a 

 deflection of one degree of the needle-galvanometer corresponded 

 to 12 millims. on the scale of the mirror-galvanometer — whence 

 the extraordinary sensitiveness of the latter may be inferred. On 

 blowing the air directly against the pile without drying or satu- 

 rating with moisture, and without covering the pile with the 

 plate, there was no alteration in the galvanometer, provided that 



* Pogg. Ann. vol. cxviii. p. 575. Phil. Mag. S. 4. vol. xxvi. p. 21. 



R2 



