Dr. Tyndall on Molecular Influences. 127 



needle of the galvanometer points to zero, when the thermo- 

 circuit is complete, the latter is interrupted by means of the 

 break-circuit key k 1 . At a certain moment, marked by the 

 seconds-hand of a watch, the voltaic circuit is closed by the key 

 h, and the current is permitted to circulate for sixty seconds ; at 

 the sixtieth second the voltaic circuit is broken by the left hand 

 at k, while almost at the same instant the thermo-circuit is closed 

 by the right hand at k'. The needle of the galvanometer is in- 

 stantly deflected, and the limit of the first impulsion is noted ; 

 the amount of this impulsion depends, of course, upon the 

 quantity of heat which has reached the bismuth and antimony 

 junction through the mass of the cube during the time of action, 

 and consequently upon the conductive power of the latter. The 

 limit of the first impulsion being noted, the cube is instantly 

 removed, and the instrument is allowed to cool until the needle 

 of the galvanometer returns to zero. To expedite the cooling, 

 the metallic surfaces of A and B are to a great extent exposed ; 

 the wood is cut away all round them, and from the space between 

 them ; they do not rest upon the wood, their sole points of sup- 

 port being the ivory i at one end and the jaws II* at the other. 

 The cube, as before explained, does not touch the brass plate 

 underneath it, but is supported on its four conical points, and 

 the plate which bears these is itself as much as possible cut away 

 to permit of a free circulation of air through the space efgh. 

 Time is a precious commodity to the experimenter, and by the 

 means described the cooling is hastened and the experiments can 

 succeed each other more quickly. To hasten the cooling further 

 I made use of a pair of small bellows during the first minute 

 after the removal of each cube, and, afterwards, a plate of thin 

 glass was placed over the junction, but not in contact with it. 

 On the glass two drops of aether were suffered to fall from a 

 pipette ; its evaporation caused a refrigeration of the air under- 

 neath, which, in virtue of its increased density, sank and diffused 

 itself around the place of junction. In this way the temperature 

 at the junction was brought a little lower than that of the sur- 

 rounding air ; the needle of the galvanometer being thus brought 

 back, not only to zero, but to a certain point at the other side 

 of it ; at this point the glass was removed and a new cube was 

 introduced; the thermo-circuit was permitted to remain closed 

 until the needle descended to zero, which it slowly did, when the 

 cause of local cooling was removed ; the thermo-circuit was then 

 broken at k', and things stood as at the commencement of the 

 former experiment. The voltaic circuit was once more closed, the 

 current permitted to circulate sixty seconds, then interrupted 

 by the left hand, the thermo-circuit being closed at the same 

 moment with the right, and the first impulsion measured as before. 

 When however these artificial means of cooling are adopted 



