1823.] Prof, Cumming on Thermoelectric Relations, S^c. 17 7 



state every thing which we have been able to ascertain with the 

 greatest certainty during the short period of time of which it 

 occupies the middle. Ihe rule for the measurement of heights 

 supposes that we know with great precision the temperature of 

 the air, and that of the instrument : both these then we must 

 discover, and when they disguise themselves, must divest them 

 of their disguises. There are times when this is altogether 

 impossible ; but it is at least something to perceive this, and 

 to know that we have to doubt an observation, well performed 

 in itself, but of which we cannot be sure that we have entirely 

 satisfied some of the fundamental conditions. 



Under some circumstances, and those very common, the local 

 temperature is so predominant, that in spite of all our care, it 

 will enter all our estimates : it is in vain to guard against it. 

 Measurements made in a hot season, and under a burning sun, 

 always tend to err in excess ; especially if the station be of such 

 a nature as to multiply the reverberations of heat. On the other 

 hand, they will err in defect in foggy or rainy weather, especially 

 if the place is so circumstanced as to concentrate the cold. The 

 error proceeds from the thermometers. To point out its origin is 

 to warn the observer to avoid, if he can, the circumstances 

 which produce it ; and if he has not the choice, to allow for 

 their influence in the opinion he forms of his measurements, 

 {Tq he contimicd.) 



Article II. 



A List of Substances arranged according to their Thermoelectric 

 Relations, with a Description of Instruments for exhibiting 

 Rotation by Thermoelectricity . By the Rev. J. Gumming, 

 MA. FRS. and Professor of Chemistry in the University of 

 Cambridge. 



(To the Editor of the Annals of Philosophy.) 



DEAR SIR, Ccmhrtdge^ July 23, 1823. 



The following tables will, I hope, be interesting to those who 

 have read my communication to you in April last. The first 

 contains the thermoelectric relations of different substances, 

 with copper wires ; the second, their relations to each other 

 taken two and two together, each substance being positive to 

 all below, and negative to all above. The voltaic series, and 

 the order of conductors of electricity and heat, are added, 

 merely to show that the thermoelectric series has no accordance 

 with either of them. 



New; Series, vol. vi. ]sr 



