Changes of Molecular Configurations. 



197 



The results of these experiments showed no essentially new- 

 points of view. Hence I will omit them in favor of the work 

 of the next paragraph. 



5. The errors discussed in § 4 induced me to repeat the work 

 with a thorough change of method, so far as the thermal 

 measurements are concerned. To obtain the data of table 2, 

 the thick wire ah, figure 1, was itself used as one of the ele- 

 ments of the thermo-couple. The other element was a Ali- 

 mentary wire a/?£, passing from a fixed point p in connection 

 with the terminal of the galvanometer, once around the wire 

 to be stretched (junction, /?), and thence to an insulated spiral 

 spring, i?, to keep it tense. At the point, /?, where the thin 

 wire lapped around the thick wire, both were carefully bright- 

 ened and good electrical contact was further insured by stiffen- 

 ing the spring R as much as the thin wire permitted. The 

 upper end a of the wire to be stretched placed in connection 

 with the other terminal o of the ballistic galvanometer com- 

 pleted the circuit. In this way the heat generated by stretch- 

 ing acts at once at the thermoelectric junction of the thick and 

 the Alimentary wire, while the latter may be chosen so thin as 

 to produce only negligible cooling. Indeed in virtue of friction 

 the discrepancy is apt to be in the opposite direction. 



In this arrangement* a special error is introduced by the 

 change of thermoelectric constants due to stretching, but this 

 error, for the present purposes at least is negligible in com- 

 parison with the thermoelectric powers copper/ iron, or brass/ 

 iron, being not greater than a few per cent. 



The notation of table 2 is the same as that in table 1. E p 

 the mean energy stored in the wire per centimeter of length 

 has been added. 



Table 2.- 



-Energy 



potentialized in 



tensile strains. Second method. 



Metal 



2p h 



P 



61/1 



E 



t 



Et/E 



Ep 





cm 



g 



(cm) 



megalergs 



°C 





megalergs 



Brass 



•166 



60,000 



•199 



11-8 



9-1 



0-55 



5-0 





•155 





•191 



11:5 



9-0 



•56 











•182 



053 



109 



8-8 



•58 

 0-81 





Copper 



•163 40,000 



612 



68 



1-3 





•151 





•153 



6-12 



6-1 



•72 











•153 



6-12 



6-8 



•81 











•143 



5-72 



6-0 



■76 









•143 



5-72 



6-3 



•80 



irliT 





Iron 



•136 50,000 



■084 



4-21 



4-0 



2-2 





•131 



•086 



4-28 



39 



•49 











•089 



4-47 



4-3 



•51 

 0^36 





Iron 



■136 



40,000 



•032 



1-26 



0-8 



09 









•031 



1-23 



05 



• 22 





* After making these experiments, I found that a similar method of thermo- 

 electric measurement has recently been employed by Wassmuth (Wiener Sitzber., 

 xcii, (2), p. 52, 1888). Wassmuth's purposes are distinct from mine. 



Am. Jour. Sci.— Third Series, Vol. XXXVIII, No. 225.— Sept., 1889. 

 13 



