OF IRON, COPPER, AND GERMAN SILVER. 



551 





Rate of Cooling. 



Temperature Excess. 





From "Grouped" Readings. 



From " Single " Readings. 



°C. 







170 



352 



352 



180 



3-87 



3-87 



190 



420 



420 



200 



456 



4-56 



210 



4-92 



493 



220 



5-28 



5-28 



230 



570 



569 



240 



610 



613 



250 



6-55 



660 



260 



7-06 



710 



270 



7-54 



7-70 



The above table is given to illustrate the results of the method of reduction; 

 for a particular reason (connected with the circumstances under which the 

 above experiment was conducted) the above values of the rate of cooling were 

 not finally employed in the deduction of conductivity. 



This process of grouping readings was found valuable in the deduction of 

 the rate of cooling at the highest temperatures employed ; since the irregular 

 falling of the mercurial column almost necessitated some equalisation of this 

 kind. This irregular descent of the mercury column is observed in the very 

 best thermometers, and is aggravated in the present case by the (necessary) 

 smallness of the bore, and also perhaps by the rapidity of the cooling at the 

 higher temperatures. 



Forbes' peculiar result as to the rate of cooling at different temperature 

 excesses has already been alluded to, as also has its explanation. A comparison 

 of results for the case of iron in the same way as has already been given for the 

 statical temperature curve, may now be given. Diagram 6 shows the three 

 final results, as they were given by Principal Forbes (marked F), by Prof. 

 Tait (marked T, and reduced, according to his statement, by 6 per cent.), and 

 lastly, the results of experiments made in connection with this work (marked M) ; 

 the last, however, having been modified as explained in Appendix I. 



The last two mentioned results clearly disprove the result of Forbes' 

 experiments, viz., that at or about 150° C. the curve should exhibit a point of 

 contrary flexure. 



The relation of this to the final results will be noted later. 



The following table gives the orclinates of the three different curves beyond 

 150° of temperature excess, up to which they practically agree. The heading of 

 the columns indicates the set of experiments to which they belong, as explained 

 above : — 



VOL. XXXIII. PART II. 4 M 



