372 



MR T. C. BAILLIE ON THE 















Lo<;. Ratio Rise 



Date of 



Initial Temp. 



Corrected Temp. 



Rise of Temp. 



Initial Temp. 



Fall of Temp. 



of Temp, of 



Experiment. 



of Water. 



of Mixture. 



of Water. 



of Nickel. 



of Nickel. 



Water to Fall of 















Temp, of Nickel. 



1897 















July 20 



21-60 



29-75 



8-15 



145-7 



116-0 



2-8467 



ji )> 



L'2-80 



30-93 



8-13 



146-4 



115-5 



2-8475 



21 



20-31 



28-73 



8-42 



147-3 



118-6 



2-8511 



11 >' 



21-12 



29-32 



8-20 



147-3 



118-0 



2-8419 



)1 >) 



21-90 



30-08 



8-18 



147-3 



117-2 



2-8438 



22 



18-78 



27-35 



8-57 



148-1 



120-8 



2-8510 



n !> 



18-99 



27-50 



8-51 



148-8 



121-3 



2-8461 



>» )i 



19 63 



28-05 



8-42 



149-1 



121-0 



2-8425 



23 



21-06 



29-43 



8-37 



149-0 



119-6 



28449 



Arithmetic 

 mean 



1 20-69 



29-02 



8-33 



147-67 



118-65 



2-8462 



Average value of specific heat of nickel, -113. 



The average value of the specific heat of the nickel turnings for a range varying 

 from just under 100° C. to about 20° C. was about '11. This shows that either the 

 thermal capacity is altered in the process of disintegration, or that there is some error 

 in the determination depending upon the size of the pieces employed. The latter I 

 believe to be the case. During the month of June, I did several determinations at the 

 same time with a bundle of copper washers, and with the piece of nickel referred to. 

 After a few trials I found the mass of copper (114*8 grammes) which had the same 

 thermal capacity as the 99*3 grammes of nickel. I tried to discover a difference 

 between the rate of rise of temperature in the calorimeter when the copper was em- 

 ployed from that when the nickel was used. The difference in the times taken to reach 

 the maximum reading was only about six seconds, the whole time being about one 

 minute to one and a quarter. Probably the lag of the thermometer behind the calori- 

 meter obscured the greater part of the actual difference. 



The effect of not receiving all the heat from the nickel would be to make the appar- 

 ent specific heat less than the true specific heat. This error would obviously be greater 

 at low temperatures than at high temperatures, and thus would make the apparent 

 specific heat increase more rapidly with temperature than the true specific heat actually 

 does. Probably this is the real reason why the specific heats of carbon and silicon — 

 so-called bad conductors of heat — have been found to be much lower at ordinary 

 temperatures than that expected from Dulong and Petit's law of constant atomic 

 heats, whereas at very high temperatures their specific heats are much greater and 

 nearly great enough to fulfil the law. Errors of this kind are reduced to a minimum 

 by using Waterman's calorimeter. A description of this apparatus, and a short dis- 

 cussion of the determinations of specific heats is given in a paper by Waterman in the 

 Physical Review (vol. iv. No. 3) for December 1896. 



§ 7. It seems to me a disadvantage of Forbes's method that its accuracy has to 



