180 Mr. H. E. Schmitz. On the Determination of [Feb. 2, 



hydrogen thermometer by subtracting 0'003 per degree ; this reduction 

 is made use of in Table IY. 



The absolute value of the Beckmann reading was given with sufficient 

 accuracy by the comparison with the secondary standard. The cor- 

 rection was approximately +14'55 at the lower extremity of the 

 Beckmann scale and + 14'60 at the upper extremity. 



Specific Heat of Water. 



The specific heat of water at 17'5 has been taken as unity, as 

 recommended by Griffiths. In all except a few of the earlier of the 

 experiments on copper, it was not necessary to apply a correction for 

 variation of the specific heat of water ; where necessary, the correction 

 was made by means of the values given by Griffiths, "Thermal 

 Measurement of Energy," Table VII. 



Method of Cooling the Metal. 



Experiments were made with a view to finding a satisfactory method 

 of cooling the metal in an air chamber, surrounded or partly surrounded 

 by liquid air, but these were not successful. The method of direct 

 immersion in the liquid air was therefore adopted. This method has 

 the disadvantage that a certain quantity of liquid air adheres to the 

 body after its extraction from the bath ; this, if carried into the 

 calorimeter, causes an error of more or less importance. Again, it is 

 possible that the temperature of the body may fall slightly after its 

 withdrawal from the bath, owing to the rapid evaporation of liquid air 

 from its surface ; a marked effect of this kind was observed in the case 

 of the platinum thermometer, but in this case the mass was small and 

 the surface large ; in the experiments described below this effect is 

 probably negligible. 



Copper. 



A number of preliminary experiments on pieces of copper of various 

 shapes led the author to the conclusion that the error, caused by the 

 adhesion of liquid air, would be comparatively small if the piece of 

 metal were massive in comparison with its surface. The amount 

 of liquid air entering the calorimeter may moreover be materially 

 lessened by shaking the piece of metal during its transference from 

 the bath of liquid air to the calorimeter. Thus a cylinder of copper, 

 suspended in the bath with its axis vertical, gave, in two experiments 

 in which it was rapidly and steadily transferred to the calorimeter, 

 0-0815 and 0-0811 for the apparent specific heat. On the other hand, 

 when it was shaken during a transference lasting about 5 seconds, 

 it gave in two experiments 0-0801 and 0'0799. It will presently appear 



