On the Steam Calorimeter. 



241 



pointer continues symmetrical, gaining only imperceptibly, perhaps, 

 one division to the right in about five minutes. This will represent 

 about half a milligram. Observed now for another five minutes, a 

 similar addition will be needed. This increment is due to radiation, 

 and the milligram thus accruing daring the ten minutes is not to be 

 included in the value, w Y , which we now have obtained. 



The experiment is now concluded. It is well, after disconnecting 

 it from the boiler, and while the calorimeter is still hot, to dry it out, 

 as the residual heat then completes the drying very thoroughly. The 

 corrections on w l7 necessary to convert it to the value w of the final 

 equation have already been considered. 



The Differential Steam Calorimeter. 



In the use of the apparatus just described, it is certain that a high 

 degree of accuracy is attained. In the thermometry and in the esti- 

 mation of the weight of steam condensed upon the substance, an 

 accuracy of one part in one thousand may, I think, generally be 

 attained. There remain certain causes of error unknown in value 

 within limits, as in all calorimetric methods. In other methods the 

 substance is transferred from a region at one temperature to a region 

 at another. The error of transference has in these cases to be con- 

 sidered. The movement of the substance, when at its highest tem- 

 peratures, through the air, may be a source of serious error ; nor can 

 this error be indefinitely diminished, for a too close approximation of 

 the heater and the cooler causes a transference of heat between the 

 two, in itself a source of error. In the method of mixtures a further 

 source of error is to be found in the continued radiation of the 

 calorimeter and evaporation, of the water contained in it. 



In the method of condensation the substance is not moved, but the 

 medium around it is changed. Is there any error comparable with 

 the error of transference in other methods ? There doubtless is 

 some error, but it is not to be expected that it is at all as great. The 

 error in this case is simply the radiation of the approaching vapour 

 to the substance. Now the velocity with which the steam can be 

 brought to fill the calorimeter is very great, and the momentary radia- 

 tion of the advancing steam upon the substance is in part com- 

 pensated by the precipitation of the most active part of this radiating 

 steam upon the substance. 



Radiation between the substance and its precincts now begins at a 

 certain rate, the value of which when the temperatures become steady 

 can easily be estimated. The first error is a minus error, the second 

 a plus error. There may be also some minus error of the second 

 sort. The errors thus tend towards a balance. All experimental 

 work indicates that the resultant error is very small. Experiments 



T 2 



