10 PROCEEDINGS OF THE AMERICAN ACADEMY. 



610, 613, 611, 612; average 611.5 ± 00.4, or almost exactly identical 

 with value 611 of the first series calculated by Regnault's method. 

 Because, however, the results of this last series must certainly have been 

 too high, the results calculated by Regnault's method must also be too 

 high ; and it is clear that one cause, if not the only cause, of this error 

 must be the lag of the thermometer. 



It is to be noted that in the last series, where irregular radiation-effects 

 took place only during two minutes, the results are more concordant with 

 one another than in the first series, where the total period of irregular 

 radiation was fifteen minutes. This fact again indicates that the calcu- 

 lated cooling correction was at fault. 



Ou studying these figures, it is clear that where only comparative re- 

 sults are required, the method of heating the environment will give 

 satisfactory results so long as the conditions in every experiment are 

 identical. Even when the environment is warmed two minutes too soon, 

 the results are constant ; but of course if this procedure is adopted once 

 in a series, it must be always adopted. 



Therefore, even if the method of warming the environment were in- 

 capable of yielding an absolute value, it would nevertheless be a distinct 

 addition to comparative calorimetry. In the following section it will be 

 shown that under proper precautious the method will probably yield not 

 only a constant value, but a result very close to the true value. 



3. THE COMPARISON OF THE TWO METHODS, AND THE EXACT 

 ESTIMATION OF ADIABATIC RISE OF TEMPERATURE. 



Br Theodore W. Richards and L. J. Henderson. 



The foregoing sections of this paper show indubitably that whenever 

 there is any considerable rate of cooling in a calorimetric experiment, the 

 lag of the thermometer will cause the result to be higher than the truth. 



It becomes now a matter of great interest to determine, if possible, the 

 method of procedure which is capable of yielding the true value for the 

 temperature increment in a strictly adiabatic reaction. The question 

 assumed especial importance because of an extended series of measure- 

 ments of heats of combustion now in progress here. In order to procure 

 further light, it was decided to execute several series of experiments in 

 which the cooling correction was calculated in the usual way, and to cor- 

 rect these results further for the lag of the thermometer. Subsequently, 

 the outcome of these experiments was to be compared with that of a 



