RICHARDS. — THE ELIMINATION OF THERMOMETRY LAG. 13 



It will be noticed that the correction for thermometric lag averages 

 about 0.008°, or about 0.23 per cent of the total rise in temperature. 

 This thermometric lag was not precisely similar in all the experiments, 

 because the rate of cooling was not always exactly the same in each case. 



It is further noteworthy that the two methods give essentially the 

 same result, the average of the six determinations with the first method 

 being 3.377° ± 0.001°, while that of the three determinations with the 

 second method is 3.378° ± 0.001°. Thus each confirms the other. 



Moreover, attention should be called to the fact that twice as many 

 determinations were needed to secure a given degree of probability by 

 the first method as by the second, although indeed the series are too brief 

 to make the application of the method of least squares satisfactory. 



After this comparison of the various methods of calculation formerly 

 used, and the correction of the results for the newly applied correction 

 for thermometric lag, it became a matter of great interest to determine 

 whether or not the final result thus obtained is identical with that 

 yielded in the same reaction by the new method of experimentally 

 eliminating the correction for cooling. The results of this new compari- 

 son, even more carefully made than before, are given below. 



The solutions previously used having been exhausted, new, less con- 

 centrated ones were prepared. The apparatus also was somewhat 

 altered. A large platinum calorimeter, capable of holding over a litre, 

 was employed to contain the reacting acid, and this was fitted into a 

 deep copper cylinder, so as to leave an annular space of air between 

 them only a few millimeters in thickness. It is obvious that this air 

 jacket should be as thin as possible in order to reduce its heat capacity. 

 The calorimeter was covered by a cardboard diaphragm, pierced with suit- 

 able holes for the thermometer, stirrer, and alkali vessel. The last was in 

 this case not immersed in the calorimeter as before, but raised above it, 

 with the delivery jet just touching the acid into which the alkali was to be 

 discharged, — a change adopted in order to make the delivery as rapid as 

 possible. The temperatures of both acid and alkali were read with great 

 care ; and because the volume and heat capacity of the latter were small, 

 it was easy to apply a small correction for its deviation in initial tempera- 

 ture, that of the acid being taken as the true initial temperature. The 

 small correction necessary is given in each case in the tables below. 



The copper cylinder was highly polished inside, and was sunk deeply 

 in a bath of water whose temperature could either be maintained at a con- 

 stant point or varied at will. A second cardboard cover protected the 

 alkali tube from accidental changes of temperature. It would of course 



