320 white: thermoelectric methods 



ing which measures the difference in the temperatures of the two. 

 It is thus possible to proceed almost as if the external tempera- 

 ture had no effect at all. This has been claimed as a second 

 great advantage of the twin method. A complete jacket, how- 

 ever, is much more effective in this direction, and, on the other 

 hand, there is a cooling correction error peculiar to the twin 

 naethod. It arises whenever the two calorimeters are not exactly 

 alike, and can easily be shown to be (Ka — Kb) * (^b — ^j) where 

 Ka and Kb are the cooling rates of the two calorimeters, 6^ and 

 dj the temperatures of the comparison calorimeter and of the 

 jacket, respectively. This error is nearly eliminated if three 

 observation periods are run (as in the usual Pfaundler Method 

 with one calorimeter) , but it can be more easily avoided by mak- 

 ing the temperature difference, 0b — Gj, small, say below 0.2? 

 Only two observation periods are then needed. 



The trouble of the temperature adjustment is largely or wholly 

 avoided by either of two schemes, both of which utilize the low 

 cooling rate of a vacuum-jacketed flask, which is used for the 

 comparison calorimeter. In one scheme, a regulator keeps the 

 jacket temperature constant. The flask is then constant to a 

 very high degree of precision (easily to 0.0001°) without any 

 further attention whatever. This has been called the constant 

 comparison body method. 



The other scheme avoids the regulator. A specially adjusted 

 (shunted) thermoelement is used in determining the cooling cor- 

 rection. The two unlike calorimeters can then be operated as 

 easily and conveniently as if they were twins, but the tempera- 

 ture adjustment is so much less exacting as to be almost negligi- 

 ble. This has been called the compensated calorimeter method. A 

 combination of the two methods can be made at will, either 

 temporarily or permanently, involves no difficulties, and reduces 

 precautions and corrections to a very low minimum. 



The gain in simplicity and ease of construction by using vacu- 

 um-jacketed bottles is evident. An effective and particularly 

 simple way of us'ng the bottle is to immerse it wholly in the 

 jacket water, with an inverted cap over it containing air, A 

 tube running down thru the cap allows the thermoelernent to be 



