TABLE LXIV. 



The anomalous differences shown in the Christchurch observations are discussed 



later. 



TEMPEBATUEE GRADIENT AND VARIATION OF TEMPERATURE. 



It will be remembered that the method of temperature observation was not iden- 

 tical throughout. With the apparatus, was provided a thermometer contained in a 

 dummy pendulum which was designed to rest on a brass block, itself resting on the 

 base of the stand. The pendulums, on the other hand, hung freely from their agate 

 planes. Under normal Observatory conditions, where the change in air temperature 

 is slow, this thermometer will represent with sufficient accuracy the temperature at 

 the mid-point on the pendulum stem. Under unfavourable conditions, however, when 

 the temperature is rapidly changing, the pendulum temperature must lag behind the 

 observed temperature by an amount dependent on the rate of change. At all the field 

 stations, except Winter Quarters and Christchurch, 1910, the question of lag does not 

 crop up. Here, however, it assumes considerable importance, and special precautions 

 must be taken either to free the apparatus from this source of error or to allow a correc- 

 tion for it The correction for lag which has been applied by the Potsdam observers is 

 +25 X 10~ 7 sec. for a rise in temperature of 1 C. per hour. This is the correction which 

 has been applied to the observations at Potsdam and at Christchurch in 1910. 



The first change was made at the conclusion of Cape Evans, Series A, by the manu- 

 facture of a cork base-plate for the dummy pendulum, so that this was no longer in 

 metallic connection with the base of the stand. This, it was thought, would cause 

 the dummy pendulum to be at sensibly the same temperature aa the working 

 pendulums, so that the observed temperature could be assumed to be correct in all 

 conditions of lag. 



* Last four observations only. 

 79 



