424 Transactions of the South African Philosophical Society, 



In the observations recorded in Table III. and Fig. 2 the effect of 

 changes of temperature were eliminated by reading the thermometers 

 as rapidly as possible in the order 714, 1,470, and 1,470, 714 alter- 

 nately, and taking the mean of the observations thus obtained. 



As the velocity of the wind at this time was thirty-four miles per 

 hour, as recorded by the Dines anemometer, the thermometers were 

 thoroughly ventilated. 



Experiments were then made with the object of determining the 

 relative coefficients of sensitiveness of three different types of 



d m 

 1259 



h m ft m 

 lb J 13 10 



h m 

 13 15 



o 



o 







' 







' 





73 

 72 

 7/ 

















73 

 72 

 7/ 



















\ ,0 



\ .0 













Y* 



\ ° 











70 

















70 

















69 

















69 

















68 

















68 

















67 













^ mm ^f 



< 



67 

















Eapid Fall of Temperature in a Rain Squall on July 10, 

 1906, as recorded by the Mercurial Standard No. 714, and 

 the Minimum Spirit Thermometer No. 1,470 respectively. 



Fig. 2. 



minimum spirit thermometers — spherical, cylindrical, and divided 

 bulb (see Fig. 4) — as compared with the Kew Standard No. 714. 



A description of the thermometers employed, together with a 

 history of their errors, will be found on pages 231, 232. 



A bath of water was prepared at a temperature of 81° + (Fahr.), 

 and the thermometers completely immersed in it for five minutes : 

 they were then rapidly removed to the comparison bath at a tem- 

 perature of 66°+ (Fahr.), and the rate of decrease of temperature 

 observed with the aid of a chronometer, care being taken to read 

 each thermometer at the moment of immersion in the low tempera- 

 ture bath. The results of the experiments are given below. 



