L'JO 



MKSSRS. R. THRELFALL AND J. A. POLLOCK 



traced to the method of observing by heating the instrument from below with a naked 

 flame, as before described. When we observed only natural maxima and minima, the 

 readings at Sydney regained their old regularity, with the exception of one observa- 

 tion on the afternoon of December 20th, which is 2 sextant minutes from the mean 

 line. In our present instrument there is a fault in design, inasmuch as the microscope 

 is fixed to a copper tube, while the thread is carried by a gun-metal framework which 

 we now try to fix firmly at both ends to the copper tube. It is only to be expected 

 that, owing to the different expansibilities of the two metals, the instrument may be 

 twisted in a most erratic way during changes of temperature, and this effect will be 

 exaggerated if the changes of temperature are rapid. Moreover, by the addition of 

 glass tie bars the character of the spring system has been entirely changed. 



Plot 8. 



10' 



az'o' 



Dayafmonthr e a id li 



Hot nsby 



u /4 



December lase. 

 Readings reduced to 23 C. 



Plot 9. 



December IBM. 

 Readings reduced to 23 C. 



Two more journeys were made to Hornsby, and the observations are shown in 

 Plots 8 and 9. It is seen from these plots that the readings are not now affected by 

 travelling, for, without correcting for daily rate, we have the following differences 

 between an observation at Hornsby and the observation before and after at Sydney : 



