in order that the measurements at Cape Evans might still be of value by reference to 

 this secondary base. 



Melbourne seemed best suited for the purpose as Hecker and Alessio had only 

 recently redetermined its gravity constant. With the permission of Mr. Baracchi, the 

 same spot in the Astronomical Observatory was used as for the observations by Hecker, 

 and the instruments set up late in March. Unfortunately, for several days it rained 

 almost continuously and no opportunity occurred to get star observations until 

 March 29th. 



Observations were started on the morning of March 31st, immediately after re- 

 winding the pendulum clock " S.C.", and were concluded on the evening of April 3rd. 



The method of observation was essentially similar to that used at Wellington. 

 The induction coil was furnished this time by the kindness of Professor Lyle. The 

 same helium tube was still in use. The distance from mirrors to scale was 280 cms. 



The pressure was read on barometer Hicks 1523 (correction+0-015 in.), and the 

 correction for aqueous vapour calculated from the readings of a wet and dry bulb 

 thermometer. 



The flexure correction was observed three times and gave the following results : 



Nos. 5 and 7. No. 21. 



March 31st .. 26- 8xlO- ? sec. 32-3X10" 7 sec. 



April 1st .. .. 26-8 



2nd .. .. 27-3 33-9 



3rd 34-7 



Mean .. .. 27-0 (0-2) XlO~ 7 sec. 33-6 ( + 0-8)xK)- 7 sec. 



The temperature was measured, as at Wellington, by thermometer No. 29110, 

 and the same corrections were again applied. The total range of temperature in the 

 dummy pendulum was from 16-31 to 16-88 C. and the maximum change of tempera- 

 ture during the 35 minutes each pendulum was under observation amounted to 0-05 C. 

 in that time. 



Time and Kate. For the observations " S.C." was. as usual, used. It was him- 

 from an iron peg driven into the brick wall and kept rigid by large blocks of wood on 

 each side. 



Comparisons were made with the standard clock, " Frodsham 991," before and 

 after each set of swings, by means of a tape chronograph. By taking the mean <>!' 

 about 60 single measurements on the tape, the comparative rates of " S.C.'' and " Fn id- 

 sham 991 " were deduced from the three-hour intervals over which the observations 

 extended. From these, the comparative daily rate of " S.C." on " Frodsham !!)! 

 was calculated, and thus the rate of "S.C." found from the comparison of star observa- 

 tions with No. 991. 



The data are put down in Table LIII. It was found that, at the beginning of the 

 observations, the rate of " S.C." was very irregular owing probably to the recent winding 



73 



