was compared with the Wellington Observatory standard on Feb. 28th, and on March 

 1st and 3rd. From these comparisons, " S.C." is found to have had the following 

 losing rates : 



Feb. 28th-March 1st 2 82 sec. daily. 



March Ist-Marck 3rd 2-45 ,, 



The Christchurch time observations gave the following losing rates : - 



Feb. 28th-March 1st .2-78 sec. daily. 



March Isfr-March 2nd 2-45 



March 2nd-March 3rd 2-50 ,, ,, 



The agreement is well within the limits of probable error of the latter observations. 

 From comparison of " S.C." with the two portable chronometers, there is evidence 

 that the latter two chronometers were keeping a consistent rate, but that " S.C.'s " 

 rate was very irregular on the 28th and on the morning of the 1st March, no doubt owing 

 to the fact that " S.C." was wound just before the commencement of the series of obser- 

 vations. For this reason, it has been deemed advisable to discard the first pair of 

 observations entirely and to apply the rate of 2-45 sec. losing daily to the four last 

 observations. 



The value for flexure was observed four times for pendulum No. 21, and three times 

 for Nos. 5 and 7, and is given below. 



Nos. 5 and 7. No. 21. 



Feb. 28th 33-9 XlO" 7 sec. 62-8X1Q- 7 sec. 



43 



Mar. 5th 51-7 45-7 



48-7 46-8 



Adopted values . . . . . . 50 46 



The values obtained were very variable, and it was decided to adopt the values 

 of March 5th and to neglect the observations made before the series commenced. This 

 point should be borne in mind when discussing the cause of the anomalous results 

 obtained at Christchurch. 



The temperature was measured by thermometer No. 29110, in the dummy pendulum 

 resting on the cork base. The temperature was very steady throughout the observa- 

 tions, remaining constant between 15-33 C. and 15-89 C., with a maximum variation 

 at the rate of 0-06 per hour. 



The same barometer, read at least once every 40 minutes during the period of 

 observation, was used as in the 1910 observations at Christchurch. 



Humidity corrections were applied from the readings of the pair of wet and dry 

 bulb thermometers. 



58 



