The magnitude of the flexure correction depends largely on the rigidity of the 

 pendulum stand, but also upon the rigidity of the pillar on which the stand is placed. 

 It depends on the tightness of clamping of the levelling screws and also on the position 

 of the screws themselves. The smaller the distance between the points of the levelling 

 screws and the base of the instrument, the less is the correction for flexure. 



The extreme values noted during these determinations were 16 and 94 units in 

 the 7th decimal place. The smallest of these was measured when the pendulums were 

 swung at Winter Quarters in a small cave dug into the ice. Here, the temperature of 

 the air was very low (from 20 to 30 C.) and the apparatus was set up on a solid 

 block or shelf of ice left unsupported only along one side. The highest value for As 

 was measured at Christchurch. 



The magnitude of this variation (from 16 to 94) might lead one to enquire with 

 what accuracy the calculated value represents the true correction. (The probable 

 error of the determination gives the calculated value from a single series as about 0-6 

 in the 7th decimal place, but, since the formula used is at best an approximation, it 

 is proper to investigate the point raised above.) 



An experimental investigation by Professor Haasemann* answers our query at 

 once. By measuring the actual horizontal movement corresponding to the application 

 of a known force at the point of suspension of the pendulums, he has made independent 

 observations of the flexure correction, and has found that, in the mean, the value 

 obtained by the latter method is greater than that obtained by the use of two pendulums 

 of equal period, by a quantity 0-7xlO~ 7 , and this is not far from the probable error 

 of the observation itself. 



The stations at which pendulums have been swung are, in order of observation- 

 Potsdam, 1910, Christchurch, 1910 (N.Z.), Winter Quarters (Ross Island Series A and 

 B in 1911, series C and D in 1912), Christchurch, 1913, Wellington, 1913, Melbourne, 

 1913, Potsdam, 1913. 



The observations at Christchurch were undertaken in order to have a comparison 

 with similar observations made by L. Bernacchi of the " Discovery " Expedition. The 

 Melbourne observations were made with the object of having a secondary comparison 

 at a spot whose value of gravity was well known. This was necessary in case the period 

 of our pendulums became changed on the long journey back to Potsdam. The Wellington 

 observations were undertaken at the request of Professor Laby, who pointed out that 

 there was no accurate determination of the value of gravity in New Zealand. 



The details of the observations and the discussion of the results will now be put 

 down in detail, in the order Potsdam I and II, Christchurch, 1910, Winter Quarters, 

 Christchurch, 1913, Wellington, Melbourne. 



* " Bestimmung der Intensitat der Sclnverkraft von Hanover," Berlin, 190'J, p. 154. 



12 



