22 



The flash apparatus was erected on a small box filled with cement, and therefore very heavy, which, in 

 turn, rested upon a larger box on the floor filled with heavy materials. The distance between the scale 

 and the mirrors was 2858 millims. 



ri-i'K.iiire. The pressure was observed by means of a large manometer open to the atmosphere and 

 therefore subject to the variations of atmospheric pressure. 



The manometer was tightly secured to a flat board, and a long millimetre scale of wood screwed alongside 

 the tube. It was then placed upright close to the pendulum case, and the level of the mercury in both 

 ends of the tube was read off by means of a T-square four times during each set of fifty coincidences. 

 A mercury station barometer placed close to the manometer was read off simultaneously. 



The correction to the mercury barometer, - 002 inch throughout, was determined at the National 

 Physical Laboratory. 



The small and very delicate manometer supplied with the apparatus, and which was most carefully 

 packed by Messrs. NEGRETTI & ZAMBRA, London, and sent out to New Zealand by mail boat, was found 



Interior of Observation Hut, showing 

 1. Marine barometer. 2. Observing telescope and flash apparatus. 



on opening the case to be broken. The manometer employed was constructed by Engineer Commander 

 E. W. SKELTON, K.N., the chief engineer of the " Discovery." 



At Kew Observatory, in 1901, some trouble was experienced with the "air-tight" case. At Melbourne 

 and Christchurch, in the same year, it was found quite impossible to reduce the pressure to anything like 

 60 millims., so that the pendulums had to be swung under atmospheric pressure. The whole weight of 

 the extremely heavy stand and case is supported by only three comparatively slender screws, and the 

 metal rim above the screws was " pressed up " and the contact between the rims was no longer perfect. 

 Mr. SKELTON, by means of a surfacing plate, actually found this to be the case. He therefore re-surfaced 

 both rims, which process considerably mitigated the evil. Indeed, it appeared at first as if we should be 

 troubled with only a very small leakage, for when the pressure within the case was exhausted in 



