﻿20 



Pressure correction = — ^ ' ; ■^' . where p is the iiressure in millimetres of merctirv at 0' C. and t 



1+0-00367/ ' -^ i ■> 



the temperature inside the receiver. 



The " pressure correction " really depends on the density, and so on the temperature as well as the 



pressure of the air in the receiver ; that is how - — AAop^-i ^''' accounted for. 



There should be an allowance made for the fact that the pressure p is due in a small degree to aqueous 

 vapour, whose density is only | of that of air under the same pressure. But to make this small allowance 

 one requires to know the vapour pressure, and of this we had no note. At low temperatures, however, 

 the vapour present is necessarily very small. 



Fiate correction = - 0' 00001 157EP, where R is the rate in seconds per day, P the period of the pendulum 

 in seconds. The rate correction is negative if E is positive or chronometer gaining. Sidereal time was 

 employed throughout. 



The correction for finite arc of oscillation is given in various forms. That of the United States Coast 

 and Geodetic Survey' is equivalent to 



- 0-01357 (w + »')(/>-'»' )p 

 logio («/"') 

 where P is the period of the j)endulum. 

 Also 



_ semi-arc in millimetres at start 



distance between scale and plane on which pendulum swings in millimetres ' 



, _ semi-arc in millimetres at end 

 distance as above 



The value of one scale-division on the instrument was 3 millims., the distance l)otween scale and mirrors 

 at Winter Quarters about 2858 millims. 



No experiments were made in the Antarctic for flexure coefficient. As it is a function of the pillar, no 

 certain allowance seems feasible. The types of j^iHar used were, however, sufficiently similar to make it 

 uidikely that the neglect of a flexure correction causes any serious error. 



The above formuliB assume 0° C. and millim. as the standard temperature and pressure, as it is easiest 

 to reduce to these even when the mean temperature and pressure of the experiments are widely remote 

 from these \'alues. 



The Melbourne Observations {Norcmljcr 10, 11, 1901). 



The pendulum apparatus \\-as erected in the cellar of the Melbourne Observatory. Swings were taken 

 with two pendulums, Nos. 37 and 39. It was found impossible to stop the leakage of air into the receiver 

 when exhausted, and accordingly the observations were made at atmospheric pressure. The rates of the 

 chronometer 6711 employed in the experiments were determined by direct comparison with the sidereal 

 clock in the transit room of the Observatory. Oliservations with the two pendulums were also made by 

 Mr, P, Baracchi, Government Astronomer of Victoria. 



The Christchurch Obsei;vations, New Zealand. 



Sets of swings were taken at Christchurch Magnetic Ob.servatory in November, 1901, before the sailing 

 of the "Discovery" for the Antarctic, and again on her return in May, 1904. The Oliservatory is in 

 latitude 43° '31' 50" S., and longitude 172° 38' 9" E, ; it is situated on a lai-go alluvial plain (Canterbury 

 Plain), and is 25 feet above the level of the sea. 



The observations were taken in the Absolute Magnetic House, a small wooil building exposed to changes 

 of temperature ; therefore the temperature during the experiments had a large range, and changed 

 vapidl)-. There was no way of obviating this difficulty, no cellar being available. 



