308 E. D. Preston — Deflection of the 



possible being observed on successive nights, the errors in right 

 ascension were also in a great measure eliminated. 



Six swings were made with the heavy end down. The pen- 

 dulum was then revolved 180° around its longitudinal axis, 

 and six more swings were made. It was then reversed end 

 for end, and the same number of swings made with the heavy 

 end up. This completed the work with one pendulum. The 

 other was observed similarly, and stars were obtained every 

 night — the mean epoch, as usual, being at the time of begin- 

 ning a new series of oscillations with the pendulum. 



The time was determined with a meridian telescope — the 

 transits of stars as well as those for the pendulum being regis- 

 tered on a Fauth chronograph. Two small cells supplied the 

 electric circuit in which were placed the chronograph, a side- 

 real chronometer, and two observing keys. The chronometer 

 broke the circuit every two seconds. 



The knife-edge plane of the pendulum head was made level. 

 The precaution was taken to hang a weight equal to the pen- 

 dulum on the projecting arm during the operation, to see that 

 no appreciable deflection was caused. A similar test was made 

 for the whole support. The barometers were carefully com- 

 pared with Signal Service standards. The thermometers 

 which were made by Baudin, and can be read to hundredths of 

 a degree centigrade, had their corrections determined in this 

 country both before and after the expedition. The zero point 

 of one of them was determined in the islands, and they were 

 compared on the mountain immediately after the experiments 

 were concluded. 



Although Haleakala is in the tropics and was occupied in 

 midsummer, we found great ranges of temperature on the 

 summit. Ice was formed at night to the thickness of quarter 

 of an inch. The clouds do not usually come within several 

 thousand feet of the top, which gave us a clear atmosphere for 

 the astronomical part of the work. Indeed many stars were 

 observed before sundown with a telescope of 2-J- inches aper- 

 ture and magnifying power of 70. One low star for azimuth 

 and four high ones for time were taken before reversing the 

 telescope. The same program after reversal, completed the 

 evening's observations — the usual level readings being made as 

 often as possible during the work. 



It was found impracticable to carry the pendulum stand to 

 the summit ; but a better natural support was found, and the 

 pendulums were swung in a cavity between two rocks. This 

 opening was closed behind by masonry laid in cement, and be- 

 fore by tarpaulins, blankets, and dry masonry. The room thus 

 formed was about nine feet high by 3| feet in diameter and 

 gave a quite uniform temperature. 



