OF ARTS AND SCIENCES. 271 



periods of storm and haze prevented the accomplishment of satis- 

 factory work. 



Some observations were obtained by Mr. Edmands from Mt. Starr 

 King, Coos Co., N. H., in October, 1878. In November and Decem- 

 ber of the same year, Mr. Edmands resided at Arlington Heights, Mass., 

 for the purpose of obtaining similar observations. 



Plans were also made for occupying simultaneously for a week 

 Mts. Moosilauk, Kearsarge (S.), and Wachusett, which lie nearly in 

 a straight line. Observations could thus be obtained at the ends and 

 near the middle of a line nearly a hundred miles in length. 



Much time was lost, owing to bad weather, when stations were occu- 

 pied which were especially adapted to the determination of the atmos- 

 pheric refraction. That a very distant point might be seen, lofty and 

 comparatively inaccessible stations were required. Moreover, in order 

 that observations might be made under the extremes of temperature 

 and other atmospheric conditions, it was necessary that the work 

 should extend over a considerable period of time. A series of meas- 

 urements was accordingly undertaken at the Harvard College Observa- 

 tory, where the absence of very distant points was compensated by the 

 ease with which observations could be made at all times and seasons 

 without materially interfering with other work. The observations 

 were begun on January 3, 1882, and continued until January 26. Dur- 

 ing this series the temperature fell as low as — 15° F. The position of 

 the level was then changed, as described on page 269. A second series 

 extended from May 6 to August 7, 1882, and included observations in 

 which the temperature was -[-91° F. All the observations were 

 made with instrument D, by Mr, Edmands and myself. 



Only the observations made at Wachusett and at the Observatory 

 will be considered in the present paper. 



I. Wachusett. 



The micrometer level was mounted on the sill of the north window 

 of room No. 19, in the second story of the hotel, on the summit of 

 Wachusett. This point is approximately lom. north, 3m. west, and 

 4m. above the United States Coast Survey bolt in the top of the 

 mountain. The effect of the unsteadiness in the support is almost 

 entirely eliminated by the differential character of the observation. 

 The observations were made by directing the telescope towards the 

 mountain to be observed, and turning the screw so that the level-tube 

 was nearly horizontal. The position of the screw, of the two ends of 



