OF ARTS AND SCIENCES. 275 



2G2, although the entire stay upon the mountain little exceeded 

 twenty-four hours. 



The most complete series of measurements relates to the compara- 

 tive altitudes of Gunstock and Joe English Hill. Table 11. gives in 

 successive columns, first, the Greenveich time in houi'S and tenths ; 

 secondly, the height of the barometer reduced to 32° F. by means of 

 the reading of the attached thermometer ; and, thirdly, the tempera- 

 ture of the outer air in Fahrenheit degrees. These quantities are 

 taken from curves drawn to represent the observations which were 

 made at intervals during the day. The fourth column gives the zenith 

 distance of the sun in degrees uncorrected for refraction. The next 

 column gives the apparent height of Gunstock above Joe English in 

 seconds of arc, being the observed micrometric distance multiplied 

 by 13".9.5, the value of one division in seconds. From the assumed 

 heights of the two hills and the apparent difference in altitude, the 

 coefficient of refraction has been computed by Mr. Edmands, and is 

 given in the last column but one of the table. A smooth curve was 

 constructed wi'Ji these values, and the last column gives the residual 

 found by subtracting the result given by the curve from that given by 

 observation. Any error in the assumed heights of the three points 

 would alter the coefficient of refraction by an amount constant through- 

 out the table, but would not affect the amount of its variation. 



Several important conclusions may be drawn from Table II. The 

 angular interval between the two hills gradually increased, and with it 

 the coefficient of refraction gradually diminished, during the first after- 

 noon. The changes were small during the last half-hour, notwithstand- 

 ing the cooling of the air in the upper regions of the atmosphere, and 

 other important meteorological changes caused by sunset. The follow- 

 ing morning the refraction was much less, increasing during the morn- 

 ing, and becoming somewhat irregular during the day. These changes 

 are much greater than the errors of observation, as is shown by the 

 accordance of successive sets, especially on the afternoon of June 22. 



Table III. gives the comparison of the observations of Passacona- 

 way, Whitefiice, and Piscataquog. The first column gives the time, 

 and the second the angular interval between the first and third of these 

 mountains expressed in seconds of arc. Tlie next column gives the 

 corresponding coefficient of refraction, and the fourth gives the resid- 

 ual found by subtracting from this coefficient the corresponding ordi- 

 nate of the smooth curve deduced from the observations given in 

 Table II. The last three columns give the corresponding quantities 

 for the line Whiteface and Piscataquog. 



