284 



PROCEEDINGS OF THE AMERICAN ACADEMY 



The first part of this table shows that no appreciable eifect is pro- 

 duced on the refraction by the varying altitude of the sun. This con- 

 clusion was also derived from the observations on Wachusett. The 

 group corresponding to a zenith distance of 96° consists of a single 

 set, No. 12. The separate settings are discordant, the observation 

 was very uncertain on account of the darkness, and B had become in- 

 visible. A variation in the refraction is obvious when either the 

 barometer or thermometer varies. Low temperatures are, however, 

 accompanied by high pressures, as is shown in the mean values given 

 on page 282. The effect of these two causes cannot readily be dis- 

 tinguished. Probably both act, since both affect the density of the 

 air. The grouping according to both temperature and pressure shows 

 that this correction is a real one, and should be applied, although this 

 IS not customary in geodetic work. 



TABLE VII. — Constants of Instrument. 



A series of measurements was made on August 6 and 7 to deter- 

 mine the steadiness of the constants of the instrument. In Table VII. 

 the successive columns give the date, the Greenwich mean time, the 

 level, the collimation, and the absolute altitude of the point observed, 



