OF ARTS AND SCIENCES. 



13 



TABLE IV. — Air. Fifth Series. 



to be in this group some considerable constant error, clue jirobahly 

 to a mistake in the reading of the themometer which was a different 

 one, from that used in the preceding groups. The deviation is so 

 large that I have thought it best to reject the sixth group wholly in 

 the computation of the empirical equation for this series on air. I 

 fully recognize that this rejection may be considered somewhat arbi- 

 trary, and also that the retention of tlie sixth group would make the 

 accordance between the fourth and fifth series greater than now ; but 

 I am so fully satisfied that there is some mistake in the rejected group 



