PHILOSOPHICAL SOCIETY OP WASHINGTON. 131 



club to which it owes its oi'igin, a chemist and physicist with 

 unusually comprehensive- knowledge, an exact experimenter, a 

 man of absolute probity and uprightness of character, and of 

 great benevolence of disposition. It is greatly to be regretted 

 that a mental organization of so high an order, combined with 

 such admirable personal qualities, was associated with a feeble 

 physical frame, that wore out when it should have been in its 

 prime. 



Resolved, That the members of the Society tender to the 

 mother and brothers and sisters of the deceased the assurance of 

 their deep sympathy and condolence, and that the Secretaries be 

 requested to transmit to them a copy of these resolutions, which 

 shall also be entered upon the records of the Society. 



Mr. G. K. Gilbert made a communication on 



A special method of barometric hypsometry. 



(ABSTRACT. ) 



In the determination of altitudes by means of the barometer 

 there are two chief sources of error. Where, as is the usual case, 

 it is sought to ascertain how much a barometer read at what is 

 called the new station is higher than another barometer read 

 simultaneously at what is called the base station, there are two 

 important factors of which it is practically difficult to take ac- 

 count. First, since the air is not in a state of equilibrium, it 

 cannot be known that a barometer suspended in the air above 

 the base station, and at the level of the new station, would record 

 the same pressure that is observed at the new station ; and the 

 discrepancy between the barometer at the new station and the 

 barometer in air, measures a factor of the problem, which may be 

 coM^di ih.Q error of gradient. Second, since the temperature of 

 the air is not constant, and since the air contains unequal pro- 

 portions of moisture at different times, there is a variation in the 

 density of the air independent of its pressure ; and the differ- 

 ence between the actual density of the air and a certain as- 

 sumed density, regarded as normal, measures a factor of the 

 problem, which may be called the error of density. 



The error of gradient will not here be considered. 



The error of density is usually sought to be measured by ob- 

 serving the temperature and moisture of the air at the base and 

 new stations, but this method is confessedly one of great inaccu- 

 racy. The observations are necessarily made at the surface of 

 the ground, Avhere the changes of both temperatui'e and moisture 

 are so rapid and great that they are not at all comparable with 



