EEPORT OF THE SECRETARY. 23 



and from them a correction obtained to be applied to the aneroid. 

 The differences appear remarkably regular, and show that the mean 

 monthly readings of the aneroid may be relied on to the hundredth of 

 an inch. The average correction was found to be .022 of an inch, a 

 quantity which, strictly speaking, is composed of two parts, namely: 

 of the true index error of the aneroid, and the specific differences of 

 the two instruments in different latitudes. In the mercurial barometer 

 a mass of metal is balanced against a mass of air, and, therefore, the 

 indications are independent of a change of gravity, for the same reason 

 that a pound weight in an ordinary scale balance is in equilibrium 

 with the same amount of the material weighed, in whatever latitude 

 the experiment may be made. Not so, however, with the aneroid 

 barometer; as this instrument, like the spring balance, indicates the 

 pressure by the reaction of an elastic material, it must be affected by 

 a change in gravity, and consequently varies in its indications with a 

 change of latitude. 



The diurnal variation in the pressure of the atmosphere in the 

 higher latitudes of the arctic regions is very small, and can only be 

 satisfactorily traced by means of the combination of a great number 

 of observations, while the fitful variations in the atmospheric pressure 

 are frequently very large. 



The minimum pressure both at Bafiin' s bay and at Port Kennedy 

 occurred at about half after four o'clock a. m., and the maximum at 

 about half-past seven o' clock p. m. The range of the diurnal fluctua- 

 tion in Baffin's bay is .028, at Port Kennedy .048, and at Rensselaer 

 harbor .010. There is, therefore, a diminution in the range as we go 

 northward, and at the same rate diurnal variations would become in- 

 sensible at about 80° of latitude. 



The average height of the barometer varies in different months of 

 the year; it is greatest in April and May. 



The occurrence of the minimum is not simultaneous at the different 

 stations; at Baffin's bay it was in January, at Port Kennedy in July, 

 and at Rensselaer harbor in September, presenting results which 

 clearly indicate that more observations are required to fix with pre- 

 cision the time at which it really takes place. The annual range of 

 the barometer at Baffin's bay was .44, at Port Kennedy .41, and at 

 Rensselaer harbor .21. 



The relative pressure connected with different winds is also given. 

 In Baffin's bay the north wind gave a maximum pressure; at Port 

 KcDnedy, and also at Rensselaer harbor, the south wind produced 



