110 



BULLETIN OF THE 



Table II. 



Mean amounts to be added to the true difference of height between the summit and 

 base of Mount Washington to obtain the computed difference. 



As the first two horizontal rows of figures apply only to obser- 

 vations for the month of May, and as it would be very desirable 

 to have results for the colder months when the fluctuations are 

 much increased, I have added a third set of figures for the summit 

 of Mount Washington, compared with the mean of Burlington and 

 Portland as the base, and computed the difference of elevation from 

 observations taken at 7 a. m., 3 p. m., and 11 p. m., Washington 

 time, during January, February, March, October, November, and 

 December, 1880. 



It is evident from Table II that during the prevalence of relatively 

 high pressure, elevations computed barometrically will, in general, 

 be too small, and, on the other hand, when the pressure is low, the 

 computed heights will be too great. This also explains the coinci- 

 dence of too great computed heights with high winds, for the reason 

 that the highest winds always occur with relatively low pressure ; 

 on the contrary, when the wind is light, the pressure is generally 

 high. 



May not this retardation be due to the effect of varying tempera- 

 ture? When a " low" has passed a station at sea level the tempera- 

 ture frequently falls steadily, thus contracting the atmosphere and 

 causing its withdrawal from the upper regions, and a still further 

 fall in pressure there. This process will continue until the fall 

 caused by the low temperature is counterbalanced by the rise due 

 to the advancing "high." The following is given as an illustra- 

 tion : 



