259 



MR. W. H. DINES ON THE VERTICAL TEMPERATURE 



the monthly values that the dates of the minimum would agree closely with those in 

 columns B., G. & H., if they were available. 



It must be borne in mind that these are independent investigations, excepting that 

 the figures in columns G. & H., and W., are mostly dependent on the same material. 

 The observations in England were nearly all made at or after sunset, so that solar 

 radiation is excluded ; they were also made with an entirely different type of 

 instrument, and refer to a different period. The observations on the Continent were 

 made in the morning ; about 8 or 9 a.m. would, perhaps, be a fair average. (With 

 regard to the greater range on the Continent, see p. 261, line 5.) 



The following facts appear from the tables. The temperature decreases steadily up 

 to a height of about 10 or 11 km. (9 or 10 km. at Pawlowsk), and remains almost 

 stationary above that height. The annual range decreases from the surface up to 

 2 or 3 km. ; it then continues nearly constant, with perhaps a small increase at 

 7 or 8 km., up to about 11 km., at which point it is abruptly reduced to less than half 

 its former value. In the strata above 1 or 2 km. the maximum and minimum values 

 are delayed for about a month, but above the point at which the vertical temperature 

 gradient ceases they come back and occur at the summer and winter solstices. 



This point is higher in summer than in winter ; it is higher in England and the 

 central part of the Continent than at Pawlowsk (59 41' N., 30 29' E.). Up to it 

 the gradient of temperature is much the same in all places and at all seasons, 

 excepting in so far as the larger annual range of temperature in the East produces a 

 modification near the surface. Above it the annual range is nearly the same, and 

 small in all four results. 



In showing the monthly values decimals of a degree have not been given, though 

 they have been used in the work. The observations are not numerous enough to give 



TABLE V. Approximate Gradient in Degrees Centigrade per Kilometre. 



