262 



MB. W. H. DINES ON THE VERTICAL TEMPERATURE 



prevail in cyclones and anticyclones, to allow for the annual temperature variation, 

 and this has been done by means of Table III. It is true that Table III. itself 

 <lr|H-nds on Table VII., and therefore Table VII. ought not to be indebted in any \v,i y 

 to Table III. The corrections might l)e applied by means of Table I. instead of III., 

 but the method of continued approximation seems the more suitable notwithstanding 

 that it suggests reasoning in a circle. The inter-relation of the two quantities 

 presents a difficulty, but it is not serious except for heights between 8 and 12 km. 

 Kadi height in each group has been treated separately, and, using the annual 

 temperature variation for each height taken from Table III., the temperature has 

 been reduced to what it presumably would have been had the ascent occurred at the 

 beginning of May or November, the times of mean temperature for the upper air. 

 As previously stated, both high and low barometric heights are common in the 

 winter, and a comparison between such as are available has been made, and the 

 result is identical with that shown in the table. 



The results are shown in Tables VII. and VIII., and the difference A-C in 

 Table VII. shows the difference of temperature at each height that is found to 

 prevail at times of very low and of very high barometer. Table VIII. shows the 

 gradients in a convenient form. 



TABLE VII. Temperature at each Height for Various Barometric Readings. 



TAHI.K VIII. Temperature Gradients at Times of Low and High Barometer. 



