MSTUIBUTION IN THE ATMOSPHERE OVER ENGLAND. 



255 



notice that the temperature of the upper air over England is largely dependent upon 

 the height of the barometer, and that above 10 km. the temperature is far more 

 dependent upon the barometer than it is upon the season. If the number of observa- 

 tions were great enough in each month, variations due to the barometric height 



TABLE I. Mean Monthly Temperature at each even Kilometre. 



A. Mean height of the barometer in mm. 



B. Mean of the temperatures at the highest point. 



The temperatures are in absolute measure with the first 2 omitted. 



would cancel out, but the column giving the mean of the barometer readings at the 

 times of the observations shows that they do not do so. The first step therefore is to 

 correct the monthly means for the height of the barometer. This has been done by 

 means of Table VII. , the formation of which will be subsequently explained. 



The variation in the monthly barometric mean is small in England, and the values 

 for the summer and winter are practically identical, hence the values in the table are 

 chance variations. The values given in Table VII. have been plotted for each height, 

 and from the curves so found corrections have been applied for each height in each 

 month so as to reduce the values to a standard value of 760 mm. at ground level. 



The corrected values are shown in Table II., and are considerably smoother than 

 those of Table I., although it is obvious that some of them, notably those for May 

 and December, cannot be accepted as correct. 



The number of observations for each month, except for February and September, is 

 over 10, but the 13 observations in May are concentrated on two dates, May 5th 

 to 7th, 1909, and May 18th to 20th, 1910, and it chanced that very unusual but 

 like conditions prevailed at both these periods. A similar remark holds for December, 

 for which there are 16 observations, but 14 of them refer to the first week of the 

 month in 1909. The temperatures shown in September are also discordant above 

 (' km., but this is not surprising owing to the paucity of observations. 



