DISTRIBUTION IN THE ATMOSPHERE OVER ENGLAND. 261 



How much the results are vitiated it is impossible to say, perhaps not greatly, for 

 the record for the fall seldom differs much from that for the rise. But any comparison 

 between the Continental and English results has got to reckon with this point, 

 because nearly all the English observations were made at or after sunset, and all 

 those on the Continent in the daytime. To this cause I am inclined to ascribe the 

 greater annual range shown on the Continent at heights above 2 or 3 km. 



It may be remarked, in passing, that the two series of hourly ascents at Manchester 

 have afforded the best possible practical test of the accuracy of the instruments. If 

 we assume that the temperature at any height at any hour, 10 a.m. say, was exactly 

 midway between the temperatures at 9 a.m. and 11 a.m., then, in general, the 

 instruments record the true temperature within about 1 C. If two separate 

 instruments of any kind are sent up, differences of temperature at a definite height 

 may be due to errors in the observed pressure, since the height can only be obtained 

 by means of the pressure. In some cases the instrument carries two thermographs 

 writing on the same sheet, and in this case the differences for a definite height are 

 solely due to the thermographs. 



Temperature and Barometric Pressure. 



The relation which undoubtedly exists between the barometric pressure and the 

 temperature of the air at various heights has been investigated by arranging the 

 observations in six groups, and taking the means. The pressure at the place and 

 time of starting the balloon, reduced to sea-level, has been taken as the standard. 

 In certain cases the balloons have travelled long distances. One sent up at Ditcham 

 Park went 800 miles to the E.S.E., and another from Pyrton Hill 660 miles. It is 

 obvious in such cases that the pressure at starting can be no guide to the surface 

 pressure at the time and place where the balloon was at its highest point, and there- 

 fore all such cases have been excluded, 150 miles (240 km.) being taken as the limit. 

 The remaining ascents, which amount to 124, have been arranged thus : 



Group I. Sea-level pressure under 741 mm. (29*17 in.). 

 II. between 741 and 750 mm. 



Ml. 751 756 



IV. 757 762 



V. 763 769 



VI. over 769 mm. (30'28 in.). 



The numbers for each group are 6, 12, 15, 30, 41, 20 respectively, so that, excepting 

 for the group under 741 mm., there are quite enough observations to form a good 

 mean. But a difficulty occurs from the fact that the low pressures are concen- 

 trate.l into the winter months, for extreme values in North-Western Europe 

 only occur in the winter, and no number of observations can cure this. It is 

 then- fore absolutely necessary, if we wish to compare the temperature conditions that 



