260 EESULTS OF THE METEOROLOGICAL OBSERVATIONS TAKEN DURING 



The curves for each season and the year are shown graphically on Plate II. The 

 annual curve is of a simple character, there being a long-drawn-out minimum, which is 

 most pronounced at 5 a.m., and a well-marked evening maximum, which culminates 

 about 8 P.M. This evening maximum is a conspicuous feature of the curves at all 

 seasons, while the morning maximum is very feebly marked. Only in Spring does 

 the morning maximum exceed the mean of the day. There is little doubt that the 

 curves for Autumn and Winter will be very materially modified, as more years' observa- 

 tions are taken, the enormously deep morning minimum being largely due to the 

 capricious occurrence of barometric minima, which may take place at any time, and 

 which it will take many years' observations to wholly eliminate. Indeed, for the purpose 

 of obtaining an approximation towards the true diurnal range, the year 1903, especially 

 as regards the Autumn and Winter mouths, was one of the worst possil>le owing to the 

 frequency of rapidly moving cyclonic system.s. 



Barometric Pressure at 32° and Mean Sea Level. 



Diurnal Barometric Range in Clear and Cloudy Weather. 



In order to ascertain the diurnal range of barometric pressure during clear and 

 cloudy weather, the hourly values of pressure have been examined for the twenty-one 

 months ending 31st December 1904. 



For clear weather, days on which the mean cloud amount was 3'0 or less were selected 

 for the Spring, Autumn, and Winter seasons ; but as regards Summer, there was not a 

 single day on which the amount of cloud fell below 6' 5, so that the ten clearest days were 

 taken for this season. The results for Summer are thus not quite comparable with those 

 taken during other seasons of the year. For cloudy weather, days on which the sky 

 was overcast at every hour of the day were selected. It is further to be noted that all 

 days on which high or squally winds prevailed were rejected owing to the disturbing 

 effect of the pumping of the barometer. The following values have been corrected for 

 " midnight differences," and further smoothed by continuous three-hour groups. Thus 

 the mean at 2 a.m. was derived from the average of 1, 2, and 3 a.m., the mean for 

 3 A.u. from 2, 3, and 4 a.m., and so on for each hour of the day. The following tables 



