THE METEOROLOGY OF EDINBURGH. 691 



Hence the greatest variability, 4° '2, occurs on January 1, and the least, 1°'9, on August 

 24. Looking at Table XIV. it will be seen that the variability is, broadly speaking, 

 above the annual average from October to March, and below it during the other six 

 months of the year, falling to an annual minimum in July and August. The variability 

 of temperature seems to depend in no inconsiderable degree on the amount of vapour 

 present in the air. When the quantity of vapour is small, the tendency for tempera- 

 ture to change from day to day reaches a maximum, but when the amount is 

 large, the conserving influence exerted by it on the temperature of the air is very 

 apparent. The variability of temperature during the winter is further increased by the 

 prevalence of rapidly moving cyclones and their accompanying anti-cyclonic systems, 

 with their different temperatures and humidities. 



Direction of the Wind. 



The number of times the wind blew from each direction has been determined for each 

 day of the year on the mean of 100 years (see Tables XV. and XVI.). The observa- 

 tions utilised were taken from 1770-79 and from 1800 to 1894, with the exception of the 

 years 1809, '10, '37, and '38, for which we have only monthly summaries. The observa- 

 tions utilised were those contained in registers III., VII., IX., XIV., XVII., and XVIII. ; 

 the hour of observation varying from 8 to 10 a.m., the direction being usually observed to 

 eight points. For some years the direction was given to sixteen points, but the values 

 were resolved to eight points by halving the eight intermediate points between the octants. 

 No attempt has been made to calculate the mean direction from the observations, 

 Lambert's formulas yielding results that are obviously erroneous unless the air movement 

 from the different octants is known. 



The seasonal variation in the percentage frequency of the various winds is well 

 marked. Perhaps the most interesting feature is the difference between the W. and 

 S.W. winds, the former of which are at a maximum in August, and the latter in 

 February. This is probably due to the fact that in February the dominating factor is 

 the low-pressure area located over Iceland, while in August the lowest barometer is 

 found over India. 



Rainfall. 

 Table XVII. gives the total rainfall collected during eighty-eight years on each day of 

 the year, the series being a composite of several registers. From 1770-76 the observa- 

 tions were made at Hawkhill House by James Hoy, who also kept a gauge during the year 

 1780. From 1785 to 1817 the rain tables given monthly in the Edinburgh Magazine 

 (afterwards the Scots Magazine) have been utilised. The station was " near the foot of 

 Arthur's Seat" till 1793, "within one mile of the Castle of Edinburgh" from 1794 to 

 1798, and thereafter at Barnton, three and a half miles west of Edinburgh, till 1817. The 

 values from 1824 to 1831 were obtained from the meteorological register kept at Canaan 

 House by Alex. Adie, optician, and F.R.S.E., the daily observations being published in 

 the Edinburgh Journal of Science. Since the establishment of the Scottish Meteoro- 



