538 MESSRS OMOND AND RANKIN ON THE 



thus throwing the preponderance of frequency to the southward, but only by a small 

 percentage. In either case, taking the observations as they stand or as grouped, the 

 relative frequency is totally different from what it is at sea-level. For comparison, 

 Tal»lc III. has been drawn up giving the percentage frequency of each wind for all 

 Scotland, as calculated by the Registrar- General from observations at 55 stations in 

 Scotland. This table is for the same six years (1884-89), and it agrees very closely with 

 a similar table prepared by Dr Buchan from twenty- three years observations (see Jour. 

 Scott. Met. Soc, vol. iii. page 293), from which we may conclude that the distribution 

 of winds during these six years was normal. At sea-level the most frequent wind is 

 W. ; and S.W., W., and N.W. include nearly half of the total observations — more 

 than half if we exclude calms. This latter order of winds is in strict harmony with 

 the distribution of barometric pressure over the British Isles, according to Buys Ballot's 

 law, by which the winds blow counter-clockwise round areas of low pressure, such 

 an area lying to the north of the British Islands ; but the Ben Nevis winds do not 

 fit in with such a distribution of pressure at all, and we are forced to the conclusion 

 that isobars drawn at the level of Ben Nevis (4400 feet) would lie in quite different 

 directions from those at sea-level. In other words, the distribution of average barometric 

 pressure which extends over the North Atlantic and North- Western Europe, and dominates 

 the surface winds over that area, does not in this country extend to a vertical height 

 of one mile. To make certain that this startling difference in the wind directions was 

 not due to any difference in the method of observing on Ben Nevis, we have, first, 

 taken the Ben Nevis winds as observed at 9 a.m. and 9 p.m. only, the hours at 

 which almost all of the observations on which the Registrar-General's reports are 

 founded are taken ; and, second, eliminated all winds of less than 1 miles per hour, 

 so as to exclude shifting breezes that might play about the sides of the hill, and 

 be merely local in character. As in neither case did we find any serious difference in 

 the relative frequency of the different winds, we may conclude that the Ben Nevis 

 winds, as given in Tables I. and II., really represent the average direction of air 

 motion at the level of the hilltop ; and, as these six years are normal in their 

 wind observations at sea-level, presumably they closely approximate to the true 

 average for Ben Nevis. Individual observations show similar differences between the 

 Ben Nevis and sea-level winds. Thus, if a cyclonic storm of small area is lying 

 to the north-eastward, the sea-level winds are W. or N. W., but the Ben Nevis 

 wind may be N.E., blowing straight out from the centre of the low-pressure area. 

 In larger storms, the Ben Nevis winds are practically the same as at sea-level, 

 indicating that in the latter case the disturbance extends to a greater height as well 

 as covers a larger area. But it is remarkable that, with a cyclone covering perhaps 

 Scotland, the North Sea, and Southern Norway, and having the usual circulation of 

 wind round it, the wind at such a low elevation as Ben Nevis should be utterly 

 different. This outflowing wind seldom or never occurs when the centre of the storm 

 is to the S. or W., but only when it lies to the N. or E., and is most marked 



