512 DR BUCHAN ON THE 



For 26th March 1903 this Table shows the sea-level pressure for the summit 

 continuously lower than at Fort- William, the difference for the mean of the day being 

 — 0*024 inch, falling to a maximum difference of — 0'048 inch at 3 p.m. and rising to 

 a minimum of — O'OIO inch at 8 p.m. The mean depression of the wet bulb for the day 

 at Fort- William was 2° - 4, the greatest depression, 4° "4, occurring at 7 a.m. and the least, 

 o, 7, at 7 p.m. The mean temperature at Fort- William for the day was 44 c, 7, and at 

 the summit 27°'9, thus giving a mean daily difference of 16°'8, the maximum difference 

 being 19°'4 at 11 a.m. and the minimum 13°'7 at 1 a.m. Much larger differences 

 of temperature, especially in the afternoon hours, are of frequent occurrence, 20° "0 

 being often exceeded. The large differences of temperature on the mean of the day 

 at the two Observatories are generally occasioned by a lowering of the temperature 

 at the top and by a fall in the humidity at Fort- William, not by an increase of tem- 

 perature there, unless when sunshine prevails at Fort- William but none at the 

 summit. It will be noted from this Table that saturation was complete and con- 

 tinuous at the summit during 26th March 1903. Since, then, during the day, 

 saturation obtained at the top throughout, it follows that sea-level pressure for the 

 top was reduced owing to the saturated and therefore specifically lighter air there 

 and at higher levels. But at Fort-William the air was differently conditioned, the 

 observations showing a mean depression of the wet bulb for the day of 2° '4. 



The examination of the inter-relations of the barometric pressure, temperature, and 

 humidity at the two Observatories soon disclosed, as it proceeded, that diurnal and 

 seasonal changes were both well marked. Further, it was also soon apparent that 

 differences of temperature between the two Observatories were closely associated with 

 equally striking differences in the humidities and sea-level pressures, and these in turn 

 were intimately linked to cyclonic and anticyclonic weather. It was accordingly 

 resolved to work out the ' Constants ' of the relations between the two Observatories 

 for the hour, the day, the month, and the year, directly by simple arithmetic means 

 calculated from the observations themselves. This laborious work has engaged most 

 of my time since August 1900. 



The daily sheets of observations examined, of which the Table at page 508 is a 

 specimen, included 4900 days. But no inconsiderable number of these days were not 

 available in the discussion owing to the frequent occurrence of stormy weather with high 

 winds at the top of the mountain, resulting in pumping of the barometer. Since in these 

 cases correct readings of the barometer cannot be made, the days on which they occurred 

 are not included in this inquiry. The number of such days of high winds during these 

 years was 1184, thus leaving 3716 days on which the whole of the observations at each 

 of the 24 hours are available. 



The distribution of these days through the year is detailed in Table II., p. 524. The 

 largest number of the days of calms and comparatively light winds occurred in July and 



