INTER-RELATIONS OF BAROMETRIC PRESSURE. 517 



hourly sea-level pressures are the simple arithmetic means deduced from the 

 observations themselves, which are uncorrected in any way except reduction to 32 o, 0, 

 and for instrumental error, the observations for the summit being reduced to sea-level 

 pressure by Table IV. 



Table II. , Group I. — The number of days available for this group are 3716, the 

 largest monthly number being 403 days for August, and the lowest 259 days for 

 February. The least hourly difference for the year was +0*003 inch at 11 p.m., 

 and the largest — 0*013 inch from noon to 2 p.m. The mean difference for the 

 year was —0*004 inch. This small difference arises from the circumstance that the 

 table of corrections to sea-level of the summit observations (Table IV.) was em- 

 pirically constructed from observations at Ben Nevis and Fort- William from December 

 1883 to December 1887, only such observations being employed when the wind at the 

 summit did not exceed 3 on the scale to 12 adopted at the Ben Nevis Obser- 

 vatory. The above small annual difference of —0*004 inch shows the substantial 

 accuracy of the Table of Corrections for height appended to this Paper, in which 

 Table, as already explained, the effects of seasons and hours are not taken into account. 



The first noticeable feature in this Group is the strongly marked diurnal range in 

 each month. Further, there is an equally well-marked variation through the months of 

 the year. During the five months, November, December, January, February, and 

 March, the mean differences of pressure give invariably the minus sign for each hour of 

 the day. In other words, the reduced means of sea-level pressure at the top give a 

 value continuously less than the mean value of pressure at Fort- William. On the other 

 hand, during the seven months from April to October, the mean differences for the 

 hours show the minus sign only from about 7 a.m. to 7 p.m. As regards the other 

 hours, the mean differences show the plus sign, the maximum being in June, July, 

 August, and October, from 9 p.m. to 5 a.m. The mean maxinum of these hours for 

 June is 0*015 inch, and for July, August, and September, 0*011 inch. These occasional 

 differences are due to the varying humidity of the atmosphere. 



The differences between the highest and the lowest mean values for the months are 

 the following : — 



Inch Inch Inch Inch 



January, O'OIO 

 February, 0-016 

 March, 0-017 



April, 0-022 

 May, 0-024 

 June, 0-024 



July, 0-020 



August, 0-018 

 September, 0-019 



October, 0-014 

 November, 0-010 

 December, 0-006 



Year, 0-015 inch. 



Table II., Group II. — An examination of the figures of Group II. with those of 

 Group I. shows a substantial agreement throughout. This result is what was to be 

 looked for, since Group I. gives averages deduced from all the available observations, 

 and Group II. averages deduced only from those observations which do not in any 

 case differ more than 3°*0 from 15 c *4, the mean difference of the temperatures at the 

 two Observatories. 



