Lyons — The Distribution of Mean Annual Rainfall. 361 



a lower level. A standard height of one foot above ground-level has been 

 recommended ; and it is estimated, on the results of many tests, that for every 

 extra foot up to about 10 feet the gauge loses 1 per cent, in catch. A 

 correction on this estimate is made by Dr. H. R. Mill in liis rainfall reports 

 on the Englisli counties. This correction seems, however, legitimate only 

 if the other conditions of exposure are quite uniform ; and, for many 

 reasons, it was considered undesirable to adopt it in the present investi- 

 gation. The two most important records, namely, Fitzwilliam Square' 

 and Fassaroe, Bray, would have needed modification. The gauge at 

 Fitzwilliam Square is placed in a confined area where there is great pro- 

 tection from wind, tliough little likelihood of the gauge being protected 

 from the rain. This fact probably explains why the record at tliis station 

 is almost uniformly greater than that kept at Trinity College, where 

 the gauge is placed (in the Fellows' Garden)^ at a lower level, but in 

 a more exposed position. With reference to the Fassaroe record, where the 

 gauge is kept at 4 feet 6 inches above ground-level, I am informed by the 

 observer, Mr. R. M. Barrington, m.a., that the "catch" at this level was 

 generalljr greater than that found for comparison gauges placed lower. 



The position of a gauge with reference to rising ground is as important as 

 the actual height above the ground. A gauge placed on a summit or slope 

 is particularly exposed to the wind effect, and the catch will be too low. An 

 example of this was found in the record at Tithewere, Roundwood, County 

 Wicklow. The mean annual value for this station at 1000 feet appears too 

 small in comparison with the numerous stations near it; and the deficiency is 

 very probably due to the gauge being placed on the windward slope of a hill. 



The measurement of rainfall when all circumstances are considered is 

 found to be necessarily only approximate. According to Dr. H. R. Mill,' no 

 one could hope to measure the annual rainfall of a place to a closer approxi- 

 mation than 0'5 inch; and it is usually recognized that the estimate of mean 

 annual rainfall is only correct to within 2 or 3 per cent. The rainfall values 

 in the tables are in consequence quoted only to the first decimal place. 



The methods adopted for estimating the average number of rain days 

 per year were in all respects similar to those employed for determining 

 the mean annual rainfall. The limits of error in the record of rain days 

 appear to be greater than in the case of rainfall. A rain day is defined 

 as one during which "01 inch or more is recorded. When more than half 



' The gauge at Fitzwilliam Square is quoted in Table I. as at 1 ft. 6 inches above ground. It 

 stands on a little mound which raises it about 4 feet above the general ground-level. 



2 I am indebted to Sir John Moore, M.D. (who is responsible for the gauge at Fitzwilliam 

 Square), for his valuable opinion on this point. 



5 Minutes, Proc. Inst. Civil Engineers, vol. civ., p. 373. 



SCIENT. PROC, E.D.S., VOL. XII., NO. XXX. 3n 



