354 



THE AGRICULTURAL NEWS. 



November 8, 1913. 



It is a matter of common observation that the fall 

 of rain in any locality measured in inches per annum, 

 varies from year to year. A study of the rainfall 

 records extending over periods of many years has 

 shown that as the number of years over which the 

 a\1eraefe is taken increases, the average of the yearly 

 rainfall tends towards a constant quantity, and that the 

 average of thirty to forty years varies but little, what- 

 ever period of thirty to forty years be taken. This rule 

 has been carefully investigated byBinnie, who examined 

 the records of twenty-six stations with records of an 

 average length of fifty-three years. He found, for 

 instance, that in the case of the mean annual rainfall 

 for Kve years, the average positive deviation was 15'3o 

 per cent, of the mean, and the average negative devia- 

 tion 14o2 per cent, of the mean, whilst in the 

 case of the mean for thirty-tive years, the average 

 positive deviation was only 1'73 per cent., and 

 the average negative deviation 18(5 per cent., these 

 latter two figures closely approximating to the 

 same thing. The stations at which the observations 

 were made were distributed over a large portion of the 

 globe, and the results may be considered applicable 

 to all insular climates,, and, indeed, to most other 

 qlimates, except those of a very extreme continental 

 type. Binnie's figures also indicate that the annual 

 deviation may slightly increase if the period exceeds 

 thirty-tive years; for, whereas the average deviation for 

 thirty-five years was found to be 179 per cent, of the 

 mean annual rainftvll, the deviation for forty years was 

 found to be 2-16 per cet^t. For a forty-five-year period 

 it was determined as 2'03 per cent., whilst it dropped 

 at the end of a fifty-year period to 1 98 per cent. 

 It would appear, then, that there is a cycle in rain- 

 fall, and that this cycle (which refers not to indi- 

 vidual rainfalls, but to the average of three or four 

 consecutive years) seems to have a period approximate- 

 ly equal to thirty-six years. The mean annual rainfall 

 may therefore be defined as the average taken over 

 a sufficiently lengthy term of years to ensure a fairly 

 constant value, and we may assume that thirty to 

 forty years is generally an adequate period. The 

 value of this rule will be readily apparent. If we are 

 in possession of a rainfall record of, say, ten to twenty 

 years, for a place which displays a fairly close relation- 

 ship with another where a longer record exists, it is 

 only a matter of simple proportion to amve at a fairly 

 accurate idea of the long-period rainfall at the first 

 locality. 



Turning to the study of yearly records in respect 

 of their absolute magnitudes, no definite relationship 

 has been found; but if, again, these be reduced to per- 



centages of the mean annual rainfall for each locality, 

 a very striking regularity will be seen. Sush investi- 

 gation has shown that the rainfall of the wettest of 

 a long series of years is about 14(i per cent, of the 

 mean, whereas the rainfall of the driest year is approxi- 

 mately 66 per cept. of the mean. 



According to Parker, if the annual rainfall, as 

 defined above, be taken as 100, the following figures 

 deduced from records for the British Isles and 

 N.W. Europe, may be regarded as very close to the 

 truth for all typically insular climates: wettest year 

 14-5 to 148; average of two consecutive wettest years, 

 l.SO to 183; average for three consecutive years 

 123 to 12(j; average for three consecutive driest years 

 78 to 7.j; average of two consecutive driest years, 73 to 

 (i6; driest year, 66 to 61; maximum number of consecu- 

 tive years with a fall above the mean, 5"52 to 3'80; 

 average fall of these years. 117 to 123; maximum 

 number of consecutive years with a fnll less than the 

 mean, o'.57 to .540; average fall of these years, 84 to 83. 



Exceptions to these rules occur in places where 

 the average rainfall is much less than 22 inches yearly, 

 or where the geographical situation is exceptional. 

 According to a communication recently received from 

 Mr. Parker, the West Indian Islands appear to be 

 exceptional so fat as the records available can show. 

 The rainfall of Trinidad is normal, but that of Jamaica 

 and the adjacent islands is quite abnormal, the figures 

 being what might be expected in a continent about 

 1,000 miles from the sea. There appears to be a 

 gradation from Trinidad to Jamaica just like what 

 would occur in working inland over a continent. 



It has to be remembered, however, that there are 

 few rainfall records in the West Indies covering a 

 longer pi'riod than twenty years, the mean of which 

 is scarcely within the limit requisite for the applica- 

 tion of the rules enumerated above. Tt will be 

 admitted nevertheless, that the whole subject is ot the 

 greatest interest and deserving of renewed investiga- 

 tion. The laws of variability of rainfall are also of 

 the greatest economic importance, for in hydraulic 

 engineering projects the rules that have been referred 

 to are commonly employed. Furthermore, the very dis- 

 covery of them is useful, since it shows the importance of 

 keeping accurate records, and what can be deduced from 

 figures which an; frequently regarded as being of merely 

 passing interest ftnd value, but which in reality are grad- 

 ually unfolding, year by year, some hidden tendency or 

 rule of nature. 



