1915.] BAILIWICK RAINFALL. 223 



Science has sought long and so far in vain for regularly 

 recurring cycles of weather — cycles, for instance, of wet and 

 dry years chat might be foretold with certainty. Apart from 

 this failure the belief grows on me — there being no proof to 

 the contrary — that, however irregularly, the rainfall balances 

 itself from time to time — that periods of drought are followed 

 by periods of wet, and rice versa ; that deficiencies are in due 

 course made up and surpluses lost. These fluctuations above 

 and below par may become excessive and the balance of nature 

 on occasions take years to redress itself. But that it does so 

 and that series of wet and dry years are not indicative of a 

 changing climate but rather of a process of compensation 

 seems to me more probable than not. The great difficulty 

 appears to be that of solving the normal quantity for any 

 district or place, for the most that can be claimed for averages 

 is that the figures merely represent the normal of the particu- 

 lar years dealt with — nothing more if nothing less. 



To theorise on a balancing of the weather in the element 

 of rainfall, years must be studied collectively, not individually. 

 Twenty-two years (the period covered by the Les Blanches 

 records) is far too short an interval in which to hope to dis- 

 cover the truth or otherwise of the theory. But in support of 

 the argument and on the supposition that the two series of five 

 years each referred to above, viz., 1898-1902 and 1905-1909, 

 were really periods of marked rainfall shortage, w r e are 

 brought face to face with the fact that four of the succeeding 

 six years (1910-1915) have had abnormally big totals as com- 

 pared with any in either of the so-called dry intervals. Do tve 

 see in this a restoring to par of a long deficient rainfall ? It 

 appears to me we do and at that I must leave it, merely giving, 

 for reference, the annual totals since 1898. These are split 

 up into three series of six years each in order to show (in 

 confirmation of what has been said) the remarkable rise in 

 the average figures in the last six years. 



1898-1903. 



1904-1909. 



1910-1915. 



in. 



in. 



in. 



32-86 



36-24 



45-54 



31-59 



33-83 



34-74 



34-74 



31-79 



45-55 



26-60 



32-49 



35-09 



32-66 



24-33 



40-07 



38-20 



32-32 



41-82 



Average 32-77 



31-83 



40-47 



