change in the curve from dry to wet seasons and there 

 can hardly be any doubt but that the curve from 1736 ami 

 following years is nowbeiiuj repeated from 1007 onwards, 

 thus demonstrating that the Nile's true period is 17J years. 

 British Rainfall.— The residual mass curve of British 

 rainfall (Plate 1) is derived from the diagram (Plate 2, fig. 2) 

 prepared by the late Mr.G. J. Symons, f.r.s., and referred 

 to by him when discussing Sir Alexander Binnie's paper on 

 "Average Annual Rainfall and the fluctuations to which it is 

 subject." 1 Mr. Symons, whose whole life was devoted to 

 the collection and study of rainfall statistics, was regarded 

 as one of the highest authorities on the subject, as may be 

 seen by the encomiums passed upon the results of his inves- 

 tigations, by those taking part in the discussion on Mr. II. U. 

 Mills' paper on the "Distribution of .Mean and Extreme 

 Annual Rainfall over the British Isles." 2 



The diagram is a very remarkable one, extending from 

 1726 to 1890, covering a period of 165 years. It does not 

 represent the rainfall at any single station, but the mean 

 rainfall over the whole area of Britain, and on this account 

 as might be expected, a difficulty presented itself owing to 

 the paucity of information concerning the rainfall in the 

 first part of the record. If the diagram be carefully studied, 

 it becomes evident that the mean shown by Mr. Symons 

 cannot be the true mean of the figures from which the 

 diagram has been constructed. Instead of it being 100, 

 as represented by him, I find the true arithmetical mean 

 of the whole record to be 95, and this is the explanation of 

 the very peculiar droop of the whole curve, when converted 

 into one showing the residual mass. As the mean for the 

 last 70 or 80 years obtained by Mr. Symons from a large 

 number of well selected and reliable stations is no doubt 

 closely approximate to the truth, I decided to adopt his 

 3 Inst. C.E., Vol. cix, p. 89. » hoc. cit., Vol. clv. p. 2M. 



