only one year (1751) in the first 25 years above the mean, 

 and one (1763) in the next five years above it. Twenty- 

 eight years out of thirty years below the mean ! We have 

 seen, in the case of the British rainfall curve, this peculiar 

 feature of declining rainfall repeated three times at regular 

 intervals of 57 years. After the year 1909, which is three 

 times 57, or 171 years after 1738, if the periodicity is to be 

 maintained, the decline above referred to should commence 

 to be repeated. The question is with what intensity, as 

 the curve shows that it varies, being greatest after 1738, 

 less after 1795, and least after 1852. 



We shall not be long in doubt, however, for the behaviour 

 of the Nile curve should afford a true indication of what 

 may be expected in England and elsewhere. Unfortunately, 

 I have no information at present to enable me to trace the 

 Nile curve beyond 1905. If it should be found to be falling 

 still, we may reasonably expect a repetition of the meteor- 

 ological conditions at each place, similar to those indicated 

 by the curves after the year 1795. On the other hand, 

 should the Nile curve be rising steadily, then we may 

 anticipate with a fair degree of probability that the 

 meteorological conditions which existed after the year 

 1738 will be repeated, in which case there may be a most 

 serious decline of British rainfall for 25 or 30 years; while 

 the Nile will indicate abundant rainfall over the Abyssinian 

 plateau during exactly the same period. 



Note.— This was written before I had received the in- 

 formation concerning the behaviour of the Nile subsequent 

 to the year 1905. It will now be seen from the diagram 

 that the Nile's curve since 1907 is on a strong upward 

 gradient, similar to that shown after the year 1736, and 

 we may therefore reasonably anticipate that the very 

 serious decline shown in the British rainfall curve will be 

 repeated, and should commence in the present year 1910. 



