286 REPORT— 1866. 



With tlie double object of cheeking these experiments in the upper por- 

 tions of the Thames valley, and also of ascertaining how serious were the 

 errors incidental to a rain-gauge on board ship, a gauge was, by the cour- 

 tesy of the Elder Brethren of the Trinity House, placed on board the Nore 

 light-ship, and its indications have been checked against the returns from 

 the land stations at Sheeruess, Shoebuiyness, and llochford. The result 

 again seems to be a less amount in mid-stream, but neither set of experi- 

 ments have gone on long enough for generahzation. 



First Approximation to Fluctuation of Rain full, 1726 to 1865. 

 In Mr. Symons's last annual Report he gave " a few particular's respecting 

 the rainfall of the last fifty yeai-s," and exhibited the smaller of the two 

 diagrams accompanying this Report * ; that investigation he expressly stated 

 was temporary only, and if he had consulted only his own inclination it 

 would have remained for a few years longer the solitary examination of the 

 question ; his reason being that the importance of the question of the secular 

 variation of rainfall is such that temporary and partial investigations are 

 unworthy of it. On the other hand, no sooner were the previous data pub- 

 lished, than he was urged to give some approximations to the truth for pre- 

 vious years, to see if older records gave any indication of droughts equal to 

 those of the ten years ending with 186.5 — to see, in short, what facts can be 

 discovered in this hitherto untrodden branch of inquiry. He thought that 

 if he persisted in maintaining that this second series, like the first, was but 

 a partial discussion of the question, curiosity would be partly allayed, and 

 there would be less occasion to hurry on the final work, he now therefore 

 submits the results of the following process. 



(1) Eight of the longest and most complete registers were selected, most 

 being for periods of over fifty years, viz. — 



Southwick, near Oundle, Northamptonshire (George yrs. 



Lynne, Esq.) 1726 to 1739 14 



Lvndon, near Oakham, Rutland (T. Barker, Esq.) . . 1737 to 1798 62 



Chatsworth, Derbyshu-e 1761 to 1813 53 



Manchester, Lancashire (Dr. Dalton) 1794 to 1840 46 



Derby, Derby (Mr. Swanwick) 1809 to 1835 27 



Eadcliffe Observatory, Oxford (the Radcliffe observers) 1815 to 1865 51 



Cobham Lodge, Sui-rey (Miss Molesworth) 1825 to 1865 41 



Mean of ten stations as given in last Report 1815 to 1865 51 



(2) The mean of each set of observations having been taken, the fall in 

 each year was divided thereby, and thus its i-elation to the mean of the 

 whole period was converted into a ratio independent of the amount of fall. 



(3) These values being therefore strictly comparable, have been tabulated 

 in parallel columns, and in order to connect all together, and to reduce all 

 to one unifonn base, the means were taken of each overlapping portion, and 

 the corrections thereby deduced having been applied, the ratios were thus 

 reduced to the following bases : — Mean of 51 years at Oxford ; mean of 31 

 years at the ten stations quoted in last Report ; and mean of 41 years at 

 Cobham Lodge. 



Of course identical results could not be expected from stations so widely 

 separated, now that we know how rarely is a year uniformly wet or dry 

 throughout the kingdom ; but Table III. showed results so similar, that he 

 felt no hesitation in constructing the large diagram, showing a " first ap- 



* See Brit. Assoc. Eep. 1865, p. 202, for copy of small diagram. 



