558 Prof. G. H. Darwin. 



as much as 3| inches lower in 1885 than in 1883, and I do not believe 

 that the numbers are accurate. 



This opinion is confirmed by even a casual examination of the results 

 of the harmonic analysis at Dover, the observations being obviously 

 bad ; for we rnay, I think, reject the supposition that both the tide 

 and the mean sea level at Dover are actually far more irregular than 

 at any other port. 



In order to test the Dover results, I have found the mean error 

 (according to the method of least squares) of the phases of the several 

 tides from the three years tabulated. I have then rejected as worth- 

 less all those tides in which the mean error of phase amounts to 

 30. By this criterion the tides S,, S 4 , S 6 , S 8 , K 2 , J, Q, T, 2SM, and 

 all the tides of long period are rejected, and many of those retained 

 will be seen to be really very bad. 



Thus the mean error of phase of M 2 is 7'3, and of S 2 , 9'5. The 

 physical meaning of this is, that it is an even chance that the principal 

 lunar high water occurs within a specified 20 minutes of time, and 

 that the principal solar high water occurs within a specified 

 25 minutes. With fairly good observations these periods should, from 

 three years of observation, be about 4 or 5 minutes for the lunar 

 tide, and 8 or 10 minutes for the solar tide. In the case of the tides 

 at New York, tabulated below for three years, it is an even chance 

 that lunar high water occurs within a specified 1J minutes, and solar 

 high water within a specified 6^ minutes. 



The Ostend results were treated in the same way as the Dover ones, 

 and compare very favourably with them, although not, I think, of 

 the highest order of perfection. 



It may thus be safely concluded that the observations at Dover have 

 been very badly made.* 



It is a pity that an expensive instrument should have been installed, 

 and that its records for many years should be rendered valueless by 

 the want of proper supervision. 



I publish the results, however, for what they are worth. 



The phases of the several tides are referred to Greenwich time. 



Ostend. 



I have no information as to the manner in which these observations 

 were taken, but, as stated above, the curves were presented by the 

 Minister of Public Works of Belgium. The Ostend M.S.L. was stated 

 in considering the Dover curves. The zero of the tide gauge is 

 8' 17 feet above the international datum. There were many interrup- 



* Captain Wharton, B.N., is of opinion that the situation of Dover is such that 

 the tides are likely to be irregular there. I cannot, however, believe that this 

 affords a sufficient explanation of the irregularity of the results. May 8, 1889. 



