338 



Prof. G. H. Darwin. 



[June 19, 



(w.) Results of Reduction. 



Mean water, 8" 54 ft. 



H = 3-98 ft. 

 K = 332 



H = 1-60 ft. 

 H = 0-44 ft. 



o 



K O 



H =: 1'04 ft. 



K = 320 



H = 0-11 ft. 

 K = 240 



H = 1-33 ft. 



* = 53 



H = 0-69 ft. 

 K = 41 



H = 0-44 ft. 



Mean of 9 yrs. obs. 

 8-223 



4-043 

 330 



1-625 

 3 



0-405 

 352 



0-997 

 313 



0-088 

 308 



1-396 

 45 



0-658 



48 



0-404 

 43 



Error of present calc. 



in inches and 



minutes. 



4 in. 



| in. too small. 

 4^ too fast. 



in. too small. 



Nil. 



in. too large. 

 22 m too fast. 



in. too large. 

 ll m too fast. 



j in. too large. 

 2 h 18 m too slow. 



in. too small. 

 32 m too fast. 



^ in. too large. 

 29 m too slow. 



^ in. too large. 

 40 m too fast. 



The second column is given because, if the calculation had beeu 

 conducted by rigorous methods instead of approximately, my results 

 should have agreed very nearly* with these. The causes of several 

 of the discrepancies are explicable. The error of mean water mark 

 is due to the necessity for neglecting the annual and semi-annual tides 

 in a short series of observations. The error in the height of S 2 is 

 partly due to my taking n = 1'034 instead of putting it equal to 1, 

 as is virtually done in the Indian tidal instrument. If I had taken 

 n = 1, I should have had H, = 1'65 nearly. The error in phase 

 in K 2 is a necessary incident of the shortness of the series of 

 observations. The error of N must be due to the shortness of the 

 series, which has not permitted an adequate elimination of the 

 evectional and variational tides. The tide L is only about an inch 

 in height, and accuracy of result could not be expected. 



The magnitude of the error in time in the diurnal tides is disap- 

 pointing, but it is clear that the length of observation has not been 

 sufficient to disentangle the O tide from the Kj tide. It may be re- 

 marked also that an error of 1 in phase makes twice as much differ- 



* I do not know the exact values of the constants used in the Bombay Tide 

 Table, which has been used as representing observation. 



