8— ON PREDICTING TIMES OF HIGH WATER AT 



DURBAN, NATAL. 



By R. Fermor Rendell, B.A., F.R.A.S., of the Observatory, 



Durban, Natal. 



[Abstract.] 



The Paper contains : — 



{a) Remarks on the subject, drawing attention to its 

 importance and urgency. The necessity for improved 

 predictions is demonstrated, the only Table hitherto 

 available being frequently more than an hour and a 

 half in error. The writer has devised a method which 

 yields good results. 



{b) An ephemeris for the whole of the year 1906 is added, 

 giving the times of High Water at Durban, predicted 

 by the new method. 



{c) A comparison is shown to prove the superiority of the 

 new method. In this comparison the following daily 

 particulars are given in five columns, commencing from 

 September ist, 1905, and including the latest records 

 to hand at time of writing : — 



(i) The predicted times of High Water given in the 

 Natal Directory. 



(2) The predicted times calculated by my new method. 



(3) The observed times supplied by the Engineer of 

 the Natal Harbour Works. 



(4) The discrepancies between the predictions in the 

 Natal Directory and the observations. 



(5) The discrepancies between my predictions and the 

 observations. 



The results show that the new method is vastly superior to the 

 old. About half of the observed times differ by more than three- 

 quarters of an hour from the old predictions (which are sometimes too 

 early and sometimes too late), while the majority of the new predic- 

 tions are within a quarter of an hour of the observed times. 



For the period of nine months, commencing September ist, 1905, 

 the average discrepancies are 46 minutes and 14 minutes respectively. 

 During this period 495 observed times were investigated. 



The following summary indicates the nature of the results : — 



Old Predictions. New Predictions. 

 Number of discrepancies over 90 minutes 27 O 



,, 45 ,, 245 12 



,, ,, not over 15 ,, 68 321 



By investigating a longer series of records, the writer hopes to 

 be able to deduce still more accurate values for the quantities in his 

 tables. 



