56 THE ROYAL SOCIETY OF CANADA 



Some illustrations were shown which gave a direct comparison 

 between the harmonic and the projection methods. Prof. Dayton 

 Miller very kindly arranged to have the harmonic analyses for the 

 purpose made in his analysing laboratory at Cleveland. It was seen 

 for estuary tides having steep graphs, that as many as ten Fourier 

 constants which varied from tide to tide might be required to express 

 the curve even approximately — whereas the projection method 

 correlated it with other tides up and down the estuary with the aid of 

 only three constants in each case. 



An account was given of a new method of prediction which may 

 be based on this projection analysis. If the projection relations 

 between the tides at certain stations, A, B, C, and D, all on the same 

 estuary, are known for any one day, and if the tide at A is known for 

 another day, then by projection at angles determined by the known 

 relations it should be possible to determine the tides for B, C, and D 

 on the second occasion. If the variation at A for the season is known, 

 then the tides for the season at the other stations should be deducible. 

 A description was given of the experiments performed to verify this 

 method. 



The writer is greatly indebted to Dr. Bell Dawson and to Dr. 

 A. N. Shaw for their keen interest, for the assistance rendered by 

 furnishing the tidal data and for numerous suggestions. 



McGill University. 



