62 ROYAL SOCIETY OF CANADA 



of high water is often uncertain, and where the variation in the luni-tidal 

 interval with the course of the month is certainly less than other irregu- 

 larities which can thus be better eliminated. The resulting differences 

 are as follows : — 



Alberton and St Paul I. October 16th to 24th ; 13 tides. 

 Difference in time of H. W. (by above method) 2h. 3.3m. earlier. 



Rustico and St. Paul I. October 20th to 24th ; 7 tides. 



Difference in time of H. W. by above method, 



corrected also (by 5m.) by interpolation ; 2h. 31m. earlier. 



St. Peter's and St. Paul I. October 27th to November 23rd ; 31 tides. 

 Difference in time of H. W. (by above method) 2h. 10m. earlier. 



This shows the tide to be nearly simultaneous along the north shore 

 of Prince Edward Island, whereas the Establishments given in the Admir- 

 alty list range from 6^ 10™ to &" 30'°, and therefore require a large correc- 

 tion. This error may have resulted from the difficulty of obtaining good 

 results where the range of the tide is often so small, or from the observ- 

 ation of the day tides only, where the diurnal inequality is large. 



In Northumberland strait, a number of comparisons were made to 

 ascertain to which of the principal stations the tides in that region could 

 best be referred. It was here found that the difference in the time of 

 high water with other stations, especially those on the Atlantic coast, 

 was itself subject to an alternation in value with successive tides, of the 

 same character as diurnal inequality in the luni-tidal interval. There wa& 

 thus a large second difference when the moon's declination was high j 

 whereas near the nodes, the consecutive differences were nearly or quite 

 equal. This will be illustrated by the following example :— 



