Iw. BELL DAWSON] TIDES IN GULF AND RIVER ST. LAWRENCE 65 



form of tidal interference ; and probably also, in part at least, to wind 

 <li8turbance, which appears to be considerable in the inlets leading to 

 Charlottetown, and in the tidal ])ortion of the Hillsborough river which 

 runs beyond it. 



In order to ascertain the best method of obtaining the tide at Pictou 

 Itself with reference to the principal stations, further compai'isons were 

 made with Halifax and St. Paul island. As the greater part of the 

 range in the difference is evidently due to diurnal inequality, the differ- 

 ences in the time of high water from Halifax to Pictou, during the five 

 months of the observations, were tabulated in draconitic month s^ or in 

 sets beginning and ending at the^corresponding lunar nodes ; and taking 

 •care to keep the differences for upper and lower transit tides respect- 

 ively, in alternate columns. The mean values of the differences through- 

 out the draconitic month were then plotted as successive ordinates. and a 

 pair of undulating curves was obtained, corresponding with the upper 

 and lower transits of the moon respectively. These curves are sym- 

 metrical, and intersect each other on their horizontal axis at about three 

 tides after the nodes, as they should. The ordinate of the axis, measured 

 from the zero line, corresponds with the average difference in the time 

 of high water. The semi-amplitude of the curves, . at the maximum 

 declination of the moon, is V 14:'", which is half the diurnal inequality 

 in the difference. The variable difference between Pictou and Halifax 

 thus obtained (varj'ing with the moon's declination) enabled the diurnal 

 inequalit}^ to be allowed for. To test the accuracy of the result, a tide 

 table for Pictou for one of the months of the observations them.selves, 

 was calculated by means of this variable difference with Halifax. The 

 range of error in this table, as compared with the observed tides, was 

 then found by omitting the greatest and least erroi", and taking the 

 averages of the 2nd and 3rd greatest and^ least, to allow for any 

 ■exceptional values, in the same way as in the case of the tidal differences 

 themselves. The error was thus found to range from 49" early to 22'" 

 late ; and the average error during the month, without reference to sign, 

 was 15"". The tide at Pictou for the same month, as found by the 

 <*onstant ditVerence given in local almanacs now published, showed an 

 iiverage error of 45'" during the month. 



The difference between Pictou and St. Paul island is less than with 

 Halifax ; and the corresponding declination ^cui'ves obtained by the 

 method above described, were found to be unsymmetrical ; but had a 

 semi-amplitude of only ,37'". As the tidal data for St. Paul island have 

 not yet been deduced from the observations, the variable difference from 

 Halifax was made use of, in the preparation of 'provisional tide tables 

 for Pictou for the season of navigation of 1897. The observations of the 

 previous season are thus made immediately available ; and these tables 

 will be sufficiently accurate to be of practical service in the meantime. 



