78 Transactions of the Royal Canadian Institute 



and Declinational months are treated separately, each in its own period. 

 The values for the same missing day (December 17) as obtained by the 

 system adopted, are shown in the table next following. The whole four 

 days, before and after, are here placed together, as the arithmetical 

 result is the same as though they were averaged in pairs. The com- 

 parison of the two results shows how very small any error is likely to be. 



Possible errors outstanding. — The only remaining sources of appreci- 

 able error which seem possible, are those due to any want of balance in 

 other periods. This may occur in respect to the position of perigee; but 

 it would only be appreciable in regions where the effect of the moon's 

 distance is large, and also when the date of the missing tide curve was 

 near to perigee. The forms of the curves on days a month earlier and 

 later, being displaced 1)4 days relatively to perigee where change is 

 rapid, might be affected. At any other time in the anomalistic month, 

 the error from this cause would scarcely be appreciable. 



Another influence that may be unbalanced, is the declination of the 

 Sun; especially since in the Declinational type of tide, this solar effect 

 is so evident. The examples here under consideration are both near the 

 solstices, when the solar element in the diurnal inequality is at its maxi- 

 mum. At the break in January 1919, the interval of 29 days before one 

 of the missing days, falls practically on the solstice; and 29 days after, 

 falls two months from the solstice. 



The possibility of these sources of error can always be seen from the 

 astronomical conditions at the time. To obtain a check upon them, the 

 Tide Table values may be resorted to. It is to be noted that in these 

 prediction values, all the influences affecting the tide, due to every move- 

 ment of the sun and moon, are taken into account; and consequently the 

 relative positions of the points of high and low water on different days, 

 as given by their co-ordinates on the prediction curves, may be taken as 

 correct. On the other hand, the absolute positions of these points, with 

 relation to the observed points, may be considerably out of place ; 

 especially in early Tide Tables which are based on the first year or two 

 of harmonic analysis, when the variation in the 19-year declination 

 cycle may not yet be well determined. 



A point at high or low water on a missing day, can therefore be plotted 

 on the tide diagram by placing it in the same position relatively to the 

 observed points that the Tide Tables indicate. The relative and abso- 

 lute positions of these points are shown in the Plates attached; one of 

 these giving the sets of points on an enlarged scale. 



The Tide Table values may also be utilized if a check comparison is 

 thought necessary, where the pair of curves between which the inter- 

 polation is made are too far apart, or in dealing with the half-tides of 



