HARMONIC ANALYSIS OF TIDAL OBSERVATIONS. 109 
values which A and B would have had if the two epochs had been 
identical. A and B are, of course, the component semi-ranges of the tide 
of short period at the epoch chosen for the tides of long period; to 
determine them it is necessary to multiply R by the cosine and sine of 
V+u—k at the epoch. 
[Q.] 
Schedule of Coefficients for Clearance of Daily Means in the Final Equations. 
1 -| o-w | 20 | 2(¢—n) | n | 2n 
(M,) n=2(y—-<¢). 
[A, n, l,cos]} —0:05557 | +0-00302 | +5:7393 | —0-10410 | —0-01465 
[B, n, t,cos]] —0°17036 | —0-03773 | —2-9228 | —0-07525 | —0-07546 
[A, n,1, sin]] —0:17075 | +0-04170 | —2:8400 | —0-00176 | —0-00353 
TB, n, 1, sin]] +0-04410 | +0-01052 | —5:7271 | +0:00476 | +0-00958 
(N) n=2y—30+4+c. 
CA, n,1, cos]} —0-05884 | +0-03680 | +.0-02938 | —0-01760 | —0-01760 
'B,n, l,cos]| —0-07758 | —0-22337 | —0-19384 | +0-00254 | +0-00254 
(A, n, 2, sin]] —0:02059 | —0-15245 | —0-12210 | +0-00020 | +0-00041 
'B, n, 1, sin]} +011381 | —0 08544 | —0-08081 | -+0-00007 | +0:00015 
(O) n=y—2e. 
(A, n, 7, cos} —0:06485 | +0:01673 | +0:01582 | —0:19240 | —0-19340 
[B, n, J, cos]} —0°34765 | —0-07788 | —0-°08158 | —0°18260 | —0-18311 
fA, n, l,sin]} — 0°34523 | +0-08418 | +0:08748 | —0-004C0 | —0:00926 
[ B, m, 1,sin]} +0°04052 +0°08379 | +0:03295 +0°00897 | +0-01802 
It may happen from time to time that the tide-gauge breaks down for 
a few days, from the stoppage of the clock, the choking of the tube, or 
some other such accident. In this case there will be a hiatus in the 
values of éh. Now, the whole process employed depends on the 
existence of 365 continuous values of ch. Unless, therefore, the year’s 
observations are to be sacrificed, this hiatus must be filled. If not more 
than three or four days’ observations are wanting, it will be best to plot 
out the values of 6h graphically on each side of the hiatus, and filling 
in the gap with a curve drawn by hand, use the values of ¢h given by the 
