124 
SIR G. B. AIRY ON THE TIDES AT MALTA. 
between the vertical lines was intended to correspond with the hours marked above 
the space.” 
The consecutive sheets were pasted together, forming a band nearly 60 feet long. 
For easy control of this I mounted it on two rollers. On examination the record 
appeared to be everywhere in general good order, with the one exception, that on 
April 5, 1871, there are 25 hour-spaces, as if the barrel had been slipped by hand 
while the pencil was in contact. I suppressed one hour-space in the part which 
appeared to require it, and the record of that day then accorded sufficiently with 
those adjoining. 
Section II. — First Treatment of the Registers for Removal of Oscillations of Short 
Periods , and for Measures referring to the Tides at London. 
Through many parts of the register the course of the pencil, as it would have been 
guided by luni-solar tides only, is disturbed by vertical oscillations (for which \ 
anticipate the name seiches ), with a sensibly uniform period of about 21 m , and with 
magnitude sometimes considerably exceeding that of the genuine tides. To liberate 
the tides from the seiches, the highest and lowest points of each oscillation were 
marked, and the intermediate point was carefully determined by compasses, and was 
marked ; a pencil curve was then drawn, through all these intermediate points, which 
was considered to be the true tidal curve. The number of intermediate points thus 
determined was nearly 1100. 
The base adopted in the numerical evaluations of the height of the water (to be 
shortly treated) was 1 0 inches below the zero line of the printed register sheets ; this 
base was adopted in order that all records of tidal elevation might have the 
positive sign. 
For the treatment of the measures in regard to time, I determined to lay aside all 
consideration of harmonic sequence of arguments, and to compare every day’s tide 
with the corresponding tide at London. The reasons for doing so were mainly these : 
that in fact each day’s tide is sensibly independent of every other day’s tide ; that the 
tide of the Thames is a very convenient standard, because it contains no sensible trace 
of diurnal tide, and that in the Admiralty Tide Tables we have (as is generally 
understood) a fairly accurate statement of the real tides at London. 
Interpreting the scale of times upon the sheets in conformity with Sir Cooper 
Key’s description above, and then considering those times as London times (thus 
introducing an error depending on difference of longitude, which is unimportant, 
because it can be corrected at the end), I laid down by points on the printed base- 
line the time of every London high water. The intervals between these points were 
divided each into 24 equal parts, and these were taken as corresponding to genuine 
tidal half-hours. At every one of these points the elevation of the tidal register was 
measured on the sheet. The number of these measures is about 1900. 
Now, considering that the quantities of which we are in search are either constant 
