10 THE ANTARCTIC MANUAL. 
that observations should be made at several places, and that, whilst 
a fortnight should be looked on as a minimum, yet even the results 
for a week would not be without value. 
Tidal observations with a tide-gauge are not appropriate for an 
expedition such as that now in preparation, and reliance must be 
placed on observations by eye. The ingenuity of naval officers is 
never wanting in the construction of an apparatus for observations of 
this kind. Until recently it has been the practice in the navy to 
observe the times and heights of water at high and low tide, and the 
times have often been tabulated in apparent instead of in mean time. 
But experience has now shown that it is far more convenient for the 
determination of the law of the tide that the observations should be 
taken at equidistant intervals of the mean time of the place. In 
their work at Campbell Island the French Expedition took observa- 
tions every half-hour of the day and night. But, in my opinion, all 
the needed accuracy may be attained by means of hourly observations. 
If it is deemed desirable, observations of high and low water might 
also be made, but I am not inclined to consider this as necessary. Dr. 
Haughton, in his ‘Arctic Manual,’ speaks of observations every four 
hours during the night as perhaps sufficient, but I find it hard to 
coincide with this opinion. If the exigencies of the service render 
hourly observations impossible, we may have to put up with two- 
hourly observations. In the subsequent numerical treatment, the 
first step would be the interpolation of the missing hourly values, 
and thus the strength of the results would be much weakened by the 
intermission of several hours. 
At each tidal station meteorological observations will, no doubt, 
be made, and thus the materials will be at hand for the empirical 
correction of the observed heights, if such correction should be found 
possible. 
It has been observed that the rise and fall of the tide is apt to 
occur by a series of impulses, so that the trace drawn by a tide-gauge 
is usually more or less notched. The degree of irregularity is found 
to vary much according to the weather, and according to the site of 
observation. In estuaries and land-locked bays the zigzag is often 
very pronounced, but even in such sites the degree of perturbation 
varies much with the exact position of the tide-gauge. It is highly 
desirable that a position should be chosen where the irregularity is 
not pronounced, but it is not always easy to decide on the most 
favourable spot by mere inspection. During the last thirty years 
much attention has leen paid to these secondary undulations, and 
they have heen shown to be local vibrations of the partially enclosed 
