services generally adopt quite a simple marigraph. The apparatus
consists in principle of a counterpoised float whose rising and
falling motion, reduced to a tenth, by means of a system of toothed
wheels, is transmitted to a pencil which moves in front of a vertical
cylinder. This cylinder itself moves around its axis by means of a
clockwork mechanism, and accomplishes one entire revolution every
twenty-four hours. By this means is obtained a curve of the tide in
which the times are taken for abscisses and the heights of the sea for
ordinates. However little such marigraphs have had to be used, great
defects have been recognized in them. When we come to change the sheet
on the cylinder (and such change should be made at least once every
fifteen days), there is an interruption in the curve. It is necessary,
besides, to perform office work of the most detailed kind in order to
refer to the same origin all these curves, which are intercrossed and
often superposed in certain parts upon the original sheet. In order to
render such a disentanglement possible, it is indispensable to mark by
hand, at least once every twenty-four hours, upon each curve, the date
of the day corresponding to it. It is equally useful to verify the
exactness of the indications given by the apparatus by making readings
several times a day on a scale of tides placed alongside of the float.
Nine times out of ten the rise of the waves renders such readings very
difficult and the control absolutely illusory.

All these conditions united, as well as others that we neglect in this
brief discussion, necessitate a surveillance at every instant. The
result is that these marigraphs must be installed in a special
structure, very near the bank, so as to be reachable at all times, and
that the indications that they give are always vitiated by error,
since the operation is performed upon a level at which are exerted
disturbing influences that are not found at a kilometer at sea. It