NOTES ON TIDE-GAUGES. (7 
As all material will, in time, respond to tension, it follows 
that the distance between the indicator on the gauge and 
the float is constantly increasing, and as the accuracy of 
the gauge depends on the constancy of this distance, it will 
be at once seen how serious the matter becomes. Any 
change that does take place is so gradual that it is not 
noticeable until some serious difference between the sight 
gauge and the recorded height has accrued, and as the 
error varies in every foot of the tide, and as we cannot say 
when the change in length commenced, it is impossible to 
accurately correct the readings. 
For the general purposes for which tide gauges are used, 
such as for reducing soundings, and for current observations 
in hydrographic work, a difference of an inch or two does 
not matter, but where the gauge readings are to be ultim- 
ately used to compare the relative levels of the land and 
the mean sea level, it will be conceded that the utmost 
accuracy is desirable. 
Another source of trouble with tide-gauges is the float. 
This should of course be made of glass, in which case it is 
frail and liable to injury, which means stoppage of the 
gauge at any time. It is generally made of copper, and 
becomes pitted with holes either from chemical or electrical 
action, and water gradually leaks in and so changes the 
zero of the gauge. This is another insidious enemy, as we 
do not know of its existence until the harm is done, and, 
as in the case of the gradual lengthening of the chain, we 
cannot correct the record when the error is discovered. In 
one case which came under the author’s notice, the stretch- 
ing of the chain was just counterbalanced by the sinking 
of the float for many months, but it is not often that errors 
are compensated in this manner. 
In all gauges it is necessary to register the rise and fall 
on a paper of suitable size for convenient handling, and 
