562 Scientific Proceedings, Royal Dublin Society. 
when taking the reading, there seems no reason to expect a less 
degree of accuracy by this method than in the case of an altitude 
referred to the true horizon. 
In rough water a correction, however, becomes necessary, 
which in extreme cases might render the results uncertain within 
three or four minutes. 
S1 
Fig. 2. 
The diagram, fig. 2, shows that the dip &, which we read if 
we observe the signal when it is on top of a wave, is less than the 
true dip 6, or the angle that would be read if the signal floated at 
the mean-level of the sea. A little consideration shows that this 
is brought about by the fact that the wave brings the signal 
more nearly into the horizontal plane of the observer’s eye, or, 
in other words, reduces the vertical height of the eye above the 
level of the signal by an amount equal to half the wave-height. 
Neglecting the Harth’s curvature the tangent of the angle 
observed is no longer 
OA but C4= OA — half wave- -height 
OS OS 
The correction consists in deducting the estimated HAL¥ wave-height 
Jrom the * height of eye,” and then calculating the dip; or, if refer- 
ence is made to the table of dips, entering this with the reduced 
height of eye and taking out the dip for the correct distance. 
As there is generally difficulty in arriving at a close estimate 
of wave-heights, even in daylight, and at night, when only by 
observation of the signal itself, when close to the ship, any 
estimate could be made, this difficulty is much increased, it is 
