i. 
240 Coast Survey of the United States. 
ble local variation, and also suggested the means of its correction. 
This was the first occasion on which this subject had been in- 
vestigated on the work. It is the design of the superintendent 
to apply the beautiful and simple mode of correction proposed by 
Mr. Airy. It is impossible to allude to his experiments on board 
the iron steamer Rainbow, in the basin of the Deptford dock-yard, 
and the calculations based upon them, without noticing this as a 
remarkable case where a theory purely scientific leads to the most 
useful practical results. Mr. Airy’s combination of the theory of 
M. Poisson with regard to the equal attraction of solids and hol- 
low spheres with his own hypothesis of the magnetic direction 
and intensity of the particles composing an iron steam-vessel, es- 
tablished the safe use of compasses where it was before supposed 
to be very insecure if not impracticable. 
The subject of tides is now treated in a manner alidoeeetied 
new upon the survey. It seems to have been imagined, that the 
only purpose of tidal observations was to correct local soundings, 
and to ascertain the time of high water on the very days of full 
and change, or the vulgar establishment, and for this the solar 
interval alone was sought. 
Even in this humble attempt, time appears to have been kept 
very loosely. Observers have been sometimes employed who 
had little idea of the importance of their office The reforma- 
tion introduced is thorough. The times of high and low water, 
and the period of slack water, are found by watching the register 
unceasingly from half an hour before to half an hour after the 
change, and noting the time and height at very short intervals. 
It was first proposed to lay down these times and variations in 
height upon a large scale, and drawing, at the extremity of the 
curve traced through them with a free hand, a tangent parallel 
to the line of abscissas, to take the ordinate perpendicular to this 
tangent asthe nearest approximation to the time of high water. 
Bat upon trial, this was not found to differ perceptibly from the 
numerical means of the times. The latter is therefore used as 
by far the most convenient. In the reductions, Dr. Bache has 
adopted the methods rendered familiar by the papers of Prof. 
Whewell and Mr. Lubbock. Corrected establishments are de- 
rived from the mean of the lunitidal intervals. As soon as ob- 
servations accumulate sufficiently, at any prominent point, the 
curves of semimenstrual and diurnal inequality will be projected, 
