186 
LINES OF MAGNETIC DECLINATION IN THE ATLANTIC. 
up to St. Helena, gradually though interruptedly, diminished the southerly dip, which 
at St. Helena is about —20°. The 2nd of January may therefore be regarded as 
dividing this part of the voyage into two portions in respect to the changes of B. As 
the daily rates of change deduced above for the passages from the Thames to Port 
Praya and from St. Helena to the Cape ('00059 and ’00052) differ so little from each 
other, we may not unreasonably take their mean as applicable to the first division of 
this part of the voyage, or for that division in which the change of dip was continuous 
and uninterrupted. This gives as the value of B on the 2nd of January +*0074. 
Now at St. Helena we find it by experiment +‘0073 ; on this assumption consequently 
the magnetism of the ship would have remained nearly stationary from the 2nd of 
January to the arrival at St. Helena, which is by no means an improbable supposition. 
We may derive intermediate values of C in a similar manner*. This coefficient is 
however of very minor importance. 
It will of course be understood that this mode of deriving these coefficients is one 
which would only be adopted in the absence of more satisfactory data ; and fortunately 
in the part of the globe for which the corrections are required the values of B and C 
are less significant than in the higher latitudes. The observations themselves, how- 
ever, furnish a test by which the appropriateness of this or of any other hypothesis 
proposed for their correction may be judged, viz. by the measure of agreement into 
which the corrections bring observations made on the same day or near the same 
spot with the ship’s head on different points. Without entering into details, it may 
be stated, generally, that the corrections computed by the formula (6.) with the value 
of the coefficients as above stated appear to bear this test very satisfactorily ; the 
observations thus corrected becoming much more accordant with each other than 
either when uncorrected, or than when corrected by the same formula with its 
variable coefficients made to vary in accordance with the dip. 
Determinations in H.M.S. Erehus, in 1842 and 1843, between Cape Horn and. the 
Cape of Good Hope. 
For the corrections of the declinations observed in the Erebus in 1842-43 between 
Cape Horn and the Cape of Good Hope, we have seen, p. 184, that the value of the 
constant coefficient D as derived from the experiments in the River Thames and Port 
Praya in 1839, and at St. Helena and the Cape of Good Hope in 1840 was +22'; for 
the experiments at the Falkland Islands in August 1842, D=+23', and from those 
at the Cape in 1843, +24'; I have made no alteration therefore in the general table 
of corrections which was computed, as already noticed, with +22'. 
For the variable coefficients B and C we possess no other data, for the period 
now under consideration, than the values derivable from the experiments in the Falk- 
* Values of C derived from the harbour observations : — 
Gillingham — '0036 
Port Praya — '0046 
St. Helena -'0033 
Cape of Good Hope — '0044 
