158 ANNUAL OF SCIENTIFIC DISCOVERY. 



iron was principally confined to paddle steam-ships, this was not impor- 

 tant; but now, when so many screw steam-ships and sailing ships are 

 built of iron, this arrangement ought never to be used. I know not 

 whether the compasses of the Tayleur could have been affected by either of 

 these causes. 



5. The question, however, which immediately presses, is : What (under 

 all circumstances) is it best to do now ? In answer, I assert in the first 

 place, and I am supported in this by Dr. Scoresby's experiments, that the 

 source of local disturbance and its laws are perfectly well known ; that 

 the disturbance can be neutralized, by well-known means, to the greatest 

 exactness ; and that this neutralization is perfect during change of time 

 and change of place, until the ship herself undergoes an organic change. 

 In the next place, I protest strongly against the system, now in use (I 

 believe) in the lioyal Navy, of using a table of errors, and thus constantly 

 making numerical corrections instead of once making a mechanical cor- 

 rection. (1.) It is bafiiing to the mariner. (2.) It is liable to exactly the 

 same errors, in the event of a change in the ship's sub-permanent magnet- 

 ism, as the system of relying on the mechanical correction. (3.) It is liable 

 to errors peculiar to itself, which would be entirely avoided by the use of 

 mechanical correction. In illustration of the last remark, I will refer to 

 the table, in page 104 of the late Capt. Johnson's book, on the " Devia- 

 tions of the Compass," second edition, a work in many respects highly 

 valuable. Capt. Johnson has given the observed deviations of the com- 

 pass on board three iron steam-vessels in different parts of the world ; and 

 I select the last, (the Trident,) because its deviations were the largest. 

 The deviations in the Thames ranged from 22 15' E. to 21 12' W. The 

 deviations of the same compass at Malta ranged from 15 29' E. to 14 21' 

 W. Now, the proportion of the terrestrial, horizontal magnetic forces, in 

 the Thames and at Malta, is as 52 to 75, very nearly. Therefore, if the 

 ship's sub-permanent magnetism remained unaltered, the tangents of the 

 angles of deviation in the Thames and at Malta would have been in the 

 proportion of 75 to 52. On computing the Malta deviations from those in 

 the Thames by this proportion, we obtain 15 50' and 15 3', agreeing 

 with those observed more nearly than observations can be made with a 

 ship's compass. The whole of the deviations recorded by Capt. Johnson, 

 for the Bloodhound, the Jackal and the Trident, at Lisbon, Constantinople, 

 the Piraeus, and Malta, can be computed in the same way from those in 

 England, and the results are equally accordant. (The terrestrial horizon- 

 tal forces, on the same scale of proportion, are, Lisbon, CO ; Constantinople 

 77 ; Piraeus, 76.) It follows from this, that the ship's sub-permanent 

 magnetism, in each case, was unaltered, and its effect would have been 

 exactly compensated, at every locality, by a permanent magnet. And thus 

 the captain of the Trident, using Capt. Johnson's table, would have had 

 errors of nearly seven degrees ; whereas, if he had used my correcting 

 magnets, he would have had no perceptible error in the whole voyage. I 

 pointed out this result to Capt. Johnson ; I know not with what effect. 



