Astronomical and Nautical Collections* 127 



other hand, that this force, remaining parallel to itself in the whole 

 extent of the vessel, the change of its absolute direction can pro- 

 duce no other effect than the change of the situation of the vessel, 

 with regard to this same direction. It must, however, be remarked, 

 that we here suppose that the ship must be imagined to have been 

 submitted, at the time that the adjustment is made, to a rotatory 

 motion, not only in a horizontal plane round a vertical axis, but 

 also in a vertical plane round a horizontal axis ; not that the com- 

 pass can ever be afterwards employed in such a situation, but since 

 the inclination of the magnetic force may change without limit 

 during the voyage, so that the relative change of direction is the 

 same. [And it is unnecessary to remark, that such a previous ex- 

 periment as this was never contemplated by the inventor of the 

 apparatus.] 



According to these considerations, the question, which is the 

 subject of the last paragraph of the memoir, is reduced to the in- 

 quiry, whether it is possible to destroy identically, that is to say, 

 for all possible directions of terrestrial magnetism, the deviations 

 of a horizontal needle, derived from bodies magnetized by the in- 

 fluence of the earth, by adding to them a piece of iron, which is to 

 be magnetized by the same cause. This requires, first, that the 

 magnetism should have the same degree of mobility in the piece of 

 iron, and in each of the other bodies ; but we may admit that the 

 coercive force may be very weak, so that the distribution of the 

 magnetism may be always conformable to the actual direction of 

 the action of the earth ; a supposition which seems to be but little 

 removed from the truth, with regard to the magnetic substances 

 which are found on board of ships. With respect to the form of the 

 bodies acting on the compass, I have supposed, in order to arrive at 

 a complete solution of the proposed question, that all these bodies are 

 spheres, either solid or hollow, of any given diameters and thick- 

 nesses, sufficiently distant from each other to allow us to neglect 

 their mutual actions, and disposed in any assignable manner 

 around the magnetic needle. Of such a system of spherical bodies, 

 magnetized by the influence of the earth, I have determined the 

 action on a given point, in order to tee if, by properly fixing the 



