repetition of M. Arago's experiments on magnetism, &c. 4*9 1 
circle, whose axis is the axis of rotation, and whose circum- 
ference passes through the extremities of the diameter parallel 
to the dip. The extent of this declination will depend on the 
velocity of rotation and the diameter of this small circle, and 
will be proportional to both, that is, to the velocity of rotation 
multiplied into the sine of the angle made by the axis of rota- 
tion with the direction of the dip. It will therefore be a maxi- 
mum when the axis of rotation is perpendicular to the magnetic 
meridian, and vanish when the shell is made to revolve on 
an axis parallel to the line of dip. These consequences are 
perfectly consonant to the results obtained by Mr. Barlow 
in his paper ; and in fact, the general result announced by him 
in (page 326 of this volume) comes to the very same thing 
as above stated ; for it is obvious, that the new axis of polar- 
ization there spoken of, acting in combination with the origi- 
nal, or, as we may call it, the primary axis developed in the 
quiescent state of the shell, will exert a compound force on 
the needle, such as would be exerted by a single equivalent 
axis situated intermediately between them, but much nearer to 
the more intense than to the more feeble one. The position of 
this equivalent axis will necessarily be in the great circle 
passing through the two component ones. Now the small 
circle described by the point which was first the pole of the 
stronger or primary axis about the axis of rotation is a tan- 
gent to this great circle, and the equivalent axis (being but 
little removed from the primary one, by reason of the small 
intensity of the other), will therefore have its pole situate 
indifferently in either circle. Or conversely, the single axis 
produced in our view of the subject being resolved into two ; 
one of which is that corresponding to the quiescent state of 
MDCCCXXV. 3 S 
