On the Hydrodynamical Theory of Magnetism. 43 



dynamics of steady motion it follows that the fluid will have 

 greater velocity and less density within the cylinder than with- 

 out, simply because of the contraction of channel by the occupa- 

 tion of space by the atoms. There will be confluence of the 

 lines of motion towards the extremity at which the stream enters, 

 and equal divergence of the lines of motion from the extremity 

 out of which it issues. These lines, as well as the velocity and 

 density along them, will be symmetrically disposed about the 

 axis of the cylinder prolonged in both directions, and also with 

 respect to a plane transverse to the axis through its middle 

 point. Under these circumstances there is no acceleration of 

 the mean current, the quantity of fluid which crosses any unli- 

 mited plane transverse to the axis being the same as if the stream 

 had not been interrupted by the cylinder. 



The above description of the courses of the lines of motion 

 applies to any solid cylinder whether or not it be magnetic. If 

 it is not magnetized, but susceptible of magnetism, the modifi- 

 cation which the original stream undergoes by passage through 

 the cylinder is proper for magnetizing it. For it is evident that, 

 by reason of the variation of the density of the sether from point 

 to point, the atoms of the cylinder, especially those at and near 

 its extremities, will be caused to vibrate ; and it appears from 

 experiment that the magnetizing of a substance is effected when- 

 ever a magnetic stream traverses it while its particles are in a 

 state of vibration. This is remarkably indicated by the well- 

 known experiment in which a plate of iron, placed with its faces 

 in the direction of magnetic dip, is magnetized by being repeat- 

 edly struck with a hammer. Possibly the permanent magnetism 

 of the loadstone may have been gradually induced by the setherial 

 streams which relatively pass through it in consequence of the 

 earth's motion in space. 



Supposing that the cylinder, either by the process above men- 

 tioned, or by some other, has been magnetized, let us inquire 

 what influence this circumstance will have on the stream which 

 traverses it. But it is first necessary to define the magnetized 

 state. According to the theory of magnetism I long since pro- 

 posed, this state depends solely on a small and regular increment 

 of atomic density from one end to the other of the cylinder, the 

 equilibrium of the atoms being maintained by the equality, at 

 each point, of the atomic repulsion towards the rarer part, and 

 the molecular attraction towards the denser part. Conceive now 

 the setherial stream to traverse the cylinder in any direction. At 

 exit and entrance there will be the same cause of disturbance of 

 the lines of motion as in the previous case of a cylinder of uni- 

 form density ; and, besides, the gradation of density will have 

 the effect of generating new streams, which for distinction I shall 



