44 Prof. Challis on the Hydrodynamical 



call secondary streams. The particular mode of generation of 

 these streams is next to be considered. 



The incident stream being supposed to have originally the 

 same velocity and density at all points of any section transverse 

 to its direction, it follows,, by the laws of steady motion, that 

 after entrance into the cylinder its resulting mean velocity will 

 be greater and mean density less, the greater the atomic density. 

 This is an immediate consequence of the contraction of channel 

 by the atoms. Hence the fluid will be impressed at all points in 

 the interior of the cylinder by a constant accelerative force acting 

 in the direction from the rarer towards the denser end. The 

 consequent effective accelerative force will, by reason of the 

 inertia of the fluid, accelerate a given particle towards a trans- 

 verse plane through the middle point of the cylinder, and equally 

 retard it after it has passed that plane. Thus there will be a 

 maximum of velocity at the points where the plane is cut trans- 

 versely by the lines of motion. Also as there can be no transfer 

 of the whole fluid mass, supposed to be of unlimited extent, by 

 means of an accelerative force impressed on a limited portion of 

 it, there will necessarily be return currents at different distances 

 from the cylinder, such that the lines of motion of these secon- 

 dary currents will be reentering. The courses of these lines will 

 be symmetrical with respect to the axis and the above-mentioned 

 transverse plane, and will cross this plane outside the cylinder at 

 right angles. Such is the general character of the secondary 

 streams to which the theory attributes 'the phenomena of the 

 magnet. 



It will be seen that the intensity of the secondary stream is the 

 same whatever be the direction of the primary, so long as the 

 latter is of given intensity. Also it must be admitted that the 

 secondary stream, as generated by the interior gradation of den- 

 sity of a magnetized body, is dynamically far more effective than 

 that modification of the primary stream which was above de- 

 scribed as being produced whether or not the body be magne- 

 tized ; for otherwise magnetic streams would be perceptible in 

 the case of a non-magnetized body. The great intensity of the 

 secondary streams is to be attributed to the efficacy of the im- 

 pressed accelerative forces by which they are generated, the 

 equation p = a 2 p showing that, on account of the great magni- 

 tude of a 2 , the extremely small variation of p due to the grada- 

 tion of density may cause a large change of p. In the subse- 

 quent reasoning the above-mentioned small modification of the 

 primary stream is left out of account. 



It may be supposed that the whole mass of the fluid in which 

 the secondary streams are generated partakes of the primary 

 motion. In that case, if the primary velocity were impressed in 



