66 Prof. Challis on a Theory of Magnetic Force. 
the atoms, and the contraction of the channel by the atomic 
occupation of space. At the parts of the boundary of any sub- 
stance where the stream enfers, there will be a sudden mcrement 
of velocity, and at the parts where it zssues, a sudden decrement ; 
but not to a large amount, because there is reason to conclude 
that even in dense bodies the space occupied by atoms is very 
small compared to the vacant space. Under all circumstances 
the motion remains steady, if there be no extraneous force, and 
no variation of the primary current. If, throughout the interior 
of the body, the atoms within a given space (as one thousandth 
of a cubic inch) occupy a portion of it which has a constant ratio 
to the vacant space, they produce no acceleration or retardation 
_ of the velocity of the stream. But if from any cause there should 
be a gradation of internal density of the atoms, the atomic com- 
position remaining the same throughout, secondary streams will 
be produced in the following manner :—Conceive the external 
primary current to be cut at right angles by a plane in a given 
position. Then since it is supposed to be steady and uniform, 
the velocity will be always the same at all points of the plane. 
Hence, tracing two contiguous and equal filaments of the stream 
into the interior of the body, it will be seen that if the density 
of the atoms be greater at certain portions of the course of one 
filament than at the adjacent portions of the course of the other, 
the velocity will be greater in any element of the former portions, 
than in the adjacent element of the datter, assuming, as may be 
done, that in other respects the channels of the two filaments 
are at adjacent parts of equal dimensions. Hence by the general 
hydredynamieal equation for steady motion, the density and 
pressure of the ether will be /ess in the former elements than in 
the latter. Consequently there will be an accelerative force of 
the fluid always tending from the rarer to the denser parts of 
the body, and taking effect, whatever be the direction of the 
original stream, in the directions of normals to surfaces of equal 
density, because in these directions the change of density of the 
atoms in a given space is greatest. These accelerative forces 
produce secondary currents, the velocities of which will depend 
on the magnitudes of the forces, and on the extent through 
which they act. Considering the vast elasticity of the ether, as 
shown by the rate of the propagation of light, a difference of its 
density equal to a ten-millionth part of the whole density would 
correspond to an enormous difference of pressure. 
2. According to these views it might be expected that a lamina 
of metal, heated unequally at the two extremities, would become 
electro-dynamic, as Volta found to be the case with a lamina of 
silver. In this stance the heat produces local disturbances of 
the condition of superficial atoms, and from these disturbances 
