1817.) On Magnetism. 139 
opposite to S, and circulating round k, will enter the magnet C, by 
the channel a, at s; and thus & becomes an element of C. ‘The 
austral fluid of the magnetising bar C, when sliding over the space 
b d, at the same time drives off the austral fluid lying upon the 
facet of k opposite to a, and also that on the facet of /, next to k, 
into the common circulation, and it entersC at z. The loreal fluid 
coming from m impels the loreal fluid on the other surface of /, and 
that also on its facet opposite m, and it is driven in the circulation 
through the channel lctos. Thus k and / become elements of 
the magnetising bar. It isnow evident how the remaining molecule 
mopgqrtv may in like manner be converted into so many ele- 
ments of the magnetic bar; and that the two series of molecule 
al, &c. and kl, &c. are entirely under its influence. But when 
the bar C is removed, the channels al, kb, &c. will be stopped, 
the repulsions of the fluids at the angles in the oblique channels 
balancing each other. The fluids belonging to each series will only 
pass through the straight channel N S, and continue to circulate 
round their proper molecule. 
We can now explain the cause of the surprising phenomenon, 
which presents itself when a magnetic bar is cut, or suddenly 
broken, soas to detach a portion from it ; however short, that por- 
tion instantly becomes a magnet. This fact I consider as a serious 
objection to the theory of Coulomb ; for on his hypothesis, if a bar 
were broken, we should not have two magnets differing only ac- 
cording to their size; but we should find the larger portion possessed 
of one pole extremely weak, .while the other pole possessed all its 
original power. The austral fluid, in short, is nearly all confined 
to the one portion, and the loreal fluid to the other. M. Haiiy 
endeavours to remove this difficulty by supposing that the fluid 
within the separated portions distributes itself to the extremities. 
But still there would be such an excess of boreal fluid in the one 
piece, and of austral fluid in the other, as would render the polar 
energy in each very sensibly different, which is contrary to expe- 
rience. On the contrary, if we suppose any portion of the bar D 
broken off, we at once perceive how each portion is a complete 
magnet. 
The cause also of what are called consequent points or poles of 
smaller power is well accounted for by this theory. These poles 
are situated in the oblique channels, and may arise from an irregular 
arrangement of the molecule, or from some error in the operation 
of magnetising the bar. Ina future communication I shall explain 
by this theory a curious magnetic experiment related by the late 
Professor Robison, 
Lam, Sir, your obedient servant, 
. ‘ AnDRew Horn. 
