Tiie magnet which forms the subject of this paper consists 
of a soft iron bar with a fiat plate attached to one end and 
surrounded by a coil of Vvire in the same way as the 
ordinary electromiagnet. Outside this coil is placed a tube 
of soft iron of the same length as that portion of the interior 
bar which projects beyond the plate; this tube lias flat ends 
■ — one of which is in contact with the plate, while the other 
comes up flush with the end of the bar—so that a plate or 
keep placed over the end is in contact with both the bar and 
the cylinder. The magnet is excited in the ordinary wa}^, 
by connecting the ends of the wire which forms the coil with 
the poles of a battery. When thus excited this magnet 
exhibits certain peculiarities as compared with a common 
ma^gnet. In the first place the magnetic field is very limited, 
being confined to the space in front of the open end of the 
tube, there is little or no magnetism along the tube or at 
the closed end. The magnet retains its keep with greater 
force than the simple bar. Mr. Faulkner has some magnets 
of this kind which retain the keep with 100 times more 
force when the outer tube is on than when it is removed. 
The ratio of these retaining powers appears, however, to 
depend on the relative diameters of the bar and the tube ; 
the larger the bar in proportion to the tube, the greater is 
the difference. Some magnets, made especially to test the 
relative powers, give an increase of only double as compared 
with the simple bar magnet. This magnet has a greater 
sustaining power than the horse-shoe ma^gnet. This was 
shown by putting tubes round the poles of a horse-shoe 
magnet, by which means it was made to sustain greater 
weights than it would without the tubes. 
The object of the paper was to suggest explanations of 
these phenomena. They were attributed to three principal 
causes. 
1. The tube surrounding the bar unites the poles and 
converts the magnet into a kind of horse-shoe magnet, the 
