RESEARCHES IN MAGNETISM. 
537 
times their diameter. In hard iron and in steel a smaller ratio of length to diameter 
would no doubt give an equally good approximation to the condition of endlessness. 
It should not be forgotten that want of perfect homogeneity introduces a self- 
demagnetising force even in the longest rod or in a ring, by causing some of the lines 
of induction to escape from the substance of the metal at intermediate points, and 
it does not appear that we can easily, if at all, approximate more closely to the ideal 
condition of uniform magnetisation than by the use of a straight piece of wire in 
which the ratio of length to diameter is (say) 400. If it be desired to deal with 
a larger section of metal than a single wire, a good plan is to form a multiple ring by 
winding a long continuous wire into a coil of as many turns as may be wished, and 
cutting the ends so that they abut against one another. This gives a more homo¬ 
geneous structure than is got by welding a rod into a ring, or even by turning a 
ring out of a solid forged piece. One or two experiments made with multiple rings 
have only served to confirm the general accuracy of the results got by the use of 
straight wires. 
§ 18. Direct Magnetometric Method of Experiment .—In addition to the ballistic 
method, a direct magnetometric method has been employed in many cases in examining 
the magnetisation of long wire rods. The wire was placed in the vertical position, 
with its upper end nearly level with, and magnetically east of a small mirror magneto¬ 
meter, formed by suspending the mirror of a Thomson galvanometer by a silk fibre 
about 8 centims. long. Two coils were wound round the rod : through one of these 
a constant current was maintained which just sufficed to neutralise the vertical 
component of the earth’s magnetic field. The other formed the magnetising solenoid, 
and the current in it was varied not by sudden steps as in the ballistic method, but 
