90 
PROFESSOR R. THRELFALL AND MR. J. H. D. BREARLEY 
neutral field. The diagram (in Maxwell, ‘ Elec, and Mag.,’ vol. 2, fig. 15), shows 
how very far from uniform the field is, at or near the neutral point. If the magnetic 
system he rotated through a very small distance, it comes to a place where the 
resultant field is differently directed, and then sails away till a new position of 
equilibrium is attained. The usual practice of considering the field as only influenced 
by the earth and controlling magnet, and of regarding the suspended magnet as 
without influence, is fallacious. The direction of the lines of force depends on three 
things, the earth, the controlling magnet, and the suspending magnet. Maxwell’s 
diagram, fig. 15, represents the case of the discovery well, but the matter is compli¬ 
cated by the fact that the resultant field of earth and control magnet is far from 
uniform even before it is distorted by the suspended magnet. 
It seemed likely that much of the difficulty would be removed by making the 
residual field more uniform, and with this object two additional and similar control 
magnets were introduced below the suspended system, and adjusted by trial till they 
were as nearly as possible symmetrically placed with respect to it and the upper 
adjustable system. The magnets were of the dimensions and at the distances shown 
in the drawing (figs. 1 and 2), and it was expected that they would produce a field of 
surface uniformity sufficient for the purpose, after the manner that the Helmholtz 
coil arrangement produces a solid uniformity (Maxwell, vol. 2, fig. 19). The arrange¬ 
ment figured was set up as a trial, but as its success was immediate and complete, 
we never went on to the selection of magnets of proper length—a step we expected 
to have to take. Since using the arrangement all the wearisome instability and 
trouble has been diminished so much that we now adjust our sensitiveness and take 
observations with no more concern than if we were operating with a wooden galvano¬ 
meter for a rough resistance balance.* 
We must not forget to mention the prime importance of having the fine movements 
of the adjustable part of the control—the upper magnets in our case—as good as 
workmanship can make them. The rough adjustment for distance is made by sliding 
the magnets, screw system and all, up and down. The fine adjustment is by means 
of an excellent nut and fine screw thread—a device described in 1889. The adjust¬ 
ment in azimuth is by means of the lever and screw device, the rough adjustment by 
releasing a clamp and twisting the magnet carriage. As the magnets and nut together 
weigh several pounds, we should have done better had we arranged a counterpoise, so 
that the freedom of rotation of the screw should not be affected by the great pressure 
on it, as is the case at present ; for, though the bearing surface of the nut was well 
ground, and was kept well oiled, still, on screwing the magnets up, the friction was 
enough to slightly twist the wide and thick brass tube, and so to displace the magnets 
in azimuth. 
* Oct., 1893.—After more than a year’s experience we have never liad the slightest difficulty in 
obtaining the • eusitiveness desired since we made the field uniform, , 
