478 Mr. Babbage and Mr. Herschel's account of the 
' At the opposite extremities of two diameters at right 
angles to each other, four equal small bar magnets were 
fixed in a vertical position, having alternately their north and 
south poles downwards. This position promised to present 
two material advantages ; first, that in neutralizing the system 
we have not the whole polarity of the magnets to contend 
with, but only the small remains of directive tendency which 
arises from the magnetic axis in each not being precisely 
coincident with its axis of figure, since it is evident that an 
infinitely thin magnetic cylinder placed perpendicularly to 
the horizon, would from that cause alone be indifferent as to 
situation ; sdly. That in this situation their poles interfere 
with each other^s action on the plate revolving below them, 
less than in any other. Instead of four we might (and as will 
be seen ) occasionally did place a greater number of magnets 
round the circle, or within its area, but for the experiments 
now in view four were enough. 
iG. The system so constructed was found to require no 
after adjustment, being to all appearance perfectly neutral, 
so that this part of our purpose was completely accomplished, 
and the earth's action eliminated from the enquiry. The 
irregular torsion of the silk thread however still embarrassed 
us a good deal. But though this undoubtedly caused indi- 
vidual results to differ more from the mean than we had 
expected, it is not sufficient to account for a singular ano- 
maly observed not only in the mean results of a great 
number of trials, but in all individual cases ; viz. that by this 
mode of observation, zinc was invariably found to stand 
above copper in the scale of magnetic action, whereas in the 
determination by the statical method, where the deviation of 
