282 Mr. Barlow on the secondary deflections deduced 
ball will be the most powerfully acted upon ; or, which is the 
same, the centre of all the actions on the needle will be in 
that branch ; if, therefore, that branch be found between the 
ball and the meridian, the attractive powers prevail over the 
repulsive ; but if the meridian be found between the ball and 
the nearest branch of the needle, then the deflection is due to 
an excess of the repulsive forces over the attractive. 
In the first case, by deteriorating the near branch, we 
diminish the attractive forces, and in the other case, the 
repulsive ; so that in both instances the needle ought to 
approach the meridian ; as is found to be the case. But by 
deteriorating the other, that is the most distant branch, we 
increase the preponderance of attractive power in the one 
case, and the repulsive in the other, and, consequently, the 
needle will be more deflected, or recede farther from the 
meridian. 
With respect to the rather uncertain character of the 
secondary deflections near the points of change, the expla- 
nation appears to me to rest on this : that, admitting the 
attractive and repulsive principle, the centre of attraction and 
of repulsion may fall both in the same branch of the needle, 
or in opposite branches. In the one case, the needle is 
deflected by only the difference of the two forces, and in 
the other, by their sum. And it is probable that in or near 
the points of change, the degree of deterioration may pro- 
duce this uncertain result, by changing these centres from 
one branch to the other, according to the intensity of the 
deteriorating power. 
After all, however, it will not be expected that the results 
due to such a complex system of forces can be illustrated in 
