238 Mr. Michael Donovan on certain Improvements in 



Its duty is occasionally to assist the directive power of the terrestrial magnetic 

 meridian, and thus to lessen the influence of deflective forces on the needle. 

 With needles of very great sensibility, very weak deflective forces ^vill often 

 produce deflection = 90", and then the galvanometer is incapable of indicating 

 any greater effect. The magnet is graduated in inches and tenths, on both 

 prongs, in order to regulate its proximity to the astatic needle, and therefore its 

 influence in experiments which require repetition or correspondence. The 

 prongs are square, as well as the trunk in which one or the other slides ; and 

 one side of the trunk has a spring which presses on the magnet and enables it 

 to slide evenly. Were the prongs cylindrical as well as the trunk, the magnet 

 might turn to one side or the otiier, and then the other prong would ex- 

 ercise an imdue influence on the astatic needle. When the magnet is not re- 

 quired in an experiment, it should be drawn out and put aside at a considerable 

 distance. One prong is marked N, the other S ; sometimes one must be upper- 

 most, sometimes the other. It is easy to see how it should be placed in order 

 to moderate the effect of a deflective force on the needle. 



It is usual to graduate the circle from to 90°, on each side of the magnetic 

 meridian line. There is an inconvenience in this mode : in some experiments, 

 the voltaic action, which is the cause of the deflection, exists but for an instant; 

 there is no permanent effect to measure, nor any effect beyond one sudden start 

 of the needle, and its immediate return. Yet sometimes the momentum is such 

 as to carry the needle far beyond 90°, and even to whirl it round several times. 

 I think it more convenient that the graduation should be carried to 180". 



The next thing to be considered in the construction of a galvanometer is 

 the compound or so-called astatic needle, a subject of great importance and 

 curious interest. A needle may be defective in two chief points: it may have 

 too little sensibility, by having a strong directive tendency; or it may have too 

 much, by having been brought too near the state of perfect astaticism : in the 

 latter case, a weak deflecting force will produce a maximum deflection, and the 

 needle will be insensible to all higher degrees of energy. 



To avoid these imperfections, I employ the following construction, having 

 discarded the common compound needle, which consists of two haxs permanently 

 fi.red to one spindle. The lower bar of the needle is secured, in the usual man- 

 ner, to the lower end of the spindle ; but the upper one is moveable on the 



