164 HENRY A. BOWL AND 



C. Correction for length of needle in tangent galvanometer, 



C = + J^ sin (a + ')f -|r-Y(a' ~ a ) ' 

 \-A-l 



where V is half the distance between the poles of the needle and A' is 

 the radius of the coil. 



D. The resistance of the circuit was constantly adjusted to the 

 standard, but during the time of the experiment the change of temper- 

 ature of the room altered the resistance slightly; this change was 

 measured and the correction will be plus or minus one-half this. The 

 resistance was adjusted several times during each experiment. The 

 correction is Z). 



Some of the errors which are compensated by the experiment need 

 no remark and I need speak only of the following. 



No. 3. By the introduction of commutators at various points all 

 mutual disturbance of instruments could be compensated. 



No. 5. In winding wire in a groove, it may be one side or the other 

 of the centre. By winding the coils on the centre of cylinders which 

 set end to end, on reversing them and taking the mean result, this 

 error is avoided. 



No. 6. The circle was always adjusted parallel to the coils of the 

 galvanometer. Should they not be parallel to the needle, G and 0" 

 will be altered in exactly the same ratios and will thus not affect the 

 result. The same may be said of the deflection of the magnet from 

 the magnetic meridian due to torsion. 



No. 7. /? and 3 both ranged over the same portion of the scale and 

 so scale error is partly compensated. 



No. 8. The zero-point of all galvanometers was eliminated by equal 

 deflections on opposite sides of the zero-point. 



INSTRUMENTS 



Wire and coils. The wire used in all instruments was quite small 

 silk-covered copper wire, and was always wound in accurately turned ls 

 brass grooves in which a single layer of wire just fitted. The separate 

 layers always had the same number of windings, and the wire was 

 wound so carefully that the coils preserved their proper shape through- 



13 To obtain an accurate coil an accurate groove is necessary, seeing that otherwise 

 the wire will be heaped up in certain places. The circle of the tangent galvanometer, 

 which was made to order in Germany, had to be returned in this country before use, 

 and much time was lost before finding out the source of the difficulty. 



