be of value to give some particular details in regard to the best method of construction and insulation. In most of the older types of quadrant electrometers the needle system was made unnecessarily heavy. In consequence of this, if a sensibility of the order of 100 mms. deflection for 1 volt was required, it was necessary to charge the Leyden jar connected to the needle to a fairly high potential. This at once introduced difficulties, for at a high potential it is not easy to insulate the Leyden jar satisfactorily, or to charge it to the same potential from day to day. This drawback is to a large extent avoided in the White pattern of the Kelvin electrometer, which is provided with a replenisher and attracted disc for keeping the potential of the needle at a definite value. If sufficient trouble is taken in insulating and setting up this type of electrometer, it proves a very useful instrument of moderate sensibility, and will continue in good working order for a year or more without much attention.
Simpler types of electrometer of greater sensibility can however be readily constructed to give accurate results. The old type of quadrant electrometer, to be found in every laboratory, can readily be modified to prove a useful and trustworthy instrument. A light needle can be made of thin aluminium, of silvered paper or of a thin plate of mica, covered with gold-leaf to make it conducting. The aluminium wire and mirror attached should be made as light as possible. The needle should be supported either by a fine quartz fibre or a long bifilar suspension of silk. A very fine phosphor bronze wire of some length is also very satisfactory. A magnetic control is not very suitable, as it is disturbed by coils or dynamos working in the neighbourhood. In addition, the zero point of the needle is not as steady as with the quartz or bifilar suspension
When an electrometer is used to measure a current by noting the rate of movement of the needle, it is essential that the needle should be damped sufficiently to give a uniform motion of the spot of light over the scale. The damping requires fairly accurate adjustment. If it is too little, the needle has an oscillatory movement superimposed on the steady motion; if it is too great, it moves too sluggishly from rest and takes some time to attain a state of uniform motion. With a light needle, very little, if
