the Capacity of Gold-leaf Electroscopes. 247 



completely shielding the outer case of the electroscope. 

 This may not be convenient or possible with all types. It is 

 also essential to undertake a series of measurements of the 

 capacity of the outer case, thus insulated and shielded, before 

 the actual measurement of the capacity of the gold-leaf system 

 can be begun. Barratt's method is to charge a standard 

 condenser and reduce the potential by a sufficient number of 

 charge and discharge operations, using the same electroscope 

 to measure the initial and final potentials. The method is a 

 valuable one where a quadrant electrometer is not available 

 or is not in proper working order, but it is slower than Cooke's 

 and still slower than the method to be described, thus 

 increasing the effect of any leakage of the electroscope itself. 

 Partly in consequence of this, the probable error has been 

 found to be from two to four times as large. 



The present method, which is only a slight modification 

 of the ordinary method of mixtures, gives quicker (and 

 therefore more accurate) results, and is applicable to any 

 conductor of extremely low capacity, while avoiding the 

 necessity of constructing any special subsidiary apparatus 

 for each electroscope to be measured. Its practicability 

 depends on the fact that, as shown by V. H. Jackson and 

 the author, very great accuracy can be obtained in quadrant 

 electrometer measurements of capacity by adopting certain 

 modifications. These were described in a preliminary form 

 in J. A. S. B. voLx. No. 6 (1914). The final form, as shown 

 at the fifth session of the Science Congress at Lahore, has 

 not yet been described in detail, but the diagram given below 

 shows the essential arrangements. The usual troubles due 

 to failure of insulation and zero creep are notably lessened 

 by reducing the number of insulators as far as possible, and 

 by enclosing all working parts of the electrometer in an 

 earthed metal ca«e. It will be seen that there are only two 

 insulators required in this arrangement, one the usual amber 

 supports of one pair of the quadrants and the other for 

 connexion with the outside. The air inside the case is kept 

 constantly dry by means of sulphuric acid, which can be 

 easily replaced if necessary. This we have found to be the 

 only satisfactory drying agent. The case itself can be 

 instantly removed or replaced when any adjustments are 

 necessary. All connexions are established electromag- 

 netically on the key supported by the quadrant insulator, 

 which brings different mercury cups into contact bv well- 

 amalgamated copper points. In the paper referred to, it has 

 been shown that the capacity of the quadrant system of the 

 electrometer itself at constant voltage, when compared with 



