applicable to the Gold Bullion Assay. 103 



set, such as will enable us to weigh these fine gold educts. or 

 cornets as they are called, with always a single weight in the 

 pan, and with the rider on the beam. Twenty weights are 

 required, they range according to the decimal notation already 

 described from '80 to - 99 (from eight grains to nine and 

 nine-tenth grains actual weight). If the cornet weighs '9843 

 with the weight "98 in the pan, and the rider on the 42nd 

 division of the beam, we seek the additional -0001 by the 

 sweep of the index, and thus arrive at the weight of the 

 cornet ; the accuracy of the tveighment, however, depending 

 more or less on those points which concern the accuracy of 

 the beam, to which reference has already been made. 



One final explanation will prepare the way for what I 

 have to propose in amendment of the above-described 

 method. First, it is to be understood that for economy of 

 time, and for other obvious reasons, the proper weight of 

 each sample to be assayed is first prepared with tolerably 

 close approximation, by an assistant, and the assay pieces 

 thus prepared are finally weighed and adjusted by the 

 assay er. Secondly, it should be remembered that the cost 

 or care required for the preparation of weights of extreme 

 accuracy, even though the weights be multiplied in number, 

 is a matter of quite minor importance as long as the utmost 

 accuracy of the work is thereby maintained. 



I will now proceed to a concise description of the proposed 

 method. First, as to the balance suitable for this modified 

 procedure : it is essential that it be light in the beam, simple 

 in construction, rigid, with sharp and hard knife-edges, 

 which must be truly set at right angles to the length of the 

 arms. These qualifications are essential ; they are commonly 

 to be found in the best description of assay balances, but 

 certain others commonly attempted in the best assay 

 balances are non-essentials for the particular method about 

 to be described. It is not necessary that there shall be 

 absolute or even approximate equality of the two arms, but 

 prime importance is attached to those requisites which 

 determine rapidity of oscillation and sensitiveness. Only 

 one minor alteration of the pattern of the common assay 

 beam is proposed, that, namely, which concerns the range of 

 the rider. It is suggested that an offshoot from the beam 

 for supporting the latter, as shown in the illustration, be 

 formed, so as to become rigidly a part of the beam and allow 

 the rider to rest over the outer knife-edge, or on any position 

 intermediate between that and the centre knife-edge of the 



