84 MR. F. K. SMITH ON THE CONSTRUCTION OF 



and rendered as air tight as possible with rubber bungs. It should l)e observed that 

 the purity of any mercury in the vessels at a and b is unimportant. When thus 

 coupled together, the whole was rested on a framework of wood, to which when 

 filled it was strapped, and placed in the ice-bath afterwards described. For filling, 

 the lead e was turned downwards, and the tubular portions of the apparatus filled 

 under reduced pressure with twice distilled mercury (see p. 105). The attendant 

 conditions were the same as those which held in the determinations of L and W. 



The adjustment of greatest importance had now to be made, viz., the ensuring of 

 the correct relative positions of the standard and its connectors. How very important 

 this is was not realised until many measurements of resistance had been made with 

 very unsatisfactory results. However, by the close observance of the reflected light 

 from the mercury columns, it is possible to secure great accuracy in this adjustment. 

 Thus, if two cylindrical mirrors be placed with their axes parallel, parallel rays of 

 light incident on both cylinders will produce two linear images. These will be in the 

 same straight line for all positions of the observer if the axes are in exact line, but 

 not otherwise. The various displacements will be equal to the projected displacements 

 of the axes. 



In the case of the mercury columns, the bright lines of reflected light acted in 

 this way as a capital fine adjustment. They were made continuous by the rotation 

 of the connecting tube, the junction being viewed first through one pair of holes in 

 the collar C, and afterwards through the second pair. By successive approximations, 

 and often slight adjustment of the screws securing the collars A and B in position, 

 the continuity of the lines for both observations was secured. Examination with a 

 lens greatly facilitated this. The standard portion and the connectors having 

 originally been joined together, the refractive effects of the walls were equal and 

 similar, and so were eliminated. Such an adjustment is, however, very tedious and 

 may occupy as long as 20 minutes. Further remarks on the influence of this 

 adjustment on the general accuracy of the measurements are made on p. 93. 



The tube and accessories were now strapped in position on the wooden framework. 

 Mercury was then poured into a and b, the platinum wires of the leads ignited, and 

 these latter, which passed through small rubber bungs, pressed into position. The 

 insulation of external junctions was made good by applying first a thin layer of hot 

 paraffin wax, and afterwards a coat of shellac varnish. The insulation tests with 

 paraffin wax alone were unsatisfactory when prolonged immersion in ice became 

 necessary. The shellac varnish, however, was completely satisfactory. 



By the introduction of the platinum battery leads into c and f the equipment 

 of the tube was rendered complete. 



The experience of previous observers indicates that for standard purposes the 

 melting-point of ice is by far the most convenient temporal um No knowledge 

 of the temperature coefficient of mercury is then assumed, and errors in thermometry 

 are avoided, (.'oiisi-qm-ntly all standard observations were made at C. 



