188 Scientific Proceedings, Royal Dublin Society. 



plan was adopted. A brass rod, having a diameter of one-eighth inch, was 

 suspended vertically in a wooden frame. Two circular brass plates, 10 cms. 

 in diameter, with three-sixteenth inch holes drilled at their centres, were 

 ■carefully ground to a plane surface and cemented to ebonite blocks, which 

 were drilled to slide smoothly on the rod. The two discs were then fixed 

 in position, so as to be separated by '12 mm., and to be as nearly parallel as 

 possible. The ebonite blocks were each furnished with a set-screw, and it 

 was assumed that if the rod were subjected to tension the plates would then 

 be moved apart by the amount that that portion of the rod between the two 

 set-screws increased in length. The screws were approximately 1"5 cms. 

 apart on the rod. 



With this apparatus it was found that the highest sensitiveness obtain- 

 able with this ultra- micrometer was even greater than had been suspected. 

 For one particular adjustment of the inductances the galvanometer was quite 

 stable in the absence of accidental vibrations, but nevertheless an extra- 

 ordinary degree of sensitivity was displayed. Thus, it was found that a 

 deflection of four hundred and ten divisions of the galvanometer was pro- 

 duced when a load of iive hundred grammes was suspended from the rod. 

 A simple calculation shows that such a load applied to a brass rod one-eighth 

 inch diameter will produce an extension of 9-6 x 10"* cm. in a length 

 of I'o cms., assuming Young's modulus for brass to be 10^^ dynes 

 per sq. cm. As this extension corresponds to four hundred and 

 ten divisions of the galvanometer, we may, in round numbers, say 

 that a displacement of 10 x 10"° -f 400, or 2-5 x 10"° cm. was detect- 

 able. With a more sensitive galvanometer perhaps a further limit might 

 have been attainable, but in view of the possible unsteadiness of the 

 readings one cannot predict what the lowest attainable limit might be. The 

 number quoted was obtained under very unfavourable conditions, no screen- 

 inc of any kind being placed around the apparatus, and the only precaution 

 taken by the operator being to stand as still as possible while lifting and 

 lowering the weight by a string. A movement of the hand towards where 

 the apparatus lay on the table was sufficient to cause a perturbation of the 

 galvanometer. I have no doubt, however, that by enclosing the apparatus 

 as far as possible in a metal box, as is the usual practice with oscillatory 

 circuits, this trouble will be eliminated.^ 



A careful series of tests with improved apparatus is now being under- 

 taken, the results of which I hope to embody in another communication. 



1 This trouble has since been completely overcome. 



