A NEW CURRENT WEIGHER, ETC. 479 



The cylinders were prepared in the rough by Messrs. GOODY and CRIPPS, the large 

 ones being 13 inches in diameter, 11 inches in axial length, and 2 inches thick. The 

 corresponding dimensions of the small cylinders are 8, 6, and 0'5 inches. A few 

 veins run through the large cylinders, but the dark material, of which these consist, 

 is of negligible magnetic susceptibility. An appreciable quantity of the substance 

 was collected from a number of rough pieces of marble sent by the marble merchants, 

 and this was subjected to the magnetic tests already dealt with ; there was no 

 indication that the permeability differed from unity. 



The cylinders are of an inconvenient shape and size for a direct determination of 

 their coefficient of expansion ; moreover, it was inadvisable to immerse them in water, 

 and this latter operation was desirable if satisfactory observations were to be made. A 

 bar of marble, 45 x 5 x 2*3 centims., was therefore procured from the same source; 

 this was baked in an oven at 140 C. and soaked in hot paraffin wax previous to any 

 linear observations being made. The mean coefficient of expansion between 1 C. 

 and 25 C. was determined by Mr. ATTWKLL to be 24 x 10~ 7 per 1 C. 



The marble cylinders were examined for flaws and freedom from cavities ; they 

 were then turned until their dimensions were approximately correct, and afterwards 

 baked in an oven at a temperature of 140 C. for 30 hours. On the completion of 

 the baking, and whilst in a hot condition, they were immersed in hot paraffin wax at 

 110 C. No bubbles of gas were evolved from either of the four cylinders used in 

 the ampere balance, but from one part of another cylinder, which was rejected for 

 reasons mentioned hereafter, a tiny stream of bubbles escaped for a minute or two 

 after immersion in the wax. Each cylinder remained immersed for at least 36 hours ; 

 on removal it was again examined for flaws, but none were detected. Previous 

 to the turning of the marble cylinders, a long steel rod was turned on the lathe 

 set apart for this work, and its ellipticity and conicality were determined by 

 measurement. The ellipticity was very small and the lathe was adjusted until the 

 conicality was too small to be measured with certainty ; notwithstanding, the 

 marble cylinders turned subsequently are distinctly conical, and, in the case of the 

 large cylinders, those ends are the larger which were nearer to the face plate when 

 the spiral grooves were cut. We conclude, therefore, that the weight of a cylinder 

 produced a tilt, and that better results might have been obtained by turning between 

 dead centres. The two small cylinders were turned in this way. 



A cylinder was secured to the face plate of the lathe by four external dogs, the 

 space between the face plate and the end of the cylinder nearest to it being about 

 l inches. This mode of support enabled the two ends to be turned truly parallel, 

 and the interior surface to be turned normal to them. To turn the outer surface, 

 four large metal studs were turned in position on the face plate, and one end of the 

 cylinder fitted over them ; this end was pressed into contact with the face plate by 

 two long bolts passing through the cylinder to a rectangular bar of steel pressed 

 against the other plane end ; the outer surface was then turned. The inner and 



