ON THE ABSOLUTE UNIT OF ELECTRICAL EESISTANCE 165 



out. No paper was used between the layers. As the wire was small, 

 very little distortion was produced at the point where one layer had 

 to rise over the tops of the wires below. Corrections were made for 

 the thickness of the steel tape used to measure the circumference of 

 each layer; also for the sinking of each layer into the spaces between 

 the wires below, seeing that the tape measures the circumference of 

 the tops of the wires. The steel tape was then compared with the 

 standard. 



The advantages of small wire over large are many; we know exactly 

 where the current passes; it adapts itself readily to the groove without 

 kinks; it fills up the grooves more uniformly; the connecting wires 

 have less proportional magnetic effect; and lastly, we can get the 

 dimensions more exactly. The size of wire adopted was about No. 22 

 for most of the instruments. 



The mean radius having been computed, the exterior and interior 

 radii are found by addition and substraction of half the depth of the 

 coil. The sides of the coil were taken as those of the brass groove. 



All coils were wound by myself personally to insure uniformity and 

 exactness. 



Tangent galvanometer. This was entirely of brass or bronze, and 

 had a circle about 50 cm. diameter. The needle was 2-7 cm. long and 

 its position was read on a circle 20- cm. diameter, graduated to 15'. 

 The graduated circle was raised so that the aluminium pointer was on 

 a level with it, thus avoiding parallax. The needle and pointer only 

 weighed a gram or two, and rested on a point at the centre which was 

 so nicely made that it would make several oscillations within 1 and 

 would come to rest within 1' or 2' of the same point every time. I 

 much prefer a point with a light needle carefully made to any suspended 

 needle for the tangent galvanometer, especially as a raised circle can 

 then alone be used. The needle was suspended at a distance from any 

 brass which might have been magnetic. There were a series of coils 

 ascending nearly as the numbers 1, 3, 9, 27, 81, 243, whose constants 

 were all known, but only one was used in this experiment. The proba- 

 ble error of a single reading was about 1'. 



Galvanometer for induction current. This was a galvanometer on a 

 new plan, especially adapted for the absolute measurement of weak 

 currents. It was entirely of brass, except the wooden base, and was 

 large and heavy, weighing twenty or twenty-five pounds. It could be 

 used with a mirror and scale or as a sine galvanometer. It will be 



