412 Mr. R T. Glazebrook on the 



Then we have 



B + U = 8 + a? 1; 

 or 



B = S + ^-U. 



Next S is placed across the gap GG' and the Beetle moved 

 into the series position ; the combination is then compared 

 with the 10-ohm coil ; let w 2 be the shift. Then 



X=9B + U + S + # 2 , 



= 10S-8U + 9^ + av 



fSo that if. U the resistance of the contact-pieces is known, 

 the value of X is found in terms of S. 



The chief source of error arises from the difficulty of main- 

 taining the temperature of the Beetle sufficiently steady during 

 the measurements. The consequences due to this can be re- 

 duced by taking a series of measurements in order, first with 

 the Beetle in multiple arc, then with it in series. If a given 

 coil be tested by this method at varying temperatures, the 

 observations usually lie on a smooth- curve to within about 

 0°'05 C. 



The method was I believe originally devised by Lord 

 Bayleigh in order to construct a coil of 25 ohms from one of 

 1 ohm. He employed five coils of 5 ohms each ; it is the 

 method by which the 10-ohm coils of the Association were 

 constructed. 



The other methods employed to determine the 10-ohm coils 

 were both build-up methods.. Messrs. Nalder, Bros., kindly 

 put at my disposal a number of boxes arranged as described 

 below. 



Each box contained five coils in series and of the values 1, 

 1, 2, 3, and 4 units. The two ends of the whole series were per- 

 manently connected to two mercury cups, and each junction 

 of two consecutive coils was connected to a mercury cup. 



Thus the resistance of the series is 10 times that of the first 

 coil. 



The coils are enclosed in a wooden box, and are numbered 1 a, 

 lb, 2, 3, 4; the unit maybe either 1, 10, 100, or 1000 ohms. 



The arrangement may be used in two different ways ; it is 

 shown diagrammatically in fig. 2. 



Method II. — Coil 1 a is connected by contact-pieces of 

 known resistance to the Carey- Foster bridge and compared 

 with a known standard S. The value 1 b is found similarly. 



Then S is replaced by an ordinary box of coils, and coil 2 

 of this bos is compared with 1 a + 1 b in series ; coil 2 of the 



