232 
MESSRS. R. T. GLAZEBROOK AND J. M. DODDS ON THE 
Again 
S 4 y= , 001(r 1 3 — r 2 z )/r-^r z tan y 
whence 
S 4 y=—3-16". 
From the tables we find that when y=62° 36', the difference in the value of 
log M 0 /477VaA for an addition 1" to y is 77X10 -7 ; that is, since SM=SM 0 very 
approximately, 
O' 
8 a M/M=^+77xl0- 7 Xl-64 ,, X log, 10 =’000048 
S 4 M/M= —77 X 10 -7 X 3*16" X log, 10 =-’000041 
The true values of the radii are A=25’753, ct'=25’766. The correction on this 
account to the value of M is therefore 9 X ’000036 X 10 s =’000324 X 10 s . 
The corrections to he applied to b are:—In position 1, —’008; in position 2, 
+ ’016 ; in position 3, X’006 ; in position 4, —’013. 
Making similar calculations for the medium and short rods we can present the values 
of the mutual induction thus : 
Table giving values of the mutual coefficient of induction between the coils. 
Position of coil?. 
Experiments A. 
Short rods. 
Experiments B. 
Medium rods. 
Experiments C. 
LoDg rods. 
1 
1-55636 xlO 8 
1-25797 X 10 8 
■762107x108 
2 
1-55430x108 
1-25649X10 8 
•761321x108 
3 
1-55539X 10 8 
l-25727xl0 8 
•761735x108 
4 
1-55744 x 10 8 
1-25875X10 8 
•762521x10 s 
Mean of Hie four . 
1-55587x108 
1-25758X10 8 
•761921x108 
Error in M produced by an 
error - 001 in a or A . 
•000063 x10 s 
•000052 x10 8 
•000036 x10 8 
Error in M produced by an 
error - 001 in b . . . . 
-•000108 x10 s 
--000078X10 8 
-•000041x10 s 
Thus the error produced by an error of ’001 centim. in b lies between ’005 and ’006 
per cent., and the error in the measurement of b is certainly not more than ’001 inch 
or ’0025 centim. 
Since the rods used to separate the coils were not exactly of the same length, the 
median planes cannot have been exactly parallel. The difference in the length of the 
rods is not as much as ’005 centim. 
The radii of the coils are approximately 25 centims., and hence the angle between 
them is not as great as yyT or y-gVo- 
