420 H. A. Rowland—Magnetic Permeability, and the 
we obtain Q’= 1257 M(zR?), which is within the limits of 
error in measuring the diameters of the solenoids. &e. 
All the rings and bars with which I have experimented have 
had a circular section. In selecting the iron, care must be 
used to obtain a homogeneous bar: in the case of a ring I be- 
lieve it is better to have it welded than forged solid; to get 
the greatest permeability it should then be well annealed and 
afterward have the outside taken off all round to about + of an 
inch deep in a lathe. This is necessary because the iron is 
“burnt” to a considerable depth by heating even for a moment 
to a red heat, and a sort of tail appears on the curve showing 
the permeability which is not normal. To get the normal 
curve of permeability, the ring must only be used once; and 
then no more current must be allowed to pass through the 
helix than that with which we are experimenting at the time. 
If by accident a stronger current passes, permanent magnetism 
is given to the ring, which entirely changes the first half of the 
curve. 
In the following tables Q= sone — bas been meas- 
C.T.G. Number of coils of tangent-galvanometer used. 
D.T.G._ Deflection of tangent galvanometer. 
D.B. Deflection from helix on breaking current. 
Q. Magnetic field in interior of bar (total). ae 
5 ‘of bar due to temporary magnetization. 
P. ‘“ ‘“ ts it “ rm ee 
anen 
Q=T+P. 
Each observation given is almost always the mean of several. 
D.T.G. is the mean of four readings, two before and two after 
the observations for Q’ and T.; D.C. is the mean of from 10uF 
to ten readings; D.F. mean of three; D.B. mean of two xs 
cept in table I, where the deflection was read only once. | : 
all the tables the column containing the temporary magne 
tion, T, can only be accepted as approximate, the experim 
having been made more to determine Q than T. ; re 
The value of n was generally varied by coiling a wire m0 
or less times around the ring, but leaving its length the sage 
he change in the value of D.C. is due to change 
resistance of the circuit from change of temperature. 
ents 
