METEORIC AND ARTIFICIAL NICKEL-IRON ALLOYS. 
39 
The constants (calculated by the method already described) by means of which 
corresponding values of p and H can be obtained in absolute units from the obser¬ 
vations are, in the case of the meteoric iron ring (first winding), given approximately 
by the equations 
p = 135*5 x i Q d ji p d 0 , 
H = 28'0 x i p , 
in which d is the inductive effect (proportional to the sine of half the angle of throw) 
in the secondary circuit when the current i p is established in the ring primary, and d 0 
is the corresponding inductive effect produced by a current i 0 in the standardising 
solenoid. The corresponding constants in the other cases are :—Meteoric ring (second 
winding). 160’5 and 25*5 ; nickel steel ring, 166*4 and 24*8. (If the values of the 
constants were calculated in the ordinary way, the first would be about 0*6 per cent, 
higher and the second about 0*75 per cent, lower in each case.) 
In the earlier experiments, of which the object was to examine how, for a constant 
value of H, p varied with the temperature, the value of II chosen was 0*476, as being- 
near the value of the Earth’s field and corresponding with a convenient deflection— 
50 divisions—on the shunted voltmeter. In the experiments after the second winding 
of the meteoric ring the value of H used was unintentionally lower, viz., 0*433 instead 
of 0*476. This happened because the shunted voltmeter reading used was still 
50 divisions, although the ring- constant had become 25*5 instead of 28*0. In the 
experiments with the nickel steel ring, and also in some comparison experiments with 
a ring of nearly pure iron (of which further description is omitted for the present), 
the primary current was, in each case, adjusted so as to give a field as nearly as 
possible equal to that used with the meteoric ring after the second winding, and 
corresponded in each case to a field of 0*434 C.G.S. unit. 
Owing to the difficulties of winding under the conditions necessary for work at high 
temperatures, the degree of accuracy of the constants calculated as above is not as 
great as is the accuracy, relative to one another, of the numbers proportional to p. 
Consequently the actual values of the latter are not given, in most cases, in the 
tables showing the results of the measurements; but only the values of i 0 d/i p d 0 , 
which are numbers proportional to them and which represent the direct results of the 
observations. 
§ 7. In some cases, however, the actual values of the permeability, and not merely 
their relative values, are required approximately for the purposes of comparison. 
Thus the permeability of the meteoric iron was carefully determined at the conclusion 
of the experiments, after it had been subjected to rej^eated heatings, as described 
below, including a final annealing at 900° C. The permeability was also determined 
before the material received any thermal treatment beyond the slight warming that 
took place when the ring was cut out of the original block. The ring windings in 
these initial and final experiments were such that relatively accurate measurements 
