21 
1908-9.] Vibrational Neutral Points in Magnetised Iron. 
when the field is positive, the differences in the series of readings obtained 
are immediately apparent, and to repeat these different readings for negative 
field the connections must be uncrossed. Compare the upper and under 
halves of the third and fifth columns of both tables. On the right-hand side 
of the tables the summations of the average readings under these two 
distinct conditions, relative to the reversal of the secondary oscillation 
connections, multiplied by the factor T54 gives the induction (B) in C.G.S. 
units plotted in fig. Vl.ab, H = 15T, and in fig. VII., H = 44, as if all the ob- 
servations had been made with positive field. These curves are typical of all 
the others. The continuous-line and dash-line curves represent for positive 
values of field uncrossed and crossed oscillation connections respectively. 
With oscillation solenoid No. 1, figs. Vl.ab, although the curves of each 
pair do not cross, they could not have been represented on the same 
diagram without unnecessary confusion. With oscillation solenoid No. 2, 
on the other hand, they can readily be shown on the same diagram, fig. VII. 
In this case the crossings of the curves of each pair, with the exception of 
those for H = 20 (the cyclic extreme), may be in the meantime noted. 
The full-line curves of fig. VIII., obtained from the same experimental data 
as those of fig. VII. (oscillation solenoid No. 2, with 15 turns per cm. of length), 
show the induction changes plotted, not against oscillations of increasing 
intensity for various values of field (as in fig. VII.), but against field as 
abscissse for various values of the oscillation intensity, i.e. against the 
number of 1 mm. sparks in series. Six curves only have been selected. 
Each of those might have been shown in pairs corresponding to crossed and 
nncrossed oscillation connections in reference to positive field. This, how- 
ever, has not been done, the average induction changes under these two 
distinct conditions having been plotted. The number of 1 mm. sparks 
in series are 1, 2, 3, 8, 9, and 10, represented by the corresponding vertical 
ordinates of fig. VII., the four intermediate values 4, 5, 6, and 7 having been 
omitted. 
On the other hand, the broken line curve of fig. VIII. is plotted from the 
same experimental data as that from which the curves of fig. V.a b were 
obtained (oscillation solenoid No. 1, with three effective turns per cm. 
of length). It shows the average induction changes produced on the 
superposition of the initial 1 mm. spark obtained from distinct series of 
experiments for the six values of H referred to above, represented by the 
corresponding vertical ordinates (1) of figs. Vl.a and Vl.b with crossed and 
uncrossed oscillation connections (with reference to positive field) respectively. 
These results are also tabulated in Table V. 
Fig. VIII. thus shows the induction changes (ordinates) due to co-direc- 
