198 Dr. C. Fromme on the Magnetism of Steel Bars. 



one sort are greater : thus the difference from I. 1 to I. 5 is 

 43*6 against 7'3 ; that from II. 1 to II. 5 is 4*35 ; while after 

 this such a regular divergence can no longer be shown, the 

 values of h follow pell-mell, and will not, even in the loosest 

 fashion, arrange themselves in a series. 



If, however, on this wo shall build no conclusions, since 

 much can be urged against the comparability of the series of 

 values on both sides, still thus much can be said — that, accord- 

 ing to what has been above maintained, an augmentation of 

 the vertical and a diminution of the horizontal component 

 would have produced the deviations observed in the second 

 series of values. 



By putting the series of values obtained with I. into a sys- 

 tem of coordinates, taking K as abscissa, k as ordinate, we get 

 a very regular curve, whether we make use of the individual 

 values or the means. The two curves corresponding to the 

 vertical and the horizontal component respectively run very 

 fairly parallel. The curves of II. proceed exactly like those 

 of I., if we merely exclude the first two of II. 4. 



§ 6. After finishing the observations with the earth-induc- 

 tor, the attempt was made to test Maxwell's theory by gradu- 

 ally raising the magnetizing force (a constant galvanic cur- 

 rent), commencing from a very small initial value, and thus 

 endeavouring to ascertain the value at which, according to 

 Maxwell, residual magnetism first makes its appearance. 



The ellipsoids were fixed in the axis of a helix which, in 

 respect of its length and diameter, satisfied the conditions of a 

 constant magnetizing force on all points of the steel body. 

 The helix was set up perpendicular to the magnetic meridian, 

 to the north of a magnetometer suspended by a cocoon-thread, 

 the magnetometer consisting of a magnetized steel ring with 

 a mirror. The deflecting action of the magnetizing helix on 

 the magnetometer was eliminated by a compensation-helix 

 placed southward of it. The current-circuit consisted, further, 

 of a tangent-compass with two windings and a mirror for 

 reading-off (described by Kohlrausch, Pogg. Ann. vol. cxxxviii. 

 in communicating a measurement with the bifilar galvano- 

 meter), and a mercury commutator for changing the current 

 at the same time in the helices and the compass. 



I moreover remark that by a suitable construction of the 

 .apparatus any influence of the current in the helices, and of 

 the magnetized steel body, upon the needle of the tangent- 

 compass was almost entirely avoided. When such took place 

 with the employment of more powerful currents, it never 

 exceeded at the utmost a deflection of the needle one scale- 

 division, and then was always ascertained by special trials. In 



