LIEUT. -COLONEL SABINE ON TERRESTRIAL MAGNETISM. 
115 
These coefficients have accordingly been used in calculating the results in the pre- 
sent Number. A careful examination of the observations at the foreign stations had 
led me to infer that Nos. 5. and 6. would probably be found to have larger coefficients 
than the other needles, which has proved to be the case. The variation in the amount 
of the coefficients in the different needles, considerable as it is, is not unprecedented ; 
it probably depends on the quality and temper of the steel, and may be particularly 
influenced by the portions of soft iron which a needle may contain. A species of 
steel has been recently employed for magnets in the Russian observatories, in which 
the coefficient for temperature has even a negative sign, i. e. the magnetic intensity 
of the bar increases with heat. In a letter which I have received from M. Adolphe 
Erman, he describes this particular kind of steel as consisting of alternate very thin 
layers of soft iron and of steel, so that when heated the soft iron layers would increase 
their magnetic intensity, and the steel layers diminish theirs; the amount and sign 
of the coefficient depending on the preponderance of the layers of soft iron or of steel, 
which is subject to much variation. It is called “ Boulat,” or “ damascened steel,” 
and is considered the pride of the Uralian forges. In a bar of this steel, kindly sent 
me by General Tcheffkine, at the request of M. Kupffer, the usual effect is thus 
reversed. Experiments made with it at Woolwich by Lieutenant Riddell gave the 
results stated in the following memorandum : — 
“ The effect of temperature on the bar of Russian steel, sent by M. Kupffer, was 
tried in the usual manner by means of the magnet of the declination magnetometer, 
the bar being placed with its axis in the line passing through the centre of the 
suspended magnet perpendicular to the magnetic meridian. The subjoined observa- 
tions furnish satisfactory proof that the ordinary effect of temperature on bars of 
steel which are hardened throughout is reversed in this bar, but the value of the co- 
efficient, or change of force for 1° of Fahr. deduced from them, must be taken only 
as a rough approximation, as in addition to the probable error of the observations 
themselves, the bar was placed with its centre at a distance of three feet, or only 1^ 
times its length from the magnetometer, in order to produce a sufficient deflection, 
the magnetism being weak*.” 
* After the completion of this experiment with the bar in the state in which it was received from General 
Tcheffkine, a portion was cut off, softened, and made into 3-inch cylinders of the dimensions used with the 
portable magnetic apparatus. The effect of temperature on one of these cylinders, hardened afresh and remag- 
netised, was tried in a similar manner, and the value of its coefficient found to be about '0003, the force decrea- 
sing with an increase of temperature, which is the ordinary effect. 
MDCCCXLIII. 
R 
