242. I will here make known a method by which the diurnal variations may be increased almost indefinitely. It consists in placing laterally, at some distance from the moveable needle, another needle either moveable or fixed, but of such a length, that the action of one of its poles shall greatly predominate. The reciprocal action of the two needles, arising chiefly from the attraction or repulsion between the contiguous poles, is combined with the directive terrestrial force, which tends to bring them to the magnetic meridian; and the position in which each is fixed, is determined by the resultant of these three kinds of forces. Now when the direction of the terrestrial component varies, this resultant will also change its direction; and it ap pears from calculation, that its change must in some cases be greater than that of the terrestrial force itself. We have, therefore, only to make this favourable arrangement; and if we are careful to use needles of very hard steel, highly tempered and incapable of acquiring a very high degree of magnetism, and consequently fitted to be more constant, I doubt not that we shall obtain an apparatus whose extreme sensibility will make known to us many curious phenomena, which the smallness of the diurnal variations has hitherto prevented our observing.†

243. It must be considered as important for the future progress of physics to determine with accuracy, the actual intensity of terrestrial magnetism, as it has been to determine the absolute pressure of the atmosphere, and the temperature of different climates. If the same observations are continued for several ages, it will be known whether there is any variation in the intensity of the magnetic forces, analogous to that which is found to exist with respect to its direction.

The first method which suggests itself is to observe the declination, the inclination, and the intensity, by means of three needles appropriated solely to these objects, and carefully preserved for the purpose from age to age. As they may lose some portion of their magnetism during long intervals, they may be restored to the same degree of magnetic power, by being subjected anew to the process of magnetizing with the aid of very strong bars, combined according to the method of double touch. Indeed, if we apply this process, the needles will by the influ

† See note on the diurnal variations.

ence of the extreme bars be immediately charged with a degree of magnetism much higher than they can retain when left to themselves. So that if their internal constitution undergoes no change, the degree of magnetic power with which they are saturated would remain constant, and of course the variations of directive force would afterwards depend solely on the changes that take place in the magnetism of the globe. We can render this method much more accurate by thus preserving a number of needles well proved; but we must always be certain that they have remained untouched. We may dispense with this condition, however, if we can discover any means of making two needles exactly similar at all times. For this purpose, we must not think of employing steel, which being a compound of carbon and iron, is necessarily variable in its proportions. But we may very successfully employ cylindrical needles made of a compound of wax and deutoxide of iron in known proportions; for this deutoxide of which natural magnets are in a great measure composed, is very susceptible of magnetism, and very little liable to change its component parts. It will be sufficient therefore to make similar magnetic needles at different epochs, and to magnetize them to saturation, and then to observe the effects produced on them by the terrestrial forces.

Practical Instructions as to the Method of observing the Elements of Terrestrial Magnetism.

244. MAGNETIC observations being one of the most important objects that can engage the attention of travellers, I have thought it would be useful to subjoin a few practical instructions respecting the processes to be employed in making such observations with accuracy. To begin with the most simple case, I will suppose that the observations are to be made on land; I will then describe the additional precautions necessary on board of vessels, liable always to be more or less agitated, and which may themselves exert a considerable disturbing force on the needle in consequence of the iron used in and about them. The first element to be determined is the declination, that is, the angle com

prehended between the magnetic needle and the plane of the astronomical meridian. The instruments destined for this purpose are called declination or azimuth compasses. Among the different forms which have been used, I give the preference to that invented by M. Cassini, to which our ingenious artist M. Gambey has added an improvement, that gives it a decided superiority over all others. It is represented in figure 123.

This compass is composed principally of a long magnetic needle of a rectangular form, suspended edgewise in a horizontal position by an assemblage of flat silk threads without any sensible torsion, and surrounded by a horizontal graduated circle, EOV, which enables us to measure the extent of its motions. The point of suspension is at C in a cross bar of copper, supported by two columns of the same metal; and these columns are inserted at their bases into a plate also of copper, which rests on a pivot in the centre of a circle; so that the whole apparatus admits of being turned about this centre, like a common surveying instrument. A branch of copper B, attached to one of the columns, carries a vernier V over the graduated circle EO, that is employed in measuring this motion. Moreover to guard the apparatus from the agitation of the air, it is completely enclosed in a glazed box of wood or copper, which rests upon the same metallic supports. Having assured ourselves of the perfect mobility of the needle, it only remains to determine the point of the horizon to which it directs itself. For this purpose, by means of the transverse axis AA', we attach to the summit of the two columns a telescope LL, which is moveable in a vertical plane, the axis AA' being horizontal. To give the axis this position, we first make the graduated circle EOV itself perfectly. horizontal by means of the adjusting screws v, v, v, and spirit levels placed upon its surface. We then suspend a spirit level to the axis A', by means of two hooks; and if it is not already horizontal, we make it so by the aid of a little apparatus of movable pieces attached to the columns and admitting of a vertical as well as a horizontal motion in one of the ends A of the axis. Now the telescope LL contains in its interior two very fine hairs or wires, situated in the focus of the eye-glass, whose point of intersection serves to fix the precise direction of the visual ray. But these wires are so placed by the instrument-maker, that the visual ray which passes through their point of inter

section is exactly perpendicular to the axis AA', and passes through its middle point. Besides, the telescope has not a simple spherical lens for an object-glsss; but two such lenses placed one over the other, very unequal both as to curvature and dimensions; one, occupying the larger part of the tube, throws a distinct image of distant objects on the wires; the other, which is much smaller, is so formed that when combined with the larger, it throws on the same wires the image of very near objects. Moreover, the direction of the visual ray, which passes through the intersection of the wires, is regulated by the two lenses in the same manner. Accordingly, if we would see only very near objects with the telescope, we have merely to cover all that portion of the larger glass for which we have no use, by attaching to the end of the instrument an opaque cover having a circular opening at the centre, as represented in figure 124; and, on the contrary, if we would look at distant objects we substitute another cover, opaque at the centre and open toward the circumference, as shown in figure 125. This being well understood, we can determine the direction of the magnetic needle in the following manner; we first turn the box until the needle attains a free and unobstructed position; and when it is stationary we direct the microscopic part of the tube LL successively towards the two ends of the needle, where are attached the cross wires which serve for signals, like those on the needle employed for the diurnal variations. It seldom happens, that the point of intersection of the wires is, on the first trial, in a line with the intersection of the wires of the telescope; but as we can move the axis of the telescope in a horizontal direction, and also turn it by means of the arm B attached to the columns, it is always very possible to bring the intersection of the wires of the telescope to coincide with the image of the signal carried by the needle; and it is morcover necessary, that this coincidence should be effected at each end of the needle. When this condition is fulfilled, the optic axis of the telescope, that is, the visual ray which passes through the intersection of the wires, will evidently be in the same vertical plane with the line drawn through the two signals, affixed to the extremities of the needle. This plane will then be that of the magnetic meridian, if the line above mentioned coincide with the magnetic axis of the needle. Let us suppose for the present that this is the case. We have,

therefore only to take from the end of the telescope the cover by which it was fitted for near objects, and to substitute instead of it that which answers to the small lens, in order that we may distinguish distant objects; then, directing the telescope to some point near the horizon, which is directly in the line of intersection of the interior wires, we shall have the position of the magnetic meridian; and thus we may discover the declination of the needle by measuring, at our leisure, the angle comprehended between this line and the geographical meridian of the place. This problem belongs to astronomy. But it is not by any means certain that the line drawn through the two signal points of the needle is its magnetic axis; here then is an occasion for applying the method of correction already explained. The process in this case is very easy; for the needle has for its cap a hollow ring to which is accurately fitted a copper cylinder that encloses it. In order to reverse it, therefore, it is sufficient to turn it upside down by shifting the cylinder; after which we observe anew the direction. If we obtain the same point in the horizon as before, we have no correction to make; but if the second direction differs from the first, as is most generally the case, we must refer them both to the geographical meridian, and take the mean of the angles thus observed. This will be the true declination.

245. This reference to the meridian can be very accurately made with the same instrument. For, when we have found the point of the horizon, to which the axis of the telescope is directed, we observe the number of degrees, &c., in the horizontal circular division to which the vernier of the arm B corresponds. This being done, without touching this circle again or deranging it at all, give free motion to the box and columns, without now regarding the needle, and turn the arm B, until the telescope is directed towards some known star then situated near the horizon. Observe, by means of a good watch, the precise moment when this star is directly behind the point of intersection of the wires, and we can hence deduce by calculation the angle comprehended between the geographical meridian and the vertical plane in which the star is situated at this instant. But having noted the point of the graduated circle to which the vernier of the arm B corresponds, we shall know the angle which this same plane makes with the magnetic meridian, in which the telescope was