whether or not there might be some slight alteration in the centring of the needle sufficient to effect, in a very minute degree, the indications of the instrument.

The experiments, I may here observe, relating to the property of steel for sustaining the magnetic energy, are matters of very great delicacy of management, because of the extreme susceptibility of that property under any extraneous magnetic influence. For not only may an instantaneous approach of one magnet within the sphere of influence of another magnet alter the power of the former, or of both; but even a change in the position of a highly magnetized bar from the horizontal to the vertical, may produce a very perceptible alteration in the deviating power of such bar.

RESULTS.

THOUGH this part of my Investigations has not been carried to the extent which I could have desired, and, though the results are not in every respect accordant with presumed analogies; yet so much appears satisfactory as to justify, I think, these general conclusions:

1. That the degree of retentiveness of magnets is directly as the hardness, and inversely as the energy.

Taking all the causes of deterioration into account, such as the action of terrestrial or other magnetism in unfavourable directions, this proposition will be found a general law; though under circumstances of favourable position, in regard to the direction of the earth's magnetism, the effect of hardness, on the one side, or of great energy, on the other, may not, in moderate intervals of time, be very perceptible.

The difference in retentiveness observable betwixt the softer and harder series of plates described at the conclusion of § 1, is corroborative of the proposition under consideration-the loss sustained by the softer series in 64 days being 9.88 per cent., and by the harder set only 5.82.

BB

2. That the loss of energy by time, in unprotected magnets, whether single or compound, is much more considerable at first than subsequently.

In the case of the two compound series of smaller plates described at page 354, we find that the harder lost two-thirds as much power the first day, as it subsequently did in 64 days.

And in the case of the several six-inch bars, the effects of time on the magnetism of which are shewn in the table at page 357, nearly one-half more deterioration occurred in the first five months than in the subsequent twelve months.

3. That combinations of magnetized plates of tempered or hard steel, appear to be fully as favourable for the retention of the magnetic energy as single magnets of similar mass, and probably more retentive than single bars of corresponding power,- especially if the plates in combination be separated to a little distance from each other so as to attenuate the intensity of the energy in the compound magnet.

This result is deduced from the considerations drawn from a comparison of the tables at page 352 and page 357. 4. That in magnets of moderate powers, though fully magnetized, if kept clear of deteriorating influences and placed in the direction that would be assumed by the suspended compass needle, the power is very enduring; and but little, if at all, less so, in soft than in hard

magnets.

In the first series of experiments described in sect. 2, consisting chiefly of plates adapted, as to weight, for compass needles, the loss of power when laid in the magnetic meridian was scarcely 2 per cent. in the first two months, which loss was not perceptibly increased during the subsequent period of a year.

Again, in the second series of experiments of sect. 2, the same result is partly verified. And in a subsequent experiment, not included in the tables, no loss of power could be at all discerned in a period of seven months, even after the bars had been re-magnetized to their highest capabilities; nor was any perceptible difference observed, in this case, betwixt the sustaining power of the soft and that of the hard bars.

Hardness, therefore, becomes most essentially important for resisting the effects of unfavourable magnetic action, whether arising from the earth's magnetism operating against that of the magnet, or from unfavourable contact with, or proximity to, any other magnet. In another respect, also, as has been repeatedly shewn, hardness is of essential moment; that is, for enabling large magnets to sustain the violence belonging to the condition of powerful combinations, or massive bars. In the experiments with the two energetic compound bars, of different degrees of hardness, described at page 354, the harder bar exhibited an advantage over the other in its relative tenaciousness; but in this case the magnets were very powerful ones, and the position in which they were laid by (east and west), had not the advantage of terrestrial magnetism for assisting in retaining the power.

5. That when the maximum power of a magnet is slightly reduced by unfavourable

proximity to another magnet, the resulting energy is still less influenced by, time.

This is a result which had been fully anticipated, and is well illustrated by the experiments exhibited in the table at page 357. For here we find (excluding the first bar in the table as an anomalous case, and apparently defective in accuracy) that whilst the unreduced bars lost, on an average, 32 per cent. in seventeen months, the bars that had been slightly reduced lost but 1·1 per cent. of-their power; and that these reduced ones, in the last twelve months, sustained scarcely any sensible loss.

END OF PART II.

London: Printed by Manning and Mason, Ivy-lane St. Paul's.