Toronto, as deduced from the monthly determinations. In doing so he remarked that "the general effect of the disturbances of the inclination at Toronto is to increase what would otherwise be the amount of that element; therefore, if the disturbances have a decennial period, the absolute values of the inclination (if observed with sufficient delicacy) ought to show in their annual means a corresponding decennial variation, of which the minimum should coincide with the year of minimum disturbance, and the maximum with the year of maximum disturbance." At Toronto, where the true secular change is very small, the effect of this superimposed variation is very visible, so that the yearly values of the inclination appear to increase up to the period of maximum disturbance and to decrease after it. At Kew the general effect of disturbances is probably the same as at Toronto--that is to say, tending to increase the inclination; but the secular change being considerable, and tending to decrease the inclination, the joint effect of the secular change and the superposed variation might be expected to appear in a diminution. of the yearly secular change for those years during which the disturbances are increasing from their minimum to their maximum value, and in an increase of the yearly secular change for those years during which the disturbances are decreasing from their maximum to their minimum.

The Kew records appear to exhibit a variation of this nature. Observations of dip were commenced at the Kew Observatory in 1854; and by comparing a good number of observations taken during the latter months of 1854, with two circles and four needles, with observations taken with the same circles and needles during the same months of 1855, we obtain a yearly secular change of 2.24.

During the years from 1856 to 1859 inclusive, monthly observations were made with a circle known as the Kew circle, two needles being always used, and the mean of the two results taken as the true value of the dip.

From this circle we have the following results :

If we take the mean of these three values of yearly secular change, and also include that between 1854 and 1855, we have a mean value of yearly secular change, for the period between 1854 and 1859, amounting to 2'-29, and this value will not be sensibly altered if we omit the observations between 1854 and 1855.

In 1859 it was resolved to substitute another circle for the Kew circle, as the action of the latter was not considered to be quite satis

been employed, and monthly observations have been made with it, generally in the afternoon-two needles being used, as before. From this circle we have the following results :

exhibiting between 1860 and 1864 a mean secular change of 2'58. It will be noticed from this, that the mean yearly secular change of dip at Kew appears to be greater from 1860 to 1864, a period of increasing disturbances, than from 1854 to 1859, a period of decreasing disturbances. Possibly the yearly decrement of dip has again begun to diminish, since the change from 1864 to 1865 is only 1'-32. It is, however, premature to assert that this is the case, and it can only be decided by continuing the monthly observations. At all events the Kew observations agree with those at Toronto in indicating that the yearly change of dip contains the combined result of two things-namely, the true secular change and the change due to disturbance; and this ought to be borne in mind by future observers of this magnetic element.

GEOLOGICAL SOCIETY.

[Continued from p. 160.]

anuary 24, 1866.-W. J. Hamilton, Esq., President, in the Chair. The following communication was read:

"Notes on Belgian Geology." By R. A. C. Godwin-Austen, Esq., F.R.S., For. Sec. G.S.

This communication related to the Upper and Lower Kainozoic formations of Belgium, in the following order :-1. The Polders, or sea-mud beds, and their equivalents. 2. The Campine sands, and Lös, or Limon de Herbaye. 3. The Boulder formation. 4. Cailloux Ardennais. 5. The Lower Kainozoic, or Crag.

The Polders, which form a belt along the sea-board of Belgium and Holland, occasionally running inland up the courses of rivers, as up the Scheldt to Antwerp, indicate an elevation of very small amount, corresponding to the raised estuarine and other beds around our own coasts. They are covered by dunes and drifted sands. A great deal of the fen-land at higher levels, with peat and bog-iron, belongs to the age of the Polders, and of still earlier times, inasmuch as the Polders very generally overlie a terrestrial surface. The Campine sands, which run parallel with the coast from North Holland to

wards Antwerp, but within the Polder-belt, were conjectured, from their composition and on other considerations, to have been derived from sands carried inland away from dunes of the Boulder-formation period. The Lös, which is of freshwater origin, resulted from the annual depositions of melted snow-waters. The dispersion of the Cailloux Ardennais was referable to another and earlier stage of a period of cold, and when the axis of the country had a greater relative elevation than at present. These views were supported by reference to the coast-section at Sangatte.

The Boulder formation proper is only slightly represented in some of the sections about Antwerp.

With respect to the Lower Kainozoic series, the author preferred the divisions proposed by M. Dumont (Scaldésien and Diestien) to the minute subdivisions of Sir C. Lyell and M. Nyst. The exceedingly narrow vertical dimensions of the Crag, and the manner in which, along the continuous sections now exposed, one bed of the Scaldésien Crag replaces another, are new facts, and preclude any systematic order of sequence, founded on percentage comparisons, from local assemblages of fossils.

:-a

The Antwerp Crag series presents two conditions of sea-bed :deepish-water and life-zone formation, corresponding to the oozedepths of existing seas; this is the Diestien of Dumont, or Lower Crag on an eroded surface of this, there occurs at Antwerp an upper series of coarser sands, shingle, and gravel, together with much which has been derived from the lower; this is the Scaldésien. The change from one to the other indicates a change as to depth over the Crag sea, and the result has been an admixture of the characteristic materials of distinct sea-zones.

The original boundary line of the Crag sea is traced, as also the great breadth of the drift-sand zone, over the Belgian area; thiscoupled with the consideration that the Crag-sea waters on the continental coast-line nowhere came in contact with any beds older than Nummulitic, such as Tongrien and Bruxellien, even as high as Denmark, whilst on the English side, from Suffolk north, its coastline was of chalk with flints-indicates a closed sea on the south, since only by such an arrangement could the flint-gravel be carried along.

The differences between the Crag-fauna of England and of Belgium were explained in accordance with bathymetrical distribution. The Scaldésien beds of Antwerp contain an assemblage which is composed in part of a littoral fauna, and in part of that of oozedepths. The Red Crag of Suffolk differs from the Scaldésien in being more littoral in its forms, as also from containing the materials of a Bryozoan zone.

The Bolderberg beds, which afforded M. Dumont his evidence in favour of his " Système Boldérien," were shown to have been

XXXVII. Intelligence and Miscellaneous Articles.

ON THE CHANGES WHICH STRETCHING AND THE PASSAGE OF A VOLTAIC CURRENT PRODUCE IN A MAGNETIC BAR. BY M. VILLARI OF naples.

MATTEUCCI states that if a bar of hard iron which is magnetized

by a spiral is stretched, the magnetism of the bar increases. If the stretching ceases, the magnetism again diminishes. If the same experiment is made with soft iron, the reverse is the case. If the magnetizing spiral is not at work, stretching also causes an increase of magnetism, and relaxation (nachlassen) a diminution.

Wertheim has repeated Matteucci's experiments, but only confirms his results in case the magnetizing spiral is at work. He adds at the same time that the deflection of the galvanometer observed is smaller each time the stretching is repeated and the more the wire is straightened. He doubts, therefore, whether with a perfectly straight wire, stretching alters the magnetism of the wire.

As no further statements exist respecting stretching, it appeared desirable to make a few new experiments on the subject, in order to explain the difference between Matteucci's and Wertheim's experiments. The apparatus consisted, like Matteucci's, of two spirals, a magnetizing and an inducing one, in the latter of which a Wiedemann's mirror galvanometer, with sliding coils, was inserted.

In this method it is not the existing magnetism of the iron or steel bar which is measured, but the inducing-action which a change in it produces. But the subesquent investigation only refers to changes; and these may very well be measured by the induction which they produce.

The apparatus employed was the following:

One spiral was firmly fixed in the other, and they were well fastened in a frame which stood on a table, and in such a manner that their common axis was from east to west. The galvanometer was at a distance of 4 to 5 metres from the spirals. The bar of steel or of iron which was to be used for experiment, was inside the inner spiral. At each end of the bar a thick brass wire was soldered, or else screwed. In order that neither these wires nor the bar to which they were fixed should move laterally, they passed through two pieces of wood which were firmly united with the frame in which were the spirals; and they also passed through corks which stuck in the ends of the inner spiral. One end was fastened to the stand by means of a screw, at the other was a rope which passed over a pulley, and by means of a lever could be stretched by 240 pounds.

The steel and iron bars, before being placed in the apparatus, were moreover straightened as much as possible, and after they were inserted were repeatedly drawn and, in order that they might be straight, were stretched by 40 pounds.

Both the spirals were coiled on brass sheaths slit lengthwise. The external magnetizing spiral, which consisted of covered copper wire 2.4 millims. in diameter, was 580 millims. in length by 225 millims. external and 110 millims. internal diameter. The interior spiral consisted of covered copper wire 1 millim. in thickness and 600 millims. in length; this had an external diameter of 30 millims., and an internal one of 19 millims.

The bars of iron and steel were all several inches shorter than the inner spiral.

The results of this investigation may be summed up as follows:With a closed magnetizing spiral, stretching and relaxing, the stretching produces an increase of the magnetic momentum up to a certain point; in like manner, opening and closing a voltaic current sent through iron or steel bars produces an increase of its magnetic momentum up to a certain amount.

If this limit has been attained, in further stretchings (as well as in openings and closing of the current which is passed through) the magnetism oscillates about this limit; in a steel or iron bar, stretchings diminish the magnetism, if it is thin and strongly magnetized, while relaxation of the stretching produces just as great an increase. But if the bar is thick and powerfully magnetized, an increase is produced on stretching, and on relaxation a diminution.

Yet, in transmitting a current, there is a difference between steel and iron when this limit has been attained. In the case of iron, there is an increase on closing the current, and on opening just as great a diminution, whatever be the direction of the current. Steel exhibits the same behaviour if the current enters at the south pole of the magnet; but if it enters at the north pole, there is conversely a diminution on closing, an increase on opening.

In the case of an open magnetizing spiral, stretching and relaxation produce a diminution of magnetism, to a certain amount. In like manner the passage of a current and its interruption produce a diminution of magnetism, up to a certain magnitude.

When this has been reached, on further stretching or on repeated opening or closing the current, the magnetism oscillates about this limit; and in soft iron stretching produces an increase, while relax. ation produces just as great a decrease. Hard steel exhibits exactly the opposite behaviour; and between hard and soft steel all intermediate stages may be observed. Closing a current produces, in the case of iron, an increase of magnetism; opening it, just as great a decrease, whatever be the direction in which the current passes through the bar. The same is the case with steel if the positive current enters at the south pole.

Up to a certain limit stretching acts like a relaxation, and opening and closing a current like a mechanical agitation. So far, therefore, these first actions of traction and of the current are connected with other known phenomena. The phenomena, however, which occur after this limit is attained are new.