3. On the Geological and Physical Development of Guadeloupe.' By Prof. J. W. Spencer, Ph.D., M.A., F.G.S.

The Guadeloupe group is separated from the Antigua and Dominica groups by depressions 2000 feet deep. Much of Guadeloupe itself consists of eruptive rocks, evidently as old as the igneous base of Antigua. The lowest beds of Grande Terre are yellow tufa, surmounted by 75 or 80 feet of volcanic sand of e irly Tertiary age. A calcareous formation conformably follows, dipping north-eastward. These two formations seem to correspond with the Oligocene rocks of Antigua. The Lafonde Gravel and Marl succeeds them unconformably, and it is possible that the limestone of the Usine of Pointe à Pitre is of about the same general age. In addition to these formations there are raised coral-reefs, consolidated calcareous sands, alluvia, the loams and gravels of the Petit Bourg Series, and various fragments of calcareous groups. The tooth of a small Elephas, allied to the Maltese type, and found in Grande Terre, is mentioned.

The land-surface during the Mio-Pliocene period appears to have been 2000 feet above the present level, but it was submerged 200 feet at the close of the Pliocene period during the accumulation of the Lafonde and Lower Petit Bourg gravels and loams. There was a re-elevation of about 3000 feet in the early Pleistocene period, and during this epoch Elephas could have crossed from the continent. This was followed by a depression to 100 feet or more below the present level, a re-elevation to 150 feet, submergence below the present level with growth of corals, and the elevation of these to 6 or 8 feet above the sea.

4. On the Geological and Physical Development of Anguilla, St. Martin, St. Bartholomew, and Sombrero.' By Prof. J. W. Spencer, Ph.D., M.A., F.G.S.

Deep channels, not less than 1800 feet deep, separate the bank on which this group is founded from the banks to the north and south. The oldest rock of St. Martin and St. Bartholomew consists of greenstone or dioritic porphyry usually much decayed, followed by altered limestones, and volcanic ashes and breccias. The calcareous divisions are associated with chert and deposits of manganese. Fossils found in these rocks in St. Bartholomew determine the age as equivalent to the Middle Eocene of Europe. A white limestoneformation, which appears to correspond with the limestone-series of Antigua, follows unconformably. The limestone is partly phosphatized at the surface and is pitted by caverns. It is apparently succeeded by upper strata, with a modern fauna, similar to that of the Pointe à Pitre Limestone of Guadeloupe. The limestones are unconformably covered by mantles of breccia, gravels, and sand, which may be regarded as the equivalent of the Columbia formation of the American Continent. The St. Martin plateau was a landsurface throughout the Mio-Pliocene period, during the earlier part of which it appears to have stood 2500 feet above its present level, and was probably connected with the now neighbouring insular

masses, from which it was disconnected by denudation during a very long period of atmospheric activity, followed by a subsidence, so as to bring the present surface of the submarine banks to a level so low that the undulating features of a base-level of erosion could be formed on them; for, during the period when the deep and broad depressions on the Antillean chain were being fashioned, the now isolated island-groups stood out as table-mountains, which were slowly being eaten away by atmospheric agents. There was next a subsidence to about 200 feet below the present level, about the close of the Pliocene period, followed by a re-elevation to 3000 feet, as shown within the area, but in reality much more. It was during this early epoch of the Pleistocene that the great rodents described by Prof. Cope reached here from South America, but the race continued to live here sufficiently long to give rise to distinct species. The submergence of the mid-Pleistocene period was to the extent of about 200 feet, and the subsequent elevation was marked by moderate denudation with the production of shallow watercourses, traceable across the sunken banks to depths of 150 or 180 feet. Again there was a moderate depression sufficient to bring the surface a few feet below the present level, succeeded by a rise during which the low shell-bearing sands were formed.

5. On the Geological and Physical Development of the St. Christopher Chain and Saba Banks.' By Prof. J. W. Spencer, Ph.D., M.A., F.G.S.

The St. Christopher (St. Kitt's) ridge rises from 2000 to 2800 feet above the submarine Antillean plateau, and is for the most part covered with shallow water, except between St. Kitt's and Montserrat, where a depression reaches 2592 feet, and between Statia (St. Eustacius) and Saba, where it reaches 1200 feet. Relics of old igneous formations are found on the islands, but in most places they are covered by more recent volcanic formations.

The Brimstone Hill Limestone is the succeeding formation, which appears to be newer Pliocene or Pleistocene, and to correspond with the Upper Marls of Anguilla and those at the Usine of Pointe à Pitre in Guadeloupe.

The St. Kitt's Gravels succeed, and in beds of apparently the same age shells of living species have been found at an altitude of 300 feet. The main volcanic activity belonged to the mid-Pleistocene period. It is inferred that the group underwent the same physical history as the neighbouring groups of islands. First there was elevation, followed by subsidence. Then came the second great elevation to about 3000 feet and erosion of the region, when the deep valleys and cirques indented the margins of the tablelands, and at the same time the great volcanic ridges were built. Next followed another subsidence to about 300 feet below the present level, and during this epoch the volcanic domes of Brimstone Hill and the 'Quill' of Statia were formed. The succeeding upward movement carried the land 60 feet or more above the present level, when ravines and small channels in the sunken shelf were excavated. Another depression to 40 or 50 feet filled up these ravines. Then came final re-elevation, and it is possible that a downward movement is now in progress.

THE

LONDON, EDINBURGH, AND DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

[SIXTH SERIES.]

OCTOBER 1901.

XXXIII. The Thermomagnetic and Thermoelectric Properties of Crystalline Bismuth. By LOUIS LOWNDS, B.Sc., 1851 Exhibition Research Scholar, Univ. Coll. Nottingham *.

§ 1.

FOR

has

OR the development or verification of a theory of the Thermomagnetic Phenomena discovered by A. v. Ettingshausen and W. Nernst † in 1886, it appears necessary to determine for one plate the several quantities which come into consideration. E. van Everdingen already determined for a plate of electrolytic bismuth the several data, but the observations were made within very narrow limits of temperature and no experiments were carried out at low temperatures where the effects strongly increase.

A plate of crystalline bismuth was placed at my disposal by Prof. Groth of München, to whom my best thanks for the specimen are due. This plate was cut from a large crystal in his possession so that the chief crystallographic axis lay in the plane of the plate and parallel to its length. It is intended to to measure the several quantities for this plate within limits of temperature and field-strength as large as possible. The present paper deals with the Longitudinal and Transversal Thermomagnetic Phenomena and the thermoelectric force with relation to copper.

§ 2. Small irregularities on the surface and edges of the * Communicated by Prof. E. Warburg.

† A. v. Ettingshausen and W. Nernst, Wied. Ann. xxix. p. 343 (1886). E. van Everdingen, Comm. Phys. Lat. of Leyden, No. 42 (1898), No. 48 (1899).

Phil. Mag. S. 6. Vol. 2. No. 10. Oct. 1901.

Ꮓ

crystalline plate were removed so far as possible by careful rubbing on a fine emery surface. In the end there remained but a slight fracture at one corner. The dimensions of the plate were: length 16 mm., breadth 13.5 mm., and thickness A chemical analysis of the bismuth has not yet been

4 mm. made.

§ 3. Suppose in figs. 1 and 2, OX, OY, OZ are the coordinate axes, OABC the crystal and OA the direction of the chief crystallographic axis. Consider throughout the lines of

force to be at right angles to the current of heat, the former being in the direction of the X axis, the latter in the direction of the Z axis. The following cases may then be investigated :

I. For the Longitudinal Effect.

case

1. The chief crystallographic axis OA in the direction of the current of heat (fig. 1). No new appears on turning the crystal round OZ through 90° (1a).

2. The chief axis OA at right angles to the current of heat but perpendicular to the lines of force (fig. 2). 3. The chief axis OA at right angles to the current of heat but parallel to the lines of force. This is

obtained by turning OABC (fig. 2) about the Z axis through 90°.

II. For the Transversal Effect.

Although 1, 1a, 2, and 3 are here different cases, only 1 and 2 can be investigated since the plate admits of the effect being measured only in the direction of its breadth and not in the direction of its thickness.

III. For the Thermoelectric Force.

Along and at right angles to the direction of the chief crystallographic axis.

§4. The Apparatus.-The apparatus was constructed so as to avoid soldering to the crystal. Fig. 3 shows diagrammatically a vertical longitudinal section through the centre at right angles to the plane of the crystalline plate. A is a cylindrical brass vessel, 10 cm. high and 3-7 cm. diameter, covered on the outside with felt, and serves to contain the freezing mixture. To its base is soldered a rectangular piece of copper B, 1.5 cm. long, 35 cm. wide, and 6 mm. thick. C, D, and E are pieces of copper of the same dimensions as B. F is of copper, and is 3.5 cm. wide, 16 cm. square cross section; through it is bored a hole G, 5 mm. radius, which serves to carry a current of water at the ordinary temperature. The piece C is soldered to F. Screws pass through the piece of wood H into the coppers B and C, and maintain them in the position shown in the figure (p. 328). The pieces D and E are each provided with three screws a, b, c ; a', b', c'; shown in fig. 4, which is a view in the direction of the arrow fig. 3. b, e; b, e' screw into corresponding holes in the coppers B and C. By this means the crystal plate K could be clamped between D, B, and E, C, the screws a and a' adjusting the