and strata in which the several forms were obtained, the author attempts some observations about the fashions and the relative age of flint implements, but they have neither any special importance nor novelty.

The latter part of the book describes some of the relics of primeval man as found on the banks of the Lea, from its source near Dunstable, in Bedfordshire, to London; with a description of the already known primeval living-place or "palæolithic floor" at Stoke-Newington, London.

In Chapter XIII. we have traces of Primeval Man elsewhere, near the Lea, in South Beds and North Herts.

Chapters XIV. & XV. Traces of Primeval Man near the Lea in Middlesex and Essex, from Waltham to Tottenham, on the BorderLine of London; and from Tottenham to the Junction of the Lea with the Thames at Blackwall: with notes on Stone Weapons and Tools-the Most Ancient Implements; Implements of Medium Age; Implements of Least Palæolithic Age; Pieces of Implements Conjoined and Flakes Replaced; How Stone Implements were Made; worked Wood and Bone; Fossils as Beads; Fossil Bones; Shells of Land- and Freshwater Mollusca; Plant-Remains; Forgeries of Implements.

Chapter XVI. Mesolithic Implements.

Chapter XVII. Palæolithic Stones found by Neolithic Men and Reworked.

Chapter XVIII. Neolithic Implements and Keltic Relics, from South Bedfordshire to London :-Implements of Stone and Bronze; Earthworks, Roads, and Trackways near Dunstable; British Hut Foundations near Dunstable; Dene-Holes; Graves, Tumuli, Skeletons, Bones; Place-Names near Dunstable.

It is evident, and indeed partly stated, that Chapters XIII. to XVIII. consist of renewals and reprints of earlier notes and memoirs by the author and others, published in the Journ. Anthropological Institute, Proceed. Geologists' Association, Quart. Journ. Geological Society, Nature,' Natural Science,' Proceed. Essex Field-Club, 'Ancient Stone Implements of Great Britain,' 'Archæologia Cambrensis,' &c. In such a compilation, however, the references to papers by fellow-workers should be complete, and careful acknowledgement of all sources of information should be made in every case, so that the reader should distinguish what is not, from what is original, in the book before him.

The printing and paper are good. "All the Illustrations are original or taken from original sources." They have been admirably drawn (some after photographs) and reproduced on wood or otherwise, by the Artist-engraver, Mr. W. G. Smith, himself. The "archæological map of the Caddington and Dunstable district" clearly defines the topography of the author's discoveries; and another map shows a part of the north-eastern neighbourhood of London. The numerous illustrations comprise skulls, bones, stone tools and flakes, sections, local views, a few shells and fossils, and four neat and expressive ideal portraitures of the aborigines. We

are sorry that two of these clever pre-Raphaelistic pictures are not lodged in the text; one is on the title-page, and one indeed is seen only gilded on the cover.

Altogether, this interesting work, partly descriptive of a new set of discoveries, and partly a compilation of more or less relative matter, is calculated to incite a taste for seriously hunting up the history of Primeval Man; and, by showing how and where the study may be practically followed, it will satisfy, to some extent, many enquiring minds, putting them on the right lines for investigation, and affording some clues to other and more important systematic works on archaic anthropology.

LIX. Intelligence and Miscellaneous Articles.

ON THE MAGNETIZATION OF IRON AND NICKEL WIRE BY RAPID ELECTRICAL OSCILLATIONS. BY PROF. I. KLEMENČIC.

BY the aid of the formulæ of Lord Rayleigh and Stefan the author

endeavoured to determine the strength of magnetization, or, in other words, the value of μ, from the disengagement of heat which occurs in a magnetizable wire when electrical vibrations are passed through (number of vibrations about 9 × 107). The development of heat was measured by means of a delicate thermoelement near the wire under experiment, and was each time compared with the disengagement of heat in a non-magnetized wire. Observation gave the following values for μ:-soft iron, 118; steel (piano wire), soft 106, hard 115; Bessemer steel, soft 77, hard 74; nickel, 29. These values agree well with those which Baur and Lord Rayleigh found for very feeble magnetic forces. As the experiments of these observers teach, the permeability is a constant magnitude up to certain values of the magnetizing force while it thus rapidly increases. The present observations show that in these experiments μ varies within a certain range. This fact can be explained either by assuming that the magnetizing forces used here are very feeble, and of the order of magnitude of those strengths of field in which μ is really constant; or by assuming that we are dealing with much greater magnetizing forces, but that the magnetization cannot follow the rapid change so quickly as to reach that part of the curve of magnetization which corresponds to the variable and far greater values of μ. An approximate estimation of the field-strengths under consideration shows that at any rate on the surface of the wire and at the beginning of the oscillations we have magnetizing forces which are several hundred times as great as that limit within which μ is constant. There would accordingly in this case be a retardation of the magnetization, which, however, must not be confounded with hysteresis. It must in this be assumed that the results of Baur and of Lord Rayleigh which refer to longitudinal magnetization are also applicable to circular magnetization.

μ

Within the limits of the constant there is no remanent magnetism; the magnetization in this region is similar to the deformation of a body within the limits of elasticity, while the further stages

of magnetization are to be compared with permanent deformations; an analogy which Maxwell had already pointed out. That part of the magnetization which can be applied technically is in the region corresponding to the permanent deformations. It is very probable, and the assumption is confirmed by experiment, that the magnetization in very rapid changes of field does not attain this region, while the molecules within the region of the constant μ can follow far more rapid oscillations than those here used. Further experiments, which will perhaps best be made by discharges of condensers, must decide this point.-Wiener Berichte, March 1, 1894.

RING-ELECTROMAGNET FOR PRODUCING STRONG FIELDS.
BY H. DU BOIS.

The large ring employed, made of the best Swedish iron (50 cm. mean diameter, 10 cm. thickness), is slit radially in one place, tangentially in another, so that the width of the former air-gap may be conveniently varied. It bears 12 coils, each having a resistance of 0.2 ohm and covering 20° () of the circumference; by means of these a maximum magneto-motive force of 108,000 ampere-turns may be applied, requiring about 6-5 horse-power; the highest value of the inductance is about 180 henries, corresponding to a "time-ratio" of 180/24-75 seconds; the mean values are of course less. Ballistic measurements are none the less out of the question.

When provided with flat pole-pieces the apparatus represents the simplest possible type of magnetic circuit, and is therefore well suited for an experimental verification of its laws. By a special method, which is described in extenso, the author was able to verify his form of the theory of magnetic circuits (Phil. Mag. Nov. 1890; Wied. Ann. xlvi. p. 491, 1892) with sufficient approximation. The leakage was roughly determined by a compass; its absolute amount was found to reach a maximum for a comparatively low magnetizing current, and to diminish considerably as the current was increased. This final decrease of leakage, which of course is more marked still when referred to unit flux of induction through the circuit, was previously found with a smaller ring by H. Lehmann (Phil. Mag. April 1893) and discussed by the author (The Electrician,' xxix. p. 450, 1892). It was explained how this simple fact constitutes an "experimentum crucis" against the fallacy of assuming-arguing from analogy with a voltaic battery immersed in an electrolyte-that leakage must always increase when saturation is being approached. It was also found that leakage is considerably diminished by the ampere-turns near upon the air-gap: the "pole-coils" producing them fulfil the useful function of keeping the flux of induction together, thus preventing the lines from spreading.

The principal object of the apparatus is the production of strong fields by concentrating this flux through properly shaped polepieces. A theory of the latter was developed almost simultaneously (1888) by Stefan and Ewing and Low, in which the

assumption of absolute saturation is made. As this condition can never be fulfilled in practice, the author redetermined the best shape for the poles in an empirical way. It was found-in good agreement with Prof. Ewing's prediction-that it is best to take a straight truncated cone of semi-vertical angle 60° (instead of the theoretical 54° 44'). The field obtainable in air with such polepieces falls short of the theoretical value by several thousand units; however, it beats the previous record (about 30,000 C.G.S.) by a considerable amount. It was found

5 mm.; H = 36,800 C. G. S. a = 3 mm.; H = 38,000

where a denotes the diameter of the small truncated faces.

To give an idea of this field the author points out that a bit of good thin iron wire would easily get saturated up to I=1750, corresponding to an induction

B 38,000+4 x 1750 = 60,000 C.G.S.

=

This corresponds to a tension of B2/8 dynes per cm.2, or 144 Kg-weight per cm. A concentrated aqueous solution of ferric chloride in a U-tube (Quincke) would be lifted about half a metre; finally, the resistance of a bismuth spiral would be nearly trebled in the field.

The author concludes that it is possible to obtain fields of, say, 40,000 C.G.S. with ring-electromagnets of reasonable size; but that stronger fields can probably only be reached by means out of all proportion with the purpose in view.-Wiedemann's Annalen, li. p. 537 (1894).

ON THE THERMAL BEHAVIOUR OF LIQUIDS.
To the Editors of the Philosophical Magazine.

GENTLEMEN,

The last number of your Journal contains a paper by Messrs. Ramsay and Young, to which it would not be necessary to reply if all your readers were familiar with the investigations on the thermal behaviour of liquids. But as this cannot be assumed to be the case, I am for the sake of my scientific reputation compelled to reply.

Messrs. Ramsay and Young reproach me with an error of observation of about 50° in the determinations of the critical point; as these authors well know, it would be impossible even for the most inexperienced to make so serious a mistake. The liquid I used had really the critical temperature stated, but I was in error in relying upon the statement of the chemist who supplied me, that the products were pure.

The error is acknowledged by myself in the book I published, and the critical temperatures I found are no longer given. But it is too much to assume that a physicist who once in a way had the misfortune to work with a mixture instead of a pure substance, has become incapable of making a correct observation.

Liége, May 7, 1894.

P. DE HEEN.

INDEX TO VOL. XXXVII.