great difficulty in the case of soft iron, and is not observed at all in the case of manganese steel. A fairly approximate numerical measurement may be made in this way: Take a block of iron or steel on which a groove is cut, and in this groove wind a coil of copper wire insulated with asbestos; cover the coil with many layers of asbestos; and finally cover the whole lump of iron or steel with asbestos again. We have now a body which will heat and cool comparatively slowly, and which will lose its heat at a rate very approximately proportional to the difference of temperature between it and the surrounding air. Heat the block to a bright redness, and take it out of the fire and observe the resistance of the copper coil as the temperature falls, due to the cooling of the block. Plot a curve in which the abscissæ are the times, and the ordinates the logarithms, of the increase of resistance of the copper coil above its resistance at the temperature of the room. If the specific heat of the iron were constant, this curve would be a straight line; if at any particular temperature latent heat were liberated, the curve would be horizontal so long as the heat was being liberated. If now a block be made of manganese steel, it is found that the curve is very nearly a straight line, showing that there is no liberation of latent heat at any temperature. If it is made of nickel steel with 25 per cent. of nickel, in its non-magnetic state, the result is the same-no sign of liberation of heat. If now the block be made of hard steel, the temperature diminishes at first; then the curve (Fig. 13) which represents the temperature bends round: the temperature actually rises many degrees whilst the body is losing heat. The liberation of heat being completed, the curve finally descends as a straight line. From inspection of this curve it is apparent why hard steel exhibits a sudden accession of brightness as it yields up its heat. In the case of soft iron the temperature does not actually rise as the body loses heat, but the curve remains horizontal, or nearly horizontal, for a considerable time. This, again, shows why, although a considerable amount of heat is liberated at a temperature corresponding to the horizontal part of the curve, no marked recalescence can be obtained. From curves such as these it is easy to calculate the amount of heat which becomes latent. As the iron passes the critical point it is found to be about 200 times as much heat as is required to raise the temperature of the iron 1 degree Centigrade. From this we get a very good idea of the importance of the phenomenon. When ice is melted and becomes water, the heat absorbed is 80 times the heat required to raise the temperature of the water degree Centigrade, and 160 times the heat required to raise the temperature of the ice by the same amount. The temperature of recalescence has been abundantly identified with the critical temperature of I am magnetism.1 not aware that anything corresponding with recalescence has been observed in the case of nickel. Experiments have been tried, and gave a negative result, but the sample was impure; and the result may, I think, be distrusted as an indication of what it would be in the case of pure nickel. The most probable explanation in the case of iron, at all events, appears to be that when iron passes from the magnetic to the non-magnetic state it experiences a change of state of comparable importance with the change from the solid to the liquid state, and that a large quantity of heat is absorbed in the change. There is, then, no need to suppose chemical change; the great physical fact accompanying the absorption of heat is the disappearance of the capacity for magnetization. What explanations have been offered of the phenomena of magnetism? That the explanation must be molecular was early apparent. Poisson's hypothesis was that each molecule of a magnet contained two magnetic fluids, which were separated from each other under the influence of magnetic force. His theory explained the fact of magnetism induced by proximity to magnets, but beyond this it could not go. It gave no hint that there was a limit to the magnetization of iron-a point of saturation; none of hysteresis; no hint of any connection between the magnetism of iron and any other property of the substance; no hint why magnetism disappears at a high temperature. It does, however, give more than a hint that the permeability of iron could not exceed a limit much less than its actual value, and that it should be constant for the material, and independent of the force applied. Poisson gave his theory a beautiful mathematical development, still useful in magnetism and in electrostatics. Weber's theory is a very distinct advance on Poisson's. He supposed that each molecule of iron was a magnet with axes arranged at random in the body; that under the influence of magnetizing force the axes of the little magnets were directed to parallelism in a greater degree as the force was greater. Weber's theory thoroughly explains the limiting value of magnetization, since nothing more can be done than to direct all the molecular axes in the same direction. As modified by Maxwell, or with some similar modification, it gives an account of hysteresis, and of the general form of the ascending curve of magnetization. It is also very convenient for stating some of the facts. For example, what we know regarding the effect of temperature may be expressed by saying that the magnetic moment of the molecule diminishes as the temperature rises, hence that the limiting moment of a magnet will also diminish; but that the facility with which the molecules follow the magnetizing force is also increased, hence the great increase of μ for small forces, and its almost instantaneous extinction as the temperature rises. Again, in terms of Weber's theory, we can state that rise of temperature enough to render iron nonmagnetic will not clear it of residual magnetism. The axes of the molecules are brought to parallelism by the force which is impressed before and during the time that the magnetic property is disappearing; they remain parallel when the force ceases, though, being now nonmagnetic, their effect is nil. When, the temperature I have only recently become acquainted with the admirable work of M. Osmond on recalescence. He has examined a great variety of samples of steel, and determined the temperatures at which they give off an excep tional amount of heat. Some of his results are apparent on my own curves, though I had assumed them to be mere errors of observation. For example. referring to my Royal Society paper, there is, in Fig. 38, a hint of a second small anomalous point a little below the larger one. And. comparing Figs. 38 and 38A, we see that the higher the heating, the lower is the point of recalescence; both features are brought out by M. Osmond The double recalescence observed by M. Osmond in steel with a moderate quantity of carbon I would explain provisionally by supposing this steel to be a mixture of two kinds which have different critical temperatures. Although M. Osmond's methed is admirable for determining the temperature of recalescence, and whether it is a single point or multiple, it is not falling, they become again magnetic, the effect of the direction of their axes is apparent. But Weber's theory does not touch the root of the matter by connecting the magnetic property with any other property of iron, nor does it give any hint as to why the moment of the molecule disappears so rapidly at a certain temperature. Ampère's theory may be said to be a development of Weber's: it purports to state in what the magnetism of the molecule consists. Associated with each molecule is a closed electric current in a circuit of no resistance; each such molecule, with its current, constitutes Weber's magnetic molecule, and all that it can do they can do But the great merit of the theory- and a very great one it is-is that it brings magnetism in as a branch of electricity; it explains why a current makes a magnetizable body magnetic. It also gives, as extended by Weber, in explanation of diamagnetism. It, however, gives no hint of connecting the magnetic proporties of iron with any other property. Another difficulty is this: When rợn ceases to be magnetizable, we must assume that the molecular currents cease. These currents represent energy We should therefore expect that, when iron ceased to be magnetic by rise of temperature, heat would be liberated; the reverse is the fact. So far as I know, nothing that has ever been proposed even attempts to explain the fundamental anomaly. Why do iron, nickel, and cobalt possess a property which we have found nowhere else in nature? It may be that at lower temperatures other metals would be magnetic, but of this we have at present no indication. It may be that, as has been found to be the case with the permanent gases, we only require a greater degree of cold to extend the rule to cover the exception. For the present, the magnetic properties of iron, nickel, and cobalt stand as exceptional as a breach of that continuity which we are it the habit of regarding as a well proved law of Nature. NOTES ON A RECENT VOLCANIC ISLAND IN THE PACIFIC. N 1867, H.M.S. Falcon reported a shoal in a position in about 20° 20' S., and 175° 20' W., or 30 miles west of Namuka Island of the Friendly or Tonga Group In 1877 smoke was reported by H.M.S. Sappio to be rising from the sea at this spot. In 1885 a volcanic island rose from the sea during a submarine eruption on October 14, which was first reported by the Janet Nichol, a passing steamer, to be 2 miles long and about 250 feet high. The U.S.S. Mohican passed it in 1886, and from calculation founded on observations in passing, gave its length as 1 miles, height 165 feet. The crater was on the eastern end, and dense columns of smoke were msing from it. In 1887 the French man-of-war Decres reported its height to be 290 feet. In the same year an English yacht, the Sybil, passed it. and a sketch was made by the owner, H. Tufnell, Esq., which is here produced. The island has now been thoroughly examined and mapped, and the surrounding sea sounded by H.M. surveying-ship Egeria, Commander Oldham. It is now 1 mile long, and of a mile wide, of the shape given in the accompanying plan. The southern portion is high, and faced by cliffs on the south, the summit of which is 153 feet above the sea. A long flat stretches to the north from the foot of the hill. The island is apparently entirely formed of ashes and cinders, with a few blocks and volcanic bombs here and there, especially on the verge of the hill. Under the action of the waves, raised by the almost constant south-east winds, this loose material is being adapted to determine the quantity of heat liberated, as the small sample rapidly removed; continual landslips take place, and used is inclosed in a tube of considerable mass, which cools down at the same time as the sample experimented upon. Commander Oldham is of opinion that the original summit was some 200 or 300 yards southward of the present highest cliff, and that the shallow bank stretching to the south represents the original extension of the island. As far as can be judged from Mr. Tufnell's sketch from the north-west and that of the Egeria from the south south-east, considerable changes have taken place in two years, the different summits shown in the former having disappeared as the sea has eaten away the cliffs. The flat to the north seems to be partly due to redistribution under the lee of the island of the material removed from the southern face. It is crossed by curved ridges from 3 to 12 feet high, which Commander Oldham considers to have been formed as high beaches during spring tides and strong winds, the flat ground between them, almost at the level of the water, being deposited under normal conditions of weather. The island is thus gaining on one side, while losing on the other, but when the high part has gone, this partial recovery will probably cease. A little steam issuing from cracks in the southern cliffs was the sole sign of activity, but a pool of water at a temperature of from 91° to 113° F., water which rose in a hole dug in the flat of a temperature of 128° F., and a temperature of 100° F. in a hole dug half-way up the slope, also show that the island still retains heat near the surface. The water is sea-water that has filtered through the loose ashes, and it rose and fell with the tide. It appears by the condition of the flat that the island has neither risen nor subsided during the past two or three years. It will be interesting to watch the ultimate fate of this last addition to the Pacific isles, but it seems probable that its existence as an island will be short unless a hard core is yet revealed. The soundings between Falcon Island and Namuka show that they are separated by a valley 6000 feet deep. Metis Island, 73 miles north-north-east of Falcon Island, is another volcanic cone that appeared a few years before the latter, but has not yet been examined. W. J. L. WHARTON. WEATHER FORECASTING. POPULAR interest in weather prediction shows no sign of abating. The January number of the Kew Bulletin is devoted to an account of Herr Nowack's socalled "weather plant," and its failure as an indicator either of coming weather or of earthquakes. Very recently a lively correspondence has been carried on in the daily press on the merits or demerits of the forecasts issued by the Meteorological Office. Accordingly, some remarks on the subject in the columns of NATURE may not be out of place. One critic says that the forecasts are little better than haphazard guesses, and that the money devoted to them would be better spent on an additional lifeboat or two on the coast. Another says the forecasts are not worth the paper they are printed on, and wishes that the Office published in the newspapers fuller accounts of the weather reported from the coasts. The fact is that the Office is compelled by public opinion to issue forecasts. The public will have its forecasts, as in 1867 it would have its storm-warnings, notwithstanding the reluctance of meteorologists to issue either the one twenty years ago or the other at present. It can hardly be doubted that, for these islands at least, conscientious meteorologists would be disposed to agree with Arago, who said in 1846, and printed it in italics in the Annuaire du Bureau des Longitudes: "Jamais, quels que puissent être les progrès des sciences, les savants de bonne foi et soucieux de leur réputation ne se hasarderont à prédire le temps." We are, of course, On the other hand, at other points the forecasts may be frequently unsuccessful. In one important particular not only our own Office but all other Offices in Europe, signally fail, and that the quantitative prediction of rain. No one is able parently, to predict whether the amount of rainfall the morrow will be a tenth of an inch or a couple of inches. No sudden floods have ever yet been foreto By this we are not speaking of predicting the approach of floods to the lower valleys from rain which has already fallen on the upper reaches of a river, for that is ne meteorological prediction at all. With the necessarily incomplete character of the m formation reaching head-quarters, the wonder is that the Office can attain such success as it does. The mar deficiency in the information is in its quantity, and this seems to lie at the door of the Postal Telegraph Ofice which insists on being paid for its telegrams. If meteoro logical messages were transmitted gratis, we might expec to hear at frequent intervals from our outposts, instead of twice, or, at most, thrice in the twenty-four hours: r fact, from several stations we can only hear once, the cost of more telegrams being prohibitive. It is self evident that such an amount of information is quite insufficient. The weather will not abstain from changing because the hour for a telegraphic report has not arrived The information contained in the telegrams is and deficient in quantity, for the reporters cannot, within the prescribed form of their messages, communicate all impressions which the ever-varying appearance of the sky may have conveyed to their minds. A skilled cloud observer, who has leisure to practise his powers, is one able to form a very correct idea of what is coming for the region bounded by his own horizon, but he is quite unable to give the benefit of his observations and expert ence to a friend in another county by telegraphing the information. The greatest want which the Office finds in its observer is skill in cloud observation, and it appears to be the that a cloud observer nascitur non fit, and that it is nev to impossible to teach the art to a new hand, at least by correspondence. Instrumental records of the phenomena taking place 13 the higher strata of the atmosphere are of course una tainable, and it is only by carefully watching the upper clouds that we can gain any notion of changes taking place up there, but, by means of such watching, M: Clement Ley is able to predict with nearly perfec certainty the weather for the Midlands-his own ne bourhood. It must always be remembered that the forecasts are drawn for districts, not for individual stations; and disregarding the amount of correctness claimed by the Office by its own checking of its work, they attain a very creditble amount of success when tested by independent observers. This happens even in the summer-time, the very season at which a recent critic said that the forecasts for one month, if shuffled about, and drawn at random from a bag, would suit just as well for the next! This is proved by the results of the hay harvest forecasts, which are deduced from the reports of the recipients, practical agriculturists. The following is the table for the season of 1888, the latest for which the figures are available : To give an idea of the difficulty of obtaining accurate opinions from outsiders as to the value of storm-warnings, which are a class of forecasts, it may be interesting to give some specimens of reports. Inquiries were made in 1882, from all the stations where gnals are hoisted, as to their correctness and general tury From Tynemouth the answer was that "these gnals have been, and will be, an inestimable boon to our afaring population." From South Shields, just opposite Tynemouth, the reply to a recent official inquiry was nat "the warnings were not a ha'porth of use, and that no one minded them." Each answer merely represented the private opinion of the person who uttered it. The reader can see that there is some difficulty in king out the actual truth from such a heap of inconTous statements as the foregoing are certain to furnish. R. H. S. THE LABORATORIES OF Bedford COLLEGE. BEDFORD COLLEGE, in York Place, Baker Street, which was one of the earliest institutions devoted to he higher education of women, is taking a leading part 2 providing facilities for their instruction in science. Founded long before Oxford and Cambridge condescended to the "weaker sex" (which has since proved strong enough to attain to the highest place in the Lassical Tripos), it is the result of the work of en husiasts who would not admit the possibility of defeat. has had to struggle not only against the inevitable difficulties due to its early foundation, but against the apathy of London. Provincial towns feel that their honour is involved in the success of their institutions. The Colleges for women at Oxford and Cambridge share in the picturesque surroundings of those old homes of learning. They attract attention and interest by their situation amid scenes and traditions of which the whole English-speaking race is proud. Bedford College has had no such advantages. London institutions are regarded as either Imperial or parochial-as too large or too small to interest its citizens as such. Bedford Square compares unfavourably with the "backs," and it is impossible to regard York Place with that gush of emotion which "the High" sets free. Thus it is that, although Bedford College has been at work since 1849, and though one in every four of the whole number of women who have gained degrees of the University of London has been a student in its classes, the work of the College does not yet receive the meed of public appreciation which it has fairly earned. Bedford College is for women what University and King's Colleges are for men. It provides, within easy reach of all Londoners, an education which is tested by the severe standard of the University of London, and bears the hall-mark of success. One-third of its students are aiming at degrees, and their presence in the class-rooms, their work in the examination-hall, guarantees the quality of the teaching they receive to class-mates who do not intend to face the same ordeal. Science has for long been taught in Bedford College, but there has been a pressing need for better laboratories and class-rooms. These the Council has now provided. A new wing has been built, dedicated to the memory of the late Mr. William Shaen, who worked long and devotedly for the College. About £2000 is required to complete and fit up this building free of all debt, and Mr. Henry Tate, who had already given £1000 to the fund, has promised to supplement it by a like amount if the Council on its part can raise the other moiety of the deficit. It is too probable that this sum will only be obtained by an exhausting effort, but surely it is not too much to hope that the public may at last appreciate the importance of promoting the higher education of women in London. In a northern manufacturing town the money would be forthcoming in a week. As regards the laboratories, it may be sufficient to say that Dr. W. Russell, F.R.S., is the Chairman of the Council, and that they have been built under his general supervision. They appear to be in all respects suited to the purposes for which they are intended. The physical laboratory and lecture-room are on the ground floor. The former has a concrete floor, and is well lighted, partly by windows, partly by a skylight. It looks out upon East Street, and is therefore removed as far as possible from the effects of the heavy traffic in Baker Street. The chemical laboratory is at the top of the house, and opens into a class-room which is fitted with all the usual conveniences for experimental illustration. It is surely a hopeful sign that a College for the higher education of women should now be regarded as incomplete unless it controls physical and chemical laboratories specially designed and fitted for the delivery of lectures and the performance of experiments. These Bedford College now possesses. We can only hope that it may soon possess them free of debt. The Editor of NATURE will be happy to receive and forward to the College authorities any subscriptions which may be sent to him for that purpose. STEPHEN JOSEPH PERRY, F.R.S. ON the evening of January 4 a telegram from Demerara announced that there had been a successful observation of the eclipse of December 22, and that Father Perry had succumbed to dysentery. Stephen Joseph Perry was born in London on August 26, 1833, and received his early education at Gifford Hall School. Having decided to enter the priesthood, he went |