January 13, 1898] 



NATURE 



261 



elected chairman, Mr. Hewitt (Liverpool) vice-chairman, Mr. J. 

 H. Nicholas (Essex) secretary, and Mr. Turner (Staffordshire) 

 treasurer. The places of the three remaining meetings for 1898 

 were fixed to take place at Birmingham, Sheffield, and London. 

 In the course of an address, the chairman remarked that the 

 Association had repeatedly expressed its acceptance of the re- 

 commendations of the Royal Commission on Secondary Educa- 

 tion and had expressed the opinion that they formed a satisfac- 

 tory basis for legislation. Latterly they had heard a great 

 deal about the essential importance of constituting a central 

 authority for education before anything else was done in the 

 matter. The constitution of an efficient central authority could 

 properly be brought about by the reorganisation of those Govern- 

 ment departments that dealt at present with educational matters, 

 a reorganisation along the lines of the Royal Commission's re- 

 port. They found it to be inconvenient in practice in many ways 

 to have two entirely separate departments, one of which they 

 were brought into contact with in connection with evening con- 

 tinuation schools, and the other more and more frequently in 

 dealing with secondary schools They wanted to see these two 

 departments merged more or less completely into one, with 

 their functions properly defined, and capable of dealing with 

 secondary education as a whole, as well as with the purely scien- 

 tific and technical part of it. The other part of the central 

 authority — the educational council — was a body to which the 

 scholastic profession naturally attached great importance. The 

 Association thought that one of the great needs of the present 

 time was a system of schools corresponding fairly closely with 

 the higher primary schools of France — schools which should be 

 a real top to the elementary education. He concluded by mov- 

 ing — " This Association considers that legislation on the lines of 

 the report of the Royal Commission on Secondary Education is 

 very urgently needed at the present time." Mr. Turner seconded, 

 and, after a short discussion, the resolution was adopted. 



SCIENTIFIC SERIALS. 



American Journal of Mathematics, \o\. xx. No. i, January. 

 — "The motion of a solid in infinite liquid under no 

 force," by Prof. Greenhill. examines the elliptic function expres- 

 sion of all the dynamical quantities involved, and explores the 

 analytical field by working out completely the simplest pseudo- 

 elliptical cases to serve as landmarks, utilising the analysis which 

 the author has developed in his paper on " Pseudo-Elliptic 

 Integrals and their dynamical applications," in the Proceedings 

 of the London Mathematical Society (vol. xxv. ) and carrying 

 out his work on the lines of his papers on the " Dynamics of a 

 Top" and on the '' Associated Motion of a Top and of a Body 

 under no Forces" in vols. xxvi. and xxvii. {Proc. L.M.S.). 

 Reference is made to the sketch of the theory in Thomson and 

 Tait's " Natural Philosophy," and to a complete solution in the 

 case of a solid of revolution in Kirchhoffs Vorlesungen iiber 

 Math. Physik, ix. ; and to the " Motion of a Solid in a Liquid," 

 by Dr. T. Craig. — " Surfaces of Rotation with constant 

 measure of curvature and their representation on the Hyper- 

 bolic (Cayley's) Plane," by G. F. Metzler. Minding (Crelle, 

 vols. 19, 20) shows that it is easy to obtain the formulae which 

 express the relations between the sides and angles of a triangle 

 of which the sides are geodesic lines on a surface of rotation 

 with constant measure of curvature. Mr. Metzler shows that 

 the same method holds for the formula expressing the area of 

 the triangle {i.e. in the ordinary spherical formulce put a- ^'i 

 for a, the radius of the sphere). — " Sur les Methodes, d'approxi- 

 mations successives dans la theorie des Equations differentielles," 

 par E. Picard, is a note, taken by the editors of the /oiirna/ (lom 

 M. Darboux's " Theorie des Surfaces (a la fin du tome iv.). — M. 

 Darboux is the mathematician whose likeness accompanies this 

 number. 



Bollettino delta Societa Sismologica Italiana, vol. iii., 1897, 

 Nos 5, 6. — The microseismographs of the Institute of Physics 

 of the Royal University of Padua, by G. Pacher. A reprint 

 of a paper giving a full account of the Vicentini microseismo- 

 graphs, already noticed in Naiure. — The Latian earthquake 

 of >Iay 8, 1897, by G. Agamennone. A note on a series of 

 slight shocks felt in the neighbourhood of Rome. — The Royal 

 Geodynamic Observatory of Catania, by A. Ricco. — Notices of 

 earthquakes recorded in Italy (April 27-May 14, 1897), by G. 

 Agamennone, the most important being the series of Latian 

 earthquakes of May 8, and an earthquake of distant, but un- 

 known, origin on May i. 



NO. T472, VOL. 57] 



SOCIETIES AND ACADEMIES. 

 London. 



Royal Society, December 9, 1897.— "On Methods of 

 making Magnets independent of Changes of Temperature ; and 

 some Experiments upon Abnormal or Negative Temperature 

 Coefficients in Magnets." By J. Reginald Ash worth, B.Sc. 



The object of this paper was to find what kinds of iron or 

 steel are least liable to changes of magnetic intensity of a cyclic 

 nature under moderate fluctuations of temperature, such as 

 take place in the atmosphere from one season to another. 

 The subject is of importance for the sake of improving 

 magnetic instruments, but apart from its practical conse- 

 quences the investigation points to some interesting theoretic 

 consequences. 



In general the effect of alternately heating and cooling a 

 magnet is to cause a large loss of magnetic intensity, which is 

 only in part recovered ; ultimately a cyclic state is established 

 in which the changes may be expressed by a formula which in 

 conformity with custom is here written 1/ = 1/ (i-o/' — /),. I 

 standing for the magnetic intensity, t and t' for the cold and hot 

 temperatures. 



Hitherto a, the temperature coefficient of the magnet, has 

 been found to be positive ; that is to say, the effect of a rise of 

 temperature is to diminish, and a fall of temperature to increase 

 the magnetic intensity. A negative coefficient must be under- 

 stood, therefore, to represent a rise and fall of magnetic intensity 

 with rise and fall of temperature, and this abnormal effect has 

 now for the first time been observed to be general in certain 

 cases. 



In the first place the influence of chemical composition was 

 sought in determining the behaviour of a magnet under changes 

 of temperature, and steel alloys severally of manganese, tungsten, 

 cobalt, and nickel were tested, as well as a series of cast irons of 

 different blends of pig irons. The results obtained show that 

 the influence of chemical constituents is subordinate to that of 

 physical condition, annealing or hardening. Thus a kind of 

 nickel steel, the same as Dr. Hopkinson found to yield such 

 remarkable thermo-magnetic results, exhibited in the glass-hard 

 state a small increase of magnetic intensity with increase of 

 temperature and decrease with fall of temperature. When 

 annealed the converse effect took place, and these effects could be 

 changed repeatedly by changing from hard to soft and soft to hard. 

 Again cast iron, in the condition received from the foundry, 

 has a very large magnetic variation for a given range of tempera- 

 ture, but when hardened by heating and rapidly chilling the 

 variation becomes exceedingly small, the coefficient now being 

 about yVth its former amount. As hardened cast-iron magnets 

 have also a very high permanent magnetic intensity, and are 

 very little influenced by shocks and blows, it will be seen that 

 they have exceptionally valuable qualities. 



However, the most interesting results were obtained from 

 some steel music wires which happened to be put to the test. 

 These wires in the normal condition and when the cyclic state 

 is established, after experiencing the usual permanent magnetic 

 loss, exhibit a negative coefficient, the higher and lower inten- 

 sities now corresponding to higher and lower temperatures, and 

 they are thus quite exceptional. But, in contrast with the similar 

 behaviour of the nickel steel alloys just mentioned, the negative 

 coefficient is destroyed and a positive coefficient established if 

 the wire be raised to a bright red heat and cooled either very 

 slowly, as in the process for annealing, or rapidly as for harden- 

 ing. If heated, however, to about a dull red and quenched, 

 the coefficient can then be rendered just zero. 



In order to gain some insight into the cause of this abnormal 

 negative coefficient, the wire was dissolved in nitric acid and at 

 difterent stages of dissolution tested. This led to the important 

 relation, which was afterwards fully established in every possible 

 way by a .series of laborious experiments, that the longer and 

 thinner the wire the more negative is its coefficient ; and so 

 much does the coefficient depend on the ratio of length to 

 diameter, or dimension ratio, that by suitably altering the 

 dimension ratio the sign of the coefficient may be changed ; and 

 for some particular dimension ratio, for a given kind of steel 

 wire, the coefficient may be made zero, and the magnet is then 

 independent of temperature changes. 



The diagram exhibits these changes graphically. Curve (i) 

 represents the variation of the coefficient from positive to nega- 

 tive for steel music wire of diameter 0*187 centimetre, and of 



