46o 



NATURE 



[March i6, 1893 



the increase in the eccentricity of orbit would also have 

 been largely reduced. 



Notwithstanding this criticism, it appears to me that 

 Mr. See fairly establishes the proposition that a high 

 eccentricity is explicable by means of tidal friction. 



Turning, then, to the question of the relative masses of 

 the components of double star systems, Mr. See remarks 

 with justice that the comparable brightness of the com- 

 ponents renders it highly probable that the masses are 

 also comparable, and he sees in certain results of M. 

 Poincard and of my own an evolutionary explanation of 

 this fact. 



Jacobi first showed than an ellipsoid of homogeneous 

 fluid, with its three axes bearing to one another proper 

 proportions, is a figure of equilibrium when it rotates 

 about its smallest axis with a proper angular velocity. 

 M. Poincare next showed that if the length of the Jacobian 

 ellipsoid exceeds the breadth in a certain ratio, the equili- 

 briurti becomes unstable, but that there is a stable 

 figure which may be described as a Jacobian ellipsoid 

 with a furrow nearly round the middle, so that it resembles 

 an hour-glass with unequal bulbs. If we trace the further 

 development of the hour-glass we find its neck gradually 

 thinning, and finally rupturing the figure of equilibrium, 

 henceforth consists of two detached masses. 



My own attack on this problem was from the opposite 

 point of view, for I endeavoured to trace the coalescence 

 of a pair of detached masses so as to form an hour-glass 

 or dumb-bell. 



Mr. See reproduces the figures illustrative of both these 

 investigations, and remarks that they both show that when 

 there is a gradual detachment from a rotating figure of 

 equilibrium, the detached portion will not norma lly be 

 a ring: but that there will ensue two quasi-spheroidal 

 masses of matter of comparable magnitude. He also 

 remarks that if the fluid be heterogeneous, the ratio of the 

 masses will be much smaller than wHen it is homo- 

 geneous. 



In the discussion of these figures of equilibrium the 

 wording of the essay appears a little careless, for it might 

 naturally be supposed to mean that increase of angular 

 velocity is a necessary concomitant of the rupture of the 

 neck of the hour-glass. Now it is a somewhat paradoxical 

 fact that, with constant density, the longer elongated 

 figures of equilibrium rotate more slowly than the shorter 

 ones, and it might therefore seem that the rupture of the 

 neck should go with retardation of angular velocity. But 

 it is the value of the square of the angular velocity divided 

 by the density which determines the length of the 

 elongated figures, and thus increase of density tells in 

 the same way as retardation of angular velocity. In the 

 history of a nebula the only condition for rupture which 

 can be specified is that of contraction. 



The probability of this view of the genesis of double 

 stars is strikingly illustrated by a number of drawings by 

 Sir John Herschel of various nebute. The great similarity 

 between Herschel's nebulae and the theoretical hour-glass 

 is obvious. It may be hoped that in the book which Mr. 

 See promises he will also illustrate this point by photo- 

 graphs. 



Annulation is usually accepted as the mode of separa- 

 tion in the nebular hypothesis, but, as already stated, this 

 is held by Mr. See to be exceptional. He thus regards 

 NO. T2 20, VOL. 4f] 



the ring of Saturn as being as exceptional in its history as 

 it now is in appearance. Where he maintains that Saturn's 

 ring will never coalesce into a satellite, he might with 

 advantage have referred to the remarkable investigations 

 of M. Roche,^ who showed that a satellite would be torn 

 to pieces by tidal action if it revolved at a distance of 

 less than 2*44 times the planet's radius. We may here 

 note the interesting fact that whilst Saturn's ring almost 

 touches " Roche's limit" on the inside, the Martian satel- 

 lite, Phobus, and the fifth satellite of Jupiter^ almost touch 

 it on the outside.^ 



In order to prove his thesis as to the highness of the 

 eccentricity and the comparability of [masses, Mr. See 

 gives a careful table of the observed elements of the orbits 

 and of the relative brightnesses of seventy-three pairs of 

 double stars. The values of the elements are of course 

 open to much uncertainty, but the mean eccentricity, 

 which is found to be '45, must lie near the truth. In the 

 few cases in which the masses have been determined,, 

 they are found to be comparable, and the comparability 

 of the brightnesses confirms the generality of this law. 

 Thus the facts of observation agree with our author's 

 ideas. 



Mr. See must be congratulated on having written an 

 essay of great cosmogonical interest, and although his 

 theory may never be susceptible of exact proof, yet there 

 is sufficient probability of his correctness to inspire us 

 with fresh interest in the observations of double stars. 



G. H, Darwin. 



MAGNETIC INDUCTION IN IRON AND 

 OTHER METALS. 



Magnetic Induction in Iron and other Metals. By J. A. 

 Ewing, F.R.S. J (London : Electrician Office.) 



IN this admirable book Prof. Ewing has brought 

 together matter which was before to be found only 

 in the journals of learned societies, and he has also 

 given a full account of his own researches in magnetism. 

 The book is written in a lucid style, and is supplied with 

 numerous references to original papers. 



In Chapter I. Prof. Ewing explains clearly the mean- 

 ing of such terms as "intensity of magnetisation" and 

 the like, which many students have difficulty in under- 

 standing. As stated in the preface, he has "endeavoured 

 to familiarise the student with the notion of intensity of 

 magnetisation (I) as well as with the notion of magnetic 

 induction (B)." When endless magnetic circuits are dis- 

 cussed, it is convenient to talk of " permeability " and " in- 

 duction " ; on the other hand, " magnetic poles " and 

 " magnetisation " are just as important when permanent 

 magnets are dealt with. The magnetisation of ellipsoids 

 and the influence of the shape and dimensions of 

 magnetised bodies upon magnetic quality are fully 

 treated. 



1 "Acad, des Sciences de Montpelier," vol. i. (1847-50), p. 243. See also 

 Darwin, Harper s Magazine^ June, 1889. 



2 The values given by Barnard (Nature, p. 377) make the distance 

 112,000 miles, and Roche's limit 107,000 miles. 



" It is proper to warn the reader that Roche's limit depends to some extent 

 on the density of the planet. For the sun it will be about one-tenth of the 

 earth's distance from the sun. Thus a body of planetary size cannot move 

 in a highly eccentric orbit, so that its perihelion distance is one-tenth, with- 

 out being broken up into meteorites ; and conversely a flight of meteorite;:- 

 with less than the same perihilion distance can never coalesce into a 

 planet. 



