January 19, 1922] 



NATU'RE 



89 



the chemical properties of the radio-elements and the 

 ring of this knowledge upon inorganic chemistry 

 : theories of the structure of the atom. Among 

 ine topics dealt with are the relation of the radio- 

 elements to the periodic system of classification, the 

 properties of isotopes, the separation and purification 

 of individual elements, and the analytical chemistry 

 of uranium, thorium, and radium. 



The spring announcements of the Cambridge Uni- 

 versity Press contain several items of scientific in- 

 terest, among which is the first volume, bearing the 

 sub-title Foundations, of a forthcoming book by 

 Prof. H. F. Baker entitled "Principles of Geometry." 

 We learn from the preface that the work seeks to 

 introduce the reader to those parts of geometry which 

 precede the theory of higher plane curves and of 



irrational surfaces. Vol. i is devoted to the 

 indispensable logical preliminaries. It assumes 

 only those relations of position for points, 

 lines, and planes which, furnished with a 

 pencil, a ruler, some rods, and some string, a 

 student may learn by drawing diagrams and making 

 models. It seeks to set these relations in an ordered 

 framework of deduction, gradually rendered com- 

 prehensive and precise enough to include all the later 

 theory ; to this end it puts aside, at first, most of 

 those intricate details which make up the burden of 

 what is generally called elementary geometry. Later 

 volumes will deal, on the basis of the results obtained 

 in this volume, with conies (and circles), with quadric 

 surfaces and cubic curves in space, and with cubic 

 surfaces and certain quartic surfaces. 



Our Astronomical Column. 



The Origin of Binary Stars. — Dr. J. H. Jeans 

 discusses this question in the January issue of Scientia. 

 He notes that binaries are of such frequent occurrence 

 (practically half the stars) that we cannot regard 

 them as freaks or abnormalities, but must seek for 

 some explanation of very wide applicability. He con- 

 siders three possible origins : (i) through fission of 

 a single mass ; (2) formation of adjacent nuclei in the 

 original nebula, sufficiently close to each other to be 

 held together gravitationally ; and (3) capture, arising 

 from the appulse of two stars originally independent. 

 The last could lead to capture only if a resisting 

 medium were present ; moreover, there would be far 

 too few close appulses to explain any appreciable fraction 

 of the existing binaries. Dr. Jeans estimates that in 

 a universe of a thousand million stars there would 

 be ten thousand captures in a thousand million years. 

 The first suggestion is shown to be possible only 

 when a certain density of the rotating star has been 

 attained (probably about that of the stars of B type). 

 It is shown that this explanation accords well with 

 the observed phenomena in the case of spectroscopic 

 binaries, notably the low eccentricity of their orbits. 

 Russell and others, however, have shown that the 

 latus rectum of the orbit cannot increase very greatly, 

 save under the action of considerable external forces, 

 which are certainly not present now, and could only 

 have been present in the past if the interstellar dis- 

 stances were then much smaller. 



A test of the fission theory is afforded by triple 

 systems, which generally consist of a close pair with 

 a distant companion. Prof. Russell showed that the 

 density of the central star would be at least 380 times 

 i^ieater at the second fission than at the first, which 

 1' ads to such an improbable figure as to throw very 

 lirave doubts on the fission theory in the case of the 

 wide pairs. Hence the second suggestion is taken 

 to be by far the most probable explanation of the 

 latter pairs. It must, however, be considered to indi- 

 cate the how rather than the why ; for the question 

 remains why there should be so marked a tendency 

 for nebular condensations to occur in pairs. 



Dr. Jeans notes that star groups with common 



motion, such as the Taurus and Ursa Major clusters, can 



' most readily explained as arising from adjacent con- 



isations in a primitive nebula; in these cases, how- 



I r, the mutual distances were so great that the stars 



were outside each other's field of gravitational control. 



NO. 2725, VOL. 109] 



The Orbit of Castor. — Dr. W. Dobcrck gives in 

 the centenary number of Astron. Nachr. an explana- 

 tion of Villarceau's method of computing double-star 

 orbits, which is analogous to Laplace's method for 

 planetary orbits. He illustrates it by revising the 

 orbit of Castor from the following four positions : 

 1719-84, 357-0° (482"); 18320, 259-0°, 4-61"; 18800, 

 2345°, 5-63"; 1920-0, 2160°, 503". The first distance 

 was not observed, but calculated. The author utilises 

 his earlier orbit to shorten the approximations, and 

 obtains the following orbit : ft 222° 7', A 67° 19', 

 7 116° 6', e 0-2875,? 477-5 years, T 1960-51, o 6-573*. 

 Predicted places, 1930-0, 210-4°, 4-58"; 1940-0, 

 2030°, 3-98". Owing to the approach to periastron 

 the motion is accelerating It should be possible to 

 obtain the relative masses of the components before 

 very long; this is desirable as a check on the result 

 suggested by the spectroscopic observations, which 

 give the faint star six times the mass of the bright 

 one. 



Spectrum of a Cygni. — This spectrum is interest- 

 ing from its relationship to the spectra of novae. Its 

 classification is A 2 (peculiar), and Dr. W. H. Wright 

 has made a special study of the ultra-violet region, 

 which is described in Lick Obs. Bull, No. 332. A 

 spectrograph with two quartz prisms was attached 

 to the Crossley reflector, and three photographs taken 

 on June 11, 192 1, two of them being on films which 

 were bent to correspond with the curvature of the 

 field. The limiting wave-lengths are 3245 to 4102, 

 and 184 lines are recorded in the table, most of them 

 being identified with known metallic lines, but they 

 are unusually sharp and narrow compared with other 

 A spectra. The hvdrogen Balmer series is com- 

 plete from HS to Ho.. Dr. Wright states that the 

 resemblance to the spectra of novae is still more 

 striking in the ultra-violet than in the visual region. 

 A curve is given of the spectral intensity on the 

 photographs: it falls verv steeply between 375^ and 

 3650, then slowly and uniformly to 3245. The paper 

 also contains some measurements of the red end of 

 the a Cygni spectrum taken on stained plates. It 

 is incidentallv proved that some lines announced in 

 this region bv Dr. Waterman from photographs 

 with a grating spectrograph reallv belonged to 

 I the blue region of the overlapping third-order 

 spectrum. 



