Deoembeb 16, 1910] 



SCIENCE 



885 



tematic errors in latitude of from 1" to 2".5 were 

 shown to exist in each ease. Furthermore, it was 

 demonstrated that it was impossible to represent 

 both series of observations by any great circle. 

 Since 1897 the planet has deviated farther and 

 farther to tlie south from both of the computed 

 orbits. The least deviation from the great circle, 

 which would account for all the observations, was 

 1", but a deviation of 2". 2 gave more concordant 

 results. Uranus also gave indications of motion 

 in an orbit well to the south of the great circle, 

 but the deviation was too small both theoretically 

 and practically to have much weight. 



On the Light -curve of R T Persei: R. S. Dug an. 

 The discussion of 14,000 photometric settings 

 being nearly in final form, the following conclu- 

 sions can be drawn : ( 1 ) The period is changing. 

 (2) The apparent slight asymmetry of the light- 

 curve is probably real. The greater part of this 

 asymmetry is due to a cause other than eccen- 

 tricity of orbit. (3) At the beginning and end 

 of eclipse, the simple geometrical representation 

 of the curve is a little unsatisfactory. 



The Spectra of some Close Double Stars: Annie 



J. Cannon. 



A large number of peculiar spectra are de- 

 scribed in volumes 28 and 56 of the Harvard An- 

 nals. Forty-six of these were classified as com- 

 posite, eighteen of which were found by Miss 

 Maury. These composite spectra may be divided 

 into two groups, according as the brighter spec- 

 trum is of an earlier or later class than the 

 fainter. In thirty-two cases, the spectrum re- 

 sembles that of a second or third type star, except 

 that H5 and H7 are more intense and the band K, 

 due to calcium, is fainter than normal. The two 

 wide absorption bands, K and H, at wave-lengths 

 3934 and 39GS, are marked features of classes G 

 to M in spectra photographed with the objective 

 prism. K is as wide, or wider than, H and any 

 decrease in its intensity is readily noted. In some 

 of these stars, as e Carina:, the band K is almost 

 obliterated; in others, as 12 Comae Berenices, it is 

 about half as wide as the band H. On photo- 

 graphs of sufficient exposure, ultra-violet hydro- 

 gen lines Ht), H9, Hf are seen as in spectra of 

 types B to A5. It thus appears that the hydrogen 

 lines are intensified and the calcium band is weak- 

 ened by the superposition of a spectrum, which 

 has stronger absorption of hydrogen and little or 

 no absorption of calcium. Ten of these stars 

 are known visual doubles, whose companions are 

 close enough and sufficiently bright to cause the 



observed peculiarity of the spectrum. Eight 

 are spectroscopic binaries: t Persei, 7 Persei, 

 f AurigiE, a Scorpii, 5 Sagittae, 31 Cygni, |3 Cap- 

 ricorni and a Equulei. Since a Scorpii is both a 

 visual and spectroscopic double, it is uncertain to 

 which companion the peculiar spectrum is due. 

 It seems probable that there are at least seven 

 spectroscopic binaries in which the two stars have 

 widely separated classes of spectrum. Taking the 

 known doubles from the list, there remain four- 

 teen stars whose companions are yet to be ob- 

 served. The following nine are brighter than the 

 magnitude 5.50: H. R. 1129, f Persei, e Carinae, 

 12 Coma; Berenices, H. R. 5667, H. R. 7031, 

 6 Cygni, 47 Cygni and X^ Aquarii. Perhaps the 

 most interesting case is that of 6 Cygni, which is 

 the brighter component of the well-known double 

 star p Cygni. The visual companion, whose spec- 

 trum is of class A, is photographed apart from 

 that of 6 Cygni and, therefore, could not be the 

 cause of any peculiarity due to superposition. 

 The faintness of the absorption band K in 6 Cygni 

 is well shown in the reproduction of this spectrum 

 in Plate XI. of the " Atlas of Representative 

 Spectra " by Sir William Huggins and Lady Hug- 

 gins. Five measures of radial velocity which 

 have been published by Kustner show no varia- 

 tion. However, in this case, as well as the others, 

 it seems highly probable that additional observa- 

 tions will confirm the existence of a close com- 

 panion. 



Publications of the U. S. Xaval Observatory in 



Press: W. S. Eichelbebgee. 



Volume VI. will contain the observations made 

 with the equatorial telescopes from 1893 to 1907: 

 positions of satellites, diameters of planets and 

 satellites, double stars, asteroids, comets, occulta- 

 tions, phenomena of satellites of Jupiter and 

 Saturn, and transits of Mercury. This volume 

 will also contain several appendices, as follows: 

 the mass of Titan, the orbits of Deimos, Phobos 

 and Enceladus, the solar parallax from observa- 

 tions of Eros, eighteen asteroid orbits, twelve 

 comet orbits; miscellaneous observations of the 

 transit of Mercury in 1894; and a catalogue of 

 the publications of the Xaval Obsei-vatory. Vol- 

 ume VII. will be a catalogue of the Washington 

 Zones, 1846 to 1852, embracing about 45,000 ob- 

 servations on 23,518 stars. About 38,000 of these 

 were reduced by nights and published thirty years 

 ago; the remaining 7.000 now appear for the first 

 time and a systematic search has been made for 

 errors in the earlier reductions. 



