NoreMBEB 19, 1900] 



SCIENCE 



727 



polarizing photometer, attached to the 23-inch 

 equatorial. Considerably more than half the act- 

 tings were read and recorded by an assistant. 

 Eleven minima were observed in whole or in part. 

 These minima, together with those ob.served and 

 published in detail by Graff, gave a graphical 

 'determination of new elements. The star at its 

 faintest phase was seldom more than just visible. 

 Under these conditions there was no evidence of 

 variability in the depth of minimum. The mean 

 •curve shows a nearly symmetrical primary mini- 

 mum consisting in a drop of 2. .55 magnitudes, and 

 also a secondary minimum of 0.0G5 magnitudes. 

 Each lasts about six hours. After recovering 

 Irom primary minimum, the curve keeps on rising 

 ■slowly for some time, and the beginning and end- 

 ing of secondary minimum are at a higher level 

 "than those of primary minimum. This would 

 indicate ellipticity and reflection. The average 

 ■surface intensity of one star is 18 times that of 

 the other, and the radius of the fainter lies be- 

 tween 0.98 and 1.86 times that of the brighter. 

 The radius of the orbit is from 3.5 to 5 times the 

 radius of the brighter star. 



The Problem of Three Bodies from the Standpoint 



of Spectroscopy : KuBT Laves. 



With the present accuracy in the determina- 

 tions of velocities in the line of sight the problem 

 ■of three bodies begins to assume importance in 

 this department of astronomy. The paper dealt 

 ■with that phase of the subject that is analogous 

 to the lunar problem. The perturbation by the 

 ■" sun " is broken up into three components, 

 ■P, T, 8, along the radius vector, perpendicular 

 "to it and perpendicular to the plane of the orbit, 

 respectively. Calling z' the velocity in the line 

 •of sight of the disturbed body, we have ds^/dt = 

 — P sin 8 + T cos in which 9 is the longitude 

 •from the ascending node. P and T may be ex- 

 pressed as functions of 8 and when t is also 

 ■expressed in terms of the same quantity we obtain 

 an equation of the form z' = i<'(9). A compari- 

 -son with the observed velocity curve leads to a 

 •determination of the inequalities involved. The 

 approximation has been carried as far as the 

 second power of the parameter involved. 



The Determination of the Moon's Theoretical 

 Spectroscopic Velocity: Kubt Laves. 

 It was shown that the four quantities V,, V,, V„ 

 Wt, in Campbell's notation, can be calculated by 

 ■means of tables computed on the basis of the 

 •elliptic polar coordinates of the earth and the 

 anoon. 



(1) 



V = K -e. sin 



will give both V, and Vj. V, can not exceed 0.50 

 km. per second, and Vj is always less than 0.04 km. 

 As the diurnal change in F, and F, is at most 

 0.012 km. we may use approximate values of the 

 longitudes of the sun and the moon. 



(2) 7,+ y. = y, co3£; + js:,(i-f e, cos 9,) 



sin E cos ft; 



this formula is easily proved with the aid of the 

 hodographic circle. The index 2 refers to the 

 moon. For £'2(1 -f ej cos B^) tables may be con- 

 structed with the argument $2, the true anomaly 

 of the moon. The angle E, which is nearly the 

 difference between the longitudes of the sun and 

 the moon, is computed thus: 



(3) tan pa^tan ft cosec (Xj — X) and 



sin JE = 8in ft cosec pj. 



Tables based on (3) are being computed for vari- 

 ous values of ft. They will be applicable to the 

 planets as well. As the " Nautical Almanac " is 

 planning to discontinue the computation of E the 

 tables here described should be of considerable 

 value to the astrophysicist. 



The Effect of Faulty Collimation of the Correcting 

 Lens on the Star Image: .J. S. Plaskett. 

 The field of the correcting lens used with visual 

 objectives for photographing star spectra is very 

 limited. A slight displacement from the axis dis- 

 perses the star image, causes a perceptible differ- 

 ence, transversely, in the po.'rition of the images 

 due to light of different wave-lengths. It was 

 shown that even the flexure of the telescope is 

 sufficient to produce this effect. The importance 

 of correct adjustment and of compensating for 

 flexure in the effect on exposure time and on the 

 accuracy of radial velocity measurements was 

 pointed out. 



The Width of Slit giving Maximum Accuracy: 



J. S. Plaskett. 



This paper was a continuation and conclusion 

 of one with a similar title presented at the last 

 meeting. It gave results for other instruments 

 of the relative errors of measurement of early 

 type spectra at various slit widths. It was shown 

 that more accordant and accurate values are ob- 

 tained at a width of about 0.05 mm. than at either 

 narrower or wider slits. Consequently consider- 

 able saving of exposure time over that usually 

 given is possible. The bearing of these results on 

 the proportions of the optical parts in spectro- 

 graphs was also discussed. 



