882 



SCIENCE 



[N. S. Vol. XXXII. No. 833 



focal length, yielding much higher linear disper- 

 sion than is usually employed in stellar work. 

 The spectrograph was placed in a constant tem- 

 perature-room and was fed by the 60-inch reflector 

 of the Mount Wilson Observatory in the coud4 

 form. These spectrograms were studied with a 

 view to obtaining some knowledge of the pressures 

 in the atmospheres of these stars. For this pur- 

 pose the relative shifts with respect to an iron 

 arc comparison spectrum were measured for the 

 enhanced lines and also for the are lines. The 

 basis for this distinction is Mr. Adams's earlier 

 work on the displacements, in all probability due 

 to pressure, of lines at the limb of the sun; in 

 this work he had found that the large displace- 

 ments were as a rule associated with the enhanced 

 lines, and small displacements with arc lines. For 

 the three stars studied Mr. Adams finds: 

 Sirius, enhanced lines minus arc lines + 0.014 A 

 Procyon, " " " " " +0.009 



Arcturus, " " " " " +0.001 



Furthermore in Arcturus, whose spectrum 

 closely resembles that of a sun-spot, it was found 

 that only the iron lines show a shift in the posi- 

 tive direction, that is 0.006 A toward the red; 

 while the lines of nickel, titanium, vanadium, 

 magnesium, calcium and hydrogen were in this 

 order shifted toward the blue, hydrogen most of 

 all. The principal inferences that the author 

 draws from these results are that the cause of 

 these systematic shifts in stellar spectra is the 

 same as that which is effective at the limb of the 

 sun, and that accordingly in all probability they 

 are due to pressure. On this basis it becomes pos- 

 sible to compute the pressures in the atmospheres 

 of these stars as compared with that of our sun. 

 For Sirius this comes out twelve atmospheres 

 (terrestrial) greater than that of the sun, and for 

 Procyon seven atmospheres. For Arcturus an 

 arrangement of the different gases similar to that 

 in the sun is indicated; hydrogen, calcium and 

 magnesium being at the level of low pressures 

 and iron in the region of high pressures. 



Note on the Spectrum of D. M. +30° 3639: 

 Waiter S. Adams. (Read by Professor Hus- 

 sey.) 



In 1892 Campbell found this star, whose mag- 

 nitude is 9.3, to be surrounded by a hydrogen 

 atmosphere 5" in diameter. With a one-prism 

 spectrograph (attached to the 60-inch reflector) 

 and an exposure of 150 minutes, Mr. Adams suc- 

 ceeded in securing a satisfactory spectrogram of 

 this faint star. The hydrogen lines from H/3 to 



Hf are visible and extend beyond the continuous 

 spectrum by an amount that precisely corresponds 

 to Campbell's estimate of 5". The bright line at 

 X 4068 also extends outward about 4"; the origin 

 of this line is unknown. This is also true of the 

 extremely bright band at X 4652, which does not 

 extend beyond the continuous spectrum. A less 

 satisfactory spectrogram liad been obtained on an 

 earlier evening; and a comparison of the two indi- 

 cates that the velocity of the star is probably 

 variable. 



Note on D^ in the Spectrum of Prominences: 

 Jeknie B. Lasby. (Read by Miss Whiting.) 

 No photographic determination of the wave- 

 lengths for the two components of D^ in the sun 

 appears to have been published. Four plates were 

 secured by Mr. Adams on August 28 and Decem- 

 ber 15, 1908, when large prominences were visible; 

 these were measured by Miss Lasby and also by 

 Mr. Adams, and the wave-lengths were determined 

 as 5875.841 and 5876.190, which are in good 

 agreement with Mohler and Jewell's visual ob- 

 servations of the sun and also with their labora- 

 tory measures. Miss Lasby advances reasons 

 against the supposition that D^ is present as a 

 dark line in the ordinary solar spectrum. 



On the Determination of the Elements of Algol 

 Variables: Henry Norris Russell. 

 In the first approximation, an Algol variable 

 may be assumed to consist of two spherical stars, 

 each of uniform surface brightness, revolving in 

 a circular orbit. If the eclipse is total, that is, 

 if there is a constant period at minimum, the 

 actual brightness of each of the stars is at once 

 known. Three unknowns have then to be found: 

 the radius B of the eclipsing star, in terms of 

 that of the orbit; the ratio k of the radius of the 

 eclipsed star to that of the other; and the in- 

 clination i of the orbit. Let 6 be the orbital 

 longitude of the eclipsing body, measured from 

 the point of conjunction. Its value at the instant 

 when any given percentage of the area of the 

 eclipsed star is obscured may be found from the 

 liglit-curve. If p is the apparent distance of the 

 centers at this moment, we have from geometrical 

 considerations p^= sin^ i sirf 6 + cos^ i; and from 

 the known eclipsed area p = Rf{K) where f{K) is 

 a transcendental function, which can be computed 

 for any given value of k. Hence R-!f(K)Y=. 

 sin" i sin- + cos" i. From three such equations, 

 corresponding to different percentages of obscura- 

 tion, R and i may be eliminated, giving an equa- 

 tion of the form 



