260
ASTRONOMY: W. J. L UYTEN
PROC. N. A. S.
parts of the Milky Way could hardly explain the shortage of high velocity...
13 downloads
345 Views
458KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
260
ASTRONOMY: W. J. L UYTEN
PROC. N. A. S.
parts of the Milky Way could hardly explain the shortage of high velocity-apices in those regions: in order to diminish the speed of a passing star perceptibly the mass of this ring would have to be of the order of 1020 stars or 1010 times the mass which Kapteyn attributes to the whole stellar system. According to the theory of relativity (De Sitter, Mon. Not. R. Astron. Soc. 77, p. 176) a ring of this mass cannot exist at a distance of the order of that of globular clusters, as the resulting shift toward the violet in the spectra of globular clusters and spiral nebulae would be much greater than the observations permit. 1 Pop. Astron., 26, 1918 (8). 2 This was first observed by B. Boss, Ann. Rep. Director Dep. Merid. Astr., Carnegie Inst. of Washington, 1918. 3 Oort, Bull. Astr. Inst. Netherlands, 1, 1922 (134). 4These PROCEEDINGS, 9, 1923 (315). 5 Shapley, Mt. Wilson Contr., 8, 1918 (134-139, 218-222, especially 314-319). 6 Mt. Wilson Contr., 8, 1918 (314), H. C. 0. Bull., 776, 1922. 7 Charlier, Lund Meddel., 2nd Series, No. 19, 1917 (Plate VII). 8 Mt. Wilson Contr., 9, 1920 (423). 9 Mon. Not. R. A. S., 83, 1923 (147-152). 10 Schilt, Bull. Astr. Inst. Netherlands, 2, 1924 (49). 1 Shapley, These PROCEZDINGs, 5, 1919 (344-351).
NOTE ON SOME STATISTICAL CONSEQUENCES OF THE LUMINOSITY LAW By WILLEM J. LuYvrzN HARvARD COLLEGE OBSZRVATORY
Communicated, May 5, 1924
The luminosity curve derived by Kapteyn and'van Rhijn is based on considerations of parallactic motions, and, as far as the part near and beyond the maximum is concerned, on the available material of numbers and parallaxes of stars with large proper motion. Although the rapid accumulation of the last two kinds of observational data will make a revision of the luminosity curve possible in the near future, it may be of interest to compare the statistical behavior of certain groups of selected stars with that to be expected from the luminosity law and some auxiliary assumptions, principally the velocity law. For some parts of the sky our meridian and astrographic catalogues give complete lists of stars, brighter than a given apparent magnitude and with proper motions exceeding a given value. It is then an easy matter to derive the frequency function of H = m + 5 + 5 log ,u for these stars. A simple theoretical' expression for this frequency curve F(H) may be found in the following way. The number of stars whose distances lie between r and r + dr, whose
VoxL. 10, 1924
ASTRONOMY: W. J. LUYTEN
261
absolute magnitudes are between M and M + dM, and whose linear velocities perpendicular to the line of sight lie between T and T + d T is given in the usual notation by: (1) dN - 47rr2 A(r)dr.A(M)dM.4(T)dT To find the total number of stars in the sky brighter than apparent mkgnitude m, and with proper motions larger than ,u, we have to integrate the above expression between appropriate limits. In o-ur applications we shall be dealing with stars of comparatively large proper motions, i.e. relatively near-by stars, and accordingly we may, in first approximation, assume that the density is constant throughout our space of integration. For cb(M) we shall adopt Kapteyn's form, and for 4(T) Schwarzschild's, i.e. a normal error curve in log T with the arithmetic mean value of log T as a linear function of M. Substituting u = 5 1 r, V = 5 log T, g = 3/(5 mod,) anrd
Vo =
a
+ #M
(2)
equation (1) reduces to dN = C.exp [gu-h2(M-Mo)2-t2(V-Vo)2]. dudMdV (3) in which C, h, Mo, t, a, and ,3 are constants of the distribution functions. N is then found by integrating (3) between the limits: M from- to+ 00, ufrom-co toB-M, VfromA +uto+0c, where A = 5 + 5 log I,u, and B = m,. However we want, not N, but F(H) for these stars. Noting that H = m + 5 + 5 log ,u = M + V, we effect the following transformation of variables: M = (H + K)/2, V = (H-K)/2, u = u. The restrictions resulting from the conditions that m]m, and ju > p, now reduce to u < B-(H + K)/2 A + u < (H-K)/2, (4) giving a double restriction for u, and none for K. Written otherwise we have: and u
F(H)dH = dHft+ dKf(-a) f(H, K, u)du for H < Hc F(H)dH = dHf+ dKf , f(H, K, u)du for H > Hc. f(H, K, u) = C.exp [gu- h2 (H+K-2M0)2/4-t2(H-K-2a-jPH- - #K)2/4] The limits (a) and (b) in the second one of both double integrals denote -
262
ASTRONOMY: W. J. LUYTEN
PROC. N. A. S.
that u has to be integrated between - c and one of the limits indicated by equations (a) and (b). The integrals are elementary, and it is easily seen that the solution may be written as follows:
H < H. H > Hc
P(H)dH = C,.j/^1r.ei2(HHa)2dH Q(H)dH = Cm.il/_.( h2t2 h2 + t2(1 + #)2
Ha
=
(5)
Ho + g/2t2 + g((1 + ()12h2 Hb = Ho-g(1 + 3)/1h2
if Ho is the value of H derived from (2), and corresponding to the maximum MO of the luminosity curve. The result is remarkable in that P and Q are quite general, and except for C.. and Cm which only convert the relative frequencies into absolute values, are independent of either A,u or mc. Equation (5) gives the solution for all values of ,uc and mc, without need of computation of C,, and Cm as explicit functions of these quantities. To obtain the result in any specific case we follow curve P up to H = Hc, and curve Q from thereon, multiplying the Q-frequencies by such a factor that for H = Hc the P- and Qfrequencies are identical. Graphically the solution is given by keeping P constant, and raising or lowering Q (by multiplying its ordinates) until it intersects P at the desired point H = Hc. It is also significant that as long as we require only relative frequencies, the solution is the same for all values of ,ic and mc which satisfy the relation mc + 5 log ,Ac = constant. It follows that the same relative curve must represent the frequency distribution of H for those stars whose apparent magnitudes and proper motions lie between such limits ml, m2 and p1, j2 that ml + 5 log I, = m2 + 5 log /2 = Hc-5. That the H-frequency curve must be "beaky" is immediately evident from the consideration that the stars used are brighter than a certain apparent magnitude and have a proper motion larger than a given value. Because the numbers of stars of all proper motions increase rapidly with decrease in apparent brightness, and likewise the numbers of stars of all magnitudes increase rapidly with decreasing proper motion the peak must appear, no matter what frequency laws are used. Before comparing the theoretical results with the observations we must choose numerical values of our constants. For h and MO we shall adopt Kapteyn's values,' 0.28 and 7.7, respectively; for t a value of 0.56 seems plausible; (3we may take2 as 0.28 and Ho as 12.5. We then have i = 0.205, Ha = 17.9, and Hb = 1.4. The resulting curve for F(H) will therefore be symmetrical only when Hc = H, = 9.65 = (Ha + Hb)/2; for
263
ASTRONOMY: W. J. LUYTEN
Voi,. 10, 1924
smaller values of Hc the ascending branch will be steeper; for larger values of HC the descending branch of the curve will be the steeper. In the Greenwich 1910 Catalogue (p. B XXXII) a list is given of all stars in the zone having proper motions larger than O".200 per annum. This list is probably complete to about magnitude 9.2; HC is then 10.7. The theoretical curve is shown in figure 1 as a full line, the dots represent the observed points from 224 stars. In figure 2 the comparison between 0
FIG.1
/\FIG
S
10
6
11
16
8
13
AFiG.4A
FiG.-3 3
1S
2/\
1
13
3
Comparison between theoretical and observed frequlency curvPes F (H) for stars brighter than a given apparent magnitude and with proper motion exceedling a given value.
theory and observation for the case HC = 11.25 is given (proper motions from the Greenwich Polar Zone, m, is about 9.5 to 10.0 andp,u = 0".200; 255 stars). Figures 3 a4d 4 show the case for H, = 8.0; figure 3 representing the brighter stars, with m, = 5.5 and iuc = O0".316 (151 Boss stars), and figure 4 representing the fainter stars with m, = 7.0, I,u = 0".158 (209 stars from various zone catalogues). The -only adjustment that has been made is the reduction of all curves, theoretical as well as observational, to the same total number of stars. The agreement in figure 1 is as good as may be expected; the other three cases may still be deemed satisfactory, considering the fact that the "beakry" form of the curves makes them extremely sensitive to small changes in'H, and the constants used in the distribution curves. On the other hand, the theoretical curves, exhibited here still depend
264
ASTRONOMY: W. C. RUFUS
PROC. N. A. S.
very largely on the absolutely brighter stars. A real test of the luminosity curve and the auxiliary laws can only be obtained if we use values of HC at least as large as 14. Thus the Wolf proper-motion stars might be used to advantage were it not for the. fact that the value of the systematic correction to be applied to Wolf's magnitudes is rather uncertain. There are indications, however, that with any reasonable value for this correction, the agreement between theoretical and observational curves will remain bad in this case, there being an observed excess of stars with large values of H. 1 Mt. Wilson Contr., No. 188. 2 Ann. Harvard Coil. Obs., 85, No. 5, p. 103.
ATMOSPHERIC PULSATION OF THE CEPHEID VARIABLE, AQUILAE' By W. CARL RuFus OBSZRVATORY OF THU UNIVZRSITY OF MICHIGAN
Communicated, April 22, 1924
A new method of analyzing variable radial velocity data by a systematic study of line displacements from different elements and at different levels of a star's atmosphere has been developed and applied. The usual method of selecting lines giving consistent displacements has masked an effect here tentatively called atmospheric pulsation. A difference in the velocities from the hydrogen and the helium lines in the star,2 B.D. + 560 2617, led to the idea of isolating the velocities of high-level elements. Application was made in the case of x Aurigae by isolating the hydrogen lines and a systematic distribution of the hydrogen velocity residuals with respect to phase was found.3 Based upon the isolation of velocities at three levels in the star's atmosphere the method4 adapted to the problem of Cepheid variation was first applied by Mr. J. A. Aldrich in a study of S. Sagittae.5 The method is also applicable to long-period and irregular variables, to secular and periodic changes in peculiar spectra, and to other astrophysical problems. The Method as Applied in -q Aquilae.-Velocity-difference curves correlated with the light variation and other periodic changes characterize the method here applied. The radial velocities of concentric layers of the star's atmosphere are determined separately by isolating certain elements, and by grouping lines of the spectrum originating at assumed levels based upon the determination of their heights in the sun by St. John6 and by Mitchell.7 The approach of two layers indicates compression of