June, igo6.j 



KNOWLEDGE & SCIENTIEIC NEWS. 



451 



same density and surface luminosity as Sirius — about 

 2,700 " lis^ht years." Let us assume with Nevvcomb 

 that the outer boundary ol the Milky Way is at about 

 3,000 light years, and see what average distance this 

 will give between each pair of stars, on the supposition 

 of an equal distribution of stars in a globular space. 

 We know, of course, that the visible stars are not 

 cqually distributed , but the computation will give the 

 average distance between any two adjacent stars. As- 

 suming a total of 100 millions, and that each star is 

 placed at the anglt of a tetrahedron,* I find that the 

 average distance between two stars would be alsout 

 21.24 I'.i^^ht years. This corresponds to a parallax of 

 o". 153. Now I find that the average parallax of 20 

 stars, for which a fairly reliable parallax has been found 

 is o''.247. If we exclude those stars with a parallax of 

 over o''.3 — which may, perhaps, be considered as ex- 

 ceptionally close to our system — we have 13 stars with 

 an average parallax of o". 155. 



From a consideration of the proper motions of two 

 groups of stars, one of 206 stars of mean magnitude 

 5.7, and mean proper motion of c".3i per annum, and 

 another of igg stars of mean magnitude 8.1, and mean 

 proper motion of o''.304, Mr. J. G. Porter finds " the 

 average parallactic motion " of all the stars to be 

 o''. 185; that is, the annual apparent motion due to the 

 .<;un's motion in space. Taking this annual motion as 

 four radii of the earth's orbit, that is, four times the 

 sun's distance from, the earth — a quantity probably 

 near the truth — the mean paridlax of the stars con- 

 sidered would be o".o46, or about 70 years' journey for 

 light. But judging from their magnitude these stars 

 would not be among our nearest neighbours in space. 



With a distance of 21.25 "ght years between two 

 stars at a distance of 3,000 light years from the earth, I 

 find that the apparent distance between such stars in 

 the Milky Way would be about 24 minutes of arc, and 

 as the faint stars in the Galaxy are, on an average, 

 much closer than this, it seems highly probable that the 

 stars composing the Milky Way are much nearer to 

 each other than a distance of 21 light years, and this 

 the crowded appear;mcc of the Galaxy would lead us to 

 suppose. 



If wc take the average width of the Milky Way as 

 20", I find that tlie volume of space contained by lines 

 drawn from ihi' eye to the edges of the Galaxy is about 

 o 1767 of the volume of the whole sphere. Hence, 

 with an equal distribution of stars, the Milky Way 

 should contain about 17A- millions of stars. Now Dr. 

 Roberts' photograph in Cygnus shows about 8,500 stars 

 to the square degree. This would give a total of about 

 61 millions for the whole of the Milky Way. Rut as 

 this is rather a rich part of the Galaxy the total may 

 not exceed 40 or 50 millions; that is, two or three times 

 that due to an equal distribution of stars. This agrees 

 with some experiments made by Mr. (iavin j. I?urns, 

 who finds that the average luminositv of the Milky 

 Way is from two to three times greater than that of 

 the rest of the sky. | 



Assuming that the Milky Way contains a total of 50 

 millions of stars, and that its limits lie between 1,500 

 and 3,000 " light years," I find that its volume would 

 be about 0.1546 of the sphere having a radius of 3,000 

 light \ears. This would give an average distance be- 

 tween two adjacent stars of 14.31) I'.glit years. On this 

 hypothesis tlic Milky W.iy would have a considerably 



* If s he tlic side (if a telrtitlicdron, its volume is S3 



— V 2. 

 12 ^ 



f .lsti:^/'li\siuil Jotii-ii.tl, (October, igo2. 



greater extension in the line of sight than at right 

 angles to that line. If we suppose that its thickness 

 in the line of sight does not on the average exceed its 

 apparent thickness; that is, that its cross section is 

 roughly circular, and assuming that its mean distance 

 is about 2,300 light years, I find that its diameter would 

 be about 800 light years, its nearest and its farthest 

 parts being at about 1,900 and 2,700 light years re- 

 spectively. In this case the average distance between 

 the component stars w'ould be about 10.7 light years. 



In the globular cluster u Centauri, an enumeration 

 made from photographs by Professor and Mrs. Bailly 

 gives a total of 6,389 stars on an area of about 30 

 minutes square. T his gives 25,556 stars to the square 

 degree, and if the distance of u Centauri is at all com- 

 parable with that of the Milky Way its component stars 

 must be much closer than in any part of the Galaxy. 

 The same may be said of the smaller Magellanic Cloud, 

 which has about 28,000 stars to the square degree. 



E!aston thinks that the stellar universe is of " a fairly 

 thick lens shape filled with stars which are much more 

 closely congregated near the edges than near the centre 

 of the lens." Professor Newcomb's views are some- 

 what similar. But this is returning to Sir W'illiam 

 Herschel's " disc theory," and it seems doubtful 

 whether such a conclusion is warranted by the evidence. 

 As is well known, this disc theory was abandoned by 

 Herschel himself in his later writings. I have never 

 seen any answer to the argument against the disc theory 

 advanced by me in " The \'isible Universe " (no. 241, 

 242). The argument is as follows : — As the thickness 

 of Herschel's supposed disc extends on both sides of 

 the earth beyond the theoretical distance of stars of 

 the gth magnitude, the stars of this magnitude should 

 be as numerous in the direction of the Galactic poles as 

 in the direction of the Milky Way itself. But this is 

 not the case. Argelander's maps show that gth magni- 

 tude stars are more numerous in the Milky Way than 

 at the Northern Galactic pole in the ratio of 2^ to i. 

 Kven the stars visible to the naked eye show a marked 

 tendencv to aggregation on the Milkv Way, and liaston 

 finds that the taintest siars of Argelanders' catalogue — 

 about gi magnitude — " present, by the manner in which 

 they are distributed, a remarkable correspondence w'ith 

 the luminous and obscure spots of the Milky Wa\\" 

 These facts seem to be inconsistent with the disc theorj' 

 as originally propounded by Herschel. From the evi- 

 dence quoted above it seems evident that if the stellar 

 system is in any way shaped like " a thick lens " there 

 must he a considerable crowding of stars along the 

 edge of the disc; that is, in the direction of the Milky 

 Way. 



Stru\e's modification of the " disc theory," namely, 

 a disc of a certain thickness, but of infinite, or, at least, 

 indeterminate, diameter, seems an improbable hypo- 

 thesis, and one not in agreement with observed facts. 

 I-lven on this theory our sidereal system is supposed to 

 be limited in the direction of the (Galactic poles. If 

 .Struve's theory were accepted we should also be obliged 

 to accept his hypothesis of the extinction of light in 

 the ether; for otherwise the Milky Way would be much 

 brighter than it is. With an infinite extension, the 

 Galaxy would shine with the brightness of the sun. 

 The comparatively feeble gleam of the Milky Way on 

 even the clearest nights should, I think, be sullicient to 

 convince the thoughtful observer that its light is rw/ due 

 to a vastly extended stratum of stars. Fven Sir John 

 Herschel's reflecting telescope of i8i inches aperture 

 (now far surpassed in space-penetrating power ijy 

 modern instruments) sufficed in some places to show 

 the eoniponent stars of the Milky Way on a /'/<h'k back- 



