3 66 



NATURE 



{August 1 6, 1888 



eventual collapse of the system would seem to be involved 

 in the first supposition, while the second implies the 

 progressive execution of majestic and profound designs. 



After the lapse of some centuries, photographic mea- 

 surements will perhaps help towards a decision as to 

 whether separatist or aggregationist tendencies prevail in 

 clusters. Allowance will, however, have to be made, in 

 estimating their results, for the possible movements of 

 recession or approach of the entire group relatively to the 

 solar system, by which perspective effects of closing up or 

 of opening out would respectively be produced. 



Inequalities of brightness, to the extent of three or four 

 magnitudes, are usually perceptible among the lustrous 

 particles constituting these assemblages. Nor are their 

 gradations devoid of regularity and significance. Gener- 

 ally, if not invariably, the smaller stars are gathered 

 together in the middle, while the bright ones surround and 

 overlay them on every side. Thus, the central portion of 

 the magnificent Sagittarius cluster (M 22) accumulates the 

 light of multitudes of excessively minute stars, and is 

 freely sprinkled over with larger stars. The effect, which 

 probably corresponds with the actual fact, is as if a globe 

 of fifteenth magnitude were inclosed in a shell of 

 eleventh magnitude stars, some of these being naturally 

 projected upon the central aggregation. Sir John Herschel 

 remarked of a cluster in the southern constellation of the 

 Altar (" Gen. Cat." 4467) : " The stars are of two magnitudes ; 

 the larger run out in lines like crooked radii, the smaller 

 are massed together in and around the middle " (" Cape 

 Observations," p. 119). A similar structure was noted 

 by Webb in clusters in Canes Venatici (M 3), in Libra 

 (M 5), and in Coma Berenices (M 53) (" The Student," vol. 

 i. p. 460). Here, again, we seem to catch a glimpse, from 

 a different point of view, of a law connecting growth in 

 mass and light with increase of tangential velocity — con- 

 sequently, with retreat from the centre of attraction ; and 

 the assumption of an outward drift of completed suns 

 gains some degree of plausibility. 



Irregularities of distribution in clusters take a form, in 

 some instances, so enigmatical as to excite mere unspecu- 

 lative wonder. At Parsonstown, in 1850, three "dark 

 lanes," meeting at a point considerably removed from the 

 centre, were perceived to interrupt the brilliancy of the 

 stellar assemblage in Hercules. They were afterwards 

 recognized by Buffham and Webb, and are unmistakable 

 in one (at least) of Mr. Roberts's recent photographs of 

 that grand object. The clusters in Ophiuchus, in Canes 

 Venatici, and in Pegasus (" G. C." 4670) are similarly tun- 

 nelled. Preconceived ideas as to the mechanism of 

 celestial systems are utterly confounded by appearances 

 not easily reconcilable, so far as we can see, with the 

 prosecution of any orderly scheme of circulatory move- 

 ment. Even if absolutely vacant, the extensive clearings 

 indicated by the phenomenon of dusky rifts, must of 

 course, in globular clusters, be partially obliterated by the 

 interposed light of the surrounding star-layers. They can 

 hence become perceptible only when their development is 

 most fully pronounced ; and, in a less marked shape, may 

 exist in many clusters in which they defy detection. 



The apparent diameter of the cluster in Hercules, in- 

 cluding most of its branches, is 8' ; that of its truly 

 spherical portion may be put at 5'. But since the sine of 

 an angle of 5' is to radius about as 1 : 687, it follows that 

 the real diameter of this globe of stars is 1/687 of its dis- 

 tance from the earth. Assuming this distance to be such 

 as would correspond to a parallax of 1/20 of a second, we 

 find that the more compact part of the cluster measures 

 558,000,000,000 miles across. Light occupies about thirty- 

 six days in traversing it. The average brightness of its 

 components may be estimated at the twelfth magnitude ; 

 for, although the outlying stars are of the tenth and 

 eleventh ranks, the central ones are, there is reason to 

 believe, much fainter. The sum total of their light, if 

 concentrated into one stellar point, would at any rate very 



little (if at all) exceed that of a third magnitude star. And 

 one third magnitude star is equivalent to just 4000 stars 

 of the twelfth magnitude. Hence we arrive at the con- 

 clusion that the stars in the Hercules cluster number about 

 4000 ; and that Sir William Herschel, in estimating them 

 at 14,000, erred considerably on the side of excess. 



If, then, 4000 stars be supposed uniformly distributed 

 through a sphere 558,000,000,000 miles in diameter, an 

 interval of 28,365,000,000 miles, or more than ten times 

 the distance from Neptune to the sun, separates each from 

 its nearest neighbour. 1 Under these circumstances, each 

 must shine with about one thousand times the lustre that 

 Sirius displays to us. Since, however, five millions of 

 stars even of this amazing brilliancy would be needed to 

 supply the light we receive from the sun, the general 

 illumination of the cluster can only amount to a soft 

 twilight, excluding, it is true, the possibility of real night 

 on any globe situated near its centre. 



At the distance conjecturally assigned to this cluster, 

 our sun would appear as a seven and a half magnitude 

 star ; it would shine, that is to say, about sixty-three times 

 as brightly as an average one of the grouped objects. 

 Each of these, accordingly, emits 1/63 of the solar light ; 

 and if of the same luminosity, relative to mass, as the sun, 

 it exercises just 1/500 of the solar attractive power. The 

 mass of the entire system of 4000 such bodies is thus 

 equal to that of eight suns. This, however, may be re- 

 garded as a minimum estimate. The probabilities are in 

 favour of the cluster being vastly more remote than we 

 have here assumed it to be ; hence proportionately 

 more massive, and composed of brighter individual bodies 

 than results from our calculation. Differences of distance 

 are alone adequate to account for the variety of texture 

 observable in globular clusters. That in Aquarius, for 

 instance, compared by Sir John Herschel to " a heap of 

 golden sand," might very well be the somewhat coarse- 

 grained Hercules group withdrawn as far again into 

 space. At a still further stage of remoteness, the appear- 

 ance would presumably be reached of a stellar throng in 

 the Dolphin (" G. C." 4585), which, with low powers, might 

 pass for a planetary nebula, but under stronger optical 

 compulsion assumes the granulated aspect of a true cluster. 

 And many such, their genuine nature rendered impene- 

 trable by excessive distance, are doubtless reduced to the 

 featureless semblance of " irresolvable " nebulae. 



But there are real as well as apparent diversities in 

 these objects. Although smaller and more compressed 

 clusters must, on the whole, be more remote than large, 

 looser ones, yet " this argument," Sir William Herschel 

 remarked, " does not extend so far as to exclude a real 

 difference which there may be in different clusters, not 

 only in the size, but also in the number and arrangement 

 of the stars." There may be globular clusters with com- 

 ponents of the actual magnitude of Sirius ; others, 

 optically indistinguishable from them, may be aggregated 

 out of self-luminous bodies no larger than Mars, or even 

 than Ceres, or Pallas. There is, indeed, a strong likeli- 

 hood that clusters and nebulae form an unbroken series — 

 that swarms of meteorites are connected by such inter- 

 minable gradations with swarms of suns, as to admit of no 

 impassable barrier being set up between them. 2 The 

 rifted structure, for instance, and truncated spectrum of 

 the Hercules cluster bring it into unmistakable relations 

 with the great nebula in Andromeda ; yet it is scarcely 

 doubtful that the one object is an assemblage of orbs 

 eich of them, quite possibly, the rival of our sun in lustre ; 

 and the other, a collection of what we can only describe as 

 cosmical shreds and particles. Further analogies emerge 

 to view through the reproduction in many nebulae of the 

 " hairy " appendages of globular clusters, and in the 

 spirality of arrangement characteristic of both classes 



1 See J. E. Gore's similar calculation, based, however, on different data 

 from those assumed above, in Journal Liverpool Astr. Soc. vol. v. p. 169. 



2 See Mr. Lockyer's " Bakerian Lecture," p. 29. 



