280 



KNOWLEDGE. 



[Decbmbeb 1, 1896. 



The mechanical arrangements of globular clusters 

 entirely baftle our narrow conceptions of what is feasible 

 and workable. Instead of the neatly finished aspect 

 betokening orderly revolution round an attractive centre, 

 they present, very commonly, ravelled edges, a radiated 

 conformation, and dark vacancies of curiously definite 

 shapes. An escape of stars is strongly suggested ; and the 

 conjecture might even be hazarded that the removal of 

 stellar material thus effected is the immediate cause of the 

 inward dilapidation manifestly progressing in visibly 

 tunnelled spheres. The two symptoms indeed appear to 

 be correlated. Three imposing groups— M 13, in Hercules, 

 M 3, in Canes Venatici, and M 5, in Serpens, may serve as 

 examples. All have curvilinear appendages, and all show 

 pierced, and, as it were, excavated interiors. On the other 

 hand, w Centauri and 47 Toucani, an exquisite ornament 

 of the southern pole, are compact within and without. No 

 perforations are visible in them, and on Bailey's plates they 

 came out almost perfectly circular. Here the twofold marks 

 of dissolution are together absent, as in other cases they 

 are together present. 



By far the most remarkable discovery, however, yet 

 made about globular clusters is that about one in five of 

 them are literally crammed with variable stars. Their 

 abundance is such, that as many as a hundred — in 

 Prof. Barnard's words •' — "have been found in a space 

 in the sky that would be covered by a pin's head held at 

 the distance of distinct vision." The pictures affording 

 this wonderful revelation were taken at Arequipa, with 

 the thirteen-inch Boyden telescope, an instrument rendered 

 available for either visual or photographic employment by 

 the adjunct of a reversible crown lens. Its fine qualities 

 are not allowed to " rust " in disuse. Hundreds of 

 exposures, from one up to six hours in duration, have been 

 made with it for the purposes of this special enquiry, the 

 results of which have been published in successive 

 " Harvard Circulars." They are most nearly complete for 

 (y Centauri. In this cluster, out of about three thousand 

 stars accessible to separate study, no less than one hundred 

 and twenty-five proved markedly, and, for the most part, 

 very rapidly variable. A large majority, in fact, run 

 through their changes in less than twenty-four hours. 

 The periods of one hundred and six have, so far, been 

 ascertained; only eight among them exceed a day's length, 

 while three fall short of seven hours. One of these belongs 

 to No. 91 of the blinking battalion, which, springing up to 

 a maximum once in six hours and eleven minutes, is at 

 present the quickest of known variables. U Pegasi, until 

 lately the claimant of that distinction, is outrun by many 

 components of clusters. 



As will be seen by referring to Circular No. 38, printed 

 in abridged form lower down. Prof. Pickering divides the 

 short-period variables in w Centauri into four classes, 

 distinguished by the forms of their light-curves. The first 

 largely predominates. The objects constituting it increase 

 with extreme swiftness, and decline by comparison slowly. 

 No. 45, for instance, with a period of 14h. 8m., sextuples 

 its brightness in a single hour, and that on the clock-stroke, 

 all these stars being characterized by exemplary 

 punctuality ; in other cases the rise may be still more 

 rapid, but closer inquiry is precluded by the needful 

 duration of photographic exposures. 



Some of the :u Centauri variables show humped light- 

 curves, indicating abortive secondary maxima, like those 

 of 'J Cephei and >; Aquilre ; and one exceptional case has 

 been noted, in which the rule of an ascent quicker than the 

 descent is reversed. 



* See his recent admirable address on " Astronomical Photography," 

 p. 26. 



Besides w Centauri, three stellar globes — M 5, M 3, and 

 M 15 — have yielded a copious harvest of rapidly-changing 

 stars. The first includes about nine hundred components 

 that can be individualized and watched ; eighty-five among 

 them are conspicuously variable. Two were visually 

 discovered as such by Mr. David Packer, of Birmingham, 

 in 1890. They have been identified by Prof. Barnard 

 as Nos. 42 and 84 of the Harvard register, and might be 

 called companion objects, since each alike fluctuates to the 

 extent of a magnitude and a-half in a period of twenty-six 

 days. Dr. Common, about the same time, obtained 

 photographic indications of variability throughout the 

 cluster* ; but so delicate an enquiry could scarcely be 



Fia. 1. — Light-Curve of No. 18 Messier 5. 



prosecuted under the muffled skies of Ealing. The perfect 

 conditions at Arequipa were indispensable to success — a 

 success enhanced by Prof. Barnard's confirmatory work 

 with the forty-inch Yerkes refractor. " These cluster 

 variables," he remarks, t " seem to form a distinct class 

 from the ordinary variable stars. It is very interesting to 

 watch one of them in a powerful telescope, and to see with 

 what quickness it passes through its light-variation. One 

 of the small stars in M 5, whose period is 12h. 31m., seems 

 to be dormant for a large part of the time, as a very faint 

 star, invisible in ordinary telescopes. It begins to brighten, 

 and in two or three hours has risen nearly two magnitudes, 

 and faded again to its normal condition ; while another 

 and larger star, quite near it, 

 seems to require a month or 

 more to go through its light 

 fluctuation." 



Most of the eighty-four vari- 

 ables in M 5 are doubtless of 

 the type dominant in w Cen- 

 tauri. They traverse their 

 cycle in a few hours, ranging 

 through a couple of magni- 

 tudes by very unequal grada- 

 tions. From a datum-level of 

 obscurity, they spring up al- 

 most suddenly, and with the 

 regularity of a flash-light, then sink back at a more 

 leisurely rate. The stationary intervals apparent to the 

 eye are, however, shown photographically to be marked by 

 the progress of slow variation. Absolute pauses are short 

 and rare. No. 18, of which the light-curve is given in 

 Figure 1, illustrates the peculiarities of the class. Its 

 period is llh. 7m. 52s., its limiting magnitudes 13-5 to 

 14-7. Probably no other star equally faint has been 

 pursued throughout its phases. During nearly five hours 

 it remains semi-extinct, but needs no more than an 

 hour and quarter for a triple gain of light, which it 



Fig. 2. — Maximum of 

 No. 12 Messier 5. A long 

 minimum is indicated. 



* Monthly Notices, Vol. L., p. 517 

 ■fAstr. Ifach., No. 3519. 



Nature, Tol. L., p. 448. 



