Fbbbuaby 1, 1895.] 



KNOWLEDGE. 



3i 



magnitude than the real diameter of the star ; whOst its 

 photographic image is still further enlarged to an un- 

 reasonable degree. Thus, in the photograph reproduced in 

 our plate (for which we are indebted to the kindness of 

 Dr. Mas Wolf, of Heidelberg), the diameters of the stars 

 \, 22, 25 and 28 Cygni average 5 of arc each, which, if 

 ■we ascribe to them the very moderate parallax of 0-1" of 

 arc, would coiTespond to three thousand times the radius 

 of the earth's orbit — say, two hundred and eighty thousand 

 millions of miles. In other words, the diameters of their 

 photographic images are not improbably a quarter of a 

 million times as great in angular measurement as the 

 actual diameters of the real stars themselves. A fictitious 

 appearance of connection is thus created, which would 

 instantly vanish if we could produce a picture of the region 

 in which the diameters and distances of the stars were 

 both represented on the same scale. 



It cannot be denied that there is a great deal of force 

 in these three objections. They do not, however, apply to a 

 feature — the dark rifts or " lanes " — which photographs of 

 the Milky Way show in abundance. Of these the 

 accompanying plate shows some examples, though by no 

 means of the most striking character that could have been 

 selected. 



Sketch ilap of Plate, skomng direction of Dark 

 the Milky Way. 



Laues" in 



This photograph shows the MQky Way in the neighbour- 

 hood of y Cygni, and overlaps the photograph, also taken 

 by Dr. Max Wolf, which was reproduced in Knowledge for 

 October, 1891. The two photographs, with a third, which 

 appeared in Kxowledoe for December, 1891, should be 

 studied together. The star -/ Cygni, which is in the centre 

 of the present plate, is easily recognizable in the south- 

 west corner of the plate published in October, 1891. It 

 must, however, be noted that the scale of the two 

 photographs is not precisely the same. 



A very cursory inspection of the plate shows that it is 

 intersected by a number of narrow rifts or channels, 

 traceable with equal distinctness over the richer and more 

 nebulous region on the west, and the poorer district on 

 the east. 



That these rifts are not mere accidents of this particular 

 photograph can be seen by a comparison with the plate 

 already alluded to as pubUshed in October, 1891. The 

 two great lanes in the north-east comer of the present 

 plate are shown with even more distinctness in the 

 south-west comer of the earlier photograph, which was 

 taken with a different telescope, on a different scale, with 

 a different exposure, and on an occasion three years earlier. 

 The longer of the two rifts is traceable on the earlier plate 



for 10° in a direction which scarcely deviates appreciably 



from the rectilinear, and ceases to be visible, not because 

 the rift itself has come to an end, but because the photo- 

 graph does. The chief rift on the present plate is not 

 quite so conspicuous, but may be traced from the south side 

 of the plate, a little to the west of A. Cygni, to a point near a 

 bright star, where it forks off into two directions. Of these 

 the westerly rift passes south of y Cygni and tums nearly 

 due west, its course being traceable to the edge of the plate. 

 The other branch runs more nearly northward, passes very 

 close to y Cygni on the east, and then, turning in a direction 

 nearly parallel to the first branch, gradually loses itself. 



The courses of these and other principal rifts are roughly 

 indicated in the index map. 



The objections brought against the reality of Unes and 

 chains of stars do not hold good when applied to these 

 singular rifts. The tendency of the eye, or rather of the 

 mind, to form patterns from disconnected dots has no 

 application to the case of a long narrow strip of sky, 

 practically void of stars. The enlargement of the discs 

 of stars from photographic diffusion would tend to conceal 

 such rifts when they do occur, not to counterfeit them 

 when they do not ; and the existence of any other stars 

 in the line of sight, either nearer than the rifts or beyond, 

 would interrupt or conceal them. 



There are not a few cases in the plate before us where 

 it appears as if a rift were rendered less obvious by the 

 intervention of a small cluster of stars or, in one instance, 

 of nebulous matter. But the course of the rift is so clearly 

 the same on both sides of the interruption that there can, 

 I think, be little doubt that we have in these cases merely 

 the accident of a few luminous objects in the same apparent 

 direction from us, but wholly imconnected with the rift. 

 We have the right to expect that sitch would sometimes 

 occur, and their occurrence is therefore no argument 

 against the reality of the rifts. 



But what can the rifts or ' ' lanes ' ' themselves be ? Seeing 

 that they extend for distances which must be reckoned, 

 not by billions of miles, but by the million billion — for a 

 length of 8° or lO'can hardly involve less — it is impossible to 

 conceive that we have here a case of absoi-ption by streams 

 of dark bodies. It must be borne in mind that many of 

 the rifts must be far longer than they appear to be ; for 

 though our position within the circle of the Milky Way wiU 

 necessarily tend to represent rifts parallel to its mam axis 

 as practically of their true proportions, many others must, 

 no doubt, be seen by us as more or less foreshortened. 

 Eefer again to the a Cygni plate, of October, 1891, and note 

 how the greatest rift of all rises sharp and dark against the 

 very brightest part of the great nebula on the east of the 

 plate, and following a sinuous course turns up northward 

 in one branch along the east edge of the plate, and in 

 another westward across the centre of the plate just north 

 of a Cygni. The boundaries of the rift in both its branches — 

 and, indeed, also in a third branch which springs from the 

 same point of divergence, and running through a sparser 

 region can be traced south of a Cygni to a junction with the 

 second branch near the west side of the plate — are defined 

 sometimes by nebulous matter, sometimes by stars._ And 

 all three are traceable alike in nebulous regions, regions of 

 closely-clustered stars, and where they are thinning off. 

 The reality of the rifts cannot, then, be challenged, for they 

 are independent of the character of then- bordering. 



If not dark absorbent matter — and I do not see how that 

 idea is tenable in the case of markings so prolonged and so 

 comparatively narrow — what other conclusion can we come 

 to but that the rifts mark regions of real barrenness of 

 stellar and nebulous material ? So regarded they become, 

 I think, eloquent of a process of condensation going on in 



