278 



♦ KNOWL.EDGE ♦ 



[Oct. 3, 1884, 



any case where the water is required for washing purposes. 

 The apparatus consists, virtually, of a series of tanks in 

 duplicate, in which the chemicals are mixed, and these 

 enter a vertical pipe in proper proportion to the water to 

 be softened, and which communicates with the bottom of an 

 upright chamber divided by a series of sloping shelves 

 through which the water gradually works up in a zig-zag 

 path. These shelves slope in one direction, and are of V 

 shape, so that as the deposit takes place, it accumulates at 

 one point, at which there is an opening ordinarily closed 

 by a tap, and when any tap is open the deposit on the 

 sloping shelf communicating with it is washed out. The 

 apparatus appears to be extremely simple in its design, but 

 its efiiciency has yet to be tested, although it is at work at 

 Messrs. DuncanV, Victoria Docks." 



OPTICAL KECREATIONS. 



By a Fellow of the Royal Astronomical Society. 



{Continued from j). 199.) 



SO far, we have spoken as though vision were confined 

 to a single eye ; but, of course, under all ordinary cir- 

 cumstances, we have, and employ, two. Let us then see 

 whether this makes any difference (and if so, what diffe- 

 rence) in our perception of external objects. Speaking 



discover that one of these images is referable to the right 

 eye, and the other to the left one. The fact is, that by 

 adjusting the eyes for distant vision we have thrown the 

 images of the pencil upon different points of the two 

 retinse, and in order that they should coincide they must 

 be made to fall upon identical points in each eye. A more 

 instructive experiment still, however, in connection with 

 binocular perspective, of which we are immediately about 

 to treat, may be made with a thin book. Holding such a 

 one as near to the eyes as is consistent with distinct vision, 

 exactly half-way between them, and with the back of the 

 book towards us, and first shutting the right eye, we shaJ 

 see the image L (Fig. 27) with the left one — i.e., the back 

 and left-hand side of the book. Closing the left eye nov 

 in turn, and opening the right one, we shall see the 

 image R, or the back of the book and the right side of it 

 Now here we have made a discovery. It is this : that ths 

 images of an object, at a short distance from the observer, 

 are dissimilar, and yet that the eyes unite these two dis- 

 similar images into one, which appears soiid as well a» 

 single. A little attention will show us what it is that gives 

 us this impression of solidity ; and this being compre- 

 hended, the principle of that pretty toy, the stereoscope, 

 becomes intelligible at once. If, first, as in the case of oui 

 pencil, we look at the opposite wall of the room, or at the 

 distant landscape, we see two separate books. Next, if wfr 

 look at the part of the book furthest from us, the two 



FiK. 27. 



Fig. 28. 



off-hand, anyone who has never had his attention especially 

 directed to the subject would say that vision with two eyes 

 is essentially the same as that with one, for that we only 

 see one image of any given object, whether we regard it 

 with the right eye, the left one, or both together. But a 

 little attention will suffice to convince us that this is by no 

 means the case, and that binocular vision supplies us with 

 a means of estimating solidity, distance, &c., which is 

 wholly wanting in the case of sight with a single eye. Ob- 

 serve especially though that llie two eyes 7nnsi act together 

 f(s one inntrument. A dragon-fly has an immense number 

 of eyes (some 24,000), but they are looking different waysand 

 forming images of different objects, so that the advantage 

 and peculiarity of binocular vision proper is wholly lost. 

 In order that it may be enjoyed, the images formed by 

 lioth eyes must be superposed, and caused to coalesce into 

 one. Now, so entirely accustomed are we to this coalescence 

 i>f the optical images formed by our two eyes, that it 

 requires a certain amount of thought to perceive and con- 

 vince ourselves that we really do see everything in reality 

 double. As a preliminary experiment to show this, let the 

 reader hold a common drawing-pencil a foot or eighteen 

 inches in front of his face, and look, not at the pencil 

 itself, but at the further side of the room, or at the 

 horizon, and he will at once see two pencils. By first 

 closing one eye and then the ether, he will immediately 



images will unite, but the book will seemingly have two 

 backs. While, lastly, if we direct our attention fixedly to- 

 the back of the book, we shall see only one back ; but the- 

 part furthest from the eye will now become double. If, 

 however, we take, so to speak, a general view of the book, 

 the eyes will range rapidly over it, and the result of suc- 

 cessively combining the images at various distances will be 

 to impress us with the conviction that the book is a single, 

 solid object In making this experiment carefally, the 

 student will note that the eyes converge, or he has to 

 squint, more to cause the images of the part nearest to his 

 eye to coalesce than he has to cause the combination of the 

 more distant one ; such convergence decreasing as the parts 

 recede from the eye ; so that at a certain distance no con- 

 vergence at all is needed to unite the two separate images 

 of any given object. It will also be noticed that the farther 

 an object is from the eyes the less dissimilar the images 

 they respectively form become, until beyond a certain 

 distance they are sensibly identical. And here we are 

 in sight of the principle of the stereoscope. As 

 far as the eyes are concerned, it matters nothing whether 

 they see two slightly dissimilar images of the same object, 

 or whether two images from two separate objects are 

 presented to them, so long as the latter resemble the 

 former in all respects. All that is needed is that by appro- 

 priate means these two images should be caused to overlap 



