550 



SCIENCE. 



[N. S. Vol. XXI. No. 536. 



advantage to bring the mirrors as near to the 

 eyes as possible, as was done in Sir Charles 

 Wheatstone's original arrangement. If we 

 think of the mirrors, as we may from an op- 

 tical standpoint, as simple openings through 



c p' 



m' 



OCO' is the stereoscopic card. MN and NM' 

 are the two mirrors. //' are the superimposed 

 images of OC and CO', and E, E', are the positions 

 of the two eyes. 



which we see the image, it is evident that the 

 eye gains much in freedom of position and 

 largeness of field, if it is brought near to the 

 mirror, precisely as we go close to a window 

 if we desire a more extended view. Also, since 

 the edge dividing the mirrors is thus brought 

 between the eyes, rather than in front of them, 

 it is no longer seen, or at least is no longer 

 troublesome, as it is to some extent when the 

 mirrors are farther from the eyes. Indeed, 

 the mirrors need not meet at all nor need they 

 be of any regular shape. 



With this arrangement two views may be 

 combined which are considerably wider than 

 those used in the ordinary stereoscope. I have 

 found no difficulty with drawings six inches 

 wide. The height of an object which can be 

 successfully used is limited by the condition 

 discussed below. But with views of the or- 

 dinary dimensions this stereoscope is entirely 

 satisfactory in its performance, possesses a 

 considerable range of adjustment, and is con- 

 venient for laboratory experiment, as it is 

 easily and quickly put together with two bits 

 of mirror and a little wax. 



All forms of reflecting stereoscope using a 



single stereoscopic card have this imperfection 

 in common, that the images formed by the two 

 mirrors do not coincide, but intersect at a 

 considerable angle. The images of any object 

 formed by two mirrors lie, as is well known, 

 on the circumference of a circle, the center of 

 which is at the junction of the mirrors, and 

 the images are separated by an angle equal to 

 twice the angle between the mirrors. Since 

 the relations between object and image are 

 reciprocal, it is plain that if the images of 

 two objects are to be superposed by means of 

 mirrors, forming one image, the objects must 

 lie on the circumference of the circle, and at 

 the angular distance occupied by the images 

 in the previous case. 



If this condition is not fulfilled, but the ob- 

 jects are in the same plane, as when they are 

 on the same card, the images, while nearly 

 superposed, will intersect at an angle equal 

 to twice the angle between the mirrors. In 

 my stereoscope, as commonly used, the angle 

 between the mirrors is about two and one half 

 degrees, so that the images form an angle with 

 each other of five degrees. The images are 

 over three inches wide. If their planes inter- 

 sect at the median line, the edges to the right 

 and left are separated in the direction of the 

 line of sight by more than an eighth of an 

 inch. This is hardly noticeable across the 

 breadth of the view, where the line of sight is 

 nearly perpendicular to the intersecting edge 

 of the mirrors, but becomes so near the top 

 and bottom, where the slight deficiency in 

 sharpness of the horizontal lines is easily trace- 

 able to their inclination. If the center of the 

 card is pushed back so that the card forms an 

 arc, approximately that of the circle on which 

 the views should lie, the improvement in defi- 

 nition is strikingly evident. 



It is an interesting illustration of the ease 

 with which the eye is satisfied in such matters, 

 that the stereoscopic result is excellent over 

 the whole view, hardly failing at all even at the 

 extreme edges, though formed by two images 

 so disadvantageously placed. 



Frank P. Whitman. 

 Western Re.serve I'niversity, 

 Cleveland, Ohio. 



