so Dr. Tyndall on the Progi^ess of the Physical Sciences : 



mixes itself with the principal image reflected from the silvered 

 surface behind. To prevent the appearance of such images, 

 total or metallic reflexion is necessary. In stereoscopic investi- 

 gations, where the contours of the images do not exactly coin- 

 cide, a doubt may arise whether this be not due to the imperfec- 

 tion of the drawings made use of. The desire to set aside this 

 possible source of error has led to the construction of the follow- 

 ing stereoscopes. 



a. Prism-stereoscope J consisting of one Prism and one Drawing 

 intended for a single eye. 



The condition of complete identity can only be fulfilled when 

 one object, and not two, is observed. An object seen in a look- 

 ing-glass appears reversed as regards right and left. In all 

 stereoscopic drawings, which difiier only in the circumstance that 

 one is a reflected image of the other (and to this class belong the 

 greatest number, even the most complicated of those hitherto 

 executed), the reflected image of the object can take the place of 

 the second drawing. Such drawings may be named simple rever- 

 sions, in contradistinction to those in which a plane perpendicular 

 to the line which joins both eyes and passing through the centre 

 of the body does not divide the latter symmetrically. If an erect 

 object be viewed through a prism whose ends are right-angled 

 isosceles triangles, the prism being so placed that its plane of 

 refraction is horizontal, that is to say, with its hypothenuse sur- 

 face vertical, the object will remain erect ; but with regard to 

 right and left, it will be found to have changed sides. Through 

 such a prism, therefore, the arranged type in a printer's work- 

 shop may be read as the page of a common book ; the letters 

 change places right and left, and thus appear as when they are 

 printed. In the same way a profile is reversed. If a common 

 stereoscopic drawing intended for the left eye be viewed by the 

 right eye through such a prism, then if the projections be simple 

 reversions of each other, the image will appear as if designed for 

 the right eye. To the left, or naked eye, the drawing appears 

 unchanged. It is easy, by turning the prism round one of the 

 edges which stands perpendicular to the plane of refraction, to 

 cause both images to coincide ; the moment this takes place, the 

 relief starts into existence with surprising sharpness. 



If the position of the drawing remain unchanged, and the 

 prism be held before the left eye, then if the object be conceived 

 to be transparent, so that both its exterior and interior surfaces 

 are risible, the forward surface will appear converted into the 

 surface behind, and vice versd. When, on the contrary, a draw- 

 ing intended for the right eye is viewed through the prism by 

 the left eye, it will appear exactly as in the first case, always 

 provided that it is viewed at the same time by the naked right 



