ing but a single eye, that looks into the very depths of the observer's Cyclopean eye.
The conception of this subjective union as the product of the experience of the race in interpreting sensations, and the consequent necessity of distinguishing between realities and their visual representations, seems never to have been appreciated until long after the invention of instruments for use in the analysis of vision. Much confusion has resulted from the attempt to explain what are really subjective results of retinal sensation by the application of geometric principles, irrespective of the illusive union of the two eyes when employed together. In 1604 Kepler stated that the distance between the eyes constituted a baseline, which we employ for measuring the distance of objects by a species of visual triangulation. This idea was subsequently greatly elaborated by Sir David Brewster and others; and in most, if not all, of our textbooks of physics today it is applied in a very familiar diagram to explain the principle of the stereoscope. On this theory the apparent position of every point in the stereoscopic field of view is determined by the meeting of separate visual lines, which converge in front. An obvious consequence is that this localization should become impossible if the visual lines become parallel or divergent. But, in truth, there can be no perception of locality by this method. If the eyes are subjectively united, the visual lines become subjectively united along with them; if, indeed, such language is at all applicable to lines that are mere abstractions. In its application to stereoscopic vision, therefore, the diagram is worthless; for such vision is much easier to most persons when the visual lines are parallel, or very slightly divergent, than when they are strongly convergent, and in no case can there be any recognition of intersection between lines which, if subjectively perceived at all, would be coincident throughout their whole extent.
The error just mentioned has undoubtedly sprung from the assumption that stereoscopic vision is always perfectly normal. If this be so, it should be as painless as the reading of this page, even when continued for hours in succession. Every one who has tried the experiment with an ordinary stereoscope, and a large, miscellaneous collection of stereographs, knows how wearying it is, and how in some cases distinct vision is found impossible. To indicate the real differences between normal vision and that which is attained in most stereoscopes, it will be necessary first to study the development of this instrument.
The duality of human vision of near objects, and the consequent dissimilarity of retinal pictures in the separate eyes, was apprehended and more or less vaguely discussed by Euclid (b. c. 300), Galen (a. d. 200), Baptista Porta (1593), Leonardo da Vinci (1584), Aguilonius (1613), and by Smith, Harris, and Porterfield during the eighteenth century. No practical results were wrought, however, until 1838, when Sir Charles Wheatstone read before the Royal Society his