hy the Light of the Electric Spark. 



Upon each arm of the stereoscope 

 frame, pivoted centrally over a divided horizontal circle (fig. 

 1), so that the plane of the card that was fitted into it could be 

 made to assume any desired angle {<p) with the direction of the 

 arm. A pair of cards on which were similar 

 trie circles were then introduced, the arms being arranged for 

 parallel vision, and the frames directly across them. The 

 binocular resultant was of course a circular flat plane vertically 

 across the combined line of sight. The manipulator then 

 turned each frame on its pivot through an angle, whose natt 

 whether positive or negative, was unknown to the observ 

 and then passed a spark'. For values of this angle less than 

 30° or 40°, the first spark was usually sufficient to enable the 

 observer to determine whether the binocular resultant, due to 

 Opposite obliquity of projection upon the two concave reti 

 was itself convex or concave. This was tried successively 

 independently by Professor Eood, Mr. Share, and the wri 

 with uniform results, the only difficulty consisting in the pre- 

 vious attainment of proper adjustment for the position of the 

 head, and in adaptation of the ciliary muscles. For larger 

 values of the angle between card and visual line, the degree of 

 dissimilarity between the two retinal images sometimes caused 



circle being 8 cm. and the sum of the incident and reflected 

 rays from its center to the eye of the observer being 25 cm., it 

 becomes possible to calculate the maximum difference horizon- 

 tally between the two retinal images. Let m and m' (fig. 1) be 

 the points of incidence for rays from the centers c and c', of the 

 circles whose horizontal diameters are ab and a'b', the cards 



