Mr. C. J. Monro on a case of Stereoscopic Illusion. 17 



and I presume the error due to this substitution would be some- 

 thing quite imperceptible to the lateral parts of the retina. The 

 eyes are converged indeed upon points very near them, and 

 focused for points further off; and though the disproportionate 

 recession of the optical centres and advance (if there is any 

 advance) of the retina due to this inconsistency would affect all 

 points of the picture alike, it is intelligible, on Mr. Abbott's 

 principles, that a flattening of the retina, apparently inseparable 

 from its advance, might produce the effect of convexity by draw- 

 ing lateral images a little nearer to the yellow spots. But any 

 illusion depending upon the properties of lateral parts of the 

 retina would disappear upon running the eyes rapidly over the 

 field. Meanwhile, having repeatedly made the experiment both 

 before and after meeting with Sir David's observations, I have 

 never detected the slightest curvature in the apparent surface, 

 except in two cases — first when the real surface is curved, and 

 secondly when the eyes are at different distances from it. In 

 the cases described the real surface is a flat wall, and the obser- 

 vers are evidently supposed to be looking straight at it : besides, 

 if they had looked obliquely the curvature would have been con- 

 cave, according to the theory ; for the surface the theory gives is 

 a hyperbolic cylinder asymptotic to the wall. 



It is difficult, if not impossible, to see this in patterns which 

 vary vertically as well as horizontally ; but if the pattern consists 

 simply of vertical lines, nothing can be more vivid than the object 

 sometimes suggested. I combined in this way the images of a 

 straight row of flat vertical bars, about an inch and a half wide 

 and an inch and a half apart, so as to form a stereoscopic image 

 beyond them : my nearest eye was about six inches from 

 their plane, to which the line joining the two was inclined at a 

 considerable angle. The bars immediately formed themselves 

 into the alternate faces of a vertical prism, of which the circum- 

 scribing cylindrical surface was apparently asymptotic to the 

 plane of their real position, and came up to a sort of apse quite 

 close to me, and then stretched off indefinitely away from the 

 plane as far as the binocular field extended. The sudden changes 

 of azimuth made by the flat surfaces as they turned the corner 

 of greatest curvature gave the image an effect of solidity as per- 

 fect as that of real nature. Now every test of distance and posi- 

 tion, except that of the intersection of the visual rays, was against 

 the production of the result produced. The very points which 

 the focusing of the eyes affirmed to be the nearest stretched 

 away in the further sheet of the surface ; and while the nearly 

 equal illumination of the bars by a single lamp affirmed that they 

 were parallel, they seemed to stand round a very convex surface. 

 So even without positive measurement, which seems impossible, 



Phil Mag. S. 4. Vol. 29. No. 193. Jan. 1865. C 



