BINOCULAR VISION 593 



front of their previous position, in fact that the closer they are placed to- 

 gether, the nearer do they appear to the observer. But the indices do not 

 show double images, unless they are moved a considerable distance to- 

 gether, and then the effect of distance ceases. If one of the index slides 

 be removed and the other be moved towards and away from the axis of 

 the instrument, the index is not found to shift its plane towards or 

 away from the instrument. This shows that for position to be appreciated 

 both images must be presented to consciousness simultaneously without 

 appearing double. 



THE ACUITY OF STEREOSCOPIC VISION has been investigated in such a 

 way that other factors which normally assist distance perception were excluded. Two 

 methods have been used : (1) to adjust the position of a thread which lies between and 

 parallel with two other threads until they all appear at the same distance from the 

 observer ; (2) to observe the fall of small coloured bodies of unknown size, and then to 

 state the position of the line of fall in relationship with a fixation mark. The former 

 method at 2 metres distance shows an average error of 1-5 mm., the latter method at 

 the same distance an error of 40 mm. The difference between the results of the two 

 methods is considerable ; but it should be noted that in the fall method the object 

 is seen only for -02 sec. If in the thread method the threads be placed horizontal it 

 is found that the appreciation of distance is greatly impaired. The greatest acuity 

 is found when the threads are vertical. If however the head is turned so that the line 

 joining the two eyes is vertical, the greatest acuity is found when the threads are 

 horizontal. This would be expected if the appreciation of distance is greatest when the 

 parallax of the objects at the two eyes is greatest. Experiment shows that the recog- 

 nition of position in relationship with a definite fixation mark is much more accurate 

 than recognition of absolute distance in which there is no point of reference. Thus it 

 is well known how inaccurate the estimation of the distance of a single source of light 

 at night may be. 



HYPOTHESES OF DEPTH PERCEPTION. Javal's view was that the move- 

 ments of the eye muscles, which are necessary in order to direct the gaze from objects 

 in one plane to those in the next, caused impulses to travel to the brain which are 

 interpreted in terms of distance. This view was ruled out by the fact that images 

 which are formed on the retina for a short length of time only (-02 sec.), are able to 

 be perceived in relief. 



HERING'S HYPOTHESIS was that it is the formation of similar images on points 

 of the retinae that do not correspond which causes distance perception. If the 

 disparation is crossed the object appears nearer than the fixation mark by an amount 

 which depends on the amount of the disparation. If on the other hand the disparation 

 is uncrossed the object is recognised as being further away. Hering supposed further 

 that crossed disparation acts as a stimulus to convergence and accommodation, while 

 uncrossed produces the reverse effect. We may now inquire how this hypothesis 

 fits in with the facts. To commence with, if depth depends on disparation it is clear that, 

 when we perceive objects lying in different planes, we must subconsciously group them 

 according as they fall on corresponding retinal points, or on points which are discrepant 

 by one, two, three or more cone widths, and whether the discrepancy is crossed or 

 uncrossed. The amount of the discrepancy must be some whole number of cone widths 

 because it is clearly impossible to stimulate half a cone with one impression and the 

 other half of the same cone with a different one and obtain two distinct sensations. It 

 is clear that space must be divided so. far as stereoscopic vision is concerned- 

 into a number of concentric shells, the centres of which correspond with the position 

 of the observer. Now the thickness of these shells can be readily calculated : at 1 metre 

 they are found to be 2 mm. thick, at 10 metres 200 mm. thick, and at 100 metres 17 

 metres thick. If we are looking at a fixation mark 10 metres away, objects 



38 



