72 



SCIENCE. 



[Vol. XI. No. 262 



But now, if we take a fine piece of wire, a knife-blade, a needle, 

 or a sharp pencil-point, such objects being used in order to get 

 double images more easily, and place it a short distance farther off 

 than the apparent position of the central circle while we keep the 

 attention upon some point of the circumference of the circle, at a 

 very short distance beyond the point of fixation the needle or piece 

 ■of wire will appear double, and represents the ordinary homony- 

 mous images, which are the images localized beyond the horopter. 

 We may increase this distance of the needle from the point of con- 

 vergence, and the distance widens as usual between the images. 

 There is perhaps nothing new in this fact. But if we keep the con- 

 vergence of the eyes perfectly fixed for the combination of the two 

 ■circles to form the central one, and turn the attention to the two 

 homonymous images apparently beyond the point of convergence, 

 and without allowing the convergence to change so as to combine 

 ■the images of the needle, we shall find, by very close attention, that 

 .they wiW instantaneously spring into the position of heteronymous 



images, nearer the eyes than the circle, and without either becom- 

 ing really heteronymous, or in the least approaching each other. 

 Rivalry often takes place between the two positions, so that the 

 images of the needle will alternately seem nearer and farther than 

 the central circle at the point of convergence. 



A beautiful way of testing the same result is to place the knife- 

 point or needle upon the sheet of paper, and coinciding with any 

 point in the circumference, but always allowing the length of the 

 object to lie in, or parallel to, the vertical meridian. If the atten- 

 tion is fixed strongly upon the knife-blade or needle while conver- 

 gence combines the two circles, the two images of the needle or 

 blade seem to coincide with two of the circles, the central and 

 combined circle, and one of the outer circles. But the central and 

 combined circle seems in the same plane with the sheet of paper 

 and the other two circles. This may vary, however, with rivalry, 

 as experience will show. But if now we begin to move the object 

 toward the eyes, and therefore toward the point of convergence, 

 without altering the latter, and without changing the attention, the 

 two images of the needle or knife-blade will appear nearer than the 

 -central circle, and also seem to approach the eyes until they reach 

 a certain point, where they instantaneously assume the homony- 

 mous position beyond the central circle. The feeling of surprise is 

 very marked at this sudden appearance of tlje images at a greater 

 -distance than they had just seemed. 



If, again, we draw the circles upon a plate of glass in order to 

 combine them by fixating beyond it, and try the experiment as we 

 have described it, the images at first appearing beyond the central 

 circle and homonymous, by close attention will suddenly appear in 

 the heteronymous position, nearer than the central circle, as be- 

 fore. It must be remembered, however, in both cases, that the 

 images do not become really heteronymous, as can be proved by 

 suddenly closing and opening ojie of the eyes. The same image 

 vanishes in both apparent positions of the double images. The 

 single interesting fact, both when we combine by convergence 

 and when we combine by fixating beyond, is that the two images 

 of the object really beyond the point of fixation will appear at times 

 to be nearer, and will not assume a fixed homonymous position 

 until the attention upon them is relaxed. Now for the explanation. 



It is clear that the double images of the needle or knife-blade are 

 simply the ordinary homonymous images, and hence are localized 

 beyond the horopter, or point of fixation. So far the phenomenon 

 only accords with the ordinary law. The anomaly appears when 

 their relative position is changed and they seem translocated into 

 the heteronymous position. But if we revert to the influence of at- 

 otention in all sensory processes, we may discover a cause for the 



effect we have described. It is known that we may so absorb our 

 attention as to be unconscious of a severe pain in the tactual sense. 

 Or in vision we may be so occupied with a particular object as not 

 to notice the presence or approach of another. We may even lose 

 entire sight of all objects except the one in which we are interested. 

 Again, it is a universal fact that attention directed to any object in 

 the field of view, at once and automatically sets the eyes into the 

 proper movement for adjustment to produce single vision. At the 

 same time the visual tension of the eyes is relaxed for the object 

 from which the attention is turned. With these simple facts, we 

 may turn to the experiments we have described. Here, when we 

 keep the adjustment for combination constant, but direct the at- 

 tention to the two homonymous images, the tension for binocular 

 localization is relaxed by the change, and we are left to monocular 

 principles for the localization of the images of the needle as well as 

 that of the central and combined circle. The latter appears in the 

 same plane as the sheet of paper, or approximates it in proportion 

 to the relaxation of binocular tension, and thus introduces monoc- 

 ular influences into the localization of combined images, while only 

 monocular functions are left to localize the homonymous images of 

 the needle or knife-blade. Hence it appears as it really is ; namely, 

 nearer than the central circle. We may test whether it is due to 

 the prevalence of monocular over binocular innervation by moving 

 the needle far enough off to make its images coincide, or nearly 

 coincide, with the circumference of the combined circle at the ter- 

 mini of the diameter, and, while they seem in the heteronymous 

 position, suddenly close and open one of the eyes. We shall see 

 the remaining image of the needle apparently nearer than the cir- 

 cle, and in the same position, without change, which it occupied 

 before closing the other eye. The eye must be closed and opened 

 as quickly as possible, so that the other eye will not have time to 

 resume the parallel position, and hence there will be no apparent 

 motion of the circles. This will enable us to determine more ac- 

 curately the monocular character of the localization of the homony- 

 mous images. We see the image of the needle and the circle in 

 the same relative positions as before closing the eye ; and, since 

 this can be only monocular, we can best suppose that the trans- 

 location we have described is due to the prevalence of monocular 

 functions over the binocular by the withdrawal of attention from 

 the latter. 



It is a still more interesting fact that the writer has been able, 

 by considerable practice, to localize one of the images of the needle 

 homonymously under the circumstances described, and the other 

 heteronymously. I have been able to alternate them to some ex- 



tent, although generally it is the left image that appears nearer, and 

 the right image farther, than the point of binocular fixation. In 

 such cases evidently one eye can keep up .the binocular innerva- 

 tion, while the other becomes monocular in it. Astonishing and 

 presumptuous as such a supposition may seem, it is entirely con- 

 firmed by the following second experiment, which also illustrates 

 the rivalry between binocular and monocular functions, as in Fig. I. 

 Take the circles A and B, with the smaller circles a, b, and c, as 

 we have drawn them, and combine them by convergence. It is 

 plain that the fusion of b and c will take place at the same time 

 with that of A and B. But a has no corresponding circle in B 

 with which to fuse. If b were absent, the binocular effort at con- 

 vergence would automatically tend to combine a and c, so that they 

 would appear nearer than the fused image of A and B in the pre- 

 cise ratio of the convergence required for their combination. We 

 have elsewhere worked out the explanation of all such localization 



