Nov. 4, 18S1.] 



KNOV/LEDGE * 



11 



strictly, in order to produce an image of the star on sucli 

 and such precise pait of the retina. 



But there are many cases in which the senses seem 

 actually to convey false information, the eye telling us 

 wrongly about the shape, size, position, <tc., of an object, 

 the touch deei;i\ing us as to its form and qualities, the 

 hearing, the smell, the taste, each in its own way deceiving 

 us. It is such cases as these that I propose to examine. 

 It is most important for the student of science that he 

 should be aware of the various forms of error into which 

 the direct action of the senses may lead him. I could cite 

 instances where, for want of precisely such information, 

 the student of science has been at the pains to explain a 

 phenomenon which had no real existence, or which was 

 precisely the reverse of reality, ilr. Proctor mentions 

 somewhere a case where a French astronomer (Chacornac), 

 was careful to explain v:]iy the edge of Jupiter's disc is 

 brighter than the middle, the fact being that the middle is 

 brighter than the edge, the apparent brightness of the 

 edge being a mere delusion. It may be added that the 

 discussion of the class of illusions referred to is full of 

 interest The reader will tind that the careful considera- 

 tion of the cause of illusions will generally suggest other 

 illusory experiments, sometimes more striking than those 

 here described. There are few occupations more interest- 

 ing as pastimes, and at the same time more instiiictive, 

 than the invention and testing of various forms of illusion. 

 It should be mentioned at the outset that such illusions 

 as these are ^^Tongly called sensory illusions. As Huxley 

 points out, " there is no such thing as a fictitious or 



delusive sensation But the judgments we form 



respecting the causes and conditions of the sensations of 

 wliich we are aware, are very often erroneous and delusive 

 enough ; and such judgments may be brought about in the 

 domain of every sense, either by artificially-contrived com- 

 binations of sensations, or by the influence of unusual 

 conditions of the body itself." He adds, "mankind woidd 

 be subject to fewer delusions than they are if they con- 

 stantly bore in mind their liability to false judgment. 

 Men say, ' I felt,' ' I heard,' ' I saw' such and such a thing, 

 when, in ninety -nine cases out of a hundred, what they 

 really mean is, that they judge that certain sensations of 

 touch, hearing, or sight, of which they were conscious, 

 were caused Viy such and such things." It is precisely this 

 lesson wliich 1 want to enforce in the present series of 

 papers. 



The simplest illusions are those 

 afiecting straight lines, causing 

 them to appear crooked, or curved, 

 or vice versa. 



Take, for instance, the case 

 illustrated Ln Fig. 1. Here the 

 parts of the broken line AD 

 appear to be parallel, but not — 

 as they really are— in the same 

 straight line. The part CD seems 

 lower than the true prolongation 

 of the part AB. The explana- 

 tion of this is not so obvious as 

 in some less simple cases which I 

 shall consider further on. ilani- 

 festly, the mind recognises that, 

 owing to the interposition of the 

 space EH, the point where the 

 prolongation of AB should appear 

 must lie lower down than B, but it underestimates 

 the allowance to be made on this account. This 

 is equivalent to saying that the eye under estimates 

 the breadth of the strip EH. Why this should be 1 It 



seems to me it is because the space EH appears t© lie 

 above the plane containing AB and CD, so that the real 

 breadth of EH is unconsciously judged to be less than it 

 really is ; for, of course, a strip nearer the eye than EH 

 is would seem broader than EH if really of the same 

 breadth, and if seeming no hroiuler tiian EH, would really 

 be of less breadth. The eye judges that EH is nearer, and 

 infers unconsciously that it is of less breadth than EH 

 really is. 



Let us test this explanation, first seeing whether a strip 

 which breaks a curve, as EH breaks the straight line AD, 

 really secins nearer to the eye than the curve. 



Describe a semicircle AD (Fig. 2) broken at BC, and 

 between B and C draw the outline of strip BF as in the 

 figure. We see at once that the space BF appears to be 

 nearer the eye than the portions ABE, CFD. Doubtless 

 the judgment unconsciously reasons that the strip which 

 breaks the outline ABCD must lie upon the semicircle and 

 hide the portions wanting. 



Again, it seems to me and to some others, but not to all 

 to whom I have shown Fig. 2, that the figure formed by 

 carrying on the curves AB, DC, to meet over space BO 

 would not be a complete circle, but somewhat contracted 

 horizontally. This would correspond with the above ex- 

 planation of the illusion illustrated in Fig. 1. As, however, 

 all eyes do not recognise this second illusion of Fig. 2, let 

 us modify the experiment, noting that the same eyes which 

 might not recognise an illusion apparently affecting breadth 

 only, would readily be aflected by an illusion afiecting 

 direction. (We may remember what Brewster points out 

 in dealing with one of the illusions mentioned in his Natural 

 Magic, tliat in many cases it is the most observant eye that 

 is aflected bj' an illusion. In the present case the eye which 

 has the best power of estimating breadths would be the 

 one most affected by illusion as to the breadth of the space 

 BF.) 



Draw then the strip GH, Fig. 3, crossing the semicircle 

 ABDF as shown. Here the curve CD seems to belong to 

 a smaller circle; than that of which it really forms part. It 

 looks as though the curves AB, FE caried on beyond the 

 strip GH would lie about considerably outside CD. 



It will be found that if the parts CD, BE of the sides of 



