A <^ THE SPACE-THRESHOLD OF COLOURS; AND ITS DEPENDENCE UPON CONTRAST BY W. B. LANE, IW.A. THE SPACE-THRESHOLD OF COLOURS AND ITS DEPENDENCE UPON CONTRAST. FIRST ARTICLE BY VV. B. LANE, M.A. I. INTUODUCTOHY. In an artichi by Mr. J. O. Quantz, B.A., published in the Ameri- can Journal of Psychology, Vol. vii., No. 1, the dependence of size estimation upon colour is discussed. This problem naturally sug- gested a converse inquiry into the dependence of colour perception upon size. But colours are not ordinarily seen in complete isolation. There is always an environment which exercises, through contrast, a decisive inHuence upon the perception of colour. It therefore appeared desirable not merely to investigate the question under con- ditions of contrast, which in some form was unavoidable, but to examine into the specific nature of the relation between particular kinds of contrast, e.g., pure colour contrast and magnitude of coloured surface. The tield of optical contrast is of great extent and includes a considerable variety of distinguishable influences. It will be neces- sary to discriminate between them both for the purpose of investi- gating the bearing of contrast upon the problem in hand, and in order to define the particular kind of contrast with whicli we are concerned. Optical contrast has been divided into successive and simultaneous contrast. The former, successive contrast, is identical with th phenomenon of after-images. If one turns away after look- ing at a dark surface on a bright ground, one will have a series of alternating positive and negative after-images, in the latter of which the ground will appear dark and the dark surface bright. This inver- sion of the relationship between surface and ground in the negative after-image is a case of so-called successive contrast. Tliat part of the object of consciousness which at one moment was bright appears at a subsequent moment dark, and the dark portion appears bright, a fact which seems to indicate that the relationship between the parts of the after-images in respect to brightness is inversely dependent [1] 2 upon the same relationsliip as it exists between tlie parts of tlie preceding original images. Hehnholtz in his Phjis'iolofjicai Optics states the law thus : " After looking at a colonr .1 of moderate intensity then look at another B. If the after-impression is not sufficient to produce a positive image it produces a negative imao-e of A on B. The parts of li which are in the saiiu' pi, .) as A are dimmed. If A and B are of the same tone then B is rendered more whitish, if tliey are complementary /I's suouration increases. If B is between A and its complementary it passes for a neigidjourino- tone which is further from ^4 and nearer its complementary. Otherwise B becomes dark as much as A becomes more bright." It was from this kind of contrast that we sought to free our experiments as far as possible. Simultaneous optical contrast embraces all, and more than all the phenomena of contrast whose connection with the space threshold of colour I seek to trace in this paper. It includes all such phenomena as the modiheation in apparent size, colour, etc., which one visible surface exerts upon our visual impression of another surface at the moment of that impression. There iway be several kinds of modifying influence by which to classify the facts of simultaneous contrast. It may consist in (a) modification of the apparent size of the surface (exten>^ion contrast), {h) modification of the apparent ])rightness of the surface (brightness or intensity con- trast), (c) modification of the saturation in the case of a coloured surface (saturation contrast), {d) modification of the colour tone of the surface (colour contrast), or {c) modiHcation in the emotional tone accompanying the perception of the surfaces.* In our usual ex|)erience these kinds of contrnst appear for the most part in com- bination, but they are for all that quite distinct, and, in scientific abstraction, sepaiable features of ordinary contrast effects. The only one likely to be obscure is saturation contrast, to which alone therefore I will briefly refer. By saturation as distinct from colour tone and brightness is mear; "■ the degree of colour quality as com- pared with the absence of the same. By colour tone, on the other hand, is meant the tlegree of transition in a closed manifoldness of * For a discriinination of the various kinds of siinultiineous opticul contrast see American Journal of Psychology, Vol. iv., No. 4: "Some EiFects of Con- trast," by Dr. A. Kirschmann, and the same author's inaugural disserbiticm, " Ueber die quantitativen Verhjiltnisse des simulbinen Helligkeits- und Farben- Contrastes," in Philosophische Studien, Bd. vi., p. 417. colour wherein the tone is deterniined.not by relation to the absence (if colour quality, but by relation to other colour qualities of the same manifoldness. For example, green, red, orange, etc., are trans- itions in a closed series of colour which begins and ends in colour, but lietvveen each of these, taken separately, and the disappearance of colour altogether there is a series of transitions whicli rise from the zero point of colour (quality to the highest possible degree of it. These latter transitions ate called degrees of saturation of the parti- cular colour, whether it be green, red, orange, or any one of the infinite possible transitions in the closed colour manifold. In other words if we represent the transitions of colour tones graphically as a circle we must convert our two-dimensional circle into a geometri- cal figure of three dimensions in order to make it also represent the manifold of possible saturations. That saturation, and therefore saturation contrast, are not mere matters of speculative formula, which can practically lie neglected in considering the possible con- tingent influences to be eliminated in an exact examination of colour contrast, is .shown by the fact that it is possible to vary the satura- tion of coloured surfaces without changing the light intensity or colour tone. Dr. Kirschmann has succeeded in doing so by means of colour disci^. For a description of the means employed I refer to his article " Colour Saturation and its Quantitr»tive Relations " (American Journal of Psychology, Vol. vii., No. ti), in which he has given a preliminary account of the subject. For my purposes it suflSces to point out that saturation contrast, which means contrast between ditierent degrees of saturation, all of ecjual light intensity and the same colour, is a possible coTitingency which had to be taken account of in the prosecution of our contrast study. Since in any ordinary case of colour contrast there are these different influences at work, it was necessary to disentangle them as far as possible and to confine our investigation to one alone. Accord- iagly in our inquiiy we sought to limit ourselves mainly, if not absolutely, to colour contrast, by which is meant contrast of colour tone, not in the ordinary careless rse of the term but in its strict significance as distinguished from saturation and intensity. Our problem then in its liearing on contrast phenomena is maiidy to examine the relation which holds between colour contrast and the magnitude of coloured surface in the initial stages of the perception of a coloured object. By magnitude is meant not the absolute size of the surface but the visual size, if we may so speak, that is, the size as relative to the perc(>iving eye, which the absohite size oi tlie surface and its distaiico from the eye conjointly deterniino. The visual size of an object is measured by the' angle whicn that object subtends at the centre of the pupil of the eye, an angle which i» named the visual angle and which for any object of constant magni- tude varies with the distance from the eye. Of course for eviny object at any distance there are as many visual angles as it ha.-; dif- ferent diagonals and diameters. If the visual angle (selecting for the purpose either a diagonal or a (^anieter angle) at which a col- oured surface is Hrst distinguishable in its proper colour tone may be called the spatial threshold of that colour, we define our present problem as an inquiry into the spatial threshold of colours, especially when those colours are subjected to the contrast influence of other colours. Incidental to the examination of this problem will come naturally also an investigation into the influence of colour contrast upon two other possible spatial throholds, viz., (1) the threshold at which a coloured surface appears first as light, and (2) the threshold at which such a surface apjjears coloured though not in its proper colour tone. These will form subordinate problems whose elucida- tion will necessarily take place concurrently with that of the main problem. For it has long been well known that a colour, whether under the contrast influence of other colours or not, does not retain its characteristic colour (piality at very small visual angles, nor does it, by the gradual enlargement of the visual angle from zero, emerge at first upon the vision fully formed in its proper tone. On the contrary almost all colours lose their special colour quality at determinable small visual angles and appear either only as light or as some other colour. A very familiar example is the greyish appearance of a distant autumnal hillside, clad in leaves which are beginning to change their tints At a distance of three or four miles it is not only impossible to distinguish the colour qualities of th° variously tinted surfaces, which on closer inspection are seen to bt massed into a mosaic of peculiar beauty, but their joint eflfect at a distance is fi-e.[uently that of a characterless russet grey. The reason is that each of the colour surfaces subtends so small a visual angle that its stimulus is below the threshold of perceptibility for that colour : and while at nearer view they would make good to the spectator their many differences of colouring, yet in the greater distance the purple, orange, yellow, red, green, brown, with their many shades and hues to be found in such a landscape, are indis- tinguishably blended in one poorly .saturated reddish colour, or even reduced to a plain colourless ^rrey. Anoth.-r illustration of the same tPct 18 our nial.ility to distinguish the violet colour of a pansy at a distance which is not great enough to render the object itself quite invisible. On withdrawal of the .spectator to .sufficient distance the flower becomes indistinguishable from a dark grey or black spot and at still greater distance disappears entirely. When we approach again it gradually becomes visible, first as a point of grey light and then as coloured though not yet violet. Still nearer" the violet character of its colouring is distinctly perceptible. From these familiar examples the fact is illustrated that most coloured objects have three different stages of perceptibility wh^ch are dependent upon their spatial relations to the eye of the percipient. The first is when they are merely visible but without colour at all ; the second when they are seen to be po.s.sessed of colour though not the proper colour ; an.l the third when they become visible in their cor- rect colour quality. The exact delimitation so far as possil)le of these three thresholds of colourless, chromatic and characteristically coloured light for all the variety of colour presented by the .spectrum iH the aim of the present paper. The problem with which we started was to determine the last of the above, but in dealing with it the question of the two former naturally arose. In the percep- tion of a coloured object the lower thresholds must be passed through before the third is reached, and therefore they also properly fall within the .scope of this paper and were investigated at the same time as the main problem. I do not mean to imply that every colour will have three thresholds absolutely distinct from one another without the po.ssi- bility of a coincidence of any two of them or of all three. There are many facts apart from those which have made their appearance in the course of our experimental treatment of the problem which would bear strong evidence against such an assumption. It has per- haps frequently been noticed that a red light, for in,stance the port light of a ve.ssel at sea or the Janger light of a train, remains visible as red almost or quite as long as it is visible at all, which goes to indicate that its achromatic, chromatic and characteristic colour thresholds as distinguished above are not separated but coincident. Our experiments have corroborated such observations, showing in a more exact and quantitative way that red is a colour which does not easily or at smallest surface sizes lose its colour quality, but is for 6 the most part distinguishable as red just as lon^,' as it can be dis- cerned at all in the diminution of its extension. For most col- ours, however, it will be found that the three stajres of disceniment are distinct, and consequently will recjuire separate examination. ■ II. HISTORY OF THE I'ROliLEM. The phenomenon of a space threshold of colour was noticed in the early history of physiological and psychological science although it did not receive careful investigarion. It was noted and passed by withouf any thorough inquiry into the details or quantitative relations of the pro olem. Such cursory treatment practically amounted to nothing -.uore than calling attention to the existence of a subject that demanded .scientific investigation. Within recent years, however, some attempts at explanation have been made which yet, as we shall see, have not completely satisfied the conditions of such an inquiry. As early as 1823, the dependence of perception of colour upon the visual angle of the coloured surface was noticed by Purkinje.* He called attention to the fact that both intensity and visual angle play an important rfile in the perception of colour •' Sensibilitas oculi in specificam colons cuiusdam qualitatem ad diversas dis- tantias et sub certis gradibus luminis examinari poterit, nam notum est qualitatem illam colorum in objectis affjitim minutis ad justas distantias evanescere."f Plateau was the fir.^t, perhaps, who noticed and recorded the fact that colours disappear as colour at very small visual angles of the coloured surface.: However, he made no further use of his observation, and did not attempt any detailed inquiry into the subject. In fact none of the early investigators sought to deter- mme with any exactitude the nature of the relationship which they observed between the visual size of a coloured object and its per- ceptibdity as colour, and, so far as I can find, no fruitful advantage was taken of the fact which these early scientists had noted. Not untd comparatively recent times, by Von Wittich and Aubert.was any attempt niade to detenaine quantitatively the relation of colour perception to the size d the coloured surface. * Commeutatio de examine orcfani visus, etcT Bresl!^^823. (P 15 ) t Quoted by Auhert in his Phymoloyische Optik. t Poggendorff s Annalen, Bd. 20, 1830. (P. 327.) Von Wittich records in the Konigsberg n.e.Jical Year Book some mtere,st.ng oxperuner.t. made to ascertain what were the snmirs at all and m any form, or as of tlieir proper quah-ty. The results obtained were twofold, the objects being viewed mom'entarily on the one hand and on the other continuously for sufficient tin. to rende possible a secure, udgment. The following tables give his reTul winch will subsequently be used for con,paHson. Black Ground Colour Momentary view Visible Coloured Continuous view Red Orange . Yellow oranae i Yellow ........'.'.'.'.'.]',[ } Pure green i Dark tjreen o Pure blue i ^ark blue ... . ,, Rose . . . f Violet ....'.'.'.'.'.'.'.'.[".'.'.[[[] I " ' n 23 1 58 4 1 32 9 1 32 23 1 23 23 1 43 17 6 53 14 2 17 17 7 38 1 14 2 17 43 1 6 53 i Vi.sible Coloured 1 4 1 4 1 4 1 4 1 14 1 \\ 1 4 1 4 1 4 1 43 2 17 3 16 1 14 1 43 1 58 3 26 1 3 32 26 White Ground. Colour Red Orange Yellow orange ... Yellow '.'.'.'.'.'.'.'.'.'.'.'.'.'." Pure green ....!...! 2 Dark green Pure blue Dark blue Rose Violet ' ' " ' As Plateau had previously seen, when he said that coloured objects at small visual an<;les appear as a scarcely perceptible cloud, Von Witticli noticed that colours at very small visual angles lose their characteristic tone and appear either as ([uite colourless li^visible at r .8' and coloured at ,' 26". There a^ other in 'n - plc^e exceptions such as violet which on a white ground is ..en un oloured earlier than on a ,Iark groun.I, although Uie dl^.c. unient of as coloured comes at a very much greater visual angle fo white ground than for dark. The probable explanation o? the-' exceptions is to be found, it seems to me. in the fact that the black which Von Wittich employed was black cardboard, by no mem' good .specimen of non-reflecting surface. As compLd with the black of a good velvet, black cardboard is grey. This ...es to show that surfaces which we call black are o.dyrelativei; black and coiisequent y that a haphazard choice of them for employn.ent as black grounds in experimental work might easily lead to a con- dition where the ground actually reflected .scarcely less light than some of the colour pigments. It is likely that smnothi.:. of the kind has been the case in Von Wittich's experiments. The violet 10 and blue pifjinents wliicli he used were pigments whicli relatively to the other colours reflected very little light and conseciuently were much more akin to black than the rest. The contrast of brightness intensities between ground and pigment in the case of these two would be much less when the ground was black cardboard than when it was white, whereas with the brighter colours the exact opposite would i)roi)iil)ly occur. We should therefore naturally expect that violet and blue would prove themselves at variance with the other pigments in regard to their colour thresholds on white and black grounds, that they would in fact have lower thresholds on tlie while ground than on the black, while the others had lower thresholds on the black ground. On the other hand, had the black ground used been not merely a comparative black, such as card- board or even velvet, but actually a surface reflecting no light what- ever, then it is possible that the blue and violet would have followed the same rule as the other pigments. Aubert, in the work A'hich he has done, so far as it bears upon the problem of the space threshold of colours has largely followed in the footsteps of Von Wittich. He has indeed reflned somewhat upon the means employed but in the total outcome he has made vi'ry little advance. He employed coloured squares of two milli- metres in diameter placed equidistantly from the spectator in diffuse but clear daylight. The observer withdrew to a distance at which the squares could no more be seen. He then gradually approached to such positions as would respectively enable him just to see the objects as uncoloured, advanced further until he could see them as coloured, whether properly or not, and still furthei- until he .saw tliem in their proper tones of colour.* By measuring the distances of the observer's .several successive positions he had the material for a calcidation of the visual angles which coloured square sur- fncos nmst subtend in order to be, Hrst, seen at all, secondly, seen coloured and thirdly, seen coloured correctly. In order to rest the vision Aubert also used in his experiments darkened tubes for the eyes with a mask to close off the diffuse daylight which illuminated tlie colour pigments.+ The results which Aubert attained were considerably different from those of Von Wittich. This difference he ascribes to three *Abliiimllun<,'en dor sehlesLschen Gesellschaft (Breslfvu, 18(51), and Physiolo- gisclio Optik. tAiibert, Physiulogie der Netzhaut, p. 15. , - - , , ^ .. 11 2 n cau es. (1 ) the place of the illumination, (2) the pigments used and 3) the subjective uncertainty concerning the extreme limits which can be set to perceptibility. Ho found* that orange appears coloure.1 from the first, at 39" as red, at 59" as orange; red on a black ground at 59" is seen as red, while on a white ground at 59" It IS dark, almost black, and becomes coloured at 1' 43"; ultramarine blue on a black ground at 1' 14" is grey, at 4' 17" blue, while on a white ground at 1 ' 7" yellow at 41". " ' From the above figures it will be plainly seen that the discrep- ancies which Aubert himself noticed between his own and Von Wittich's results and between his own at different times are not inconsiderable, so far at least as the absolute numerical values of the visual angles are concerned. Whether the relative values among the various colours show any similarity or regularity as between the results of the two investigators we are unable to pronounce upon, because the colour list which Aubert used and that of Von Wittich have few points in common. Had they both employed the same set of colours it might have been shown that although there were considerable discrepancies in the actual figures given by the two for any one colour, yet the mutual relationships among the various colours were fairly constant. Aubert has made one material advance upon Von Wittich's work in the investigation of this problem. He has contributed to the more definite distinction of the mere chromatic from the characteristic space threshold. Von Wittich combined the two in the records of his observations although he did not fail to notice the existence of the distinction. Aubert not only definitely enunci- ates the distincMon, but also gives it a (ju antitative expression and *Aubert, Physiologische Optik, p. 537. ~ 12 notes the visual angles of the colours quite as carefully when they are first seen coloured, though incorrectly coloured, as when they are seen in their proper tone. The v.-ork which Von Wittich and Aubert have done upon this problem of the space-thresholds of colour perceptibility cannot but be seen to be incomplete in method and results. Its • deficiencies, howe\ or, are largely those which attend pioneer efforts in any line of work. For them the (luestion nrose as a curious fact of no small interest in the midst of other problems more immediately engross- ing, with the inevitable consequence that the investigation was not completely carried out in all its various and obvious divisions. The subject which thay dealt with in a somewhat incomplete way is capai)le of considerable extension. They reoooiii/ed in some degree the influence which contrasting surfaces have upon the per- ceptibility of colours. Thus they took into account the difference of white and black surfaces in this regard, examining their colours when subject to the contrasting influence of black and white grounds. But although it is quite as frequent in every-day experi- ence for colours to be seen upon backgrounds of other colours as upon the colourless grounds of white and black, yet it never seems tt) have occurred to these early investigators to examine the per- ceptibility of colours under the ordinary influences of colour con- trast. It is not improbable that colour contrast would introduce quite as considerable variations into their results as did the contrast with white and black. And indeed our own experiments have confirmed this prima facie probability as will be seen subsecjuently in the exposition of our results. Apart from the omission of Von Wittich and Aubert to extend their investigation to the full .imits of the problem we cannot fail to notice the unsatisfactoriness of tlieir method oven within the cir- cumscribed sphere of their in(iuiry. Uliat precautions did they ado])t to insure with fair reliability that the thresholds which they found for the perception of colour were really assignable to the colour perception per se rnther than to e(.lour perception aided or retardet. by some other accessory conditions ? For example, can we be sure that the threshold of 39", at which Aubert finds that orange is seen as red, is really the chromatic threshold of orange, unaffected by contrast of light ? It appears to be quite probable that this result is in part due to the influence of the contrasting intensities of the colour pigment used and its surroundings. It is 13 nd ed true that these mvestigators do not attempt to maintain that their results indicate anything but the perceptibiHty of colours when seen upon certain specific grounds, namely black and white. Such results may have served their purpose and observations on that basis may be of considerable practical value. But I must nnist that the inere fact of their performing their experiments under certain conditions of contrast i. far fro.u being a recognition or uiscovery of the exact part which cont.asf play, in the matter. Again It appears to me to be scientifically desirable to remove in some way ri possible the contribution which i,i tensity contrast brings to the results recorded, and to isolate the co editions that produce pure colour perception. Neither Von Wittich nor Aubert made any attempt to separate the factors that entered into the colour perceptions which they examined. I do not mean to stiggest that It would be possible to remove the factor of light intensity altogether from the field in order to get at the perceptL of colour pure and simple. This would be to eliminate colour itself because every colour must have a certain intensity as light But It 18 not neces.sary to do this in order to secure the separation of the two factors. What is necessary is to prevent inequality of light intensities and thus to remove the disturbing influence of light contrast. This would be effected if it were possible to equalize the intensities of ground and colour. It is very likely that neglect to eliminate the influence of light contrast between the pigment surfaces examined and the "rounds on which they were seen contributed largely to the disagreement wluch Aubert finds between his own results and those of Von Wit- tich and also to the discrepancies which occur iu his own observa- tions; not that they are altogether due or even mainly due to this cause, but such an influence must indisputably have had a -n-eat effect in modifying results, and in the nature of the case must have been brought into play in diflbrent degrees not only in Von Wittich's experimontH, as compared with Aubort's, but likewise in the various examinations of each pigment made by either of them. This is plain from the fact that scarcely imy two pigment surfaces have exactly the same light-reflecting power. And the same is true of the black and white surfaces. It can be photometrically shown that some surfaces which we call black reflect many times more light than others which we likewise call black. The terms black and white are in fact elastic in their significance, each including an 14 . indefinite iminber of surfaces of varyin-,' degiees of light-reflectiixr power.* Since such is the case it becomes plain that the (juan^ titative relation between ground and colour in respect of liglit intensity must vary not only according tc the blacks j.nd whitts used for grounds but also according to the .several colours examined ; and not only will this variation depend upon the essential differ- ences between the pigment colours, such as green, red, etc., in light- reflecting power, but it will exist no less concomitantly with "the diflierence in the pigment paper employed for each colour by two or more experimenters, or by any one of them at different times Accordingly, the observations of Von Wittich and Aubert may be expected to show greater discrepancies for some colours than for others, and for some tones of the same colour than for other tones In this manner the threshold differences among the various colours would be exaggerated. Moreover the natural divergencies for the same colours wiiich arise from varying individual sensibility among observers would be accentuated by the use of slightly different pigments, which would render them unduly and in indeterminable degrees greater or less than if exactly the same conditions of lioht- intensity-contrast obtaine.l. It is therefore apparent that from'this uncertain variable, arising from contrast of light intensities not being converted into a contrast by equal distribution to all colours a means was afforded for the entrance of considerable discrepancies into the results. *Dr. Kir,sc!.mann l.as^^^^^iTl^i^sThig experhiients which illuJ^^^Tuii; 8t.te,nent very wel . He records then in his article " Ei„ photonu-trisches Apparat zu psych.. physi.schen Zweclcen " publi.shed in vol v. of Wundfs Philo sophische 8tudien. He conipares photometrically various blacks with a standard white— I. Paris Black : !1) in lamplight (petroleum) -V-9 of the intensity of white (2) in gaslight sV-s of the intensity of white. (3) in diflFuso daylight 5V., of the intensity of white. II. Chma Ink : (1) lamplight jVo of white. (2) diffuse daylight j^., of white. III. Graphite (Faber BB): (1) lamplight i.„ of white. (2) diffuse daylight J. „ of white. IV. Graphite (Faber B): (1) lamplight i. 2 of white, (2) diffusrt daylight j|.„ of white. This goes to show the photometric variability ia what are ordin.riW accepted as good blacks-a range broadly speaking from ^ toTof tlJ wh te- chosen for standard of comparison. ffa ff "* wt wnite 15 From another standpoint their method of investiijation appears to be un.satisfactory and to reds of (omparatively poor saturation and liy Di". Kirschnmnn* which brings o>it tliis fact. It consisted in ende ivouriiifj to detect brown surfaces by loc>king tlirouf discrete points, as in the pl.eno.uen.m which Fick deHcribe.l* Not only will uicreased intensity .secure the perceptibility of the colour of a smaller coloured surface hut it will also increase that „f a s.naller nu.nher .,f discrete colour points, which according to Fick's observation cooperate with one another to produce a colour impres- sion althou-h they may sevrally be indistinguishable as coloured Von Kries has also noted that by decrease of the visual anoles coloured surfaces pass through transitions of tone until tinally at very small visual angles they lose all colour qualitv. He has traced the transitions for some colours. Red, he has noted, becomes colour- less, l)ut under circumstances that make it very ditKcult to observe the hmits of visibility usually coinciding with the lo.ss of its colour (|uality. This is quite in conformity with our own experimental results. Orange, he says, appears red before becoming colourless. With us this took place generally in regard to orange on a black ground, although there were cases where particular observers .saw orange a.s coloured jii,st as long as it was visible at all. Yellow .simply pas.ses into white. Green no. 1 becomes white without the intermediate stage of yellow. Green no. 2 at the smallest visual angle appears greener (with some blue) and goes into white by way of yellow-green. Blue becomes colourless without change in tone. Violet becomes reddish at a small angle. These changes of course are for direct vision ; for indirect vision the transitions are in some cases quite different. Charpentierf has done .some work which, though not bearing directly up(Mi this problem, is suggestive. He has in the article referred to examined into the relationship of light intensity of coloured surface to the size of surface necessary for the perception of the surface as coloured. He sought to free his colours from the adulteration of ordinary pigments and to render them as nearly as possible spectrally pure. For example, he obtained his blue by the interposition of C(jbalt-coloured glass and glass coloured with oxide of copper. The first lets pass only the blue and the red rays, inter- cepting the green, while the second lets pass only the blue and the green, intercepting the red rays ; the combined effect was to produce a blue resultant of tolerable purity, He had difficulty in producing ti^pectrally pure yellow (a difficulty which we also encountered in ■* Pfliiger's Archiv, Bd. vii, p. 152. t "Sur la qufuitite de hnniero necessaive pour percevoir la couleur d' objets de diffurentes surfaces." Comptu.s Kendus, 1881, ire Seinestre, p. 92. ... 20 our work), and Hiially luloptod an .standanl yellow a coiulnnation of all the rays which pivo a preibminant yellow element. This endeavour of Cliarpentier'a to obtain pure coloui-s was an advance in the treatment of this and kindred problems beyond the unscruti- nizing haphazard adoption of j)i<^in<'iits whieli had characterized some of the early work. The principle urcc o , lununafon for one o^ th. surfaces, is in.praeticaMe if use uayhght. Two sources of illun.ination are therefore necessary ana <,ne or both ot these n.ust be n.oveable to enable us to rule out all intensity contrast that is irrelevant to the problem and a .listurbin. influence. Ihese conditions, we thou^d.t. could be best fulHlled b? carrying on the who e investigation m a room fron. which dayli.hl was wholly exclude.l, and where our sources of illununation were incandescent electric lamps of measurable illu.ninating power and aiTanged m a certain n.anner to b,> presently described. (SeeFi- 1 ) I he central part of the apparatus upon which the two set" of contrasting colours were set up consisted of a blackened table with two parallel vertical planes, also black, standing upon it at either end and about two feet apart. On the further plane was fastened a revH.lving disc covered in sectors by the dittarent pigment papers whicb were used as the foundation for producing, with various com- binations, the several approximately pure .spectral colours that we wished to examine. The nearer plane had an opening cut in it which was covered by a brass diaphragm (Fig. 2) with an adjustable ,^,uare opening. The coloured light from the pigment papers on the further plane wa.s admitted thro„gh the opening in varying quantities according to its size. The diaphragm consisted of two brass plates laid together, each having a sfpiare opening of two and a half inches diagonal, cut so that the diagonals were horizontal and vertical I he front plate was made to slide in metal grooves attached to the second p ate. which was itself permanently fixed to the standi.u. plane. A stationary .screw, attached to the permanent plate and pa.ssing through a cylindrical nut on the sliding plate, furnished the means of moving the latter and of thereby controlling the size of the aperture from the zero point, where the two square openin-s .[ust ceased to overlap, through the length of one diagonal, until they became completely coterminous at the maximum opening The screw was manipulated by a crank handle and was so constructed that every complete turn moved the sliding plate just one-twentieth ot an inch, or in other words increased the length of the diagonal ot the square opening by one-twentieth of an inch. Each rcTolu- 22 tion of the screw, or one-twentieth of an inch increase in the diagonal, was indicated npon a graduated scale by a brass finder attached to the sliding (,hie. furthermore, to the handle end of the screw was fastened a brasr disc of about two inches in diameter, marked oft' at the circumference into fifty equal parts. As this disc revolved with the screw a peripl^eral movement through the length of one degree of its circumference would indicate thaf. the sliding plate had moved one-fiftieth of orie-twentieth, or the one-ti.ousandth part of an inch. Starting from the zero point of opening we could thus produce in the diaphragm a square aperture, of which the diagonal measured one-thousandth of an inch. At any stage, besides the integral number of revolutions of the screw, which might be found from the index h"nger and its graduated scale on the groove, we could read the extra fractional part (in fiftieths of a revolution) from another indicator fixed above the revolving brass disc at a point corresponding to the zero point of its scale, Ihat is, at the point where a whole revolution of the disc, and hence of the screw also, was just completed. We had in this simple arrangement a means of varying by gradual transitions the size without the form of the opening through which che coloured light under examination came, from zero up to a square opening of two and a half inches diagonal. Moreover the gradu»' increase or decrease was measur- able with accui-acy to the one-thousandth part of an inch, or to the two-thousandth or even the four-thousandth part, if we estimated halves anil (juarters of the degrees on the brass disc, which had con- siderable magnitude, in addition to this extreme accuracy of measurement it will be noticed that the apparatus artbrded a simple and speedy method of ascertaining the diagonal size of the opening, for a mere glance at the two scales was all that was necessary to read the registration in terms of revolutions of the screw. On the diaphragm were placed the ground or inducing colours on which we examined the colour tlireshold under contrast hifluence In .some experiments this diaphragm surface was transformed into a colourless ground, white, black or grey, by means which will be nuhcated further on. It may b.^ mentioned that the two brass plates of the diaphragm were not laid exactly together, but that there was left a very thin interspace so as to permit of the insertion of pi-mient papers. The object of this was that the inner plate of the diaphragm might be given the same colour as the outer plate and thus a con- tinuous coloured surface be presented throughout the movement Fi(!. 2. 23 Here may be mentioned a nicety of construction which was of material importance to the reliability of the results. It became apparent that in putting the coloured paper on the diaphragm we must avoid exposing any ragged or white edges of paper at the margin of the square openings, lest the clearly defined edges of the cut on the outer plate should indicate too conspicuously the exact position of the aperture al)out to appear, and thus by leading to an anticipation of its appearance vitiate the results. To preclude this chance of error the pigment paper was first pasted on the brass plates> and then by two continuous cuts with a sharp knife, downwards only and slightly towards the brass edge all round each opening, the portion cover' tig the holes came away leaving no uneven edges to the paper and no trace of white. Through the observing tube in front it was now impossible to distinguish lines where the surface of the upper plate ceased and that of the under plate seen through the square hole of the former began, but the whole appeared to be one con- tinuous and uniform coloured surface. So complete was the dis- appearance of the lines, that very frequently, in fact generally, it was necessary to point out to the observer the neighbourhood on the coloured surface where the point of light would appear, lest he fall into the opposite error of not seeing the emergence of light when it was already visible, from his attention being directed to another portion of the field of vision. It happened, moreover, not seldom, that after being once seen the point of light would disappear again, in consequence of a slight deviation of the eye's fixation, and it would be only necessary to point out the neighbourhood again for the light to be seen. So much help to the obsei'ver was absolutely required to prevent his attention from wandering to all parts of the field of vision. But this could not endanger the results in the same way as the careless exposure of lines intersecting at the exact point where the spot of light was to be expected. It involves all the ditt'erence between giving the region in which the thing pyppcted mnst lii> sought and locating the very point with mathe- matical exactness. We had now our two contrasting surfaces, the inducing surface on the diaphragm structure, the induced seen through the opening of the diaphragm and variable in colour by revolution of the colour disc on the further plane. It has been indicated above that each of these surfaces was illuminated aiiificially and from separate sources. I will now explain the means by which this was done in such a 24 manner as to rule out tlie disturbance of intensity contrast. The source of light in each case was an incandescent electric lamp of thirty-two candle power placed in a rectangular elongated box which was blackened inside and out, about five feet in length, with a square opening of eight inches to the side. The top was so con- structed as to leave a slit along the full length of each box wide enough to contain the neck of a lamp in a wooden sliding frame- In this way the lamp could be moved through the entire fength of the box. The open end of each box was directed towards the pigment surface, either the inducing or the induced, so as to per- mit the light from the lamp to fall obliquely upon it. The opening on top through which the lamp slid was covered by a lid hinged to the large box, for the purpose of preventing unneces.sary escape of light. Since either or both the lights could be moved we had a very easy means of ruling out differences of light intensity in the two contrasting surfaces. All that whs necessary was to alter the positions of the lights gradually until with a tolerably open diaphragm the light intensity of the two surfaces appeared equal. The point of equality was approached from both sides, first from above the equal point, where one surface was decidedly brightest, and then again from below, where the same surface appeared evidently less bright, so as to arrive as nearly as possible at the evact mean. At this stage a diflSculty presented itself, which it was of great importance to overcome, namely, how to get pure colours to operate upon. I have already pointed out that the work ot Aubert and Von Wittich was gravely defective upon this score and that ordinary pigments are by no means spectrally pure CO ours, but are the product in all cases of a mixture of spectral colour tones, and may even, as I have indicated, contain no rays of the kind after which they are nameil. Moreover our incandescent lamp light is always somewhat yellowish and m being used to illuminate the pigments mu.st cause a decided adulteration of colour tones even if the pigment colours were otherwise spectrally pure. Hence it was incun.bent upon us to procure colours for investigation as free as possible from admixture with other colour tones. ]n order to meet this difficulty we decided to illum^inateour pigment surfaces only by light transmitted through coloure^d media such as coloured glasses and gelatine films. Such media have the convenient property of absorbing some of the rays falling upon them and of allowing others to pass through Of 25 course ditferent media will absorb and transtnit diffei-ently and in varying degrees. Some will only weaken instead of totally absorb- ing the elements they interfere with and will conse(|uently transmit the entire spectrum with weakened elements in some part* In such a case an effective remedy is to increase the number of the interposed media until the disturbing elements are completely eliminated or at least eliminated to such a degree that the residue transmitted makes no material difference in the result. Moreover when a film or glass has the power of absorbing * part of the blue rays, and it is placed before a coloured surface which reflects -' part of the blue rays which fall upon it, the joint effect will be that only ~ of y of the blue rays is reflected. If, for example, we have a blue pigment which reflects 10% of the red, 8% of the orange, 5% of the yellow, 10% of the yellow-green, 25% of the green, 98% of the blue and 40% of the violet rays, and we interpose before it a com- bination of blue gelatine and glass which transmits 6% of the red rays falling on it, 5% of the orange, 2% of the yellow, 6% of the yellow-green, 10% of the green, 100% of the blue and 20% of the violet, we get as a final colour result 11% of the red rays, iZ of the orange, iV/o of the yellow, >;% of tlie yellow-green, 2i% of the green, 98% of the blue and 8% of the violet, which indicates a very fair removal of the colour elements that obscured the spectral purity of the blue. In this way by various comlnnations of films and coloured glasses, using different combinations for the different colours desired, we were able closely to approximate to pure colour for the purposes of our experiments— a feature which was absent in the work of Aubert and Von Wittich. For each colour separately we found by actual trial, the colour combinations, which under spectroscopic examination seemed to produce a colour most free from foreign elements. And although in no case did we get an absolutely pure colour, yet we succeeded so approximately that the colours which resulted may fairly be said to have been as good specimens of the various spectral colours as could be produced. At least it is safe to say that tin; errors whicli must have arisen from the use of pigment colours simply and without any correction were immensely reduced by this device and that our results are to that extent the more i-eliable. In the accompanying table (pp. 27 and 28) are shown the various combinations by which the colours employed for our *Kirsclinifinn, Uebor die Herstellung monochromatisches Lichtes. 2G experiments were i)r.)(lncc.l according to the n.ethod indicated an.l also the synopsia of their .spectroscopic analysis. There remained a further consideration. The ob.server must be at some distance away from the coloured surfaces to be inspecte.l, and that distance, as 1 have hinted in tlie beginning of the paper, must be constant : otherwise errors will creep in from the fact that the visual size of a surface depends not merely upon its absolute magnitude but apon its di.stance from the centre of the pupil of the eye as well. How then eouM we secure immunity from the danger of a disturbance caused by the dim light of the room being inter- posed between the observer's eye and the illuminated pigment surfaces ? It is impos.sible to so darken a room that there will not remain some smal' reflected light capable ol introducing a disturb- ance into the observations. It would be undesirable moreover to have the room absolutely dark, because in that case a fre.sh source of error would arise, from the fact that in total darkness the eye is incapable of holding its fixation for any length of time but moves unconsciously within a range of ninety degrees. This would be especially mischievous in experiments on the non-contrast light threshold of colours, where the front colour surface is not illuminated: since the observer might be looking completely in the wrong direc- tion and not discover the emergence of the small point of light until long after the threshold mark. To overcome this .2 >— s c P S 0) ".JJ s| 2 C8 . X «S ; ■^ 10 n is -M 1 3 'Ji ^ ■« s M ^ -^ a. a. 50 1 3 if t i 4. "5 ^— s ''-"S. ^-s • tJ. ^1 s a. a. a. 3 a. a. a. a. 'w i» M lA -4^ hi i '•3 » .0 fit c '0- ^'S in (N © isible I of the pectrui 1 i 2 = •M -4- ^1 a.-" a.o in Tf Si s > " •"-^ *- g in t 'S Tf .n^ in^ =• ^ 3- '■O in in. ^^ ^-^ ^-' -a t <" i 3-^3 *.S J ^ is is J c 2 s S '©'«« a "a, ^ ^ c _c "X to 7i •" c c .a Q 60 3 ;. "a "a >■" 2 5*3 tf Ou u. > > 4-1 c5 g S "i '■f t*-r .^ ^ c ^H OJ to 2 "a > "a 1 2 1 >. SO »^ -4.3 3 & D Si 6D C >t „ ^ B S S rf :4 •—.1 03 CO '0 e C S V c Q >,■ iM 1-1 iH i-( •ag .^§S 1 _3 -a S o- u p— 4 ^••3 2. I.S 03 t4 a) Sh g ZI4 -4J § a § ^ fe is a aj d 3 5 ,2 ^ ^ So = 3 3 £ "a ® "S £ s s Pi >H >H >* e 3 h— 1 2 1— 1 CH tc "o -kil 4 s ^ fe i a> g 1— ^ _o S, s "o ^ ^x ^ •—1 1— ■S S u M k |1< >^ >H >H 1 1-^ N M ■*■ «£> d 1> 28 a. a. J5 a, a. 'SS si as O U o >J S3 O ^00 ^-o O 'o a.00 a^?;^ |-,n ;-• o cJ 3 .mS^ aSJ ^"^ 5 t; t- ■S -^ t: -i-i ic5 *^ M o S o «- s K O bo o is lO so c a s o O o o I o 3 to 3 1>- -d 2 2 «J .-30) |> « CL B 01 c o 0} 3 •^ '-i n pit I c o . St> 1) 4) u o o ^ A ^ OD 1 o SD >> w^ 1-^ 1— 1 >, >. O ?* u U o o f-l 04 2!) Hliutt(>r, which could ho usorl to sluit ott" all the lij^'ht from the ilhuiiiimted pigniontH iind thu.s to ;,'ive the ohHcrvcr's eye a chance to rest hetween observations. Li our experiments it was of course necessary to take tlie further precaution, after openinjr the shutter, of pausing' briefly to allow the eye to accustom itself again to the bright light and colour. Not to use this precaution would be to introduce the disturbance of successive brightness contrast, just as to neglect to close the shutter freciuently would admit the vitiating influence of retinal fatigue. The small end of the observation tube was kept stationary at the cho.sen distance from the brass diaphragm. But because of the fact that the opening in the brass diaphragm could be enlarged only on one side, it became necessary to make the larger end of tlie observing tube also moveable, in order to keep the brass diaphragm opening always in the centre of the held of vision as debmited by the diaphragm of the ob-servation tuV)e. It would not do, however, simply to make the end of the tube move at the same speed as the sliding diaphragm plate, for this would be too fast for the purpose, which was to keep the induced surface marked out by the brass diaphragm opening always in the centre of the held of the inducing or iiround surface. For it is evident that the centre of the increas- ing square aperture in the changing diaphragm moves exactly half as fast as the moving plate which effects its increase This we can see if we consider that at the zero point the centre coincides with the moving angular point of the plate, but that when the latter has moved any distance the centre is ' xactly half-way between the zero or starting-point of the moveable plate and its latest position. In order to secure a movement of the observation tube of exactly the desired speed we introduced into the appanitus a \ ory simple leverage device which accomplished the result automatically. We employed an arm twice as long as the distance of the diaphragm plate from the observation tube ; one end was fastened to the slid- ing plate of the brass diaphragm, while the other was pivoted to the table upon which tlie observation tube stood. From the middle of this arm an attachment by pivot was made with the end of the observation tube. Then as the plate was moved by the operator the arm turned round the pivot on the table, drawing from its middle point the end of the observation tube. This contrivance completely served the purpose of securing an automatic steady shiftintr of the tube, so that the colour transmitted through the 30 openiiit,' iti tlu' hiasH (liiiphni^in was kept prccisoly in the ccntiu of tho iiulucinu' ^'loimd as (Icliniitcd by the diapliraj^m of the observa- tion tube. Wo were tbii.s ciialded to escape from tbe irretjulaiitics of our preliminary nu'thod. wliieli was indtpciidcnt ninnipulatioii by tho operator. Tho iuran^rcnicnt worked in all respects .satisfactorily. 80 little did the extra weight of the tid)e ohstruci. the rotation of tho screw which moved the (liaphra;^Mii plate, that the difference in the ease of the movement was scarcely perceptible to the operator. There are a few other considerations which I must brielly dwell uix)n in order to <^ive a full ex])Iani\tioii of our experimental ecpiip- ment. We trade some experiments with a black anying tables, the general relationships which obtain among the achromatic thresholds on a black ground will be manifest. Tables I, II and IV contain, under section I, the results of the achromatic thresholds for three oUservers respectively, Dr. Kirschmann, Mr. Preston and 36 Mr. McCrtlluia. Tho tables representinj,' the observutions of Dr. Kirschniann and Mr. Preston were based upon the results obtained from several sets of observations, as many as six in the case of Mi'. Preston and three in that of Dr. Kirschniann, whereas Mr. McCal- lum's table represents only one series of observations. It is obvious, therefore, that where there are any serious discrepancies between Mr. McCallum's observations and the joint testimony of the other two, the presumption is that their sensibility is more accurately represented than is that of Mr. McCalluni. It will be noticed by reference to Tables I, II, III and IV that the colours which, when viewed on a black ground, have lowest achromatic space thresholds are blue and blue-green. The maximum point is attained by the red, while the colours intermediate between red and blue show a somewhat gradual transition from the highest threshold point to the lowest. The most abrupt break is at the yellow -green, where a sudden deviation upward occurs, more marked in the case of Dr. Kirschmann than in that of Mr. Preston. There is not complete agreement between the observations of Dr. Kirsch- mann and Mr. Preston in regard to violet, which for Dr. Kirschnmnn (in our marking) has an achromatic threshold considerably higher than purple, but lower than purple for Mr. Preston and also for Mr. McCallum. It will be noticed also that the achromatic threshold of grey no. 1 is lower than that of grey no. 2 for all observers. A peculiarity of red, as distinguished from the other colours when viewed on a black ground, is that it is always and by all observers seen from the very first as coloured. In other words, whereas the other colours can for the most part and with the majority of observers be seen as colourless points of light or simply as something different from the ground, at sufficiently small visual angles, red is conspicuous as having, at least on a black ground, no strictly achromatic threshold. We have, however, used the term achromatic threshold to embrace the visual angle at which a colour is first seen at all, and hence it may be applied to red as well as to the other colours which are first seen as cjlourless light. It will be found that all the observations recorded agree in this peculiar behaviour of red on the black ground. Some of the other colours are occasionally seen also coloured from the first, but these instances are probably due to accidental circumstances, since even the same observer will at other times see them colourless. In the case of orange, Mr. Preston sees it always coloured, but Dr. Kirschmann and M' 37 Mr. McCalhim botl\ see it first as colourless light in the majority of their observations, though in some cases, like Mr. Preston, they see it coloured from the first. It is a noteworthy fact that the region of highest achromatic space thresholds is the red end of the spectrum, and the region of lowest thresholds is that which embraces the other end of the ordi- nary spectrum, from blue-green to violet, with purple as a mean between the two. This position of purple coincides very appropri- ately with its known spectral relationship to red and violet, as a transition between these two ends of the ordinary spectrum. This is shown in the " inverted spectrum,"* where purple, which does not appear at all in the ordinary spectrum, is seen as the middle colour, while red and violet are its immediate neighbours on either side. It might at first seem rather peculiar that the lowest achro- matic thresholds shonld oe found in the blue, blue-green and violet region, as these colours are usually regarded by us as less bright than such colours as orange and red, which, according to our experi- ments, have a much higher achromatic mark. But we must not forget that the colours were reduced so far as possible to the same brightness, so that their comparative positions as regards threshold visibility do not rest upon a basis of brightness, but are due to some other cause inherent in the quality of the coloured light and in the nature of the sensibility to which it appeals. A graphic representa- tion of the relations of the achromatic threshold on black ground is triven in Curve I of Figs. 4, 5 and 6, the latter representing the results of an average of observers. (2) Grey Ground.— If we consider the achromatic threshold results on a grey ground we notice in the first place that the thresholds are, as a whole, considerably higher than those obtained on a black ground. This was to have been expected from the great brightness contrast which was in play in the experiments with a black ground. The sole object, indeed, of performing the experi- ments on the grey ground was to secure as far as possible an im- munity from this contrnst in brightness between the colour and the ground. Hence, if our measures were at all effective in achieving the end for which they were devised, we should look for an increase in the threshold sizes. *Colour Saturation and Its Quantitative Relations, by A. Kirschmann. (American Journal of Pfiychology, Vol. vii., No. 3, page 389.) - - -^ 38 Examininj; Tables V, VI and VII (Sec. I) we notice that the relationships of the achromatic thresholds anionff the colours are considerably altered from those which we found to obtain on the black ground. Instead of red being the region of highest achromatic space thresholds and blue that of the lowest, with the intermediate spectral colours forming a graduated series between, we (ind now a tendency to decrease among the red thresholds and to increase among the blue. In fact, so great is the change that red has the lowest threshold of all the colours on the grey ground, though not much lower than some colours of the blue-violet region. We notice, however, that in the region from orange-yellow to gi-een, inclusive, the same nmtual relations of the parts hold good as on the black ground. The only serious alterations occur in the threshords of the col- ours at the ends of the spectrum. What can be the explanation of this discrepancy between the results on a grey and those on a black grounds I have already pointed out that in both cases the colours were of equal intensity as light, so that their brightness in relation to the ffround should in both cases be uniform for all the colours. It cannot, therefore, be attributed to any change in relative brightness. The only difference in the two sets of circumstances was that by substituting a grey for a black ground we removed the inequality of brightness intensities between ground and colour, while retaining the uniform intensity of the latter. But all that we should expect from such a change would be a general elevation of the achromatic thresholds, without disturbance of their mutual relations. We know that the change to the grey ground from the black would modify the brightness intensity of the colours relatively to a percipient, though not absolutely. For if the existence of contrast of brightness means anything it means that the surface observed by the percipient ap- pears to possess a greater intensity than it otherwise would have. Consequently to do away with light contrast would be equivalent to lowering the light intensity of the colours all round— so far at least as the achromatic space thresholds are concerned. But it does not appear that these considerations offer satisfactory explanation of the actual facts observed. It is conceivable that such a change might introduce some modifications into the system of achromatic threshold relations, since equal increases or decreases in the liglii intensities of colours are accompanied by unequal degrees of change in saturation, and the presumption is that the achromatic space 39 threshold of a colour is quite as much a function of the saturation as of either of the other variables of light phenomena, viz., bright- ness and colour tone. This supposition, however, even if admitted, does not account for the fact that when a grey instead of a black giound is employed the red region of the spectrum (including orange) is converted from the region of highest to that of lowest achromatic thresholds, while the thresholds of the blue-violet region are raised. Possibly the explanation of this phenomenon depends largely upon the fact that in spite of our care we were unable to exclude completely the colour element from our grey ground, while keeping it at an in- tensity equal to that of the colours under examination. The grey which we u.sed was, as has been said, slightly bluish, more conspicu- ously after exposure to the yellow light of the incandescent lamps, and this bluish tinge in the ground may account for the increase in the blue violet thresholds and the comparative lowering of those of red and orange. The assumption is here involved that the achro- matic space threshold of colours is dependent to some degree upon the contrast relations of colour qualities. Why blue should tend to influence red more than orange is a matter upon which light will be thrown by the exposition of the contrast relations of the actual characteristic and chromatic space thresholds. In the meantime it is sufficient to anticipate the results there reached, and say that we have found blue and red to influence each other's space thresholds more than blue and orange or red and green, the ordinary comple- mentaries. Accordingly, ray suggestion is that since the achromatic space threshold may be considered as simply the zero point of per- ceptible saturation, the same modes of behaviour may obtain here as in the higher degrees of visible saturation. This is not offered as an a priori deduction from the relationships of colour contrast, but merely as a plausible theory in explanation of the phenomenon ob- served ; its validity will be strengthened if it is confirmed in the case of achromatic space thresholds on coloured grounds. Before leaving the observations with a grey ground I wi.sh to mention a fact that is noted in our results. Red, which uniformly appeared coloured when first visible on a black ground, has usually on a grey ground a distinct achromatic threshold. The achromatic threshold is represented for some observers by the curve I in Figs. 7 and 8. (3) Red Ground. — On refei'ence to Tables VIII, IX, and X (Sec. I), which embody the i-esults of our observations of achromatic 40 thresholds on a red ground, it will be noticed that the highest thresh- olds are at the red end of the spectrum, being those of orange, orange- yellow, and purple, while the lowest are in the regions of green and, more especially, blue. The threshold of purple was unusually high, for purple on a red ground is not always seen as a point of light but requires to become of considerable size before it can be distin- guished from the red of the ground. Some of our observei-s, however, did see the purple as a spot of colourless light, and consequently their curves show a lower achromatic thi-eshold of purple than that of Dr. Kirschmann for either eye. The greys occupy a middle place on this ground, and grey no. 1 has a tendency to be somewhat higher than grey no. 2, a significant fact when we remember that grey no. 2 was slightly bluish in tone. It will be seen that the supposition by which we sought to explain the disturbance of achromatic space threshold relationships among the colours in changing from the black to the grey ground is, according to our anticipation, verified by the results with a red ground. We find that on a red ground the blue and the green have the lowest thresholds, the blue even lower than the green. (4) Blue Ground.— The results with a blue ground (Sec. I of Tables XVI, XVII, XVIII, XIX, and XX,) are similar to the fore- going. The regions of lowest achromatic space threshold are those of red and orange-yellow. The threshold of orange itself is uniformly somewhat higher than that of either red or orange-yello .v. The reo-ions of highest threshold are those of the colours nearest to blue on both sides of the spectrum, viz., green, blue-green, violet, purple, and yellow-green. The greys are closely related, as we might expect, grey no. 1 having a lower threshold than grey no. 2. The princi- pal fact to be noticed in these results is the lowness of the threshold of red which in two out of five of the tables is the lowest point and in the other three is but slightly above that of orange-yellow, the actual lowest for those observers. This would indicate the same tendency already observed in the case of the grey and the red grounds, which is that red and blue ai-e in close correspondence in reference to the facility with which each can be seen achromatically on a ground of the other. It is also rather a peculiar circumstance that orange itself is not influenced to the same degree by blue as either red or yellow no. 1 (orange-yellow) in the direction of low achromatic space threshold. It is natural to suppose that if the achromatic perceptibility of colours is dependent upon the colour of 41 the ground it will be determined by tlie ordinnry complementary colour relationships. But actual obscirvation does not bear out this ^ supposition, and thus shows us how precarious is all such f the colours on coloured grounds is represented in Figs. 9, 10, and 11 by the dotted curves. Chromatic ThreHholch. (1) Black Ground. — In our chromatic .space thresholds on a black ground as shown in Section II of the corresponding tables, it will be seen that the agreement among the different observers is not by any means complete, althougli with one exception (Mr. Preston) they are not seriously at variance Apart from Mr. Preston's testimony the results are substantially alike, from purple up to yellow-green inclusive. In the case of green there is considerable disagreement, but after passing green the results again agree fairly well up to blue. Mr. Preston's observations, however, contradict those of the others almost at every point from the red to the yellow-green, inclusive. \\"\i\\ so little unanimity among the other observers as to the chro- matic space thresliolds on black ground we can scarcely find sufficient data for a secure judgment. It is, therefore, difficult to speculate as to the cause of Mr. Preston's noticeable disagreement with his fellow- observers. But we may remark that the strongly emphasized dis- crepancy in the matter of orange is very probably due to an unusually keen sensibility for orange on the part of Mr. Preston 4:i His six observations on the cliroinatic tliresliold of ornnrre were exceptionally close to each other in threshold nia■■■■■■ — ___^ — — ^ *Philos. iStudien, vol. viii, p. 199. 66 Such a condition of tliinfi;^ was, of course, rather reumrkable, both because of the fact that the one eye was perfectly normal in its appreciation of colours and because the iuipaired sensibility of the colour-blind eye showed neither more nor less than normal appreciation for red and blue, while at the same time these two colours acted as con»plementaries. Such an tinorthodox behaviour on the part of this colour-blind eye was ijuite in contravention of the component theories of colour, and among the adherents of the component theories there was a disposition to deny the accuracy of the experiments,* to minimize the importance of the facts disclosed, or even to ignore them altogether. For it is plain that if white light is composed of three constituent colour elements, red, blue and green, as the Young-Helmholtz theory assumes, it is inexplicable how this calour-blind eye could see colourless light at all, having sensibility for only two of the three, red and blue. Again, if red is a sensation which arises from the destruction of a certain kind of visual substance, and blue a sensation arising from the con- struction of a totally different kind of visual substance, as Hering's theory propounds, it is inexplicable how this par- ticular colour-blind eye could continue long to have sensi bility for more than blue alone. For the blue and red of this abnormal eye are exactly the .same as those colours in the normal eye, as was shown by careful test. Hence the destruction of the red substance could never be made up by the (onstruction of the blue, since according to the theory tliey are totally different, with the inevitable result that the red substance must ultimately become exhausted, and, eo ipso, the appreciation of red he destroyed. There is no obvious way of adroitly escaping from the difficulty by sup- posing tlii.s a special case where the red and blue are the dissimila- tive and assimilative aspects of the .same visual substance, because this is to assume that red and blue for this eye are different from red and blue for the normal eye, which was experimentally shown not to be the case. My object in referring to Dr. Kir.schmaiui's paper is to call attention to the remarkable coincidence between the colour sensi- 'Piofessor El)ingliau.s showsthisattitude toward the matter in his article on p. 215, Band v, of the Zeitschrift fur Psyclwlogie. The importance of the above case of colour-bhndness can scarcely be set aside in the summary way that Pro- fessor Ebinghau.s is disposed to use. 1 he observations and experiments were not only carefully and accurately conducted, but Professor A. was an expert optici'^t, and thoroughly competent to judge the facts presented to him. 57 bility for MiU' untl n-d in this enlom-bliiid tlichromate ami our space threshold resnltH. I have frci|iu'ntly in this papor laid stre.sH on the fact disclosed in our results that coloured ((rounds do not aeeni to lower the space threshold of tlieir ordinary coniplement- aries so much as they do that of sonu( other colours, as red and blue. We have noticed that on red j^fi-ound almost all the colours appear first as blue, and on blu- nearly all appear first as redi including even the complemijntaries themselves of the coloured grounds. W(! have also noticed that in regard to characteristic space thresholds there appears to be the same disturbance of the ordinary complementary relationships of tlie colour system as in the other thresholds, and that the evidence for it is even more em- phatically unanimous. In our experiments for determining space thresholds red and blue seem to act exactly as complementaries might be expected to act. It thus appears, so far as our experiments go, that for small angular sizes of coloured surfaces there is a disturbance of ordinary complementary relations, and that for red and blue grounds at small visual angles a condition of things obtains which is some- what similai- to that present in the colour-blind eye of Professor A. The coincidence is not quite complete because there is not absolute failure to appreciate other colour qualities besides red and blue. I would enunciate our conclusions on this question in the following way : — On red and blue grounds, below the limits of the character- istic space thresholds of blue and red respectively, ih I'e is a lack of ability in the normal eye to make definite discriminations of the other spectral colour tones and a tendency to confuse them with either red or blue. Thus in a limited sphere embracing only small angular sizes are practically reproduced the conditions of colour sensibility exhibited by the colour-blind eye of Professor A., which form a colour system of one dimension founded on the two colours red and blue. Further, it is interesting to note that in this peculiar comple- mentary relation between blue and red we have come across what in our estimation is a formidable difficulty to Hering's colour theory, by which all colour phenomena are explained by a threefold antag- onism of fundamental colour processes. Black, red and yellow are by him set over against white, green and blue respectively, the former three being the outcome of the destruction of three distinct kinds of visual substance, while the corresponding latter three are 58 the outcome of the construction or aNsiniilation of the sjuno kinds of nervous substance respectively. A coiuplenientary rehition of colours is inherent in the nature of the optical organism and is the issue of tlie natiiral balanced activity of fatigue anci repair, destruc- tion of tissue and consequent recuperation. The different sets of complementaries belong to the activity of the different kinds of visual substance. Fi-om this theory it would follow that the red process set up in the optical apparatus should originate the green process in a contiguous surface relieved from the stimulation of the red. Applied in particular to our course of experiments we should expect that the red ground, when watched for the emergence at its centre of a small surface differently (|ualified, would tend to induce upon the emerging surface a green colour. The theory apparently demands so mucli by its physical resolution of the ])henomena of complementary relationships, rendering them rigid and inviolable. But the facts of the case as shown in our experiments are that the red does not induce green but blue. And it appears to us that this actual disturbance of the complementary relationships is a difficulty which Hering's theory is utterly unable to account for. It might be offered as a plausible explanation from Hering's point of view, and according to his terminology, that the matter is an aspect of irradiation or light induction or negative contrast, by which is meant that a coloured surface t(!nds not merely to induce upon a contiguous surface its complementary colour but also to spread its own peculiar colour qualit}' over it. But our rejection of such an explanation is uncjualitied. In the tirst place, if the red tends not only to induce green upon the surface revealed by the opening aperture of the diaphragm but also to irradiate a positive influence of red itself, we are no nearer an explanation than before, having on our hands a weakened saturation of induced green instead of the blue which we wished to explain. For it follows that if the negative induction a.ssumed is slight and not quantitatively eijual to the green induced, then as far as it goes it will neutralize the colour quality of some of the green induced, and produce colourless light. J'his, which is a strict deduction from the theory, would still leave a weak saturation of green and not blue, because of the admixture of the colourless light anil the residue of green. Again, if the negative induction is equal in quantity to the green induction, we should have no colour induction at all but simply an induction ot colourless light, which is as little like blue as ever. Finally., if .59 we admit the iiia o ^ ;?; ^ ,s o 3 '3 o . I— I h 4) O o Pi s o o hh'.H o u < [viio8«!P JO a'jSiU! pmsi^Y •Jiuiundo luSiuiidrnp jo JO ST11O0OT "! Siiuwdo niSif.nidinp jd tvuofitiiq •SuTitndo m8iMt[dwip jo IiiuoSinp JO eiJiin! i«uBr\ ■Sinundo luaiandiiji- jo aa^.Muvip JO o[Siu! iwitiiA •t[DUl llti JO siponOT "! Sinuodo mgiuiiihnp JO pmoHniQ •Suiuodi) ui"i.'.u[dvip JO (vuoTiinp JO Olefin! piusT \^ « n3 m -M § •S ^ [3 S ► o i^ (M rH 'M ^ ec w e<) . ■* o « m ■* X in CO i« « ;^ •^;occ05 e'5T-too-*iM_i.^o» ■ ci ?r ^ ?t ■*■ d rH •* o in o eg in- inaooifflOooo^M-jt-QO rH «q l>. (M (M o; 1-1 ^ f- CO J;r JO CO CC !■- »0 ao •* O t CO-*-* Mcom ■— -< d o) c> d -^ o rH CO G« 61 a o 03 ^ < »-J u PQ ^ 55 .. o 1 > 1 u l~>l 0) »--l ii « o bl s o o u , •n o ■s -5 S » n; ki -*-> ,£3 *3 2 cS ^ o o -a ! (M O (M ;C t- OS lO IS -H CO « rH 1-1 •Jiiiuiodo niSv.Ti|iIvip Jo .lo^ouiifip JO o[8m? l«ii8i \^ OS QO » M la ■* O •>! IS •* » aj •>! lo ■.-" -H •* in •* rt ■<* o t- Its o 00 1 1-- -H « CO >J •Suniodo mSiuqdtirp jo pnioSiiip JO r^(3u« t«nsi^\^ 00 th s-1 ri' -H CI 3i oi •■£ rxi i6 ts n o-i c-i co ■^ i— I -^ ■>! ^7 CO lO - CO (M O^ (M ! -H •Suiiiodo tUiiii.n[dtiip JO jo^ouiiijp JO o[S>ui! l'>(n8ij\^ ■([.nil in? JO snjoooi III "iiuiodo iuSn.ii[(h.'ip JO [liuoSiiiQ r-l 05 O l^ " I !>. — iffl oi 00 o ~' r-H i4 00 o" ■-I if5 CO rC ■* -^ t 0-1 Oil— l'-'T-(r-liHi-l^ rH lO -* -ii 1^ O O !>. •-J i.O ■* O "It IS iS CO •S?iiuiodi> iiiSi(,u[dvip JO jwuoSvip JO o[7fiii; pfiisi \ ■Siiiiiodo iit"(!.n(di,'ip JO .lajoiiivip JO o[j)iu.' [iinsi^\^ 50 M< •* -f CO 'M •w 1-1 (M 3 X t^ iS CO CO » IS iS — I -H l>. Tt tN-ti^COCO-^-J-H » i-i5n'^ti'''""^'-*'^ti>t:i"^:i"t l^ IS -"t <-< O CO 05 "* 5-) iS ■* ^5 CO CO O^ (M 01 X CO CD05 t>l CO iS (M CO 01 .o l~-rt Oi CO in IS ^ M IM ■to -* I - -H r-, (M •^ X X iS ■ in ■* X « .k» &^'9 esSSaajaaoaJoj sa>£;(i>ii3Jt4^r:^'-^(Mhi il, tf O >H >* >- O P3 PQ f> O O G2 Q O hD OS 03 r'^ u 4) W > (; < ■^ ^ "■^"^ m •*^ O a) 1 m 1 ^ ►-I O 1— H S 'O OJ W >»H^in«oitvooO'«toroo ^rtoc->*i>eo'^'woo«ocr'oo •Suixiado uiSujqdtfip jo aa^auiuip jo ai^uw [«nsi^ 1^-00 OS C5 Tt< «i o « » •*' eo ?o im' — ' i-i ■>* 1-i o o lO i>^ 00iH(NClS!OTl<'»tlOt-lr-lC0 •qoui n« JO sti:K)O0T "! Suiuorlo iu3BJi[(h)!p JO |«uo8t!T(j (N eO(N»-<'^i-Hi-(i-iC0SOcc •Suipado ui8i).iiid«ip jo j«U()3«ip JO ajSun lensij^^ •Suuiado uiSu:i|dvip jo .i3^9uiT!ip JO oiSuu l«nsi^ ■qDui ire JO smooox wi «"tiifdo luSuJijdflip JO itiuo8«i(j eo'cxooecit^.co «ooo cq-nJodioifflt-^oot^'-ioMM t^ in !>• r-( t- 50 -H W iH 05 00 <>J ■* Tf !0 CO CO im' ro c o ?c ici tc o 00 1> o 05 r-1 5^1 CO CO CN .«iTHicc".co •Suiuodo tiiSiundnip JO pjuoSiJip JO 0[SuB l«n8i;\^ I CO cc Tt X 00 Tt Tj" ^ lo Tji tH CO T*i CO CO CO CO CO - rt 1-1 C<1 »*^''*"0505 » Ui 05 CO ^' 00 05 00 00 ifliC-HktCMXTgQC ■ "itji. 1-/J c(«— H» 00'VS>lOOiM-«ti-( — (MOI'N rtco-^-'j'-^cococococococo 3 O O etts 3 « B D I— ( 60 CO 1—1 1—1 »— * ^ .« .01 .0) v bo OJ o a> « -^ !*> >» a, 3 s - - • bi .^ ^ O * aj pkp3o>^t«tHC!Q3mkoo 68 A IB, D o g A S O "eS o < o >A M i* S5 S O p 0) {^' > '--( 0) CO ,a -is o J3 s o I— I o O o CO a o •r: O c cc * »«500<-lt»-iO 30 C} « t>- »n 1ft 00 1>- 50 o (N ia'«)i»H-«jii-(i-ii-(ia •Suiuodo uiSuaiidtiip jo IBUoSntp JO 8i3uu l«n8t^v 00 00:3 ■«*' tH '*' 'ft O 'ft 50 ■^ rH CO l 1-i d •* *-o t£ d CO d CO (M -^ CO CO *J tC «rs •l[OUl u« JO smOOOT "! Suiuodo ui8«jq45^WiM c Ph C4 O >^ >H >H o PQ 64 Representation of the Spaoe-Turesholds for Black Ground. Sfal'J : 1 inch = 10 minutes. Curve 1— Achromatic Threshold, " II — Chromatic Threshold, " III— Characteristic Colour Threshold. R O oY Y YC, G bCi R V P Fia. 4 (Table I). R O oY Y YC C bC B V P Fio. 5 (.Table II). P O oY V YO C bG B V P Fig, 6 (Table III). 65 3 . — ^ .2 sc 55 O o iz; O c o O u > noi?ot--i-icc^xoio O'MC0->!f00t^-*• 1-1 Tjl lO CO •* o « oi -* O X c-i >-< 1-^ ad r-i-. OC lO t- t- 00 iH !>• lO IS « ^eo"-! 05(Noooo* »ft_ '^ 1>1 1>1 OJ C^ N^ '* !0' -H X ift CO TUTjtiMiOCOTttJ-li-H - eO ©1 IN CO (N H >- O M W >• 66 3 go S g o e-a o m t3 O EC O ft O n s u a , o (6 u O u > o o m a> o o n 2 a 2 •8uiu9do tu3wji{d«ip jo IBUoSflip JO 8i8u« iBnsijV •qoui ut! JO sqioOOT "! 8utuado mSBJiidBip JO i^uoSbiq •Suiuado luSuJud'utp jo ^wioSeip JO aiSuB l^nsi^. JO smoooT uiSunidwip ipui UB ui 3uiuado JO {wuoSwiQ •Suiuado ui3«.iurI«Tp jo l«uoS«ip JO afSuv \vn9\\ JO sq^oOOl "? iu3«Ji{d«ip JO •qoui UB Suiuado [UUOXWIQ lO ^ IS lO gs ■* t- 00 ■* *£r* OJ lO !>• 00 t^ 00 t-- 1- X X CO CO tc 50 00CC(NOJtCiO'-*0-** "^Mt OlXOpfHOSMOiM-* I- 11.^ l- 00 t- t- t'- I- CO t^ 3 « 9 It c « ^ ^ S i ® ^ g"aSS»330 aj 0) » t. J ^ •-« 67 u 0) o g a "1 O h-1 >- PQ H , P4 ^ !J5 O 1 1 P ^ K, 0) > S m 03 ^ o « :e5 o la « 3 O ^1 u eS O t4 o 2 -«1 •Suiuodo iu3«JtidT!ip JO juuoSBip JO 3iSu« l«n8!A ~ ;c ©J ■^ lo ec «s OiOO (Mi-l©lt>-''ijt-.«> V 8 &» rH •ipUT UB JO sqqoOOT "! Sup^do iu/ivjvicl«ip JO jBuoSBia •Sumado uiSBaqdwip jo |uuoS«ip JO aiSuB i^nsiA ■ipui uv JO smoOOl "I Suuiado mSBJiidnip JO ivuoSBid •auTiiado luSuuidBip jo , i«uo8tiip JO 3\Suv pmsiA JO sii^oOOT "I luSwiidnip JO •llOUT u« guniodo ^duoSbiq C; ^ O S S m1 55 Oi « M OS OS '-©■ d in «g i« --j t^ Jg >* cQ »n . t- oo ■* M lO 00 00 (3 ^1 53.2 68 Rkpresentation of the Space-Thkesholds for Grey Ground. Scale : 1 inch = 20 minutes. Curve I— Achromatic Threshold. " II— Chromatic Threshold. " III— Characteristic Colour Threshold. R O oY Y YG C bG B V P Fig. 7 (Table V). R C <■> "1 vO G to B V P Fig. 8 (Table VI). 69 o O o S o et cS a o M u > u 93 o ^ s o .2 J •Sutuado ui8«JndBtp jo JO 8^0001 uiSwji[dwip ui ipni uv SuTuado JO jnuoSniQ o -H.U B o o •Suiuado iu3«ai[d1 •Suiuado iuS«aiid«ip jo l^uoSrsip JO aiSuM i^nsi^v Haui u« JO 8^0001 "? Suiuado luSBJlldBip JO ItJUOSBIQ: I u SO ^* Ji^a . o : : : i i 1 I •J3 . ^ . . . iM OO (M 05 IfS -^ = M rj t-^ t-^ -"t 53 g d d M ■*X35O'^(M00t>.iJ5t- - 01 -t ; lO t>^ «' CO >C ■* "H -H 5-1 Tj< lO lO i-H 1^ !>. •<* 'M M OO "H "O IS •* irf iH OC -^ I- ^i 05 Tt< ^ ■^ 10 S-l S^ ift 5'I Tl" ift t-t <—• ift- q ^ a H » n fcifS,^'- 70 1 .a .a a> 55 D bO (-1 O a i o ■ fH ua V ;^ fO 5^ Si (T ^ o 1 II i.8 o 1 c" « IJ, 09 60 I .a a S o V- 00 « V « V 03 0! o o IwuogBip JO 9|Su« l^nsi^ 0?P (N W lO iQ lO CO Ifj •*■ t-i "H c«i (M lO ■^ •«»■ M iffl cc 00 ■^ « 002 «o «o eo 1-1 o ■^ ^' I>^ lO jo M CO W OS OS C5 'C t? J" sipoOOT iuS«.njdwip •qoui UB in Suiuado JO i«uo8biq I t-oo«5t- oseci-i'^i-- -4 -< > — '^ o *-" .2 -^ a o o ITiuogtiip JO aiSuw I«n8ij\^ S X T»< OS '^ . . . ■.f O OS O O 00 OS oo' a^ iffl lO iC iH mo iHt>t>r» xooiaos luSwaqduip in •ipui UB «UIU9(Io JO i«uo8biq 1-lTt-'^ l^Ql-^tO (NiOlOtA ;C«OiHCS IS o o •jT!iioS«ip JO a]8in! lTm8i^\^ ■^ iO ^ ^ ■^ M M ■* . O iC b" M eo 1:^ 00 L-- V.'' l^ 1-- I a 60 ,^ G- d O - « i 4) ,2 >» >> 7X g D § o I-H Q s H 0) P5 fe »5 o 1 ^a H -aj S .0 « 6-1 3 m 0) S.2 o o s .a I 99 s 2,% a .a 8 - >.« 14-1 «*H •8uiu9do uiSBJtidBip JO •qoui UB JO ai{^oooi ui Suiuado tuSHJlldflip JO IBUOSUIQ •8uiU9do mSBjqdBip jo ^BnoSiBip JO eiSuB l^nsi^Y •qoiii UB JO BinoOOT "? 8iiiu9do ui8Bjqd«Tp JO 1BU08BIQ •Sutuado m8BJi{dBip jo IttuoSBip JO 9]%uv l^nsi^ •l[OUl UB JO sinOOOT "I 3uiuado uiSujiidvip JO ibuoSbiq % o » 0^ in Tf t- 00 Q ©i ■^ t-^ "H «)' to IN ifli rt M w eo 'H <-« (M-^OO-HiHiap—c-HlO 10 i£5iN'^e*5©5io<-ieo i(5n:»-- t^ O CO t>. 50 ift CV C 05 M ctN'itoaoi'-ccOMOo -^ ■* •? S "5 S inaocox»«05 0st^«o lO T)< Ift iH iH M •^ - ■Tji 05 © 00 00 « 00 as !0 o j«He a-- _ lO-H-^OSCO-HOt^PTj 050S-^ift«D«0«0'---iWW (NiHIM>Hi-<.-(.-HtHiHiH S8 -M 05 © lO t- CO 00 !>; 05 Oi «■ 00 OO" to '2 CJ CO - t^ M ec M '^:i'<;>-2« OOO0Q0^O0O0««t-:C2 gsCOWtClft'Olffl'ftift'" M Ph o >^ >^ o P5 pa > 00 72 Representation of the S.-ace-Threshollh for Rbi, Orouni.. Sralo : i of nn inch = 20 minuteH. Curve. I (dotted lines) -Aehrom. ...Threshold. ^. rru u m .. II (lines coinposea of nlternate dashes and J ats) = Chromatic Threshold Curve lII-ChHracteri«tie C. Th. for Observer K.'s left eye (Table VIII). ,, _ .. .4 " " right eye (Table IX). ^^ y_ .1 .' for 10 observers averaged (Table X). O oY Y YG G bG B V P Fig. 9 (Tables VIII, IX and X). 73 C3 r c tf o C -C be 55 "3 H t4 (d o tf r*^ ;m S 55 O b* (U > K, fc4 !N ^ u o •o « ^ s o • CS 2 01 ^.H o o s 1 = 9) bo a 2 o a (V V h to B s e » V « » « » 02 CC JC •Siuundo mSiMijdiiip jo IwuoSrnp Jo ojSmi pMisij^ JO smooOT iuSw.ii[«iA JO sipoooi luSwJiidBip •qaui iiB UI 8inuodo JO |Tiu<)8m(j •Siuuodo ui8T?jiidBip JO Itmogvip JO ojSm! nm8ij\^ Unui lit! JO sii^ioooi wi Sum'ado mSiuiidwip JO iBuoSuiQ ^ -4* ^ 00 oo © f Ift -f IM to 1» OI l« 1 ^ |^i --j t-^ r: cc -n* 'I' I" ■* -^ in CO -^ T* i;; o '* C5 Tf -J ec 50 tc ,H-i S^M-H'M'-i -< 1^ o o ■-; O in eo oi 'N 00 o; s> rts-nnct-fifliM'-i'* K ?o ci t-- -^ ■^^ c » 05 ^_ "* O'^'sooo^occoda^oiop S lO lO -c5iNiff oi t>- 1-- 'ft lo 50 1'- ^ ^ 'S £; 1-- to I- t^ 00 a; o >o t^ »« o S M 'J « § S X '* >« « '*: _ _ .,....> lO C5 op ^ 1(5 in " -H rH • e> : ,^,H(N5-)t0irJ0^'H»H ,Hin:oooO'-<'0£2£2Si2 1-1 £ S fe « • ^ ^ ^ ^ eij r, ■ £ PS o kh >^ >* m PQ r» o o 74 Q S5 & O « O 0) S5 ^ M , se; (^ O 1 a) > wi 0) l»H »j 1-^ Xi H C ■» -*o -O ^S er-< a 3 "3 ;'-3 2 -I 1-1." O o n o S o <5 • Suxuado iu3«Jij Jutp jo IwuoSeip JO ejSuB l^nsi^ 3i L "f 5g ?*' t^Mift-*'HCCQ0 ^ -^ (N -^ (M eo <-H JO sti^oOOT IuSl,MlJ^ O C5 0-) -t 50 1-H ^^ »— I ^H •qoHi uu JO smoOOT "! Siiiuodo t(i-«ii[dvip JO i«uoS«i(i -♦;i-*r:tK: *j s^ X ^ »"• -< '2 ^ * "^ 2' «C5H»(MX5SOCD^'-21 ^ ^ ^ _H l.ot-50C005 ■qoin iin J" sqioOOI "! Sutuodo tuSnaqduip jo [buo^biq l^ !M O Q O iM O -H 35 eg g^ *i X ^ w *i in r- ■^ < >< >i « w > c c 75 3 J: 2 in cS C d . 3 cS to ^ o §-« ^ 3 ■w c« eS ^ ^ O h 60 (V f« ?n je c IH u 0>H d SB & , o a o: a O «3 a 2; -c 04 ■ rH o fcii Q 1 u t-^ 0) e: ^ J V o pO w &4 3 O •— ( c ■^ S 5l » Tf *< ■* «! -H — I 1-1 l-l J3 2 u -4J •8att?f)do uiSwaiidrnp jo IBtioSwip JO oiSuu l«n8!A ■^ t~- 01 05 -Hsqweocoeot-- Os lO Tf 00 in in Tji Tc » — . CO 00 o (M Tf-^(MiHin'<*. -^ 'X '5' i.'i OS COXXt--0-t--COI-- « 1-1 ?' c Is ^ fe so o o o ■ ^^ « 33^ >>>> 3) 76 O 55 O >1 o ■J} o c o o •3uiu3(I() uiSKjqihitp oi{4 JO piiioS -inn rtijj jo 3[Sub p'risi^ •ll.lllt IM! Jl) smpinwiioin III Suiiioilo uiSii.iqflrip aijq jo paioSwiQ •"luuodo uiSwJijdBip oqj JO imi(j8 -Btp 0H4 JO oj^iiTi junst^ •IJOtll IIV JO si[;pui."Bnom ui Suiuodo uiiiuitjditip oq; jo {wuo3i!i(j •Suniado uiSuiqduip oq; jo [i!uo3 -■Bip oqq JO 9i8u« i^nBi^ ■qoin u« JO smpuTisnoq; ui 3uiuo(Io ui8«jq(I«ip aq^ jo itiuoSviQ eo ta CO : !>• _ «S CO ©1 N l «" -^ oi d »c oi 35 00 TjJ - . 05 t, CO •^ 50 O 'P"2i-S<»'^'^ot-i«05 e'irHCOcoi>.co-Hcoi!3;oos ■O lO (M X -H 1-H 1C5 _ 1ft m 05 o © 0^ (M C?Xt^-f|~O'-i'^O05i-l ift t Tf -fl- m t^ CO ^ lO X CO -H X 05 X e ■* -5 -M o ti 10 CO !■- « U3 ■^f 1^ CO ift t^ .-H X ■* »oto m X ■^ i-i X •— I d 05' •T X> ^ :£ O: ■^' If; « C; i-( 3^ e-l 1-1 it ■^ ■^»^3-♦^^ -»!-<« — 05 05 C'1 3 -H lo r- la ^ CO X » I- X 05 »< O X X iS CO CO c^ -^ u 3 O O a - o o o Q, ■ c O O O r-^Sass . d* tf O >• tH >< OD pq t>- a C5 o V V 77 as O Q > a> O a o o ill 01 t4 2 o •73 i-'.a o o o (0 0) a o u u ^•.' ■ 1 ^ ^ « so I 8 g ' o X •Sinuodo luSiuijtliiip oi}:} JO paioS -Dip aij^ JO o{Smi ifiiiHijY •l[OUl UT! JO smpuTJsnom ui Sui'iodo iu8«ji{(Ib!P ail'; jo pjuoSiiiQ ?0 O (M lO «) «•? W (N S ■»(• rH >— I •^ (M lO la '— I 01 ■>* i-( ^ lo lO X 'M >-i 1ft Tf 10 o; 05 C5 1-- C fO S: lO OS — * Tf rH -^ 00 CO iH CO •Sinuodo itiSn.nid'Uip om JO ii!uo3 -wip om'jo oiHiU! pnisi^v lO ira ift CO lOCO'-^'^^-^^t^'*'* ~ i>. r^ CO i» op *T-2 >« i- »-r 35 la lO T Tji -^ ■* CO --I lO i» . CO C-l CO CO CO o ift ■* •^ t- CO o: cc o CO!|iT(l CO (Nrt CO CO C-l t- -t S-l M ©1 CO (M 00 CO Q •»*' w •-< CO 'O ^BUoSBip JO aiSiui {M\wi\ •l[0tlt UB JO yipoOOr "iSluiuodo UlS^t!.Ill- 3 3 3 t«H M^ ^1 21 ■ - o 2 3 O '3 o o >, s> hi 00 s s ■3 l-H &I -1^ c« 2§ 2 § a fl •~0 lOrc— 'X CO --i •-0 CO CO 50- 00 X 00 •* IT) ift 00 t> 00 © ■* M M CO ei n Tji »-( in CO CO CO 'i ! "M ;S '-3 «0 ?3 !0 cs e-i itj ift 'f 06 o '■£ 00 in 00 op j; 2 ex 10 CO r: CO a-) iM fo f -t -* s-i ^ irTc: ■* 10 1- 'S * ■; *< 2 r* 1" i« t- in i« «o 00 X X •* t- a) , , c o ^ ^ » 3 dj C aj 4) » fcc ^ . _ _ jg e5 O >- >H >^ O « >• O O C Pi tC- a 80 a; O P eQ O S T. o O > a; 3 o o o2 I— I r o — ' u ^ C o 2 "o ^ '-S ce S o u u <>3 -as tk C-: S 9 4|.« 1%-^ * p^-d 3 2 •*< 1 -tf 4) e« be ^S a> S fc. o '^JTJ -^ c d « S rt hi T3 : ^ ^ 3) hi s o o .JJ t «3 -. ■a 4 c l]> 0) m •i|r)tii iiv JO sti^ooOT "! S'liiuodo mfl\!.n[(lt!ip JO [vn jSviq •Suiuailo luSiJ.iijcIiiip JO t«uo8L>ip JO o[8inj p.'iisi^^^ •tlOUI 111.' JO sipoOOl "} '"utuodo UlStUlldlJip JO J 1! II O S 1! I Q •Suuiado iu8uji[dBip JO [tJuoS -uip oip JO f>iaini pinsjA 'Hoiii \w. JO HipoOPT "! SiH"'''!'^ uiSBjqdtJip JO J « u o 8 1! I (J irt . to •^ cc si ci '-C o4 c: 05 O t^ O >-H 5-4 rf Tf i-( lO m 3J IS ^5 CO CC t-r lO t- —I lO 5^ (M -< '^i rH o" (M (M CO rS ?1 CO iC '3 '* -ooX'tt- '5 a n hi to ^ hi '^ CQ "^ ^^ r-^ 3 « S U ® ^ e il - sri^ "-I '"' « © pji; t»> r»> 0) 3 o o o hi ^ '•« hi hi pHtfO>'KH>-lO«>'OC3 81 Q ^ W D -.4.:) O 05 To o C« Cd & fl >-] G w cS !25 S O "3 IB 1 Si • fH ^ i4 S N <» ^0 > ^ r- oo30»toot«os(N(r3i« ' fH c! oi 05 cc ci t>^ eo «i x -h ^ ■* M •^ M lO (M W C<5 . © W -^CO « O X 00 M O ■«*" ©5 F- O'tCM-HCMIMTjl •3uuiodo ni3tui[(l«tp jo 00C5 OO-^lM-.OCO'J'CO — O CC 00 C5 .-< '>] t>. (TJ © M 05 ■ t; 05 -^ oi o in" 00 to cc ^ 00 -H -^ »H Ct -1 ■* (M 5^ •>* . r: 00 « ^^ -; -* 'Ji « oq lO "t l^ C5 ■•*■ id t-i -^ 05 05" d C5 1 I- ^ iR lO O C © n CC t 05 rl 1-1 l-H •Suiuodo uiStmjdiiip jo » 1* «Q0-*XC^O5» ■^ _ fj 't •'i; M ift CO ;5 « t-. ■^' oj «' id C'i r: 1—' x ©i X 05 •H TT CO CC M W -* T(< rl ■* - in cs CO cc •* I- 1-- 05 -^ M ;o •qou; n« JO «mO()OI "• "(iiiiado ui5^if.n[ih!ip JO |i!uoi5«!Q •o T>i in © — c in *i — > tM o T X O ~ ^J rt C 1 ^ C-O 05 05 rH rt 1-1 -IC-I S-1 51 -H iH 2 'o 2 l-H U i -a o •iSuiU.)(Ic) UlSl.',IT[(Il.'ip JO p.'uoSwip JO ojnuii pii'sij^ 1-1 lO <» 1-1 -H 1-H CO 05 -J •>!»< CO X 05 t- O CO © ©' 05 CO 1-H CO rt s-1 r-i 01 r: oi co . ioe0'S ■»i< in "O -^ ')•_ X « o; -H X !•-. «j '*■ ct cc' ^' CO x" c^' 05 00 ;d coco C4COCN5-ItCCO-<*CO ■ cooicooicot'J'^coeo-* •IJDUl un JO ^q^OOOT "! fluiuado luSe.tqdtitp jo [uuoSbiq © 1-1 lO X © OJ TT t- i-< « X ©l-X«O> o M K -u ij cS § O ^ O C a s t3 cS 1-3 M m ^ O S5 £ o 1 s' "^ • • ^H tH o 3 O CD o o t4 o o 58 s o hi u hi s [uuoSBip JO oi3iit! [vnsi^Y •i{aiii in; j(i sqjoooi ui Suiuado mStumltiii) 9114 ji) jwiioSiiiQ •Suiuadi) uiSvji{iInip jo jfuoSBip JO oi^uc niiisiA •ijoui inj JO smooOI "I ''^uui'odo uiSwji{(I«ip JO [t!uo3«iQ •Suiuado ui8«ji{dBip jo puioSwp JO f)[8uB pinsi Y ■qaui ui! JO «"nOOOI "? io iC 'M lO lO — < O TT Tjl (N la •H ^H »H -H ■M (NO-* C5 O ^ i>^ •»!< x' ci l>^ 50 «o •ftC IMCOtMCOlOi-l Ix -f "^ i!5 00 © OJ lO « lO fH C5i«:£i-HOino»«©i>.cN xc;e<;occ©csi©i-^iN ^ ^^,^(>j^eo»Hi-((M ift lO O lO Ol <-< 'I' : ®-*ioeo-* gs 50 X © o X "O o". © m iji oj in t- !>. a. CO (M X © «c X -H M (M -H B 60 s kl 'd 3 a* w lu hi .::< -rK hi hi 00 ^ o cc O 00 H S t) N X >1 '^ 1 »[ « ■ a o , 1 s a " » 1 a 3 O S o u e •r . c ' 3 u •Sumado ui3wji{(I«ip JO I V u o 3 -Btp am JO ajiun {wnsi^ ya c ■nji t- _ •'^ 00 w oo -Hf^Q-^aisooJoioo© M -H la in C5 (M (N th S e2;oo«oo«o«QQ.^ift •ipui u« JO 8i{')0001 «> 3iiiuodo uiSunjcfuip JO jiiiio8bi(j i-j** JO o n . o: «' i>i -^ ci 05 05 w -H o lo 1-1 1-1 iH i-( — ll>-CCl--.lMO05«5(M(N(J1 OtOOitoe-IMMi-trtCs^ -H rt (N ©) :^ rt M o 2 2 u mSBjqdBtp -Bip aij:} JO •Sir' .ado JO JBIIOS ejSuB iBnsrY -«o:c:(NQoeO(M >a©j C:OT.l-«J''^'^05.TfOO M5 «J t-^ to Ift -H W ■^ ® 05 im' ■^•^'(J'CC.'.CC^ CO iH CCiM(M<>JMeO-*iOM(M..J( ijout urj JO smooOT "! 8uuiado iu3BJi{dBip aqi jo jbuoSbiq ■•*■<* 1ft .-(O0'*05.-(«^M !.■«. Ift ift lO t-- to r-t 05 i>. •^f 00 : : • c : : : ■ : g II gree een f u g. ? fc te a 60- -H S i3 3a>(. ^ i>a>(-i^.^bii4 (X■ >- o m ?• o o 84 Rbprbsentation of the Space-Thrhsholdh kok Blue fjRorND. Scalo : J of an inch— 20 minutes. Curve I— Achronifitic Threshold (avorage of 8 observers). " II— Chromatic •* •« " III— Characteristic C. Thr. for Obs. Dodds (Table XVII). " IV— " " « Kirschmimn, left eye (Table XIX). " V- " " " " right eye (Table XVIII). " VI— " " 8 observers, averaged (Table XX). m i\r fJ O oY Y yG G Fig. 11 (Tables XVI to XX). bC 85 SUMMARY. /. — Black Ground. 1. The chromatic space-throHhold has two decided maxima, in the yellow and in the blue regions. 2. The characteristic colour threshold has three str'mgly marked maxima, at orange, yellow-green and violet; and three decided minima, in the yellow, blue-green and red-purple regions. 3. For red the achromatic, chromatic and characteristic thresh- olds coincide; and for yellow the chromatic and characteristic thi'esholds coincide. //. — Grey Ground. 1. Similar results to those noted for black ground are obtained for the maxima and minima, except that the second minimal ragion includes yellow no. 1 (orange-yellow), and that the second maximal region includes green. We may also notice that the maxima are considerably raised while the minima retain in general their former position. ///. — Coloured Ground.'^. A very peculiar change of the antagonistic colour relations seems to take place. In small areas it is not the complementary colour which finds its most favourable conditions for being recog- nized on coloured grounds. On a red ground the minimum of the characteristic colour threshold is decidedly in the blue ; and on a blue ground the curve has an exceedingly marked minimum in the red. Thus for small surfaces on coloured grounds blue and red act as complementaries. A CASE OF ABNORiMAL COLOUR-SENSE EXAMINED WITH SPECIAL REFERENCE TO THE SPACE-THRESHOLD OF COLOURS By J. W. BAIRD, B.A., AND R. J. RICHARDSON, B.A. J^^ ^^ ABNORMAL COLOUR SENSE EXAMINED WITH SPECIAL REFER ENCE TO THE SPACE THRESH- OLD OF COLOUR. BY J. W. BAIRD, B.A. AND R. J. RICHARDSON, B.A. Notwithstanrlii.g the extensive literature already before the world upon the subject of colour-blindness, we hope t!,at an account of our examination of the case about to be described may bo of interest. The classification of the human family into cofouv- blmd and non-colour-blind is at best a matter of rouo-h and somewhat arbitrary judgment. The colour sense is%alled normal when expressions and actions lead us to suppose that the relations of the colour sensations are the same as with the majority of people, while persons whose expressions and actions show us that these relations differ considerably from those ot the majority are called colour-blind, or persons of abnormal colour-sense.* Colour-blind persons may, according to Dr. Kirschmann be classihed on the basis of the degree of their abnormality into (l)achromates-who are totally devoid of colour-sense and see objects only in various degrees of grey, (2) dicl.romates-who see only two antagonistic colour qualities, such as red and green or blue and yellow, and (3) abnormal polychromates-who disthiguish more than two colours, though the relations of their colonr-sensations differ from those of the normal. The results of the investigations of colour-blindness can only approximate to accuracy, owing chiefly to the difficulties which we always encounter when we attempt to ascertain and describe the psychic states of others. Colour-blind individuals u.se the ordinary colour vocabulary, which is quite inadecpiate to express their sensa- tions of colour. The experiments which furnished the data for this paper were conducted in the psychological laboratory of the University of Toronto at the recjuest of Dr. Kirschmann, director of the labor- atory, to whom our thanks are due for cooperation and suggestions. *This is the definition of normal colour sense and colour-blindness which Dr. Kirschmann gives in his lectures on Psychological Optics. [87] 88 We are also indebted to Mr. W. B. Lane, M.A., now Fellow in Psychology in the University of Wisconsin, to Mr. F. S. Wrinch, M.A. and Mr. A. H. Abbott, B.A., Assistant in the psychological laboratory, for valuable assistance in our work. This case of colour-blindness was investigated in the following ways : I. Spectroscopic examination. II. Stilling's Pseudo-isochroniatic tables. III. Colour equations. IV. Experiments on space-threshold of colours under contrast influence. I. A series of observations was made upon the solar spectrum by observer R. (whose colour-sense is abnormal), with a view to determining the nature of his deviation from the normal. The wave-lengths are calculated by graphical interpolation, the num- bers for the principal lines being taken from the tables of Rowland, Mr. R. always claims that he sees three distinct colours, red, green and blue, in the spectrum. Using the left eye Mr. R. saw the light, 920 ^ju to G70 /Lifi, dark brownish red 670 nn to 540 |UjU, uniform red (o90 fifi, red of deepest saturation) 540 fi/i to 493 fi/x, green (513 fifjt, best green) 493 fxfi to 384 nn, blue (4?.9 fi/n, deepest blue). The maximum of light intensity was seen at 590 ^/i. At 702 ju/u. the line A was seen as distinctly as the other lines, whilst observers B and K could see it, but with diiRculty. In the red end of the spectrum in addition to the lines A, B, and C, other lines were clearly distinguishable at points corresponding to the following wave-lengths : 603 /uju, 626 luu, 644 ^ju, 728 jUju, and ut tivo points beyond the A line in the tdtra-red spectrum. The w^ave-lengths corresponding to the latter points are not computable from any data within our reach, but wo estimate them from our Interpol ation- curve to be the Z line and '^he X^ line (Abney), corresponding to which are the wave-lengths 822 jUju and 880 f.ifx respectively. For the right eye the colours and their positions were given as follows : — 900 jUju to 687 nn, dark brownish red (726 jUju, most characteristic brown-red 687 fxn to 541 fifi, red (590 jUju, red of deepest saturation) 89 541 fin to 501 fifi, green (527 fjfi, best green) 501 ftn to 384 ujn, blue (427 jiiju, blue of deepest saturation). The lines in the ultra-red spectrum (especially that at 822 it/.i) were .leen as distinctly as with the left eye, whilst observers B and K could see absolutely nothing beyond 762 fi/ji and 770 jn/x respect- ively. Tn addition to the examination of the solar spectrum with the spectroscope; obs'^rvations were made upon an inverted spectrum* projecteil on a screen by means of an arrangement recently devised by Dr. Kirschmann for the purpose of comparing the ordinary and the inverted spectrum. One of these spectra had its colours arranged in the usual order — purple being absent. The other was thrown upon the screen parallel with and contiguous to the first, but with the order of the colours inverted ; that is, the colours proceeded through yellow, orange, red, purple and violet to blue — green being absent. A photograph of the two spectra is given (p 101.) The only difference noted by observer R. in these spectra was that in the inverted form the dark brownish red was missing. He could distingithh vo (llrference in colour tone in the centred section of the two t'pectra, that is, between the purple of the inverted and the green of the ordinary spectrum. The statements of observer R. were essentially the same when the two spectra were divided by means of interference bands. 11. — A series of tests for colour- blindness by means of Stilling's Pseudo-Tsochromatic Tables (third edition, 1889) yielded the follow- ing results : Plate I (orange-red figures on brown ground) was read slowly and with a great deal of uncertainty. The ground was called green, the darker patches brown or purplish. The figures were called blood-red, their lighter patches yellowish. The figures were more easily distinguished through a red glass, but best of all through blue. Plate II (orange-red figures on brown ground) was more easily read than Plate I, but also with considerable uncertainty. The ground was called green, the darker patches murky red like the extreme red end of the spectrum. The figures were said to be com- * For description of the inverted spectrum see the article hy Dr. Kirschmann on Colour Saturation in the American Journal of Psychology, Vol. vii., p. 387, and note at the end of these articles. 90 poHod of patches of rod nnd oran