: let et cared hat ne Pe LAP Wes is er PP dutse) arse, ~~ Bo vir ” helt a8 it 4 we 5 Oa A Baty teat an, ha ye 4, pie vate Merrit as F phi iba aie is EWP ah pet t robb: oan Apa nat Minn nn sere HE MOLAR E at a DU Beet et SH 1 ee Se es ise ¥ i Lt bee Pee yi Gp c7sga THY et Te weg ete ts te As eherarks tt oon ties i iN A ep adeade RAS qatey Pea et | varies it te ue Y . % teydea ets Mia phos fe 1 shy a I i j 7 ae enh STAC Lat CLP ea Gr oes eT ea ate hy oF eS sc 1 "53 in at nets 45 ts cats by ri + a Set eee! sag T3532 rae. s Seats ptt rete be > hy onthe Hie te seat Asie oo tg Peeters Hh EAC =e 7c. ey ates Sec 523 a A OAS pi She ee My ; 2 £2 a HE AE 4? He Ak uy behets ie; figs ess tf Ete if weg . aioe tae ite eit ge i en dae ‘ 4 4 beat 4 tf | 43) ti : 4 4 , 7 : 4 4 bk Ny aaa ely tit 204 ones e | + i iF tt 4 ae Nanna tran i te Py Re MIL aE ae nen RUN NA ait ht GN ab tae in ti ' Vouiss TRF ait id pe f 1 ee ee abd atk Testeeas ROMAN Latit ih eth i i aK Aes NOL ees i 12) erie Print eh cea aN EN PS Ay nites Bl ts iL Aut a ‘ iy dei At r Ra ete ri Gi ft vi ee reales t Ped MMe Wey ae eri ink Bini ‘ lesan} ROH as AVE sath ban he ‘ i) k hy ‘ kal ; ‘ 59992 iti val ie si gidensd ‘yee Hes ne yd vie thot d thy : H ‘a ia jangle toy ' hectgut tee 24 1A ttt ire) TOR are +e UA i at ie preity ry na i ihe ri rae Hitate ms ‘ . Les f 4 i li { Hl Ranh # rita tatate (} peur ent) : i oi a Af) hy paid ra +40) pil y 4 Uy ef +) Wee u Bia! M4 ity «s) Te ae POEM ar te ed Sete ves ee it ites 7 Hythe u favpee id + + hal aAee4 | ; > } et putt o ad) \4 tad 4 $3 hae oy ey ithe paehieitat Me t ‘geen PEPRT ELS) be Fray 5 " ri a gt Er MG UEC RL : L. 15!) A Si PE ET It} a AY AN EULA iF sft ye yee mr) fs ; [ f 4 Cane dhnagt ea 8 cyan : N ere ah fet Hehe wae, Pres fhe N) tnt Ane TOA CMH HO roe he! i if si dpby) Hl aend yh) sat ejay neat Seay egg ARLE Litany bated by ares * , : wie Tarbes lay at pea! a seta iA 5 EAD ped as PH he BARE pjaisbebed W rt i erties ¢ i ‘ Teak] ee } hata lt AR Rt 4 FI payee ey ! ae TReY if " i tie pasARbr gsi folkd doe i aly ub Weal By rt ee tenitas ad sensi aires RIN at Hie paties Ua) hh ; 7 i 104 Agyrarats a bit roryey tence b sand t i ihe mise if! ee ” oT 4) nnd gc Bo ned | det og ein wrote tl a bees pede es Od paelet sha bree jit ty By debt pti Te ae denlb salads he state aa) aa bhai sy Hg y OANA IAD Poaiatr Hib 8g AHR myer eeniter Pe ltiny Meck atte) Hie driers we be tie 2910108 Bats bt dy ttt PAWN beat en! i} hie sie a atibiee tight TAS ea Ai ge vitl Vee ttt ag ele Ati oh ane eis ht sitaaiiyta E fan Beh) liek 4 rhs : ‘ yr ATT A) ha : 3 Mice Pico We ifkew? ae ey ERA | yas Ll Ah pes eae gait Fibs a eee i ak on AN Bow Pee, ranch dav : eM RAMS Rede ec 7 5c peat atone. chohe Coat ‘ . te titan UCase ict ae a? rad Linge: ats Seah: , “4 : Ty Hi ahs 7 _ ; 4 f COE aie aL ere HCA woe Lee 4 Dy ies f a 4 fe , ety ‘ td ’ oS ae ott ss opel y . i - ‘ , ‘ 1 ca a oy ves 7 2 ) 7 ha erry Se? 4 f- 4 + P iy j . yi 5) eige' Pate D AL ie y ret yen, et mph ra ite res j j / x ve We *, ss Yiahdele Ue Mig ss isi wy fan. & pAlte Yay 2 us tia & ; iy aiare ‘asin ae 4 é Shuey . tv mere ify Pate tars: - wen 7 . ene tS 7 ’ Fea it a" } aS tet {ied ZZ os : ine Leong: pet } ; ‘ no | - Fh . ad é mA, : CG Bote el! rh ey ae . y ‘ sh, 4 . , a e i 1 ~ ‘ ‘ 4‘ ; r */t ¥ . “at t + ’ ; t Mi +. ’ Pee | i , aa b f y 4 * ae | . ’ ; ~ / : - t ® { id o ; « A a | ™ } ay ‘ * ae ee oy MS ” ‘ i * é * ne i é > ‘ = | t ve { 4 ia 4 a . : ; [ es a4 ese ies oe : ae r 4 i t ecre Ky aM vitcleeee?: Merete: i 4 ire, nth te 1 ‘ ) = : + ‘ fs ra i ) ™ \s ; , r \ f /f A ccc "PSYCHOLOGICAL REVIEW PUBLICATIONS Rn eere oat 158 Psychological Monographs EDITED BY SHEPHERD I. FRANZ, University or CattrorniA, So. Br. HOWARD C. WARREN, Princeton University (Review) JOHN B. WATSON, New York, N. Y. (J. of Exp. Psych.) MADISON BENTLEY, University or Itirno1s (/ndex), and S. W. FERNBERGER, University or PENNsyLvania (Bulletin) The Energy Value of the Minimum Visible Chromatic and Achromatic For. Different Wave-Lengths of the Spectrum By MARGARET M. ‘MONROE Bryn Mawr COLLEGE PSYCHOLOGICAL REVIEW COMPANY PRENCH TO NEN se AND ALBANY, N. Y. Acents: G. E. STECHERT & CO., Lonpon (2 Star Yard, Carey St., W. C.) Leripzic (Hospital St., 10); Paris (76, rue de Rennes) Digitized by the Internet Archive in 2022 with funding from Princeton Theological Seminary Library https://archive.org/details/energyvalueofminOOmonr PREFACE In the work reported in this dissertation three important deter- minations have been made at seven points in the spectrum: the minimum visible achromatic and chromatic and the photochromatic interval. or the first time these determinations have been made by direct measurement in absolute energy terms. The inves- tigation is one of a series of studies the object of which has been primarily to lay a foundation for the exact measurement of human responses in terms that are quantitative or numerically comparable. Such work has been possible only within the last few years. The importance of the study as a foundation for the more scientific phases of medical work on the eye should also be noted. Further work on this and other applications of functional testing to the study and diagnosis of diseases of the eye will be carried on by the author in a research position in the Graduate Medical School of the University of Pennsylvania. The subject of this dissertation was suggested by Professor C. E. Ferree and Dr. G. Rand of the Department of Psychology of Bryn Mawr College and the dissertation was prepared under their direction. The writer wishes to acknowledge her deep indebtedness to Professor Ferree and Dr. Rand for their careful supervision and constant help during the work on the dissertation and for their great kindness and support during the whole of her graduate course. She takes pleasure also in expressing her appreciation to Dr. Rand for her kindness in making the energy measurements given in Table V. Thanks are likewise due to Dr. Luther C. Peter, Associate Professor of Ophthalmology, Philadelphia Polyclinic and College for Graduates in Medicine, University of Pennsylvania, for his courtesy in sending patients for the section on pathological cases. iil , be ebeise (agi eis Lekie, ME EY ® } , y { } CONTENTS PAGE IRL TRODU GLUON GE aU MARE: ce co digas a Rocka a/6 oo sacar TORR ae 1 Mined ESTORICAIS UM MARY afrika sinter oe bits ca's so Lae 5 PILE PARATUS(AND EROCEDURE. oF Gy 45.0.0 cs,08 ose a od oe eae 15 EAM SELEY OUT CE OLE LACIE OE ON aay cle cif alee sn mists ee oS Sere Larter S POCUP ORCI PEN ae a Tre Ske sere aay Wis ais a Se 16 ‘C. Apparatus for Presenting the Light to the Eye.... 19 D. Means of Reducing the Intensity of Light........ Os Wel nepkilterse eset enw. ra weeeeitels sue tn 22 UAL HE) CVECEOT OCS DO TSCRu te eon el crt Ss oye tee 23 Die Dea edecaie. cakes sips cit choi Daas amt 3 23 rl ne adtometriccA pparains git os as nga hie 24 FAT heik nergy COsUremnent san oy cried oan adie sashes 24 Cruel he ethos wt ODServaHon. ix eis «ching » Vises 27 LV. STATEMENT AND DISCUSSION OF RESULTS. . 8... f. 062.6) 28 FA CRvOMae LIM ESHOLAS A). es ave e so OEE) eee 28 SC UITOIMCTICML LT ESHIOLIGS oe oe x ind Ca eae elven 34 C. The Photochromatic Interval .05.0....0000 0 eee ee 43 D. Comparison with Previous Determinations of the Le SHOLGs CASHEL Ctr en wd oe ord eunane Vaio de meter” IVE SPATE be ESE DIC cc Let kis voces Wecm,e een: ale Soke 48 PeMIOL OL OCACOL, CSG 5 Nyc eer lei Ae nla Re i8ed (edgy Mx, 54 meena ease ta Fie? One et he Seek te Se UREA Ee or sae 59 Pe Wey 7 ‘; Ay ’ TR Sek ) ¢ oe pears Hy i. Wits Mid oS SAIN RTS serait ae TT ae ? ‘ At OAs ‘ 4 ae Ati, Eel te aay, nen \ ny viata ihe aan din ¢ bay I INTRODUCTION By the minimum visible is meant the least amount of light radiation to which the eye is capable of giving a visual response. Obviously the term may be used, broadly, to mean the least amount of light of any composition whatever to which the eye gives a just noticeable response; or, narrowly, as the least amount of the wave-length to which the eye is the most sensitive. Apparently the purpose of a recent group of writers has been to favor the narrower usage of the term although it has been variously applied by different members of the group to determinations made with the light of the stars, the light of a tungsten lamp, and to wave- lengths in the mid-region of the spectrum. The broader usage, which is the older, is more compatible with the purpose of this investigation and will be conformed to in the statement and dis- cussion of results. Used in this sense, the minimum visible is synonymous with the absolute threshold or limen. Numerous points of interest may attach to the determination of the minimum visible. (1) There may be a scientific curiosity to know the least amount of energy to which the eye is capable of giving a response and to compare this with the least amount to which the ear or some other sense organ gives a noticeable reaction. Wien, for example, was led to attempt a determination of the minimum visible by his previous work on the minimum audible. A natural extension of this interest is a comparison of the sensitivity of the eye with that of the physical instruments which respond to light. Coblentz has found, for example, that it has, roughly speaking, 300,000 times as great a sensitivity to light radiations as the most improved type of thermopile. (2) In the evolution of the sense organs, many complex char- acteristics and adaptations have developed presumably in the interests of functional efficiency. The eye, for example, has 1 2 MARGARET M. MONROE developed to a high degree a selectiveness of response to light radiations. Obviously in adapting light to the service of the eye in problems of illumination and in making a correct use of the eye in the many scientific and technical ways in which it is employed, it is highly important, therefore, to know minutely this selective- ness of response both as to kind and amount. It is also interest- ing from explanatory and technical points of view to be able to compare the sensory with other known and better understood types of response such as the photoelectric, the photochemical, and the thermoelectric reactions, the action of light on selenium, etc. For example, explanatory theories of the eye’s response have already been developed in terms of two of these types of reaction—the photochemical and photoelectric—and all of them, including the sensory reaction, have been utilized at various times for rating light intensities for scientific and technical purposes. (3) It is generally recognized that the most sensitive means of detecting the eye’s abnormality due to natural causes or its subnormality due to pathological conditions and processes is in terms of its lack or loss of sensitivity or power to give response, relative and absolute. There are many practical applications of this principle. For example, (a) one of the ways in which ocular fitness for vocational purposes is rated is a determination of the light and color sense. Eyes vary a great deal in light sensitivity, particularly in their range of sensitivity to light intensi- ties. Many, for example, who are able to qualify for work at medium and high intensities of illumination are disqualified for vocations requiring keen power and quickness of vision at low illuminations. Others lack in color sensitivity, varying from a complete absence of power to sense one or more colors to slight deficiencies which disqualify only for work which requires special powers such as great keenness of discrimination, speed of dis- crimination, etc. And (b) the most pronounced and earliest manifestation of pathological conditions of the sensory mechan- ism is the loss of light and color sensitivity. This varies from a slight deficiency to a total loss of function, depending upon the severity and stage of advancement of the pathological condition. The application of the testing of light and color sensitivity to THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 3 diagnosis, however, is in its infancy, partly because of the lack of instruments and methods which are feasible for office and clinic work and partly because of a lack of knowledge of norms of sensitivity for the healthy eye and of the deviation from these norms which are characteristic of disease and which differentiate one diseased condition from another. Some of the known effects of pathological conditions are a diminished light sensitivity, a particular and differential manifes- tation of which is a greatly diminished or entire absence of power to see at low illumination, absolute and partial loss of color sensitivity, changes in the relative sensitivity to the different colors, and changes in the interval between the achromatic and chromatic thresholds of sensitivity to colored lights—the photo- chromatic interval. As has already been stated, but little use has been made of the testing of light and color sensitivity of the central retina by the medical profession. Their work up to the present time has been confined chiefly to the mapping of the fields of light and color sensitivity. Even this has been almost pro- hibitive because of the care required to obtain an acceptable precision of result, and the time consumed in making the determinations in a sufficient number of meridians. It is a matter of great interest and importance, therefore, to see how far central sensitivity testing, which with methods amply sensitive for diag- nostic work need consume but little time, can be substituted for field taking. As already indicated, field taking is at best a time consuming and difficult performance from the standpoint of the patient, the physician, and the apparatus and controls required. While the medical aspects of the problem have made the strong- est appeal to the writer’s personal interests, it has been deemed advisable to subordinate them at this time to an endeavor to build a sounder groundwork on which to rest the applications. It is important in every applied field that there should be standards of reference as to apparatus, methods, and results against which the deviations made in the interests of convenience and feasibility can be checked and evaluated and from which fresh starts can be made. There are two important aspects of the testing of retinal sensitivity: the testing of the absolute sensitivity to the different 4 MARGARET M. MONROE wave-lengths of light and a determination of the comparative or relative sensitivity to these wave-lengths. Differences in both of these regards are fundamental in all of the fields in which sensitivity testing and its results may be applied. To serve as a standard of reference it is obvious that absolute sensitivity should be expressed in absolute terms both with regard to the composition of the stimulus and to its intensity, i.e., for the eye, in terms of wave-length and energy value of light. It is equally obvious that if the sensitivities to the different wave-lengths are to be compared, the intensity values of the stimulus of which sensitivity is taken as the reciprocal must be expressed in terms that are numerically comparable, i.e., in terms of the physical intensity or energy value of the lights employed. Norms either of chromatic or achromatic sensitivity have not as yet been determined in absolute terms. Fragmentary attempts have been made on the basis of results obtained from a few observers to express achromatic sensitivity in relative terms. The investiga- tion of chromatic sensitivity, however, has not progressed even this far. The work of establishing norms of absolute and rela- tive sensitivity for both of these types of response will involve expenditure of a great deal of time and effort and the cooperation of many people. Even to contemplate it seems overambitious at. this time. However, every work must start from small begin- nings. It has been the purpose of this study to make such a beginning of the study of the thresholds of chromatic and achromatic sensitivity. The following determinations have been made at seven repre- sentative points in the spectrum: (1) The achromatic threshold. — (2) The chromatic threshold. (3) The photochromatic interval. Twenty-one observers have been used in making these deter- minations and in every case a direct energy measurement was made of the light used to stimulate the eye. In addition, determinations were made in a limited number of pathological cases. II HISTORICAL SUMMARY As early as 1888 Hermann Ebert (1) attempted to ascertain numerically the relative achromatic sensitivity of the eye to wave- length. Two lines of evidence had led him to suppose that the eye might have a maximum of sensitivity to green. In his con- clusions the emphasis is on the position of this maximum rather than on the details of the threshold visibility curve. The first probleni, his more immediate incentive, was the explanation of the striking simplicity of the spectra of the gaseous nebulae which in most cases consist merely of lines in the green and the blue- green, A 500,495 and 480 wu. Two theories of this phenomenon had already been advanced: first, that only these wave-lengths are emitted, and second, that there is selective absorption in intra- nebular space. Ebert thought it more probable that the cause lies within the observer. That is, if the eye should prove to be most sensitive to green, it would be reasonable to expect that in a weak spectrum, such as that given by the gases, only the green would be visible. His second, more general motive, was the reopening for discussion of the whole psychophysical question, a subject made pertinent by the work of Weber (2) and Stenger (3) on Draper’s law during the previous year. In discussing the results obtained in this work, Weber had assumed a direct proportionality between the responses of the eye and the energy of the light wave. Stenger had pointed out the incorrectness of this assumption but his discussion of the matter had been slight enough to warrant a more complete investigation by Ebert. Draper’s law refers to the order in which the wave-lengths of the visible spectrum emitted by incandescent solids reach the threshold of sensation with increase of temperature. Draper had stated that all metals begin to glow at the same temperature, about 525° C., and that the development of the light emission runs the following definite course: at 525° the light emitted gives 5 6 MARGARET M. MONROE a spectrum which reaches from line B to line b; at 625° from B to F; at 718° from B to G; and at 1165° it gives a spectrum approximating that of the sun in extent. According to Draper, then, the spectrum of a glowing solid whose temperature is gradu- ally increased develops practically in one direction only—from red to blue. . In the course of some experiments on the relation between the brightness and consumption of energy in carbon lamps, Weber had noted some entirely unexpected phenomena which caused him to doubt the correctness of this law. Watching the development of the light emission of the carbon filament with increase of temperature in a dark room, he noted that the red glow was not the first light visible, but that another light appeared and under- went an entire series of changes before there was any sign of color. This light he designated as “ gespenstergraue’’ or ‘“dusternebelgraue.”’ At first it was very unstable, although whether its rapid appearances and disappearances were due to change of intensity of the light because of fluctuation in the temperature of the filament, or whether they are due to an eye phenomenon, he did not know. With a slight increase of intensity the light became a little less variable although it retained its “ dustergraue”’ quality. With greater increase the gray light- ened, and the coloration gradually changed through an ashgray to a definite “ gelblichgraue.’’ By the time the red glow was visible, the light had become fixed, losing the flickering quality which it had had throughout the previous changes. A prismatic analysis of this first gray light with the complete spectroscope was not possible because of the weak intensity, but Weber was able to observe the changes through a direct vision prism, and, at slightly higher intensities, through a grating. The spectrum of the ‘“ dusternebelgraue”’ light, when first strong enough to be seen through the prism, consisted of a homogeneous gray strip which occupied the position where at higher intensities yellow and yellow-green appeared. As the temperature was increased this strip broadened and brightened. When the tem- perature had reached a degree such that with the naked eye the light looked “ gelblichgraue,” its spectrum was seen as a broad. THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 7 band, yellowish-gray in the center and shading to a weak gray on either side. When, viewed by the naked eye, the light was red- dish, there appeared in the spectrum at one side of the gray strip a small “ feueroth ”’ space, and almost simultaneously on the other side a gray-green strip. As the light became whiter with increase of temperature the spectrum continued to develop until finally the entire spectrum was present. This description of the light emis- sion of a glowing solid was verified by Stenger, who repeated Weber’s experiment. It is in his conclusion that Weber makes the assumption that visibility is proportional to energy, an error which, as was stated above, was pointed out by Stenger. Weber states his conclusions as follows: ~“Das Spectrum des glithenden Kohlenfadens wachst also bei steigender Temperatur nicht einseitig, in der Richtung vom Roth nach dem Violett, sondern entwickelt sich, von einem schmalen Streifen ausgehend, genau von seiner Mitte aus, gleichmassig nach beiden Seiten. Die dem Auge zuerst erscheinende, den Ausgangspunkt der Spectrumsentwickelung bildende Strah- lung ist dieselbe Strahlung, die im vollstandig entwickelten sichtbaren Spectrum dem Auge mit der grodssten Helligkeit leuchtet und in dem schwarzen Flachen der Thermosdule und des Bolometers die maximale Energie entwickelt. “Daraus ist wohl der Schluss zu ziehen, dass diese Strahlung mittlerer Wellenlange deswegen dem Auge am friihesten sichtbar wird. weil sie auch schon bei der Temperatur der beginnenden Graugliiht die maximale Energie besitzt, infolge dessen ihre lebendige Kraft am friihesten jenen Schwellen- werth tbersteigt, welcher vorhanden sein muss, um eine Lichtempfindung zu veranlassen, und dass die iibrigen Strahlungen kleinerer und gr6sserer Wellenlange dann bei steigender Temperatur der Reihe nach dem Auge sichtbar werden, sobald deren lebendige Kraft einen Schwellenwerth ahnlicher Grosse tiberstiegen hat.” Stenger, referring to Langley’s (4) results on the energy dis- tribution of the sun’s spectrum, showed that since the greatest energy is in the red, not in the green or yellow-green, the eye must be selectively sensitive, and that the maximum of this sensitivity must lie in the green. It was at this point that Ebert took up the problem. The details of his experiment are as follows: The source of light was a gas flame which illuminated from behind a screen of oiled paper. This screen was assumed to be evenly illuminated and was focussed on the slit of the spectroscope by a lens 12 cm. in 8 MARGARET M. MONROE diameter placed at 125 cm. from the slit. The observer, looking through the objective slit, saw the face of the prism filled with spectral light. The intensity of the light was then reduced until it was no longer visible. For each wave-length used, determina- tions of the threshold were made in ascending and descending: series, and the mean of the two series calculated. The variation amounted to between 2 and 3 per cent. His values are for only five different wave-lengths and for two observers. The length of the adaptation period is not given; Ebert states merely that the experiment was performed after “ sufficient ” dark-adaptation. The reduction of the light intensity was obtained by a dia- phragm, .07 cm. in diameter, placed between the focussing lens and the collimator slit, and so mounted that it could be moved over the entire distance—125 cm. As this diaphragm is moved nearer to the focussing lens, it decreases the diameter of the used portion of the lens, thus reducing the intensity of light focussed at the slit. If E is the distance of the diaphragm in cm. from the slit, and D the diameter of the used portion of the lens, then 125 X .07 Da E The intensity of any part of the spectrum is then proportional touD*: The distribution of energy in the spectrum of the light source used was calculated indirectly by combining the results of Langley mentioned above, with those of Meyer (5), who had checked the energy distribution of a gas flame, the source used by Ebert, against that of the sun. Langley, using a bolometer, had deter- mined the energy distribution of the sun’s spectrum in relative terms, 1.¢., in terms of galvanometer readings—no absolute values are given. Meyer had compared photometrically the spectrum of a gas flame with that of the sun, obtaining a series of ratios representing the energy distribution of the gas flame relative to that of the sun. Ebert, to get the energy distribution of his own source, multiplied the values given by Langley by the appropriate THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 9 ratios as given by Meyer. The results of these calculations are shown in Table I, quoted from Ebert. TABLE TL Helligkeit Gaslicht E Mittlere Helligkeit Sonne Sonne E Farbe Wellenlange (Meyer) (Langley) Gaslicht Oth Irae clea ee 675 pp 4,07 62 252 Gel Die anette ee hore 590 uy 1,00 45 45 Grits ee es 530 pp 0,43 28 12 isrunblauneed. ir oe Ale 500 py 0,43 22 10 (VEG Geond Baleege oor 470 py 0,23 14 3 In this table the second column gives the middle wave-length of the various spectrum bands used; the third volumn gives the ratio of brightness of gas flame divided by brightness of sun for the particular wave-length; the fourth column gives the relative energy of the same wave-lengths as determined by Langley; and the fifth, the product of the third and fourth, shows the relative energy distribution in the spectrum of a gas flame. Such a calculation would, at best, yield results only approx- imately correct, and in this case it is still further open to criticism in that Ebert did not use exactly the same wave-lengths as Meyer. As was said previously, however, he lays little stress on the shape of the visibility curve, insisting only that the maximum lies in the green. The final threshold values were obtained in relative terms by multiplying the values given in column 5 of the above table by D?. Since sensitivity is taken as the reciprocal of the threshold value of the stimulus, the relative sensitivities of his observers were as the reciprocals of these products. The values of these reciprocals are given in Table II. TABLE II Observer Red Yellow Green Blue-Green Blue S. 1/25 1/15 1 1/1.3 1/3 oN 1/34 1/17 1 1/2 1/4 In 1889 Langley (6) himself published results bearing on the visibility of radiation. He wished “to make the novel calcula- tion as to the actual amount of energy either in horse-power or any other unit, required to make us see.” 10 MARGARET M. MONROE Unlike Ebert, Langley determined the minimum visible by means of reflected light. A piece of white paper on a black screen was illuminated by spectral light of different wave-lengths. As source he used the sun—reflecting its rays into the spectroscope by means of a siderostat mirror. The gross reduction of intensity was accomplished by reducing the effective area of the collimating lens by placing before it a metal plate pierced by a minute aperture, by a sectored disc, and by a glass screen very lightly smoked. The finer reductions were made by varying the width of the collimator slit. By actinometric meastirements, Langley found the solar radiation to be 1.5 calories per square centimeter per minute. Knowing the reductions made, and esti- mating the absorption through the optical system, he calculated the values of the minimum visible for four wave-lengths to be as shown in Table III. TABLE III Color Wave-length Energy (ergs) Ratio of Energies Crimson 750 pe 1/780 450000 Scarlet 650 um 1/1600000 230 Green 550 um 1/360000000 1 Violet 400 up 1/1500000 240 These results are for one observer only. The fourth column shows the relative energy values of the threshold for the four wave-lengths in question. Langley says of these results that they are subject to variations of a wide range, and may perhaps be in error by as much as 100 per cent. The next results bearing on the subject are those of Konig (7), who, as part of an extended experiment on the relative brightness of different bands in the spectrum at various, intensities, deter- mined a threshold visibility curve. He used the Helmholtz spec- trum color mixer, reducing the intensity of the light by means of slit width. Like Ebert, he used Langley’s figures representing the relative energy distribution in the sun’s spectrum, multiplying them by ratios previously determined by himself and Dieterici in a comparison of the energy distribution of the spectrum of a triplex gas burner (the source used) with that of the sun. His data are for two observers and fourteen wave-lengths. Table IV THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS Il shows the relative sensitivity of the two observers to the wave- lengths employed. The method of calculation is similar to that of Ebert, described above. TABLE IV Wave- Observer Wave- Observer length A.K. R.R. Av. length A-Kiai kerk. Av. 670 pee .00019 .00017 .00018 535 up Aa .60 .68 650 wy .00047 .00056 .00051 520 wm .98 .83 .90 625 um .0038 .0048 .0043 505 wu 1.00 1.00 1.00 605 pu .015 .019 .017 490 up .86 5 .80 590 wu .045 .033 .039 470 wp .50 ait) .50 575 wm 12 uy “12 450 pp ves .26 25 555 we .36 ooo ‘35 430 wm .047 .059 .053 Pfluger (8), in taking up the problem in 1902, adopted a pro- cedure similar in principle to that of Ebert. Unlike Ebert, how- ever, he determined the energy distribution of the source directly with a thermopile, although only in relative terms. Also a greater number of observers was used. A Nernst filament was used as source, the light from which was focussed on the slit of the spectrometer by a condensing sys- tem composed of two triple achomatic lenses. This seemed to satisfy best the need for a source whose intensity was both con- stant and sufficiently great to permit of direct measurement in the violet. The energy measurements were made by means of a Rubens thermopile mounted at the objective of the spectro- scope, and a DuBois-Rubens galvanometer. The radiometric apparatus was not, however, calibrated against a standard, and the energy curve is, therefore, in terms of galvanometer readings, not absolute units. In order to allow for the reduction of the light to the threshold of sensation, slight changes in the arrangement of the apparatus had to be made. The ocular slit was shortened to 34 mm. in height, which with the breadth of 14 mm. was considerably smaller than the observing pupil. Milk glass was placed over the collimating slit to give evenness of field brightness, and the focussing lenses removed. The energy measurements were cor- rected for the absorption of the milk glass and an attempt was made to compensate for the absorption of the lenses by inserting glass plates in the beam of light. A bar four meters long was 12 MARGARET M. MONROE added to the collimator arm. The illumination of the milk glass could then be varied by changing the position of the Nernst fila- ment on this bar. 7 So III APPARATUS AND PROCEDURE In any quantitative determination of the amount of light needed to arouse a just noticeable sensation for any given group of wave-lengths there are two essential requirements: (1) We must have a means of presenting to the eye the desired range of wave-lengths free as nearly as possible from alien wave-lengths ; and (2) we must have some means of measuring the energy of the stimulus thus presented and of reducing its intensity by known amounts. The greater part of the spectroscopic and radio- metric apparatus needed to fulfil these two requirements was already in use in the Bryn Mawr laboratory when the present work was undertaken.” The description of the apparatus as modified to meet the needs of the present investigation, together with the necessary additions, is given under five headings: the source of light; the spectroscope; the apparatus for presenting the light to the eye ; the devices for reducing the intensity of light ; and the radiometric apparatus. The procedure is described under two headings: the energy measurements; and the methods of observation. A drawing showing the path of the beam of light and the arrangement of the apparatus is given in Figure IV ; a photograph of the assembled apparatus in Figure V. A. The Source of Light. The source of light was a Nernst filament operated at 0.6 ampere. This source was chosen because when properly seasoned it gives a light very constant in both intensity and radiometric composition, and at the same time sufficiently intense to permit of direct energy measurement. Its shape also well adapts it for use with the slit of the spectroscope, 1.e., the shape is such as to make it possible to utilize for the illumination of the face of the prism a relatively large part of the light emitted. When in use the filament is placed directly in front of the slit and as close to it as is possible. This placement 2For description of this apparatus see References (9) and (10). 15 16 MARGARET M. MONROE of the filament, however, presents two difficulties. In the first place the Nernst material must be heated before it will conduct the electric current. This requires that the filament be moved from its position in front of the slit prior to each period of work. In the second place the terminal wires, which are of platinum and very pliable, give little stability of position to the filament. Because of these difficulties a special mounting had to be devised which would provide for the adjustments required for the removal and precise resetting of the filament and would give the rigidity of support needed to prevent sagging or other displacement of the filament from its position in front of the slit. On this latter point it may be noted that if care is not taken that the light which enters the slit come always from the same part of the filament, variations both in its composition and intensity may occur. This mounting is shown in Figure IV. B and C are two metal arms at the ends of which are attached the terminal wires of the filament N. B and C are supported by a piece of asbestos A, which is in turn fastened to the rod D by a pin E. E serves a double purpose: by its use B and C are supported in a manner which not only provides for both heat and electric insulation, but which also allows a slight rotary movement necessary for perfect alignment of the filament with the slit. The height of D is adjustable and F is fastened by a collar to a round rod attached to the collimator arm, thus permitting move- ments of the mounting back and forth, right and left, and up and down. Around the whole is a metal housing ventilated at top and sides, but in such a manner as to remain light-proof. The filament is connected in series with a Weston ammeter graduated to 0.02 ampere; a ballast which both reduces the current and compensates for the change in the resistance of the Nernst with change in temperature; and two adjustable rheostats, one coarse, the other fine. The former is used to cut down the current to approximately the desired value and the latter to correct for the fluctuations in the line. B. The Spectroscope. A diagramatic representation of the spectroscope is shown in the drawing given in Figure IV. Si is the collimator slit; Li, the collimator lens; P, the prism; Le, the THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 17 objective lens; Sz, the objective slit. Ls to Ls is the system for focussing the light on the eye. The collimator slit S:1 is 12 mm, high, and its width can be varied by means of a micrometer screw fitted with a head graduated to read to thousandths of an inch. This slit was set at a width sufficient to allow of the radiometric measurements being made with precision, and was kept constant throughout the experiment. Lenses Li and Le are both Zeiss triple achromats, 60 mm. in diameter; the collimator has a focal length of 180 mm., the objective of 240 mm. A carbon bisulphide prism 105 mm. high, with a refracting angle of 60 degrees, was used. With the exercise of a reasonable amount of precaution to keep the CSe free from impurities and to maintain a uniform temperature in the room, this prism has given satisfaction. If the temperature is not kept constant the change of refractive index of the CSe, resulting from the change in temperature, neces- sitates frequent checking and resetting of wave-length. The objective slit is 0.342 mm. wide and adjustable in height. For the radiometric measurements a height of 10.4 mm. was used; for the work with the eye this was reduced to 1.85 mm. The greater length was necessary in order to obtain an intensity suffi- ciently high to make the energy measurement; for the eye work, however, a much smaller slit is needed in order that the image which is focussed on the eye may fall entirely within the pupil. This slit is mounted on an independent base screwed to the table in a fixed relation to the base of the spectroscope. In order that the distance of the slit from the lens Lz may be adjusted for the different focal distances of different wave-lengths, the frame on which the slit is mounted is furnished with a rack and pinion R. Lens Ls, which serves as collimator in the system for focussing the light on the eye, is mounted on the same rack and pinion so that the distance between Se and Ls (the focal length of Ls) remains always constant. In order to obtain automatically minimum deviation for all wave-lengths falling on the objective slit, the spectroscope was fitted with a special attachment for the purpose. K in Figure IV is a rod fastened to the prism table in such a position as to be continuous with the radius of the table which bisects the refracting 18 MARGARET M. MONROE angle of the prism; X and Y are two rods of equal length fastened at one end to the two arms of the spectroscope at points equidistant from the center of the prism table, and at the other to a collar Z, which is free to play back and forth along rod K. M is a micrometer screw with a graduated head which is used to move the collimator arm through the small angles needed to change the wave-length. Opposite this screw is a plunger work- ing against a spring. The collimator arm is held between the screw and the plunger so that it responds to a movement of the screw in either direction. By this attachment the prism is always turned through half the angle traversed by the collimator arm in changing the wave-length. Therefore if the prism is once set for minimum deviation for the D-line, there will also be minimum deviation for any other wave-length. That is, when the prism is set for minimum deviation, the line bisecting the refracting angle of the prism also bisects the angle made by the incident and emergent rays, hence if in changing the wave-length the angle between the incident and emergent rays be changed a given amount by a movement of the collimator arm, the prism must be moved through half that angle in order that the line which bisects its refracting angle will also bisect the angle made by the incident and emergent rays. In all quantitative work on color sensitivity it is very important. that the light employed be as homogeneous as possible as to wave- length. The presence of alien visible wave-lengths affects the determination of chromatic sensitivity in two ways: (1) It decreases the amount of the color response through physiological inhibitions and interactions, and (2) it increases the value of the energy measurements. In the work in this laboratory determina- tions made with and without provision for absorbing impurities SOUIIIYIP MOYS ‘939 ‘SUOTPOHoI [eUIOJUT ‘GYSI] poi9}jeVos 0} onp large enough to be considered significant. In the present investi- gation the aim has been to obtain a degree of purity such that any portion of the spectrum used should show only one band when examined with a second spectroscope. In order to secure purity of light the following precautions were taken. The spectroscope was provided with the minimum deviation attachment already THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 19 noted. Great care was employed in eliminating as far as possible all stray light and internal reflections. The source was housed as described above and a screen was placed at the objective lens to prevent any stray light from the prism from reaching the farther parts of the system. As far as could be all surfaces that might either admit or reflect extraneous light into the path of the refracted beam were blackened, and as a final precaution a light- tight housing was built around the whole apparatus from just in front of the collimator slit to just beyond the lens which focussed the light on the eye. This compartment was large enough to permit of the experimenter working inside. Even with these precautions, however, some impurities remained, chiefly those due to reflections from the surfaces of the lenses. These were absorbed out by very thin gelatine filters, carefully selected with reference to the bands to be eliminated. The gelatine filters were held in place over the objective slit by small clips fastened on either side of the plate containing the slit. The filters were used both for the radiometric measurements and for the threshold determinations. C. Apparatus for Presenting the Light to the Eye. ‘There are several methods, by which we may obtain a homogeneously illumi- nated surface suitable for determining thresholds of sensation: (1) Spectrum light of the desired range of wave-lengths may be allowed to fall on some diffusively reflecting surface, such as magnesium oxide, which in turn is viewed by the eye; (2) spec- trum light may be allowed to fall on some diffusely transmitting surface; or (3) spectrum light may be focussed directly on the pupil of the eye by means of a double convex lens. The third of these possibilities was chosen for this work. The use of either of the first two methods necessitates the assumption that the reflecting or transmitting surface is absolutely nonselective to wave-length. Both methods, moreover, are very wasteful of light—not only is there a comparatively high percentage of absorption, but also only a small percentage of the light coming from any point of the stimulus surface enters the pupil of the eye. Under these conditions it would be very difficult to specify accu- rately in radiometric units either the unit density or the total MARGARET M. MONROE KH Wt 9. § | $4" 40g Pare etree EPEC VED EERE EOCENE \ THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 21 amount of the light at the eye. That is, the energy value at the eye could not be measured; it could be calculated only approx- imately. An accurate energy specification is possible, however, when the light is focussed on the eye. The light, by this method, does not spread from the stimulus opening as if emanating from a source, but is concentrated into an image on the pupil of the FiGuRE V eye—an image of the objective slit of the spectroscope. The amount of energy concentrated into this image can readily be determined with the radiometric apparatus to be described later. In order to present the stimulus in compliance with the above plan, the rays of light emerging from the objective slit are first rendered parallel by lens Ls (Fig. IV) placed at its focal length (150 mm.) from the objective slit, and are then focussed on the 22 MARGARET M. MONROE eye by lens La, focal length 275 mm. In this way ample space is obtained for the introduction into the system of any apparatus necessary for reducing the intensity of the beam of light, 1.e., sectored disc, filters, etc. The dimensions of the image at the eye were determined by photographing it on a plate carefully mounted in the plane of the pupil and measuring the photo- graphed image with a micrometer comparator. These dimen- sions were 3.7326 mm. X 0.6956 mm. __ Since this is well within the pupil of a dark-adapted eye, the use of an artificial pupil with its attendant difficulties is avoided. In front of Li, 240 mm. from the eye, a screen G is placed containing a circular opening O, 10 mm. in diameter, which serves to diaphragm the lens Ls to the desired stimulus aperture. When the eye is in position the exposed area of the lens Ls is seen filled with light. The visual angle subtended by this area is 2° 2.2’. D. Means of Reducing the Intensity of Light. Because of the small amount of energy required to arouse a just noticeable sen- sation of light, it would be impossible to measure this energy | directly with a thermopile. It is necessary, therefore, to measure the energy at a high intensity and to reduce this intensity by known amounts to the intensity required. In order to obtain the range of reduction necessary for all observers, three methods of cut-down were utilized. 1. The Filters. The main reduction of light was made by the insertion in the light beam of neutral filters, combinations of which allowed for a possible range of transmission of from 1148 « 10° to 141 « 10°. The filters are of gray gelatine mounted between glass, 25 mm. by 25 mm. They were made by the Eastman Kodak Company. The densities were specified by the Eastman Kodak Company and the transmission of each filter was calculated by the formula: 1 Density = log Transmission A holder designed to take any combination of filters up to eight was placed at H in contact with the screen G. This position THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 23 of the filters precluded the possibility of any scattered light not completely absorbed being transmitted through the screen opening to the eye of the observer. 2. The Sectored Discs. Any further large reduction needed was obtained by the use of a pair of sectored discs inserted at SD. These discs were rotated by a small motor suspended from above by coiled springs to absorb vibration. The range of open sector used was from 180° to 0°. 3. The Wedges. The final reduction was made by means of the two wedges WW mounted immediately in front of the objec- tive slit. Early in the work it was realized that for any accurate threshold determinations a means must be had of making very small changes in the intensity of the stimulus, and that the method employed must insure a perfectly uniform reduction throughout the cross section of the beam of light. A single wedge such as is ordinarily employed does not give this required uniformity of reduction. When placed in front of the slit, no matter how fine the gradation in density, there is always a slight difference in transmission between the opposite edges of the used portion of a single wedge. A double wedge device was therefore planned to obviate this difficulty. The two wedges were made according to specification and calibrated by the Eastman Kodak Company. Like the filters, they are of neutral gelatine mounted between glass. The wedges are identical, each being 135 mm. by 13.5 mm., and so constructed as to cover the same range of transmis- sion. They are mounted parallel to each other in holders which are operated by a micrometer screw. These holders are provided with right- and left-handed threads. As the screw is turned the wedges move in opposite directions, each wedge traveling in front of the other through a path equal, if need be, to twice the length of one wedge, that is, from a position of juxtaposition at the thin end of each wedge to a position of juxtaposition at the thick end. Since the density gradients of the two wedges are identical and the wedges move in opposite directions, it is obvious that the resultant densities will be the same from point to point throughout the overlapping section. A consideration of the range of move- ment shows further that a series of densities may be obtained 24 MARGARET M. MONROE varying by minute amounts from the sum of the minimum densi- ties of both wedges, through the sum of the minimum of one and the maximum of the other to the sum of the maxima of both. 1 The density in this case also equals the log of Transmission E. The Radiometric Apparatus. The apparatus used con- sisted of a linear thermopile of silver and bismuth couples, a Paschen small coil galvanometer especially constructed for the thermopile employed, and suitable auxiliary apparatus. These instruments were constructed by W. W. Coblentz of the Radio- metric Division of the Bureau of Standards. The apparatus has been described in an article by Dr. Ferree and Dr. Rand (10), and the reader is referred to this article for further details. All the radiometric measurements were made by Dr. Rand, to whom I am deeply indebted for the values given below. The procedure of making these measurements is quoted from the article just. mentioned. F. The Energy Measurements. The apparatus for measuring the energy is so planned that measurements may be made at the objective slit, at the stimulus opening, and at the eye. A descrip- tion at one of these places, namely, the objective slit, is sufficient to show in a general way the method employed. “The thermopile to be used was placed in position immediately behind the slit and a blackened aluminum shutter was interposed in the path of the beam of light between the slit and the end of the objective tube of the spectroscope. Preliminary to the exposure of the thermopile to the light to be measured, -the current sensi- tivity of the galvanometer was tested by means of a special device provided for this purpose in the construction of the galvanometer. With regard to this procedure it may be pointed out that the current sensitivity of the galvanometer varies with the period or time of the single swing of its needle system. Since it is not possible to control the field so as to get this period always the same, it is necessary, if results are to be compared, to take some sensitivity as standard and to convert all readings into deflections THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 25 for the standard sensitivity by means of a correction factor deter- mined at each sitting. (For a detailed description of the method of determining this factor, see PsycHoL. Rev. Monoa., 1917, 24, No. 2, pp. 60-65. ) “The thermopile was next connected with the galvanometer and the light allowed to fall on its receiving surface until a temperature equilibrium was reached (ca. 3 sec. for our thermo- pile). The deflections were read by means of the telescope and scale and the readings are corrected to standard sensitivity by means of the factor previously determined. The final step in the process of measuring was the calibration of the apparatus, 1.e., the value of 1 mm. of deflection in radiometric units was deter- mined for the area of thermopile exposed. To do this a radiation standard, the value of the radiations from which is already known, had to be employed. The standard used by us was a carbon lamp specially seasoned and prepared for the purpose by W. W. Coblentz. This lamp was placed on a photometer bar 2 meters from the thermopile and operated at one of the intensi- ties for which the calibration was made, in our case 0.40 ampere. The thermopile was exposed to its radiations with the same area of receiving surface as was used in case of the lights measured, and the galvanometer deflection was recorded. From the deflec- tions obtained the value of 1 mm. of deflection, or the radiation sensitivity of the apparatus under the conditions given, was com- puted from the known amount falling on the surface of the thermopile. Having the factor expressing the radiation sensi- tivity of the apparatus, the deflections produced by the wave- lengths of light measured were readily converted into energy units.” The radiation sensitivity of the linear thermopile as used in the present investigation was computed from the following data. The energy value of the radiations per square millimeter of receiv- ing surface from the standard lamp at a distance of 2 m. operated at 0.40 ampere was 90.70 & 10°° watt. The deflections of the galvanometer produced by this intensity of radiation falling on the same area of receiving surface as was used in measuring the lights employed as stimuli, when corrected (a) to a sensitivity 26 MARGARET M. MONROE of i 1 10°'° ampere, and (b) for the absorption of the glass cover of the thermopile, was 323.85 mm. The area of the sur- face exposed was 3.5657 sq. mm. The value of 1 mm. of gal- vanometer deflection, or the sensitivity of the instrument for the area of receiving surface used, was, therefore, 998 & 10-™ watt, By means of this factor the galvanometer readings produced by the different wave-lengths of light were readily converted into the . energy value of light falling on the receiving surface of the thermopile. For the purpose of the present investigation, how- ever, it is needed to know also the energy values of the light entering the eye. These are sufficiently great only in the case of red and orange to be measured directly with the required pre- cision. It was necessary, therefore, to measure all the wave- lengths used at the objective slit and only the red or orange at the eye, and from the comparative values of the red or orange at the two places to determine a correction factor which will repre- sent for all the colors the reduction of the light from objective slit to eye. In order to determine this reduction factor with precision a larger area of receiving surface of the thermopile had to be exposed to the light than the actual area of the image enter- ing the eye for the threshold determinations. It will be remem- — bered that the height of the objective slit and consequently the height of the image focussed on the pupil of the eye was adjust- able. In the present work the receiving surface of the pile used to measure the light at the eye was 11.2548 & .895 mm., or 10.073 sq.mm. The value of 1 mm. deflection of the galvanometer for this area of receiving surface was 88 < 10°'° watt. Since the focussed image is of uniform density, the energy for the area of the image at the eye, 3.7326 mm. & .6956 mm., or 2.5967 sq. mm., could be readily determined. From these values a single factor was calculated which would convert the amount of energy for the different wave-lengths measured at the objective slit into the amounts of energy entering the eye. These values are given in Table V. A division of these values by 2.5967 will give the energy density at the eye, a division by 78.54 the energy density at the stimulus opening. THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 27 TABLE V Red 655 uu = 300 xX 10-9 Orange 616 wp = 133 yy Yellow 580 npu= 16.2 Zn Yellow-green Shey Pees 4g Green 522 uu= 14.8 nf Blue-green 489 pu= 9.86 “ Blue A653 op O19 G. The Methods of Observation. The experiments were con- ducted in a light-tight dark-room, the walls, floor, and ceiling of which were painted black. In the preliminary work it was found that an adaptation period of twenty minutes was sufficient to give constant results in the determination of chromatic thresholds. For the achromatic thresholds a much longer adaptation period was needed. The observer was seated in front of the apparatus, the eye being at the focal length of the lens Ls. To ensure steadi- ness of position the head was held rigid by means of a wax- coated mouthpiece in which the impression of the teeth had previ- ously been made and hardened. The other eye was lightly ban- daged with a black cloth. The determinations were made in ascending and descending series. The edges of the stimulus opening were touched at suitably spaced intervals with luminous paint to enable the observer to take and hold the correct fixation. A rough adjustment of the filters and sectored disc was made until an approximate value of the threshold was obtained, and then a rest period of several minutes was given and the exact value of the threshold accurately determined. In order to prevent a pro- gressive loss of sensitivity from fatigue short rest periods were given after each observation. A number of independent deter- minations were made of the threshold value for each wave-length. A similar procedure was employed for the determination of the achromatic threshold for each wave-length. In this case, how- ever, even greater care had to be exercised to guard against progressive loss of sensitivity. IV STATEMENT AND DISCUSSION OF RESULTS A. Achromatic Thresholds. The achromatic thresholds of the seven wave-lengths were determined for twenty-one observers. A complete statement of the results is given in Table VI, Parts A and B. A graphic analysis of Table VI is given in Figures VI-XIV. The minimum visible for the different wave-lengths used was found to be as follows: 1. Red (655 up) Average = 626.99 watt X 10-16 Median 6. + 37 .67 1209.27 — 186.48 ¥ Range = 2. Orange (616 up) Average= 130.28 < Median = 118.08 4; Range == 273.77 — 70.72 $ 3. Yellow (580 up) Average 27.96 € Median = 25.74 pe Range = 49.04 — 11.56 . 4. Yellow-Green (553un) Average 2.203 . edian == 2.298 © Range) )==) 3242 —'), af718 7 5. Green (522 up) Average= 1.953 Median = _ 1.868 oe Range = 3.469— 1.295 , 6. Blue-Green (489 up) Average=— 8.11 * Median = 8..00 Hs Range = 13.8 — 2.96 7 7. Blue (463 pu) Average= 15.62 + Median = 15.00 > Range = 29.53 — 6.28 +: The total range of minimum visible of the wave-lengths used is therefore from 1209.27, in the red, to .718 (watt X 107%*) in the yellow-green, a ratio from highest to lowest of 1684. The average sensitivity of the twenty-one observers was great- est in the green. 1.953 (watt X 107°). The average threshold for this wave-length is There is not, however, complete uni- formity as to the position of maximum sensitivity. Nine observ- Part A of ers show a maximum sensitivity in the yellow-green. Table VI gives the results of those whose maximum sensitivity 28 THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 29 TABLE VI ACHROMATIC THRESHOLDS Amount of light entering the eye (watt X 10-16) Yellow- Blue- No. Observer Red Orange Yellow Green Green Green’ Blue 1. M.M.M. 922.41 103.53 45.74 2.208 1.405 9.58 15.23 2a 787.50 192.70 11.56 3.145 1.981 10.97 25.95 3. Md.B. 753.63 70.72 44.61 2.596 1.471 11.25 16.37 4. M.B. 689.28 91.14 32.13 2.107 1.295 3.83 8.16 S.C. 667.50 129.52 35.06 3.242 2.446 7.16 19.55 Gavel: 648.83 A I ds 59.50 3.220 1.868 8.00 11.06 » ola Bal ASE 637.67 273.77 25.74 2.665 lee 7.74 16.00 8 3.G.L. 626.22 106.24 31.08 2.586 2.116 8.46 14.14 9. M.O.L. 613.90 136.22 21.94 3.024 1.939 7.65 23.302 LUN Ce Gea 6 609.84 123.17 49.04 2.512 2.268 10.50 18.31 135 T WL. 498.96 141.90 23.08 1.661 1.513 6.04 20.56 12. Md.B.L. 461.16 115.64 28.73 2.298 2.059 9.91 8.90 Mean— 659.74 141.76 32.02 2.605 1.840 8.42 16.46 Median—A 643.25 126.35 31.61 2.591 1.904 8.23 16.18 ; B 13.:J.G: 1209.27 118.08 25.56 2.480 3.450 8.26 15.06 14. T.W. 787.20 176.08 21.83 1.293 1.353 eT 13.18 15. M.MC. 642.33 108.10 32.06 2.264 2.539 5.07 11.04 163 Ei: 641.28 88.20 20.74 818 1.548 6.36 12.20 17. M.G.R 597.00 119.36 21.14 2.203 3.060 13.80 23.19 18. M.O. 593.45 98.42 29.52 2.550 3.469 10.66 29.53 19. R.N. 305.71 135.42 14.38 1.622 1.718 7.79 13.00 AY. WO Le 287.30 79.72 18.18 718 1.395 2.95 7.07 Zipetivks 186.48 111.55 19.66 1.066 1.416 3.10 6.28 Mean—B 583.34 114.99 22.56 1.668 2.216 7.68 14.51 Median—B 597.00 PEE5S 21.14 1.622 1.713 7.79 13.00 Mean—A and B 626.99 130.28 27.96 2203 1953 811 15.62 Med.—A and B_ 637.67 118.08 25.74 2.298 1.868 8.00 15.00 lies in the green, Part B the results of those whose maximum sensitivity lies in the yellow-green. With two such distinct types a single measure of the general tendency of the group is mislead- ing—the mean and median of the two groups were therefore calculated separately. As shown by these measures, the two types are quite different not only with regard to the position of maximum sensitivity but also with regard to the absolute value of the thresholds. With the exception of the green, the average threshold values of the yellow-green type are smaller than those of the green type. There is, however, on the whole, a wider range of values in the yellow-green type than in the green type. In order to show more clearly the difference between these two groups, as well as to show the relative average sensitivity through- out the spectrum, the results of Table VI, A and B, were plotted 30 MARGARET M. MONROE in the form of sensitivity curves. The reciprocal of the energy value was taken as a measure of the sensitivity. The maximum sensitivity for each observer was then made equal to unity and the other six values for the observer calculated as ratios of that value. Figure VI is the composite of the twenty-one sensitivity curves. In this composite we can distinguish not only the two groups Achromatie Sensilivily Sada ee a oy Filla SS Wate Wave Length 655 b 16 580 553. 522 9463 Ficure VI . under discussion, but a possible third group. Two observers, although belonging to the type having maximum sensitivity in the yellow-green, are also very sensitive to green, as is indicated by the great breadth of the curve between wave-lengths 522 pu and 553. Similarly a few observers belonging to the type having maximum sensitivity in the green are very sensitive in the yellow-green. It is possible that the true maximum in both these THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 31 cases may be neither green nor yellow-green, but some intermedi- ate wave-length. If the investigation in hand had been for the purpose of determining the exact shape of the threshold visibility curve for all types of observers, the region between 522 pp and 553 u» would have been carefully explored for all cases differing from the average in this way. The purpose of the study was, as has been stated, the determination of the minimum visible at 1.0 8 : | a | 6 : = al = { 5} « rs H c : a” ‘ f ep Meet .f = : f — Average Tw 6 h : 3/5 \ |i ----Median be : <— \ #\< ‘ J Z ee” Wave Length 2 655 er zig 580 §&53 522 489 463 pH Ficure VII certain representative points throughout the spectrum. The pos- sibility of a third type is offered, therefore, merely as a suggestion as yet unproved. The mean and median values of the two groups were also calculated in the form of ratios of sensitivity. Figure VII gives these average sensitivity curves. The curves for the median and mean values agree very closely for the green type, and also for the yellow-green type except in wave-length 522 py. 32 MARGARET M. MONROE For all the wave-lengths employed the distribution of values around the average is approximately symmetrical. The fre- quency graphs for the seven parts of the spectrum used are given in Figures VIII-XIV. The size of the class interval for the vertical codrdinate, representing the number of observers, is the same in all the curves. That along the horizontal coordinate differs according to the absolute value of the minimum visible in the particular color. Thus in red, where the value of the minimum visible is large, the interval is 400 (watt x 10°), 16 4 12 “a os f @& UNS Bites wo n =) §}° — o = 6) 5 = > on 4 2 -/6 ) Energy -(Walt x 10 | 1250 850 450 50 Ficure VIII. Distribution of threshold values for red, 655 wn (Table VI) while in green it is only .8 (watt X 10°°*). The intervals were further chosen so that the median threshold value of the group in question would fall approximately at the center of some interval. Since the number of observers was small, it was thought that a truer picture of the distribution of values was obtained by the use of rather large intervals. As has been said, the frequency graphs are very nearly symmetrical—each shows a large average group and two smaller, almost equal groups, .one superior, one THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 33 inferior to the average. Since this is true even for a small group, it seems reasonable to suppose that the same would be Number of observers Energy - (watt x107!®) 216 184 q ° Ficure IX. Distribution of threshold values for orange, 1616 uu (Table V1) Number of observers Energy-C Watt 107") 50 34 13 2 Ficure X. Distribution of threshold values for yellow, 580 yu (Table VI) 34 MARGARET M. MONROE true for normal observers as a whole. In other words, it is probable that there is no uniform “normal”’ sensitivity to light of different wave-lengths, but a very wide range of sensitivity within which an individual threshold may be considered normal. B. Chromatic Thresholds. It was originally hoped to deter- mine the chromatic and achromatic thresholds of the seven wave- Number of observers ™N Energy- (Watt x107!®) 4.6 3.2 1.8 A Ficure XI. Distribution of threshold values for yellow-green, 4553 us (Table VI) Number ot observers Energy-(Watt x 107!) 3.5 2.5 1.5 5 Ficure XII. Distribution of threshold values for green, A522 uu (Table VI) THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 35 lengths for the same observers. Most observers, however, were unable to give the time required for this, and comparison is pos- sible, therefore, only between averages. The values of the chromatic thresholds obtained are given in Table VII. Since in only a few cases did the same observer determine the chromatic threshold for all seven wave-lengths, each column of Table VII gives the twenty-one values obtained arranged not according to nn — qa > x. fob) wo oe mej o i °o a 1 & © = => = ; Energy- (Watt x 1o7!*) 14 10 6 2 Ficure XIII. Distribution of threshold values for blue-green, \ 489 pu (Table VI) Number of observers 16) Energy- (Walt x 10° 30 20 10 0 Figure XIV. Distribution of threshold values for blue, 4463 uu (Table VI) 36 MARGARET M. MONROE observers, but according to order of magnitude. The results are presented graphically in Figures XV—XXIII. The minimum visible—chromatic—for the different wave- lengths used was found to be as follows: 1. Red (655 up) Average __.1778 watt X 10-12 Median = ..159 ~ Range = .460 — .0597 “ 2. Orange (616 up) Average .562 a? Median = __.166 A Range = 3.11 — .054 3 3. Yellow (580 up) Average 9.57 Median = 4.32 Range =52.8 —.190 . 4. Yellow-Green (553 un) Average .0856 . Median — _ .0127 a Range = _ .621 — .00771 fe 5. Green (522 uu) Average .143 5 Median = .0436 re Range = _ .603 — .0042 - 6. Blue-Green (489 un) Average _ .643 3 Median = __ 460 - Range = 3.16 — .0299 5 7. Blue (463 uu) Average __—.812 x Median = _ .329 e Range = 3.07 — .0298 og TABLE VII (CHROMATIC THRESHOLDS Amount of light enternig the eye (watt < 10-12) Yellow- Blue- No Red Orange Yellow Green Green Green Blue 1 460 Zire Ol 52.8 .621 .603 3.16 3.07 2 .370 2.44 49.5 413 .587 2.91 2.54 3 316 1.84 38.9 9 .504 986 2.48 4 301 833 1333 0968 401 925 Dae 5 228 820 8.57 .0959 398 714 bP 6 188 522 7.31 .0870 0859 48 1.14 7 171 432 7.07 .0268 0684 504 814 8 168 242 6.09 0148 0493 482 566 167 227 4.95 0142 0474 473 433 10 163 170 4.61 0130 0449 460 355 11 159 166 4,32 0127 0436 460 329 12 155 163 998 .0125 0429 366 284 13 129 149 896 .0119 0420 351 274 14 127 135 596 .0117 0344 343 257 15 117 118 546 .0108 0218 321 252 16 101 0931 519 .0107 00818 233 17 100 0859 516 .0100 00628 0477 213 18 0999 0786 516 00959 00497 0473 0574 19 0832 0667 463 00890 00449 0404 0425 20 0715 0568 461 00853 00420 0321 0392 21 0597 0540 .190 00771 00420 0299 0298 Mean .1778 .562 9.57 .08555 .14314 643 .812 Med. — .1590 . 166 4.32 01270 .04360° .460 329 THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 37 The total range of the chromatic minimum visible of the wave- lengths used is therefore from 52.8 in the yellow to .0042 (watt < 10°?) in the green, a ratio from highest to lowest of 12571; The average chromatic sensitivity is greatest in the yellow- green. The average threshold value for the yellow-green is Average — Median --- FIGURE XV 08556 (watt x 10°*). The relative sensitivity to the other wave-lengths is best shown by Figure XV, which gives the mean and median values in the form of sensitivity curves. These curves were plotted in the same way as those in Figure VII. In both the mean and median sensitivity curves there is an extremely sharp drop from the yellow-green to the yellow. Since the chromatic thresholds for all colors were not determined throughout on the same observers, a composite of the individual sensitivity curves, similar to the composite of achromatic sensi- 38 MARGARET M. MONROE tivity curves could not be made. All seven thresholds, however, were determined for six observers, and a composite of their sensitivity curves is given in Fegure XVI. There is much more variation in these curves than in those representing achromatic sensitivity. Four only have a maximum sensitivity in the yellow- green; of the other two, one shows a maximum in the green, the other in the blue. The relative sensitivity to the longer wave- 1.0 mS Chromatic Sensilivily lengths is more uniform—all six curves show the drop in the yellow, and all are fairly close together in the orange and red. * For the observers tested, the distribution of chromatic threshold values around an average is not symmetrical. In each color the threshold value of the median is much smaller than that of the corresponding average—that is to say, the curves are skewed to a THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 39 the upper end. This means that the range of values below the median is much greater than that above the median. For instance, there is a difference of 48.48 & 10°” watt between the Number of observers Wx 10 ~!2 W x 1907'3 Figure XVII. Showing the threshold values for red, 4655 uu (Table VII) an” Kas aw > 5 ae wn 2 (ae) S Lee aw pt ew ] e —_ z WX 10TH WX 10TH W107 Figure XVIII. Showing the threshold values for orange, \ 616 uu (Table VII) 40 MARGARET M. MONROE median and the largest threshold in yellow, while there is a differ- ence of only 3.13 between the median and the smallest threshold value in the same color. Because of this very wide range it was thought advisable to plot the logarithms of the energy values of Number of observers WES 10 5fe (NX EO A WX 107! Figure XIX. Showing the threshold values for yellow, 4580 un (Table VII) an ft @ > . au 1 a i=) Co oa u wy _ f= | Zz WX 107% WW. X10T Wx 107% Figure XX. Showing the threshold values for yellow-green, \553uu (Table VII) ‘ THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 41 the chromatic thresholds rather than the values themselves. Figures X VII—XXIII shows the frequency graphs of the seven wave-lengths. Along the ordinates is plotted number of cases. As in the curves for achromatic sensitivity, the class interval Number of observers WX 1072 WXI073 Wx 107! Figure XXJI. Showing the threshold values for green, 4522 uy (Table VII) 2 c wy > , = @ wv” 2 ° cio o ae qw 2 e > 2 W.X 107" WX 107% W.X 1078 Figure XXII. Showing the threshold values for blue-green, \ 489 wu (Table VIT) 42 MARGARET M. MONROE for the abscissae differs for the wave-lengths. Thus the first interval of the abscissa in blue-green, blue and orange contains all cases of the order 10°" watt, the second interval all cases of the order 10° watt, and the third all cases of the order 107% watt. In green and yellow-green the intervals contain cases, of the orders 10”, 10°, and 10°“ watt, respectively, and i in the yellow 10°*°, 10°", and 10°?” watt. It will he noted that the graphs of red and yellow differ from those of the other five wave-lengths by showing skewness even ey | 10 Me = wv > g p a ce) _ 6 oS a co) i ioe) + —< & > 2,2 Energy W. X10 7" WX {O72 WX 107!3 Figure XXIII. Showing the threshold values for blue, \ 463 uu (Table VII) in their logarithmic form. In red there is no inferior group— the values fall into a large average group and a smaller superior group. In yellow there is no superior group, but there is an additional very inferior group. It might seem reasonable to account for this on experimental grounds because of the differ- ence in the abruptness with which the chromatic component comes into the sensation in the two cases. The chromatic threshold for red is by far the easiest to determine. The transition from achromatic to chromatic is abrupt and sharply marked and there is little or no hesitation on the part of the observer as to whether or not there is any color present. The ease of judgment resulting from this small and clearly defined transition interval probably ae THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 43 accounts for the absence of the inferior group. In yellow just the opposite condition obtains. The color comes in very slowly and gradually and the exact point of appearance of the chromatic component is difficult to determine. At low intensities light of wave-length 580 wz has much the appearance of ordinary artificial light, such as that emitted by a carbon lamp, which an unpracticed observer is accustomed to consider as white. In such a case much greater intensities are needed for the light to be called colored. This may account for the additional very low group in yellow not found in any other of the colors. C. The Photochromatic Interval. By the photochromatic interval is meant the colorless interval between the chromatic and the achromatic thresholds. In the present study the actual energy of both thresholds has been measured. This makes it possible to express the value of the photochromatic interval in absolute terms—terms that admit of a numerical comparison from wave- length to wave-length. Previous to this the photochromatic interval has been discussed only in the most general terms. Its existence, and the fact that it varied under different conditions, were noted by Tschermak (11), Fick (12), v. Kries (13), Her- ing (14), and others. There has been only one numerical state- ment of the value of the colorless interval between the two sets of thresholds, that of Charpentier in 1888 (15). Charpentier focussed light from the spectrum of the sun on ground glass, using the colored surface thus obtained as the stimulus for his determinations. The intensity was reduced by a diaphragm. No measurement, either photometric or radiometric, was made of the actual intensities used. The relation between the chromatic and achromatic thresholds was taken as equal to the ratio of the square of the diameters of the diaphragm openings necessary for the two thresholds at any given wave-length. The diameter of the open- ing required to reduce light in the yellow to the achromatic threshold was 1, the opening for the chromatic threshold of the same color was 3.1. The chromatic threshold is then, according to Charpentier, 9.6 times as large as the achromatic threshold. The ratios found are as follows: 44 MARGARET M. MONROE Rouge "extremes... Sere ek eee eee re 3.6 Orange sts sey. elles haar ene ib areas fee arene 5.5 Jarre ys is vs ce ates « Le een Oe ag el tame ina 9.6 Vert (mMoyennsles t's sae peige hase eingeiteree eee ee 196.0 Blew franc, région moyenne: sa... ss.01- sence. 635.0 The wave-length is not stated. Such ratios, obtained without reference to the absolute or even the relative intensity of the lights used, have little value other than to indicate that there is a photochromatic interval. . In 1892 Abney and Festing (16) determined the achromatic and chromatic thresholds throughout the spectrum and attempted to give a photometric evaluation to the results obtained. A mono- chromatic beam and a comparison white beam were so reflected as to fall on adjoining portions of a white screen. The chromatic thresholds were determined by reducing the intensity by means of a sectored disc introduced into the path of the monochromatic beam. Relative luminosity values were calculated by the use of a spectrum luminosity curve previously obtained. Relative lumi- nosities, however, change with change in intensity of light, and this curve, while a low intensity curve, was not determined for threshold intensities. Achromatic thresholds were similarly determined, the relative luminosities of the different wave-lengths being calculated from the sectored disc values and the luminosity. curve just mentioned. Abney and Festing did not, however, attempt to assign any values, either relative or absolute, to the photochromatic interval as such, and since their values for the thresholds are only relative, it is not possible to calculate it from their data. The values of the photochromatic interval for the seven wave- lengths used in this study are shown in Table VIII. Column 2 gives the energy value of the minimum visible achromatic, column 3 that of the minimum visible chromatic, and column 4 the differ- ence between the two. In column 5 the values of column 3 are given in the form of ratios, the value for yellow being made equal to unity. The values are very large, ranging from 853.29 & 10°* watt in the yellow-green to 95672.04 1077® watt in the yellow. There is, apparently, a great difference between the development of the color sense of the eye and that of the light sense. THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 45 TABLE VIII Wave Achromatic T. Chromatic T. Difference Ratio length (watt X 10-16) (watt 10-16) (watt X 10-16) Yellow=1 655 ue 626.99 1778 .00 1153.00 01 616 up 130.28 5620.00 5489.72 .06 580 uu 27.96 95700.00 95672.04 1.00 553 um 2.203 855.50 853.30 .009 522 pp 1.953 1431.40 1429.45 .02 489 uy 8.11 6430.00 6421.89 .07 463 wp 15.62 8120.00 8104.38 .08 The value of the photochromatic interval throughout the spec- trum is represented graphically in Figure XXIV. Along the 106,000 99000) | | } gaged | ! 14,000 60,000 50,000 40,000 Energy (Watlx 1079) 38,000 20,000 10,000 Wave Lenath 655 616 580553 522 409. $63 ee Ficure XXIV. Showing value of the photo-chromatic interval (Table VIII) vertical coordinate are plotted energy values (watt X 10°"). The resulting curve bears to some extent an inverse relation to the curve of chromatic sensitivity. Thus, the greatest chromatic sensitivity being in the yellow-green, the smallest photochromatic interval is in the yellow-green; similarly the largest photochro- 46 MARGARET M. MONROE matic interval is in the yellow, to which there was the least chromatic sensitivity. D. Comparison with Previous Determinations of the Threshold Visibility Curve. In section II is given a summary of various investigations that have been made of the relative sensitivity of the eye to lights of different wave-lengths. In no case was the apparatus and procedure identical with that employed in the present study, but it is, however, of interest to make the general —Ebert —= Langley ane Koemg —— Green Type —- Yellow-green Type Achromatic Sensitivity 190 100 650 606 560 500 450 +00 rr Figure XXV. Comparison of results of threshold visibility curve by various observers comparison between such curves and the curves presented here. Figure XXV shows graphically a comparison of the results of Ebert, Konig, Langley, and the two types found in this experi- ment. In each case the maximum sensitivity was made equal to unity. The curve for Ebert represents the average relative sensi- tivity of two observers. It will be remembered that these two sets of results were very similar. Konig also gave values for only two observers, and these, too, have been averaged here, since they show comparatively small differences. The curve for Lang- ley is plotted from the results of one observer. The range of maximum of the five curves is from A 505 to 553 pu. Koénig’s THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 47 curve is shifted toward the blue end of the spectrum; Langley’s curve, very similar to that of the “ yellow-green type’”’ found here, is shifted toward the red end. Ebert’s curve and the curve of the “ green type ” fall between these two extremes. So differ- ent were the conditions of experimentation in each case that it would be futile to attempt to analyze disagreements. Difference in source, difference in number (and probably type) of observer, difference in methods of energy determination—any one of these would seem to be sufficient to account for the variation found. Indeed, it is surprising that there should be as much agreement as there is. It will be noted that the results of Pfluger are not represented on this composite graph. The individual and diurnal variations in Pfluger’s results were so large as to make averaging for the purpose of such a comparison impossible. The position of the maximum according to him ranges from A 495 to 524 up, and this deviation is found even in the results for a single observer—EI. A portion at least of this extreme variation can be explained. In the preliminary work of the present study the same difficulty obtained. It was soon found, however, that much of this diurnal fluctuation could be traced to physiological factors which could be controlled—at least in part. A cold, lack of sleep, or general fatigue changed the character of the results greatly—not only was the absolute sensitivity lessened, but the relative sensitivity was altered. Table IX gives the threshold values for the same observer taken on different days—once when much fatigued, the other when rested. In many colors the difference amounts to several hundred per cent. Besides being very insensitive, a fatigued observer is erratic, often makes entirely contradictory judgments, and is much more subject to troublesome after- images and idio-retinal light. Curiously enough, general fatigue was found to be more disturbing than eye strain. As has been stated in the section on apparatus and procedure, before making the final threshold determinations each observer was carefully questioned and a few trial settings made as a check on fatigue. It is believed that the influence of this factor on the final results is negligible. 48 MARGARET M. MONROE TABLE IX Yellow- Blue- Condition Red Orange Yellow Green Green Green Blue Fatigued 2257 .94 676.25 48 .67 4.618 4.558 16.00 19.04 Rested 305.71 135.42 14.38 1.622 1.718 7.79 13.00 E. The “ Minimum Visible.’ In the work described in the preceding chapters the amount of light required to arouse both the chromatic and the achromatic response in different parts of the spectrum, the minimum visible for those wave-lengths has been measured directly in energy terms for the first time. As was pointed out in the historical section, the determinations of comparative sensitivity by Ebert, Pfluger, and others were made only in relative terms. Absolute values were not assigned to the light intensities on which the values were based. Moreover, in all cases but one the galvanometer deflections used in compiling the ratios were not obtained with the stimulus actually used in producing the eye’s reaction. Following a different line of development, however, attempts have been made to calculate the energy value of the minimum visible. The photometric and radiometric data used in making these calculations, however, were assembled from different sets of observations and experiments. One of the first of such estimates is that of Wien (17), who in his dissertation, ‘‘ Ueber die Mes- sung von Tonstarken,’ 1888, sought to compare the absolute sensitivity of the ear with that of the eye, his data on visibility being taken from the observations of the astronomers. He assumed that brightness of stars of the sixth magnitude could be taken roughly to represent the limit of visibility. By a comparison of available photometric and radiometric data he estimated the light from those stars to have an approximate value of 4 X 107% watt. Drude (18), some eleven or twelve years later, calculating also from stellar data, obtained a smaller value for the minimum visible, 6 X 10°'® watt. He assumed the brightness of a star of the sixth magnitude to represent the limit of visibility, a brightness which he estimated to be equal to that of the Hefner lamp at 11,000 meters. Angstrom (19) had determined experimentally the energy value of the Hefner unit of illumination (the light THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 49 emitted in a unit solid angle, the lumen) to be 8.1 & 10* erg/sec. The relation of the unit of illumination to the unit of energy or power Drude called the mechanical equivalent of light. This relation of the photometric to the energy unit has been much employed by later investigators in attempts to’arrive at approx- imate energy values from measurements made in photometric terms. The energy value of the lumen thus, in terms of the Hefner standard, would be 8.1 X 10* erg/sec., and the intensity of the illumination from a Hefner lamp at 11,000 meters would equal 1 X 10° lumens per square meter. Assuming further a pupillary opening of 3 mm., Drude calculated: Silex 10" 407 ==.6 X10 erg/sec} = 6 X 1071° w. This value of Drude has recently been recalculated by Cob- lentz (20), who had himself measured the total radiation of a Hefner lamp. Coblentz found the light density at 1 meter to be 7 X 10° g.cal. (or 29 & 10° w) per cm? per second. With this value the minimum visible would be: 29 1Oe 07 =~ 17 >< 10°'* w. . 11000? Another calculation of the least radiation visually perceptible from astronomical data is that of Ives (21). Ives, too, accepted the brightness of stars of the sixth magnitude as representing the minimum visible, but recalculated the photometric value of this brightness on the basis of later data on the relation of stellar magnitude to the candlepower scale. He also used a different value of the mechanical equivalent to convert the photometric into radiometric terms. Drude had based his determination of the mechanical equivalent on the lumen as representing the total radi- ation in a solid angle from the Hefner standard. Ives used the lumen to represent the photometric unit of radiation from the region of the visible spectrum to which the eye is the most sensi- _ tive, taken from previous determinations of the relative sensitivity of the eye to wave-length. These determinations, however, were not made at threshold intensities. If the minimum visible be taken to represent the least amount of radiation visually percep- 50 MARGARET M. MONROE tible, of wave-lengths to which the eye is most sensitive, the lumen selected by Ives is the more compatible with the purpose of the problem. However, the light selected by Drude on which to base the ratio of the photometric to the radiometric unit is more nearly of the same composition as the light used to produce the eye’s reaction. Both methods are in error, but on different points. Obviously if the minimum visible is to be determined in absolute units for the wave-lengths to which the eye is most sensitive, these wave-lengths should be used to determine the threshold of sensa- tion and the energy value of the light should be measured directly. Using .00159 as the lumen value in watts, Ives made the following calculation: 1 meter-candle = 1 lumen per sq. meter == 0.0001 lumen per ~ sq. cm. == .000000159 watt per sq. cm. == 1.59 ergs per sec. per sq. cm. The meter-candle value of a star of the sixth magnitude had been found by Russell to be 0.849 x 10°. Ives, adopting this value, obtained 1.59 X 0.849 X 10° w= 1.35 X 10° erg/sec. Assuming a 6 mm. pupillary opening instead of a 3 mm. dasa the minimum visible would equal 38 X 10° erg/sec. = 38 X 107" w Russell (22), supplementing the work of Ives, adopted the same method of calculation, but used what he thought to be more correct figures for the breadth of the pupillary aperture and the stellar magnitude of the faintest visible object. He accepted the pupillary value of 8.5 mm. proposed by Steavenson and a stellar magnitude of 8M.5. The resulting minimum visible was 77) Kyl Ow The above method of calculating by means of stellar data and the mechanical equivalent has been varied by the use of an arti- ficial star. In this way the photometric determination of the threshold can be made experimentally, thus avoiding to some extent the uncertainties of stellar observations. Buisson (23) measured the distance at which phosphorescent screens of different THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 51 sizes could just be seen. The brightness of the screen was then transmuted into the stellar magnitude scale. Using the same values for the mechanical equivalent and pupillary aperture as did Ives, he found the minimum visible to be 12.5 K 1077 w. Reeves (24) used the transmitted light of a modified Nutting sensitometer, the intensity of which could be controlled. The source of light was a tungsten lamp. He also measured the pupils of his observers instead of taking a standard average value. With a visual angle of 1.17 (1 mm. star at 3 meters), the total energy entering the eye at threshold intensity was calculated to be 19.5 & 10°"? w. (mean of three observers). The mechanical equivalent was taken equal to .00159 watt per lumen. Among the investigation in which some magnitude of star is assumed as the limit of visibility, direct energy measurements of that source have been made only in one case. Coblentz (20) assumed stars of the sixth magnitude to have the least visible brightness and measured directly the visible radiation from these stars. He states his procedure briefly as follows: ‘‘ The sensi- tivity of the eye may be estimated—from direct measurements of the heat from stars. The calibration of the radiometer was 1 mm. deflection = 34 & 10™** g.-cal. per cm.” per minute = 85.5 & 10°78 w. per cm.” per second. The sixth magnitude stars gave deflec- tions of 0.5 mm. for blue stars to 1.5 mm. for red stars (say, 1 mm. on an average), depending on their color. From measure- ments made on the transmission of stellar radiation through a cell of water the radiant luminous efficiency may be 0.2 (0.1 for red stars to 0.4 for blue stars). Hence the luminous energy inter- cepted by a pupillary opening of 0.07 cm.? is (85.5 & 10°"? x ee 0.0/ i 1.2.x, 105° w.” Here, then,, is,.a direct radio- metric measurement of the source—there is, however, no corre- sponding direct determination of the threshold. The above summary of work done by previous investigators shows disagreement both as to procedure and as to what shall be called “ the least radiation visually perceptible ’’ or “ the minimum visible.” Lights differing greatly in composition have been used, spectrum and approximate white; and with one exception the energy value has been calculated indirectly from measurements 52 MARGARET M. MONROE made on some other source or computed by the use of a mechani- cal equivalent which expresses the relation of the photometric to the energy unit for some other source than the one used for the visual stimulus. And in case of this one exception the limit of visibility was taken from astronomical data on the visibility of stars, the estimates of which range from the sixth to the eighth magnitude, rather than having been experimentally determined. It would seem reasonable to assume that the minimum visible should mean the least amount of radiation visually perceptible of the wave-length or range of wave-lengths to which the eye is the most sensitive measured in absolute units. So defined, the essential conditions for its determination would be a careful search of the spectrum for this wave-length to which the eye is the most sensitive, a determination of the limit of visibility with this wave-length, and the direct measurement of its energy. In the investigations cited above, one and sometimes two, but never all of these conditions, have been satisfied. Moreover, in no case, whatever the source of light chosen, has the energy been measured and an actual determination of the limit of visibility been made for the same light." 1 By a still more rigid interpretation the determination of the minimum visible might also involve the satisfying of other and more difficult require- ments such as the use of the most favorable time of exposure and size of field, the use of the most sensitive part of the retina, etc., all of which features would in all probability differ in value both for the wave-length of light employed and for the observer. Von Kries and Eyster (25) sought to determine the achromatic threshold not only with the wave-length to which the eye is most sensitive but also under the most favorable conditions of time of exposure, size of field and portion of the retina used. Their selection of 507 wu as the proper stimulus to use was apparently based not on a determina- tion of their own of relative sensitivity but primarily on the fact that Trendelenburg had found that this part of the spectrum gave the maximum bleaching of the visual purple. According to the Diplicitats theory the wave- length that would give the maximum bleaching of the visual purple should be theoretically the most correct. Also this wave-length fell within the range to which Konig had found the eye to be the most sensitive for a group of observers. Their description of procedure leaves one in doubt whether the © long and exceedingly difficult systematic survey was made which would be needed to determine what part of the retina is the most sensitive to the light in question, and what is the optimum size of field and time of exposure. The statement is made that the periphery of the retina was used and that several sizes of field and times of exposure were used, but no information is given -THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 33 The determination of the minimum visible has been by no means the purpose of the present investigation. The energy values of the achromatic and chromatic thresholds have been determined for seven representative parts of the spectrum. There has not been a graded, minute search for the wave-lengths to which the eye is the most sensitive. However, two of the points used, 522 ym in the green and 553 ym in the yellow-green, fall within the range of wave-lengths to which previous investigators have found the eye to be the most sensitive at threshold intensi- ties for different types of observers. To this extent the require- ment that in the determination of the minimum visible the wave- length should be used to which the eye is the most sensitive at the threshold intensities has been satisfied; the other requirement, that the energy measured should be of the light used to produce the eye’s reaction, has been fully satisfied. It would seem then that our results are entitled to consideration together with those purporting to represent the minimum visible, although such a determination formed no part of the original purpose and its relation to the present work came out only in a study of the results obtained. A brief summary of the various values that have been obtained is given in Table X in order that the values of previous investi- gators may be more conveniently compared with those obtained here. The apparent closeness of agreement of the results as to the approximate meridian or degree of eccentricity of the area of the retina stimulated. The source of light was a Hefner lamp the radiations from which were reflected from a magnesium oxide surface into a spectro- scope. The energy values were calculated from Angstrém’s data on the dis- tribution of energy in the light from a Hefner lamp. Von Kries sums his conclusions as follows: © “1, Fiir eine merkliche Erregung des Sehorgans ist bei Herstellung der glinstigsten Bedingungen hinsichtlich Adaptation, Strahlungsart (507 wu) raumlicher und zeitlicher Verhaltnisse eine Energiemenge von 1,3-2,6 X 10-10 Erg. erforderlich. “2. Fiir die Sichtbarkeit dauernd exponierter Objekte ergibt sich bei giinstigster Strahlungsart und giistigster raumlicher Anordnung eine Energiezufthrung von ca. 5,6 X 10-10 Erg. pro Sekunde.” Boswell (26) repeated the experiments of von Kries and Eyster, using an amyl acetate lamp as source and the fovea rather than the more sensitive peripheral retina. He calculated the minimum visible to be 23.7 X 10-17 watt. 54 MARGARET M. MONROE obtained may not at first glance seem compatible with the rather wide disagreement in plan and method of making the determina- tions. On this point, however, two comments may be made: (a) The percentage disagreement is not small, and (b) a great deal of disagreement is doubtless masked by the insensitivity of the instrument used to measure the energy values as compared with the eye. Coblentz (20), for example, has estimated the eye to have 300,000 times the sensitivity of the thermopile. TABLE X Wien 40.00 X 1 10-l6w Reeves 1.95 X 10-16w Drude 6.00 X Von Kries .20 X Drude-Coblentz 1.70 “ Boswell 2:0/ Ne Ives 380, (Present study) Russell O27 ae: Average (553 zu) -27560) 1 Buisson Ligon ye Average (522up) 2.30X “ Smallest value 64 aes F. Pathological Cases. The importance of light and color sense testing as an aid to diagnosis in pathological conditions of the eye is well recognized. There has been, however, little or no quantitative study along this line. Although the present appa- ratus is, of course, unsuited to clinic work, it was thought of interest to determine the achromatic and chromatic thresholds to certain wave-lengths in a few typical pathological cases. The following patients were sent me by Dr. Luther C. Peter, Associate Professor of Ophthalmology, Philadelphia Polyclinic and College for Graduates in Medicine, University of Pennsyl- vania, whom I wish to thank for his kindness both in cooperating in the finding of suitable cases and in furnishing the history of each case. CASE I Mr. Charles P., age fifty-six, cabinet maker. Chief complaint dimness of vision. Has been a heavy user of tobacco and alcohol. Vision O D= 20/100, O S = 20/150, plus correction = 20/40 partly in the right and 20/70 in the left eye, now corrected to 20/40 in each eye. Fundus shows atrophic condition of the optic nerves with low grade retinitis. Fields are considerably contracted for colors, and both ey spots are enlarged for both form and color. Diagnosis: Toxic amblyopia (tobacco and alcohol) ; papillo-macular buidie is undoubtedly involved in the toxemia. Toxic amblyopia—weak vision due to chronic toxemia—may be brought on by any toxic agent, but its chief causes are tobacco THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS 55 and alcohol, either singly or combined. Samelsohn, according to Fuchs (27), was the first to discover the anatomical changes that take place in the disease. ‘He showed that they were limited to the papillo-macular bundles, whose position and course within the optic nerve he was thus able to determine. In the course of this bundle it is found that the nerve fibers have disappeared and nothing but glia tissue is present, while the connective-tissue septa lying between the nerve fibers are thickened. Samelsohn regarded this as the outcome of an interstitial inflammation of the optic nerve, the inflammation affecting the connective tissue portion and especially the septa which convey the blood vessels and which because of the inflammation become thickened. Others, however, think that thickening of the connective tissue is a primary lesion of the optic-nerve fibers by the poison, and that if a thickening of the connective-tissue septa was found, this was a secondary change. Lastly, there are some who believe that even the destruction of the nerve fibers is not the primary affection, but that, in analogy with the conditions of acute poisoning (by quinine, etc.), this consists in a lesion of the ganglion cells in the retina and that an ascending atrophy develops in the nerve fibers as a secondary affair.’ The values of the chromatic and achromatic thresholds obtain2d are givenin Table XI. The loss of sensitivity for both color and light sense is very great. The loss is not, however, uniform throughout the spectrum—there is much irregularity shown from wave-length to wave-length. The extremely insensitive reaction to blue (right eye) and red (left eye) is particularly noticeable. TABLE XI ToraL Amount oF Licut ENTERING THE Eye (Mr. Charles P.) A. Achromatic Thresholds (watt X 10-16) Color Right Left Average (Normal) Red (655 wm) 3754.76 1101394 .80 626.99 Yellow (580 up) 433.73 4888 .48 27 .96 Green (522 up) 27 .64 34413 .28 1.953 Blue (463 uu) 61695 .27 2789 .47 15.62 B. Chromatic Thresholds (watt X 10-12) Red (655 wm) doe 297 .640 .178 Yellow (580 up) 86.530 9.406 9.570 Green (522 up) 3.820 74.140 143 Blue (463 uu) 156.440 201.430 .812 56 MARGARET M. MONROE CASE II Ruth L., age thirty-eight, single, housework. Bilateral glaucoma when twenty-eight years of age. Present condition of the right eye—phthisis bulbi and blind. Left eye — vision= 20/15 partly. The left eye was operated on twice, first operation about eight years ago, an iridectomy; second operation sclero-corneal trephining about eight months ago. Tension normal. Fields show some contraction for both form and color. Up and to the nasal side of the field is a large angular scotoma extending from and includ- ing the blind spot of Mariotte to the periphery including about 1/6 of the circumference of the field. The blind area passes in a horizontal line above the point of fixation. The macular fibers are not involved by the pathologic process in so far as clinical methods can determine. Diagnosis: Chronic congestive glaucoma; partial loss of the upper nasal field. The threshold values for the left eye are given in Table XII. Chromatic sensitivity is normal and the achromatic threshold for green is normal; the achromatic sensitivity to red and blue, however, is very low. The threshold values for yellow were not determined, as the eye fatigued rapidly. TABLE XII ToraL AMouNT oF LiGHT ENTERING THE EYE (Ruth L.) A. Achromatic Thresholds (watt X 10-16) Color Left Average (Normal) Red (655 wu) 2871.54 626.99 Green (522 up) oroZ 1.953 Blue (463 um) 70.15 15.62 B. Chromatic Thresholds (watt X 10-12) Red (655 pm) 287 .178 Green (522 um) .407 .143 Blue (463 um) .687 .812 CASE III Mr. Wm. A. M., age fifty-one, carpenter. Eyes have been failing for seven years. Hypermature cataract in the right eye, incipient cataract in the left. In addition the patient is suffering from bilateral sclerosis of the choroid. Patient has tubular vision in the left eye, 20/30 partly, central vision, improved to 20/20; vision now reduced to 20/40. Light perception only in right eye. Form fields are reduced to 10° in the left eye, and red and green varies from between 5° and 7°. Light perception and light projection is feebly present to the extent of 20°. Wassermann is positive. Clinical diagnosis—sclerosis of the choroid and incipient cataract. The threshold values obtained are given in Table XIII. Both eyes are much below normal, the right more so than the left, as would be expected. In this case again the relative sensitivity is quite different from that of the normal eye—sensitivity to blue ae THE ENERGY VALUE OF SPECTRUM WAVE-LENGTHS _ 57 being disproportionately low. The chromatic responses are most erratic, although they approach the normal more nearly than do the achromatic. In these, too, is shown the great loss of sensitivity to blue. TABLE XIII TotaL Amount or Licht ENTERING THE Eye (Mr. W. A. M.) A. Achromatic Thresholds (watt X 10-16) Color Right Left Average (Normal) Red (655 up) 164307 .31 35229 .61 626.99 Yellow (580 up) 496082 .50 5969 .26 27 .96 Green (522 uu) 61211 .66 7651.46 1.953 Blue (463 nu) 3759525 .00 1051104.50 15.62 B. Chromatic Thresholds (watt X 10-12) Red (655 um) 12.63 Bs fare .178 Yellow (580 up) fr ¥ 9.570 Green (522 up) 9.294 Teist 143 Blue (463 wu) 463 .84 158.49 .812 - * Yellow was called colorless even at full intensity. + Green was always called blue, no matter how high the intensity. CASE IV John H., age sixteen. Secondary atrophy of the optic nerve following injury to the orbit by an automobile. Vision O D = 20/20 partly, O S= 20/70. This has since been reduced to 20/500 in the left eye. The ophthalmoscope shows a white nerve head with much contraction of arteries and veins. The form field is reduced to about 10° and green is visible 3° beyond the point of fixation (by clinical methods of study). Clinical diagnosis: Progressive secondary optic atrophy (traumatic). The energy values for the achromatic thresholds cannot be given—both eyes fatigued very rapidly, and as the first rough settings showed that the threshold values would fall well within the normal range, the more accurate determination was not under- taken. The chromatic thresholds values for the injured eye are given in Table XIV. Sensitivity to green and blue is normal; sensitivity to red is greatly reduced. TABLE XIV ToraL Amount or Licht ENTERING THE Eve (John H.) Chromatic Thresholds (watt * 10-12) Color Left Average (Normal) Red (655 wu) 1700.00 .178 Green (522 up) .653 143 Blue (463 pp) .106 .812 CASE V John B. H., age twenty-three, student. Convergent unilateral squint since early childhood. Vision without glasses: O D= 20/500, O S= 20/300. Vision 58 MARGARET M. MONROE with glasses*i:0O D450», S°O \ 25) Ciax90—=20/300,5.0.5==4008 S OF50 Cax 90—=20/12. Right eye shows convergent squint of 20°. Peripheral vision good; macular vision in the squinting eye amblyopic. Diagnosis: Amblyopic ex anopsia as the result of the squint. Amblyopia ex anopsia—defective vision attributed to lack of use—“ may occur on account of obstruction to the rays of light : falling upon the retina—e.g., congenital corneal opacities, con- genital cataract, and impervious persisting pupillary membrane; or in an eye which from early infancy has squinted, and has, therefore, not been concerned in the visual act.’ Parker (28) gives the following description of the process: ‘‘ Because of defect in the fusion faculty, aided, perhaps, by hypermetropia (far-sightedness) and possibly from debility from disease, one eye shows an occasional transitory squint. This at first produces a diplopia (double images) from the fact that the two images do not fall on corresponding spots of the retinae. The eyes right themselves by muscular effort to parallelism to avoid this diplopia, but this power is soon lost, and the image of the squinting eye. is suppressed, at first much as we would suppress the images falling on the left retina when we are looking through a micro- scope with the right eye; finally the squint becomes constant, diplopia no longer is noticed, and the retina of the squinting eye ceases to functionate. This condition is properly called amblyopia ex anopsia.” As is shown in Table XV, both the light and color sense of the defective eye were found to be normal. This is what was expected by Dr. Peter, who believed it to be a matter of defective spatial perception. TABLE XV. Tota Amount or LicHt ENTERING THE Eye (John B. H.) Achromatic Thresholds (watt X 10-16) Color Right Average (Normal) Red (655 pu) 560.00 626.99 Yellow (580 uu) 26.46 27 .96 Green (522 um) 3.811 1.953 Blue (463 uu) 16.48 15.62 Chromatic Thresholds (watt X 10-12) Red (655 up) 17 .178 Yellow (580 uu) 4.68 9.570 Green (522 uu) 1.219 143 Blue (463 up) 2.102 .812 10. 11. 12 13. 14. 15. 16. WWE BIBLIOGRAPHY Historical Summary . Epert, H. Ueber den Einfluss der Schwellenwerthe der Lichtempfindung auf den Charakter der Spectra. Ann. der Phys., (3) 33, pp. 136-158, 1888. . WeserR, H. F. Die Entwickelung der Lichtemission gltthender fester Korper. Ann. der Phys., (3) 32, pp. 256-270, 1887. . STENGER, Fr. Zur Lichtemission gliihender fester Korper. Ann. der Phys., (3) 32, pp. 271-275, 1887. . Lanciey, S. P. The Selective Absorption of Solar Energy. Amer. Journ. of Sc., 25, pp. 169-196, 1883. . Meyer, O. E. Ueber die Farbe des elektrischen und des Gaslichtes. Zeitschr. f. angew. Electricitétslehre, 1, p. 320, 1879. . Lanciey, S. P. Energy and Vision. Phil. Mag., (5) 27, p. 1, 1889. . Koenig, A. Ueber den Helligkeitswerthe der Spectralfarben bei vers- chiedener absoluter Intensitat. Abhandlungen zur Physiologischen Optik, pp. 144-213, 1903. . Prrtcer, A. Ueber die Farbenempfindlichkeit des Auges. Ann. der Phys., (4) 9, pp. 185-208, 1902. Apparatus and Procedure . Ferree, C. E., and Rann, G. A Spectroscopic Apparatus for the Investi- gation of the Color Sensitivity of the Retina, Central and Peripheral. J. of Exper. Psychol., 1, 3, pp. 247-283, 1916. Ferree, C. E., and Ranp, G. Radiometric Apparatus for Use in Psycho- logical and Physiological Optics. Psychol. Monographs, No. 103. The Photochromatic Interval TSCHERMAK, E. Ueber der Bedeutung der Lichtstarke und des Zustandes des Sehorgans ftir farblose optische Gleichungen. Arch. f. d. ges. Physiol., 80, pp. 297-328, 1898. Fick, A. E. Studien tiber Licht-und Farbenempfindung. Arch. f. d. ges. Physiol., 43, pp. 441-501, 1888. v. Kries, J.. and Nacer, W. Ueber den Einfluss von Lichtstarke und Adaptation auf das Sehen des Dichromaten (Griinblinden). Ztsch. f. Psychol. u. Physiol. d. Sinnesorg., XII, pp. 1-38, 1896. Herinc, E. Ueber das sogenannte Purkinje’sche Phanomen. Arch. f. d. ges. Physiol., LX, pp. 519-542, 1895. CHARPENTIER, A. La Lumiere et Les Couleurs. Paris, pp. 213-216, 1888. ABNEY, W. dE W., and Festine, E. R. Colour Photometry. Phil. Trans. Roy. Soc., London, 183, pp. 531-567, 1892. The “ Minimum Visible” Wien, M. Diss., Berlin, 1888. Ueber die Messung der Tonstarke. Ann. d. Phys., 26, pp. 834-857, 1889. 59 60 18. . ANGSTROM, K. Ueber absolute Bestimmungen der Warmestrahlung mit 27. 28. MARGARET M. MONROE Drune, P. Lehrbuch der Optik. Leipzig, p. 444, 1900, 2d ed., p. 471. dem elektrischen Compensationspyrheliometer, nebst einigen beispielen der Anwendung dieses Instrumentes. Ann. d. Phys., N. F. 67, pp. 633- 648, 1899. . CopteNtTz, W. W. Relative Sensibility of the Average Eye to Light of Different Colors and Some Practical Applications to Radiation Prob- — lems. Bureau of Standards No. 303, p. 225, 1917. . Ives, H. The Minimum Radiation Visually Perceptible. Astrophys. Journ., 44, pp. 124-127, 1916. . Russet, H.N. The Minimum Radiation Visually Perceptible. Astrophys. Journ., 45, pp. 60-64, 1917. . Bursson, H. The Minimum Radiation Visually Perceptible. Astrophys. Journ., 46, pp. 296-297, 1917. . Reeves, P. The Minimum Radiation Visually Perceptible. Astrophys. Journ., 46, pp. 167-174, 1917. . V. Kries, J., and Eyster, H. Ueber die zur Erregung des Sehorgans erforderlichen Energiemengen. Z. f. Sinnesphys., 41, pp. 373-394, 1907. . Boswett, F. P. Ueber die zur Erregung des Sehorgans in der Fovea erforderlichen Energiemengen. Z. f. Sinnesphys., 42, pp. 299-312, 1908. Pathological Cases Fucus, H. E. Text-Book of Ophthalmology. Philadelphia and London, p. 630. Parker, H. C. Hand-Book of Diseases of the Eye. Philadelphia, p. 93, 1910. ow Ve St eh ee a eter ty nA pel 4 nu mi ap aif We ‘ Taek a a ae BF21 .P96 v.34 The influence of tuition in the OT 1 1012 00008 5474