Journal of the Optical Society of America and Review of Scientific Instruments Vol. 7 JANUARY, 1923 Number 1 THE ELECTRICAL RESPONSE OF THE RETINA UNDER STIMULATION BY LIGHT By E. L. CHAFFEE, W. T. BOVIE, AND ALICE HAAIPSON CONTENTS I. HISTORICAL REVIEW ............................................. 2 II. SUMMARY OFPREVIOUS WOR . ...................... 10 III. PREFACE ...................................................... 12 IV. APPARATUS ..................................................... 13 1. General arrangement ........... ........................... 13 2. Eye chamber . ............................................. 13 3. Amplifier ................................................ 15 4. Galvanometer and photographing apparatus .................. 18 5. Stimulating light .......................................... 18 6. Shutter ................................................... 19 7. Device for indicating the exact duration of exposure ........... 19 V. DISCUSSION OFRESULTS .......................................... 20 1. Changes occurring in the response of the eyeball with time ...... 20 2. Explanation of the misunderstanding between (1) Waller and (2) Einthoven and Jolly ......................... ........ 21 3. Improved method of making contact with the eye and the fine structure of response curves obtained with its use ........... 21 4. Analysis of the experimental curves in terms of assumed funda- mental reaction curves, correlated with the responses to the two types of receptor organ, rods and cones..................... 22 5. Discussion of abnormal curves ........... ................... 29 6. Changes occurring in the response of the posterior half of the eye- ball with time ........................................... 29 7. Mathematical laws giving the relation between the response of the eye and the intensity of the stimulating light. Confirmations of psychological phenomena ............................... 31 VI. SUMMARY OF RESULTS ..... ....................... 36 VII. DATA FOR THE PLATES III-VIII ............................ 37 VIII. BIBLIOGRAPHY ............................ 40 1
Transcript
Journalof the
Optical Society of Americaand
Review of Scientific InstrumentsVol. 7 JANUARY, 1923 Number 1
THE ELECTRICAL RESPONSE OF THE RETINAUNDER STIMULATION BY
LIGHTBy E. L. CHAFFEE, W. T. BOVIE, AND ALICE HAAIPSON
II. SUMMARY OF PREVIOUS WOR . ...................... 10III. PREFACE ...................................................... 12IV. APPARATUS ..................................................... 13
1. General arrangement ........... ........................... 132. Eye chamber . ............................................. 133. Amplifier ................................................ 154. Galvanometer and photographing apparatus .................. 185. Stimulating light .......................................... 186. Shutter ................................................... 197. Device for indicating the exact duration of exposure ........... 19
V. DISCUSSION OF RESULTS .......................................... 201. Changes occurring in the response of the eyeball with time ...... 202. Explanation of the misunderstanding between (1) Waller and
(2) Einthoven and Jolly ......................... ........ 213. Improved method of making contact with the eye and the fine
structure of response curves obtained with its use ........... 214. Analysis of the experimental curves in terms of assumed funda-
mental reaction curves, correlated with the responses to the twotypes of receptor organ, rods and cones ..................... 22
5. Discussion of abnormal curves ........... ................... 296. Changes occurring in the response of the posterior half of the eye-
ball with time ........................................... 297. Mathematical laws giving the relation between the response of
the eye and the intensity of the stimulating light. Confirmationsof psychological phenomena ............................... 31
VI. SUMMARY OF RESULTS ..... ....................... 36VII. DATA FOR THE PLATES III-VIII ............................ 37
VIII. BIBLIOGRAPHY ............................ 40
1
CHAFFEE ET AL
I. HISTORICAL REVIEW
The electrical response of the retina when stimulated by lighthas been studied by many investigators. It is not the intentionhere to give a complete rsum6 of their work but to review themore important results which apply to the present experiment. Abibliography is appended to this paper which gives references toall the previous papers on the subject which have been found.
Holmgren2 * in 1866 first observed that an excised eyeball of afrog when illuminated, gives an electric current through a gal-vanometer. One terminal of the instrument was connectedthrough a non-polarizable electrode to the optic nerve and theother terminal through a similar electrode to the cornea. Asteady current due to the injury to the nerve and tissues flowsfrom the cornea through the galvanometer back to the opticnerve. This current, which is called the current of injury ordemarcation current, shows an increase both when the stimula-tion by light begins and when it ends. These changes in currentare spoken of as the positive on and off effects. Holmgren alsoused the posterior half of the eyeball and obtained similar resultsshowing that the electrical response is not due to the iris muscles.He stated that within limits, the change in current is proportionalto the intensity of light.
Dewar and M'Kendrick4 in 1873 and later Dewar6 aloneadded materially to the knowledge of the reaction of the retina.The typical curve for a frog's eye obtained with a Thomsongalvanometer shows a diminution in current after the first sharprise of the on-effect. They proved beyond question that theelectrical effects are due to the retina and not to any other portionof the eye and that the electrical variations are conducted to thebrain. They obtained responses with varying intensities oflight indicating that the Weber-Fechner law is followed. Manykinds of eyes including those of fish, lobsters and pigeons wereused, both when excised and in situ. Their general conclusions aresummarized below:
1. On exposure to light, an increase in demarcation current is obtained fol-lowed by a diminution. The off-effect is another increase.
* Citation numbers here and hereafter refer to Bibliography at end of this paper.
[J.O.S.A. & R.S.1.1 VII2
RETINAL RESPONSE
2. These effects are shown by simple and compound eyes, the eyes of birds,mammals, fish and frogs. The same effects are observed with the eye in a livinganimal when one electrode is connected to the optic nerve or brain and the otherelectrode to the cornea.
3. The change is proportional to the logarithm of the intensity of light,showing that Fechner's law is followed.
4. Yellow light gives the greatest effect, violet the least.5. Sensitiveness to light is essentially dependent upon the retina.6. The change may be followed into the optic lobes of the brain.7. Polarized light has the same effect as.ordinary light.8. The latent period of response is less than one-tenth sec.
Kiihne and Steiner in 188010 and 1881,11 using a slowly moving'aperiodic galvanometer, confirmed the main points of the pre-vious work, but obtained curves in which the diminution after thefirst rise of the on-effect was so great as sometimes to cause thecurrent to decrease below the normal current of injury. Theyconcluded that this excessive negative variation denotes a dyingor injured eye. These investigators worked with the isolatedretina.
Waller5 in 1895 performed some experiments on the frog'seye using a D'Arsonval galvanometer which indicated that thecurrent variations follow the Weber-Fechner law.
Beck 6 in 1899 worked with the eye of the cephalopod Eledonemoschata, an eye which has no cones. Placing the electrodes invarious parts of the eye he obtained very simple response curvesconsisting merely of a rise of current on exposure to light followedby a return to normal when the stimulation terminates.
Waller 17, 18, 19 in 1900 reported some experiments on the retinal
currents from a frog's eyeball under stimulation by light and byelectricity. He observed a negative variation which precedes thepositive rise of current. This is a relatively slow negative varia-tion recorded by a slowly moving galvanometer. He also calledattention to a step-like rise of current during the on-effect, which,in the light of present work, proves to be due to the superpositionof the typical quick rise of the on-effect upon a long gradual in-crease later emphasized by Ishihara, and Einthoven and Jolly,and pointed out by them as being a part of some normal responsecurves.
Jan., 1923] 3
[J.O.S.A. & R.S.I., VII
Fuchs1 421 in 1901 continued his investigation which was begun
in 1894, of the latent periods of the various portions of the retinalcurrent fluctuations and obtained smaller values than thoseobserved by others. He concluded that with constant intensityof stimulation, (1) the time values of the photoelectric* phen-omena of the retina remain constant in whatever manner thecurve forms and (2) the restitution of the changes which arebrought about by this intensity, is completed in an almost con-stant length of time. His method of illuminating the eye wasby the use of a sparking rheotome, which presupposes the reactionof the eye to be of short duration like that of muscle or nerve.
DeHaas27 in 1903, using a slowly moving d'Arsonval galva-nometer, investigated the relation of the response to the productof intensity of light and the time of exposure. His results showthat for exposures varying from 0.01 sec. to 0.8 sec. equal responsesare produced by equal products of intensity and time.
Gotch25 ' 26, 2 in 1903 and 1904 performed some extremely inter-ing and valuable experiments with the frog's eye. He used acapillary electrometer as a measuring instrument, which, becauseof its quick response, made his method of experimentation muchsuperior to that of previous observers. He obtained a muchmore accurate history of the time relations of the response of theretina. The general shape of his curves is similar to thoseresulting from former experiments, less modified, however, bythe inertia of the measuring instrument. The more rapid fluctua-tions in response were brought out and for the first time curveswere obtained showing a quick negative reaction preceding thepositive on-effect. The latent periods of the on and off-effectswere measured for white and colored lights and they appeared tobe different for the three colors, red, green and violet. He calledattention to small irregular oscillations which are occasionallyobserved in the level portion of the curve before the off-effect. Heconcluded that the photoelectric* changes are all of the samegeneral monophasic type, and that the retina contains two photo-
* The word "photoelectric" is used in the bibliography because it is the termused by the earlier investigators. The term "photoelectric" has now come to have avery definite meaning in physical phenomena. The authors believe that the term"electrical response" is a preferable term.
4 CHAFFEE ET AL
Jan., iyzij RETINAL RESPONSE 5
chemical substances, one reacting to light and the other todarkness.
Piper30 in 1904 repeated Beck's work with the eye of thecephalopod and obtained in general confirming results.
The work of Ishihara32 in 1906, although quite extensive, addedlittle to the results of the earlier investigations. He did recognize,however, that a second rise after the first positive on-effect whichhad been previously observed, without appreciation of its impor-tance, is a part of the typical response under some conditions.His curves could not give the true relation between response andtime because of the use of a slow Deprez-d'Arsonval galvanom-eter.
Work of considerable importance was done by Brilcke andGarten33 in 1907 who used at first a capillary electrometer andthen a string galvanometer. Many different kinds of eyes were
= ~~~~~
FIG. 1. (Typical crve front Einthoven and Jolly) "The combined reaction of thethree substances." Flash 0.01 sec. Dark eye, Green light. (Ordinate-Height of B=150 microvolts: Abscissae-Distance to maximum of C=16 sec.)
studied and curves were obtained showing the typical responseof the several eyes. The latent period was measured and theeffect of fatigue studied. They observed, as did Gotch, the quickbut weak negative reaction preceding the positive rise of the on-effect. Briicke and Garten were unable to observe any oscilla-tion in the continuous effect during illumination claimed tohave been observed by Gotch.
Einthoven and Jolly3 5 in 1908, in a very important and pains-taking research, pursued further the investigation of the effectof light on the retina of the frog. In these experiments theyused the whole eyeball and obtained photographic records of theresulting deflections of a string galvanometer. Fig. 1 shows a
IA 1 _ . - 1
[J.O.S.A. & R.S.I., VII
curve copied from their paper and represents a typical responseto a flash of strong light of 0.01 sec. duration. Fig. 2 is a response
obtained by them to a time exposure. The curve of Fig. 1 con-sists of three parts all of which had been previously observed byothers and designated by Einthoven and Jolly as follows: thesharp negative reaction, "A"; the rapid positive rise followed by adecline, "B"; and the later gradual increase and subsidence,
OFF EFFECT
aI
FIG. 2. (Typical Curve fromt Einthoven and Jolly.) "Conflict between the reactions
of the three substances." Tine exposure of 9 sec. Dark eye. White light. (Ordinate-
Height of E=208 nicrovolts. E, a control curve.)
B
FIG. 3. Diagrammatic representation as three separate curves of the reactions to
light of the three substances, A, first; B, second; and C, third substance. 1, light; d, dark-
ness. (Einthoven and Jolly.)
"C." They found the same three, characteristics occurring innearly all responses whether for a flash of light or for a prolonged
exposure, except that in the latter case the off-effect is superposedon the "C" rise, and the three parts have different relative magni-tudes. Einthoven and Jolly attributed the three parts of thecurve to the reaction of three substances in the retina, each onegiving its particular form of reaction curve. Fig. 3, taken from
CHAFFEE ET AL6
RETINAL RESPONSE
their paper, shows their idea of the separate reactions of thethree substances, "A," "B," and "C." The sum of these threecurves would give a typical response. The lowest line in Fig. 3indicates the duration of exposure, I denoting light and d darkness.
They also investigated the variation of latent period withintensity of light, and showed that with strong light it is of theorder of 0.01 second while with very weak light it may be as muchas 2 seconds.
Attention is called to a curve which shows irregular variationsduring exposure which they thought might be due to a rythmicalresponse of the eye, but which they suggested may have beencaused by variations of the arc light used to illuminate the retina.
Waller37 in 1909 attacked the three-substance theory of Ein-thoven and. Jolly and proposed a two-substance theory. Aresponse curve observed by Waller is shown by Fig. 4. His
FIG. 4. "Retinal response of the second stage showing the preliminary negativeeffect at the 'make' of illumination." (Waller)
analysis of this response shown by dot and dash curve of Fig. 5 isindicated by the two full-line curves of the same figure. The curveshown in Fig. 4 is not, however, a typical response but, accordingto Waller, one indicating a modified or ageing eye. The work ofEinthoven and Jolly was carried further by Jolly38 alone in 1909.He published curves, one of which is reproduced in Fig. 6, showingresponses of dying eyes and the effect of massage and tetaniza-tion. The paper contains a very good historical outline of pre-vious work.
Jan., 1923] 7
[J.O.S.A. & R.S.I., VII
Piper39 ' 40, 41 in 1911 worked with various eyes and proposed atheory very similar to. that of Einthoven and Jolly. Piper's
FIG. 5. (Analysis of te response of a modified eye, made by Waller to demonstratehis two-substance theory.) "Diagram to illustrate the effect upon a galvanometer(broken line) of a simultaneous larger positive current and smaller negative current, thelatter commencizg and ending more rapildy." (Waller.)
A_ I d
FIG. 6. "A response to strong green light of a dark eye which has remained for sixhours between the electrodes. The second substance is little in. evidence. The response isto an exposure of 13.2 sec." (Jolly.)
theory supposes three reactions: exposure reaction, a dark reac-tion of opposite sign, and a third long secondary reaction lastingbeyond the stimulation.
8 CHAFFEE ET AL
RETINAL RESPONSE
In 1911 and 1912 Nikiforowsky4 ' observed the retarding effectof lowering the temperature of the eyeball.
Brossa and Kohlrausch46 in 1913 and later47, 48 in 1914 showedthat there is a qualitative difference in the shape of the responsecurve of the pigeon's eye for various colored lights supporting thetheory that cones are color sensitive. No qualitative differencewith wave length was observed with rod vision of the owl.
Frohlich4' in 1913 observed very marked oscillations in theresponse curve of the cephalopod and showed that the frequencyis a function of the wave length of light, and the intensity of oscil-lation a function of the intensity of light. He thus claimed tohave found a basis for light and color sensation but it must beremembered that the cephalopod possesses only rods.
Day5 ' in 1915 studied the electrical response of the retina understimulation of light in the eye of the fish. The eye was used insitu and was connected to a d'Arsonval and later to an Einthovengalvanometer, through non-polarizable electrodes making contacton the cornea and on the eyeball. Photographic records of theresponses were obtained. The source of illumination was a 32-candle-power incandescent light at a distance of one meter.The general conclusion reached was that the photoelectric*phenomena in the fish eye are similar to those obtained by Ein-thoven and Jolly in the frog eye. The latent periods of the variousparts of the response curve were measured and also the times tothe various maxima. The first maximum is reached after 0.26sec. and the summit of the off-effect in 0. 165 sec. Intermittentlight produces separate responses up to 25 flashes per sec. At28 per sec. the curve resembles that of an exposure to continuouslight.
Sheard and McPeek 4 in 1919 performed experiments withdog eyes, connected to a Thomson galvanometer, the small deflec-tions of which were made visible by a system of mirrors. Theyinvestigated the gradual changes in potential difference after alter-nate exposures to lights of different colors by exposing an eye to acertain portion of the spectrum for from four to five minutes, wait-ing the same length of time, and then exposing the eye to anotherportion of the spectrum. They called attention to the fact that
Jan., 19231 9
CHAFFEE ET AL
previous experimenters had paid little or no attention to the gen-eral rise or fall of potential after stated periods of exposure, andarrived at the general conclusion that the shorter wave lengths oflight produced negative potentials (from retina to nerve) as shownby the general downward trend of the galvanometer deflection,while the longer wave-lengths produce positive potentials (fromnerve to retina), shown by the upward trend. Complementarycolors have opposite tendencies, and each may neutralize the po-tential changes produced by the other. They found agreementwith the Herring theory of color perception. Their records showvery marked irregular maxima and minima which were explainedby the authors as being due to fatigue and after-image effects andnot to any external causes.
II. SUMMARY OF PREVIOUS WORK
Although various observers disagree regarding some detailsof the response of the retina to stimulation by light, certaincharacteristics of the observed curves are fairly well established.Fig. 7 diagramatically embodies these characteristics which
Lighk on Lig oif
a Normol Current J-Inlun ________________
r , ~~~~~~~~~~~Timed
FIG. 7. A diagrammatic curve which embodies the characteristics of the normalelectrical response of the eye as established by previous observers.
consist of a quick decrease in the current of injury immediatelyfollowed by a rapid increase to a maximum, and a second gradualincrease of current on which is superposed the rapid increase ofthe off-effect. The typical curve undergoes a change as the eyeages and eventually takes forms shown in Fig. 6.
[J.O.S.A. & R.S.I., VII10
RETINAL RESPONSE
The early observers and some more recent ones used slowlymoving galvanometers which are totally inadequate to show anyfine structure in the response curves or to give the true time rela-tions of the reaction. The capillary electrometer used by Gotch,BrUcke and Garten, and others is more rapid than moving-coilgalvanometers, but when adjusted for high sensitivity is then muchtoo slow. The string galvanometer first used by BrUcke and Gar-ten and later by Einthoven and Jolly and others is more rapid, butan examination of the calibration curves of Einthoven andJolly shows that, in order to get sufficient sensitivity, the stringwas used under small tension so that it required about 0.2 sec.for the string to attain its final deflection under the influence ofan instantaneous change in current.
All instruments previously used, including the capillary elec-trometer, require some current for their deflection so that themeasurement of potential changes is modified by the current flowand consequent drop in potential through the resistance of thetissues.
Several observers speak of oscillations or perturbations duringthe declining portion of the "B" or during the "C" part of thecurve. These oscillations are irregular and were observed in onlyfew instances.
Some experimenters used the whole eyeball, while others usedthe posterior half of the eyeball or the isolated retina.
In most cases the intensity of illumination of the retina wasextremely great. Einthoven and Jolly used stimulation rangingfrom 120X 106 meter candles to 0.01 meter candle. A quotationfrom Einthoven and Jolly says: "It is sufficiently clear that forthe photoelectric reaction of an isolated frog's eye, there isrequired much more light than for the development of light per-ception in the human eye." (See ref. 35, p. 414). With themore sensitive apparatus to be described, this statement no longerholds.
The only theories of action that have been found by the authorsare the nearly equivalent three-substance theories of Einthovenand Jolly and of Piper, and the two-substance theories of Gotchand of Waller, all of which theories have been outlined above.
Jan., 1923] 11
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
III. PREFACE
Vision includes several separate processes. First, radiantenergy in the form of light, through the agency of the crystallinelens, forms an image on the retina, the receptor mechanism.*The receptor mechanism under this stimulation sets up disturb-ances or nerve impulses which are conducted through the complexstructure of the retina to the optic nerve and thence to the brainwhere the impulses are interpreted. These nerve impulses inthe rods and cones of the retina, and in the optic nerve are ac-companied by electrical changes which can be measured by suit-able electrical apparatus. These electrical changes are un-doubtedly "action currents" which are inseparable from thefunctional responses of the conducting tissues and as a workinghypothesis, it is assumed that a study of these electrical varia-tions gives some knowledge of the nerve impulses which go to thebrain. If this be true, a study of the electrical changes accompany-ing variations of stimulation of the retina affords powerful meansof determining what goes on in the receptor organ. No attemptis made to study the interpretation processes in the brain. If
*The retina of the vertebrate eye is composed of three concentric tissue layersand certain supporting elements. The receptor mechanism is the outer of these con-centric layers. It consists of radially arranged highly differentiated cells of two kinds-the rod cells and the cone cells. The rods terminate at their inner ends in globe-shaped enlargements, the cones in pyramidal-shaped enlargements. The retinaltissue next beneath these visual cells, the middle concentric layer, consists of bipolarnerve cells also radially arranged. They end outwardly in dendritic processes justbeneath the enlarged ends of the rod and cone cells. One unbranched process fromeach bipolar nerve cell is extended in between the visual cells where it terminates in asmall thickening. The other dendritic processes bifurcate repeatedly so as to forma close sub-epithelial mesh-work just below the visual cells. At the inner ends, thebipolar nerve cells also end in dendritic processes which form synapses with similarprocesses from the ganglion cells of the optic nerve. The optic nerve ganglia form theinner concentric layer of the retina. Each nerve ganglion cell sends out a nerve fibreto become a part of the nonmedulated optic nerve.
There are other nerve cells lying between the bipolar cells and some of these senddendritic processes outwardly towards the bases of the visual cells and others sendprocesses inwardly to form synapses with the dendritic processes of the ganglion cellsof the optic nerve. The exact course of the processes of some of these cells is stillin dispute.
This brief description of the retinal tissues is sufficient to emphasize the complexityof the biological mechanism with which this experiment deals.
12
RETINAL RESPONSE
the findings can be correlated with psychological observations,the method of investigation is further justified.
Besides this possible opportunity of adding to the knowledgeof vision, a study of the electrical variations in an optic nervewhen the retina is illuminated, provides an extremely valuablemethod of investigating the propagation of an impulse in asensory nerve. For not only is the impulse initiated by a normalexcitation of the receptors, but, using stimulation by light, it ispossible to control not only the intensity factor but also thequality of the exciting agent.
The experiments reported in this and subsequent papers werebegun in January, 1920, and consist of a study of the electricpotentials produced in the retina under various conditions of theretina, and for many variations in the stimulating light such asvariations of intensity, time of exposure, wave length, etc. Theobject in view in attempting this investigation is to carry further,with improved apparatus, the work of previous experimentersbriefly outlined above, with a hope that a better insight into themechanism of vision can be obtained.
The work has expanded to an extent that precludes the pos-sibility of reporting it all in a single paper. This paper contains adescription of the apparatus, some general observations, thechange of the response as the age of the eye increases measuredfrom the time of excision, and a brief report on the effect on theretinal response, of varying the intensity of stimulation.
IV. APPARATUS
1. General arrangement.-The general arrangement of theapparatus is shown in photographs a and b of Plate 1 and sche-matically represented in Fig. 8, which is self-explanatory. Adetailed description of the various parts follows.
2. Eye chamber.-The excised eye is placed in the eyechamber shown in section in Fig. 9. In the early experiments,the eye was removed from a pithed frog and the whole eyeballwas used, electrical connection being made through non-polariz-able electrodes to the cut end of the optic nerve and to the cornea.Latterly the frog is etherized before excision and the eye ball,
Jan., 1923] 13
14 CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
after excision, is cut in half and only the posterior half is used.
Etherization did not appear to have any effect upon the results
except that it was necessary to allow the eye to recover from theSHUTTER
NEtTTIHN_ SHIELD I
EYE CHAMBER STIMULATINGLIGHT SOURCE
AMPLIFIER
PHOTGRAPHING DRUM
MARKER
EINTHOVENGALVANOMETER
FIG. 8. The general arrangement of the apparatus
S
FIG. 9. The eye chamber
effects of the anesthetic before taking observations. After re-
moving with filter paper some of the vitreous humor, the half
of the eyeball is set up on a pad of cotton, moistened with Ringer's
RETINAL RESPONSE
solution, on a small pedestal B capped with wax W, as shown inFig. 9. Connection is made with the pad and thence with theoptic nerve through a cotton thread t also moistened with Ringer'ssolution. This thread dips in the U-shaped glass tube E1 con-taining zinc sulphate solution. An amalgamated zinc wire formsthe other terminal of the. non-polarizable electrode and connectswith the insulated terminal of the eye chamber T. The otherconnection to the eye is made through the moist thread t2, whichis permitted to touch the retina at one point only. This threadconnects with the other terminal T2 of the eye chamber througha second non-polarizable electrode E2. Water in the bottomof the chamber maintains a moist atmosphere. On the coveris a shutter S and a reflecting prism P.
3. Amplifier. The eye chamber is connected by electro-statically shielded wire to the amplifying system, which com-prises the principal improvement in the apparatus. Theamplifying system consists of a two-stage, thermionic-vacuum-tube amplifier commonly used in radio communication.* Suchan amplifier is a potentially operated device, so that, besides in-creasing the sensitivity of the measuring apparatus, it affords theadditional advantage of making measurements of potential
+~~~~ + epX~~~~~ _+
el~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~A -'U
FIG. 10. The thermionic amplifier tube
possible without drawing current or energy from the retina.Resistance of tissue and other resistances are without effect onthe measurements.
* Forbes has used a single stage amplifier in the measurement of action currentsfrom nerves (61).
Jan., 1923] is
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
The thermionic amplifier tube consists of an evacuatedchamber C, Fig. 10, into which is sealed a filament F which canbe heated, a plate P, and between the filament and plate a wiregrid, G. Electrons emitted by the filament pass through the gridto the positively charged plate and thence through the externalcircuit R. Any variation of potential of the grid with respect tothe filament causes a corresponding change in electron current tothe plate and hence a change in current in the external platecircuit. This change in plate current resulting from a changeof grid potential beg could be caused by some larger change inpotential e, if applied directly in the plate circuit. Since epis greater than beg, the tube acts as an amplifier and the fraction
Le is the amplification factor yt. Another useful constant ofbeg h
the tube is the filament to plate resistance defined as - rp,
where here i is the plate current and es is the voltage betweenthe plate and filament.
The connections of the amplifying system are shown in Fig. 11.
-TO~~~~~
FIG. 11. The connections of the amhplifying system
V1 and V2 are thermionic amplifiers. The tubes used in thisexperiment are standard types made by the Western ElectricCompany. Tube V1 is known as type D and has an amplificationfactor of about 40 and a plate resistance of 200 000 ohms. TubeV2 is type J, and has a j, of about 8 and a resistance of 30 000
ohms. The filaments are heated by the common 6-volt battery
16
RETINAL RESPONSE
Ef. In the particular arrangement used in these experiments E1is a storage battery of 200 volts and E2 a similar battery of 100volts. R and R2 are chosen to be at least equal to the filament toplate resistance of the tube with which it operates, and are 250 000and 30 000 ohms, respectively. e is a small constant cell usedto polarize the grid of V to a proper negative value. P is anarrangement which, when the key is depressed, impresses on thegrid a known small potential for calibration. The adjustable tapa, is attached to the battery at a point such that the grid potentialof the second tube is zero or slightly negative as indicated by theabsence of a deflection of the galvanometer b. G is an Einthovengalvanometer provided with an adjustable shunt. The adjust-able tap a2 is attached to the battery E2 so that there is practicallyno current through the galvanometer G, perfect adjustment beingobtained by means of the potential divider P. The amplifica-tion of the whole system, defined as the ratio of the deflection ofgalvanometer G using the amplifier to that if the amplifier werenot used, is given by the expression;-
i e 4112RIR2Rt
/l e (R+rl) (R2r2+R2G+r2 G)where i is the current through the galvanometer when potentiale is impressed on the grid of the first tube, and i' is the galvanometercurrent which would result if the same potential e acted in thecircuit containing only the eye and galvanometer. /li and 2 arethe amplification factors and r and r2 the plate to filament re-sistances of the respective tubes V and V2. R and R2 are therespective external plate circuit resistances as shown in Fig. 11.G is the resistance of the shunted galvanometer, and Rt the re-sistance of the eye and connecting electrodes.
In the particular arrangement used in these experiments theconstants are approximately as follows:
On account of the high amplification of the system, certainprecautions must be observed. The plate battery E1, the tubes,and resistances R1 and R2 must be shielded from outside electricaldisturbances with sheet metal or netting. The batteries mustbe of large capacity and charged at a low rate in order thattheir potential be constant. Mercury switches are necessaryin the filament circuit to insure good contact, for the slightestvariation of filament current causes irregularities and drift of thegalvanometer deflection. The tubes V1 and V2 must be speciallyselected as some have enough internal leakage to prohibit theiruse.
4. Galvanometer and photographing apparatus.-The galva-nometer is a Cambridge Einthoven instrument provided with afield coil which operates on a 12-volt storage battery. The fibre
is of quartz gilded by cathode spray.The photographing camera consists of a camera bellows fastened
to a box B, Fig. 8. (See also Plate I.) On this box slides a
chamber C in which rotates a drum driven by a motor. Aboutthis drum is wrapped the photographic bromide paper on whichthe records are taken. A mirror and a ground glassviewingscreen are so arranged that the deflections of the fibre can be
watched while the record is being photographed. The timemarker, shown in Fig. 8, consists of a special clock driven wheel
carrying spokes which serve to momentarily intercept the lightonce every five seconds.
5. The stimulating ligh.-The stimulating_ light source isshown in section in Fig. 12. The outer part consists of a com-
T
FIG. 12. Te stimulating light
mon three-cell flash-light casing T. At one end is a plug of woodW on which is fastened a'socket carrying the small 6-volt light L.A ground glass G diffuses the light which then passes through ahole in the center of the diaphragm D. Various diaphragms are
18
RETINAL RESPONSE
used with holes ranging from 0.010 inch.,to 1 inch in diameter,each one having double the diameter 'of the preceding one. ExLcept when the large' holes are used, the source of light is con-sidered to be at the diaphragm D since the light diverges from thispoint. The intensity of light at the retina is then proportional to
(d)2 where d is the diameter of the hole in the diaphragm and x
is the distance from the diaphragm to the retina. The intensityis then reduced to foot candles or meter candles by a factor deter-mined by a footcandle meter. Using this source of light, a con-tinuous range of intensities from 6 x 10-6 to 5000 meter candlescan be obtained. For higher intensities a bare 400-watt nitrogen-filled lamp is used.
6. Shutter.-The shutter shown in photographs a and bof Plate II consists of a motor-driven pair of disks, the speed ofwhich is given directly by a tachometer. Each disc has 'anopening extending through nearly 180 degrees of arc so that bychanging the relative position of the disc, various openings can beobtained up to half a revolution. These discs rotate in front ofthe light source just described. On the opposite side of the discsald in line with the light is a compound photographic shutterwhich is normally closed. By a special cam on the shaft of therotating discs the photographic, shutter can be made, whentripped, to open automatically for one half a revolution. Inthis way one exposure can be obtained whose time is determinednot by the photographic shutter but by the angular opening ofthe two discs and their speed of rotation. By varying both the,angular opening and the speed of rotation any accurately deter-
1mined time of exposure can be obtained from second to 2seconds. 5000
7. Device for indicating the exact duration of exposure.-An improvement recently added to the apparatus consistsof a device to record automatically on the photographicpaper the exact instants of commencement and termination ofexposure. In essence it consists of a small string galvanometerplaced in box B, Fig. 8, so that the string casts a shadow on the!
Jan., 1923]
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
bromide paper. This galvanometer is operated through a com-mutator on the shaft of the rotating shutter discs. The com-mutator consists of two metal segments one permanently fastenedto one disc and the other to the second disc. The segments ofthe commutator are designed in such a manner that the circuitthrough this time-recording galvanometer is opened and closedin exact synchronism with the opening and closing of the shutter.
V. DISCUSSION OF RESULTS
In work of this kind it is necessary to study a great manyresponse curves in order to determine the characteristics whichare typical. Up to the time of writing this paper over 5000photographs of the response of the retina of the frog's eye havebeen taken under various conditions. In these experimentsseveral hundred eyes have been used.
The earlier experiments, following the procedure of previousworkers, were made using the whole eyeball with one threadelectrode touching the cornea and the other, the optic nerve.The resulting curves are similar in the main points to thoseobtained by the previous investigators.
1. The shape of the typical curve for white light stimulationundergoes marked changes as the eye ages after excision. Expo-sures selected from an "age run" on an eyeball are shown in PlateIII. As will be seen on examination of photograph a of Plate III,a fresh eyeball gives a response curve for time exposure whichembodies the quick negative depression called by Einthoven andJolly the "A" portion, a rapid rise in the positive direction to amaximum known as the "B" portion, and the long positive risecalled the "C" rise. In the figures of Plate III this "C" rise isinterrupted by the positive "off" effect when the excitation ter-minates. It may be difficult to detect the "A" dip on some ofthe figures of Plate III but they are easily distinguishable on theoriginal.
As the eye ages the earlier part of the response undergoes achange in shape so that the potential may actually decreasebelow the normal value for a considerable time as shown in curvefof Plate III. It will be noticed, however, that even in this
20
RETINAL RESPONSE
modified condition the first rapid negative variation still persistswith an apparent increase in size.
2. The changes shown in Plate III explain a misunderstandingwhich existed concerning the first negative response. Wallerclaimed that a negative reaction signifies a modified eye (see ref.37, pp. 171-175) while Einthoven and Jolly asserted that a nega-tive deflection is a characteristic of a normal response (see ref.35, p. 390 and ref. 38 pp. 378, 379). The difficulty is due to thefact that there are two types of negative reactions. The slownegative reaction obtained by Waller and shown in Fig. 4, repro-duced from his paper, was obtained using a slowly movinggalvanometer and does indicate a modified or dying eye. Thecurve of Fig. 4 is practically the same as curve f of Plate III ifthe first quick dip of the latter is smoothed out. Waller con-fused this relatively long negative reaction with the quite distinctvery rapid negative variation first observed by Gotch and indi-cated by Einthoven and Jolly as the "A" portion of the normalcurve (see Figs. 1 and 2). However, Fig. 6 reproduced fromJolly's later paper embodies the two types of negative reactionsand is very similar to curvef. of Plate III.
3. Many experiments were made to determine the effects onthe response curves of changing the intensity and time of expo-sure but it was soon realized that these curves were of doubtfulvalue because of the disadvantageous position of the cornea elec-trode. The seat of the electromotive forces is in the retina and anelectrode situated as far away as the cornea is unable to pick upthe fninute changes in electrical potential and will give onlyaverage gross effects from the whole retina. Moreover, con-siderable electrical leakage through the tissues and over the out-side of the eyeball undoubtedly takes place under such condi-tions which materially modifies the results. Consequently in alllater experiments only the posterior half of the eyeball was used,and the end of the moist thread which previously made contactwith the cornea was made to touch the retina at only one smal Ipoint. It was expected that using this method the reaction of asmall number of visual elements would be obtained and that itwould be little modified by leakage. Most of the vitreous humor
Jan., 1923] 21
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
was removed from the divided eye. This arrangement of elec-trodes gives much greater responses and brings out, as will beexplained later, much fine structure in the curves which can notbe obtained using the whole eyeball. If the thread, instead oftouching at one point, lies along the inside surface of the retina fora short distance, much inferior results are obtained showing littleof the fine structure.
It should be remarked that the arrangement of thread elec-trodes described above, would, because of the flow of currentthrough their high resistance, give distorted records if connecteddirectly to the string galvanometer, but used with the amplifyingsystem described above, gives the true potential variations irre-spective of the resistance of the eye and electrodes. One testshowed that each thread electrode had a resistance of approxi-mately 50 000 ohms, and the eye itself a resistance of from 1000
to 10 000 ohms according to the position of the inside threadterminal.
When the posterior half of the frog's eyeball, arranged withelectrodes as last described, is illuminated with flashes of whitelight of various intensities, the series of response curves obtainedare much more complex than those obtained using the whole eye-ball and show structure not observed by previous workers.
4. A few typical response curves are reproduced in Plates IVand V and are selected to show the variety of normal shapesobtained under various conditions. All curves except e,f, g, and hof Plate V were taken with a flash of the exciting light. Curvese, J, and g give the typical response for a time exposure of a! fewseconds. Curve h is due to two separate exciting lights; thefirst maximum indicates the beginning of the first light, thesecond maximum the beginning of the second light, the thirdmaximum the termination of the second light, and the last maxi-mum the termination of the first light.
The vertical white lines, where they appear, indicate five-second intervals. The short constant deflection at the begin-ning of some of the curves is the calibration of the galvanometerand was produced by a potential change of 0.14 millivolt.
22
RETINAL RESPONSE
An examination of normal response curves reveals certainoutstanding typical characteristics. Except for very weakillumination, the curves for a flash of light show a quick negativevariation followed by a positive rise to a sharp maximum. Thisfirst positive maximum is usually followed by one or more sec-ondary maxima. The time from the beginning of the responseto the first positive maximum is practically independent of theintensity and time of illumination and is approximately 0.27second. This first positive maximum is shown most clearly incurves a, b, c, and d of Plate V.
A definite second maximum is usually discernable but is some-times masked by the larger first maximum. The time to thesecond maximum is, as with the first maximum, practically aconstant independent of intensity and duration of excitation andabout equal to 0.70 second. The second maximum is clearlyshown in curves a, b, and c of Plate V and also in some of the otherphotographs.
Following the second maximum there is occasionally seen athird maximum, the time of which varies from something of theorder of 0.6 second to several seconds. This third maximumis not always visible for it may be masked or fused with the secondmaximum giving a broad hump as shown in g and of Plate IV.The third maximum is clearly shown in curves i, j, k, and I ofPlate IV. These four curves belong to a series in which the timeof excitation due to a light of constant intensity was very grad-ually increased from 0.00128 to 0.0135 sec. In i the first maxi-mum is only a pause in the upward ascent to the second maximum.As the time of illumination increases, the first maximum increasesin size relative to the third as though the increase of the firstwere at the expense of the third. The second maximum, on theother hand, shows a gradual increase in height. These progres-sive changes shown in the series i, j, k and I have been observed inmany other series and indicate that the first and third maximaare related, and that the second maximum is quite independent ofthe first and third.
Many of the curves show a very slow positive rise to a maximumwhich may occur even 10 or 15 seconds after the flash of light.
Jan., 1923] 23
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
This fourth maximum may be easily confused with the third andcan be separated sometimes from the third only by plotting thetime to its peak as the intensity of illumination is increased. Thetime to- the third maximum decreases while that to the fourthmaximum increases as the intensity of illumination is increased.The fourth maximum is present only for the higher intensitiesand is identified with the "C" rise of Einthoven and Jolly. Curved of Plate V shows the fourth maximum, the second and thirdbeing absent or masked. The fourth maximum is present whenthe eye is subjected to time exposures and shows in curves f, gand h of Plate V.
At very low intensities of excitation the separate peaks are morefused and less pronounced. In curve b of Plate IV the first,second and third maxima are evident. In a of the same plate it isdifficult to tell whether the first rise is due to the first or secondmaximum or both, although the third maximum is clearly seen.The same uncertainty prevails in c and d of Plate IV.
The typical responses described above can obviously not beanalyzed into the three fundamental curves assumed by Ein-thoven and Jolly nor into the two curves of Waller. Becauseof the complexity and the great variety of the normal responsecurves it is difficult to obtain an analysis which is able to buildup all possible forms obtained and which has at the same timesome reasonable correlation with the recognized performance ofnerves and their connected mechanisms.
Since the duplicity theory of v. Kries (see ref. 56, p. 203) hassurvived while other theories of vision have come and fallen, it isreasonable to assume that the response curves will show twomore or less independent variations, one due to the cones or colorsensitive organs, and the other to the rods or achromatic receptors.Attention has already been called to the fact that the first andsecond maxima are apparantly more or less independentsand also that the third maximum seems to be related tothe first. Experimental evidence has not yet been able todetermine definitely which of the variations in the responsecurve is to be attributed to the cones and which to the rods, butindicates that the first maximum is attributable to the cones and
24
RETINAL RESPONSE
the second to the rods. In the two upper rows of Figs. 13, 14, 15and 16 are shown some simple fundamental forms of responsewhich when added, as shown in the lowest row, give curves closelyresembling the experimental results. The forms in the toprow of each figure are assumed responses of one type of receptororgan, rods or cones, and those in the middle row are assumedresponses of the other type of receptor organ, both responsesbeing considered to be independent. In the bottom row ofFig. 13, is shown a series of summation curves the shapes of
a b c d
FIG. 13. Synthetic response curves (third row) built up from assumed fundamentaresponse curves of he two receptor organs, cones and rods. The effect of varying relativemagnitudes of the two fundamental curves, their shapes remaining similar.
which progressively change as the assumed forms of the firstmaxima curves (top row) and second maxima curves (bottomrow) have different relative magnitudes. Such a progressive
Jan., 1923] 25
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
change is observed experimentally as the intensity of illuminationis increased. In Plate VIII are reproduced two series of responsesshowing this progressive change in shape. The first series a-mwas taken with the intensity of excitation decreasing from583 X 10-5 to 2.7 X 10-5 meter candles, the second series n-z wasobtained with increasing intensity of illumination from 32 X 10-5to 18930 X 10-5 meter candles. Fig. 14, and Fig. 15 a and b, showother possible variations of the curves. In a and b of Fig. 14
C d1
X
_X
:I
e
FIG. 14. Synthetic response curves (third row) built up front assumed fundamentalresponse curves of the two receptor organs, cones and rods. Effect of latent period andnegative variation in second fundamental curves.
and in a of Fig. 15 the second maximum is represented as havinga latent period. The result in b of Fig. 14 is a curve very similarto the observed curve n of Plate IV although it is not assertedthat such a shape can only be explained by the occurrence of alatent period. Fig. 14 c shows the result of assuming a negativevariation preceding the second maximum; d and e of the same
26
9
RETINAL RESPONSE
figure and b of Fig. 15 give the results of adding a latent periodto the negative variation. The summation curve of Fig. 14 e isvery similar to i andj of Plate IV, but since the pause in the up-ward ascent of i and j develops into a definite maximum asshown in the following curves k and 1, it can hardly be explainedby the summation of the two curves of Fig. 14 e. This lattercurve is, however, very similar to curves r and s of Plate VIILCurves a and b of Fig. 15 are similar to f of Plate IV.
a b c
N
FIG. 15. Synthetic response curves (third row) built up from assumed fundamental
response curves of the two receptor organs, cones and rods. a and b are forms obtainableassuming latent period and negative variation of second fusndamental curve. c shows effect ofaddition of fourth inaximum.
Fig. 15 c shows the sum of three separate curves, the thirdbeing added to give the fourth gradual rise observed in so manyof the experimental curves. Such a third curve could obviouslybe added to all of the curves of Figs. 13, 14, and 15. Examples ofthe response curves containing a fourth maximum are representedin Plate V, c, d, e, f, g and h and in a-f of Plate VII
us
27Jan., 1923]
II
I
i
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
Fig. 16 shows assumed forms to explain the third maximum,and as is evident, summation curves are obtained which closelyresemble curves i, j, k and I of Plate IV.
In the preceding paragraphs an analysis of many of the observedexperimental curves has been given in terms of some assumedfundamental reaction curves shown in Figs. 13, 14, 15 and 16.A great variety of forms are obtained in experiment, but as is
a
Il
a
FIG. 16. Synthetic response curves (third row) built up frot assuned fundamentalresponse curves of the two receptor organs, cones and rods. Explanation of third maximumtassunting complex reaction for first fundamental curve.
evident from the above discussion, they can all be resolved intothe sum of simpler forms. It is clear, however, from Figs. 13-16that a great variety of summation curves can be obtained bychanging slightly the shape of the assumed fundamental curves,.their duration, or latent period. It is sometimes difficult todetermine whether a particular experimental curve involves alatent period or a preliminary negative variation in one of the
I~~I
28
II
RETINAL RESPONSE
fundamental curves, for as will be seen, the summation curvesof Fig. 13 b and Fig. 14 a and c are very similar. Although theexact shape of the fundamental responses has not yet beendetermined, the shapes assumed are reasonable and have beenworked out to correlate the experimental results with the knownperformances of nerves57
58, 59, 60, 63 A further discussion fromthis point of view will be given in a future paper.
5. Plate VI shows some abnormal curves obtained withinjured or modified eyes. Curves a to e inclusive were taken withfrog eyes and the remaining curves with guinea pig eyes. Curvesa and b of Plate VI were taken with an eye which, when freshlyexcised, appeared much less rigid than a normal eye. It wasapparently diseased. These curves were taken soon after exci-sion and show characteristics more or less typical of the responsecurves of the old eyes. Curves c, d, and e show the effect ofprogressively lowering the temperature of the retina. Thechanges that take place due to decreasing the temperature aresimilar to the changes that occur as the eye ages. Curves f, g, hand i, show the modification in the response curve of a guinea pigeye as it ages. The eye of a warm blooded animal ages in a muchshorter time than does that of a cold blooded animal. Al-though curve g was taken only 37 minutes after excision, thecurve is not the same as that of a normal frog eye, but shows thesame characteristics as an aged eye. The fact that the eye of awarm blooded animal dies much more rapidly than the eye of acold blooded animal leads one to wonder if the normal responsecurve for the guinea pig eye would not be much the same as thatof a frog eye if it could be taken soon enough after excision or if itcould be prevented from dying so rapidly.
6. Although the whole eyeball dies after a few hours as shownin Plate III, when only the posterior half of the eyeball is utilized,not only is the fine structure more clearly brought out and largerresponses obtained but the rate of dying is much less rapid.
In Plate VII are given a few curves selected from a series takento show the changes which occur as the posterior half of the eyedies. The series covered a period of twenty two hours after
Jan., 1923] 29
CHAFEE ET AL [J.O.S.A. & R.S.I., VII
which no response could be obtained. A set of three exposures of0.040 seconds each were taken periodically for light intensities of1.07 X 10-4, 1.71 X 10-3 and 0.11 meter candles. In Fig. 17 and 18the heights of the first, second and third maxima for the twohighest intensities of stimulation are plotted against time. Therecovery from etherization and from the operation is shown bythe curves and requires about two hours, after which the eyeremains fairly constant in sensitivity for a time but soon begins todie.
2400
> 1600
400 ~~~80 1 20AGE IN MINUTES
FIG. 17. Te heights of the 1st, 2nd, and 3rd maxima (in microvolts) of the responsesof the eye are plotted against the respective ages of the eye (in minutes). H, H2, and Haindicate tke curves obtained from the 1st, 2nd, and 3rd m7axiima respectively. The intensityof illumination =1.71 X 104 meter candles. The data was obtained from te series, apart of which is reproduced in Plate VII.
The curves of Fig. 17 show that the first and third maximadecrease before and more rapidly than does the second maximum.This would be more marked if the curves could be analyied intothe several component parts and the actual heights then measuredindependently of the other parts. Since this cannot easily bedone the ordinates on the plots are the actual heights of the totalcurve at the points of maxima. Therefore, each measurementconsists of the maximum of one component curve added to theheights of the other component curves at that point. The factthat the first response dies more rapidly than the second is wellshown by the change in shape of the curves shown in Plate VII.
30
RETINAL RESPONSE
The curves of Fig. 18 do not show the differential dying as wellas the curves for weaker illumination (Fig. 17) because, as will beshown in the following portion of the paper, the responses to strongstimulation are little altered by slight variations of the stimula-tion and therefore it may be inferred that small changes in thesensitivity of the eye due to dying produce little alteration in theresponses under such strong stimulation.24
H.
1600 __
AGE IN MINUTES
FIG. 18. The curves were obtained in the same manner and from the same seriesas those in Fig. 17, for the intensity of illumina tion, 1. 1 X10-1 meter candles.
7. The relation between sensation and stimulation has beenstated by Weber in a law bearing his name. (See ref. 56, p. 20and ref. 62, p. 505.) The just appreciable increase of stimulusbears a constant ratio to the original stimulus. This law stated inmathematical form is
SS = SIy
where S is the measure of the sensation, I the measure of thestimulus and K is a constant. Fechner, with somewhat doubtfuljustification, integrates the above expression obtaining Fechner'slaw which supposedly gives the law connecting sensation andstimulus. Fechner's law is,
S=K log I+Cwhen C is a constant of integration.
Psychological experiments with vision have demonstratedthe fact that Fechner's law does not hold for very weak or strong
Jan., 1923] 31
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
illumination but holds approximately over a limited range ofmedium intensities of stimulation of the retina.
Several of the early experimenters who measured the electricalresponse of the retina for variations in intensity of light believedthat there was a correlation of their results with Fechner's law.These early results were, however, obtained over very limitedranges of intensity and with imperfect apparatus so that they donot constitute noteworthy confirmation of the law connectingsensation and stimulus. This line of investigation has, however,been carried further as described below.
If it be assumed that the sensation is proportional to somecharacteristic of the response curves, then this characteristicshould vary with intensity according to the above law within thelimits between which the law applies to psychological observa-tions. In this case the mechanism which follows the law wouldnecessarily be in the retina and not in the central organs of per-ception. Conversely, if some characteristic of the response curvesis found to follow the Weber-Fechner law, it is possible, althoughnot certain, that sensation is proportional to that particularcharacteristic of the response curves. All confirmations ofpsychological observations which can be found in the electricalresponse curves give added evidence to the hypothesis that theresponse curves depict the course of sensation.
Many series of response curves have been obtained for widelyvarying light intensities. The question arises-which character-istic of the curves measures the sensation? The most naturalcharacteristic to select is the maximum height, and for want of abetter choice, this was used.
A series of 5 second exposures for light intensities at the retinaranging from 2.52 X 10-5 to 2.5 X 10-3 meter candles was made on aneye four hours old. The maximum height of each on-effect wasmeasured and plotted against the corresponding intensity oflight 1. The result is shown by the curved full line of Fig. 19, acurve resembling a logarithmic curve. If log I is plotted againstR, the response, the curved dotted line, which becomes straight atits upper end, is obtained, but if the logarithm of each coordinateis plotted, the straight line over the lower portion is the result.
32
RETINAL RESPONSE
The slope of this last mentioned line is 1.91 giving as the derivedlaw of response for the lower intensities of stimulation-
R1-91 cl or R oc 5- 2 3
or R c Vi, approximately.It is then evident that no simple law applies to the whole curve,but over the portion for weak intensities the power law verysatisfactorily holds but breaks down for intensities at the retinaabove 700 micrometer candles. Above this point the R-log Iline is nearly straight showing that Fechner's law is approximatelyfollowed.
L 6 2 X .lXM . ... . . . .. . . . . . . ..
FIG. 9. Curves showing the relation of the maximum height of the r . o the
usd (1 I-R (2 logg I-lg R, (3 lg iRweob y g t a s o w a
invariably ocr an arp cag in so of he IlIo
shown inF. urve soiThe eaio of thes au eig f the eponete
weis fo8 ien exoue ni very casel from th2.kihs
When higher intensities of stimulation were employed, thereinvariably occurred an abrupt change in slope of the log.I-log R
Jan., 1923] 33
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
curve, this break coming always at practically the same intensityof light (6.3 X10-4 meter-candles.) This abrupt change in lawbegins to show in Fig: 19 but is more clearly apparent in Fig. 20.
FIG. 20. Curves obtained in a similar manner as those in Figure 19 fromn a seriesin which I was carried to a greater value. The plot shows the break in the curves, whichmeans a change in law, at approximately the point where log I=2.4. (The value of 1varies for any given intensity, with the different eyes, but the break in the curve is prac-tically always near the point where log =2.4.)
As has been pointed out, the part of the curve above the changewhen plotted to coordinates log I and R usually yields a straightline, showing that for this part of the curve the Weber-Fechnerlaw is obeyed. For very high intensities, the law again fallsdown. This state of affairs is in entire agreement with psychologi-cal observations alluded to above stating that Fechner's law holdsfor a limited range of medium intensities.
An examination of the original response curves shows that forthe intensities below the first break in the law, the second maxi-mum of the curve was higher than the first and consequently wasthe height measured, while for intensities above the first break,the first maximum had increased until it had become larger thanthe second maximum and was, therefore, the height measured.Thus. the break comes at the point where measurements for the
34
RETINAL RESPONSE
maximum response changes from the one maximum to the other.It might be said, then, that for intensities of stimulation abovea certain well defined value Fechner's law of response holds, butfor stimulations below that the first law holds. This, however,must be considered with limitations because if there are two sep-arate mechanisms contributing to the total observed electricalresponse, neither can be measured separately, hence, the true law
2A
2.0
1.8
1.6
1.2
I.
I
0
FIG. 21. Diagram showing the values of the exponents of the log I-log R plots ob-taimedfrom various series. The average of these is 1.98.
of response of the separate mechanisms has not been determined.On the other hand, the total sensation must be due to the com-bined effect from both rods and cones so that, after all, is not thecombined effect of both mechanisms, and hence the total deflec-
Jan., 1923] 35
IAt
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
tion, a better measure of real sensation than either one separately?Perhaps the very fact that there are two simultaneously actingmechanisms, each possibly with its own simple law of response, isthe very reason why the observed psychological observations donot follow any one simple law.
Other characteristics of the observed response curves have otherpossible psychological confirmations. It is a natural inference toassociate the long fourth maximum with the positive after-image. The outstanding reason for this association is thesimilarity in time to the observed maximum of the responsecurve and the time to the maximum sensation of the after-image;also the fact that the time to the observed maximum in bothcases increases with intensity of exposure.
The sharp increase in the response curve of the "off effect"can be observed psychologically under proper conditions when adecided increase in sensation is obtained upon termination ofillumination.
It is interesting to note that it has been demonstrated byexperiment that in a general way the same changes in the shapeof the response curves take place for recovery from etherization,recovery from light adaptation, increase in intensity of light, andincrease in time of exposure.
Another point of interest is the extreme sensitivity of thecombination of a retina and the electrical amplifying and record-ing system. Deflections are obtained for light so weak that it isat or below the threshold of human vision. Deflections areobtained.for light intensity at the retina equal to that falling on asurface 1200 feet from a candle. All measurements have beenmade for light intensities within the range of normal vision.
VI. SUMMARY OF RESULTS
1. An improved apparatus is used in studying the electricalresponse of the retina. This apparatus makes use of a two-stagethermionic amplifier which makes possible the measurement ofsmall potentials with the absorption of no energy.
2. In preliminary work the electrical response was obtainedwhen using the whole eyeball and the resulting curves are similar
36
RETINAL RESPONSE
in principal characteristics to those obtained by previous experi-menters. See Fig. 7 and Plate III.
3. The response curve of the whole eyeball undergoes markedprogressive changes as the eye ages, as shown in Plate III. Themost notable change with age is the decrease in first positive in-crease in potential of injury and the development of a secondrelatively long negative deflection as shown in Plate III.
4. Much superior results are obtained by using only theposterior half of the eye-ball and making direct connection to asmall point on the surface of the retina.
5. Using the improved method of connection to the retina,the responses are greater and the response curves reveal muchfine structure showing that the reactions are more complex thancan be explained by the two or three substance theories.
6. The complex curves are analysed into four definite andtypical components called the first, second, third and fourthmaxima. The first and second parts appear to be independentof each other. The time from the beginning of response to thefirst maximum is practically constant and equal to 0.27 seconds.The time to the second maximum measured in the same manneris also essentially constant and equal to 0.70 seconds. Thetimes to the third and fourth maxima vary greatly according tothe energy of light stimulation. The time to the third decreasesand the time to the fourth increases as the energy increases. Itis believed that the first and third maxima are related.
7. The apparently unrelated first and second maxima can bereasonably attributed to the reactions of the two types of visualcells in the retina, the cones and rods.
8. Synthetic curves resembling many of the observed curvesare built up from two assumed types of fundamental curves.See Figs. 13, 14, 15 and 16. The two distinct types of funda-mental curves are considered to be the responses of the two typesof sensory receptors the cones and rods. The shapes of thefundamental curves have been chosen with due regard to theperformance of nerves.
9. The changes with time in the response curves of theposterior half of the eyeball are studied and graphs obtained.
Jan., 1923] 37
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
(See Figures 17 and 18). The shape of the response as well as itsmagnitude undergoes a change with age. Responses have beenobtained for twenty-two hours after excision of the eye.
10. The changes in maximum height of the response curves,used as a measure of the intensity of response, are plotted againstthe corresponding intensities of excitation. For low intensitiesthe law
Sensation cc -VStimulationapproximately holds: for medium intensities of excitationWeber-Fechner law (S cc K log I+C) is obeyed.
the
VII. DATA FOR THE PLATES III-VIII
All the curves except those of Plate III andf, g, t and i of PlateVI are responses of the posterior half of the frog's eyeball.
I=intensity of the stimulating light.T= time of the exposure of the eye to light.
Age = time from the excision of the eye and for all the givencurves in between the limits of one and four hours unlessotherwise indicated.
PLATE III. Eyeball Age Series.I= 500-watt lamp 50 cm from a piece of ground glass, which in turn was 42 cm
from the eye.T= 30 secs.
Plate letterabcdef
Age in minutes206095
150180210
PLATE IV.
Plate letterabcdef
gh
Typical Response Curves-Flashes.I in meter-candles T in seconds
.00128.00205 consecutive exposures.00445.013501.04.025 IC.025 I
PLATE V. Typical Response Curves.-Flashes and Time Expo-sures.
I in meter-candles T in Seconds1.34X10-3 0.103.45X10- 2 0.0252.61X104 0.108.22X10- 1
0.0251.10X10- 3
12.252.76X10- 1
10.00frosted 16 cp lamp 5.0042 cm from the eyeExposure: I, for 30 sec.; + 2 for 20 sec.; -I2 for 20 sec.,darkness; where
Il=500-watt lamp shielded by a piece of thin whitepaper, 542 cm from the eye and12 =a frosted 16 cp lamp 42 cm from the eye.
PLATE VI. Abnormal Response Curves.Plate letter Age in min., I in meter- T in sec.
candles,90 1.595 35.0
17.0\23.01
consecutive exposures
180 .2 0.1 Eye at room tem-Iperature Effect of
210 - .44 3.0 Eye below room loweredperature temperature
225 .44 3.0 Eye at still lowertemperature
67 870.037 54.455 870.0
105 96.7
0.15.0
10.05.0
IResponse of a guinea-pig'seyeball
Age Series of the Posterior Half of the Eyeball.
Age in minutes35
114162
I=.1X10- 1 meter-candles T= .04 sec.Plate letter
abc
Jan., 1923]
j
k1
n0
Plate letterabcdef
g
h
abC
d
e
f
ghi
PLATE VII.
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
d 520e 874f 966
g 1073h 1162i 1205
PLATE VIII. Responses for Varying Intensities of Stimulation.Series 1 T= .2 sec. Slope of the log I-log R line=2. 7
Plate letter I in meter-candlesa 583.0X10-5b 410.0X10-5c 168.0X10 4
d 106.OX 10-5
e 68.2X10-5f 42.8X10-5g 26.9X10-5h 18.6X10- 5
i 13.5X10-5j 8.1X10-5k 7.1X10-51 5.4X10-5
m 2.7X10-5
Series 2 T= .04 sec. Slope of the log I-log R line= 1.95Plate letter I in meter-candles
n 32. X 10-5o 54. X 10-5
p 89. X 10-5q 171. X10-5
r 312. X 10-5s 516. X10-5t 865. X 10-
5
u 1420. X10-5
v 2740. X10-5
w 4550. X 10-5x 6670. X 10-5y 10200. X10-5z 18930. X10-5
VIII. BIBLIOGRAPHY(a) ELECTRICAL RESPONSE
1. du Bois-Reymond. Untersuchungen iber thierische Elektricitat., 2: Abt. 1, p.
256; 1819.2. Holmgren. Method ot objektivera effekten af ljusintryck po retina.
Upsala Likaref6renings Forhandlingar, 1, p. 177; 1866.3. Holmgren. Centralbl. f. med. Wiss., p. 423; 1871.4. Dewar and M'Kendrick. Recent Researches on the Physiological Action of
Light. Nature, 8, 204; 1873. Also Journ. Anat. and Physiol. 7: 275; 1873.
40
Jan., 1923] RETINAL RESPONSE 41
5. Dewar and M'Kendrick. On the Physiological Action of Light. Trans. Roy.Soc. Edin., 27, 141; 1874.
6. Dewar. The Physiological Action of Light. Nature, 15, pp. 433, 453; 1877.7. Holmgren. Physiol. Untersuch, Heidelberg, 2, p. 81; 1879.8. Holmgren. Ibid. 3, p. 308; 1880.9. Chatin. Sur la valeur compar6e des impressions monochromatiques chez les
invert6br6s. C. R. 90, p. 41; 1880.10. Kuhne und Steiner. Physiol. Untersuch, Heidelberg, 3, p. 326; 1880.11. KlUhne and Steiner. Ibid. 4, p. 564; 1881.12. Holmgren. Physiol. Untersuch., Heidelberg, 2, 1882.13. Engelmann and Grijms. Ueber elektrische Vorginge im Auge bei reflectorische
und director Erregung des Gesichtsnerven. Helmholtz'sche Festgabe; 1891.14. Fuchs, S. Untersuchengen iber die im Gefolge der Belichtung auftretenden
galvanischen Vorgange in der Netzhaut und ihrem zeitlichen Verlauf. I Pflugers'Arch., 56, p. 408; 1894.
15. Waller, A. D. Points relating to the Weber-Fechner Law. Retina, muscle, nerve.Brain, 18, p. 200; 1895.
16. Beck, A. Ueber die bei Belichtung der Netzhaut von Eledone moschata ent-stehenden Actionsstrome.Pfluger's Arch., 78, p. 129; 1899.
17. Waller, A. D. On the Excitability of Nervous Matter with Special Referenceto the Retina. Brain; 1, 1900.
18. Waller, A. D. On the Retinal Currents of the Frog's Eye, excited by Light andexcited Electrically. Phil. Trans. R. S. and B., 193, p. 123, 1900.
19. Waller, A. D. The Retinal Response to Light. Brit. M. J., Lond., 2, p. 840;1900.
20. Waller, A. D. On the "Blaze-Currents" of the Frog's Eyeball Phil. Trans.,R.S.B. 124, p. 183; 1901.
21. Fuchs, S. Untersuchungen ber die im Gefolge der Belichtung auftretendengalvanischen Vorgiinge in der Netzhaut und ihren zeithichen Verlauf II.PflUger's Arch., 84, p. 425; 1901.
22. Himstedt und Nagel. Berichte der Naturforschenden Gesellschaft, 11, p. 149;1901.
23. Himstedt und Nagel. Versuche iber Reizwirkung verschiedener Strahlenartenauf Menschen-und Tierauge. Festschrift der Univ. Freiburg, p. 166; 1902.
24. Gotch, F. The Sub-maximal Electrical Response of a nerve to a Single Stimulus.J. Physiol., Lond., 28, p. 395; 1902.
25. Gotch, F. The Time Relations of the Photo-Electric Changes in the Eyeball ofthe Frog. J. of Physiol. 29, p. 388; 1903.
26. Gotch, F. Further Observation on the Photo-electric Response of the Frog'sEyeball. Proc. Physiol. Soc. Lond., 1, 1903.
27. de Haas. Lichtprikkels en retinastroomen in hum quantitief verband. Disser-tation, Leiden; 1903. Also Onderzoekingen Physiol. Lab., Leiden, 2nd ser. vol. 6.
28. Waller, A. D. The Signs of Life. London; 1903.29. Gotch, F. The Time Relation of the Photo-electric Changes produced in the
Eyeball of the Frog by means of Colored Light. J. of Physiol., 31, p. 1; 1904.30. Piper, H. Das elektromotorische Verhalten der Retina bei Eledone moschata.
Arch. f. Physiol., 453; 1904.
CHAFFEE ET AL [J.O.S.A. & R.S.I., VII
31. Piper, H. Untersuchungen fiber das elektromotorische Verhalten der Netzhaut
bei Warmblutern. Arch. f. Anat. u. Physiol., Suppl.-Bd., p. 133; 1905.
32. Ishihara, M. Versuch einer Deutung der photoelektrischen Schwankungen am
Froschauge. Pfluger's Arch., 114, p. 569; 1906.
33. Brucke, E. Th. und Garten, S. Zur vergleichenden Physiologie der Netzhaut-
strome. Pfiuger's Arch., 120, p. 290; 1907.
34. Kreidl und Ishihara. Photoelektrische Schwankungen an embryonalen Augen.
Zentralbl. f. Physiol., 21, p. 502; 1907.
35. Einthoven, W. and Jolly, W. A. The Form and Magnitude of the Electrical
Response of the Eye to Stimulation by Light at Various Intensities.
Quar. Journ. Exp. Physiol., 1, p. 373; 1908.
36. Westerlund, A. Studien fiber die photoelektrischen Fluktuationen des isolierten
Froschauges unter der Einwirkung von Stickstoff und Sauerstoff.
Skandin. Arch. f. Physiol., 19, p. 337.
37. Waller, A. D. On the Double Nature of the Photo-electrical Response of the
Frog's Retina. Quar. Journ. of Exp. Physiol., 2, p. 169; 1909.
38. Jolly, W. A. On the Electrical Response of the Frog's Eyeball to Light.
Quar. Journ. of Exp. Physiol., 2, p. 363; 1909.
39. Piper, H. Die Aktionsstrome der Vogel und Saugenetzhaut bei Reizung durch
kurzdauernde Belichtungen und Verdunkelung. Archiv. f. Physiol. Leipz; 1910,
also Suppl. Bd., p. 461; 1911.
40. Piper H. Verlauf und Theorie der Netzhautstrome. Zentralbl. f. Physiol.,
Leipz, 24: 1041; 1910-1911.41. Piper, H. Ueber die Netzhautstrome. Arch. f. Physiol., 85, 1911.
42. Trendelenburg, W. Die objectiv feststellbaren Lichtwirkungen an der Netzhaut
III Die Atktionsstrome des Auges. Ergebnisse der Physiol., 11, p. 21; 1911.
43. Nikiforowsky, P. M. Ueber den Verlauf der photoelektrischen Reaktion des
Froschauges bei Abkuhlung. Zeitschr. f. Biol., Munchen u. Berl., 57, p. 397;
1911-1912.44. Jolly, W. A. Onderzoeckungen gedaan i.h. Physiol. Laborat d. Univ. Leiden II.
8, p. 81; 1912.45. Frohlich, Fr. W. Vergleichende Untersuchungen fiber Licht- und Farbensinn.
Deutsch. med. Woch, 39, p. 1453; 1913.
46. Brossa, A. and Kohlrausch, A. Die Aktionsstr6me der Netzhaut bei Reizung
mit homogenen Lichtern. Arch. f. Physiol., p. 449; 1913.
47. Brossa, A. and Kohlrausch, A. Die qualitativ verschiedene Wirkung der ein-
zelnen Spektrallichter auf die Tiernetzhaut mittels der Aktionsstr6me unter-
sucht. Zentralbl. f. Physiol., 27, p. 725; 1913. Also Ibid. 28, p. 126; 1914.
48. Kohlrausch, A. and Brossa, A. Die photoelektrische Reaktion der Tag und
Nachtvogelnetzhaut auf Licht verschiedener Welleniange. Arch. f. Physio., p.
421; 1914.49. Kohlrausch, A. Die experimentelle Analyse der Netzhautstrome an der Taube.
Zentralbl. f. Physiol. Leipz. u. Wien, 28, p. 121; 1914.
50. Kohlrausch, A. Die Aktionsstrome der Wirbeltiernetzhaut bei Reizung mit
Lichtern verschiedener Wellenlange. Zentralbl. f. Physiol. Leipzig u. Wien. 28,
p. 759; 1914.51. Day, E. C. Photoelectric currents in the eye of the fish. Am. J. of Physiol.,
38, p. 369; 1916. Also Science, Lancaster, Pa., n.S. 43: 144, 1916.
42
RETINAL RESPONSE
52. Westerlund, A. Light Stimulus and Photo-electric Response. Lunds Univ.Arssirift, N. F., p. 12; 1916.
53. Riedel, A. H. Ein Beitrag zur Kenntnis der photoelektrischen Reaction desHummerauges. Ztschr. f. Biol., 69, p. 125; 1918.
54. Sheard, Chas. and McPeek, C. On the Electrical Response of the Eye to Stimu-lation by Light of Various Wave Lengths. Am. J. of Physiol., 48, p. 45; 1919.
55. Beuchelt, H. Die Abhingigkeit der photoelectrischen Reaktion des Froschaugesvon den ableitenden Medien. Zeit. Biol., 73, p. 205; 1921.
(b) OTHER REFERENCES56. Parsons. An Introduction to the Study of Color Vision.57. Lucas. The Conduction of the Nervous Impulse.58. Troland, L. T. The Physical Basis of Nerve Functions. Psychol. Review, 27,
.p. 5; 1920.59. Troland, L. T. The Progress of Visual Science in 1919. Am. J. of Physiol.
Optics, 2, p. 232.60. Troland, L. T. The Nature of the Visual Receptor Process. J. of the Optical
Soc. of America, 1, No. 1, 1917.61. Forbes, A. and Thacher, C. Amplification of Action Currents with the Electron
Tube in Recording with the String Galvanometer. Am. J. of Physiol., 52, p. 409;1920.
62. Burch, G. J. On Light Sensations and the Theory of Forced Vibrations.Proc. Roy. Soc. Lond., 86: sB., p. 490; 1913.
63. Troland, L. T. The "All or None" Law in Visual Response.J. of the Optical Soc. of Amer.; May, 1920.
Jan., 1923] 43
[J.O.S.A. & R.S.I., VII
PLATES I AND II
1. Taking Light.2. Electrostatic Shielding.3. Einthoven Galvanometer.4. Time Marker.5. Photographing Apparatus.6. Adjusting Galvanometer.7. Shielded Amplifier.8. Eye Chamber.9. End of Photometer Bar.10. Mercury Switches.11. Shielded Motor for Driving Photographing Drum.
12. Shielded Control Switches and Leads.13. Shielded Plate Batteries.14. Wedge Used in Later Work.15. Monochromatic Illuminator Used in Later Work.
