i${i,t'#/q#rBrit. J. Ophthal. GgZ6) 6o, r35

Clinical assessment of rhodopsin in the eyeUsing a standard fundus camera and a photographic technique

U. B. SHEOREYFrom the Department of Visual Science, Institute of Ophthalrnology, London

In recent years, the technique of fundus reflecto-metry (see Weale, tg74), has been shown to beparticularly helpful in locating defects in the pri-mary stage of the transduction of light into vision.It has thus been possible to assess the availabilityof retinal pigment in cases such as night blindness.When the results are considered together with thosefrom colnplementary electrophysiological or psycho-physical tests, a much clearer diagnosis concerningthe site of the lesion is possible. The basis of themeasurement is the marked change in the absorp-tion spectrum of the visual pigment rhodopsinwhen it is bleached by strong exposure to light.Thus if a defined part of the retina is bleached-that is, an imprint or an 'optogram' is formed onthe retina-a comparison of the reflectances of theunexposed and exposed areas can be used as anindicator of the availability of pigment in thatregion. The measurement can be carried outonly for a double transit of light through theretinal layer provided that the bleaching of theeye has affected only the retinal pigment.

In most previous clinical studies, complexapparatus r,vas used and the procedure was involved.We have been able to use a simple procedure and aconventional fundus camera, albeit slightly modi-fied. The use of a standard fundus camera and aflash bleaching technique has obvious clinicaladvantages-for example, the co-operation re-quired of the patient is minimal. These modifica-tions are described in this paper and the use of thistechnique in assessing the pigment concentrationin rod rich parts of the retina is illustrated. Thesemodifications were designed so that they couldeither be easily removed from the light path in thecamera or left in place provided they did nothinder the original application of the camera. Asvery little machining is required on the main bodyof the apparatus, its absence from the clinic isshort. An approximate assessment of the densitycontrast of the optogram can be made immediatelyafter developing and fixing the latent image on thephotonegative. A value for the density contrastAddress for reprints: U. B. Sheorey, Department of Visual Science,Institute of Ophthalmology, Judd Street, London WCIH qQS

can be obtained later by means of a microdensito-meter. The modified apparatus thus provides asimple procedure for routine and rapid clinicalassessment of rhodopsin in the eye.

MethodThe spectral absorption of the visual pigment is differentfor the bleached and the unbleached states' and theintensity of the light beam which has traversed throughthe retinal layer containing the pigment will be modifiedaccordingly. We may compare the intensities eitherfrom two adjacent areas' one of which is bleached, orfrom a given area before and after it is bleached. Ifthisdifference is due orly to a change in the retinal pigrnent,we can relate the changes in the concentration of thatpigment to changes in optical densities. The change inoptical density is a function of the wavelength of lightused for the measurement; it is largest in bluish-greenlight for the rod pigment.

The present method uses a photographic techniqueto record this difference from adjacent areas of theretina after a defined part of the retina has been bleached.In order to relate the recorded density differences totrue differences, the photonegative also carries a cali-bration which is superimposed at the same time asthe recording. Thus, distortions which may be intro-duced during the process of the development of thelatent image on the film can be compensated (Arnold,Rolls, and Stewart, ry7t).

APPARATUS

The basic apparatus consists of a standard Zeiss*fundus camera and its associated electronic powersupply unit for the xenon flash tube. The paths ofthe light beams which are used to provide theretinal illumination for general viewing, flashbleaching, and photography are unfortunatelylargely common. In order to produce an optogramby bleaching a defined part of the retina, an aper-ture in this light path is required, but it has to beremoved for general viewing of the fundus andduring photography.

These modifications are illustrated schematicallyin Fig. r. The original tubular holder which sup-ports the objective lens L is replaced by a modified*C. Zeiss, Oberkochen, Germany

L- ...sk,

136 British Journal of Ophthalmology

housing. The axial length of the housing and themechanical dimensions at both its ends are exactlythe same as for the original tube so that the twocan be interchanged. The housing accommodatesa moveable carriage which in turn accommodatesan aperture-plate. The apertures can be movedalong the optical axis by rotation of the shaft S, sothat they can be made conjugate with the patient'sretina. Either aperture may be brought into thelight path by means of the knob Kr. A shieldedpress-switch, Sw, is mounted at the end of K, toactivate the xenon flash B, for retinal bleaching.This switch is wired electrically in parallel withthat on the mechanical shutter of the 35 mmcamera. This duplication of the switch is desirableto enable the flash to be activated independentlyof the camera and to reduce the number of handmovements by the operator during the procedure.

The aperture-plate P carries two apertures, A,for normal viewing of the retina and for photo-graphy, and 42 for defining the retinal area that is tobe bleached (Fig. rc). The plate is mounted perpen-dicular to the optical axis of the fundus camera andin the moveable carriage. Both apertures are madefrom thin sheets of brass and can easily be removedand exchanged for other shapes and sizes-forexample, as might be required for the investigationof the density of the retinal pigment along thevertical meridian. The aperture A, is currentlyz8 mm diameter, equivalent to a retinal view of35'. Access to these apertures may be obtainedsimply by removing the lens L. All internal sur-faces of this housing and the additional componentswhich are mounted in the fundus camera arepainted matt-black so as to prevent stray reflectionswithin the camera reaching the patient's eye or therecording camera.

F1 is the deep-red colour filter that is requiredin the path of the white light from the bulb B, toprovide a safelight for retinal viewing. It is con-structed from a sheet of coloured gelatine (Ilford,no. 6o8) which is clamped between two thin platesof glass and held rvithin a metal plate. A rectangularslot is required in the side panel of the apparatusfor the insertion or removal of the filter from thelight path. Note that F, does not need to be handledsince it does not interfere with either the bleachingor the recording steps. Two miniature light-emitting diodes* are also provided within thehousing as additional fixation targets. These diodesare required because the patient cannot see theinternal target T once the aperture A1 is changedto A, if his centre of fixation is outside the areadefined by Ar. Furthermore, since the paths of theviewing and the bleaching lights are largely com-mon, T is not illuminated when the bluish-greenfilter is introduced into the light path before*Motorola, type MLED zo

photography. Few people, however, have anydifficulty in maintaining fixation during the briefinterval required to introduce the bluish-greenfilter and photograph the optogram.

The safelight path is combined with the path ofthe xenon flash light by means of the half-mirrorBS. In order to remove the infrared content ofthe intense xenon-flash, BS was remade from apiece of heat-absorbent glass. The bluish-greencolour filter F, which is required during therecording of the optogram, is placed in the lightpath after BS. The rest position of this filter isout of the light path as it is manually pushedinto place immediately before photography andthen retracted. Within the same mount, neutral-density filters may also be incorporated so as toadjust the retinal illumination independently ofthe power supply to the flash tube. It is therebypossible to achieve a common setting for bleachingand photography, thus considerably simplifyingthe procedure and avoiding waiting for the elec-tronic power supply to discharge and rechargebetween changes in settings.

The calibration is superimposed on the film byplacing a density-graticule DG in the recessbetween the film and the mechanical shutter of thecamera; therefore, it is only in the path of thelight reflected from the fundus. A typical graticuleis illustrated in Fig. fi. The part of the film onwhich the fundus image is recorded is indicatedby the broken circle; the inner rectangle indicatesthe position of the optogram image relative to thegraticule and one frame of the film. The graticulesare made by exposing a series of strips-for ex-ample, (zo mm long, z mm wide, spaced z rrrrnapart) on to photographic paper*. For present use,the exposure time was simply varied in steps duringa trial run and then the appropriate exposure timeswere selected so as to produce strips of opticaldensities in the range o'o5-o'r5. The position, size,and orientation of these calibration steps are deter-mined by the particular densitometric investigationthat one wishes to perform. For example, for in-vestigations of the changes in the pigment con-centration along the vertical meridian, it would beuseful to orientate these strips horizontally-that is,perpendicular to the bleaching rectangle. It isessential that an adequate calibration is recorded onthe same film in the vicinity of the region of interest.

CALIBRATION OF THE FLASH SOURCE

The flux incident upon the patient's retina duringthe flash bleaching and photographic steps wascalculated from the radiometric measurementswhich were taken on the xenon .lish. To do so,a model eye was formed in front ;f lens L and a*Ilford N+ESo

sfl.}..' Clinical assessment of rhodopsin in the eye r37

(c )

FIG. r (a) Schematic side aiew of Zeiss fundus camerashozuing princip al c omp onent s and mo difi c ations'Tubular housing at front houses objectiue lens L, andmoaeable holder for aperture plate P. Also shown are :adjustable, internal fixation target T; mirrors M;combined heat-filter and beam-splitter BS; filter F2for photographic step ; heat-absorbent filter HF;deep-red filter Fr; white zsiewing light Bt; xenon flashtube B2; calibration graticule DG1' camera shutter S;additionat fixation target, LED' Cylindrical assembly toteft of lens L is model eye used to measure flux arriuingat simulated retinal plane. During use, assembly is pushedoaer mounting tube of lens L and distances are fixedby spacing rings Sp. Output of photodetector D is fedinto ampffier A for measurements

(b) Schematic plan of modified fundus camerashowing additional fixation targets LEDs and aariouscontrol knobs and shafts. Rotation of Krmoaes plateholder P along optical axis for focus adiustments.Bleaching aperture is brought into light path by pullingK2 outwards. Flash can be actiaated by szuitch Sw.This is electrically connected to be in parallel withthat on shutter release switch of camera

(c) Front uiew of aperture plate holder P whichcarries tzpo apertures, A1 for normal aiewing of fundusand A2 for bleaching a defined area of fundus

(d) Schematic illustration of density-graticule DGu:hich is placed be4uteen the shutter of camera andfilm so as to impress a calibration on film duringphotography. B.tken line circle, corresponding toaperture A1, indilates area within zuhich retinal imageis formed. Inner rectangular region is area, correspondingto A2, within which optograrn is present

photodiode was placed within it. This assembly isshown in Fig. ta to the left of the objective lens L.The lens of the model eye had a focal length ofr7 mm and an aperture of 8 mm diameter. Anobjective radiometric method using a photodiodewas used in preference to a photometric methodusing an eye because it is easier and more preciseto use routinely in the clinic. The difficulty inmaking routine photometric measurements by eyewith this fundus camera should be noted' Theflash source is imaged in the pupil plane of theobserver's eye; this image is not uniform and isalmost as large as a fully dilated pupil. Furthermore,the output flux of the flash varies according to theinterval between firings and the age of the tube'

The waveform of the xenon flash (measured at1:5oS nm) consisted of a fast rising front ofabout-o'r ms (measured from 5 per cent of peakto the peak) followed by an exponential decay with acharacleristic decay time of about o'4 ms (from thepeak to r/e of the peak, where e: 2'72).For photo-graphy of the optogram, the spectral distribution andLt"nsity are defined by the transmittances of theocular media, the filters, and the optical componentsr,r,'ithin the fundus camera. The retinal energydensity during photography is calculated to be7 Jm-' when using setting no. + on the flashcontrol unit and filters Fr*Htrr* zElo'+ ND' Theocular transmittance has been taken as 7 5 percent for this calculation; details of this measure-ment and the calculation are given in the Appendix'

For the bleach, F2 and the neutral-densityfilters are removed from the light path. The increasein the spectral bandwidth and transmittanceresults in an actinic retinal energy density which isexpressed by Te x 9oo Jm-', where Te is thetransmittance of the ocular media. Details of thesecalculations are also given in the Appendix' Takinga mean value of T":6o per cent over the spectralrange 43o-63o nm, the bleaching flash is estimatedto pto"iae an actinic retinal energy density of

05'4X ro" Jm ".- In previous studies on fundus densitometry,the retinal fluxes have been specified in photo-metric units. These may be related to the valuesgiven above by means of the following:

r (scotopic) troland second: r'or X ro-* Jm-',for ).:5o5 nm and a pupil diameter of 8 mm'Thus, while the value given above for photographycompares well with the value of 4X roa troland swhich was used by Highman and Weale (rg7Z),the bleaching flash currently used is about threetimes less intense than their figure of r'8zX ro7troland s. The relationship between the fraction ofpigrnent bleached and the retinal illuminance(pig. g, Ripps and Weale, ry6g) indicates that thepresent flash will bleach about 55 per cent of thepigment. Thus, if the maximum optical density

(o)

-;;;:)'1r',,'.t,t'jt:')' I| !:iL.:;: I

(d)A2

Sw

K, --l

r38 British Journal of Ophthalmology

difference that can be achieved in the eyes ofnormal subjects is taken to be o.16-o.18 D, wemay expect to measure a density difference ofo'og-o'ro D in these subjects with the presentequipment.

The alignment and focus adjustment of thecamera were carried out by using the red filterF. in the path of the white light from the bulb Br.On setting z of the Zeiss power supply the retinalillumination was calculated to be 35o troland. Theluminance of the image of the source formed in thepupil of the model eye was measured by meansof a photometer*. The luminances of the additionalLED fixation targets were adjusted by varyingtheir input power so that they appeared to havethe same luminosity as the surrounding field. Thepower spectrum of these LEDs is negligible atwavelengths shorter than 63o nm and their emissionis confined to a narrow band in the deep red partof the spectrum.

ProcedureThe patient is seated at the fundus camera in a darkenedexamination room. The pupil must have been fully dilatedand dark adaptation allowed to proceed for at least r 5 minbefore this. The plane of the film and the apertures areadjusted rapidly so as to be in focus with the retina.Since it is difficult for the operator to see much retinaldetail in the dim red light, it is helpful to begin theadjustments while viewing the region around the opticdisc which has a larger reflectance and is rich in struc-ture. Note that a slight re-adjustment will be necessaryto view different parts of the fundus but no correctionfor chromatic aberration is required. The requiredretinal area is then brought into the centre of the fieldof view with the moveable internal fixation target or oneof the LED targets and the focus checked. A retinalarea in which there are relatively few blood vesselsshould be selected as far as possible, so that the tracesrecorded by the scanning densitometer are smooth andthus easier to analyse. The aperture A2 is brought intothe optical path by pulling the shaft Sz, the xenon flashis activated by means of the switch Sw, and the shaftis pushed back; this completes the bleaching step.If the eye is still correctly placed, the filter F2 is pushedinto the light path, the retina is photographed, and F2is pulled back to its resting position. The camerashutter is set nominally to rfz5 s. The film is thenadvanced by one frame so that the camera is ready foranother photograph or the next sequence. A readilyavailable sensitive filmt is used for this reading. Thefilm is processed in complete darkness and accordingto the manufacturer's instructions.

Careful alignment of the patient's eye to the apparatusis particularly important in this technique, because it isessential to transmit all of the available flux into thepatient's eye for bleaching. During photography, thespurious reflections and scatter from the eye that canso easily mask the image of the optogram, must be*Model 4oA, made by UDT Inc, California, USAtKodak z+Zs

avoided. Note that with the subject's fixation on theperiphery of the apertures A, or Ar, the pupil willappear to be elliptical as viewed along the optical axisof the apparatus. Thus there is a reduced margin forthe transverse alignment of the image of the light source,formed in the subject's pupil. A misalignment u'ouldmean that a part of the beam would be occluded andthis could cause a serious reduction in the retinalillumination. For a given patient, the focus and thefixation target require little, if any, adjustments onrecalling the patient after a period of rest or furtherdark-adaptation. In order to ensure accurate fixationduring the course of this procedure, the fixation targetshould be presented to the same eye as that used forrecording the optogram.

Results and the analysis of the optogramA positive print of a typical retinal photographtaken with this modified fundus camera is shownin Fig. z. This was recorded from a subject inrvhom dark-adaptation and vision were normal.The rectangular optogram can be clearly seen inthe centre of this photograph as an area of loweroptical density. This area is crossed by the regularlyspaced stripes caused by superimposition of thedensity graticule. It is important to note that theoptical densities are reversed during the process ofthis photographic printing-that is, the bleachedarea in reality reflects more bluish-green lightthan the surrounding unbleached region so thatit is more dense on the photonegative. The aper-ture A, used for bleaching was centred about r5'temporal to the fovea, along the horizontal meridian.The backs of the two LED fixation targets which

FIG. 2 A positiae photographic print cf typicaloptogram recorded from normal fundus. Subject'spoint of fixation is marked by LED target at left edge

#,"$/

are provided within the lens housing can also beseen; the subject's fixation is marked by the illu-minated LED target at the left of the photograph.

In order to obtain quantitative results from thefundus photograph, the photonegative has to bescanned in a microdensitometer; a double-beamrecording microdensitometer* was used for thepresent work. The displacement of the trace-recorder, due to a change in the optical transmit-tance of the scanned sample, has to be initiallycalibrated; this u'as done by introducing knownvalues of neutral-density filters in the path of themeasurement beam. Then, using common experi-mental conditions for both measurements-forexample, the same spot size and scan rate-thedensity-graticule and the photonegative are scannedin turn. The difference between the scans takenthrough the bleached area and the adjacent un-bleached area is proportional to the true opticaldensitlr difference. The proportionality constantis the transfer characteristic of the particular partof the film (Arnold and others, rgTr). For a givenregion of the film if the step change due to thesuperimposed densit-v graticule is the same forboth scans, then this constant is unity. Thusdeviations from unity can be noted and the resultsscaled accordingly. The procedure can best beillustrated by referring to Fig. 3 which shorrys thetraces (corresponding to the scans taken throughthe sections indicated in Fig. z). The traces aredrawn about a datum line which is taken to be thebase or fog level of the film. A scan of the particulardensity graticule which was used for this photo-graph-is also shown in Fig. 3. Note, however, thatthis third trace has been drawn inverted so as toillustrate the point that the exposure received bythe film r,vhich lies under the calibration strips isreduced. This local reduction of exposure can alsobe seen in the other two traces which are pulleddown towards the datum in the regions containingthe calibration strips. The calibration for the grati-cule is indicated in the centre of this figure. Thetransfer characteristic of this photonegative can'be*Model 3c, made by Joyce-Loebl Ltd

Clinical assessment of rhodopsin in the eye r39

seen to be slightly above unity. The density contrastof the optogram formed in central part of thisphotograph-that i., t5-2o" temporal-is abouto'r D. The density difference can be determinedmore easily, in practice, when all three traces areavailable on sparate sheets of transparent graphpaper. The sheets containing the scans taken out-side and through the bleached region are thenviewed against the sheet for the graticule. Thegraticule is once again held inverted and used as atemplate.

These traces illust:ate several additional interest-ing features. The traces are noisy and the reflectancevaries across the retir:a. This variation is due to avariety of reasons, partly in the instrument andpartly in the subject's eye-for example, thepresence of blood vessels in the path of the scanor a non-uniform retinal reflex. Note, however,that the scale of this variaticn is either much smalleror else much larger than that due to the modulationimpressed by the graticule on the fundus photo-graph. Thus it is easy to discriminate between thetrue density changes and the artefacts. The tracesalso illustrate the variaticn cf the density differenceacross the retina. This variation is practically zerofor the part of the retina nearer than roo temporalto the fovea. It rises rapidly beyond this to abroad maximum in the region r5-3o' temporal,in general agreement with the distribution of rodsin the retina. For routine clinical investigations,one would be interested chiefly in the centralportion of the photograph. Thus, as indicated above,a graticule is made so that it provides a densitystep of a value which is of the order of the densitydifference that one expects to measure from normalsubjects. A marked reduction in the density differ-ence recorded in a patient's eye would thus bereadily seen to be much smaller than this step. Thetwo outermost strips are well outside the expectedrange. llowever, they illustrate the fact that evenwhen the illumination on the film is reduced byo.35 density units, an appreciable density differencecan still be recorded. This difference can of coursebe achieved only from optograms formed in regions

>.=

cq,-9.E.v

.o_o

FIG. 3 Trace recordings produced by scanningmicrodensitolneter. Traces t and z are due toscans taken through corresponding sections whichare indicated in Fig. z. Lowest trace is scan ofparticular density graticule used to superimposecalibration. Scan has been drazun int-terted so as toindicate that exposure of fiInt underlying densitystrips is reduced. Ordinate represents opticaldensity ; abscissa represents distance alonghorizontal meridian on fundus

------!>

*!:.

-t:

*br4o British Journal of Ophthalmology

of the retina distal from the macula. Thus thefilm as such appears to have an operational rangewhich encompasses the gross variations of retinalreflex across the retina. Note also that in thisgraticule the densities of the steps increase awayfrom the vertical midline. Thus a graticule suchas this may be used to photograph the optogram incorresponding regions of either eye without requir-ing the replacement of graticules during measure-ments on both eyes.

DiscussionThis example of the use of the modified funduscamera was typical; a clinical study of variousretinal disorders using this apparatus is in progressand full results will be described in due course.Thus the comments below refer mainly to pointsof technique. In regular use, this modified funduscamera has proved to be no more difficult to usethan is the normal camera for routine fundusphotography. As the method provides a rapid,convenient means of supporting the measurementsof associated retinal tests in diseases such as night-blindness, routine densitcmetric measurementsmay also be carried out in the same clinic with thisapparatus.

The maximum flux available from the presentequipment is insufficient to bleach all the visualpigment within a given region. The flux could beincreased but only by modification, perhapsextensive, of the optics or the flash unit. Flowever,as establisheJ by Highman and Weale (1973) incases such es retinitis pigmentosa, the densitydifference; in a given retinal region are much lower,and the vis,ral thresholds of the same retinal regionare much higher, than those in normal subjects.Therefore, for the assessment of such cases theapparatus may be used without any further modifi-cations. It shculd be noted that simply increasingthe incident flux rnay not give a markedly denseroptogram, because short intense flashes of lightcan bleach only up to a limit of 7o-8o per cent ofthe available pigment because of photoregeneration(Williams, rg74. Furthermore, the associatedheating effect of the intense retinal illuminationmay well produce a deleterious effect-that is,oedema. The optical effect of this oedema ispartially to counteract the density difference dueto the bleach (Weale, r968). Its magnitude, how-ever, is said to be independent of the wavelengthof light used for the measurement so that it may bequantified by measurements outside the absorptionspectrum of the pigment. In order to ascertainwhether oedema had been caused, a series of fouror five fundus photographs were taken in deep-red light immediately after the retina had beenbleached and at half-min intervals subsequently.

The filter F2 was replaced by a suitable deep-redcolour filter*. In these photographs, no densitydifference corresponding to the bleaching aperturecould be seen in the region of the optogram. Theexperimental conditions in the earlier fundusdensitometric measurements were markedly dif-ferent from those used currently-for example,Highman and Weale GgZi used a continuous ro sintense exposure for bleaching-and thereforecomparative assessments of the various proceduresare difficult to make. In the previous studies, onlythe photometric effect of the bleaching light wasspecified, whereas in the present approach, boththe actinic and the total retinal flux densities areestimated. Thus the total flux entering the eye canalso be calculated (see Appendix).

This apparatus cannot be used in its presentform at large viewing angles-that is, no larger thanabout 15' in either the vertical or the huizontalmeridians. In order to do so, the image of theflash source which is formed in the subject's pupilwould need to be reduced without reducing theincident flux. This reduction in the pupillary areaas a function of the viewing angle was meajured byJay (r96r). A smaller pupillary image wouldadditionally reduce the variation in the retinalillumination due to the involuntary eye movements.These are particularly noticeable in older patientsin whom more care is required for alignment. Notethat a sli3ht variation in the retinai illumination-for example, as a result of ocular absorption orsmall eye movements-does not restrict the use ofthis reflectometric method of measurement becausea differential measurement is involved. The reduc-tion in the retinal illumination during the bleachingstep, however, could be significant since thedensity contrast of the optogram would be reduced.The reduction in the flux as a result of absorption-for example, in the lens-could be compensatedby an appropriate increase in the incident fluxusing the data established by Said and Weale(rgSg), or by a specific measurement in a particularcase.

SurnrnaryA technique based on the method of differentialfundus reflectometry is used to assess the avail-ability of rhodopsin in the eye. A defined part of thedark adapted fundus is bleached by a short intenseflash of light. The fundus is subsequently photo-graphed in order to record the flux reflected fromthe bleached area, the optogram, and the surround-ing unbleached region. This procedure requiresonly a few simple modifications to a Zeiss funduscamera before it can be used routinelv in theclinic.*Ilford No. 6o8

*'

dr

Filter

The use of the Zeiss fundus camera for fundus densito-metry was suggested by Professor R. A. Weale whosekeen interest during the modifications and the clinicaltrials was invaluable. The modifications to the camerawere carried out by Mr G. Mould.

Thanks are also due to Mr P. E. Cleary for collabora-tion with the clinical development.

AppendixMEASUREMENT AND ASSESSMENT OF THE RETINAL FLUX

The absorbances ,A.1 of the various spectral filters usedin the procedure and for calibration were measured ina dual-beam scanning spectrophotometer*. The resultsare tabulated below:

Clinical assessment of rhodopsin in the eye r4t

amplitude points) was less than zo ps, a value con-siderably less than the rise-time of the xenon flash. Acapacitor of ro pFarad was used in the feedback pathof the operational amplifier in order to measure thetotal charge of one pulse of photocurrent. UsingF2+HFe+zE-lr':ND filters within the light path ofthe fundus camera and setting no. 4 on the flash controlunit, the output was r3of 3'7 mV (mean of five flashes,ignoring the first flash). Thus the charge accumulatedwas r'3 x ro-6 C per flash. The response sensitivity ofthe photodiode is o'zr7 AlWatt, at tr: 5o5 nm, and theactive area is 5'r X ro-0 m2. Thus the energy densityreceived by the photodiode at the simulated retinalplane is lr8 Jm-2. Note that when the flash is firstfired after a rest period of a few minutes, the flux fromthe first flash is about 5 per cent higher than that of thesubsequent flashes which are fired at regular intervalsof, say, 5 or ro s.

During the retinal photography step, the filtersFr+HFr+zB*t+ ND are used in the fundus camera.Taking the transmittance of the ocular media as 75per cent at tr:5o5 nm (Table z'J,Wyszecki and Stiles,1967) the retinal energy density is 7'o4 Jm-2. Theretinal energy density may then be calculated for anygiven combination of filters inserted in the funduscamera, given the spectral distribution of the xenonflash-discharge source. Although there are severalnarrow peaks in the 46o-49o nm band, in most applica-tions the distribution may be taken to be essentiallyflat over the range 4oo-63o nm. Thus the use of abroad spectral band for bleaching, say, from 43o-63o nm,increases the total flux by about 3oo times more thanthat used during photography. If we take the meantransmittance of the ocular media as 70 per cent, theretinal energy density within this spectral band isabout r'5 x ro3 Jm-'.

In order to calculate the increase in the actinic fluxduring the bleaching step, the absorption spectrum ofrhodopsin was used (Crescitelli and Dartnall, r953).In the spectral range 43o-63o nm, the increase is 9.6times more than that in a ro nm bandwidth centredat 5o5 nm. Thus, the removal of the neutral densityfilter and F2 from the light path will increase the actinicflux at the retinal plane to: r'r8 x9'6xlogro-t (r'9) xTeor 9oo x T" Jm-2.

Absorb-ance(density) .A1 (at 1:5o5 nm)

F2HF,'Wratten zE

o'6oo'07o'o3

(Half power bandwidth: ro nm)

The flux density in the retinal plane which wassimulated within the model eye was measured by afast-response photodiodef. The photodiode was oper-ated in the photoconductive mode, using a bias of-45 V on its guardring electrode. The photocurrentwas fed directly into an operational amplifier which wasconstructed from an integrated circuit, type LM 3o8.This was wired so as to perform either as a linear cur-rent-to-voltage converter or an integrator. Thus theenvelope of a photocurrent pulse resulting from thedetection of a xenon flash could be viewed on an oscillo-scope or the integral of the current pulse could bemeasured by means of a voltmeter. The linearity ofthe operational-amplifier circuit was checked by inject-ing pulses of current from a pulse generator and throughsuitable known values of resistors. The response time(measured between the ro per cent and 9o per cent*Perkin Elmer, model 4oztSGD rooA, a silicon PIN type, supplied calibrated by E. G' andG. Inc, Boston, IJSA

ReferencesARNoLD, c. R., RoLLS, p. J., and srrwaRT, J. c. n. (rg7r) 'Applied Photography', Focal Press, LondoncREScITELLI, F., and DARTNALL, H.J. A. (tsS:) Nature (Lond'), T72' tgsHIGHMAN, v. N., and wEALE, n. e. (r973) Amer. J. Ophthal',75r 822JAY, B. (r96r) Vision Res., t, 4r8Rlpps, H., and wEALE, n. e. (1969) J. Physiol. (Lond.),2oo' r5rsAID, F. s., and wEALE, n. e. (rg59) Gerontologica (Basel),3t 2r3wEALE, n. a. (1968) Nature (Lond.), zr8,238

GgZ+) Ophthalmologica ( Basel) , 169, 3owrLLrAMS, T. p. Gg7+) Vision Res., r4r 6o3wrrszEcKl, c., and srILES, w. s. (1967) 'Colour Science', Wiley, New York