Partial Performance Degradation of a Remote Sensor in aSpace Environment, and Some Probable Causes

John J. Horan, Daniel S. Schwartz, and James D. Love

The Multispectral Scanner (MSS) was launched on the Earth Resources Technology Satellite (ERTS-1)

23 July 1972. The MSS has two calibration systems, one internal and one external. Both calibrationsystems have shown strong, spectrally dependent performance degradation since launch. This paperpresents details on the optical system of the MSS and data on the performance degradation as a function

of both spectral interval and time in orbit. The history of the MSS during tests is traced, and it is

shown that hydrocarbons from an external source may have been deposited on optical surfaces in the in-

strument. It is postulated that these contaminant coatings may have polymerized as a result of the ex-

posure to uv light from the sun, increasing their blue absorbtion and accounting for the observed perfor-mance degradation. Arguments supporting this postulate are presented, and other possible sources of

the performance degradation are discussed.

Introduction

The Multispectral Scanner (MSS) was launchedon the first Earth Resources Technology Satellite(ERTS-1) 23 July 1972. This instrument is a veryhigh resolution line scanning imaging radiometer(Fig. 1) operating in the visible and NIR spectra. Inaddition, it contains two on-board calibrationschemes one of which utilizes direct solar energy.We have measured a significant change in the mag-nitude of these calibration signals after launch.This paper will briefly describe the spacecraft'sorbit and the solar input to the instrument, the MSSinstrument itself, and the calibration schemes. Wewill then discuss the ground calibration and presentdata on the orbital performance. As a result of thisorbital performance, a mechanism for the degrada-tion has been postulated and certain ground testingperformed to verify this mechanism, which is de-scribed.

Description of Multispectral Scanner

The Multispectral Scanner (MSS) consists of anoptical radiometer scanner and a signal multiplexermounted on the spacecraft as shown in Fig. 1, addi-tional ground equipment to process the video datafor suitable photographic reproduction, and comput-

J. D. Love is with the Hughes Aircraft Corporation, El. Segun-

do, California 90245; the other authors are with the General Elec-

tric Company, Space Division, Valley Forge Space Center, P.O.

Box 8555, Philadelphia, Pennsylvania 19101.

Received 26 March 1973.

er processing of input corrections. The MSS oper-ates in the visible spectrum, has open reflective op-tics, on-board calibration, and solar input calibra-tion. A later model will add an ir window channel.

The spacecraft is in a near polar sun-synchronousorbit wherein the orbit plane precesses at a rateequal to the earth's motion about the sun. Thus,the sun-earth line intersects the orbit plane at virtu-ally the same angle throughout the year. As thespacecraft is stabilized so that its base is alwayspointed toward the nadir, the sun rotates about thespacecraft as shown in Fig. 2. Over the poles thesun-earth line is parallel to the base of the space-craft, and it is in this orientation that sun calibra-tion takes place.

The MSS gathers data in the solar-reflected spec-tral region of 0.5-1.1 /m by scanning cross-trackswaths of 100-nautical mile width, imaging six scanlines (Fig. 1) across in each of the four spectralbands.

The optics system of the scanner consists of a scanmirror, a telescope, a rotating shutter, associatedcontrols for sync, an on-board and solar calibrationsystems, an array of twenty-four optical fibers andtheir associated detectors.

The object plane is scanned by means of an oscil-lating flat mirror between the scene and the tele-scope optical chain. Twenty-four optical fibers (sixin each of four bands) are arranged in a 4 X 6 matrixin the image plane of the telescope. Light impingingon each glass fiber is conducted to an individual de-tector through an optical filter, unique to the spec-tral band served. An image of a line across theswath is swept across the fiber each time the mirror

1230 APPLIED OPTICS / Vol. 13, No. 5 / May 1974

MSS MULTIPLEXER(LOCATED INSIDE ACOMPARTMENT OFSENSORY RING)

MULTISPECTRAL SCANNER

ENTRANCE APERTURE

able neutral density filter. All these elements areshown in detail B of Fig. 3, with the optical fibersmoved away from the array for clarity.

During the scan retrace interval, after the scanmirror has completed an active scan, a shutter wheelcloses off the optical fiber view to the earth. An in-ternal light source is then projected onto the fibersvia a mirror. A continuously variable neutral densi-ty filter (NDF), which is attached to the shutterwheel, is swept across the light path so that eachvideo channel contains a nearly right triangular pulseof light, which begins with a rapid transition fromblack to white and descends at a lower rate monoton-ically to black. The electrical wave thus separated bythis variation in irradiance on the detector is shownin Fig. 5. The optical components used after eachlamp concentrate the light output on the fiber arrayas illustrated in Fig. 3, detail C.

The neutral density filter varies continuouslyfrom 100% transmission on one end to approxi-mately 1% transmission on the other and is usedto attenuate the light that falls on the fiber op-tics array to provide all levels between zero and fullscale. This assembly is referred to herein as the cal-ibration wedge, or cal wedge.

Fig. 1. MSS on ERTS-1 and scanning pattern.

Iq - sJYsOUBIT DIRECTION

- Ii~~i~~ -~I~~i-~ii- I8~~ EARTHIS SHADOW/(X -- aSUN CAL ERTS ORBITMIRROR

DURING DARK

, EARTH

N

Fig. 2. ERTS-1 sun/orbit relationship.

scans, causing a video to be produced at the scannerelectronics output for each of twenty-four channels.

The major characteristics of the MSS are given inTable I, and an optical layout is shown in Fig. 3.We shall be discussing the various calibration tech-niques, and these are shown as inserts on this figure.

In-Orbit Calibration TechniquesThe MSS has two in-orbit calibration techniques,

one utilizing a calibrated lamp and stepped neutraldensity wedge in the instrument, and the second uti-lizing the sun itself. To help clarify the differencebetween these techniques, we have constructed asimple model of the instrument in block diagramform (see Fig. 4).

Internal Calibration System

The internal calibration system consists of an in-ternal light source, a lens, and a continuously vari-

Table I. Optical Characteristics

Item Characteristics

Telescope optics 22.8-cm (9-in.) aperture diameter,f/3 .6 Ritchey-Chretien

Scanning method Flat mirror oscillating -42.9 at13.62 Hz

Scan (Swath) width 11.50 (100 nautical miles at 496nautical miles altitude)

Scan duty cycle 44%Instantaneous field of 86 grad

view (IFOV)Number of bands FourNumber of lines Six

(detectors) scanned perband

Limiting ground 40 mresolution from 496-nautical miles altitude

Spectral band wavelength:Band 1 0.5-0.6,umBand 2 0.6-0.7,umBand 3 0.7-0.8,umBand 4 0. 8-1. 1 m

Band 1 Band 2 Band 3 Band 4Sensor response:Detector PMT PMT PMT PhotodiodeNominal input for 4V 24.8 20.0 17.6 46.0

Scanner Output(10-4W cm-2 sr-')

Scanner and multiplexer 50 kg (111 pounds)weight

Scanner size Approximately 36 X 38 X 89 cmCommand capability 72Telemetry channels 97Signal channels 24

May 1974 / Vol. 13, No. 5 / APPLIED OPTICS 1231

FOCAL

E- SUN CAUBRATE ME R.(DETAIL A)

SOURCEA/,SSEMBLES

NEUTRAL

DERIVER teSHUTTER \ TD-E LESCOPEOPTICAL AXIS

OPTICAL ROTATINGFIBER ARRAY SHUTTER WHEEL

DETAIL B

MO.-IRROR__FILTERS

,' ENSIT LIl FILTER

OPTICAL FIBER ARRAY

1 . M

NOMINAL 0.127 CM 0.0081CM1E| | |]|D|1 z NOMINALSPACING

.,0051 CM BETWEENFIBER S I BANDS OP0.007CM| I I | lrr- i- 3 FIBERSSQUADE 0.0031 CM

DETAIL D

Fig. 3. MSS optical layout.

TYPICAL CHANNEL VIDEO OUTPUT

,SUN PULSE (II [CALIBRATION WEDGE

SUN PULSE n < ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I4 SCAN .. I RETRACE * SCAN * RETRACE __

AnL< i + IST SCAN *

Fig. 5. Output waveform during sun calibration.

Fig. 4. Calibration model block diagram.SUN RAY

Sun Calibration System

In addition to the internal calibration system,there are provisions in the MSS for a sun calibrate(absolute) when the spacecraft is at a nearly polarposition (Fig. 2).

The sun calibration system consists of a four-facet-ed mirror and an entrance aperture for each of thefour facets that directs the solar energy through theaperture onto the scan mirror and thence throughthe primary MSS optical system. Figure 3, detail A,shows the characteristics of the sun calibrate system.

I

/ NI

k /

VIEW LOOKINGDOWN FROMABOVE SCANNER

Fig. 6. Calibration geometry.

1232 APPLIED OPTICS / Vol. 13, No. 5 / May 1974

APERATUR.S

SURFACES

DETAL A

k - i TH SCAN

ONE OF FOURFACETS ONSUN CALIBRATE

/ MIRROR

2 l

( I

Table II. Sun Cal Mirror Alignment Test Results

Corrected exoatmospheric average voltagePredicted

Spectral Table Mountain Valley Forge Theoretical Percent ofband Signal Deviation Signal Deviation signal full scale

1 2.77 -40.06 3.01 -40.20 2:72 0.682 2.72 40.05 3.08 40.12 2.88 0.72

-0.043 2.44 4±0.06 2.63 ±40.04 2.56 0.644 1.75 ±0.04 1.45 4t0.20 1.84 0.46

The sun calibrate apertures are 0.051 cm in diam-eter. The four facets are required to compensate forseasonal variations in the sun-earth line with respectto the orbital plane. As the spacecraft orbits theearth, coming from the dark side, the spacecraft is il-luminated by the sun before the earth's surface atnadir. Thus, the sun calibration can be performedwith dark earth as a background. This occurs about18° before the terminator. This can be seen fromFigs. 2 and 6. To better illustrate the sun and in-ternal calibrations, the electrical output wave formsgenerated by each is shown on Fig. 5. Note that thecal wedge is outputted only every other scan line.The sun pulse location in the scan is a function ofspacecraft attitude, time of year, and geographicallocation of the spacecraft subpoint.

The sun cal system was measured both in theSpacecraft Assembly Facility at Valley Forge, Penn-sylvania and in the Jet Propulsion Laboratory atTable Mountain, California. Estimates were madeof the atmospheric transmission, and from these theexoatmospheric irradiance was computed. The re-sults of these measurements are given in Table II.

Tests at a General Electric Spacecraft AssemblyFacility, after spacecraft thermal vacuum test, weredone to determine sun cal mirror alignment, using aHe-Ne laser. The measured outputs in Band 2 wereof the order of 1-3 V, which is the expected valuerange. Final visual inspection was done at theSpacecraft Assembly Building a few days before theERTS-1 spacecraft was launched to ensure thatthere was no blockage of the sun calibrate apertures.

Orbital Performance

Initial turn on of the MSS occurred during Orbit20 followed by a sun calibration on Orbit 21. Theoutputs were of lower magnitude than expected,especially for the Band 1 sensors, and exhibited aspectral dependence as shown in Table III. It isclear from the data in Table III that the degradationis spectrally selective with increasing degradation atthe shorter wavelengths.

Since Orbit 21 we have made measurements onnot only the sun calibration but also the internal cal-ibration wedge. These data are presented in Figs. 7,8, and 9. The data in Fig. 7, the sun cal data, arefrom a computer generated histogram and have beencorrected for changes in cal wedge system gains.

Table IlIl Sun Cal Run Results

Ground Percent-calibration Orbit 21 age of

Spectral Theoretical (average data expectedband signal (volts) volts) (volts) output

1 2.72 2.89 0.2 72 2.88 2.90 0.7 243 2.56 2.54 1.6 624 1.84 1.68 1.4 76

29

28

27

26

25

24

23

22

7

I I l I I I I

2000

2.0

1.8

Ui 1.6- IU

d 1.4

, 1.0

1 0.8

U 0,60.

0.4

0.2

1O000

ORBITS

Fig. 7. Sun calibration as a function of time in orbit.

NOTE - OSCILLOSCOPEREADINGS (NOM E0.2VACCURACY)

z BAND 3

BAND 4

BAND2

BAND 1

500 1000 1500

CONSECUTIVE ORBITS

Fig. 8. Sun calibration vs orbit.

May 1974 / Vol. 13, No. 5 / APPLIED OPTICS 1233

BAND 3

BAND 4

BAND 2

D~

0

U)

.1

3

BAND I

0~

63[ FULL SCALE

40 RAN BAD

B I

I I I I I I I I I I I I .T . I - AT M M I.

INOTE .-RcZTE. ATA F A

GIVIE TANSMISIO OF

31

108 2000 30

C-NSC.TIVE -S.ITS

Fig. 9. Cal wedge vs orbit.

The ordinate is in digital counts with sixty-threecounts being full scale. Figure 8 is uncorrected datataken from oscilloscope readings (ordinate is volts)and is presented only to show that there was littlechange in the sun cal pulse prior to Orbit 600 whenthe histogram studies were started. The data in Fig.9 are the output normalized at Orbit 100 at forty-fivecounts. (Note: Cal wedge varies between 0 and fullscale as shown in Fig. 5. Figure 9 data are based ona level near forty-five counts.)

The data given in Fig. 8 are in output volts, whileFig. 9 are given in relative output normalize at Orbit400. As can be seen from these figures, the sun cali-bration has remained relatively stable in orbit. Thecal wedge, whose output was not particularly differ-ent in initial orbits than ground testing, has slowlydecreased with time. All detectors in Bands 1 and 2showed large initial cal wedge degradations. Band 3detectors showed small changes, and the Band 4 de-tectors were relatively stable. Band 1 at the wordnumber plotted, which is equivalent to a specific ori-entation of the cal wedge filter and hence a specificoptical transmission, decreased from 45.7 counts tothirty-eight counts or about 12%.

After Orbit 1000, the trend was toward stability ofall cal wedges with increasing time. Note that, as inthe case of the sun calibration, the greatest effect isseen in the shortwave channels.

Evaluation

We can now postulate some theories as to whatmay have caused the initial dramatic change in thesun calibrate and the slower changes in the calwedge. To do this we refer to Fig. 4, the MSS blockdiagram model. Thus, scene radiance comes fromthe ground to the scanning mirror, to the telescope,to the fiber optics, to the detectors, where it is con-verted to an electrical signal and outputted. Weshall consider first the changes in the internal cali-bration wedge.

The outputs from the calibration wedge input

have been noted to decrease over orbit life in Bands1 and 2. The output in Band 4 is relatively stableover all orbits to date. A number of direct causescould be responsible, for example: (1) irradianceinput shift to the fiber optics; (2) change in spectraloutput of the calibration lamp due to aging, (3)input excitation changes, (4) in the high voltage tothe PMT detectors; (5) voltage distribution acrossthe dynodes; (6) Van Allen radiation. We postulatea seventh cause, that is, the existence of a spectrallyselective contaminant, which has coated one or moreof the optical surfaces. We shall discuss this in moredetail later in the paper.

Cal Lamp Irradiance Distribution

This is the most likely cause, other than our pos-tulated one, since it has been previously determinedthat the fiber optics input is nonuniform, the varia-tion observed being less than 5%. If the cal wedgelamp energy at the fiber optics changes, the calwedge output would shift, but the shift would becongruent in all channels. Since no changes havebeen noted in Band 4, or has the shift been the samein all channels, it is expected that this has not hap-pened.

Cal Lamp Excitation

The cal lamp is powered from a constant currentsource that is monitored via telemetry. Fromlaunch to date no variation in lamp current has beennoted. Further, a lamp current change should affectthe Band 4 output (as well as all the others), but theBand 4 output, as noted previously, has been rela-tively constant. If the filament metal were to evap-orate and deposit on the bulb envelope, the filamentresistance would increase, thus increasing the powerto the filament (P = 12R) and hence its brightnessand color temperature. The evaporated metalwould act as a neutral density filter. If these effectsjust balanced each other in amplitude (a highly un-likely situation), there would be at least a spectralshift to shorter wavelengths. If the effects were notin balance, we would expect there to be both an irra-diance level and a spectral shift. Neither has beenseen.

PMT High Voltage Excitation

If the high voltage supply voltages were to change,we would expect the sun cal, the cal wedge, and theearth scene signals in all three PMT Bands (1, 2,and 3) to change more or less proportionally. Asnoted, Bands 1 and 2 cal wedge data change is muchdifferent from Band 3.

Van Allen Radiation

Figure 10 is a cutaway photograph of the instru-ment. From this figure it can be seen that the tele-scope housing is a casting and that the shutter wheelis located within the casting and behind the primaryand secondary mirrors. This casting is of variousaluminum thicknesses. The cal wedge lamp and fo-cusing lens are further contained in a 0.1-cm thick

1234 APPLIED OPTICS / Vol. 13, No. 5 / May 1974

Fig. 10. Cutaway view of MSS showing sun cal mirror.

steel housing. Although a detailed study of poten-tial radiation browning of the glass has not beenmade, it is estimated that the level at the cal wedgeis less than 103 rad and at the lamp and lens lessthan 102 rads.

Further, it should be noted that all six detectors ina given band have decreased their output propor-tionally to one another, most likely indicating acommon cause. Thus, we have probably eliminatedall the potential causes except the spectrally selec-tive contamination.

Sun Calibrate Degradation

The sun calibrate outputs are of much lower mag-nitude than predicted by modeling or measurementsmade on the earth's surface. The outputs in Band 1are especially low, while that in Band 4 is closest tothe predicted value. There is a remote possibilitythat something has partially blocked the solar cali-bration entrance apertures (Fig. 10), but it is diffi-cult to conceive of any available material that wouldhave the proper spectral characteristics to give themeasured orbital data. Further, in orbit, the sunhas illuminated more than one facet and apertureswith no appreciable change in output. Thus, thesecond aperture would also have to be blocked withthe same material. The sun cal mirror, as can beseen from Fig. 10, is supported on beam with little ornothing near it so that it is most unlikely that thereis anything blocking the solar ray bundle. Furtherchanges in solar mirror alignment angle would only

result in the sun pulse occurring at different parts ofthe scan line or location orbit.

Additional Orbital Data

Another measure of instrument performance is itsoutput from given ground targets as a function oftime. The instrument is set up with a transfer func-tion for earth inputs. Targets such as clouds orsnow saturate the amplifiers, and thus they cannotbe used for calibration. However, by densitometericreading of a small area of the same earth scene overa period of time, one may get an idea of the instru-ment transfer function. This was done, and the re-sults are given in Fig. 11. You will note that the in-

0.5 I

0.4 -u:

.0

5

J

a

0.3

0.2

0.1

BAND I

BAND4

I I I I I I I I I I I I I I I I

I I1 21 30 10 20 30 10 20 30 9 19 29 9 19 29 8

AUG SEPT OCT NOV DEC

I Fig. 11. Radiance history taken from 14.1-cm film with a densi-tometer (1-mm aperture).

May 1974 / Vol. 13, No. 5 / APPLIED OPTICS 1235

0

After this coating, the mirrors exhibited a similarmilky appearance to those in the spacecraft test.An ir reflectance spectrum was run from 2.5 Am to

COATEB MIRBOR AFTER SCI.ANIZATION

50 Am on both the mirror samples and a sheet of thesame Mylar. The correspondence of the strongbands in both spectra clearly indicates that themilky substance is simply Mylar transported fromthe tape to the mirror.

The spectral reflectance of this test mirror wasalso measured from 0.4 Am to 1.2 Am. The test mir-

0, 5 0, 1 0, 7 0, , ,~ , , , ror was then exposed to VUV radiation. The irra-0.4 diation simulates the solar irradiance but is limited

by windows in the source to radiation longer thanFig. 12. Reflectance of sample mirror. 164 nm. It would be expected that the higher ener-

gy photons of the shorter wavelengths would producemore degradation, but facilities to accomplish this

rment data agree quite well to the radiance ex- test were not readily available.The results of the tests are shown in Fig. 12. Ased, assuming a Lambert law diffuse reflector. cnb enfo hsfgrterfetneo hn this we may conclude that there is no block- can be seen from this figure, the reflectance of theor obscuration of either the earth or sun image in solarized mirror is degraded in a spectrally selectiveimagep lane of the instrument, manner, being some 50% of the unsolarized sample

and 45% of an uncoated mirror. Figure 13 is a plotctraily Selective Contaminant of the solarized mirror reflectance and orbital sun cal

ie postulate that the sun cal mirror was contami- signals expressed as the percentage of the expecteded before launch with a material that, under nor- orbital valuesconditions, was relatively transparent in the While in Fig. 13 the spectral match between the

spectral range. In orbit, the unfiltered uv curves is not particularly good, one must observe1an hsectralanore ti oitamtheunitd a that laboratory irradiation did contain the highly ex-ation has transformed this contamination into a cited photons below 164 nm. Nor can we be abso-tance exhibiting selective spectral reflectance. lutely positive that the so called milky coating waslring the spacecraft thermal vacuum test some not a mixture of several different hydrocarbons, all ofer insulating tape was overheated and left a which would tend to cause a mismatch between theIy deposit on the MSS collimator mirror and sev- spectra.

-: -_ -,_-.. m,-.I. -P i-u r%'rQ _7 _~7erasl Willlubb 111111U10. X COU U1 lu VI inliuwvu 11U

degradation in performance of the sensor or calibra-tion systems, but none of these tests involved the useof sunlight. The sun mirror, scanning mirror, andinternal calibration optics were not cleaned prior tolaunch. The sun calibrate mirror is located suchthat it receives direct solar irradiation (Fig. 2) in-cluding the strong hydrogen emission bonds.

The scanning mirror and internal components areshielded from direct uv radiation from outer space,but scattered uv can impinge on these surfacesthrough the input aperture. The transformation ofthe contamination to the selective spectral materialwould then necessarily proceed at a much slower ratethan possibly occurred with the sun cal mirror.Note that the sun cal output has remained relative-ly stable in orbit, while continued changes in calwedge output have been observed.

We believe that some evaporated Mylar coatedvarious optical surfaces, and this Mylar has polymer-ized under uv radiation into new substance with therequired spectral characteristics. In an attempt toprove this hypothesis, some postlaunch laboratorytests were run at the Goddard Space Flight Centerunder the direction of Warren Hovis.

Contamination Tests

The same kind of Mylar tape used in the space-craft thermal vacuum test was heated in a vacuumin the presence of two fresh first-surface mirrors.

Conclusions

The argument presented here regarding the perfor-mance of the internal calibration system seems to in-dicate that the decrease in signal was due to thespectrally selective contaminant being deposited onother optical surfaces, such as the scanning mirrorand the fiber optics, and slowly lowering their opti-cal quality. There are however other argumentsthat seem to contradict this hypothesis for the calwedge. For example, if the material were deposited

PERCENTAGE DF EXPECTED10 ORBITAL RESPONSE

L... 2 SURFACE

0_ /

0.4 0.5 0.6 0.7 0.8 0.9 1.0 .1 1.2

WAVELENGTH (MICROMETER)

Fig. 13. Solarized mirror and orbital sun cal signals vs expectedorbital values.

1236 APPLIED OPTICS / Vol. 13, No. 5 / May 1974

stru:pectFroring the i

Spe

wnatemal0.5-radisubs

DiMylmill--- I

Too

so

on any of the series elements (see Fig. 4), as the calwedge decreased, so also would the sun cal and thescene video, which also pass through this same opti-cal chain. This has not been observed. One mightpostulate that the contaminant was on the cal wedgesystem only. However, there is no line of sight forthe contaminant to travel from the entrance aper-ture to the cal wedge which does not include the se-ries optical elements. Possibly some reoutgassing ofthe Mylar in space or some other internal contami-nant source adjacent to the cal wedge could explainthis dichotomy.

We believe that the sun cal mirror was contami-nated and did degrade as indicated herein. We feelthat something similar also happened to the calwedge but with much less confidence. We also rec-ognize that it is not possible to prove or disprove thishypothesis, but nonetheless the data are valuable toothers working with space optical systems.

We thank Warren Hovis of the Goddard SpaceFlight Center for his permission to publish the re-sults of his measurements. This work was per-formed under NASA contract NAS 511320.

NEW CATALOG LISTS SPECTRUM SYSTEMS

STANDARD OPTICAL COMPONENTS AND COATINGS

A new 62-page catalog describing standard optical components andcoatings is, now available from Spectrum Systems, Barnes EngineeringCompany, Stamford, CT 06904.

Approximately 20 major categories of optical components and coat-

ings are divided into some 40 section headings. Among these are quartzlenses, interferometer flats and windows, spherical mirrors, Littrowprisms, Brewster angle windows, retardation plates, polarizers, arsenictrisulfide lenses, matched-pair Fabry-Perot interferometer plates, laserreflectors, beamsplitters, etc. Each family of entries is arranged intoa table listing dimensions, optical characteristics, part number, priceand delivery information.

For a free copy of the Standard Optical Components catalog write toSpectrum Systems, Barnes Engineering Company, 44 Commerce Road,

Stamford, CT 06904.

May 1974 / Vol. 13, No. 5 / APPLIED OPTICS 1237