Z6). - Plant physiologyThe 135-volt battery is composed of three Eveready No. 482 bat-teries or equivalent. The counters are Mercury (Production Instrument Company,Chicago, Illinois)

PHOTOTUBE-TYPE INTEGRATING LIGHT RECORDERS:A SUMMARY OF PERFORMANCE OVER

A FIVE-YEAR PERIOD

G. FRED SOMERS AND KARL C. HIAMNER1

(WITH EIGHT FIGURES)

Received June 28, 1950

In recent years considerable attention has been given to a study of therelationship between the amount of light a plant receives and its vitamincontent (1, 2, and 3). Such a relationship can be studied under laboratoryand greenhouse conditions, but ultimately it has to be evaluated under fieldconditions. This requires some measure of the amount of sunlight. Variousinstruments are available for recording the amount of sunlight. One, usedby the Weather Bureau, is the Eppley pyrheliometer, which is a multiple-junction thermocouple combined with a millivolt recorder. While this instru-ment gives a continuous and reasonably dependable record of the amountof solar radiation, it has certain drawbacks which limit its use under fieldconditions. It is expensive and requires an external source of electricityto operate the recorder. The record obtained must be integrated in somefashion, which either requires additional time in making use of the dataobtained or requires the purchase of a rather expensive integrating unit.Furthermore, the light receiver is a horizontal disc, whereas in certain casesan opal glass globe receiver would appear to be more desirable (7). Aninstrument which appears to be more suitable for field experiments wasdescribed by SPRAGUE and WILLIAMS (5, 6). This instrument is simple toconstruct, is portable and is battery operated. It uses an opal glass globereceiver of the type WALLACE (7) found to be desirable for use in plantphysiological experiments.

Some years ago we were confronted with the problem of measuring sun-light under field conditions. After considering various possibilities wedecided to construct a number of light recorders similar to the one describedby SPRAGUE and WILLIAMS Z6). In constructing these instruments we hadseveral aims: (1) To obtain dependable daily values for the amount of sun-light with a minimum of effort, (2) To be able to obtain these records underfield conditions without being dependent upon an external source of elec-tricity, (3) To obtain instruments which were readily portable and could beshipped readily to various cooperators.

Details of construction

The circuit which was adopted after making various preliminary testsis shown in figure 1. This differs in several details from the instrument

1 Present address: Department of Botany, University of California, Los Angeles,California.

318

www.plantphysiol.orgon April 5, 2020 - Published by Downloaded from Copyright © 1951 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

SOMERS AND HAMNER: LIGHT RECORDERS

described by SPRAGUE and WILLIAMS (6), but the principle of operation isessentially the same. It will be noted that eight counters were provided.By means of a commutator these give a record for the total amount of lightper day for a period of a week. The counters can be read any time on theeighth day, at which time the commutator is set for another week. Thepower for the counters is supplied either by 110-volt AC or from a 6-voltDC source (four 1A-volt No. 6 Eveready "Ignitor" batteries). The ratingsfor other components of the circuit are as follows: Condensers: C1, 0.02 uF;C2, 9 to 12 ,uF; Resistors: R1, 50,000 ohms; R2, 250,000 ohms; R3, dependsupon the relay. The sum of relay-coil resistance and R3 was 2500 to 3500ohms. A sensitive relay with a 6-volt coil and 115-volt AC, 6-ampere con-tacts (No. 29XAX, Struthers Dunn, Inc., Philadelphia) is satisfactory forthe recorders which use 115 volts AC to operate the counters. For the

RCA RT SCOUNTERS9R2 C2f '--

1 4 5

3 6

I 2 ~~~~~~~~~~~~~~7

I ~~~~8

Sp ~6V. D.C. OR+ IJ5V. A.C.

FIG. 1. Diagram of circuit used in phototube-type light recorders. RCA922 is avactuum type phototube. OA4G is a cold cathode relay tube.

battery-operated units a more sensitive relay with a 3500-ohm coil is moresatisfactory (Series 5 DC, with S.P.D.T. #1 contacts, Guardian ElectricCo.). The 135-volt battery is composed of three Eveready No. 482 bat-teries or equivalent. The counters are Mercury (Production InstrumentCompany, Chicago, Illinois) five-digit, non-reset counters with either 6-voltDC or 115-volt AC coils, depending upon the power supply.

The phototube is housed in a brass case (see fig. 2) on which an opalglass globe such as is used in light fixtures is mounted with a vapor-tightseal of caulking compound (not shown in the figure). A brass sleeve ex-tends into the throat of the globe, and in this are mounted two glass filters(Corning Nos. 2962 and 3850) and a ground-glass plate. These are held ina special support which has a one-inch aperture. On top of the glass filtersis an aluminum disk (not shown in the figure) in which were punched holes(3/32 inch diameter) at regular intervals. The number of holes varies

319



PLANT PHYSIOLOGY

some from recorder to recorder. The number is varied to give the desiredresponse, but usually six are used. The phototube housing is kept dry bymeans of a container of anhydrous calcium sulphate (indicating Drierite).Originally the phototube was connected by means of a shielded weather-proof antenna cable to an aluminum box in which were housed C1, C2, R1,the OA4G relay tube, and, initially, the relay. This box was also kept drywith calcium sulphate. Later the relays on the DC units were mountedexternally so they could be serviced readily. Also, it was found desirableto use two, separate, shielded, heavily insulated cables to connect the photo-

GLASS CLOSE

R c A 9 !5

PHOTOTUBE C

FIG. 2. Vertical cross section of phototube housing as originally built. In latermodels two phototube cables were used.

tube to the recorder. An even better arrangement apparently would be toinclude C1 and the relay tube OA4G in the same housing with the photo-tube.2 The commutator for the DC-operated units is coupled through anappropriate set of gears with the clock drum of a hygrothermograph andmakes one revolution in eight days. This commutator and clock are syn-chronized each week when the counters are read.

The construction of the commutators used on those recorders which

2 Personal communications from Radio Frequency Laboratory, Boonton, NewJersey, and W. H. Brittingham, Texas Agricultural Experiment Station.

320




required AC power supply is essentially the same, except that in this casethe commutator is driven through a friction clutch by an AC clock motor.The rotor of this commutator also is driven at the rate of one revolutioneach eight days, and its position is indicated by an 8-day dial. Since thedrive is by means of a friction clutch, the setting of the commutator can bechanged readily at any time. The usual practice is to set the commutatorto the correct time for Day No. 1 on Monday of each week at which timethe counters are read.

The recorders using a DC power supply can be placed in weather instru-ment cabinets in the field adjacent to the experimental plots. Those usingan AC power supply must be near a source of electric current and are ordi-narily placed in a building with the phototube housing outside. Both types

100

8;0

20 _-J

3600 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500WAVE LENGTH - ANGSTROMS

FIG. 3. Approximate relative sensitivity of the 922 phototube to light of variouswavelengths when combined with the filter combination used.

of recorder have been shipped by express for long distances at various timeswith little or no damage.

The transmission of the filter combination used in the phototube housingwas measured with a Beckman spectrophotometer. These data were com-bined with data supplied by the Radio Corporation of America on the spec-tral sensitivity of the phototube to obtain a curve representing the spectralsensitivity of the RCA 922 Phototube in combination with these filters (seefig. 3). Due to the variations in filters, phototubes, and opal glass globes,this curve is probably only an approximation of the spectral sensitivity ofany one instrument. The phototube alone has a sensitivity response withpeaks of relatively high sensitivity at about 3600 and 8000 A. These peaksare eliminated by the filters and the response is limited essentially to thevisible region of the spectrum. Our measurements indicate that the opalglass globe influences this relative spectral sensitivity little, if any.

321



PLANT PHYSIOLOGY

Performance

The criteria used to evaluate these instruments were: (1) linearity ofresponse, (2) sensitivity to light of low intensities, (3) ruggedness, (4) sta-bility of the calibration between recorders, (5) performance and durabilityof mechanical components and (6) comparison with the Eppley pyrheli-ometer over extended periods of time. This comparison was made becausethe pyrheliometer is an accepted standard for measuring solar radiation.

Linearity of response was checked in two ways. A limited test wasmade by shading the receiver with single or double thicknesses of cheese-cloth, timing the counting rate, and measuring the light intensity with aWeston Sunlight Illumination Meter (Model No. 603). The observationsare summarized in table I. In all cases the light intensities were the maxi-

TABLE ICOUNTING RATE OF RECORDER No. 3 AT VARIOUS LIGHT INTENSITIES

Relative RelativeShade Number of Light light countingobservations intensity intensity rate

(f Q)None 4 6190 100.0 100.0Single Cheesecloth 3 3230 52.2 50.8Double Cheesecloth 2 1700 27.5 27.7

mum readings obtainable by directing the target of the Weston light metertoward various portions of the sky. To further test the linearity of responsethe light intensity was measured in the open with the Weston meter onseveral days which varied in cloudiness, and simultaneously the countingrates of the integrating light recorders were timed with a stopwatch. Theperformance was considered satisfactory in -this respect when a linear re-sponse to the light intensity measured by the Weston meter was obtained.In some cases, minor adjustments had to be made to obtain the desiredresponse. A typical response curve is shown in figure 4.

The sensitivity to light of low intensities was determined by measuringthe light intensity (by Weston meter) at which the recorders stopped record-ing in the evening. This proved to be about 100 fc when measured in July,1945, and a similar value was obtained in October, 1946.

The ruggedness of these recorders is indicated by the fact that duringfive years various recorders have been shipped by express to various partsof the United States and have been transported repeatedly over shorter dis-tances by truck and car. In only one instance has a receiver been broken.In one or two other cases minor damage occurred, but the results demon-strate clearly that the instrument can be transported easily and withoutserious damage.

These recorders were used in field experiments by allowing them to oper-ate simultaneously at our laboratory for a period, during which time the

322




relative number of counts per day was obtained for each instrument. Therecorders were then sent into the field, one to each location. After theexperiments were terminated, the recorders were returned to our laboratory,where they were again operated simultaneously to obtain their relativecounting rates. When comparisons of this kind were made repeatedly overa period of years they provided some measure of the relative stability of

9000

6000

7000

-Iaz< 60000

I~-0o 5000Ua.a

- 4000

z

I-z 3000

0- 2000

1000

00 1 2 3 4 5 6 7 8 9 10

COUNTS PER MINUTE

FIC. 4. Calibration of phototube-type light recorder with Weston light meter afterinstalling dual phototube cables.

the different recorders. Some comparisons of this kind, obtained duringthree summers of operation, are summarized in figure 5. The periods ofobservation are indicated in each case by a heavy bar. The length of eachbar indicates the length of the period used for the comparison. The dataare typical of the results obtained with all of the recorders. They showthat, while there is some variation among the recorders from time to time,they give reasonably reliable data. The variation in the relative counting

323

-0



PLANT PHYSIOLOGY

rate from time to time is apparently less than ± 5%o in all cases, and con-siderably less than this in most cases. This is sufficiently accurate for mostfield experiments at the present time.

The only significant mechanical weakness encountered in these lightrecorders is the failure of the counters. The counters were inexpensiveunits (about $6.00 each) but the performance obtained was poorer than hadbeen expected. Occasionally a counter will stick. This causes an errone-ous record. This difficulty has been minimized by providing a ninth counterwhich is arranged to count each time any of the other eight counters regis-ters. In this way, a total is obtained directly for each week. This can be

x 150 NM 100 140 _

r No. 5

No..2 No.1IC.

c 100

0 90. NO. 120

so.0s 70a 60_Z 50

! 6.9.45 9.1 12.1 3-1-46 6-1 9-1 12-1 3-1-47 6.1- 9 .1 1i-DATE TESTED

FIG. 5. A summary of relative response of six light recorders during various periodsof observation. The duration af the periods of observation are indicated by the heavybars. The response of each recorder (= Ax) is compared with the response of recorderno. 3 (= A3).

compared with the sum of the counts for each day during the week. Anydiscrepancy indicates a failure in one or more counters. Table II sum-marizes some comparisons between the total counts of the ninth counterand the sum of the other eight. It is seen that for the most part errors dueto counter failure are small, but too frequently they are of serious magnitude.Counters which should prove to be more dependable are now available.

Some attempt has been made to obtain a comparison between these lightrecorders and an Eppley pyrheliometer installed on the Plant Science Build-ing of Cornell University, which is a few hundred yards from our labora-tory. Comparisons made between the pyrheliometer and the phototube-typerecorders, remodelled with heavily insulated, separate phototube cables,during the years of 1947, 1948, and 1949 are summarized in figure 6. Thesedata indicate that over this period of time the two types of instruments had

324




a rather constant relationship in their responses. There is no appreciablechange from year to year. The correlation coefficient between the two setsof observations is + 0.968 (n = 285).

Some further conclusions can be drawn from these data. It is obviousthat there is some day-to-day variation in the response of the two types ofinstruments. Part of the observed variation may be due to errors such asthe failure of counters, or errors in obtaining the data from the instruments.

TABLE IIPER CENT. DIFFERENCE BETWEEN SUM FOR THE INDIVIDUAL COUNTERS AND

THE TOTAL OF A NINTH COUNTER.£A1+A2..... 8-A9*9 X 1009 WIEREA INDICATES E NUMBER OF COUNTS

Week ending Recorder Recorder Recorder(1948) 1 2 5

May 31 ... + 3.3 ...June 7 ... + 1.7 ...

14 ... + 0.60 + 0.9221 ... - 1.3 0.0028 ... . + 0.46 ...

July 5 ... - 1.7 + 1.712 ... - 1.19 - 0.02519 ... + 0.014 + 0.0126 + 2.58 - 0.87 + 3.7

August 2 +2.6 +16.3 + 0.129 ... ... + 0.68

October 11 + 0.050 - 0.1218 +4.04 -.10.625 - 0.014 - 0.018 -16.13

November 1 + 0.017 - 0.007 - 4.848 +9.75 - 2.12 - 4.4415 -0.054 - 6.7 + 6.9522 -0.012 - 8.1 ...29 -0.18 0.00 10*-

December 6 -0.012 0.0013 - 0.027 - 0.9320 -0.19 - 4.50 *--

However, part of the variation observed may result from real differences inthe response of the two types of instruments. For example, since the pyr-heliometer uses a horizontal disc receiver and the phototube-type recorderuses an opal globe receiver they may respond differently to diffuse lightfrom the sky. Also, the disc-type receiver will be more sensitive to thedeclination of the sun than will the globe-type receiver. A linear regressionline is drawn in figure 6. While this fits the data rather well it is probablethat the relationship is not strictly linear. This is particularly true at thelower light values. Many of these latter values were obtained late in theyear when the sun was at a relatively low angle above the horizon. Itseems probable that this may account for the observed curvature.

325



326 PLANT PHYSIOLOGY

An attemiipt was made to test this possibility. The values obtained fromthe Eppley pyrheliometer were adjusted for the expected effect of the timeof the year upon the response of the instrument. A comparison of theseadjusted values with the responses of phototube-type recorders is sum-marized in figure 7. This adjustment has resulted in somewhat more scatter-

800 _

. ~~~~zxX)A X zZ

700 X NK N

600 _ Z N KZZz ~~A

Xu. ~~~~~~~~~~~~~ZKA x K

500 AN

w A

>. ~~~~~~~~~~NAA

4 400 -

AAA z0. NU4N o

0) 300 A z

,q 200 N K;Zz

0~~~~

00 1000 2000 3000 4000 5000 6000

COUNTS PER DAY

FIG. 6. Comparison of total daily radiation (gm. cal. per day per cm.2) as meas-ured by an Eppley pyrheliometer with the total visible radiation measured by photo-tube-type light recorders (counts per day). The solid line is a regression line. Theletters indicate observations made with various recorders at different times as follows:A, recorder no. 11, summer, 1948; K, recorder no. 5, spring, 1949; N, recorder no. 5, fall,1947; 0, recorder no. 2, fall, 1948; T, recorder no. 1, summer, 1947; X, recorder no. 12,summer, 1948; and Z, recorder no. 5, summer, 1948. All values for the phototuberecorders are adjusted to a comparable basis.

ing of the points at the lower light values. The regression of the numberof counts per day on the unadjusted daily values for the pyrheliometer(fig. 6) is:

Counts = 7.46 gm. cal./cm.2 + 428.

This indicates that the phototube-type recorder would give about 400 countsdaily in the absence of any radiant energy measurable by the pyrheliometer.Repeated observations show that this is not the case. On the other handthe regression of the counts per day on the adjusted daily values for thepyrheliometer (fig. 7) is:

Counts = 4.60 gm. cal./cm.2 - 92.




This relationship fits observations on the response of the phototube-typerecorder at low intensities much better. Mloreover it agrees with the ex-pected response of the two types of instruments as influenced by the timesof the year. The globe type of receiver would be expected to be somewhat

1400

0~~~~~~~~~~~~~~~~1200w~~~~~~~~~~~~~~~~

So ioootX X.~~ ~ ~~~~ .

0 0~~~~~~~~~~

8o400 1'I | I |

xx

'. O 1000 2000 3000 400 5000x60x0

x.x

6.00 xO

x x 0 *

4200a. ~~x 00

0 1000 2000 3000 4000 5000 6000* COUNTS PER DAYFIG. 7. Total daily radiation measured by an Eppley pyrheliometer and adjusted

for the expected influence of the time of year upon the response, compared with countsper day obtained with phototube-type light recorder. The regression of counts per dayupon the adjusted response of the pyrheliometer, using all the observations, is indicatedby the line. 0 represents values obtained in late spring and summer. X representsdata obtained from October to December.

less sensitive to the angle of the sun (7) and the lower light values wereobtained principally late in the summer and early fall.

When the response of the phototube-type recorder was compared withthe adjusted values for the pyrheliometer it was observed that the relation-ship between the two instruments during the fall months was different fromthat obtained in the late spring and summer (fig. 7). The two regressionsare:

April to September: Counts per day = 4.461 gm. cal. per day/cm.2 + 117October to December: Counts per day = 3.459 gm. cal. per day/cm.2 - 68

A difference in response between the two instruments at different times ofthe year probably is to be expected. The receiver on the pyrheliometer isa horizontal disc which measures the total radiation received per unit ofarea tangential to the earth's surface. As a consequence this instrument

327



PLANT PHYSIOLOGY

is rather sensitive to the angle between the receiving surface and the sourceof radiation. On the other hand, the phototube-type instrument with anopal globe receiver and light filters, as described above, gives an integrated

5000

4000

3000

4

hicCL.

co 2000I.-z00

1000

01 3 S 10 15 20

DAY OF MONTH25 30

FIG. 8. A comparison of the total amount of visible radiation daily in the openduring July and in the greenhouse in December. Both series of observations were madein Ithaca, New York, and are adjusted for differences in the counting rate of therecorders used.

measure of the total amount of visible radiation received by such a body.With such an arrangement, the angle between the receiver and the radi-ating source is of little or no consequence. Hence the two instruments mayrespond somewhat differently to reflected light from the sky, and the amountof such light probably varies with the season of the year. Ordinarily thereis considerable cloudiness in Ithaca during the fall months. In addition,the two instruments would be expected to respond differently to changes inlight quality, which will be influenced inter alia by the amount of moisturein the atmosphere. In view of such variables as these it is not surprisingthat there are differences in response between the two types of instruments.The day-to-day variation may be rather large, as is indicated by figures 6and 7, but over a long period of time the correlation between the total radi-ation as measured by the pyrheliometer and the visible radiation as meas-

328




ured by the phototube-type of instrument is good. Too much significanceshould not be attached to individual daily variations since they may arisein part from mechanical errors in the instruments, particularly the failureof the counters used on the phototube-type of instrument. One of the mostimportant conclusions to be drawn from this comparison between the twotypes of instruments is that the phototube-type recorder gave values whichcorrelated closely with those given by the pyrheliometer over a period ofyears. The latter instrument has a reputation of dependability and thiscomparison speaks well for the phototube-type of instrument.

These recorders have provided a number of interesting comparisons ofthe relative amount of light under various conditions. A comparison of therelative illumination obtained at various locations in the United States dur-ing the summer of 1945 has been published (4). A comparison of the rela-tive amount of light in various locations in one compartment of a smallgreenhouse for a period of about three weeks in May, 1946, is shown intable III. Data comparing the illumination outside in the summer withthat in the greenhouse in the winter are summarized in figure 8.

TABLE IIIILLUMINATION, RELATIVE TO FULL SUNLIGHT, IN THREE LOCATIONS OF ONE

COMPARTMENT OF A SMALL GREENHOUSE.

Location Illumination relativeto full sunlight

Center of north bench 46.8Center of east bench 32.7Center of west bench 34.0

DiscussionThese observations indicate that while the phototube-type integrating

light recorder has certain drawbacks, it gives useful data and is reasonablydependable. The principal drawback in our present recorders is the occa-sional failure of the counters. Some investigation has been made of variouspossibilities for alleviating this difficulty. Two approaches now appear tobe feasible. (1) A counter manufactured by Western Electric Company(Type 5S or 5AA message register) appears to be constructed in such a waythat it would give more dependable service. (2) Most of our field experi-ments have indicated that daily values for the amount of light are reallynot necessary. If this is the case then the eight or nine counters used atpresent could be replaced with a single, more dependable counter. Suchcounters are available. Their use would simplify the construction of therecorders and would reduce their cost appreciably. Probably it is safe tosay that, in view of our past experience, rather dependable and satisfactorylight recorders of the phototube-type could be built easily and without

329



PLANT PHYSIOLOGY

great expense. For some purposes a rather inexpensive unit could bedesigned.

SummaryA number of integrating light recorders have been built using a circuit

similar to that described by SPRAGUE and WILLIAMS (6). When properlyadjusted and constructed these give a linear response to sunlight at intensi-ties greater than about 100 fc. These instruments have proved to be ruggedand satisfactory in many respects. They are portable, and the battery-operated models can be placed directly in a weather instrument cabinet inthe field where, if they are read and set once a week, they give an inte-grated value for the daily amount of light. These readings are closely cor-related (r = 0.968, n = 285) with the integrated radiant energy values ob-tained by an Eppley pyrheliometer. A more nearly linear correspondencebetween the two types of instruments is obtained if the response of thepyrheliometer is corrected for the effect of the declination of the sun.

The authors wish to express their appreciation to Dr. W. C. Kelly, whoassisted in the construction of these recorders, and to Dr. C. S. Brandt, whodeduced the relationship used to correct the pyrheliometer readings for theeffect of the declination of the sun.

U. S. PLANT, SOIL, AND NUTRITION LABORATORYAGRICULTURAL RESEARCH ADMINISTRATION

U. S. DEPARTMENT OF AGRICULTUREITHACA, NEW YORK

LITERATURE CITED1. HAMNER, K. C. and MAYNARD, L. A. Factors influencing the nutritive

value of the tomato. A review of the literature. U. S. Dept. ofAgriculture Misc. Publ. 502. 23 pp. 1942.

2. MAYNARD, L. A. and BEESON, K. C. Some causes of variation in thevitamin content of plants grown for food. Nutrition Abst. andRev. 13: 155-164. 1943.

3. SOMERS, G. F. and BEESON, K. C. The influence of climate and fertilizerpractices upon the vitamin and mineral content of vegetables. Areview of the literature. Advances in Food Research 1: 291-324.1948.

4. SOMERS, G. F., HAMNER, K. C., and KELLY, WV. C. Further studies onthe relationship between illumination and the ascorbic acid contentof tomato fruits. Jour. Nutr. 40: 133-143. 1950.

5. SPRAGUE, V. G. and WILLIAMS, E. M. An inexpensive light recorder.Plant Physiol. 16: 629635.. 1941.

6. SPRAGUE, V. G. and WVILLIAMS, E. M. A simplified integrating lightrecorder for field use. Plant Physiol. 18: 131-133. 1943.

7. WALLACE, R. H. Methods of sampling visible radiation. Plant Physiol.12: 647-666. 1937.

330



Documents

Z6). - Plant physiologyThe 135-volt battery is composed of three Eveready No. 482 bat-teries or equivalent. The counters are Mercury (Production Instrument Company,Chicago, Illinois)