Group Title: Citrus Station mimeo report - Florida Citrus Commission ; CES 64-3
Title: A Colorimeter for grading citrus juices
Full Citation
Permanent Link:
 Material Information
Title: A Colorimeter for grading citrus juices
Series Title: Citrus Station mimeo report
Physical Description: 7 leaves : ill. ; 28 cm.
Language: English
Creator: Hunter, Richard S ( Richard Sewall ), 1909-
Citrus Experiment Station (Lake Alfred, Fla.)
Florida Citrus Commission
Publisher: Florida Citrus Commission :
Florida Citrus Experiment Station
Place of Publication: Lake Alfred FL
Publication Date: 1963
Subject: Citrus juices -- Color -- Florida   ( lcsh )
Citrus juices -- Quality -- Florida   ( lcsh )
Colorimetry   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Includes bibliographical references (leaf 7).
Statement of Responsibility: Richard S. Hunter.
General Note: Caption title.
General Note: "October 8, 1963."
 Record Information
Bibliographic ID: UF00072415
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 75959034

Full Text

Citrus Station Mimeo Report CES 6h-3
October 8, 1963

A Colorimeter for Grading Citrus Juices

Richard S. Hunter
Hunter Associates Laboratory, Inc.
McLean, Virginia

I. Introduction

For some time now, there has been a need in the citrus industry for a simple,
objective instrument to measure juice color. The instrument should not depend on
human eyes, yet it should give results that correlate accurately with the averages
of grade scores by skilled human observers. Primarily, such an instrument would
be used for orange juice, but there are needs for color grading of grapefruit juice,
tangerine juice and other citrus juices.

The Florida Citrus Commission last spring placed an order with the Hunter
Associates Laboratory for a prototype citrus colorimeter. It was specified that
construction of the instrument should be simple and rugged so that a device suit-
able for the processing laboratory could be built economically.

The present paper describes the development of the first prototype instrument.
The objectives of the development, the plans to carry it out and the experimental
work involved in testing and calibrating the new device are described.

II. Background

The importance of the color of citrus juices and concentrates has been dis-
cussed by a number of investigators. In orange juices, the desirable color is
reddish orange. Most of the work in the present paper deals with orange juice.

The color of orange juice tends to vary (in Florida at least) with variety and
time of harvest. Mixtures of juices from different varieties are customarily blend-
ed in order to produce juices having acceptable color.

At present, the color scoring of orange juice is guided by a series of one-
inch diameter glass test tubes containing colored plastic jells. These were pre-
pared by the Fruit and Vegetable Division, Agricultural Marketing Service. They
were carefully formulated to have the translucent appearance of orange juice. The
series of tubes, currently numbered 1, 2, 3 and h, represent yellow limits for
color scores of 38, 36, 31h and 32, respectively, of reconstituted frozen orange
juice. That is, tube No. 3 represents the poorest juice that can receive a score
of 3h. These tubes were prepared visually. It is believed that colorimetric
specifications for these tubes are not available.

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
O00 10/8/63-RSH


The bulk of published data on the color of orange juice has been prepared by
Wenzel, Huggart and others at the Florida Citrus Experiment Station Y/. Most of
their measurements have been made with the Hunter Photoelectric Color Difference
Meter 2/. This instrument is designed to measure the colors of reflecting surfaces
on three readily understandable, visually uniform color scales:

L measures lightness and varies from 100 for perfect white to zero
for black,
a measures redness when plus, gray when zero, and greeness when minus,
b measures yellowness when plus, gray when zero, and blueness when

As an alternative to the L (lightness scale), there is an Rd (diffuse reflectance
scale) which is frequently used for comparisons with results obtained by other
instruments. Wenzel and Huggart used the Rd scales for all their work.

In a paper presented a year ago 3/, Wenzel and Huggart showed the following
correlations between values on the Color Difference Meter and numerical ratings
on the U, S. Department of Agriculture color grades:
a +0.91
Rd -0.82
b +0.49

The a scale which measures redness gives the best correlation with visual rating.
This would be expected since the redness of orange juice color is visually the
attribute which corresponds most closely with score. Figure 1 is a scatter diagram
from the 1962 paper 3/ showing USDA color scores and Color Difference Meter a
values for a large number of reconstituted frozen juices.

III. Selection of Measurement Scales

Orange juice owes its color to a mixture of pigments, the concentration of
which vary with variety and other factors. Because there are a number of pigments,
it is necessary, if results are to correlate with visual appearance, that a tri-
stimulus method be used for color measurement. This means that when photoelectric
cells are used, they must be combined with filters converting spectral responses to
those of the average human eye. The Multipurpose Reflectometer, the Color Differ-
ence Meter and other instruments developed by the writer employ the method of
photoelectric tristimulus colorimetry 4/.

The Color Difference Meter a scale is designed to measure redness of surfaces

a = 175 fy (X-Y)

where X and Y are tristimulus reflectances and fy is a nonlinear function of Y.
Because of this nonlinear fy function, a is not a simple quantity to measure.

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
400-10/8/63 RSH









* *

:. O C


y 39.44+0.87 x




"a" Values

For Reconstituted Orange Juices

f1or 91S' OCv$ e jAits

0 A *
* -~e
0 Qe# 0

36 F



.* *

*.1I *2*

* *. .

* *

33 F-







-- 1 -- -- ---

*. "

- -


Co 16

,3. I


Simpler functions tending to correlate with Color Difference Meter a are X-Y and
X/Y, With liquids X/Y is preferred because X-Y will vary significantly with
lightness as well as with redness. (It should be remembered that the Color
Difference Meter a scale was designed for visually uniform measurements of the
colors of opaque surfaces.) In appearance, orange juice is neither an opaque
surface nor a clear liquid; it is intermediate. For its measurement, however, the
X/Y ratio scale is definitely closer to what is needed than would be an X-Y scale.

In Wenzel and Huggart's paper on the colors of 215 samples of orange juice 3/
are given linear equations for the best fit between Color Difference Meter scales
and USDA grade scores for reconstituted frozen juices. From these equations were
computed average Color Difference Meter color values for each of the USDA grade
levels. From these Color Difference Meter values it was possible to compute
average values of X and Y. (Y is simply Rd.) In Figure 2, the X/Y ratio is
plotted as a function of standard USDA color score.

In Figure 2 is a second curve showing the relation of 0.8 (A/Y) to USDA color
score. In this equation, A refers to amber reflectance; i.e., reflactance for
only the amber, or long wave portion of the X tristimulus function. This A function
has been widely used in tristimulus colorimetry by Hunter and others h/; indeed, the
Hunter Color Difference Meter is the only instrument of this type which measures the
complex X function rather than only the amber part of it.

The A function was studied in the present case because it was suspected
(correctly) that A/Y might give better grade differentiation of orange juice color
than the X/Y function. It is possible to convert from Color Difference Meter
values to A/Y by using the following equation:

A/y = 1 + .0071a .6 ..00357b
fyY fyY

This equation may be derived from equations in two of Hunter's papers 2, h/.

It can be seen in Figure 2 that in comparison with X/Y, the A/Y ratio gives
about a 50% improvement in steepness of slope when plotted against USDA grade.
Since the simple A filter is easier to build than the complex X filter and because
it further gives better grade separation, the A/Y ratio was selected for orange
juice grading in the Citrus Colorimeter.

Figure 3 compares values of the A/Y ratio with Color Difference Meter values
of a and b at Y = 30%. This is an average Y value for orange juice. In this
same figure are hue and value contour lines for Munsell renotation colors at
Value 6 (Y = 30%).

IV. Plan of the Instrument

It was because one-inch diameter test tubes are already used for visual in-
spection and other tests of citrus juices that it was decided to use one-inch
tubes to hold juices in the new colorimeter.

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
00-10/8/63 RSH


D0 DDo


.8 (A/ Y

Fig. 2 i/I and 0.8(A/T) plotted against USDA scores derived from Wensl-Huggrt
data on 215 reconstituted frozen concentrates.









;:A ,,. !4 L
5t ICI

ii t-4, -"4ir '~
T7 L i;~i~- i -

*-i-- -H- r 1 i ** i i i t+- : *

I -

... I .i

i-i i .. t- .... 2 :T....i --ii~-
,$ =- T 1- I t l-. i I o a f J 1 t 1 r .' t ? o. 1 : L 4.' : A -



-trt 4-- _- ,_ .i t

~~7j~~~~~L I, :Iz~


.' '; -:1


-;": -. .7,!^ !- t .. ... ;:

- 1 ; ,
\:: \[ L

I -1 -I

L/LT IJ.J2 -

iltli iS~i .ili: FtiigS-Iffi-lM
. l ..'. .. .. .. .

-'1 .-

j4 .. .. .. .. ." I 1 __ I I J. .;

T -. .-,. ,. .... .-
-I. 1 ...,

4"![ !!:[ :'. +I Z:-,h-,i+: it~i'-: )t-4L: ] : J :- ': .: i:- t L:; T='! !44 Il; !;\: b::;tt-L~i: -- ;ti4 :, -:f[Zz t }'I! E6 :t;, f : tI$ ..!l, i''i_
":" f,- % L ,1 .. .. .. I .f /i,.I -.-



7S F

L. :F ,

4 t4-h

, I ,--

ig. 3 Color diagram on which m sell renovation colors at Value 6 and A/- ratios are plotted against A and b.
Fig. 3 Color diagram on utich !hnell renovation colors at Ya,.ue 6 and A/I ratios are plotted against a and b.





,;t7 -tt

L:^ ^ l tI~ rfeI;i fi';^ '!
I; -,i i T~~fFr 7 iIA`


-'..-1 T--- -i
,,., ~ ~ ~ .. .. ..: .,- [. :,'i ;" ? F"-- rT ,_ L :


-- --------- -- -------


t--+ I t-h*> *i-+---I-4- t-** I ;-111H.;- l- t 1 '*** *" '*'r .; I~ r.:l -I I_ f .1 *IM :; I* ^ .n :.' I L .


~t~l ;j ~TF~;-~~`~;F;~;;;iir~Ti;~rit;"~ .....
hittt t!~H-~r~( Et;

~ ~---

- I I .V


1 II rl I -i i IrNNL 31 1

" I if" 1 : 11111 L I i z i

~ I~:-~1-i ~t

7 ---


-~~~~-~--~ --~--~

_4_ : .. ', : : .. -- ,. -.-

S. -r ......... .-4 [ -
-c^ ^ n1-f:^-( ~;:^^^;^

L '' I i. '-:)' -' ,,__-C__ :1L- --I -

I.1.L L

:r;i .~I;-.-
I .-r j:;l h: t-

------- --- -- -- - -- ----- ------- -



~ -- ~-- -+-


r:_,j j ,'


Figure h is a diagram of the new device; Figure 5 is a photograph of the
first prototype.

To make the instrument small and compact and at the same time to receive min-
imum heat, the lamp was mounted outside the instrument housing at the rear. A
heat-absorbing glass and condensing lens are arranged in the manner shown in Figure
h to project a 1 x 1/2" spot of light on the specimen tube. For accuracy and stabil-
ity, the light pipe and thermostatted phototube block of the D25 instrument were

For redness measurements, only Y (green) and A (amber) filters and photo tubes
are required. However, for yellowness measurements of grapefruit and other juices,
a Z (blue) phototube can be added. The standard phototube block of the Color
Difference Meter has positions for three phototubes.

For ratio measurements, the Wheatstone bridge circuit in Figure 6 was used.
In this bridge, the ratio A/Y is measured by the reciprocal of the relative re-
sistances required to produce balance (Ry/RA). It will be seen in Figure 2 that
only 18% of the total 0.8 (A/Y) ratio range is required to cover the scale of
juice scores from 30 to hO. This 18% was expanded to 20% of the full range of
ratios. The 80% of the ratio range is added, as is shown in Figure 6, by fixed
resistance. Settings in this 80% range are not needed nor are they possible.

V. Affect of Translucency on Measured Color

Although much data on the color of orange juices including that used for
Figures 1 and 2 has been obtained by the Florida Citrus Commission, it was decided
to make Color Difference Meter measurements of typical juice colors at Hunterlab
before starting to calibrate the new instrument. Orange juices are notably trans-
lucent and therefore it seemed important to find whether measurements by the Florida
Citrus Commission had been affected by the geometric conditions of measurement.
Fresh juices were measured and supplied by the Florida Citrus Commission for this

Figure 7 shows how markedly the geometric conditions of measurement can affect
the results of color measurements of orange juices. This light-penetration effect
on measured values of food color was discussed by the writer in a paper before the
Instituted of Food Technologists in 1960 V/. Where light can travel within a color
specimen, some of it may, if the specimen window is small, be trapped behind the
edge of this window and not be permitted to escape for measurement. Where light is
thus trapped, values of Rd or L measure lower than otherwise. In addition, the
measured chromaticity will be affected because the wavelengths not absorbed in the
material will travel farther and be more likely to be trapped than those wave-
lengths strongly absorbed. Because the red wavelengths are least absorbed, orange
juice measures greener and darker than it. appears wherever the instrument window
traps reflected light.

The changes are large. Those shown in Figure 7 amount to about six units on
the USDA scoring scale. The problem is serious for photoelectric colorimetry of

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
0OO 10/8/63- RSH







Fig.1, Diagram of Citrus Colorimeter









Fig. 6 Wheatstone bridge circuit used to measure







Fig. 7 Rd' a color diagram ahowibg change of measured color with
specimen window area.



many foods because the standards normally used are opaque and not subject to this
color shift. It will be decidedly advantageous to use standards such as the USDA
plastic tubes which are not opaque like most standards, but instead are optically
like the product measured.

VI. Calibration and Standardization

Munsell glossy papers in the color range of orange juice and having reflect-
ances of about 20% (Value 5) and 30% (Value 6) were used for the colorimetric
calibration of the new apparatus. These were obtained from the Munsell Company
together with their colorimetric specifications obtained by the GE Spectrophoto-
meter. Figure 8 is a graph in which spectrophotometer values of A/Y are compared
with those obtained from the new Citrus Colorimeter. The Munsell papers were
measured in two ways: (1) they were held flat against the interior specimen
window and (2) they were mounted on a strip of cardboard which was inserted di-
agonally inside a one-inch test tube to simulate the approximate position of the
sample during measurement. It will be seen in Figure 8 that the colorimetric
accuracy of the new device is quite good and that results between the two methods
of mounting the Munsell papers are in substantial agreement.

Figure 9 shows an actual comparison of scale readings of the new instrument
with color scores by the present scoring system. Three groups of samples were
measured to obtain data for this figure. Five of the USDA standard color tubes
were measured at the Fruit and Vegetable Division of the USDA in Washington. The
A/Y ratio representing each of these five tubes is plotted as an "S" midway be-
tween the mean score which each represents and the next score below. This was
done because these tubes are used as color separation points, not average colors
for the scores which they represent. The dots in Figure 9 represent the A/Y
ratios of a number of. juices which were graded at the Florida Citrus Commission
and then shipped frozen to Hunterlab for use in the construction of the new
instrument. Note that both concentrates and reconstituted juices were measured
in these series. The third lot of specimens were shipped frozen from the Minute
Maid Oorporation. These were also measured in both reconstituted and concentrated
form. The lines in Figure 9 were drawn to give visually the best fit between ob-
servations and measurements. Computed A/Y ratios are given on the left side of
this drawing. Corresponding readings from the actual dial of the instrument are
shown on the right.

Figure 9 shows excellent agreement between measured values of the USDA tubes
and color scores assigned the juices supplied for the present study. Color inter-
vals between the five USDA tubes appears to be quite uniform although tube No. 3
seems to be slightly low, tube No. h slightly high in color. The tube between
No, 39 and h0 is actually the 00 tube. The point representing this tube is below
the line which curves upward on this right side of the chart because of the one
sample of California Valencia concentrate which was given a score of 40. The score
for this juice should probably be higher than its assigned value of hO.

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
0oo 10/8/63 RSH

1.10 1.20 1.30



Fig. 8 Colorimetric Accuracy of orange juice colorimeter
measured by Munsell calibrated paper standards.






I I I 1000

1.30- 9I0(
^ x

S* 700

JS 6C0
>1.20 500Q

FCC ~ 300
x MM
x TUBES 200

I I I I I I I I 0 O
32 33 34 35 36 57 38 39 40

Fig. 9 Relation of A/I to color &core for USDA standard tubes, reconstituted juice, and concentrates.

VII. Effect of Pulp

When design of the present instrument was first discussed, the writer asked
whether variations in pulp content might not affect instrument measurements in a
manner not in accord with the present visual scoring. A series of 12 specimens
was prepared at the Florida Citrus Commission to explore this possibility. To a
base juice was added first 4% then 8% and finally 12% of coarse pulp by volume.
Specimens of each of these samples as well as of the original juice were retained.
The remainder of each sample was then passed through a colloid mill to divide the
pulp into finer particles. The affects of these treatments on measured color are
shown in Figure 10. It will be seen that after dispersal in the 0.24" colloid
mill, parts of each sample were further homogenized by passage through a colloid
mill at 0.010" clearance.

Figure 10 shows little change of color by the colloid mill for the sample to
which no pulp was added. However, addition of the coarse pulp made the sample
lighter and redder. Dispersion of this pulp in the colloid mill increased the
lightness much further, but caused the sample to go green in chromaticity. The
fine .010" dispersion caused the samples to go still greener, though not lighter.

The effects shown in Figure 10 are like those which appear in colored glass
as it is ground into finer particles. There are more particle surfaces to reflect
light and shorter distances of travel within the particles where coloration occurs.
As a consequence, making particles smaller causes materials to be come lighter in
color and to differ in chromaticity from the characteristic color of the pigment.
The amount and dispersion of the pulp in the juice does affect its color.

The 12 juice specimens studied for pulp content and degree of dispersion in
Figure 10 were graded for USDA color score in Florida by four persons. These
scores are compared in Figure 11 with the A/Y ratios of these specimens. It will
be seen that the juices dropped in A/Y ratio as they were passed through the
colloid mill. In the upper left-hand corner of Figure 11 is a small part of the
Figure 9 curve relating score to A/Y readings. All four of the specimens were a
little below this line before passage through the colloid mill, but those contain-
ing pulp all dropped well below after passage through the colloid mill.

This decrease in score with dispersion of the pulp is undoubtedly related to
the marked increase in reflectance shown in Figure 10. Wenzel and Huggart found a
negative correlation of -0.82 between reflectance and score. A pulp is penalized
in score when it is light. This is the affect demonstrated in Figure 11.

The Citrus Colorimeter would have to be increased in complexity in order to
make it respond to Rd as well as A/Y. Whether this increase in complexity is
justified by practical requirements is a matter for possible further study.

VIII. Summary

A simple chromaticity-sensing ratio colorimeter has been designed and built to
measure the redness of orange juice. This instrument takes specimens in the standard

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
4oo 10/8/63 RSH










0 p











Fig. 11 Change of A/T and score with addition of coarse pulp and subsequent dispersal
with colloid mill

12% -


AT .024"
AT .010"


32 -




one-inch diameter test tubes regularly used in the laboratory testing of orange
juice. There is a single digital dial on the instrument reading to three places.
Provision has been made to add a Z phototube and to measure yellowness by the
ratio Z/Y. This modification will be used to grade grapefruit and other citrus

IX. References

la. R. L. Huggart and F. W. Wenzel. Measurement and Control of Color
of Orange Concentrate, Proc. Florida State Hort. Soc. 67, 210

lb. R. L. Huggart and F. W. Wenzel. Color Differences of Citrus Juices
and Concentrates Using the Hunter Color-Difference Meter. Food
Technology 2, 27 (1955).

lc. F. W. Wenzel, R. W. Barron, R. L. Huggart, R. W. Olsen and M. D.
Maraulja. Comparison of Color and Flavor in Frozen Concentrated
Orange Juice. Proc. Florida State Hort. Soc. 71, 274 (1958).

2. R. S. Hunter. Photoelectric Color Difference Meter. J. Opt. Soc.
Am. 48, 985 (1958).

3. F. W. Wenzel and R. L. Huggart. Relation Between Hunter Color-
Difference Meter Values and Visual Color of Commercial Frozen Con-
centrated Orange Juice. Proc. Florida State Hort. Soc. 75, 331

4. R. S, Hunter. Photoelectric Tristimulus Colorimetry With Three
Filters. Nat. Bur. Stds. Circ. Ch29 (1942).

5. R. S. Hunter. Standardized Color Measurements of Food Products.
Unpublished paper presented before Inst. Food Tech. (May 18, 1960).

Florida Citrus Commission and
Florida Citrus Experiment Station,
Lake Alfred, Florida.
400-10/8/63 RSH

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs