476
7
~~ ~
Table 11. Recopery of 2,4-D from Prepared Rlixtures by
Titration Method
Petroleum
Sample
Oleic
Acid 011 dr%d 2,4-D
Added
2,4-D
Found
% % % % 70
A 4 09 44.48 46 55 4 88 4 78
B 8.20 22.60 59;lO 10 10 10 33
cipitate. Place a 100-ml. aliquot in a 250-ml. separatory funnel,
acidify with hydrochloric acid, and proceed as in (1) with the
ether extraction, evaporation, and titration.
4. 2,4-D in Presence of Oil, Soap, and Alcohol. Dilute a
sample weight equivalent t o 0.5 gram of 2,4-D t o 75 ml. with pe-
troleum ether and transfer to a 250-ml. separatory funnel. Ex-
tract with six consecutive 10-nil. portions of a 50% alcohol-water
solution, t,o each portion of which have been added a few drops
of 407, potassium hydroxide solution and a fcm drops of phenol-
phthalein indicator. Draw off the combined alcohol-water
extracts into a 250-ml. beaker ahd place on a steam bath to re-
niove the alcohol. Transfer the aqueous residue to a 200-ml.
volumetric flask and cool t o slightly below room temperature.
Proceed as in (3), beginning with “add 1 t o 1 hydrochloric acid
to the disappearance of the pink color.”
Determination of Flavanones in Citrus Fruits
W. B. DAVIS
Laboratory of Fruit and Vegetable Chemistry, c‘. S. Department of Agriculture, Los .4rigeles, Calif.
This paper describes a new colorimetric method
using alkaline diethylene glycol for the determi-
nation of the bitter rhamnoglycoside naringin and
other fla\anones that may be present in grapefruit
in particular, as well as in other citrus fruits (6).
.Although not specific for naringin, it is a rapid pro-
cedure of practical value, which is particularly
applicable to the assay of naringin in the jhice and
colored flavedo of grapefruit, and of hesperidin in
other citrus fruits. The possibility of other sub-
stances interfering w-ith the method is discussed.
Citral, furfural, and geraniol produced color with
ETfIODS already- described for estimating naringin are
either SIOK or not specific. Poore ( 9 ) ,who used a crystalli-
zation procedure, found 0.0ec6 naringin in grapefruit juice; how-
ever, this method is not adaptable to rapid determinations. The
ferric chloride method (4,10) used for the determination of narin-
gin in the comparatively colorless albedo tissues of grapefruit
cannot be applied to the juice, because citric acid and other hy-
droxy compounds present develop a strong yellow color with fer-
ric chloride. This color masks the weak reddish-broFn color pro-
duced by such small quantities of naringin as occur in grapefruit
juice. Although the boric acid method (11) for hesperidin is ap-
plicable to naringin and is sensitive to small quantities, it is slow
and inconvenient. Gibbs’ test ( 3 ) gives no distinctive color with
naringin, while the procedure of Folin and Ciocalteu ( 2 ) produccs
rcsults that are clearly too high (0.3 to 0.8%).
A pink color is produced when a solution of naringin in alcohol
is boiled with hydrochloric acid and magnesium. However, at-
tempts t o use the color formed as a basis for a colorimetric test
have been unsuccessful. An investigation of the addition of al-
kali to solutions of naringin and hesperidin led to the observation
that a stable yellon color n-as formed on the addition of strong
sodium hydiovide to a solution of the flavanonr in diethylenr gly-
col. Fairly concentrated sugar solutions, glycerols, other glycols,
and even ~ a t e develop
r color with naringin when made strongly
V O L U M E 19, NO.
5. 2,4-D in Alcohol or Glycol Carriers. Dilute a sample
nTeight equivalent to 0.8 to 1.0 gram of 2,4-D to 75 ml. with 9570
ethyl alcohol and titrate directly with 0.1 N sodium hydroxide
solution.
AYALYTICAL DATA
In Table I are results obtained on some relatively pure com-
pounds and samples of commercial herbicides containing 2,4-D
by the titration and total chlorine methods. Differences in re-
sults obtained by these methods may be attributed t o experimen-
tal error in the analyses performed by the Parr bomb where
0.1 ml. of 0 1 S silver nitrate is equivalent to an error of about 1%
on 100-mg. samples containing 2,4-D 99%. .4nalytical data on
samples prepared in the laboratory containing 2,4-D dissolved in
varying amounts of oleic acid, petroleum oil, and amyl alcohol
are prcbr)ntcti in Table 11.
LITERATCRE CITED
(1) ParrInstiumeiit Co., Moline, Ill., Diwction Booklet 116,
(2) Stepanow, A., Ber., 39, 4066 (1906).
(3) Umhoefer, R., ISD I?%. CHEY.,A N < LED..15, 383 ( 1 9 4 3 ) .
(4) Zimmerman, P. IT., Iiad. Eng. Chem., 35 596 (1943). .
alkaline diethylene glycol but did not interfere with
the method when added to grapefruit juice in
quantities larger than those ordinarily found in
that juicc. A n interesting difference between the
behavior of flavanones afid flabones in the extracts
of certain other plants is pointed out and the sug-
gestion made that the method may be useful in the
determination of flavones. The method has been
used to determine the distribution of flavanones in
the various tissues of citrus fruits and the recovery
of pure naringin added to various mixtures, and to
follow the course of naringin hydrolysis.
alkaline, but are less suitable and effective than dietkiylene glycol
as a reagent.
Color intensity appears to be specifically related to the flava-
nones present when the reaction is carried out in the manner de-
scribed below. This specificity is indicated by the fact that, when
extra flavanone glucosides are added to various extracts of citrus
fruits, not only niay the addition be accurately determined, but
the results may also be extrapolated to give highly probable val-
ues for the original flavanone content of the extract.
PROCEDURE
Ten milliliters of 9 0 7 diethylene glycol are placed in a Iilett-
Summerson colorimeter tube, and 0.2 ml. of an unknown extract,,
or of the solution to be tested, is added and mixed. The colorim-
eter is then adjusted to its zero reading (using a blue filter) to give
a starting point that allows for the natural color of the extract.
Spectral transmittance curves for the color developed by naringin
and naringenin with the reagent show a narrow minimum near
420 mp, thus necessitating the use of the blue filter. Thereafter
0.2 ml. of approximately 4 N sodium hydroxide is added and the
increase in color is read after 5 minutes. The observed color in-
creases are compared with standard curves prepared from the
pure flavanone, or its glucoside, known to be dominantly present
in the species of fruit used-e.g., naringin in grapefruit juicc, and
hesperidin in orange juice. Readings of duplicate analyses in the
Klett colorinictcr usually agree n-ithin 1 or 2 scale divisions. The
development and reading of the color should take place at a defi-
nite temperature.
JULY 1947 477

I slight increase 0x1 standing), the maximum coloration took only a

few seconds to develop, in contrast to solutions of naringin and
"z
u)
120 -
hesperidin that required several minutes. This behavior was also
exhibited by rutin (melting point, 189 "), the only pure flavonol
a
a- / derivative available for testing. The present method may there-
wu)
a w i fore be useful in the determination of flavones and flavonols.
E 5100-
w I' (Further studies have shown that pure rutin glycoside supplied
by the courtesy of J. F. Couch, of the Eastern Regional Research
Laboratory, may be determined by this method. The test was
made by the Pharmacology Laboratory in the Kestern Regional
Research Laboratory.)
Ascorbic acid also develops color with alkaline diethylene gly-
col, but this color decreases within a &minute period. For this
reason there is less difference between readings made a t the end of
1- and 5-minute periods for citrus juice than for a solution of purr
naringin. The color developed by ascorbic acid is in the neigh-
borhood of 5 t o 6 q of the total color produccd by the grapefruit
juices tested.
The folloFving substances, some of xvhich are possible constitu-
ents of grapefruit juice, do not interfere with the method: stachy-
drin, betaine, phloroglucinol, isolimonin (5), linalool, methyl an-
50 100 150 200 250
MICROGRAMS PER IO ML. REAGENT AT 25' C
thranilate, octyl alcohol, octyl aldehyde, decyl aldehyde, methyl
heptenon, cineole, citranellal, toluene, riboflavin, and the caroten-
Figure 1. Standard Curves oids present in the juice. The following substances produced
color with alkaline diet,hylene glycol but did not interfere when
Standard Solutions, A convenient method of preparing a added t o grapefruit juice in quantities larger than those ordinarily
standard solution of naringin is t o add about 40 ml. of distilled found in grapefruit juice: citral, furfural, and geraniol. Eriodic-
water to 25 mg. of recrystallized naringin dried a t 100" C. The tyol glycoside from Eriodictyon glutinosa, supplied by the Califor-
suspension is then heated t o about 70" 6.
(naringin in this sus-
pension may decompose a t temperatures above 83" C.), cooled, nia Fruit GroLVers FJxchange, \vas found to color slor~1yTvit,l1 the
and diluted to 50 ml. Naringin does not precipitate from this reagent.
solution for several hours. To prepare a standard solution of Recovery of Added Naringin. Tablr I shotvs the recovery of
hesperidin, 1 t o 2 drops of 10 N sodium hydroxide are added to a pure naringiii added t o various mixturrs and the effect upon its
suspension of hesperidin in water or alcohol. After dilute acid recovery of several mc.thodr of treatment. The velocity of the
(citric acid may be used) is added until the solution is slightly
acid, the clear solution is diluted to volume without warming. color development is shown. The fact that readings a t the end
of IC minutes differ only slightly from those taken at the end of 5
Standard curves for naringin and hesperidin, made with the minutes indicates that the reaction is nt.arly complete in the
above standard solutions, are shown in Figure 1. shortrr time. I t is furthcr seen that treatment with alcohol or
neutral lead acetate dors not remove naringin from grapefruit
EXPERIMENTAL
juice. Recovrry of added naringin is satisfactory.

Interfering Substances. Many plant substances give yellow APPLICATIOE OF THE METHOD
colors when mixed with an alkali, and therefore this interference
must be taken into account. Fortunately, in the case of citrus Flavanones in Citrus Fruits. Determinations were made of the
extracts and juice, such interference appears to be minor under distribution of naringin and hesperidin in tissues of available
the reaction conditions. However, l l a y e r ( 7 ) and Perkin and citrus fruits. Table I1 gives the distribution of the flavanones in
Everest (8)list all such flavones, including the flavanones naringin grapefruit, Valencia orange, lemon, and tangerine. The juice
and hesperidin, benzophenones,
including phloridzin, and the
xanthones, such as genrtisin,
which are less common ( 1 ) . Table I. Recovery of Added Naringin from Citrus Juices and Plant Extracts
A study was made of the de- (Temperature 25' C.)
Colorimeter Readings, Minutes % Recovery
velopment of a yellow color in Description 1 2 3 4 5 10 of Naringin
alkaline diethylene glycol by 1. Fresh undiluted grapefruit juice 120 150 162 169 171 171 ...
alcoholic or aqueous extracts 2. Fresh undiluted grapefruit juice precipitated by
60% alcohol and centrifuged 120 151 162 167 168 ... ...
of some plants reported t o con- 3. 25 ml. of fresh juice diluted to 100 ml. 32a 36 41 42 42 42 ...
4. Solution of pure naringin 15 mg. per 100 nil. 500 67 72 74 76 76 ...
tain flavones. Extracts of the 5. Mixture of 3 and 4 i n oriLinal dilution 83a 105 113 114 115 115 97.5
following p l a n t s were e s - 6 . Mixture 5 precipitated with neutral lead acetate.
Lead removed with sodium oxalate 86'7 104 110 112 112 112 95.7
amined: potato tuber, onion 7. Canned grapefruit juice. Dilution n i in 3. Carot-
enoids removed 47 55 ... 60 63 ... ,..
skin (cont,aining quercetin), 8. Saringin solution, 25 mg. per 100 nil. ,. , ,. . .. .., 126 , .. ...
petals of POPPY (containing 9 . Mixture of 7 and 8 in original dilution . . . . . . , . 186 . . . 9 8.4
Fresh orange juice 35 38 . . , 40 40 40 ...
rutin), peach bark (containing 11. ZIixture of 8 and 10 in original dilution 81 119 149 160 160 . . 96.4
20 24 . . . 25 25 . . .
phloridzin), and the leaves of ~ 1 2dilution
~ ~ & & ~ ~ \ ji n ~original 78 133 136 158 158 . lb6:O
Robinia (containing acacetin), 14. Solution of pure naringin, 15 mg. per 100 irrl. 42a 61 67 71 72 7i ...
15. Extract of California poppy petals (rutin) 34a 34 34 34 33 30 ...
parsley ( c o n t a i n i n g a p i i n ) , 16. Solution of hesperidin, 15 mg. per 100 ml. 10a 12 14 15 16 18
17. Mixture of 14, 15, and 16 in original dilutiou 865 107 115 120 121 120 1bb:O
Rumex (containing quercetin), 18. h-aringin solution, 15 mg. per 100 mi. 47 63 72 76 76 77 ...
senna (containing kaempferol), 19. Mixture of extracts of flavedo, albedo, and section
membranes 43 55 62 64 64 65
apricot, dandelion, gentian, and 20. Jfixture 19 with naringin 18 in original dilutions 89 122 136 143 , 144 . .. lO2:l
walnut. In every case except hverage of duplicates.
one (gentian root extract gave a
478
Table 11. Distribution of Flavanones in Citrus Fruits
Juice Tissue
Fresh, centri- Sediment Total Juice Section
fuged (1) (2) +
(1) (2) sacs membrane Albedo
Glycoside p e r 100 ml. Glycoside p e r 100 grams fresh weight
Grapefruit ‘ I 0.038 0.003 0.041 0 140 1.250
Lemon h 0 046 0.008 0.054 0.600 1.900
Orangeh 0.080 0.038 0,118 ... 1.500
Tangerine b 0.065 ... ... ... 0,500
a Calculated as naringin.
b Calculated as hesperidin: eriodyctiol glucoside may be present and would be included in the calculation as
hesperidin
C Determination made on whole peel.
d Fibrovxsridar and adhering tissue (tangerines have no core).
was q u e r z r d out in such a way that it came in contact only with
the juice sac membranes, and the time of contact therewith was
minimized. Each natural division of the fruit was secured sepa-
rately from other parts, as far as possible. Although the bitter,
rather dry fibrovascular core contains the highest amount of
flavanontJs, it represents but a small part of the whole fruit.
Since both the juice and membranes contain naringin, there was a
possibility that some of the flavanone contained in the juice was
really of membrane origin. Hoviever, determination of naringin
in juirr removed from the juice sacs of grapefruit by means of a
glass nredle showed a concentration but little smaller than that in
juice expressed directly in contact with sac membranes-that is,
squeezed through cloth from separated juice sacs. One experi-
ment i n which 1-minute contact with macerated section mem-
b r a ~ ~doubled
=% the concentration of naringin in centrifuged juice
s h o w that t,he juice quickly extracts naringin from section mem-
branes. I t is therefore expected that with this method it may be
possible to correlate differences in the bitterness of juice with thc
method of its extraction and with the condition of the fruit.
Course of Naringin Hydrolysis. Naringenin, the aglucone of
naringin, reacts Tvith the alkaline reagent to form much lees color
Device for Estimating the Height of Polarographic Waves
JOHN KEENAN TAYLOR
National Bureau of Standards, Washington 25, D . C .
4 simple device which facilitates the measurement
of current-\oltage curkes can be used for measuring
diffusion currents by either graphical or exact meth-
ods. P hen the slope of the portion of the current-
toltage curve following the wale is approximately
equal to that of the initial segment, the graphical
method may be used with confidence and the de-
vice measures the diffusion current automaticall>
corrected for increase in residual current. When
Q CASTITATIVE polarographic analysis depends upon es-
timating the heights of iyaves or steps on current-voltage
curves. I n comparative methods the relative magnitude of the
waves found for an unknown and a standard solution are required,
while in the absolute method the heights of the waves must be
measured to determine the values of diffusion currents to be sub-
stituted in the IlkoviE equation or some modification of it.
The usual practice has been to determine the vertical displace-
ment (increase in current) of points on, or portions of, the cur-
rent-voltage curve immediately preceding and succeeding the
wave, and to correct for increase in residual current over the same
voltage interval by graphical means or from measurements on a
VOLUME 19, N O . 7
than naringin (Figure 1). Sam-
ples taken a t intervals t o the
end of hydrolysis show a con-
Flavedo Core tinuous decrease in color on re-
acting with the reagent. This
2.100 0.600 2.600 makes it possible t o follow the
3.000 2.500 course of naringin hydrolysis.
1.600 1.000 41500
1.700C 1,570d 9 study of enzymic hydrolysis
is being made using the new
method for flavanones.
ACKNOWLEDGMENT
The author is indebted to
C. IT, Wilson, California Fruit
Growers Exchange, for details of the boric acid method for
hesperidin, and to E. J. Eastmond, Western Regional Research
Laboratory, for the determination of the spectrophotometric
curves for naringin and naringenin.
LITERATURE CITED
(1) Abderhalden, E., “Biochemische Handlexikon,” Band 6, Berlin,
Julius Springer, 1911-33.
(2) Folin, O., and Ciocalteu, V.,J . Biol. Chem., 7 3 , 627 (1927).
(3) Gibbs, H. D., Ibid., 72, 649 (1927).
(4) Harvey, E. M.,and Rygg, G. L., Plant Physiol., 11, 4G3-5
(1936).
(5) Higby, R. H., J . Am. Chem. Soc., 60, 3013 (1938).
(6) Matlack, M.B., “Bibliography on the Chemistry of the Genus
Citrus (1931),” Supplement, U. S. Department of Agriculture,
1945.
(7) Mayer, F., “Chemistry of Natural Coloring Matters,” Kew
York, Rheinhold Publishing Corp., 1943.
(8) Perkin, A . G., and Everest, A . E., “Natural Organic Colouring
Matters,” London, Longmans, Green and Co., 1918.
(9) Poore, H. D., Ind. Eng. Chem., 26, 637 (1934).
(10) Rygg, G. L., and Harvey, E. >I., Plant Physiol., 13,571 (1938).
(11) Wilson, C. IT., J . -4m.C‘hem. SOC.,61, 2303 (1939).
the slopes are significantly different, values obtained
by the graphical method should be used only for
empirical relation of w-ave height to the concentra-
tion of the reducible ion. If, in this case, the wave
height is measured by the exact method, several
points, both preceding and succeeding the w-ave,
should be selected for measurement and the agree-
ment of the diffusion current measured at each used
as a criterion of ideality of the electrode process.
curve taken for the supporting electrolyte alone. I n either proce-
dure, lines are ordinarily drawn through the selected portions of
the curves, using parallel rulers, drafting triangles, etc., and the
vertical displacement is determined with a suitable scale or a pair
of dividers.
The practice of draxing lines on polarograms is somewhat ob-
jectionable. Marks so placed tend to prejudice any subsequent
estimation of the wave height by the same or a different observer.
I n addition, when using a triangle or other transparent straight-
edge as an aid in averaging the curve to determine 1% here to draw
the representative line, the optical asymmetry of the operation
makes the measurement difficult and unsatisfactory.