You are on page 1of 10

Components of Flavour

I
Interactions among salty,
opposite is often true, as when two weak compounds

sour and bitter compounds enhance one another2,‘J6.


Moskowitz” carefully explored whether the enhance-
ment or suppression seen in many mixtures was either
true synergism or true masking (see Box 1 and Fig. 3).
That is, he determined whether the shape of the con-
Paul AS. Breslin centration-intensity function (‘power function’; i.e. the
perceived intensity plotted as a function of the con-
centration of tastant) for a single compound changed
when tasted against a background taste. He found that
The human gustatory system is capable of responding to and the power functions shifted laterally but the exponents
processing the taste of solitary compounds in water. How- (slopes) of the functions did not change. In other words,
the rate of growth of the intensity as a function of con-
ever, the taste system rarely contacts solitary compounds out-
centration was unaltered by the presence of a back-
side the laboratory and has surely evolved to process complex ground compound. A leftward shift of the power func-
mixtures of sapid chemicals, such as occur in virtually all tion is defined as enhancement by the background taste.
foods. This review will focus primarily on the lesser-studied A rightward shift is defined as suppression by the back-
interactions between pairs of salty, sour and bitter compounds. ground taste. When the slope of the function changes to
become either steeper or shallower (a rare occurrence in
Pair-wise interactions among these three taste qualities should
tastant mixtures), then the background compound is af-
be of interest because they constitute a significant proportion fecting the rate of growth of taste intensity as a function
(-30-50%) of possible binary taste interactions. In general, of concentration, which is defined as synergism or
salts and acids enhance each other at moderate concen- masking (see Box 1 and Fig. 3).
There are several reports of binary tastant mixtures
trations but suppress each other at higher concentrations.
resulting in the enhancement of individual taste qualities.
Bitter compounds and acids can either enhance or suppress Reliable enhancement has been reported with mixtures
each other depending on the concentrations, the food stimuli such as combinations of sweetenersi0*i8.i9, mixtures of
and the experimental methods involved. Sodium salts and monosodium glutamate (MSG) with 5’-ribonucleotides9~*o,
bitter compounds generally interact so that bitterness is or mixtures of sucrose with dilute NaC121.Interestingly,
the enhancements seen in mixtures of 5’ribonucleotides
suppressed to some variable degree and the saltiness is un-
and MSG and in certain combinations of intense sweet-
affected. As will be described below, there are exceptions to eners are the only apparent cases of synergism that have
all of these generalities. been identified, in which the taste of the two com-
ponents in mixture is greater than any linear model
would predict based on their tastes when unmixed9*‘0~‘9.
When two compounds that elicit different taste qualities One of the more striking features of the literature on
are mixed in solution at moderate or strong concen- heterogeneous taste mixtures is the lack of agreement
trations, the mixture will often yield a taste sensation that across laboratories. There are several reasons for this that
is less intense than the simple sum of the component should be kept in mind when evaluating the literature:
tastes. In two-component mixtures, each of the taste
(1) Differences in individual subjects’ sensitivities to vari-
qualities is usually suppressed (perceived as less intense)
ous taste compounds can be quite pronounced and this
than when it is tasted separately’-4 (see Box 1 and Fig. 1).
results in very different responses to mixtures of the
Binary combinations of certain taste stimuli may also
compounds22,23.
result in extremely asymmetrical suppressions. For ex-
ample, Schifferstein and Frijters”, and Breslin and (2) There is a tendency within the field to expect all
Beauchamp12 reported that when quinine hydrochloride compounds that share a common taste quality (e.g. all
(QHCl), which usually elicits a bitter taste, was mixed bitter-tasting compounds) to interact with other com-
in solution with sodium chloride (NaCl), which usually pounds (e.g. all sour-tasting compounds) in a similar
elicits a salty taste, the saltiness of the NaCl was rela- way. This has resulted in questions such as ‘How do
tively unaffected, whereas the bitterness of the QHCl sour stimuli interact with bitter stimuli?’ being asked.
was suppressed by 50-70% (see also Ref. 13), depend- This question cannot be answered simply, because even
ing on the concentrations involved (contrast with Refs 3 if all sour stimuli elicit the same sourness quality and all
and 14). Although taste suppression is commonly ob- bitter stimuli elicit the same bitterness quality, they are
served, in no instance has a parallel been found in taste likely to do so through multiple transduction mecha-
to the color opponency phenomenon, in which two nisms24-26;and enhancement and suppression may, in
lights cancel each other’s qualities15. In fact, in taste, the the case of some interactions, involve transductive path-
ways directly27-29. Because the effects of some taste
Paul AS. Breslin is at the Monell Chemical Senses Center, 3500 Market mixtures will be mechanism specific, one cannot expect
Street, Philadelphia, PA 19104, USA (fax: +l-215-898-2084; e-mail: all bitter-tasting stimuli to behave the same way in the
breslin@cattell.psych.upenn.edu). presence of an acid (or acids) and vice versa.

390 Copyrlght01996. Elsewer Science Ltd. All rights reserved. 0924 -2244/961$15.00 Trends in Food Science & Technology December 1996 [Vol. 71
PII:50924-2244(96)10039-X Special Issue on Flavour Perception
Box 1. Definitions of types of potential interactions between two mixed co~~~pounds
of dierent tastes
linear their individual thresholds that, when totalled, sum to less than
WW’
I”U 10 ‘I\
rthlj.
a 45% for A and 50% for B, or 45% for A and 45% for 8,
Suppression: (for suprathreshold concentrations) etc.)7. WY+~-~,~.
*or*Adi+i\re mixing bears a direct parallel to the supra-
Suppression is a linear (or subtractive) process that bears direct A”,“t...r.l p~~~~~“lllcnon
A.‘.,,...,
u~~~~~~“Iu of enhancement.
analogy with subthreshold hypoadditivd mixing as well as with
subtractive mixing (see below), except that it is applied here to Hypoaddiivemixing:(for subthreshold concentrations)
suprathreshold compounds. As shown in Fig. 1, it is most clearly Hypoadditive mixing is defined in the same way as additive mixing
I depicted by a rightward shift of the concentration-intensity func- except that the threshold for the nmixture is obtained when the
tioh along. the concentration axis in the presence of a second compounds are at concentrations corresponding to percentages of
compound, so that every concentration point along the curve is their individual thresholds that, when totalled, sum to more than
per&&d as less intense. One should no& that the-function itself 100% (e.g. 55% fur A and 50% for B, or 55% for A and 55% for B,
is not finear and will not result in the same magnitude decrease etc.17, Hypoadditive mixing bears aI direct parallel to the supra-
at different positions along the concentration-in&&y curve (see thr&hoId phenomenon of suppression.
Fig. 2P. Subtractivemixing:(for subthreshold and suprathreshold mixtures)
1Enhancement:(for suprathreshold concentrations) Subtractive mixing involves the determination of a threshold for
Enhancement is a linear (or additive) process that bears direct anal- compound A in the presence of a suprathreshold background con-
ogy with subthreshold hyperadditive mixing (see below), except centration of compound 6 (Ref. 7). This usually results in com-
that it is applied here to suprathreshold compounds. As shown in pound B elevating the threshold for compound A.
Fig. I, it is most clearly depicted by a iefhrvard shift of the concen-
tration-intensity function along the concentration axis in the pres- Nonlinear
ence of a second compound, so that every concentration point A&I&~: (for suprathreshold concentrations)
along the curve is perceived as more intense. Again, one should Masking is a nonlinear process in which the addition of compound
note that the function itself is not linear and will not result in the
A decreases the intensity of compound B in a manner that goes
same magnitude increase at different positions along the concen-
beyond iinear reductions in intensity. As shown in Fig. 3, it is most
tration-intensity curve (see Fig. 2)jr6,
clearly depicted by a decrease in the slope of the function and a
Additivemixing:(for subthreshold concentrations) rightward shift of the concentration-intensity function along the
Additive mixing may be the same process that governs enhance- concentration axis in the presence of compound B, so that every
ment and suppression. Additive mixing involves the determination concentration point along the curve is perceived as less intense.
of the threshold concentration for compound A in the presence of One should note that the function itself is not linear and will not
a subthreshold concentration of compound B. Simple (completet result in the same magnitude decrease at different positions along
additivity means that, when the concentrations of A and B are re- the concentration-intensity curve (see Fig. 2P-“‘.
duced below their individual threshold concentrations and mixed Synergism:[for suprathreshold concentrations)
together, the detection threshold for the mixture will occur Synergy is a nonlinear process in which the addition of compound
when each compound is at 50% of its concentration threshold (or A increases the intensity of compound B in a manner that goes
30 : 70; 70 : 30; 40 : 60; 60:40, etc.), so that the sum of the per-
beyond linear increments in intensity. As shown in Fig. 3, it is most
centages of the individual threshold concentrations for A and B
clearly depicted by an increase in the slope of the function and
equals 100%. This usually results in nearly additive combinations
a leftward shift of the concentration-intensity function along the
of compounds A and B (Ref. 7). rnnrontrs*ion axis in the presence of compound B, so that every
C”II\.~Ilil~I

Hvwradditive mixine:(for subthreshold concentrationsl concentrat ion point along the curve is perceived as more intense.
Hypradditive mixin; is defined in the same way as a;fditive mix- One should note that the function itself is not linear and will not
ing except that the threshold for the mixture is obtained when the result in the same magnitude increase at different positions along
compounds are at concentrations corresponding to percentages of the concentration-intensity curve (see Fig. 2P”‘O.

(3) Several (perhaps all) strong suprathreshold taste mix- experiments) than when just the -3O-40% of taste buds
tures display non-monotonic interactions (the direction on the end of the tongue are engaged30 (as in dorsal-flow
of the interaction changes, sometimes increasing taste experiments). Furthermore, even when comparing re-
and sometimes decreasing taste, depending on the initial sults from several studies that used the same dorsal-flow
magnitude of the taste in question). For example, NaCl design, the rate of the flow and the temperature of the
enhances the sweetness of sucrose when the NaCl is solutions can both markedly affect the magnitude of the
present at a low concentration, has no effect when pres- result@. It has even been observed that the type of
ent at a moderate concentration, and suppresses the experimental design can change a mixture interaction
sweetness of sucrose when present at a high concen- from weakly suppressing to strongly suppressing6, pre-
tration2’ (see ‘Conclusions’ section). Given that many sumably owing to changes in the shape of the concen-
researchers work with one or only a few concentrations, tration-intensity power functions of the unmixed com-
and different laboratories select different concentrations pound&6,31 (see Fig. 2). In addition, both the experimental
with which to work, disagreement seems inevitable. context in which ratings are made and the number of
scales on which a subject rates simultaneously can have
(4) Different experimental protocols will often yield dis-
an impact on the magnitude of the ratings”,32-36.
crepant results. For example, whole-mouth sipping ex-
poses the stimuli to a much greater array of potential (5) There are very large differences in the types of mixture
taste receptor interactions (palatine and epiglottal as well interactions depending on whether near-threshold (concen-
as pharyngeal) than would a simple lingual dorsal-flow trations that are barely detectable or not detectable at all)
design; that is, compounds will taste differently when or obviously supratbreshold concentrations are examined.
-90% of taste buds are stimulated (as in ‘sip-and-spit’ This point is distinct from point 3, which discusses only

Trends in Food Science & Technology December 1996 [Vol. 71 391


x
.ti
2
3
E

,‘-

Concentration

1 Nochang; inslope 1

Enhancement Suppression

,
.
.
_ _ - - .

:.......
......
,i
L

presence
Linear scaling.
concentration-intensity
like a Weibul
of a background
shift the function linearly
The upper

or logistic
Concentration

function
function.
panel depicts

concentration
for a tastant
The lower
a typical

of an added
left or right along the x-axis,
that is shaped
panel shows how the
compound
indicating
Fig. 1
psychometric
roughly

may
that the
1J I
Expansive,
Concentration

linear and compressive portions of a psychometric


Fig. 2

background additive is enhancing or suppressing taste, respectively.


function. All three panels depict the exact same psychometric
Note that the shape of the function, however, is unaltered.
concentration-intensity function for a tastant that is shaped
roughly like a Weibul or logistic function. In the upper panel, the
interactions with easily detected suprathreshold concen-
low-concentration portion of the function is outlined to indicate the
trations of tastants. Subthreshold mixtures of qualitatively
region where the function is expansive. Thus, two solutions of low
different-tasting compounds have generally yielded additive
combinations7, whereas almost all very strongly supra- concentration should combine hyperadditively. In the middle panel,
threshold mixed-quality interactions yield suppression1-4. the moderate-concentration portion of the function is outlined to show
that here it is linear. In this concentration range, two solutions should
In what follows, I review information about various types combine additively. In the lower panel, the high-concentration portion
of mixture interactions between pairs of salty, bitter and of the function is outlined to indicate where it is compressive. Thus,
sour compounds, with an emphasis on the enhancement or two solutions of high concentration should combine hypoadditively.
suppression of taste qualities in binary mixtures. Most
Note that all three of these types of interactions occur whenever the
of the reviewed research focused on single judgements of
function is ‘S’ shaped, and therefore will occur for a function that is
quality magnitude, rather than multiple magnitude ratings
over time, such as in time-intensity experiments. The shifted along the x-axis but maintains its ‘S’ shape as in Figs 1 and 3.
present focus on individual ratings of sourness, bitterness
or saltiness in binary mixtures is not meant to imply that both simultaneous (split-tongue)29 and successive con-
combinations of qualities in mixtures may not elicit new trasts among taste combinations of different qualities.
taste qualities that individual compounds cannot elicit37.
However, many subjects appear to be able to relate to the Mixtures of salts (NaCI) and acids
experimenter’s request to introspect about single attributes Much of the work on salt-acid mixtures is reviewed
of a complex taste stimulus, as they might do during every- in Table 1. The following points briefly summarize the
day experiences, such as when they report that vegetables conclusions that can be drawn from this table.
need salt, or fruit taste too bitter, or lemonade is too sour.
In this review, I will not fully evaluate the literature Subthreshold compounds mixed with subthreshold
regarding the location of mixture interactions (i.e. distin- compounds
guish among effects taking place in the oral epithelium, Salts and acids tended to display slight hyperaddi-
peripheral nerves and the brain). Nor will I review the tivity when subthreshold concentrations were mixed. The
various types of adaptation studies that have examined approximate additivity of these different compounds (for

392 Trends in Food Science & Technology December 1996 [Vol. 71


and saltiness, respectively, in a variety of media with a
range of methods.

Strongly suprathreshold compounds mixed with strongly


suprathreshold compounds
1 Suppression was much more commonly observed
when strongly suprathreshold concentrations of acids
.--
Change in slope
and salts were mixed. There were also several obser-
,/-.-’ vations of no interaction in mixtures of compounds at
these high intensities. Enhancement was rarely seen.

*‘I
J ,
.,’
__-’ Mixtures of bitter-tasting compounds and acids
Much of the work on bitter-acid mixtures is reviewed
in Table 2. The following points briefly summarize the
conclusions that can be drawn from this table.
Synergy
Subthreshold compounds mixed with subthreshold
compounds
Bitter-tasting compounds and acids tended to display
/, slight hypoadditivity when subthreshold concentrations
were mixed.

Subthreshold (or threshold) compounds mixed with


Concentration suprathreshold compounds
, In one study, Pangbom42 found that caffeine and citric
Fig. 3 acid suppressed each other at very low concentrations.
Nonlinear scaling. The upper panel depicts a typical psychometric In contrast, von Skramlik41 found that the threshold for
concentration-intensity function for a tastant that is shaped roughly
tartaric acid was lowered by quinine HCl. In other
words, sensitivity was enhanced.
like a Weibul or logistic function. The middle panel shows how the
addition of a background concentration of a compound may alter the Weakly suprathreshold compounds mixed with weakly
shape of the function nonlinearly, rendering it steeper or shallower. (or strongly) suprathreshold compounds
The lower panel shows how the presence of the background additive The sourness of citric acid was enhanced by caffeine2,
may shift the function along the x-axis, leftward and steepening it, or whereas the bitterness of quinine sulfate was suppressed
rightward and shallowing it, indicating that the additive is interacting by low concentrations of tartaric acid16.
synergistically or is masking the taste, respectively.
Strongly suprathreshold compounds mixed with strongly
all three mixture types, see also Tables 2 and 3) suggestssuprathreshold compounds
that, at subthreshold concentrations, their effects summate The results with strongly suprathreshold concen-
in a pathway common to both chemicals somewhere in trations of acids and bitter-tasting compounds were
the nervous system, although most likely not in the same mixed. There is an account of sourness being enhanced
peripheral taste receptor cell. by mixture with a bitter compound [H. Gerigk (1955)
iiber die Miigh’chkeiten der Sinnlichen Unterdriickung
Subthreshold (or threshold) compounds mixed with auf dem Gebiete des Geschmacksinnes (On the Possi-
suprathreshold compounds bilities of Sensory Suppression in the Area of the Taste
When subthreshold concentrations of a salt (NaCl) Sense) (PhD thesis), Berlin, Humboldt Universitat,
were mixed with strongly to weakly suprathreshold con- Berlin, Germany], and an account of bitterness being
centrations of various acids, the salt tended to suppress enhanced by mixture with an acid2. On the other hand,
the perceived sourness (or elevated the threshold for the there are two accounts of acids and bitter-tasting com-
acid) (see Table 1). pounds suppressing each other, with the exception that
Similarly, acids suppressed the perceived saltiness of moderate concentrations of tartaric acid are not sup-
salts, with two notable exceptions: subthreshold concen- pressed by quinine sulfate 16,45.Frank et aE.32found that
trations of many acids enhanced the saltiness of NaC13*, citric acid either enhanced or suppressed the bitterness
and tartar% acid lowered the threshold (i.e. enhanced the of quinine depending on how the subjects rated the
saltiness) for NaC141. bitterness.

Weakly suprathreshold compounds mixed with weakly Mixtures of salts and bitter-tasting compounds
(or strongly) suprathreshold compounds Much of the work on salt-bitter mixtures is reviewed
In general, weakly suprathreshold concentrations of in Table 3. To briefly summarize, the following conclu-
acids and salts tended to enhance one another’s sourness sions can be drawn from this table.

Trends in Food Science & Technology December 1996 [Vol. 71 393


Table 1. Mixtures of salts and acids

Components
Mixture intensity level Salt Acid Outcome Ref.

Subthreshold t subthreshold NaCl Citric Slightly hyperadditive Stevens (1995)’

Subthreshold [or threshold) NaCl Many Subthreshold concentrations of NaCl moderately suppressed Fabian & Blum (1 943)38
t suprathreshold the sourness of acetic, HCI and citric acids without
altering the pH (suppression)

NaCl Many Subthreshold concentrations of NaCl greatly suppressed Fabian & Blum (1943)‘”
the sourness of lactic, malic and tartaric acids without
altering the pH (suppression)

NaCl Many Subthreshold concentrations of many acids increased Fabian & Blum (1 943)38
the saltiness of NaCl (enhancement)

NaCl HCI Seven out of ten panelists reported no interaction; three Fabian & Blum (1 943)38
out of ten reported that subthreshold HCI reduced the
saltiness of NaCl (suppression)

NaCl HCI HCI increased the threshold for NaCl (subtractive mixing) Anderson (1 950)3g

NaCl HCl Thresholds were raised when suprathreshold concentrations Heymans (1 899)40
of one were added to the other (subtractive mixing)

NaCl Tartaric Most subtractive mixtures elevated both thresholds, but tartaric von Skramlik (1962)“’
acid lowered the threshold for NaCl (subtractive mixing)

NaCl Citric Subthreshold concentrations of NaCl reduced the sourness Pangborn (1960)“’
of citric acid (suppression)

Weakly suprathreshold t weakly NaCl Citric Weak NaCl and weak citric acid resulted in enhanced Pangborn & Trabue
(or strongly) suprathreshold saltiness and mixed effects on sourness (enhancement) (1 964)23

NaCl Citric Citric acid enhanced the saltiness of NaCI, when NaCl Kamen et a/. (1961)j
was weak (enhancement); the sourness of weak citric acid
was enhanced by NaCl (enhancement)

NaCl Tartaric High concentrations of tartaric acid enhanced the saltiness lndow (1 969)16
of weak NaCl (enhancement); high NaCl concentrations
enhanced the sourness of weak tartar? acid (enhancement)

NaCl Lactic NaCl enhanced the sourness of weak lactic acid (enhancement) Hellemann (1 992)43

NaCl Acetic NaCl enhanced the sourness of acetic acid in rye bread Hellemann (1 992)43
(enhancement)

,Strongly suprathreshold NaCl Tattaric NaCl and tartaric acid suppressed each other Hambloch & Puschel
t strongly suprathreshold (suppression) (1928)+’

NaCl Tartaric Tartaric acid had no impact on the saltiness of NaCl (no effect) lndow (1 969)16

NaCl Tartaric Tartaric acid and NaCl suppressed each other (suppression) Gerigk (1955)”

NaCl Citric Citric acid reduced the perceived saltiness of NaCl (suppression) Pangborn (1 960)42

NaCl Citric The sourness of citric acid was suppressed by NaCl Pangborn & Chrisp
(suppression); the saltiness of NaCl was enhanced by (1 964)22
citric acid - but variability was very high (enhancement?)

NaCl Citric The saltiness of NaCl was unaffected by citric acid (no effect); Kamen et a/. (1961)*
the sourness of citric acid was suppressed by NaCl (suppression)

NaCl Lactic NaCl suppressed the sourness of lactic acid (suppression) Hellemann (1 992)43

NaCl Acetic NaCl suppressed the sourness of acetic acid (suppression) Hellemann (1 992)43

NaCl HCI NaCl and HCI were mildly enhanced or mildly suppressed Kiesow (1 894)45
depending on the subject (mixed effects)

NaCl HCl The sourness of HCI was unaffected by NaCl (no effect) von Skramlik (1962)41

NaCl HCI NaCl and HCI did not interact with one another (no effect) Bartoshuk (1975)’

NaCl HCI NaCl had no effect on the sourness of HCI (no effect) Cragg (1 937)46

“PhD thesis; see text for details

394 Trends in Food Science & Technology December 1996 [Vol. 71


[able 2. Mixtures of bitter-tasting compounds and acids

Components

Mixture intensity level Bitter Acid Outcome Ref.

Subthreshold + subthreshold Quinine HCI Citric Slightly hypoadditive Stevens (1995)’

Subthreshold (or threshold) Quinine HCI Tartaric Most subtractive mixtures elevated thresholds, von Skramlik (1 962)4’
t suprathreshold but quinine HCI lowered the threshold for tartaric
acid (subtractive mixing)

Caffeine Citric Subthreshold concentrations of caffeine Pangborn (1 960)42


suppressed the sourness of citric acid
(suppression); subthreshold citric acid suppressed
the bitterness of caffeine (suppression)

tieakly suprathreshold Caffeine Citric Caffeine enhanced the sourness of citric acid, Kamen et a/. (1961)2
t weakly (or strongly) especially at low acid concentrations
suprathreshold (enhancement)

Quinine SO, Tartaric Low concentrations of tartaric acid suppressed lndow (1 969)16
the bitterness of quinine SO, (suppression)

Strongly suprathreshold Variable Variable Bitterness and sourness did not interact Kiesow (1 894)45
t strongly suprathreshold (no effect)

Quinine HCI Tartaric Quinine HCI and tartaric acid suppressed Hambloch & Puschel (1 928)44
each other (suppression)

Quinine HCI Tartaric The sourness of tartaric acid was enhanced Gerigk (1955)’
by quinine HCI (enhancement)

Quinine SO, Tartaric High concentrations of quinine SO4 suppressed lndow (1969)r6
the sourness of low and high concentrations
but not moderate concentrations of tartaric
acid (suppression)

Quinine HCI Citric Citric acid either enhanced or suppressed the Frank et al. (1 993)32
bitterness of quinine depending on how the
ratings were made

Caffeine Citric Citric acid greatly enhanced the bitterness Kamen et al. (1961)*
of caffeine for all concentrations of caffeine
(enhancement)

‘PhD thesis; see text for details

Subthreshold compounds mixed with subthreshold not affected2,‘2.47.Notable exceptions to this trend were
compounds MgS04 and amiloride HCl, which were the only two
Salts and bitter-tasting compounds tended to display bitter compounds to suppress the saltiness of NaCl at
slight hypoadditivity when subthreshold concentrations moderate salt concentrations12.
were mixed. The magnitude of the suppression of bitter taste by
NaCl varied greatly for the different bitter-tasting com-
Subthreshold (or threshold) compounds mixed with pounds tested, ranging from -5% suppression of the
suprathreshold compounds; weakly suprathreshold most bitter concentration of MgSO, tested to -80% sup-
compounds mixed with weakly (or strongly) pression of the most bitter concentration of urea. Non-
suprathreshold compounds; and strongly suprathreshold sodium chloride salts (with the exception of LiCl) were
compounds mixed with strongly suprathreshold not able to suppress the bitterness of any compound to
compounds any great degree.
All three intensity levels yielded comparable results There was a single example of bitterness enhance-
and are therefore considered together here. Although ment by NaCl when mixed with iso-c.u-acids48.
quinine and caffeine are the two most commonly em-
ployed bitter-tasting stimuli in the psychophysical litera- Conclusions
ture, there is at least one report in which a larger array ‘How do bitter and sour tastants interact in mixture?’
of salts and bitter compounds was used’*. In all but one ‘Does sourness increase saltiness when in mixture?’
instance, sodium salts suppressed bitterness to some de- ‘Will bitterness affect saltiness?’
gree. The opposite was not always true. Although there I argue here that such questions do not really make sense.
are some accounts of bitter compounds suppressing the Mixture interactions will be dependent on the specific
saltiness of NaCl, there are others in which saltiness was chemicals in question, the magnitude of their perceived

Trends in Food Science & Technology December 1996 [Vol. 71 395


Table 3. Mixtures of bitter-tasting compounds and salts
I-
Mixture intensity level Bitter Salt Outcome Ref.

Subthreshold t subthreshold Quinine HCI NaCl Slightly hypoadditive Stevens (I 995)’

Subthreshold (or threshold) Caffeine NaCl Subthreshold and threshold caffeine reduced the saltiness of NaCl (suppression); subthreshold Pangborn (I 960)42
t suprathreshold and threshold concentrations of NaCl suppressed the bitterness of caffeine (suppression)
Quinine HCI NaCl Quinine HCI increased the threshold for NaCl (subtractive mixing) Anderson (I 950)39

Weakly suprathreshold t weakly Quinine HCI NaCl Low concentrations of NaCl suppressed the bitterness of quinine (suppression) Hopkins (I 953)47
(or strongly) suprathreshold

Strongly suprathreshold Quinine SO, NaCl All concentrations of NaCl reduced the bitterness of all but the weakest quinine SO, lndow (1 969)16
t strongly suprathreshold solutions (suppression)
Quinine SO, NaCl Quinine SO, had no effect on the saltiness of NaCl (no effect) Hopkins (1 953)47
Quinine SO, NaCl The bitterness of quinine SO, was suppressed by NaCl (suppression) Kroeze (I 980)*8
Quinine HCI NaCl NaCl and quinine HCI suppressed each other (suppression) Hambloch & Puschel (I 928)”
Quinine HCI NaCl Quinine HCI suppressed the saltiness of NaCl (suppression) Cerigk (I 955)”
Quinine HCI NaCl The bitterness of quinine HCI was suppressed by NaCl (suppression) Bartoshuk (I 975)’
Quinine HCI NaCl NaCl suppressed the bitter taste of quinine HCI at the same tongue location (suppression) Kroeze & Bartoshuk (1985)2q
Quinine HCI NaCl NaCl suppressed the bitterness of quinine HCI (suppression) Schifferstein & Frijters (1992)”
Quinine HCI NaCl NaCl suppressed the bitterness of quinine HCI by 41% (suppression) Breslin & Beauchamp (1995)‘*
Caffeine NaCl Caffeine and NaCl had no effect on each other. Although a couple of significant interactions Kamen et a/. (196I)j
were observed, the authors questioned their own findings owing to large error terms (no effect)
Caffeine NaCl NaCl suppressed the bitterness of caffeine by 55% (suppression) Breslin & Beauchamp (1995)12

MgQ NaCl NaCl suppressed the bitterness of MgSO, by 4% (suppression) Breslin & Beauchamp (1995)‘*
Amiioride HCI NaCl NaCl suppressed the bitterness of amiloride HCI by 69% (suppression) Breslin & Beauchamp (1995)i2
KCI NaCl NaCl suppressed the bitterness of KCI by 78% (suppression) Breslin & Beauchamp (I 995)12
Urea NaCl NaCl suppressed the bitterness of urea by 76% (suppression) Breslin & Beauchamp (I 995)12
Urea Na acetate Na acetate suppressed the bitterness of urea by an amount comparable to that of NaCl (suppression) Breslin & Beauchamp (I 995)”
Quinine HCI Na acetate Na acetate suppressed the bitterness of quinine HCI by an amount comparable to that Breslin & Beauchamp (1 995)12
of NaCl (suppression)

Urea Na gluconate Na gluconate was as effective as NaCl at suppressing the bitterness of urea (suppression) Breslin & Beauchamp (I 995)”

Urea LiCl LiCl suppressed the bitterness of urea by 81% (suppression) Breslin & Beauchamp (1 995)12

WQ NaCl MgSO, suppressed the saltiness of moderate concentrations of NaCl (suppression) Breslin & Beauchamp (I 995)i2
Amiloride HCI NaCl Amiloride HCI suppressed the saltiness of moderate concentrations of NaCl (suppression) Breslin & Beauchamp (I 995)12

Iso-a-acids NaCl NaCl enhanced the bitter taste of isoa-acids (enhancement) Yokomukai et a/. (I 994)48

‘PhD thesis; see text for details


intensity, the particular concentrations examined, whether Kroeze’” points out that enhancement and suppression
they are subthreshold or suprathreshold, and the particular may also often be predicted by the side tastes that many
method that is used to answer the question32*36. compounds elicit, such as the sweet taste of weak NaCl
Despite the need for these specificities in order to an- concentrations, which may play a significant role in
swer such questions, trends in binary mixture interac- suppression or enhancement. For example, the sweet
tions may be seen from the literature. Foremost among taste of weak NaCl concentrations may be responsible
them is that mixtures of subthreshold, threshold and for the enhancement of sweetness in mixtures with
even weakly suprathreshold compounds behave quite sucrose21. Kroezer4 boldly goes on to conclude that be-
differently from mixtures of medium to highly supra- cause of the existence of such side tastes, enhancement
threshold compounds. Mixtures of compounds that elicit in moderate- to high-concentration mixtures is almost
different taste qualities are reported to be additive on always spurious (with the possible exceptions of
average - usually either slightly hyperadditive or hypo- MSG and S-ribonucleotides, certain intense sweetener
additive - when a subthreshold concentration of one is mixtures, and NaCl mixed with iso-o-acids). In other
mixed with a subthreshold concentration of another’. words, the same taste quality is in effect being added
Subthreshold, threshold and weakly suprathreshold to itself, resulting in apparent ‘enhancement’. Dis-
compounds mixed with threshold and weakly supra- agreement among laboratories may be accounted for by
threshold compounds may result in suppression, but differences in the concentrations employed or in the
often instead in enhancement, depending on the par- compounds selected, as just described. However, there
ticular compounds in question and their precise inten- are even simpler arguments to account for disparities in
sities (for example, see Ref. 40; also see tables). observations. Pangbom42 suggested that many of the
Highly suprathreshold mixtures usually suppress each discrepancies among laboratories may be due to the em-
other, sometimes symmetrically and sometimes very ployment of expert panels in some studies, and naive
asymmetrically12~42~49. panels in others. Gregson and McCowen50 also attribute
The linear decreases (suppression) or increases (en- inter-laboratory differences to variations in the sensitiv-
hancement) seen in most suprathreshold mixtures are ities of the different populations of panelists. However,
predicted by the shape of the concentration-intensity there are several reliable interactions on which different
power function for each of the individual compound#. laboratories concur. The rather impressive interactions
Compressive power functions for tastants (i.e. those that among:
have exponents less than one - where doubling the con-
S-ribonucleotides and MSG, which enhance each
centration results in less than doubling of the intensity)
other;
result in mixture suppression for that compound (it is
less than purely additive). Expansive power func- certain intense sweeteners, which also enhance each
tions (i.e. those that have exponents greater than one - other; or
where doubling the concentration results in more than
sodium salts and certain bitter-tasting compounds, in
doubling of the intensity) yield enhancement in mixture
which the bitterness is suppressed,
(it is greater than purely additive)31. Note that the effects
of these mixtures are a direct linear prediction from the may be separated from all other forms of mixture inter-
shape of the function. action in that these three are reliable, large-magnitude
This idea may be extended. Because many psycho- effects that occur at relatively high concentrations and,
physical concentration-intensity functions display a sig- in the case of MSG and ribonucleotides, and certain
moidal or Weibul-like character, they will, by their very intense sweetener mixtures, are not predicted by the
nature, display both expansion and compression for shapes of their power functions6,10J9.Interestingly, it has
lower and higher concentration ranges, respectively (see been suggested that all three of these interactions have
Fig. 2). This extension of Bartoshuk’s idea may account a peripheral transductive/receptor locus and are not cog-
for why salts and acids usually display enhancement at nitive interactions.
weak concentrations, when functions are expansive, and One may surmise functional and adaptive reasons to
suppression at high concentrations, when functions are explain why the inter-quality interactions discussed here
compressive (see Table 1). This idea may also account vary as a function of the different intensity levels. First,
for why intense sweeteners often enhance one an- the additivity of compounds at the weakest intensity level
other-18,19.That is, their mixture interactions are rarely (subthreshold mixtures) has many practical implications.
tested at concentration ranges where the concentration- At the weakest mixture concentrations, when we experi-
intensity functions show saturation (compression), but ence complex mixtures of subthreshold compounds in our
rather are tested where the function is expansive (begin- dally lives, we may use subthreshold additivity effects to
ning of function) or linear (middle of function), where detect their presence. For example, it is possible that many
enhancement is predicted. Were they to be administered different mineral waters are distinguishable, despite the
at the highest testable concentrations, they would be ex- fact that their mineral and salt contents are in the parts
pected to show suppression. In fact, Frank et aLlo found per billion range, because the subthreshold minerals and
that intense sweeteners exhibited hyperadditivity at low salts combine additively to generate distinguishable tastes
concentrations, additivity at moderate concentrations, (assuming there are no olfactory cues)‘. Second, the di-
and hypoadditivity at high concentrations. versity in mixture interactions at low concentrations,

Trends in Food Science & Technology December 1996 [Vol. 71 397


which generate suppression in some instances and en- compounds and concentrations. However, the magnitude
hancement in others, may be utilized to modulate fla- of the bitterness suppression varied greatly from com-
vors. A good example of this is the role that salt plays in pound to compound, and one bitter compound was en-
breads. An acceptable bread does not usually appear to hanced by NaCl. Non-NaCl salts did not tend to have as
taste salty, although it contains moderate concentrations great a general suppressive capacity, with the exception
of salt; however, when the same bread recipe is fol- of LiCl. Only two bitter-tasting compounds suppressed
lowed without the added salt, the bread is markedly the saltiness of NaCl clearly.
worse in taste, almost unpalatable to most. Glutamate,
and possibly other amino acids, may play a comparable Acknowledgements
modulatory role in soup flavor. Third, the high-concen- I wish to thank Gary K. Beauchamp and Bruce P.
tration interactions, which are usually suppressive, may Halpem for reading and commenting on an earlier draft
serve a more general inhibitory function. For example, of this manuscript.
many fruit and plants (e.g. citrus fruit, tomato, pine-
apple) are extremely acidic when unripe. As they ripen, References
1 Bartoshuk, L.M. (1975) ‘Taste Mixtures: Is Mixture Suppression Related to
the sugar concentration increases (and the acid concen- Compression!’ in Physiol. Behav. 14, 643-649
tration may decrease). The increase in sugar concen- 2 Kamen, J.M., Pilgrim, F.J., Cutman, NJ. and Kroll, B.J. (1961) ‘Interactions of
tration serves to suppress the perceived sourness, in the Suprathreshold Taste Stimuli’ in ). Exp. Psycho/. 62, 348-356
3 Kemp, S.E. and Beauchamp, C.K. (1994) ‘Flavor Modification by Sodium
same way that sugar suppresses the sourness of lemon-
Chloride and Monosodium Glutamate’ in ). Food Sci. 59,682-686
ade. This strategy may allow fruit to inhibit microbial 4 McBride, R.L. (1989) ‘Three Models of Taste Mixtures’ in Perception of
growth with high acid concentrations, yet retain a de- Complex Smells and Tastes (Laing, D.C., Cain., W.S., McBride, R.L. and
gree of attractiveness to certain seed-carrying animals. Ache, B.W., eds), pp. 265-282, Academic Press
5 Kroeze, J.H.A. (1989) ‘Is Taste Mixture Suppression a Peripheral or Central
Interactions among the three qualities salt, bitter and Event?’ in Perception of Complex Smells and Tastes (Laing, D.G., Cain, W.S.,
sour should be of great interest to the food industry be- McBride, R.L. and Ache, B.W., eds), pp. 225-243, Academic Press
cause they shed light on such major issues as: 6 Bartoshuk, L.M. and Gent, J.F. (1985) ‘Taste Mixtures: An Analysis of Synthesis’
in Taste, O/faction, and the Central Nervous System (Pfaff, D.W., ed.),
l how to modulate or reduce bitterness in foods, oral- pp. 210-232, Rockefeller University Press, New York, NY, USA
care products and pharmaceuticals; 7 Stevens, J.C. (1995) ‘Detection of Heteroquality Taste Mixtures’ in Percept.
Psychophys. 57, 18-26
l how to maintain the effects of NaCl on the flavor of 8 Stevens, S.S. (1966) ‘Power Group Transformations Under Glare, Masking, and
foods without adding sodium (or while reducing Recruitment’ in ). Acoust Sot. Am. 39, 725-735
concentrations); 9 R&in, B. and Bartoshuk, L.M. (1980) ‘Taste Synergism Between Monosodium
Glutamate and Disodium 5’.Guanylate’ in Physiol. Behav. 24, 1169-l 172
l how to alter the perception of sourness without chang- 10 Frank, R.A., Mize, S.J.S. and Carter, R. (1989) ‘An Assessment of Binary
ing the levels of acid present. Mixture Interactions for Nine Sweeteners’ in Chem. Senses 14, 621-632
11 Schifferstein, H.N.J. and Frijters, J.E.R. (1992) ‘Two-stimulus Versus
One-stimulus Procedure in the Framework of Functional Measurement:
Summary A Comparative Investigation Using QuinineHCVNaCl Mixtures’ in Chem.
Salts and acids Senses 17, 127-l 50
12 Breslin, P.A.S. and Beauchamp, G.K. (1995) ‘Suppression of Bitterness by
Interactions among salts and acids were non-mono-
Sodium: Variation Among Bitter Taste Stimuli’ in Chem. Senses 20, 609-623
tonic over different concentration ranges. Subthreshold 13 Frijters, J.E.R. and Schifferstein, H.N.J. (1994) ‘Perceptual Interactions in
salt concentrations tended to suppress sourness and vice Mixtures Containing Bitter Tasting Substances’ in Physio). Behav.
versa. However, there are two examples of subthreshold 56,1243-l 249
14 Kroeze, J.H.A. (1982) ‘The Relationship Between the Side Tastes of Masking
concentration of acids enhancing saltiness. Moderately Stimuli and Masking Binary Mixtures’ in Chem. Senses 7, 23-37
weakly suprathreshold concentrations of acids and salt 15 Hurvich, L.M. and Jameson, D. (1957) ‘An Opponent-process Theory of Color
tended to enhance one another. Strongly suprathreshold Vision’ in Psycho/. Rev. 64, 384-404
concentrations of salts and acids tended to suppress one 16 Indow, T. (1969) ‘An Application of the T Scale of Taste: Interaction Among the
Four Qualities of Taste’ in Percept. Psychophys. 5, 347-351
another. 17 Moskowitz, H.R. (1971) ‘Intensity Scales for Pure Tastes and for Taste
Mixtures’ in Percept. Psychophys. 9,51-56
Bitter-tasting compounds and acids 18 Ayya, N. and Lawless, H.T. (1992) ‘Quantitative and Qualitative Evaluation of
High-intensity Sweeteners and Sweetener Mixtures’ in Chem. Senses
Interactions among bitter-tasting compounds and 17,245-259
acids were also non-monotonic over different concen- 19 Schiffman, S.S. eta/. (1995) ‘Investigation of Synergism in Binary Mixtures of
tration ranges, and seemed particularly variable across Sweeteners’ in Brain Rex Bull. 38, 105-l 20
laboratories. There is some indication that weakly supra- 20 Yamaguchi, 5, Yoshikawa, T., Ikesa, U.S. and Ninomiya, T. (1971)
‘Measurement of the Relative Taste Intensity of Some t-u-Amino Acids and
threshold sourness is enhanced by bitterness but that 5’.Nucleotides’ in ). Food Sci. 36, 846-849
weakly suprathreshold bitterness is suppressed by sour- 21 Bujas, Z. (1934) ‘Quelques Remarques sur le Contraste et I’lnhibition a la Suite
ness. At higher concentrations, acids and bitter compounds d’Excitations Gustatives Simultanees’ in C. R. Seances Sot. Bio/. Paris
either suppressed or enhanced one another depending on 116,1304-l 306
22 Pangborn, R.M. and Chrisp, R.B. (1964) ‘Taste Interrelationships. VI. Sucrose,
the stimuli in question and the experimental methods Sodium Chloride, and Citric Acid in Canned Tomato Juice’ in ). Food Sci.
employed. 29,490-498
23 Pangborn, R.M. and Trabue, I.M. (1964) ‘Taste Interrelationships. V. Sucrose,
Sodium Chloride, and Citric Acid in Lima Bean Puree’ in ). FoodSci.
Salts and bitter-tasting compounds 29,233-240
Sodium salts almost universally suppressed per- 24 Cowart, B.J., Yokomukai, Y. and Beauchamp, G.K. (1994) ‘Bitter Taste in
ceived bitterness across a wide range of bitter-tasting Aging: Compound-specific Decline in Sensitivity’ in Physiol. Behav.

398 Trends in Food Science & Technology December 1996 [Vol. 71


56.1237-l 242 of Variable Line and Category Rating Methods’ in Percept. Psychophys.
25 Fischer, R. and Griffin, F. (1963) ‘Quinine Dimorphism: A Cardinal 54,477-484
Determinant of Taste Sensitivity’ in Nature 200, 343-347 37 Erickson, R.P., Priolo, C.V., Warwick, Z.S. and Schiffman, S.S. (19901
26 Spielman, A.I., Huque, T., Whitney, G. and Brand, J.C. (1992) ‘The Diversity ‘Synthesis of Tastes Other Than the “Primaries”: Implications for Neural
of Bitter Taste Signal Transduction Mechanisms’ in Sensory Transduction Coding Theories and the Concept of “Suppression”’ in Chem. Senses
[Corey, D.P. and Roper, S.D., eds), pp. 307-324, Rockefeller University Press, 15,495-504
New York, NY, USA 38 Fabian, F.W. and Blum, H.B. (1943) ‘Relative Potency of Some Basic Food
27 Bartoshuk, L.M. and Seibyl, J.P. (1982) ‘Suppression of QHCI in Mixtures: Constituents and Their Competitive and Compensatory Action’ in Food Res.
Possible Mechanisms’ in AChemS Abstracts, 4th Annual Meeting, p. 1, 8, 179-l 93
Association for Chemoreception Sciences, Tallahassee, FL, USA [Abstract] 39 Anderson, R.J. (1950) ‘Taste Thresholds in Stimulus Mixtures’ in Microfilm
28 Kroeze, J.H.A. (1980) ‘Masking in Two- and Three-component Taste Mixtures’ Abstr. 10, 287-288
in Olfaction and Taste Vl/(van der Starre, H., ed.), p, 435, Information 40 Heymans, C. (1899) ‘Untersuchungen iiber Psychische Hemmung’ in
Retrieval Ltd, Press, Washington, DC, USA Zentralbl. fhysiol. 21, 321-359
29 Kroeze, J.H.A. and Bartoshuk, L.M. (1985) ‘Bitterness Suppression as Revealed 41 von Skramlik, E. (1962) ‘i.&er die Erscheinungen der Positiven und Negativen
by Split-tongue Taste Stimulation in Humans’ in Physiol. Behav. 35, 779-783 Unterdrtickung beim Ceschmackssinne’ in 2. 6iol. (Munich) 113, 266-292
30 Miller, I.]., Jr and Bartoshuk, L.M. (1991) ‘Taste Perception, Taste Bud 42 Pangborn, R.M. (1960) ‘Taste Interrelationships’ in food Res. 25, 245-256
Distribution, and Spatial Relationships’ in Smelland Taste in Health and 43 Hellemann, U. (1992) ‘Perceived Taste of NaCl and Acid Mixtures in Water
Disease(Cetchell, T.V., Doty, R.L., Bartoshuk, L.M. and Snow, LB., Jr, eds), and Bread’ in Int. 1. Food SC;. Technol. 27,201-211
pp. 205-233, Raven Press 44 Hambloch, H. and Puschel, J. (192%) ‘tiber die Sinnlichen Erfolge bei
31 Bartoshuk, L.M. (1977) ‘Psychophysical Studies of Taste Mixtures’ in Olfaction Darbietung von Geschmacksmichungen’ in Z. Psycho/. Physiol. Sinnesorg.
and Taste V/ (LeMagnen, J. and Macleod, P., eds), pp. 377-384, InformatIon 59,136-l 50
Retrieval Ltd, Press, Washington, DC, USA 45 Kiesow, F. (1894) ‘Beitr;ige zur Physiologischen Psychologie des
32 Frank, R.A., van der Klaauw, N.J. and Schifferstein, H.N.J. (1993) ‘Both Ceschmackssinnes’ in Philos. Stud. 10, 523-561
Perceptual and Conceptual Factors Influence Taste-Odor and Taste-Taste 46 Cragg, L.H. (1937) ‘Sour Taste’ in Proc. Trans. R. Sot. Can. 31, 131-140
Interactions’ in Percept. Psychophys. 54, 343-354 47 Hopkins, J.W. (1953) ‘Laboratory Flavor Scoring: Two Experiments in
33 Schifferstein, H.N.J. and Frijters, I.E.R. (1992) ‘Contextual and Sequential Incomplete Blocks’ in Biometrics 9, 1-21
Effects on Judgements of Sweetness Intensity’ in Percept. Psychophys. 48 Yokomukai, Y., Breslin, P.A.S., Cowart, B.J. and Beauchamp, C.K. 11994)
52,243-255 ‘Sensitivity to the Bitterness of Iso-a-acids: The Effects of Age and Interactions
34 Schifferstein, H.N.J. (1994) ‘Contextual Effects in the Perception of with NaCI’ in Chem. Senses 19, 577
QuinineHCliNaCl Mixtures’ in Chem. Senses 19, 113-l 23 49 Amerine, M.A., Pangborn, R.M. and Roessler, E.B. (1965) ‘The Sense of Taste’
35 Schifferstein, H.N.J. (1994) ‘Sweetness Suppression in Fructose/Citric Acid in Principles of Sensory Evaluation of Food, pp. 28-l 44, Academic Press
Mixtures: A Study of Contextual Effects’ in Percept. Psychophys. 56, 227-237 50 Gregson, R.A.M. and McCowen, P.J. (1963) ‘The Relative Perception of Weak
36 Stillman, ].A. (1993) ‘Context Effects in Judging Taste Intensity: A Comparison Sucrose-Citric Acid Mixtures’ in 1. Food Sci. 28, 371-378

How are bitter and


Sweet taste and bitter taste are both apparently mediated by
G-protein-coupled receptors. In this review article, connec-
tions between bitter taste and sweet taste are examined. In
sweet tastes related?
addition, several ways in which sweet taste may be more
effectively used to mask bitter taste are discussed.
D. Eric Walters
Although sweet taste is usually regarded as pleasant and
bitter taste as unpleasant, there is much evidence to sug-
gest that the two tastes are quite closely related with re- vert sweet compounds to bitter ones, and vice versa, as
spect to their transduction mechanisms. In this article, shown in Fig. 1. ‘Sugar’ and ‘sweet’ are practically syn-
I will summarize much of the experimental work pointing onymous, yet many sugar esters and halogenated sugars
towards such a relationship, review a theory that would are intensely bitter. Other sugar derivatives, such as the
account for the experimental results, and discuss some chloro-sugar sucralose, have sweetness potencies that
practical ways in which the proposed relationship may aid are hundreds of times that of sucrose. Numerous classes
us to use sweet taste to mask bitter taste more effectively. of structures (e.g. peptides, oximes, terpenes, ureas,
guanidines and saccharin derivatives) include examples
Evidence linking sweet and bitter tastes that taste sweet, examples that taste bitter, and examples
Structure-taste relationships that taste both bitter and sweet’. Neohesperidin is a bitter
The relationship between chemical structure and taste terpene glycoside that is found in citrus peel; cleavage
provides many clues that sweet and bitter tastes may be of a single carbon-oxygen bond produces the intensely
very closely related. Small changes in structure can con- sweet compound neohesperidin dihydrochalcone.

D. Eric Walters is at the Department of Biological Chemistry, Finch Sensory experiments


University of Health Sciences/The Chicago Medical School, 3333 Green Sensory experiments also point towards a connection
Bay Road, North Chicago, IL 60064, USA (fax: tl-847-578-3240; e-mail: between bitter and sweet tastes. Mixture suppression is
walterse@mis.finchcms.edu). commonly observed: when sweet and bitter compounds

Trends in Food Science &Technology December 1996 [Vol. 71 CopyrIght 81996. Elsewer Science Ltd. All rights reserved. 0924 -2244/96/$15.00 399
PII: SO924.2244196)10040-6
Special Issue on Flavour Perception

You might also like