Chemical Senses and Flavour Volume 5 Number 3 1980

Lipid solubility and the sourness of acids: implications

for models of the acid taste receptor
R.J. Gardner
7 Ham Barn Cottages, Farnham Road, Liss, Hampshire, G33 6LB, UK*
(Received 29 November 1979; revised 5 August 1980; accepted 2 September 1980)

Abstract. Taste threshold concentrations (or equi-sour concentrations) of organic acids are
significantly related to their octanol/water partition coefficients. Since the latter appear to model

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

aqueous/membrane partitioning in biological systems this suggests that absorption of an acid into the
taste cell membrane plays an important role in the perception of the sour taste.

1. Introduction
Models of the sour taste receptor have been reviewed (Price & Desimone, 1977);
the major theoretical models (Beidler, 1967, 1971; Makhlouf & Blum, 1972) de-
pend on interaction of hydrogen ions with surface receptors. However, in little
quoted papers Taylor and co-workers (Taylor, 1928, 1930; Taylor et al, 1930)
have suggested that for acids to produce the sour response they must penetrate
the taste cell membrane (see discussion).
If solution of the acid in a membrane is important in eliciting the sour response
then taste thresholds, or equi-sour concentrations, of different acids should be
related to their membrane/water partition coefficients. The lipophilicity of drugs
in biological systems is well modelled by their octanol/water partition coefficient
(as log P); Hansch, 1971. Many types of drug activity (Hansch, 1971, 1978),
odour potentials and thresholds (Wolkowski et al, 1977; Greenberg, 1979) and
flavour (combined odour/taste) thresholds (Gardner, 1979b) are highly
significantly correlated with log P. I now show that acid thresholds (and equi-
sour concentrations) are correlated with their log P values. These results, with the
observations of Taylor and co-workers (Taylor, 1928, 1930; Taylor et al, 1930)
suggest that for an acid to be tasted as sour it must penetrate the taste cell mem-
brane.
2. Methods
Thresholds
Taste (or flavour) thresholds for organic acids in various media, were obtained
from the sources quoted in the Appendix. For correlation studies thresholds (T)
have been calculated in Moles/litre and expressed in the form log 1/T; this
parallels the format normally used in property-activity studies.
Partition Coefficients
Octanol/water partition coefficients (P), as log P, were obtained from the com-
pilation of Leo et al, (1971), the work of Wolkowski et al, (1977), or estimated
using the approach described by Hansch (1971). Values of log P for the named
acids are given in the Appendix.
© IRL Press Limited, 1 Falconberg Court, London Wl V 5FG, U.K. 185
R.J. Gardner

Correlation Studies
All correlation equations were derived using ADP's 'TSAM' statistics package.
The fit between log 1/T and log P was tested for two equations:
log l/T = a(logP) + b ...(1)
log 1 /T = a(log P) + b(log P)1 + c . . . (2)
The following correlation statistics are given for all quoted relationships of
Tables 1 and 2:
r, or R : Correlation Coefficient
s : Standard error of estimate; calculated as \fL(yc-y+y/N-2 where yc
and y+ are the calculated and measured values of the dependent
variable (in this case log 1/T) and N is the number of degrees of

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

freedom.
3. Results
Tables 1 and 2 give equation coefficients and correlation statistics for linear (1)
and parabolic (2) relationships between log 1/T and log P respectively.
Thresholds, or equi-sour concentrations, of acids are significantly correlated with
log P by linear, or parabolic, equations.
4. Discussion
For a biologically active compound to produce it's physiological effect it must
first reach the appropriate receptor. Once at the site of action the compound must
have the appropriate groups, in the correct stereo-alignment, to interact with this
receptor. In pharmaceutical research this model of drug action has been quan-
tified by the Hansch approach of relating biological activity in series of com-
pounds by multi-parameter regression equations to lipophilic, steric and elec-
tronic parameters (Hansch, 1971, 1978; Ariens, 1971). In studies of chemorecep-
tion, however, it has been largely assumed that taste receptors are situated on the
surface of the cell. (There is strong evidence that this is the case for the sweetness
receptor (Price & Desimone, 1977) but not for other receptors). Hence study has
been concentrated on steric effects and lipophilic effects have until quite recently
been overlooked (Wolkowski etal, 1977; Gardner, 1978, 1979a, b, c; Greenberg
1979).
The log P value (P = octanol/water partition coefficient) appears to be a good
model for aqueous/membrane partitioning (Hansch, 1971) and equations (1) and
(2) have been widely used to correlate drug activity with lipophilicity (Hansch,
1971, 1978). Correlation of activity with log P suggests that transport of the
molecules across membrane barriers is an important factor in determining their
biological effect. The results of Tables 1 and 2 show that taste thresholds and
equi-sour concentrations of acids are significantly correlated with their
lipophilicity, as measured by log P. However, the best form of equation (linear,
or parabolic) cannot be decided from these results, probably because of the
limited number of samples of high log P value. Greenberg (1979) found a similar
problem with relationships between odour potential and log P, but in studies of
flavour (combined odour/taste) thresholds of compounds in beer (Gardner,
1979b) a parabolic relationship is statistically more significant.

186
 LJpld solubility and the sourness of acids

In the equations of Table 2 the percentage variation in log 1/T 'explicable' in

.terms of variation in log P (i.e. R2x 100) ranges from 28% to 92%. It is likely
therefore that other parameters are also of importance. However, the data
available does not lend itself to a complete Hansch (1971) multi-parameter regres-
sion analysis.
Thus steric (e.g. the Taft steric parameter, EJ and electronic (e.g. the acid
dissociation constant, K) parameters are, for the relatively small sets of data con-
sidered here, to some degree co-linear with log P. For example, in the data of
Rothe et al, (1972) Es values were available for six of the acids and these were
linearily related to log P with r2 = 0.64. Similarly log K and log P are linearily
related, for thirty-five of the acids in Appendices 1 and 2, with r2 = 0.51.

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

Correlation of the thresholds of acids with log P supports the contention of
Taylor and co-workers (Taylor, 1928; 1930; Taylor et al, 1930) that for an acid to
taste sour it must penetrate the cell membrane. These workers assumed that only
undissociated acid could penetrate the membrane and, once inside the cell,
dissociation occurs so that the hydrogen ion released can interact with some
receptor. This implies that for equi-sourness (or at the thresholds) the acids must
be producing equal concentrations of hydrogen ions inside the cell. (I would sug-
gest that the receptors for the sour response are in the cell-membrane, rather than
in the interior of any cell, but otherwise the reasoning and equations of Taylor et
al, (1930) remain unchanged). This view immediately rationalizes the observation
that, although an available hydrogen ion appears to be necessary for a compound
to taste sour, concentrations of equi-sour solutions of different acids are not
always related to their pH (Taylor et al, 1930; Kenrick, 1931; Beatty & Cragg,
1935; Beidler, 1967, 1971; Makhlouf & Blum, 1972).
Taylor et al, (1930) noted that the free energy change (AF) governing passage
of an acid from the external aqueous solution into the cell membrane is given by
equation (3):
AF=-RTlog e (HA) o /(HA) 1 ...(3)
R is the gas constant, T the temperature and the parentheses denote activities of
the acid (HA) 'outside' (subscript o) and 'inside' (subscript i) the membrane. The
ratio (HA)o/(HA),, the 'activity gradient', is a. measure of the ease with which the
acid can penetrate the membrane. By assuming that the dissociation constant of a
given acid was the same 'inside' and 'outside' the membrane Taylor et al, (1930)
showed that (HA)o and (HA); could be related to the squares of the hydrogen ion
activities 'outside' (H + ) o and 'inside' (H+), the membrane by equation (4):
(HA)/(HA), = (H+)y(H+)f ...(4)
The relative activity gradients of two acids (represented as HxA and H A) may
thus be written as:
(H»A)o/(H,A), = (H
(HyA)o/(HyA)i ( H ; ;
(Note: subscripts x and y merely differentiate the two acids, they do not indicate
the number of protons). When the hydrogen ion concentrations of the two acids

187
OO
OO

O
3.
Table 1. Equation coefficients and correlation statistics for relationships between thresholds, or equi-sour concentrations of acids (as log 1/T, in moles/litre)
and log P (P = octanol/water partition coefficient) using the linear equation: log 1/T = a(log P) + b
Equation Number
Source of data Modality' Medium2 r (significance)
number of samples
Siek et al, (1969) oil 10 0.50 4.71 0.91(P'<0.001) 0.62
Patton (1964) oil 5 0.81 5.82 0.93(P'<0.05) 0.54
9 Patton (1964) water 5 0.34 3.48 0.89(P'<0.05) 0.34
Norwegian
10 Engan (1974) 18 0.49 3.23 0.90(P'<0.001) 0.49
beer
11 Rotheet al, (1972) water 9 0.57 4.11 0.69(P'<0.05) 0.70
12 Harrison (1970) f Irish beer 11 0.65 3.08 0.74(P'<0.01) 0.76
f
American and
13 Meilgaard (1975) 21 0.45 3.03 0.71(P'<0.001) 0.74
Mexican beer
Taylor et al, (1930) 0.24
14 water 0.11 2.64 0.56 (n.s.)
(Taster B) e
Taylor et al, (1930) 0.14 2.61 0.76(P'<0.05) 0.10
15 (Taster T) e water
d from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

Taylor et al, (1930) 0.09 2.84 0.51(P'<0.05) 0.18

16 water 15
(Taster F) e
Notes: 1. Modality: t = taste; f= flavour (taste/odour combined); e = equi-sour concentrations.
2. Medium: solvent in which tastings were carried out.
 Table 2. Equation coefficients and correlation statistics for relationships between thresholds, or equi-sour concentrations, of acids (as log 1/T, in moles/litre)
and log P (P = octanol/water partition coefficient) using the parabolic equation: log 1/T = a(log P) + b(log P)2 + c
Equation Number
Source of data Modality1 Medium2 r (significance)
number of samples
17 Siek et al, (1969) t oil 10 0.46 -0.01 4.68 0.91(P'<0.001) 0.66
18 Patton (1964) f oil 5 1.45 0.13 6.41 0.94(P'<0.02) 0.62
19 Patton (1964) f water 5 0.77 -0.11 3.32 0.96(P'<0.02) 0.26
Norwegian
10 Engan (1974) f 18 0.50 -0.01 3.24 0.90(P'<0.001) 0.51
beer
21 Rotheet al, (1972) t water 9 0.94 -0.16 4.07 0.73(P'<0.05) 0.72
22 Harrison (1970) f Irish beer 11 0.70 -0.04 3.11 0.74(P'<0.01) 0.80
American and
23 Meilgaard(1975) f 21 0.54 -0.05 3.13 0.73(P'<0.001) 0.74
Mexican beer
Taylor et al, (1930)
14 e water 7 0.13 0.17 2.37 0.76(P'<0.05) 0.21
from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

(Taster B)
Taylor et al, (1930)
15 e water 7 0.14 0.00(2) 2.61 O.76(P'<0.05) 0.11
(Taster T)
Taylor et al, (1930)
16 e water 15 0.10 -0.02 2.86 0.53(P'<0.05) 0.19
(Taster F)

Notes: 1. Modality: t=> taste; f= flavour (taste/odour combined); e = equi-sour concentrations.

2. Medium: solvent in which tastings were carried out.

oo
'•O
R.J. Gardner

inside the membrane are equal (e.g. at equi-sourness or the thresholds) equation
(5) simplifies to:
Gradient for acid x _ (H+)£ ,^\
Gradient for acid y (Hy+)£
Taylor et al, (1930) measured the pH's of a range of acids at equi-sour concen-
trations and, by using equation (6), calculated relative gradients for pairs of
acids. These relative gradients demonstrated that for organic acids the ease with
which they penetrated membrane rose with increasing chain length but fell if
hydrophilic groups (e.g. hydroxyl) were introduced into the structure. These
workers also noted that equi-sour concentrations of acids paralleled both their
water/xylene and water/chloroform partition coefficients.

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

In comparing the approach of Taylor et al, (1930) with the correlations
presented here it should be noted that their 'activity gradients' are the reciprocals
of the membrane/solvent partition coefficients. Hence the observed effects of in-
creasing chain length and introduction of hydrophilic groups on the ease of
penetration of membranes by acids may depend on the lipophilicity of the acid.
However, in terms of Taylor et aPs (1930) view of acid taste, two effects must be
considered - the ease with which the acid penetrates the membrane (measured by
it's lipophilicity) and the degree to which it dissociates once at the receptor site. It
should be borne-in-mind that the threshold at the receptor probably involves only
a minute quantity of material, perhaps orders of magnitude less than the
measured threshold. This means that to reach the threshold concentration at the
receptor only a small amount of acid need penetrate the membrane. Consider
first relationships between log P and thresholds: essentially the greater the ease
with which the acid can penetrate the membrane the lower will be the measured
threshold. Consideration of lipophilic effects also provides a simple explanation
of the fact that the same compound can have different thresholds in different
media. Hence a lipophilic compound will partition from an aqueous medium to a
cell membrane much more readily than from an 'oily' medium. Therefore, if
penetration of a membrane is of importance, to achieve the same concentration
(threshold) at the receptor a taster must be presented with a higher concentration
of the tastant in oil than in water. This effect can be seen (Table 1) for the
thresholds of acids in oil and in water; i.e. as the lipophilicities rise the thresholds
fall in water (equations 9-13), but rise in oil (equations 7 & 8). (Similar reasoning
applies to aroma, but here the first step is partitioning from solvent to air, a fact
utilized by Amoore & Buttery (1978) to calculate, using appropriate partition
coefficients, the aroma thresholds of a series of acids, in a substrate of safflower
oil, from their thresholds in water.)
When comparing equi-sour concentrations, which are larger than the threshold
value, different considerations apply. Equi-sour concentrations are generally
measured against a strong acid (e.g. formic). Although such acids may have low
lipophilicities and do not easily penetrate the cell membrane any which do will
yield a high concentration of hydrogen ions at the receptor. This means that
concentrations of weak acids which are equi-sour with strong acids need to be
disproportionately large in terms of their relative lipophilicities. Hence equi-sour

190
Ljpid solubility and the sourness of acids

concentrations of acids tend to increase with increasing lipophilicity (i.e. log 1/T
v log P has a negative slope, as in equations 14-16 of Table 1).
Crozier (1943) has discussed sourness in terms of the rate of penetration of
tissue by acids, but McBurney (1976) could find no differences in the growth of
sensation among acids that differed in penetration rates by 10:1. In fact he
observes that it 'seems clear that the functions are not limited by any process
before the chorda tympani'. This does not mean, however, that adsorption of the
acid into the membrane is not an essential feature in the production of the sour
response, only that such a process cannot be the rate determining step. As noted
earlier the concentration of tastant required at the receptor to produce a given
response is very small. Hence Taylor et aFs (1930) 'activity gradients', and the

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

partition coefficients considered here may be considered as indicating the relative
ease of membrane penetration by the acids without implying anything about the
rate at which such a process might reach equilibrium. Taylor et al, (1930)
recognised this in their comments on the observation that relative gradients of
acids (compared with acetic acid) decreased with chain length: 'whether these
results indicate that the higher members of the homologous series actually diffuse
faster through the membrane itself than do the lower members, or whether this
sequence points to the relatively greater ease of adsorption of the higher members
by the sensory receptor mechanisms cannot be decided, but the latter would seem
to be the more probable assumption'.
The present results are of interest from the point of view of the proposal
(Koyama & Kurihara, 1972) that sour taste is induced by binding of hydrogen
ions to a phosphate group of phospholipids in the gustatory membrane. In a
wider context, these results, with the earlier considerations on bitterness (Gard-
ner, 1978, 1979c) and olfaction (Wolkowski et al, 1977; Gardner, 1979, 1979a,
1979b; Greenberg, 1979), suggest that transport of the tastant, or odorant, to the
site of action plays an important role in chemoreception.

*This work was performed whilst the author was Group R. & D. Chemist, Harp
Lager Ltd., Manor Park, Alton, Hampshire, U.K.
References
Amoore, J.E. and Buttery, R.G.: 1978, 'Partition coefficients and comparative olfactometry',
Chem. Senses and Flavour 3 ,57.
Ariens, E.J. (Editor): 1971, 'Drug Design, Volume /'Academic Press, New York and London.
Beatty, R.M. and Cragg, L.H.: 1935, 'The sourness of acids'. J. Amer Chem. Soc. 57, 2347.
Beidler, L.M.: 1967, 'Anion influences on taste receptor response', in: Olfaction and Taste II. edited
by T. Hayashi, Pergamon Press, New YOrk, p.509.
Beidler, L.M.: 1971, 'Taste receptor simulation with salts and acids', in Handbook of Sensory Phy-
siology. IV. Chemical Sences. 2. Taste edited by L.M. Beidler, Springer-Verlag, New York, p.200.
Crozier, W.J.: 1934, 'Chemoreception', in: Handbook of General Experimental Psychology edited
by C. Murchison, Clark University Press, Worcester, Mass., p. 987.
Engan, S.: 1974, 'Organoleptic thresholds of some organic acids in beer', J. Inst. Brew. 80, 162.
Gardner, R.J.: 1978, 'Lipophilicity and bitter taste', J. Pharm. Pharmacol. 30, 531.
Gardner, R.J.: 1979a, 'Application of property-activity relationships and structure-activity relation-
ships to flavour research. Part I: Introduction and Background'. Tech. Quart. M.B.A.A. 16, 106.

191
R.J. Gardner

Gardner, R.J.: 1979b, 'Application of property-activity and structure-activity relationships to

flavour research. Part II: correlation of calculated octanol/water partition coefficients with the
flavour thresholds of compounds in beer', Tech. Quart. M.B.A.A., 16, 148.
Gardner, R.J.: 1979c, 'Lipophilicity and the perception of bitterness', Chem. Senses and Flavour
4(4), 275.
Greenberg, M.J.: 1979, 'Dependance of odor intensity on the hydrophobic properties of molecules.
A quantitative structure odor intensity relationship', J. Agric. Food Chem. 27, 347
Hansch, C : 1971, 'Quantitative structure-activity relationships in drug design' in: Drug Design
Volume 1 edited by E.J. Ariens, Academic Press, New York and London, p.271.
Hansch, C : 1978, 'Recent advances in Biochemical QSAR', in: Correlation Analysis in Chemistry-
Recent Advances' edited by N.B. Chapman and J. Shorter, Plenum Press, New York and London,
Chapter 9, p. 397.
Harrison, G.A.F.: 1970, 'The flavour of beer-a review', /. Inst. Brew. 76, 486.
Koyama, N. and Kurihara, K.: 1972, 'Receptor site for sour stimuli', Nature 239, 459.

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

Kenrick, F.B.: 1931, 'The sour taste of acids', Trans. Roy. Soc. Can. 25, 227.
Leo, A. Hansch, C. and Elkins, D.: 1971, Partition coefficients and their uses', Chem. Rev. 71, 525.
Makhlouf, G.M. and Blum, A.L.: 1972, 'Kinetics of the taste response to chemical stimulation: A
theory of acid taste in man', Gastroenterol. 63, 67.
McBurney, D.H.: 1976, 'Temporal properties of the human taste system', Sensory Processes 1, 150.
Meilgaard, M.: 1975, 'Flavor chemistry of beer Part II: flavor thresholds of 239 aroma volatiles',
Tech. Quart. M.B.A.A. 12, 151.
Patton, S.: 1964, 'Flavor thresholds of volatile fatty acids', J. Food Sci. 29, 679.
Price, S. and Desimone, J.A.: 1977, 'Models of taste receptor cell stimulation', Chem. Senses and
Flavour 2, 427.
Rothe, M., Wolm, G., Tunger, L. and Siebert, H.J.: 1972, 'Schwellenkonzentrationen von Aroma-
stoffen und ihre Nutzung zur Auswertung von Aromaanalysen', Nahrung. 16, 483.
Siek, T.J., Albin, I.A., Sather, L.A. and Lindsay, R.C.: 1969, 'Taste thresholds of butter volatiles
in deodorized butteroil medium', J. Food Sci. 34, 265.
Taylor, N.W.: 1928, 'Acid penetration into living tissues', J. Gen. Physiol. 11, 207.
Taylor, N.W.: 1930, 'The nature of the nerve receptor for the acid taste as indicated by the absorp-
tion of organic acids by fats and proteins', Protoplasma 9, 98.
Taylor, N.W., Farthing, F.R. and Berman, R.: 1930, 'Quantitative measurements on the acid taste
and their bearing on the nature of the nerve receptor', Protoplasma 10, 84.
Wolkowski, Z.W., Moccatti, D., Heymans, F. and Godfroid, J.J.: 1977, 'A quantitative structure-
activity approach to chemoreception: importance of lipophilic properties', J. Theor. Biol. 66,
181.

192
Appendix 1 Thresholds (as log 1/T, in Moles/litre) and other data for acids in different media.

log 1/T (Moles/litre)

Acid logP 1
Harrison Engan Meilgaard2 Rothe et al Patton Patton Siek et al
(1972); water (1964); (1964); (1969); oil D
(1970); (1974); (1975);
beer beer beer water oil
Acetic -0.17 2.48 2.48 2.54 3.08 3.05 3.93
Propionic 0.33 2.87 2.87 2.69 4.39
Butyric 0.79 4.94 4.64 4.60 4.47 4.11 5.17 5.12
Iso-butyric 0.85 2.64 3.25 3.47 4.67
Valeric 1.20 3.46 4.31 4.11 5.31
lso-valeric 1.32 4.11 4.61 4.83
Hexanoic 1.88 4.37 4.37 4.16 4.37 4.33 4.67 4.67
3-Hexenoic 1.58 (e) 4.94
Octanoic 3.00 4.56 4.46 4.01 5.50 4.40 2.61 2.61
Nonanoic 3.50 4.50
2-t-Nonenoic 3.20 (e) 3.96
Decanoic 4.09 4.93 4.24 4.69 2.93 2.93
Undecanoic 4.14 (e) 5.27
Dodecanoic 4.20 5.60 2.46 2.46
Tetradecanoic 5.20 (e) 1.66
Hexadecanoic 6.20 (e) 1.41
from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

Octadecanoic 7.20 (e) 1.28

Oleic 7.00 (e) 1.55
Lactic -0.62 2.35 2.35 2.35 3.95
Pyruvic - 1 3 8 (c) 2.55 2.47
Citric -1.72 2.74
Tartaric -1.84(e) 2.70 2.40
Malic -1.26 2.58
Malonic -0.70 (e) 2.62
Succinic -0.59 2.77
Phenylacetic 1.41 5.13 4.74 6.13
Glyoxylic -0.99 (e) 1.57
Oxalacetic -2.05 (e) 2.42
Fumaric 0.27 (e) 2.46
Vanillic 1.16 (e) 3.32
Gallic -0.16 (e) 2.67
p-Coumaric 1.49 (e) 2.50
R.J.Gardner

Appendix 2 Equi-sour concentrations and other data for organic acids in water (Taylor et al, 1930).
Acid log 1/T logP 1 Acid log 1/T logP 1
Observer F Observer B
Adipic 2.64 0.08 Azelaic 2.70 1.57
Benzoic 2.55 1.87 Oxalic 3.00 -1.34
Citric 3.01 -1.72 Phenylacetic 2.32 1.41
Chloracetic 2.77 0.07 (e) p-Hydroxybenzoic 2.70 1.58
Glycollic 2.85 -1.11 Propionic 2.48 0.33
L-Malic 2.89 -1.26 Pyruvic 2.70 -1.38 (e)
Maleic 3.05 -0.58 (e) Trimethylacetic 2.24 1.08 (e)
Malonic 2.89 -0.70 (e) Observer T
Mandelic 2.42 0.60 (e) Acetic 2.55 -0.17

Downloaded from http://chemse.oxfordjournals.org/ at Emory University on August 25, 2015

Mesaconic 3.00 -0.24 (e) Butyric 2.46 0.79
Phthalic 3.19 0.37 (e) Formic 2.75 -0.54
Sebacic 2.72 1.52 (e) Lactic 2.55 -0.62
Suberic 2.84 0.71 (e) Succinic 2.80 -0.59
Tanronic 3.02 -2.20(e) Iso-Valeric 2.43 1.32
Tricarbyllic 2.94 0.93 (e) Glutaric 2.65 0.16 (e)

Notes to Appendix:
1. (e) indicates that the value is estimated; the others are measured values (Leo et al, 1971; Wolkowski
et al, 1977).
2. These are combined results for determinations in American or Mexican beer, thresholds in the two
beers are similar (Meilgaard, 1975). Some isomers (e.g., D- and L-Tartaric acids) have been
deleted from consideration, because of uncertainty in log P.

194