Wine Control Praxis in Germany

CARSTEN FAUHL-HASSEK

Federal Institute for Risk Assessment

Head of the Senior Expert Office for the Import Control of Wine
Thielallee 88-92
D-14195 Berlin
GERMANY
E-Mail: carsten.fauhl-hassek@bfr.bund.de

Information has been prepared on the actual state of knowledge with great care. BfR in no
way guarantees that the information made available on this report is complete, accurate or
up-to-date in all cases. The reflection of the legal situation is not approved by any legal
service and therefore does not provide legally binding character.
Bundesinstitut für Risikobewertung

Table of Contents

1. Introduction.................................................................................................................................... 3

2. Major aims in Wine Control .......................................................................................................... 3

2.1 Control of wine quality/fair merchantable quality: Sensory, “off-flavours” ............................... 3
2.2 Control of statements on the label: Alcohol content, quality, grape variety (blending),
geographical origin, vintage................................................................................................................. 3
2.3 Chemical adulterations: illegal acidification, addition of water, glycerol, alcohol, illegal sugar
addition, dyes, sweeteners, preservatives, flavours............................................................................ 4

3. Samples .......................................................................................................................................... 4

4. Legal background of objections attributed to the wine composition ...................................... 4

5. Wine Analysis ................................................................................................................................ 5

5.1 Accreditation - a necessity....................................................................................................... 6

6. Investigation intensity................................................................................................................... 6

7. Control limits.................................................................................................................................. 7
7.1 Regulative (EU) ....................................................................................................................... 8
7.2 Regulative (national)................................................................................................................ 8
7.3 Recommendations (O.I.V.) ...................................................................................................... 9
7.4 Experience limits ..................................................................................................................... 9

8. Appreciation................................................................................................................................. 10

9. Measurement uncertainty ........................................................................................................... 10

10. Use of databanks ..................................................................................................................... 11

10.1 General considerations for the use of databanks.................................................................. 11
10.2 Stable Isotopes Analysis ....................................................................................................... 12
18
10.3 δ O Databank (German wine control) .................................................................................. 14
10.4 Shikimic acid (SA) ................................................................................................................. 16
10.5 Anthocyan composition ......................................................................................................... 19

11. Special analysis ....................................................................................................................... 21

11.1 Glycerol.................................................................................................................................. 21

12. Pyrazine .................................................................................................................................... 23

13. Recommendations .................................................................................................................. 24

14. References ............................................................................................................................... 25

Page 2/25
Bundesinstitut für Risikobewertung

1. Introduction
The objective of this report is to highlight the situation and practice of official wine control in
Germany and to provide recommendations on best practice to FERA (formerly CSL).
Particular emphasis is given on the actual use of databanks assisting authentication of wine
including the proof of origin. However, state of the art of the official control practice is
characterised by its possibly forensic utilisation. Due to its high demand for reliability an
official objection - standing “beyond reasonable doubt” – the actual practise does not reflect
totally state of the art research particular as it concerns the proof of origin. Therefore
conservative data assessment and evaluation is done in practice.
German wine control has a long tradition in authentication of wine. In addition to national
products, also European wines, but as well third country wines are tested for their conformity
with actual regulations. BfR is the senior expert office for the import control of third country
wines and provides therefore long experience in the analysis and assessment of wine. In its
function BfR manages different databanks for control purposes in mission of the national
official wine bodies.
The report has its focus on the analytical aspects relevant in wine appreciation rather than on
the more formal aspects of wine control of such as label design or font size for example. The
report does not cover the administrative structure and organisation of the wine control in
Germany which is due to the federal system complex and different for each federal country.

2. Major aims in Wine Control

2.1 Control of wine quality/fair merchantable quality: Sensory, “off-flavours”

The first and major point to be judged is whether a wine complies with the expected standard
or not. This can be already a very crucial point because the assessment is mainly based on
the sensory evaluation of the wine. The judgement whether a wine defective is still
acceptable or not is not often straight forward. Analyses of fusel oils, certain flavours, and the
volatile acidity or other indicators of spoilage may support the sensory perception.

2.2 Control of statements on the label: Alcohol content, quality, grape variety (blending),
geographical origin, vintage...
Labelling on a wine bottle is the most important source of information for the consumer and
also for the analyst. All statements on the label - representing the wine identity - should be
verifiable by analytical data, however some of them are very challenging. Already the
“simple” parameter alcoholic grade can cause problems; particular the application of
appropriate methods, the reference methods in official control, shall be assured in official
control. In addition the provisions on the tolerances of the alcoholic grade statement and the
additional consideration of the analytical precision (measurement uncertainty) must be
respected appropriately.
If a grape variety is stated on the label its verification - apart from the analytical difficulties -
must consider the appropriate legislation. For European wines the so called 85 % rule will
apply but for third country wines different conditions, documented in Trade contracts may
apply. The EU commission has incorporated particular contracts with Chile, South Africa, the
United States of America, Canada and Australia.

The most challenging part of authentication here is the proof of the geographical origin.
Actually what is proofed at state of the art wine control in Germany is whether the wine in
question does fit in the chemical profile established for the origin stated or not. It is not
reasonably possible to determine its correct origin. Concerning the objection it might be also

Page 3/25
Bundesinstitut für Risikobewertung

the case that the wine in question does not comply with the authenticity range established
because it was adulterated with other wine constituents.

2.3 Chemical adulterations: illegal acidification, addition of water, glycerol, alcohol, illegal
sugar addition, dyes, sweeteners, preservatives, flavours.....
The chemical adulterations in question below this point are commonly non authorised
oenological practices and therefore the determination of them leads to an objection. The
proof of chemical adulterations depends on the possibilities of the control facility.

3. Samples
Samples to be controlled can be divided into four categories:
• National products
• EU Wines
• Third country wines
• Custom samples (import process)

4. Legal background of objections attributed to the wine composition

The old wine law offered directly the objection of wines by article 45 (EC) 1493/1999, the
following products may not be offered or disposed of for direct human consumption:
1. which have undergone oenological practices not authorised by Community rules (e.g.
addition of water, glycerol)
2. which are not of sound and fair merchantable quality; (e.g. sensory)
3. which do not comply with the definitions shown in Annex I. (e.g. semi-sparkling wine
CO2 > 2.5 bar, use of “hybrid” grape varieties)

The wine legislation was changed fundamentally with certain transition periods but for wines
marketed after the 1st August 2009 the new legislation is valid.

To 1.
Article 45 of regulation (EC) No 1493/1999 was partially transformed in the Council
regulation (EC) No 479/2008 on the common organisation of the market in wine, by which
article 27.4 excluded the application of non authorised oenological practices. This regulation
never really came into force but it was implemented with regulation (EC) No 491/2009 into
regulation (EC) No 1234/2007 establishing a common organisation of agricultural markets
and on specific provisions for certain agricultural products (Single CMO Regulation):

"Article 120c (Regulation (EC) No 1234/2007)

4. Products covered by this Regulation, which have undergone unauthorised Community oenological
practices or, where applicable, unauthorised national oenological practices or which contravene the
restrictions laid down in Annex XVb, shall not be marketed in the Community."

Although it must be noted that the number of authorised oenological practices was extended
because the new EU wine legislation is strongly linked to the O.I.V. rules (International
Organisation for Vine and Wine: Code of oenological practices and Oenological Codex)1.
Commission Regulation (EC) No 606/2009 lays down certain detailed rules for implementing
Council Regulation (EC) No 479/2008 as regards the categories of grapevine products,
oenological practices and the applicable restrictions.

1
www.oiv.int

Page 4/25
Bundesinstitut für Risikobewertung

To 2.
With enforcement of Regulation (EEC) No 479/2008, respectively with the amendment of
Council Regulation (EC) No 1234/2007, in frame of the wine market reform particular the
application of the paragraph (article 45.3 of regulation (EC) 1493/1999) related to the sound
and merchantable quality was removed. Official German wine control will refer - at least as it
concerns sensory defects - to article 14.2b in combination with 14.5 of Regulation (EC) No
178/2002, which does not exclude wine.

"Article 14 Food safety requirements (Regulation (EC) No. 178/2002)

1. Food shall not be placed on the market if it is unsafe.
2. Food shall be deemed to be unsafe if it is considered to
be:
(a) injurious to health;
(b) unfit for human consumption…..

5. In determining whether any food is unfit for human consumption, regard shall be had to whether the
food is unacceptable for human consumption according to its intended use, for reasons of contamination,
whether by extraneous matter or otherwise, or through putrefaction, deterioration or decay. "

In addition specific national law („Weingesetz“) was adapted in terms that products of the
wine law generally must be of sound and merchantable quality (§16.1).

To 3.
The marketing of products which do not comply with definitions is now excluded by article
113d of Council Regulation (EC) No 1234/2007:

"Article 113d (Regulation (EC) No 1234/2007)

Specific provisions for the marketing of wine
1. A designation for a category of a grapevine product as provided for in Annex XIb may be used in the
Community only for the marketing of a product which conforms to the corresponding conditions laid
down in that Annex."

For completeness: In the “intermediate” Council Regulation (EC) No 479/2008 on the

common organisation of the market in wine, article 25.2 (Production and marketing) was
equivalent.

5. Wine Analysis
In the European Union the analysis of wines is a special case because methods of analysis
were laid down in ANNEX of Regulation (EEC) No 2676/1990 (EU, 1990). The determination
of 45 analytical parameters was described and for some of them more than one method was
stated, the usual and the reference method. The results obtained by use of the reference
methods shall prevail particular where dispute arises. For almost all reference methods
precision parameters were determined in collaborative trials.
However this regulation is not valid any longer since the 1st of August 2009. The EU
legislation in the Council Regulation (EC) No 1234/2007 refers now directly to the O.I.V:

"“Article 120g (Regulation (EC) No 1234/2007)

Methods of analysis
The methods of analysis for determining the composition of the products of the wine sector and the rules
whereby it may be established whether these products have undergone processes contrary to the
authorised oenological practices shall be those recommended and published by the OIV. Where there
are no methods and rules recommended and published by the OIV, corresponding methods and rules
shall be adopted by the Commission in accordance with the procedure referred to in Article 195(4).

Page 5/25
Bundesinstitut für Risikobewertung

Pending the adoption of such rules, the methods and rules to be used shall be the ones allowed by the
Member State concerned."

The O.I.V. provides a comprehensive compilation of internationally accepted methods, the

Compendium of International Methods of Analysis, for many parameters.

5.1 Accreditation - a necessity

According to Regulation (EC) No 882/2004 accreditation will be required for official control
laboratories in food and feed analysis from the 1st January 2006 on. The European Norm
which sets the standards and requirements for accreditation for testing and calibration
laboratories is the EN ISO/IEC 17025 (ISO, 2005). Although the situation for wine is not
completely clarified, because agricultural products are somehow excepted it is evident that
the requirements of accreditation set by EN ISO/IEC 17025 (ISO, 2005) are fundamental in
wine control as well. All official wine control laboratories belong to the official food control and
are accredited in Germany.

6. Investigation intensity
The typical analysis of wine includes the sensory evaluation and the so called identity test.
The identify proof usually includes the analysis of:

• Actual alcoholic grade

• Total alcoholic grade
• Total dry extract
• Total acidity
• Volatile acidity
• Citric acid
• SO2

Further to this fundamental proof, which allows the identification in case of imported wines by
comparison with data in the V I 1 document according to Commission regulation (EC) No
555/2008) Article 41, 43 and ANNEX IX, other parameters might be tested.

"Article 41 (Regulation (EC) No 555/2008)

Contents of the analysis report
The analysis report shall include the following information:
(a) in the case of wines and grape must in fermentation:
(i) the total alcoholic strength by volume;
(ii) the actual alcoholic strength by volume;
(b) in the case of grape must and grape juice, the density;
(c) in the case of wines, grape must and grape juice:
(i) the total dry extract;
(ii) the total acidity;
(iii) the volatile acid content;
(iv) the citric acid content;
(v) the total sulphur dioxide content;….."

Page 6/25
Bundesinstitut für Risikobewertung

The list of parameters depends on the availability in the individual control laboratory and of
the thereof extracted investigation intensity. The specification of the analytical depth is done
by the expert.

• Preservatives (sorbic acid, benzoic acid)

• Dyes
• Mineral composition: Na, K, Ca, Mg, Ash
• Anions: Cl-, SO42-, PO43-, NO32-
• Glucose/fructose
• Acids: Tataric acid, malic acid, lactic acid, glucoronic acid
• Glycerol, butandiol, higher alcohols
• Glycerol by-products
• Anthocyan pattern
• Shikimic acid

• Biogenic amins (often tested in case of consumer complaint, as well as in focussed

surveys)
• Heavy metals, pesticides, mycotoxins (ochratoxin) parameters tested in focussed
survey actions, normally not in routine.

• Specialized laboratories:
o Stable isotope ratios (D/H, 13C/12C, 18O/16O)
o Flavour analysis (e.g. pyrazines in case Sauvignon Blanc)
o Enantioselective Analysis (e.g. enatiomeric raito analysis by chiral
chromatography of lactones, peach flavour)

At this point it should be noted that the always requested high number of control samples is
realised on cost of the investigation intensity. Therefore a compromise between number of
samples and the investigation intensity is practised, including the targeted control of wines
from origins which are especially suspicious at the current state of information, because
indications of adulterations or non authorised practises were “linked” previously to the
particular type of samples.

7. Control limits
In the case of wine analysis parameters limits are fixed in EU regulations, other limits are
defined by the O.I.V. and for others limits have been established on a national level.
For several relevant analytical parameters no limits are laid down by law or established by
international organisations in wine appreciation. This situation occurs often in the case of
authentication with stable isotope analysis but also in several other cases such as mineral
content analysis and evaluation. For these parameters the experience of the analyst and
availability of comparison or reference values is important, because the effectiveness of
control depends on it.

Page 7/25
Bundesinstitut für Risikobewertung

7.1 Regulative (EU)

Table 1: Regulatory limits in EU provisions (list does not claim being complete).

Parameter Limit Annex Regulation

Sorbic Acid 200 mg/l IA No 606/2009
SO2 150-400 mg/l IB No 606/2009
Volatile Acid 18-20 mval/l IC No 606/2009

Citric Acid 1.0 g/l IA No 606/2009

Ascorbic Acid 200 mg/l IA No 606/2009
CO2 3 g/l IA No 606/2009

7.2 Regulative (national)

2
Table 2: Exemplary regulatory limits in German provisions

Parameter Limit
Al 8 mg/l
As 0,10 mg/l
B 80 mg/l
Br 1 mg/l
F 1 - 3 g/l
Cd 0.01 mg/l
Cu 2 mg/l
Zn 5 mg/l
Sn 1 mg/l
Trichloromethane 0,1 mg/l
Trichloroethene 0,1 mg/l
Tetrachloroethene 0,1 mg/l

2 th
Appendix 7, (to § 13 .1.2), “Weinverordnung” from 21 April 2009, in combination to § 13.3.1 and 13.3.3 of the
“Weingesetz”

Page 8/25
Bundesinstitut für Risikobewertung

7.3 Recommendations (O.I.V.)

3
Table 3: Recommendations of the O.I.V. (list does not claim being complete) .

Parameter Content Comment

Citric acid 1 g/l
Volatile acidity 20 milliequivalents/l
Arsenic 0.2 mg/l
Boron 80 mg/l
Total bromine 1 mg/l
Cadmium 0.01 mg/l
Copper 1 mg/l
Diethylene glycol 10 mg/l
Ethylene glycol 10 mg/l
Fluoride 1-3 mg/l
Methanol 250-400 mg/l
Ochratoxin A 2 µg/l (from vintage 2005)
Lead 0.15 mg/l (from vintage 2007)
Excess Sodium 80 mg/l
Propylene glycol 150-300 mg/l
Sulfate 1-2.5 g/l
Zinc 5 mg/l

7.4 Experience limits

Table 4: Experience limits for typical parameters

Parameter Limit
Glycerol 4-18 g/l
Glycerol factor (ethanol/glycerol) Ø 8 (6-10)
Glycerol/Butandiole Ø13 (<20)
Extract (selection) >13 g/l

Chloride 20-90 mg/l

Sodium 5-50 mg/l
Potassium 35-40 % of ash
Magnesium 75-115 mg/l
Calcium 35-140 mg/l
Phosphate 150-400 mg/l
Nitrate 5-15 mg/l

Some examples of “suspicious” composition (generally accepted without explicit databank

consultation):

• High or low glycerol factor: Addition of glycerol, ethanol or water

• Magnesium < 50 mg/l: Indication of water addition
• Excess Sodium > 80 mg/l: Deacidfication with NaOH, or application of ion exchange

3
www.oiv.int (Compendium of International Methods of Analysis of Wines and Musts (2 vol) (MA-E-C1-01-
LIMMAX))

Page 9/25
Bundesinstitut für Risikobewertung

• Nitrate > 30 mg/l: Indication of water addition (well water with high nitrate content)
• Low Ca content: application of non authorised procedures

8. Appreciation
The comparison of an actual measurement value with a control limit, that can be in the case
of wine a regulatory limit at European level, national level, a recommendation by the O.I.V. or
a so called experience value for certain parameters is one major piece of the control process.

Of course the overall process of wine appreciation is more complex because many
parameters of the analysed product contribute to the total impression of an expert, but this
report will focus to the “simple” single parameter proof only. This part of the appreciation
always relies back on whether the analytical parameter determined, considering also the
measurement uncertainty, is violating the established limit or if it is still in compliance with it.
More sophisticated appreciation goes beyond the scope of the report and remains generally
a single case appreciation in the expert’s opinion.

9. Measurement uncertainty
According to the norm EN ISO/IEC 17025 the test result of an analytical measurement must
be stated with an estimate of its uncertainty in particular when it is relevant to the application
of the test results or when uncertainty affects compliance to a specification limit. For most of
the parameters which are relevant for the wine appreciation measurement uncertainty shall
therefore be stated respectively it must be considered for drawing conclusions.
From a statistical point of view five different situations, must be considered when a test result
was determined in a laboratory for wine appreciation and control.

Situation a) result plus uncertainty below limit

b) result below limit but limit within uncertainty
c) result above limit but limit within uncertainty
d) borderline situation (decision limit)
e) result less uncertainty above limit

Situations a and b do not cause any consequences since the results are in compliance with
the control limit. Already situation c may lead to different interpretations, when uncertainty is
taken into account the result is compliant with the limit, but when measurement uncertainty is
completely ignored by the analyst, the result is interpreted to be non-compliant. In this case
the terminus “beyond reasonable doubt” - that should be demonstrated in case of official
objection - does not fit perfectly and also conventions of accreditation are mistreated. On the
other hand situation e clearly presents a violation.

The particular situation d is of utmost importance in laboratory praxis because it defines if a

measurement value violates a limit truly or if it is still in compliance with the control limit and
depending on the decision further action (e.g. official objection) may become necessary.

It must be underlined that the consideration of measurement uncertainty in data

interpretation is an essential point and a demand in accredited laboratories.

Page 10/25
Bundesinstitut für Risikobewertung

Different methods for using - and previously even the estimation of - measurement
uncertainty are available for the calculation of decision limits, unless it has been
demonstrated that the final differences were relatively low (Fauhl 2006). It must been pointed
out that the appropriate consideration of measurement uncertainty is mandatory but the
exact proceeding is individually left to the expert in wine analysis and appreciation. Although
wine analysis is highly unified at an analytical level further harmonisation is needed within
data interpretation process.

10. Use of databanks

In most of the cases the use of stable isotope ratios for authentication requires the specified
comparison with reference data. For this purpose data bases are of tremendous importance
for data interpretation. Authenticity and representativeness of the reference samples are
crucial points that arise if any jurisdiction becomes involved.

10.1 General considerations for the use of databanks

In some cases it is necessary to “pre-select” data from a databank in order to define the
appropriate reference data collective. Pre-selection “sharpens” usually the authenticity range.
Stable isotopes ratios for example depend on the region and vintage considered and
therefore, when enough data are available, such a reduction is justified. The selection of
reference collectives from data bases is a very crucial and delicate matter and should
therefore left to experts only. Representativeness and comparability are only two points of
importance.

Apart from the process of selecting relevant reference data - that is not discussed in this
eport - the process after selection of the reference data set is highlighted from a statistical
point of view. After the selection of comparison samples these values are evaluated regularly
by calculating the arithmetic mean, median, standard deviation and their confidence limits.
For these operations ideally at least 30 reference samples should be available what is not
always the case of course. Calculation of the confidence limits by consideration of the
Student-Factor (t-distribution) for a low number of reference samples was described for
authenticity proof in wine analysis previously (Christoph et al. 2003; Otteneder et al. 2004)
and is accepted to be valuable for data interpretation. The t-distribution for a two-tailed
question delivers an upper or a lower limit and defines the range which is used for the
comparison with the actual value of the wine in suspect. If the actual measurement value
falls in the range then it fits statistically into the reference collective with a given probability.

Equation 1. Upper and lower confidence limit C (with X = arithmetic mean of reference samples, S =
standard deviation and t = Student-Factor for n = values and certain probability).

CUpper − Limit = X + ( S × t )
C Lower − Limit = X − ( S × t )

On the other hand naturally distributed variables like e.g. mineral content in wine would in
principle follow a two-tailed distribution, because the variation can follow two directions. From
a statistical point of view the application of one-tailed distributions requires the theoretical
exclusion of one of the two directions. The question of the hypothesis to be answered defines
the appropriate use of one-tailed or two-tailed distributions. For isotopic parameters the
consideration of one- or two-tailed distributions depends on the suspect and information
Page 11/25
Bundesinstitut für Risikobewertung

about the wine. If the addition of sugar is the clear assumption the proof of chaptalisation by
the use of the D/H ratios and the δ13C value of wine ethanol would allow the application of a
one-sided distribution. In case that water addition is the clear suspect in the expert opinion
for the explanation of a low δ18O value found - supported by other indications drawn from
further parameters - one-sided distribution probability limits are in principle applicable.
Although the sharpness of the evaluation is better if one-tailed distribution is applied from a
control point of view, reasonable exclusion of one direction of distribution remains and its
justification is left to the expert. Application of two-tailed distribution probabilities seems to be
the most appropriate if no specific information is available, because this surely covers all
doubts.
What remains to be selected by the expert for each individual case is what kind of distribution
– one- or two-tailed – is appropriate for interpretation. If a one-tailed distribution is considered
only one limit – a lower or upper - is the result for the confidence limit calculation and
consequently only exceeding or falling below is decisive. Christoph et al. (2003) depicted the
data handling and interpretation with an example for stable isotope analysis.

It is sometimes discussed that the analytical variation must also be considered additionally if
a set of reference samples was used for the calculation of the confidence intervals. However
Martin et al. (1996) have described concretely that the analytical variation can be neglected
when the natural variation is considered because this type of variation is already and
therefore sufficiently included in the natural one. Therefore the limits obtained from such a
calculation are comparable with regulatory limits from a statistical point of view.

10.2 Stable Isotopes Analysis

Due to the high validity of the data and also to the guaranteed authenticity of the samples
included in the official EU-Wine data-bank according to regulation (EC) No 555/2008, its
meaning is very valuable particular in court cases. According to our knowledge at the
beginning stage of the official databank several court cases took place in some countries in
which always the expertise of control laboratories was confirmed. Exchange of expertise and
data within the official German as well as other control laboratories is a proven practice that
contributes to a reliable data background for comparison purposes during the appreciation of
wines.

In case of the Deuterium (2H) NMR method for the determination of chaptalisation no
precision data is stated neither in Method 8 in the ANNEX of Regulation (EEC) No
2676/1990 nor in the now relevant comparable method description in the O.I.V. compendium:
MA-E-AS311-05-ENRRMN “Detecting enrichment of musts, concentrated grape musts,
rectified concentrated grape musts and wine by ²H-RMN”. However, the method has been
validated comprehensively for fruit juice, AOAC 995.17 Method (AOAC, 2000), in which the
standard deviation of repeatability (Sr) 0.25 ppm and 0.37 ppm for standard deviation of
reproducibility (SR) were obtained for the worst sample (r= 0.69 and R=1.02) that is usually
taken for the estimation of the measurement uncertainty of the NMR method.
Christoph et al. (2003) discussed in detail one wine sample coming from Franconia, that was
found to be suspicious in terms of chaptalisation with beet sugar showing a very low (D/H)I-
value of 98.4 ppm. For this sample minimum (D/H)I-values were calculated taking into
account different reference samples collectives.

The official method for the measurement of the δ18O-value of wine water was laid with
Method 43 in the ANNEX of Regulation (EEC) No 2676/1990. The relevant method, actually
valid for official wine control, of the O.I.V. was modified with resolution (OIV OENO
353/2009) recently, implementing modern aspects of analysis (“Gasbench” equilibrium

Page 12/25
Bundesinstitut für Risikobewertung

technique). The precision parameters r and R are given with 0.24 ‰ and 0.5 ‰ δ18O
respectively. Limits for δ18O-value are not laid down in any provision.
Watering of wine with tap water is detectable using the δ18O value. In that case the wine,
enriched in 18O and therefore characterized by almost always positive δ18O values, is diluted
with local tap or ground water, which always shows a negative δ18O value (Fauhl 2006). The
final 18O concentration in the wine water corresponds to the grade of watering. Thus,
depending on the information available about the wine concerned a reliable conclusion can
be drawn as to the watering of a wine. The comparison with authentic wines of the same
origin with similar climatic conditions is necessary.

The determination of the 13C/12C ratio of wine ethanol is currently described in the OIV
compendium with method MA-E-AS312-06-ETHANO “Determination of isotopic ratio of
ethanol (oeno 17/2001)”. Based on the differences during the CO2 assimilation of C3 and C4
plants the 13C-content of sugars is enriched in C4 plants. All vine species are C3 plants and
therefore any adulteration of the organic ingredients of wine with C4 plant products will lead
to an enrichment of the 13C content of, for example, the fermentation alcohol or sugar. The
addition of cane sugar is easily detectable by 13C-IRMS of the fermentation ethanol, whereas
the addition of beet sugar is not visible in this way. In addition the δ13C value may give certain
indications on the origin, in dry and hot regions the δ13C value is often more positive together
with very positive δ18O value.

In some instances, when so-called “cut-off” values are violated by stable isotope parameters
the consultation of databanks is not necessary (Christoph et al. 2003). δ13C-values higher
than -23 ‰ and (D/H)I- ratios exceeding a value of 106 ppm may indicate a chaptalisation
with C4-sugar, (D/H)I -ratios less than 98 ppm are usually caused by chaptalisation with beet
sugar. However officially accepted “cut-off” values, such as ones established by the
Association of the Industry of Juices and Nectars from Fruits and Vegetables (AIJN) in case
of fruit juice analyses do not exist in the case of wine products.

Therefore the consultation of the EU wine databank according to regulation (EC) No

555/2008 usually becomes necessary when a wine originating in the European Union is
assessed. There is a manual on the handling of data in the EU database (Annex 1) to assess
a wine on the basis of isotope parameters. This recommendation should be taken into
account. Having as much information as possible about the wine to be assessed is of
fundamental importance in the process. The responsible and correct handling of the
database rests exclusively with the expert and always requires an examination of the
individual case coupled with specific knowledge of the respective attendant circumstances.

According to the regulation (EC) No 555/2008 the data of the EU-Wine data-bank are
handled restrictively. The communication of data from the EU-Wine data-bank shall be
compliant with the regulation, in which the access is explained in detail:

"Article 85: Isotopic analysis of wine-sector products and interpretation of the results are delicate
procedures and, in order to permit uniform interpretation of such analysis results, the JRC databank
should be made accessible to official laboratories using that analytical method and, on request, to other
official bodies in the Member States while respecting the principles of the protection of private data.

Article 88 …
7. Member States and the JRC shall ensure that:
(a) data in the analytical databank are preserved;
…..
(c) the databank is used only for monitoring the application of Community and national wine legislation or
for statistical or scientific purposes;
(d) measures are applied to safeguard the data, in particular against theft and interference;

Page 13/25
Bundesinstitut für Risikobewertung

....
Article 90
Communication of results
1. The information contained in the databank shall be made available on request to the laboratories
designated by the Member States for that purpose.
2. The JRC shall draw up and update on a yearly basis the list of the Member States laboratories
designated for the preparation of samples and the measurements for the analytical databank.
3. In duly substantiated cases, the information referred to in paragraph 1, when representative, may be
made available on request to other official bodies in the Member States.
4. Communication of information shall relate only to the relevant analytical data required to interpret an
analysis carried out on a sample of comparable characteristics and origin. Any communication of
information shall be accompanied by a reminder of the minimum requirements for the use of the
databank."

Among the experts of the EU-Wine data-bank (network of Member States Laboratories,
MSL) several points related to the standardisation and improvement of data access and its
use are currently in discussion:

• Concerning fraud detection the network of the EU Wine databank should be used for
addressing information requests between MSL laboratories. This procedure could be
standardized by standard procedures and forms.

• For a first screening and information of samples in question an evaluation of

“condensed statistical data” for each year and wine growing region of each MS
should be provided to the official MSL. The final interpretation of a suspicious wine
requires the detailed data in any case. The condensed data could/should be: wine
growing region, year, number of samples, mean-values, standard deviations,
minimum- and maximum-values of all isotope ratios.

• Furthermore it is generally acknowledged that the MSL laboratory is the most

competent body for interpretation of isotope data of its own region, because it is
supposed to be aware of specific variations. The opportunity to comment the actual
case by the experts of the MSL in question shall be assured.

• Revision of the existing guidelines for interpretation is considered to be very

important, whereas these guidelines should not only refer to the detection of water
and sucrose but also to a general control of authenticity using all isotope parameters
(multi-element, multi-component evaluation) in connection with multivariate statistical
calculations as well as instruments of traceability.

However discussion on these points is in progress. The aim from a scientific point of view
should be an easier and standardised access to the databank which has been originally
established for control purposes.

10.3 δ18O Databank (German wine control)

Independently from the EU Wine databank, δ18O data about wines from third countries have
been collected in Germany in a database since 1998 within the official wine control. They
have been statistically analysed to check geographical origin and to prove watering. The data
are collected by official testing stations and are only forwarded to official testing stations. It is
up-dated twice a year.
The database of the German wine control currently comprises around 1000 entries of wines
from South America, the United States, North and South Africa, Australia, New Zealand and
East Europe. Alongside authentic samples, unsuspicious commercial samples are also
included in the database of third country wines, after having been approved by the respective

Page 14/25
Bundesinstitut für Risikobewertung

expert in charge with the help of a set of analyses and requirements that had been uniformly
defined.
Firstly, the data are examined for outliers (Pearson) specific to countries and vintages. Table
5 shows a basic statistical survey describing the 95% confidence intervals (t-distribution) for
the δ18O values of one country, calculated without outliers. Data are reported according to
country of origin and for each vintage (sub categories of country). These confidence intervals
form the basis for an assessment and frequently suffice to prove watering or a wrong
indication of geographical origin. However, it must be critically examined in each individual
case whether sufficient data are available for a characterization of the wine in question and
whether further imponderabilities such as extreme weather conditions during the harvest
period must be taken into consideration. An intensification of the control in the sense of a
reduction of the respective confidence interval can also be given in particular cases, e.g.
through different geo-climatic conditions of different wine-growing regions within a country. It
must be substantiated by corresponding data. The advices on the handling of data in the EU
database (Annex 1) to assess a wine on the basis of isotope parameters should be taken
into account of course also for using the δ 18O databank. Consideration of all information
available about the wine in question is of fundamental importance in the process. The
responsible and correct handling of the database rests exclusively with the expert and
remains always an examination of the individual case coupled with specific knowledge of the
respective attendant circumstances.
Exemplary one data set for a country or region was extracted from the databank for which
the statistical characteristics are given with Table 5.
18
Table 5. Fictive data set for calculation δ O-value confidence limits.

18
δ O-values in ‰ two-tailed

n 44
Minimum -1.26
Max 2.14
Mean 0.58
Standard dev. 0.89
Median 0.82

Student-Factor 2.02
95% CLower limit (-) -1.22
95% CUpper limit (+) 2.37

This example is also depicted in Fauhl (2006) in which the extend of measurement
uncertainty on the decision limit (enforcement limit) is discussed in detail. It was shown that
the decision limit, taking measurement uncertainty into account, varies only from -1.5 and -
1.6 ‰ between the different calculations. However it is remarkable that if a one-tailed
distribution is assumed for the data given in Table 5 the lower limit would be -0.91 ‰ instead
of -1.22 ‰ leading to decision limits between -1.2 ‰ and -1.3 ‰, so 0.3 ‰ higher than using
the two-tailed distribution.

Page 15/25
Bundesinstitut für Risikobewertung

δ18O analysis is known to be indicative for the geographical origin, but the violation of one of
the limits, upper or lower, does not automatically imply that the wine in question derives from
another origin than stated, because the following points possibly causing the deviation also
must be considered:

- In the case of water addition only the lower limit is relevant because the tap water added
usually is characterised by a very negative δ18O value and its addition causes a reduction of
the initial value in wine water.

- The product might come from another geographical origin (than stated),

- For certain geographical origins for which vintage fluctuations play an significant role (e.g.
middle Europe), high proportions of other vintages of the same origin might have been used.
(Statement of the vintage also to be verified).

10.4 Shikimic acid (SA)

Shikimic acid is found in wine in different concentrations. It derives from caffeic acid and is
precursor of different amino acid in the biochemistry of the grape plant. The concentration in
wine has been linked to its grape variety. In particular wines from Burgundy group are
characterised by a low content of shikimic acid. Already in 2001 Holbach et al. published
data of 457 burgundy wine samples by which it was shown that the maximum level of SA
found in burgundy wines was 28 mg/l. In addition comprehensive data (n=2980) on 25 non-
Burgundy grape varieties were published by Holbach et al. (2001), demonstrating that for
certain questions in grape variety identification the level of SA acid can contribute
information.

The method originally proposed by Holbach et al. (2001) has been adopted by the
International Organisation of Vine and Wine (O.I.V.) as so called Type II method (reference
method) for the determination of the acid (OENO 33/2004, MA-E-AS313-17-ACSHIK). For
that reason the method was validated in a collaborative trial. The method consists of HPLC
determination employing two separation columns (C18 and cations exchanger) coupled
consequently.

Page 16/25
Bundesinstitut für Risikobewertung

A typical chromatogram is shown in Figure 1. The determination of shikimic acid is performed

by HPLC in parallel with other organic acids and nitrate.

Figure 1: Chromatogram of a wine according to the official method (MA-E-AS313-17-ACSHIK)

DAD1 A, Sig=225,4 Ref=550,100 (OS081203\005-0501.D)

mAU

175

acid
150

Tartaric
125

nsäure

acid
acid
Wei
100

Lactic acid
Shikimic
Malic
Äpfelsäure

Shikimisäure
75

acid
Milchsäure
50
Nitrate

Fumaric
Fumarsäure
Nitrat

25

0 5 10 15 20

However discussions on the use, interpretation and reliability of this information occur. The
German official wine control started several years ago a comprehensive data collection on
the shikimic acid content in different wines and meanwhile in 2009 almost 5400 data have
been collected. Wines from many growing regions over the world are implemented, although
some sample collectives for certain varieties derive mainly from Germany. The BfR manages
the data and provides the official wine control with actualised compilations on a regulatory
basis. Already in 2003 the official wine control published reference data for the so called
burgundy group of varieties which are characterised by a low content of shikimic acid
(anonymous 2003).

Page 17/25
Bundesinstitut für Risikobewertung

Figure 2: Publication of control limits for burgundy wines by Working Group of Food Chemistry Experts
of the German Federal States and the Federal Institute for Consumer Health Protection and Veterinary
Medicine (ALS) for control purposes.

Pinot
Pinot Pinot Pinot Pinot
Precoce
blanc noir grigio Meunier
Noir

The confidence ranges published 2003 for the burgundy wines are still confirmed by the
actual data of 2009 on a broader basis of samples.

In particular the so called 85 % rule as regards protected designations of origin and

geographical indications, traditional terms, labelling and presentation of certain wine sector
products, by which it is allowed to add 15 % of another variety or origin, adds variation to the
shikimic acid distribution. In other countries the tolerated mixing rules for variety wines are
different and in some cases less defined. Therefore even quantification of grape variety
proportions is very interesting from the view point of control, but at the current state almost
impossible to be approached.

One of the main criticism of authenticity testing in general is that the data used as reference
are not covering the natural diversity and that possibly some effects - biological or
oenological - on the parameter in question have not been investigated in detail enough. If
one uses data of so called authentic samples or experimental samples which are real
authentic samples their production might differ from commercial samples. However the data
presented consists of both authentic and commercial wines. One strong argument for doing
so is that all possibly effects and variations are covered already and the data are therefore
fairly robust. Dealing with authenticity is always linked to certain level of probability and false
negative results or false positive objections can not be excluded totally.

For some questions the shikimic acid gives interesting information on the authenticity of the
wine variety. For example Riesling wines are characterised by a high content in contrast to
the burgundy wines which show a low concentration of shikimic acid. Therefore shikimic acid
is an indicator for certain varieties and can be indicative for some others. According to the
information of the German wine control the number of objections dropped down drastically
after the consideration of the confidence limits for the burgundy wines.

Page 18/25
Bundesinstitut für Risikobewertung

10.5 Anthocyan composition

Although similar types of anthocyanins are found in different grape varieties, the relative
amounts of the individual compounds differ. For example, it has been noted that Pinot Noir
grapes contain no acylated anthocyanins, demonstrated in Figure 3 which shows the
blending of a pinot noir wine (trace 1) with Dornfelder wine (trace 5) in different proportions
(trace 2-4, with 10, 20 and 30 % Dornfelder in the Pinot Noir wine respectively).

5
100 % Dornfelder
4
30 % Dornfelder + 70 % Pinot Noir
3
20 % Dornfelder + 80 % Pinot Noir
2
10 % Dornfelder + 90 % Pinot Noir
1
100 % Pinot Noir

Figure 3. Chromtograms of the anthocyanins of increasing proportions of Dornfelder in Pinot Noir wine

This feature of burgundy wines is successfully applied for their variety control. For example
the German speciality “Weißherbst” which is a rosé wine produced from 100 % Pinot Noir
grapes, should show no significant proportion of acetylated anthocyanins. Although in this
case the possibility of technologically non-avoidable residues of other grape varieties in wine
production should be mentioned. The typical variety wine in the EU may contain 15 % of
another variety, which must be considered appropriately in the interpretation of anthocyan
pattern.

Analysis of nine anthocyanins and particularly their ratios has been used as a means of
validating the identity of the grapes used during vinification to ward against adulteration or

Page 19/25
Bundesinstitut für Risikobewertung

fraud. Individual grape varieties can also be verified from one anthocyanin compound in
many cases. The acylated anthocyanins have proven to be particularly characteristic for
certain grape varieties, with considerable practical significance being attached to the ratio of
acetylated to p-coumaroylated anthocyanins (Rac/cou) and the sum of acylated
anthocyanins (Sac) in the assessment (Holbach et al. 2001). It has been proposed that the
ratio of acetylated to p-coumaroylated anthocyanins should be >3 if the wine is derived from
Cabernet Sauvignon, and this test is applied to Cabernet Sauvignon-labelled wines imported
into Germany. Although it must been noted that measurement uncertainty of the wine in
question, authorised blending (e.g. 15 %) and in case of sweetened wines also the addition
of further products (e.g. must) shall be considered appropriately before drawing conclusions
on the variety in question.
The analytical method has been adopted by the OIV as Type II method (reference method)
with Resolution 22/2003: “HPLC-Determination of nine major Anthocyanins in red and rosé
wines” (MA-E-AS315-11-ANCYAN).

Figure 4. Chromatogram of a red wine for the determination of nine anthocyanins according to the official
O.I.V. method (MA-E-AS315-11-ANCYAN)
DAD1 A, Sig=525,4 Ref=550,100(AN280706\003-0101.D)
mAU
22.561 - Mv-3-gl

Ratio acetylated / cumalyrated anthocyans:

80 (6 + 7)
R=
(8 + 9)

60

Malvidol-3-5-
diglycosid
36.802- Mv-3-acetyl-gl

40
44.676- Mv-3-cumaryl-gl
19.629 - Po-3-gl
16.113- Pt-3-gl
9.513 - Dl-3-gl

34.901- Po-3-acetyl-gl

43.424- Po-3-cumaryl-gl

20
12.712- Cy-3-gl

0 10 20 30 40 50 min

Page 20/25
Bundesinstitut für Risikobewertung

The Working Group of Food Chemistry Experts of the German Federal States and the
Federal Institute for Consumer Health Protection and Veterinary Medicine (ALS) have
published data on the natural fluctuation ranges of anthocyanins for wines of the grape
varieties Cabernet Sauvignon and Portugieser, in order to furnish wine inspection authorities
as well as wine producers with standardised principles to base their assessment of grape
variety (anonymous 2001).
Figure 5. Statistical data of the Rac/cou for wines of the variety “Portugieser” and “Cabernet sauvignon”
(anonymous 2001):

“Portugieser”

“Cabernet Sauvignon”

Features such as these have enabled anthocyanins to be used taxonomically and to detect
adulteration in wines. However, the extent of anthocyanin extraction depends on
fermentation temperature and duration and the concentration of sulfur dioxide and alcohol.

11. Special analysis

11.1 Glycerol
Glycerol is present in wine as a natural by-product of the fermentation process. During
fermentation about 92% of sugar molecules undergo alcoholic fermentation to produce
ethanol, the remaining 8% undergo glycero-pyruvic fermentation to yield glycerol.
Glycerol is believed to be responsible for the mouth feel characteristics that are often
indicative of high quality wines. It has a sweet taste but only in very high concentrations (>20
g/l) does it contribute to the perceived viscosity. However, as it is established as an important
quality-determining constituent that contributes to the sugar-free extract, it may be
deliberately added to mask a poor quality wine. Since such a practise is prohibited in
European oenological legislation, the addition of glycerol in wine is a fraudulent practice that
was and partly is prevalent.
The simplest method for the detection of illegal addition of glycerol to wines is based upon
the quantitative determination of the glycerol and ethanol content and comparison of these
values. Typically for an authentic wine the glycerol content will be in the range 6 -10% of the

Page 21/25
Bundesinstitut für Risikobewertung

ethanol concentration. However, due to the high variability of the natural glycerol:ethanol
ratio, and the natural variability of all other wine constituents which are classically used to
assess glycerol addition such as gluconic acid – a co-indicator of a Botrytis cinerea infection
on grape - or 2,3-butanediol, the usefulness of these methods is limited. It is feasible that
glycerol could be added to a wine with a low natural glycerol content and still remains within
the range for authentic samples.

More recently it has been shown that detection of adulteration with industrial glycerol is
possible by the determination of trace amounts of by-products formed during glycerol
production (Lampe et al. 1995). Industrial glycerol is mainly produced either by fat cleavage
or synthetically from a propylene feedstock. It is known that glycerol produced by
transesterification of plant and animal triglycerides using methanol contains considerable
amounts of 3-methoxy-1,2-propanediol (3-MPD). The synthesis of glycerol from
petrochemicals leads to impurities of cyclic diglycerol (CycD). As both types of compounds
do not occur in wine naturally, it is possible to determine the illegal addition of glycerol to
wine by the determination of these impurities by GC-MS analysis.

A validated method for the determination of 3-methoxypropane-1,2-diol and cyclic diglycerol,

consisting of an extraction with diethyl ether and GC-MS determination (Fauhl et al. 2004),
was adapted with resolution OENO 11-2007 by the O.I.V. as official method (Type II) MA-F-
AS315-15-GLYCYC.

Quantification of the glycerol addition is difficult due to the variation in by product content of
glycerols obtained from different producers. In addition, advances in purification techniques
have resulted in the availability of high purity glycerols that do not contain 3-MPD or CycDs in
detectable quantities. Analysis of different glycerols on the retail market, fine chemicals and
also food ingredients, showed that the concentrations of the by-products vary in a broad
range. Concentrations of up to ~900 mg/kg were found for 3-MPD and up to ~1400 mg/kg for
the cyclic diglycerols. Concentrations of up to almost 4000 mg/kg (CycD) were reported
earlier for one commercial glycerol (Lampe 1997). Both types of impurities were found in
certain glycerols indicating that mixtures are retailed composed by glycerol obtained by both
principal processing ways. As it had been already established (Lampe1997) it must be
concluded from these findings that the quantification of illegal addition of glycerols to wine is
only possible if also the added glycerol is known and available. This will be only the case in
very few exceptions.
Some enzyme preparations, authorised as fining reagents, may contain glycerol as solvent
and stabiliser. Considerations of the possible intake by application of these preparations
(highest recommended dose of preparation was 15 g/100 l) containing 50 % glycerol lead to
a maximum addition of 75 mg/l exogenous glycerol to wine. With the maximum
concentrations of the by-products of 800 mg/l 3-MPD and respectively 4000 mg/l cyclic
diglycerols mg/kg in technical glycerol the “worst case” concentrations were estimated with
0.06 mg/l 3-MPD and 0.3 mg/l CycDs resulting from an authorised enzyme treatment.

Based on this worst-case scenario and the appropriate consideration of the measurement
uncertainty the German wine control established enforcement limits of 0.1 mg/l for 3-MPD
and 0.5 mg/l for CycDs in 2008.

Page 22/25
Bundesinstitut für Risikobewertung

Analysis of by-products is still a very useful tool for the authentication of wine since positive
determination of these impurities above the decision limits, clearly indicates a fraudulent
wine-making practice, in a much more direct way than the classical assessment.

According to the information of the German wine control the number of objections dropped
down drastically, although some products are still on the market.

12. Pyrazine
Typical Sauvignon blanc wines have a characteristic cultivar character. This complex
character consists of various nuances that are described, inter alia, as green, grassy, green
pepper-like, asparagus-like, herbaceous, fig, gooseberry, litchi and tropical fruit. It is well
known that certain methoxypyrazines are responsible for the typical green
pepper/herbaceous nuances in Sauvignon blanc, Sémillon and Cabernet Sauvignon grapes
and wines. 2-Methoxy-3-isobutylpyrazine (iBMP) is by far the most important because it
occurs in much higher concentrations than the other methoxypyrazines. The particularly low
threshold value of iBMP, namely 2 ng/l in water and 1 ng/l in white wine indicates that this
component can have a huge impact on wine quality.
In late 2004 early 2005 hints arose that Sauvignon Blanc wines from South Africa were
subject of an addition of iBMP in order to improve their sensory perception. One particular
wine on the German market, already ostentatious from its sensory impression (artificially
strong green paprika), was found to contain 150 ng/l iBMP.

Typical iBMP concentration levels, as determined in a number of overseas Sauvignon blanc

wines, ranged from 5 to 40 ng/l in France, from 10 to 35 ng/l in New Zealand and from
approximately 2 to 15 ng/l in Australia. The investigation of more than 200 Sauvignon Blnc
wines of 2003 and 2004 from South Africa showed a maximum level of 10 respectively 14
ng/l for iBMP (Marais et al 2004). Comparison with these data on iBMP concentrations
revealed that the concentration of 150 ng/l is ~10 fold higher than the highest concentration
typically for South African wines and ~3 times higher than the highest concentration ever
found for this variety. Therefore it was concluded at a high level of probability that this wine
was falsified by addition of iBMP.

However it must be noted that more appropriate additions - from a sensory point of view - of
iBMP in the order of 10 - 30 ng/l would certainly require further dedicated work e.g.
establishment of reference data and authenticity ranges, in order to enable an suitable
detection of falsification.

Page 23/25
Bundesinstitut für Risikobewertung

13. Recommendations
From the experiences in Germany concerning wine control it must be concluded that wine
analysis and appreciation is very complex. Analysis itself appears rather accomplishable,
because a high level of unification is given in the field of wine analysis. Applicable validated
standard methods are available. On the contrary the interpretation of authenticity indicators
such as stable isotope ratios, but as well anthocyans or shikimic acid, requires a long term
and constantly growing expertise and experience. Relevant aspects of interpretation are
neither regulated nor standardised and thereby left open to the individual expert in charge.
Authenticity proof of wine products by means of reference data from databanks requires,
apart from the analytical expertise, certainly specific comprehensive product knowledge
which is needed for the interpretation and objection in view of a very complex wine
legislation. Getting an overall impression on the total view of parameters is one of the
complex requirements at a wine chemist in official control function and the stuff involved.
Therefore it is recognized that colleagues in wine control need longer to acquaint oneself
with the analytical and interpretational aspects.

Survey actions, in the way how they are applied often in UK, are suitable for the control of
“hard” parameters for which regulative limits exist. For many parameters which are “health”
relevant limits or recommendations are laid down. However as it regards authentication by
means of experience limits (“soft parameters”) data interpretation becomes an important
factor in the decision making process and for that reason specialised expertise - concerning
analysis, appreciation and legislation - is requested. In Germany the different control
authorities of Bundesländer have wine chemists in charge which are specialised exclusively
to wine products and which provide the expertise in the different requested fields. Although it
is unclear to which extend a similar system is installed in the UK yet, it is highly
recommended that specialized laboratories or sections should be mandated exclusively for
official wine control purposes. In order to develop such activities “training on the job” in
facilities of official German wine control authorities is suggested that also would be the bases
of an intensified co-operation. Particularly since UK is world wide the second largest importer
and one of the top ranking consumers 4 wine control including authenticity appears to be a
very interesting and important field.

4
www.oiv.int: Statistical Information: State of the vitiviniculture world report in 2007 by the Director
th
General: 6 General Assembly Verona 2008.
Page 24/25
Bundesinstitut für Risikobewertung

14. References

Anonymous (2001): Bundesgesundheitsblatt 44:748

Anonymous (2003): Bundesgesundheitsblatt 46:1114

Christoph, N., Rossmann, A. and Voerkelius S. (2003): Possibilities and Limitations of Wine
Authentication Using Stable Isotope and Meteoroligcal Data, Databanks and Statistical
Tests. Part 1: Wines from Franconia and Lake Constance 1992 to 2001. Mitteilungen
Klosterneuburg 53: 23-40.

Fauhl C, Wittkowski R, Lofthouse J et al. (2004): Gas chromatographic/mass spectrometric

determination of 3-methoxy-1,2-propanediol and cyclic diglycerols, by-products of technical
glycerol, in wine: Interlaboratory study. Journal of AOAC International; 87(5):1179-1188.

Fauhl, C. (2006): Measurement Uncertainty in Wine Appreciation, Mitteilungen

Klosterneuburg 56: 3-13

Holbach B., Marx R., Zimmer M. (2001): Bedeutung der Shikimisäure und des
Anthocyanspektrums f die Charakterisierung von Rebsorten. Lebensmittelchemie; 55:32-34

Lampe U, Kreisel A, Burkhard A et al. (1997): Zum Nachweis eines Glycerinzusatzes zu

Wein. DLR; 93(4):103-110

Marais J, Minnaar P, October F (2004): 2-Methoxy-3-Isobutylpyrazine Levels in a Spectrum

of South African Sauvignon Blanc Wines. Wynboer
(http://www.wynboer.co.za/recentarticles/0204sauvignon.php3)

Martin, G. G., Martin, Y. L. and Hertz, J. (1996): Statistische Konzepte in der

Fruchtsaftanalytik. Flüssiges Obst 63: 647-654

O.I.V. (2008): O.I.V. Compendium of International Methods of Analysis of Wine and Musts.
Edition 2008 (www.oiv.int)

O.I.V. (2005): O.I.V. Resolution (2005): OENO 09/2005. Recommendations on Measurement

Uncertainty.

Otteneder, H., Marx, R. and Zimmer, M. (2004): Analysis of the anthocyanin composition of
Cabernet Sauvignon and Portugieser wines provides an objective assessment of the grape
varieties. Australian Journal of Grape and Wine Research 10: 3-7.

Page 25/25