Chapter 5

Wine Tasting Procedures and Overall Wine Flavour

5.1 Wine tasting

In Chapter 3 we discussed the individual components in a finished wine that mainly contribute to the colour and taste of wine, while in Chapter 4 we examined the volatile components, which are likely to contribute to the smell of the wine. However, it is the combination of the wine colour, taste and smell that we appreciate when appraising wines. The word flavour is usually employed to indicate the combination between smell (or odour) and taste. The word ‘tasting’ is often used to indicate that the flavour of the wine is being judged. Judging wine is generally in contrast to drinking wine, the former is often a more or less formal occasion, while the latter is a social event, during which we mainly consume wine because we enjoy doing so.

Our evaluation of the wine during a social gathering is influenced not just by the wine, or aspects immediately related to the wine such as its image, the label, the cork or the shape of the bottle, but also by many other factors such as our mood, the friends we are with, the atmosphere of the evening, the food we are eating, how relaxed we may feel, etc. This may explain why many people are convinced the wine they drank on holiday does taste quite different at home and is often disappointing. It is not that ‘the wine did not travel’, but the ‘wine drinker is not travelling’. Thus it is not just the smell and taste of a wine which influences our choice and acceptance of wine, but a combination of psychological factors; product information, mood, anticipated sensory properties of the wine and psychological effects the wine may have (such as the effect of the alcohol in the wine) will all influence our perception and judgement. This illustrates the importance of wine tasting procedures when tasting wines, to ensure that the wine is being assessed and not the moment or the event.

Indeed wine tasting procedures are described and recommended by wine writers/experts (e.g. Robinson, 1993; Grainger, 2009). Methods for tasting wines used by scientists are standardized to give objective results, or as objective as can be, without bias. For more in-depth information, the reader is referred to Jackson (2008), and Jackson (2009) gives a detailed account of current knowledge of wine tasting. Some general objectives of wine tasting will be discussed in this chapter.

The tasting procedures for scientific purposes, appraisals by experts and judgements by amateurs can all be described in three major parts:

(1) Presentation of the sample to the person assessing the wine, including glass, sample size, sample information, temperature, light, no distractions.

(2) Accurately described task, comprising instruction and training regarding what the person tasting the wine has to do, including how to taste the wine, what attributes to the score, and what and how to score the wine.

(3) Procedure for collecting and analysing, interpreting and presenting the information.

Many influences on our sensory perception of taste and probably more so on smell, affect our overall judgement of food, and most likely also wine, over and above those environmental factors briefly mentioned above. In particular, our sense of smell is influenced by tasting techniques such as sniffing, hunger and hormonal changes. A brief overview on the mechanisms involved in smell perception is given by in Chapter 4 and on taste perception in Chapter 5.

The considerable investigations carried out on the physiological mechanisms involved in interpreting tongue sensations by the brain through the nervous system have been well described by Heath (1988). Less familiar are the concepts and mechanisms involved in the odour assessment, for example from wine, where the taster has to interpret the effect of up to 400 volatile compounds found in wine. Each compound is present in very small quantities, which by smelling from the glass or by drinking (even slurping) reaches the olfactory organ at the back of the nose. This results in a fairly complex cascade of biochemical reactions, briefly discussed in Chapter 4, giving signals that can be interpreted by the brain. The mechanism whereby the human nose detects volatile compounds for the brain to characterize its particular aroma characteristics has been the subject of several different theories, as discussed by Heath (1988). The most obvious is to relate ‘smell’ to chemical or molecular structure, and much progress has been made in this area of research but to date this has not proved sufficient to predict smell. Since the publication of the Nobel winning paper of Buck & Axel (1991) much progress has been made on odour perception, discussed briefly in Chapter 4, and has probably outdated an older theory (Turin, 1997), which assumed that molecular vibrations determine smell, so that the receptor cells in the nose behave as ‘vibrational spectroscopes’. Guti et al. (2001) have given an overview of the various theories.

There are also numerous biological influences on our perception of food and therefore probably also wine. Some factors we may, to an extent, be aware of, such as hunger, and in the case of wine also the way we sniff (discussed below). Our sensory perception of wine is based on the integration of information about many of its aspects. The sensory properties of foods are conveyed to the brain by our senses: colour and surface structure are assessed by vision, structural information is gained by tactile and kinaesthetic senses, combined with hearing while the food is chewed; volatile compounds are perceived by the sense of smell, and the water-soluble chemicals by the sense of taste. The trigeminal nerve (which has endings in lips, oral and nasal cavity) registers temperature, touch and pain, such as in wine the burn of ethanol or the presence of carbon dioxide giving a prickling sensation. Research has shown that this nerve may also enhance the perception of certain volatile components (see Maruniak & Mackay-Sim, 1985). In all, while eating or drinking our senses collect an enormous amount of information, which is rapidly conveyed to our brain. Smell is probably the sense that gives us most enjoyment when eating or drinking. Wines we tend to drink for enjoyment, not for any nutritious reason. Hence the smells we detect from well made wines will make consuming them a pleasurable experience.

5.2 Wine tasting procedure

Although there are endless possible variations, once the wine has been selected and opened, there is generally a set order of events, ranging from the choice of glass to whether or not to expectorate the sample. It is important that for more formal wine assessments, the room is comfortable, free of smells and has northern daylight, in order to assess the wine colour. Of course, sample presentation, sample identification and the order of the samples need to be carefully controlled for formal assessments. Volatile aroma compounds are best perceived if present near or above the sensory threshold for detection and recognition, although synergistic effects need to be borne in mind (see Section 5.3). Some aspects relevant to wine tasting are briefly discussed below and summarized in Table 5.1

5.2.1 Tasting glass

In all wine tasting, a very important aspect is the use of an appropriate wine glass. The glass must be generous enough in size to enable it to be tilted sideways to inspect the colour at the rim of the glass and to swirl the wine to allow a better judgement of the aroma of the wine. The wine glass should have a stem that allows the taster to hold the glass, so that the wine is not unintentionally warmed up, or the clarity of the glass obscured. Such glasses tend to be tulip shaped, e.g. narrower at the top. The glass should be only part filled to give a generous amount of air-space above the glass from which the assessor can sniff the aromas escaped from the wine and present just above the wine in the glass. The glass has to be wide enough at the top to allow the nose to be comfortable above the wine, but not so wide that insufficient of the volatile compounds remain present in the air-space immediately above the wine.

The BS ISO glass has been especially designed to fulfill these requirements, however many other wine glasses having that general tulip shape can be used for assessing wines. Reports in the popular press seemed to suggest some effects of the phenolic composition as a result of the presentation of a red wine in different shapes of glass (New Scientist, 31 August 2002, p. 23); however, the information given was inaccurate and in the short time span the wine is in the glass it is likely that the main effect is the release of the volatile compounds rather than the chemistry of the phenols.

Table 5.1 Aspects relevant in wine tasting.

Glass Tulip shape, allowing appraisal of colour and flavour.
Serving Comfortable, clean room, no smells, with adequate daylight for the appraisal of colour.
Wine temperature: light wines cool, dark wines warmer, maximum room temperature.
Decanting, only needed when wine have deposit or reductive aroma.
Avoid oxygen for delicate older wines and white wines.
Sight Judge appearance of wine (clear, no deposit).
Colour – tilt glass, at meniscus indication of age in red wine (blueish, orange/brown)
Depth of colour, indication of anthocyanin content, possibly related to other phenols
Smell Swirl wine to help release aroma volatiles from wine
Sniff (typically 200 ml in 0.4 seconds) for nasal aroma perception
Adaptation, temporary reduced ability to smell, short exposure
Taste Sample of about 30 ml in mouth, swirl round in mouth whilst sucking in air
Retronasal perception of volatile aroma compounds
Perception of mostly sour, sweet, bitter and astringency
Perception of combination of taste and aroma compounds (flavour, bouquet)
Adaptation to taste components
Role of bubbles
Perception of stringency, mouthfeel, time needed to perceive and recover
Expectorate or swallow
Interactions Psychological factors, such as expectation
Compounds, for example sugar may influence fruity perception
Compounds having both smell and taste, such as acetic acid
Interactions between aroma compounds, giving suppression or enhancement effects
Effect of food on wine perception
Appraisal Hedonic like or dislike
Descriptive, use vocabulary to describe sensory properties
Value for money

5.2.2 Serving

Serving the wine is also still a topic shrouded in some mystique. Of course the wine to be tasted from the glass needs to be at the ‘right’ temperature and wine books recommend temperatures considered best for wines (e.g. Robinson, 1995). Generally, the lighter the wine, the lower the serving temperature, for example light white wines are served the coldest, light red wines are slightly cooled and full red wines are served at room temperature. Breathing the wine is rarely necessary, and if the wine is left in the bottle after extracting the cork the access of oxygen is probably highly ineffective since wine presents a relatively small surface area for oxygen diffusion. Many modern white wines are unlikely to benefit from oxygen at all.

Decanting tends to impart more oxygen into the wine than when used directly from the bottle and is best only done carefully with wines that have a sediment in the bottle. Clear wines do not need decanting and it is thought that aged wines or very fresh young wines (especially white ones) can easily lose their aroma as a result of reactions with oxygen. The room in which the wine is tasted should be normal room temperature, free of strong smells, which could interfere with our assessment of the wine. The colour of the wines should be assessed in natural northern daylight; in particular, our perception of the red colour quality can be affected by lighting. Special daylight tube-lighting is usually fitted in professional or scientific tasting rooms, so any variation of daylight can be controlled.

5.2.3 Visual

The assessment of the quality and quantity of colour as well as the clarity of the wine is done entirely by eye, usually at the beginning of the tasting. The glass is angled against a white background and the wine should be clear, unless the wine is drawn directly from a wine barrel. The colour can indicate the grape variety and age of the wine, although the information is not precise or conclusive. A red wine with a deep-red intensity suggests Cabernet Sauvignon or Syrah wines, generally grapes from a warm region and/or wines made with long maceration. A light red wine suggests a cool climate and grapes such as Pinot Noir.

The colour quality of the wines at the rim of the glass gives an indication of the amount of ageing in particular a red wine has received. Young wines are blue or more purple at the rim but as the wine matures this changes to brick red, orange–brown or even yellow. A comparable colour at the edge of the glass of two different wines does not necessarily indicate the same age, since the colour depends on other factors, such as the aeration the wine has received during maturation and the maturation temperature. White wines can also develop a considerably brown colour, such as Sherry that has been stored under oxidative conditions. The colour of a wine can also affect our perception of wine quality.

It is thought that even a very crude indication of the alcohol content can be obtained. In such an appraisal, colourless streams of wine running down the side of the glass after the wine has been swirled around (often referred to as tears or legs) is taken as a measure of the alcohol content; presumably, differences in viscosity and surface tension of the wine assessed give different effects. However, although wine books often refer to such possibilities, there seems to be no research data confirming that such judgements can indeed be made. Very sweet wines probably have a higher viscosity from their sugar content and after swirling it may take longer for the wine to run down the glass. Again, there is no research data to confirm this speculative statement. Pure ethyl alcohol is, however, known to have a much lower surface tension (22.3 dynes cm−1 at 20°C) than pure water (728 dynes cm−1 at 20°C) but the viscosity of both pure alcohol and water are the same (1.1 cp at 20°C). However, the viscosity of alcohol–water mixtures is generally higher than that of either (e.g. 3 cp for a 40% solution).

Jackson (2008) quotes research from Neogi (1985), who attributed the formation of tears to a physical process. After swirling, water droplets form tears slowly sliding down the sides of the glass, counteracted by evaporative losses of ethanol on the sides of the glass, which offset the action of gravity pulling the drops down by pulling wine up from the glass.

5.2.4 Smell

Next, our sense of smell will be used. There are two routes for the volatile substances to reach the olfactory organ, which is placed just behind the top of the nose where smells are somehow being recognized, followed by signals to communicate this information to our brain in order to identify the smell. The two routes are referred to as nasal and retronasal. Nasal means that volatile compounds will reach the olfactory organ through the nostrils of the nose during the period of nosing the wine from the glass. Retronasal means the smells escape from the wine in the mouth and travel via the back of the mouth to the olfactory organ.

Nosing is the traditional sniffing of the air-space above the glass of wine, usually before any sample is placed in the mouth. Sniffing serves a useful purpose. Under normal breathing conditions, only 5% of the respired air reaches the olfactory organ, while during sniffing this increases to about 20% (see Maruniak & Mackay-Sim, 1985), hence only a fraction of the odour compounds released from the wine into the air will reach the olfactory organ. Thus sniffing presents a much more efficient sample of volatile compounds than normal breathing would have done. Studies have shown wide variations among sniffing patterns between people but individual patterns were remarkably consistent over time. The average sniff of subjects was 200 ml in 0.4 s, giving an inhalation of 30 L−1 min. The individual sniffing techniques do not appear to affect performance in sensory perception, hence people seem to have their own individual sniffing technique that helps them to get the most smell from the wine.

Retronasal perception means wine is in the mouth when the flavours are released for perception. There will be a temperature change, with the wine warming to body temperature. Also the wine will to an extent be diluted with saliva. Genovese et al. (2009) found that dilution with saliva in model systems significantly affected the flavours released from the wines tested. In white wines diluted with saliva, the release of esters and fusel alcohols was reduced by 32–88%, whilst the release of 2-phenylethanol increased 27% and furfural increased 155%. These differences are large enough to explain the difference between nasal and ortonasal perception. The effect was less marked on red wine, showing only a decrease of some esters by 22–51%. The authors suggested polyphenols in wines binding to the proteins in saliva inhibit the activity of the salivary enzymes.

Interestingly, we ‘adapt’ to a certain smell, which means that during prolonged exposure to a certain smell we tend to no longer notice it. In sensory terms, adaptation is always a temporary change in sensitivity. For example, if you sit in a badly ventilated smelly room, you probably do not notice it yourself, at least not after a while. However, after a break outside this room and coming in from outside, you probably do. Thus, when tasting a wine, the first sniff will give information about many smells in the wine to the brain via the olfactory system. However, we can maximize our performance by not constantly sniffing the wine but by sniffing and drinking at intervals, giving our senses a chance to recover. Prolonged sniffing can lead to adaptation and a 30–60 second interval is thought to be sufficient for adaptation to disappear. There is also cross-adaptation, whereby a compound with a similar chemical structure inhibits the perception of another.

The physical laws controlling the release of volatile compounds have been fully discussed in Chapter 4. Generally, the volatile compounds in wines will be readily released while we are tasting. The concentration of each compound released into the air depends on its physical properties. Since wines contain usually between 11 and 15% alcohol by volume, alcohol will be the main influence on the ‘behaviour’ of the volatile compounds, determining what amounts of each volatile compound in the wine will travel into the air-space. Swirling the wine in the glass will aid the otherwise slow and diffusion-controlled process to reach maximum concentration of the volatile compounds present in the wine in the air-space (equilibrium will only be apparent, since the volatile compounds will continue to escape from the space above the open glass). Thus, swirling the wine before nosing it in the glass will help release more volatiles from the wine into the air-space above it and gives us the opportunity to have a good smell of the wine when nosing. Similarly, gently warming the wine will help release more volatiles and indeed may affect the balance of volatile compounds in the air-space. A wine served very cold might have a different ‘nose’ once it has warmed up to room temperature. A very cold wine will release only the most volatile compounds, while as the temperature increases relatively less volatile compounds are increasingly released from the wine into the air-space. Apparently, no data is available on times to reach equilibrium or maximum concentration. The time is probably of the order of minutes. Calculations are possible, but not reported, on the basis of mass transfer coefficients, including the known diffusion coefficients of volatile compounds (Dj) in water–alcohol, which are of the order of 1 × 10−5 cm2 s−1 at 20°C.

5.2.5 Flavour

Next, our senses of smell and taste are used to assess the flavour of the wine. Firstly three of the taste sensations (sweet, sour, bitter) from non-volatile substances are perceived on the tongue. Many scientific, and even some popular wine books covering taste, include a sketch of the tongue showing the distribution of the sensitivity to the basic tastes. Sweet is placed at the tip of the tongue, bitter at the base, although the placements of salty and sour vary. These maps are wrong (Bartoshuk, 1993). According to Bartoshuk, the so-called tongue map has become an enduring scientific myth, based on some erroneous interpretations of research data. Data supports the view that all four classic tastes can be perceived on the sites of the tongue where the taste receptors are located, and although the intensity of taste varies over the tongue, this in no way implies that taste qualities are localized on the tongue as suggested by the maps. Heath (1988) agrees, though also states that the central region is relatively insensitive to all.

There is a temporal element in the perception of taste. Sweetness and sourness are rapidly perceived. We tend to adapt readily to sweetness and wine acids and phenols reduce the perception of sweetness. Adaptation to sourness is thought to be slower, and may cause a lingering aftertaste. To reach the maximum intensity of bitterness can take 10–15 seconds, and it takes time to disappear after swallowing or expectoration. Astringency is usually the last attribute to be detected, it takes 15 seconds to reach the maximum intensity and even longer to disappear. This information illustrates the importance of allowing time between assessing samples, so the senses can recover and overlapping effects of the samples are avoided.

Wine flavour is perceived once the wine is in the mouth and is a combination of three of the taste sensations from non-volatile substances perceived on the tongue, and the aroma (or smell) sensation from volatile substances perceived by the olfactory organs behind the nose. A generous portion of the wine has to be put in the mouth, e.g. up to about 30 ml. Once the wine is in the mouth, the wine is warmed up, moved around in the mouth and there is the option of noisily sucking air through the mouth. All these actions help the volatile compounds to escape from the wine and to travel retronasally via the back of the mouth to the olfactory organ. In physical terms, increased temperature will enhance the diffusion coefficients, affecting the partition coefficients, of the volatile compounds. Bubbles in the wine will increase the surface area from which the volatile compounds can escape. It is impossible to predict whether any such differences are large enough to give sensory differences between retronasal and nasal. Volatile compounds detected during nosing are often described separately, and may or may not be similar or identical to those detected from the palate. Again, research data is lacking.

It is important to allow sufficient time to assess the sample, since the quality of aroma often changes during sampling. One explanation for this may be that olfactory adaptation may allow the perception of other compounds.

5.2.6 Interactions

There are many possible interactions, between aroma compounds, between aroma and taste compounds, or between aroma or taste compounds and the matrix, possibly leading to enhancement or suppression of perception. Such interactions make the interpretation of wine flavour difficult to predict, even when considerable analytical data are available. Scientific debate continues regarding the extent of interaction between taste and smell and numerous sensory interactions have been determined, see reviews (Delwiche, 2004; Palaskova et al. 2008). It is interesting to note that interactions even occur at sub-threshold level, for example when a subject is given a sub-threshold concentration of an odour compound together with a sub-threshold concentration of a taste compound, subjects are able to detect the combination. This cross modal summation of selected smell and taste compounds at sub-threshold concentrations demonstrates that central neural integration of smell and taste signals is occurring. Examples of interaction causing suppression also exist, for example, the addition of sugar to fruit juice reduced the perception of sourness and bitterness, increased the perceived sweetness but also increased ratings for fruity ratings. There have been studies to determine whether the presence of, for example, sugar influences our perception of certain smells in wines, such as fruitiness. There are documented physical effects, such as at high concentrations sugar increases the partition coefficients (see Chapter 4).

There is likely to be an effect of ethanol. Le Berre et al. (2007) studied the effect of ethanol on the physico-chemical and perceptual interactions of a woody aroma (whisky lactone) and a fruity aroma (isoamyl acetate). Whisky lactone volatility was less in the presence of alcohol. Numerous interactions between the three compounds were determined, such as a masking of woody odour by the fruit odour and both odours masked the so-called alcohol odour, showing the importance of perceptual interactions even when tasting a very simple model wine.

There are other known interactions between aroma compounds; a well known example is an early study on esters in wines contributing to the fermentation aroma (see Francis & Newton, 2005). Six esters all present above threshold level were added in typical wine concentrations to a deodorized wine. However, there were no differences in odour intensity of this mixture when one of the esters was absent in the mixture, hence it was concluded that the absence of one ester was masked by the odours of the other esters. Other interactions reported are masking of the bell pepper aroma in Cabernet wines by fruity aromas, fruity aromas in red wines enhanced by C13-norisoprenoids (see the review by Palaskova et al. 2008). Volatile compounds with a similar character also are thought to enhance each other, for example two compounds present below threshold value are perceived by an additional effect.

However, cognitive associations can increase the flavour intensity scores (i.e. we expect something sweet to be fruity and therefore we think it to be more fruity than it really is). It would be interesting to study whether the sugar in sweet wines affects our perception of the flavour of the wine. Considering the large numbers of flavour compounds analysed in wines (see Chapter 4), the interactions in our sensory perception make it even more difficult to determine the contribution individual compounds make to the overall wine flavour.

In addition, there are compounds that we can taste and also smell. For example, acetic acid prevalent in vinegar has both a taste (sour) and a smell (vinegar). In wood-aged mature wines (Madeira) the acetic acid content may be high enough to contribute to both taste and smell.

The flavour or taste of drinks can, to an extent, be affected by compounds present in foods consumed with the drink – these are usually referred to as flavour modifiers or flavour enhancers. For example, it has long been said that chewing artichokes spoils the flavour of fine wine and makes water taste sweet (see Moulton, 1982). This sweetening effect lasts 4–5 minutes, although not everyone experiences this effect. Monosodium glutamate is perhaps a better-known flavour enhancer. Research in simple taste solutions showed that the threshold for sourness and bitterness were both lowered but there is no research data available for wines.

A recent paper by Bastian et al. (2010) reported interaction between food and wine. The authors reported that eating cheddar cheese just before Shiraz wine tasting reduced the flavour length and astringency of the wine, whilst the tannins were perceived as silkier. Most sensory studies are on single food or drink items, in particular with wine where there are so many well established opinions on food and wine combinations, there is scope for further research studies.

5.2.7 Astringency

The moving around of the wine in the mouth also helps to assess mouthfeel, which contributes significantly to our perception and enjoyment of wines. In particular, red wines are often referred to as ‘full bodied’ or ‘robust’, and would be described as lacking ‘body’ and ‘backbone’, without adequate astringency. This is not a taste but a sensation perceived in most parts of the mouth. This sensation is usually referred to as astringency. The mechanism of astringency perception is still debated in scientific research and probably involves phenolic compounds interacting with salivary proteins (Chapter 3). The sensation tends to be a drying, puckering experience in the mouth. The accepted definition of astringency is the complex of sensations due to shrinking, drawing or puckering of the epithelium as a result of exposure to substances such as alums or polyphenols. Astringency perception is not confined to any particular region of the tongue or mouth and gives a diffuse feeling of extreme dryness and roughness. Some astringency is needed for wines to have adequate mouthfeel. An excess is unpleasant and such wines need to mature to soften.

5.2.8 Judging the wine

To avoid having the sense of judgement affected by imbibing alcohol during multiple tastings, expert wine tasters usually ‘spit out’ (expectorate) the wine sample after the assessment. Many wine books also comment on the desirability of ‘length’ or ‘finish’, which means that the desirable smells and tastes should linger in the mouth after swallowing or expectorating the wine. The comment ‘short’, in this context, indicates that the wine flavour disappears pretty well as soon as the wine leaves the mouth. Wines with ‘good length’ are generally more highly valued by expert tasters.

There are many different tasting terms but a confusing number of definitions of these terms. No consensus seems to exist in the use of terms like bouquet, aroma, etc. and different wine writers may use these terms with different meanings. The term ‘aroma’ is most commonly used to describe the smell of the wine derived from the grapes, while ‘bouquet’ tends to refer to the smell of the wine formed as part of the development during maturation. In fact, many different terms are used in sensory analysis, the science of tasting food and drink and many of these terms are also applicable to wines. Some terms are well defined in, for example, the ‘British Standard Glossary of Terms related to sensory analysis of food’ (1992). Even so, there seems to remain some confusion, specifically with terms like bouquet, aroma, flavour, nose, where even scientists seem to give definitions whereby some terms have multiple meanings. Some terms relevant to wine tasting defined by the British Standard Glossary of Terms are:

(1) Sensations perceived via the taste organs when stimulated by certain soluble substances.

(2) Sense of taste.

(3) Attribute of products including taste sensations.

There are many lists of wine descriptors, further discussed in Section 5.5.1. For the researcher, it is important to tailor a selected set of tasting terms to the wine samples to be assessed and train the panel appropriately. However, the amateur can experiment with the information on taste terms and interestingly research has shown that having an adequate vocabulary to describe the wines helps towards becoming a more discriminating wine taster (Hughson & Boakes, 2002).

It is not really surprising that sensory information contains usually a lot of variation. There are differences in sensitivity between people, interactions between and within samples, a general lack of adequate terms to describe sensory perceptions and a wide array of information in one sample has to be assessed.

5.2.9 Reasons for wine tasting

There is a great difference between the sort of tastings that can be done and the information that can be obtained from such exercises. The rigidity of the tasting and the recording of the information depend entirely on the purpose of the tasting. Wine tasting can be done purely for the experience and left just as mental judgements at the time. However, if the information is needed for publication in magazine or newspaper articles, some notes are usually made. In the scientific tasting procedures used to analyse wine, much effort is put into designing the experiment so that bias is taken out as much as possible and the collected data can be analysed using statistical analysis, so that calculated values have a known error term.

Sensory analysis

Why do people taste wine, other than for pleasure? The answer is relatively simple. There is an abundance of instrumental measurements available, many used to ensure the wine complies with legal and safety restrictions. Many chemical constituents can be measured instrumentally but the overall impression of the ‘taste’ of wine cannot generally be predicted from instrumental analysis. Hence tasting remains a very important part of the tests performed on wine and over the years tasting has also grown into a relatively new branch of science that more commonly is referred to as sensory analysis. Over recent decades this branch of science has made valuable contributions to our understanding of wines and the effect of various changes of production on the sensory properties of wines.

Tasting forms a valuable assessment of the wine’s sensory properties and is used at many stages before the wine is enjoyed at the dinner table. The producer will taste the wine during various stages of production, from the vineyard (are the grapes ripe?), just after wine-making (young wine) and at various other stages of the maturation until the wine is sold. The brokers, the merchants, the négociants/shippers will also taste the wine to assess whether the wine is what they expect it to be. Official bodies for certifying wines with Appellation Contrôlée’ type quality denominations will also taste wine to check whether it is typical for the region. Competitive tastings have a rôle in comparing wines and making selections regarding quality judgements. The consumers will ultimately taste the wine and decide whether they like it enough to ever buy another bottle.

Quality tastings

Many of the wine tastings are done to determine the quality of the wine, especially the tastings by expert wine tasters. A pertinent question is what this quality actually is. In its broadest sense, ‘quality’ means fitness for its intended use. In many instances the wine drinkers confuse the often rather ill-defined word ‘quality’ with ‘like’; however, these words have very different meanings in the world of assessing drinks. For wine ‘quality’ is an integrated response to the sensory properties of the wine based on the expectation one has of a particular wine based on previous experience. Thus, for example, young red Bordeaux wines intended for maturation before drinking are likely to be deeply coloured and with a fairly high tannin content. Quality is an individual and subjective response, dependent on the person, based on the person’s experience, expectation and preference.

The word ‘like’ should denote how much we like or dislike a wine, a personal preference in other words, and should be independent of the ‘quality’ assessment of the wine. However, in practice it is not easy to keep our ‘like’ or ‘dislike’ of a wine from interfering with the ‘quality’ assessments. Hence most experienced wine judges will have differences in opinion about wine quality. Many quality assessments are typically to test whether the wine meets the specification set. In wine terms that may mean that the wine should conform to the expected regional distinctiveness and standards. This is of interest to a wine buyer when assessing whether to buy the wine. Or, possibly, the wine buyer may be looking for a more innovative wine, having some ‘artistic merit’, in order to market a wine with certain characteristics in a special promotion.

Even though ‘quality’ assessments are likely to contain an element of personal preference, as discussed above, the opinion of the wine expert is generally well respected. Despite the shortcomings of such expert judgements, there is generally a remarkable agreement on many attributes of the wines. Many quality judgements are made by experts within a sub-group, such as Australian red wine made from Cabernet Sauvignon. This appears to be quite a satisfactory basis for judgement. The price of wine appears to be a good arbiter, since the quality of the wine, no doubt in conjunction with its image, allows, for example, Château Lafite of 1990 to command currently £500–750 for a bottle; the price has rocketed over the decade since 2000. However, although these wines are attributed with having a high quality, this does not mean that all consumers like an expensive wine more than a well made cheaper one, as is sometimes illustrated in TV programmes, where the most expensive wine is not necessarily always the most well liked.

Identifying wines by tasting

Identifying a wine 100% correctly in a blind tasting is very difficult. These ‘blind’ tastings shown on TV wine and food programmes have excellent entertainment value, but to identify a wine correctly on tasting every time is likely to lead to many embarrassing mistakes. In fact, the chemical identification of wines to ensure they come from a certain area, grape and quality standard, the so-called authenticity of wine, forms part of a branch of science that uses highly sophisticated analysis to try and establish wine authenticity, outside the scope of this book but refer to reviews (Isci et al., 2009). Using sensory methods for authenticity tests cannot be relied upon.

Experienced experts, exposed to tasting many different wines each day, seem to be able to identify wines quite well and there is a strategy to use (Broadbent, 1979). However, with an ever increasing number of good quality wines being generally available, the task becomes ever larger. Even experts tend to think that ‘One glance at the label is worth 50 years of experience’. However, with some training and thought it is possible to analyse the wine while tasting and pick up the characteristics of the wines (colour, smells, taste, etc.). It may be possible to determine the grape variety and, combined with aspects of the style of the wine, one can make an intelligent guess about the possible origin of the wine. In fact, a study by Hughson & Boakes (2002) compared the performance of testers in a group of wine experts and a group of novices to wine. They found that a short list of appropriate variety-relevant descriptors enhanced performance. In addition, differences in performance were attributed in part to the lack of vocabulary and information on varietal types that experts employ in tasting tasks.

Sensory analyses used in research

For research purposes, the scientist hopes, usually, to answer a very precisely defined question; typically only one variable is changed, and the effect is compared with the normal so-called control wine. For example, the effect of grape maturity on the wine may be tested by picking grapes from the same vineyard at two different picking dates. The word quality would be defined in specific terms, for example, by testing if there is a difference in fruitiness between the two resulting wines.

In such a research experiment, a panel of people is used to perform sensory assessments, rather than just using one or two experts. Using a larger number of tasters allows the collected data to be pooled and analysed using statistical methods. Such panels are often specially selected to ensure their taste and smell sensitivity. In addition, they are trained, depending on the tasks they are asked to perform. For example, for wine tasting they would be trained to assess wines and are taught to recognize certain smells or tastes in wines. The sensory experiment has to be carefully controlled to eliminate any bias and to ensure that the experiment can be repeated so as to give comparable results. Samples are usually presented blind, with a random number, in identical glasses. The experimental design also needs to be compatible with the use of statistical analysis.

Consumer tasting

There are many different ways to taste foods and drinks under controlled conditions, which are used in addressing research questions. One important style of tasting is hedonic – this means, do you like it? For example, a hedonic question can be asked to compare two wines to determine how many people prefer one to another wine and possibly get some pointers for the reason why (too sweet, too bitter, etc.). Currently these sorts of tasting are often done using ‘normal consumers’ (not trained) and the methods used form an important tool of market researchers, for example, trying to establish in advance how successful a new wine may be. Depending on the information the researcher wants to collect, wines can be presented blind or in their normal packaging.

Analytical tasting

Another style of tasting is strictly analytical and the person assessing the wine does not allow his or her like or dislike to interfere with the description of the wine. For example, a person assessing Australian red wines based on Cabernet Sauvignon might normally prefer ‘fizzy’ white wines from Italy. There are three main types of analytical tasting:

(1) Difference testing. The first step is often to determine whether samples are different, established in an experiment whereby the samples are presented totally blind. A form of triangle test is often used, by selecting the odd one out in a set of three, or by matching one or two samples with a given standard. There is a good chance of guessing the answer correctly in such tests, hence there are statistical tables that should be consulted to check whether the sample difference is significant or whether chance would have given the same result. Therefore tests have been developed to reduce the so-called statistical odds, for example by selecting wines which differ from three others in a set of five samples.

(2) Descriptive analysis. Typically a well-trained taste panel uses words to describe the differences between the wines accurately. Such a sensory panel is used like an instrument and trained in the use of terms to describe the sensory properties and scoring the properties. A scoring system is used so that numeric data can be collected and subjected to statistical analysis. Typically, a so-called scalar rating is used for example, using a line marked with low and high at the extremities. The panelists mark on the line the quantity of each defined attribute, for example the amount of blackcurrant smell in the wine. Very detailed sensory information can be obtained and this data can be matched with, for example, chemical analysis related to the compounds that impart smell and taste to the wine. For example, the effect of picking date of the grapes can be correlated to differences in the smell and chemical composition of wines. This sort of tasting analysis is very informative and permits analytical evaluation of the qualitative and quantitative differences between the tested wines.

(3) Ranking tests. To determine threshold concentrations (the concentration of a compound a person detects) a ranking test is sometimes performed. Samples of increasing concentration of the test compound are presented and the person is asked to rank the series according to their strength. Of course the presentation is very important and the assessor has to be aware that if the strongest sample has been assessed first, time may be required for the senses to recover in order to perceive the weaker samples.

One example of a scientific experiment involved a trained panel to describe the properties of 24 Bordeaux wines from four different communes, whereby the panelists discriminated between the samples primarily on the basis of astringency and bitterness (Noble et al., 1983). Eight Masters of Wine (highly trained and very knowledgeable experts in tasting of wines) described the quality and the properties of the wines. The Masters of Wine were consistent in the specific descriptions of the wines; however, the quality scores reflected, more than anything, the difference in interpretation of quality among the Masters of Wine.

5.2.10 Wine tasting information and analysis

Essential information about wine and tasting should be written down to record more than just a memory of a good wine. Many wine experts also have some sort of scoring system, such as a 20-point system, with high numbers given for high quality wines. The following information may usefully be recorded (as suggested by Broadbent, 1979), some of it simply copied from the bottle, but also some comments based on one’s own experience:

Whether this information is sufficient or far too much rather depends on what you want to know. If the question of interest is whether the wine is liked or not, the information recorded needs to be quite different from when the question of interest is what the precise sensory properties of the wine are. The methods of tasting and recording tend to be tailored. There is a great difference between the type of information recorded in research and the quality evaluations typically used by experts, as the expert tasting usually has a different aim to that in research.

Statistical analysis

Numerical data recorded by a panel of trained tasters or a large group of consumers is usually too unwieldy by itself for the experimenter to draw any sensible conclusions without subjecting this data to statistical analysis. The development of computing power has allowed the rapid and cheap use of older statistical techniques but it has also led to the development of very powerful new statistical analysis. The large data sets generated by sensory analysis can be handled in several ways, and the experimental design needs to be compatible with the use of statistical analysis. Many detailed textbooks are available, usually covering both the methodology of the actual tasting procedure and the statistical analysis (Piggott, 1984 & 1986; O’Mahoney, 1985; Jackson, 2009; see also Section 5.7.3). Only some general information will be given here.

Statistical analysis forms an important and integral part of tasting with taste panels. When consumer tests are carried out to find out which wine the consumer likes best, the experimenter’s main interest is in the differences between the consumers and especially what their differences may be due to; for example, preferring sweet wines to dry ones, or preferring the bouquet of one wine to another one. Data from trained taste panels usually addresses a more analytical question about the samples; for example, is there a difference between these three wines, or what are the flavour characteristics that differentiate these wines? Even this data usually shows large differences between people, attributed to the differences in sensitivity between the assessors, even though the taste panel has been specially selected and trained. Commonly, each assessor tastes the same sample in duplicate or triplicate (all presented blind, so the person does not know the identity of the sample). Presentation order of the samples forms part of the experimental design.

A first step is often analysis of variance, to determine whether there are differences between the samples and whether the individual panelist’s assessments are reproducible. The means of the various terms used to describe a wine are calculated, together with some information regarding the likelihood of there being real differences, or whether any differences are just due to ‘noise’. Comparisons of a few wines for their mean scores (and errors) can be done on a spider plot (Fig. 5.1).

These large sets of data can now be manipulated using specialist statistical packages on computers. Usually, the statistician first decides whether the data ‘makes sense’, which means tests are done to ascertain that all panelists are judging samples more or less in the same way. Any so-called outlier data may indicate that the person had not understood the task, or some other problem. Any interference in the data is usually also tested for, to make sure the particular attribute tested is not dependent on another one. This should be known and the information will form part of the interpretation of the results. The most appropriate test will be chosen for the particular question the experimenter is addressing with the collected data set. The last stage is to draw conclusions from the analysis of the results, attached with a calculated uncertainty or error (which is present in all scientific data).

The advancement in computers has allowed the development of so-called multivariate statistics required to analyse large and complex data sets. Essentially, all the data is projected in a multidimensional space (where there are more than three orthogonal possible, somehow). The cut is determined, which represents on two or three orthogonal axes the most variation. If one is lucky, 50%, or even more, of the variation can be explained by the first two axes, and this may be enough to answer the questions asked. Otherwise, the information on a third, or possibly even a fourth, axis many need to be examined. Data from descriptive analysis can be represented graphically, which is usually easier to examine than large tables. The development of statistical techniques, including the so-called multidimensional statistical analysis, has enabled the handling of sensory data and has greatly helped to progress sensory testing.

Figure 5.1 Spider plot showing the mean values of three wines assessed for intensity of 11 attributes on a 0–10 scale (seven assessors, duplicate tasting).

image

5.3 Factors influencing sensory perception

Numerous factors that have nothing to do with our sense of smell or taste can influence our perception, as has been briefly touched on in Section 5.1. However, our sense of taste and smell are very sensitive and robust, illustrating their biological importance. Without these senses, our enjoyment of food and drink would be rather limited and our motivation to eat and drink a varied diet would be diminished. Damage to taste or smell is relatively rare and the nerve cells of the taste organs on the tongue and the smell organs located behind the nose are constantly renewed. When people say ‘I cannot taste anything’, they usually have a temporarily reduced sense of smell, commonly a result of a severe cold, causing congestion.

Although it may seem straightforward that we identify tastes and smells with relative ease, many factors affect our ability to taste and smell, many of which are relevant to our perception of wine. These influences can be physical, chemical, biological and even psychological. Temperature is an obvious physical influence and it is generally assumed that we perceive tastes best when food or drink is at mouth temperature. It seems that the perception of bitterness in red wines increases as the temperature of the wine decreases but little research data is available. The presence of other compounds can also influence our perception, a well known example is adding sugar to reduce the sourness of a drink and thus, residual sugar in wine will reduce our perception of its sourness and also bitterness. The pH (a measure of the amount of acid in a liquid) is an example of a chemical influence and can affect the perception of taste – one effect could be directly on the receptor proteins.

Numerous taste components exert more than one taste or sensory effect, for example alcohol gives a sweet taste but also creates a burning sensation and gives in wines the effect possibly referred to as ‘weight’. Another example is acetic acid, which has both a taste (sour) and a smell (vinegary). A biological example is the interaction of wine with the production of saliva. Components in wine stimulate the amount of saliva produced in our mouth; the extra saliva then dilutes the wine. In addition, the protein present in saliva can bind with some of the components in the wine, thus possibly changing our perception of the wine.

5.3.1 Threshold and sensitivity

The threshold concentration for taste (defined as the concentration at which a compound is correctly defined by its taste, for example sweetness) can be determined using single sets of dilutions. There is also a detection threshold, at which concentration a person detects the presence of a compound without being able to recognize it. The detection threshold is usually lower than the recognition threshold.

Studies have been carried out in grouping people according to their taste sensitivity, and people are divided into Supertasters (those very sensitive to the used test compounds), Tasters (those less sensitive), and Nontasters (those not perceiving the test compounds). Supertasters tend to be more sensitive to other compounds also, for example certain sweeteners. One fascinating study, investigating people with a great preference for sweetness, showed that, generally, people with a ‘very sweet tooth’ were Nontasters (Bartoshuk, 1993). However, not all studies produced consistent results. Surprisingly, the differences in taste sensitivity recorded are small in comparison with the physiological differences observed.

Our sense of smell is extremely sensitive and discriminating, with certain compounds being detected in parts per trillion in the air. Threshold concentration and detection thresholds can be measured, although it is more difficult than these determinations for taste. There can be about a hundredfold variation in thresholds for smells between people. Surprisingly, even the same person fluctuates considerably, dependent on, for example, time of day and hunger (Stevens et al., 1988). Threshold levels for smell/odour are discussed in more detail in Chapter 4.

When comparing what we smell with others during wine tasting, it has to be borne in mind that some people cannot perceive certain smells, or only at extremely elevated concentrations. It is rare for someone to be unable to smell anything at all; mostly there are certain groups of compounds for which a person is insensitive, a so-called specific anosmia (Amoore, 1977). Several diseases, such as Alzheimer’s, Parkinson’s and HIV are believed to be associated with loss of sensitivity for smells or even specific anosmias. Another sensitivity problem discussed above is that we tend to smell differences and hence suffer from adaptation (always a temporary change in sensitivity) when exposed to the same smell for some time. Hence, we tend to smell best if the smell we are presented with is different from the background smell. We ‘adapt’ to a certain smell, which means that after prolonged exposure to a particular smell we tend to no longer notice it; hence the importance of swirling and sniffing wine at intervals.

Finally, another sensory phenomenon is masking, whereby ‘stronger’-smelling volatile compounds mask the perception of other volatile compounds. Although there seems to be no data available, masking probably contributes to the wine lover’s observation that the perception of wine changes over time, when possibly some masking volatiles have diminished or larger concentrations of the masked volatile compound have been released. The above discussion shows that, besides our personal preference for certain tastes and smells, there are numerous chemical and biological processes that will affect our perception of wine.

5.3.2 Vocabulary

Although people are generally very sensitive to smells and can recognize very many different ones, it tends to be difficult to describe the different smells. A first problem is ‘finding the words’. The smell is familiar but out of context it is not always easy to remember the word. In addition, even when you recognize cheddar cheese correctly, one is sometimes hard pushed to describe it any other way than ‘cheesy’ or ‘cheddar-like’, which does not convey much information. Wine experts have developed sometimes quite extravagant sets of words to describe a wine but there does not seem a ready-made vocabulary to draw on to describe the smells of wine.

5.4 Balance of taste sensations in wine

The interactive flavour effect of the different taste sensations in wine – bitterness, acidity and sweetness – discussed above, leading to the desirable situation of ‘balance’, is also the subject of many wine experts and writers. In particular, Peynaud (1986) and Ribéreau-Gayon (1978) have given the concept of balance a formal and semi-mathematical interpretation in a suppleness index.

The first factor is the equilibrium of basic tastes, where they state that acid and bitter tastes should be balanced by the sweet taste: sweet taste ↔ acid taste + bitter taste.

They especially note the important contribution that alcohol content makes to sweetness. This equation explains why red wines rich in tannins and thus bitterness, cannot tolerate as high a level of acidity as can white wines, which are low in tannins. The inevitable high acidity of wine, compared to other fermented drinks such as beer, is tolerable because it is counterbalanced by the sweet taste of the alcohol. Sweet wines containing residual sugars need the high acidity to balance the sweet taste. Although sweet taste is the only one, which is pleasant by itself, an excess in wine is considered ‘cloying’.

Wines, especially reds, must be ‘supple’, meaning that they should not have either an excessively bitter or acid taste. A numerical suppleness index has been defined (Equation 5.1), in which a value of 5 was regarded as balanced. An increase in this index (in mixed units) corresponds to an increase in the sensation ‘softness’ and ‘fullness of body’ characterization (highly desirable in red wines), whereas a decrease corresponds to ‘firmness’ or ‘thinness’. However, Ribéreau-Gayon recognizes the limitations of such an index, which is valid only within a certain compositional range.

(5.1)    Suppleness index = Alcohol (degrees GL) − [Acidity (g L−1 H2SO4) + Tannin (g L−1)]

Other factors, especially in red wines, Ribéreau-Gayon considers as important. ‘Hardness’ (decreasing suppleness) depends upon additional compositional elements, notably acetic acid and ethyl acetate. These compounds in excessive quantities indicate bacterial spoilage and give wines particular organoleptic characteristics reminiscent of vinegar and nail varnish, generally considered undesirable in wines and reducing their quality. Even when present below the organoleptic threshold (of the order of 1 g L−1 for acetic acid and 150 mg L−1 for ethyl acetate in wines) these compounds, especially the latter, intervene in sensory evaluation, particularly of the aftertaste, reinforcing impressions of ‘hardness’ and ‘burning’. He also notes the importance of other features of tannins, according to their structure in a wide range of complexity and their capacity for change through ageing, from strong astringent characteristics to ‘smooth’, ‘supple’ and ‘velvety’, as described in Chapter 3.

5.5 Wine aromas

Wine is primarily described according to its bouquet, and to the odour/aroma element of its flavour on tasting. Peynaud (1986) prefers to use the term ‘bouquet’ rather than odour of the wine being nosed, referring to the smell arising from the surface of the wine being swirled around in a suitable glass vessel, for wines which have been aged, either ‘in-cask’ or ‘in-bottle’ (so-called ‘oxidative’ or ‘reductive’ bouquet respectively), as already described in Chapters 3 and 4. There is the implication here that, unless the wine has been aged, it will have very little aroma perceptible on nosing. Certainly, the concentration of the volatile compounds will depend on the ageing of the wine, influencing the quality of the smell, but there is no scientific evidence regarding its quality. All wines will have palate-aroma, a term used by Peynaud (1986) referring to the retronasal perception already described, so that bouquet will also form part of this aroma/odour assessment. The exact distinction is difficult to define; palate aroma (retronasal) will certainly be more intense than ‘nose’ (nasal) aroma as a larger amount of the same vapour (in compositional terms, approximately) will reach the olfactory organ via the back of the mouth in drinking (and slurping) rather than merely ‘sniffing’ for the vapour to reach the same organ, where the odour characteristic is actually perceived. Of course, in drinking, a proportion of the aroma/odour will be taken up directly from the nostrils also. There is a comparable sensory discrimination in coffee tasting, where ‘brew aroma’ is referred to, when ‘sniffing’ from the surface of a cup of coffee. However, a different physical situation relates to sniffing a dry roast and ground coffee (and instant coffee), where the volatile compounds released in amount/composition will be somewhat different from those expected to be released when drinking an aqueous solution, due to partition coefficient factors (see Chapter 4).

To consider wine aroma alone, without reference to basic tastes, a large vocabulary of terms has been generated to describe these organoleptic sensations. This necessarily uses words borrowed from the smelling and tasting of other odoriferous substances, often using other food (especially fruit) or drink, or other smells (flowers), or smells associated with some chemical processes. Descriptions are essentially by analogy, though may refer to the odour of known single pure chemical substances.

These vocabularies have undoubtedly created ‘wine-speak’ and jargon, as so amusingly criticized some time ago by Simon Jenkins (The Times article, September 1993), then quoting from British Rail Intercity wine list the following description: ‘Hardy’s Nottage Hill Cabernet Sauvignon: another Aussie stunner – loads of colour, loads of flavour, but the flavour is soft, not rasping, and the fruit is deep plums and blackcurrant, with just a hint of spice’. There are no prizes for guessing the identity of this particular wine writer. Interestingly, a correspondent to The Times, noting this Jenkins article, said that she expected wines to taste of grapes, not these and other exotic fruits; this letter indicated a massive misconception about the nature of wine. Serious scientific wine authorities use numerous terms, e.g. Peynaud (1986) in his book The Taste of Wine gives a vocabulary of some 200 different words, whilst Broadbent (Wine Tasting, 1979) lists 120 words commonly used of wine, but also has 12 pages more of complex ones. Clarke (Wine Fact Finder, 1987) is content with some 31 for the ordinary reader (some referring to taste rather than the volatile component). A later book by Clark (Introducing Wine, 2003) adopted a different approach for tasting terms. Fruity flavours were used to group red and white wines separately, whilst 50 non-fruit terms were listed to describe wine. Some of the terms, such as astringent, were related to the mouth drying effect of tannin in wines, whilst other terms were more suggesting an intensity measure, such as meaty, defined as a heavy red wine with solid chunky flavours. Wine books all seem to propose their own set of terms to describe the sensory properties of wines. Possibly this diversity in approaches towards describing wines illustrates the difficulty of the task. Increasingly wine writers have web sites publishing their tasting notes, for example, Jancis Robinson.

Wines can be praised for their ‘fruitiness’ in wine writers’ descriptions, though it is important to describe which kind of fruit flavour they are referring to. Nearly all these fruit flavours are generated during the fermentation; whilst only those best described as ‘grapey’ (also fruity with floral notes) were originally present in the grapes and remained largely unchanged during fermentation, as described in Chapter 4. Changes in all these flavours occur on ageing. Noted is the wine writer’s use of the word ‘aromatic’, which probably means the same as ‘grapey’ since the responsible constituents are terpenes present at different levels in different grapes, as opposed to neutral, as in Chardonnay (without terpenes).

5.5.1 Odour/aroma classification

Odour classification and vocabularies have much in common with those used in the perfumery industry, where perhaps they first originated. One such system with eight main groups, proposed by Peynaud (1986) for the wine industry, is derived from the foregoing and is shown in Table 5.2.

Peynaud (1986) believed that all the odour/flavour terms used in wine tasting can be placed in one or more of the groups. Each group in Table 5.2 consists of terms that can be divided into sub-groups. These terms have separate names, which are those used in practice in the wine trade and industry and other specialist organizations concerned with aroma or flavour, such as in perfumery or coffee.

Thus, the Fruity Group (no. 5 in Table 5.2), of particular interest to the wine trade and industry, can be divided into sub-groups carrying terms as follows:

In practice, the flavour term used in the vernacular or in botanical parlance, such as ‘blackcurranty’ in wines from Cabernet Sauvignon grapes, will be used first, which can then be allocated to an appropriate sub-group/main group. Similarly, flavour terms can also be slotted in other relevant odour groups.

Table 5.2 Classification of odour sensations.

Type Origin
1 Floral arising from the flower parts of plants
2 Woody associated with hard vegetable matter
3 Rustic/Vegetal associated with soft vegetable matter
4 Balsamic associated with resinous substances
5 Fruity associated with the crushed fruits of plants
6 Animal associated with the odour of animals
7 Empyreumatic arising from heating/roasting processes
8 Chemical associated with synthetic chemical manufacture
9 Spicy associated with aromatic spices and herbs
10 Etherish associated with ether substances (slightly sickly)

Some common terms may be appropriate to one or more groups, thus ‘oaky’, whilst others such as ‘soapy’, ‘rancid’, may be difficult to fit in.

These tiers, essential descriptors classified at three different levels, can be graphically illustrated as an Aroma Wheel, developed by Noble et al. in 1987 – a modified version for wine aroma terminology has been quoted in books by wine experts (Robinson, 1995), presumably indicating the use of such a system in the tastings of wine experts. Aroma Wheels have been described for coffee and for beer (Meilgaard & Peppard, 1986). Like all classifications schemes, overlaps and omissions can be found, but even so they will be found to serve well.

It is not often possible to characterize a particular wine aroma with a particular pure chemical compound. As can be seen from Table 4.7 in Chapter 4, many of these compounds have been described with a range of odour/flavour sensations, even in more than one category, and often have differing flavour impressions, dependent upon their concentrations in aqueous solution or in the air-space above. However, we can attribute their sensory effect from their presence in wines, arising as from very dilute aqueous solutions, in most cases approaching infinite dilution.

From Tables 5.2 and 5.3, some groups of chemical compounds with the olfactory sensations can be identified (described in Tables 5.5 and 5.6) for wines from specific grape varieties. Table 5.4 lists known or believed associations from the descriptions given in Tables 4.14, 4.15, 4.16, 4.17, 4.18, 4.21, 4.22, 4.23, 4.24, 4.25, 4.26, 4.27 and 4.28 (Chapter 4), though the listing is incomplete, owing to a lack of information.

5.5.2 Aroma/odour characteristics of wines from particular grape varieties

It is useful to set out the odour/flavour characteristics of wines from the main grape varieties commercially used, as described by leading wine writers and indeed by lay-persons with experience. Wines from some ten main grape varieties at 100% usage were so listed by Goolden (1991), upon which Tables 5.5 and 5.6 were based, although now extended to include many more grape varieties. The second column provides the odour terms used, with those in italic indicating a wide acceptance. The third column places them into relevant groups/sub-groups according to the principles set out in Section 5.5.1. The fourth column distinguishes, where possible, particular aromas/odours as whether primary (P), i.e. arising from original grape (varietal), secondary (S), arising from the fermentation (sometimes from the pre-fermentation stages), or tertiary (T), arising from maturation, in-cask ageing [T (oak) when oak barrels are used], or in-bottle ageing. Finally, other relevant comments are also included.

Table 5.3 Flavour/odour terms in groups/sub-groups/common names.

Group/ sub-group
Number Name Common names used for wines
1 Floral Roses, geraniums, violets, etc.
2 Woody (A term widely used in the coffee industry for green/roasted coffee: only occasionally used for wine).
3 Rustic/vegetal Fresh – herbaceous, potatoes, peas, green bell peppers, eucalyptus, etc. ‘Grassy’ is often used equivalently to ‘herbaceous’ as a flavour defect.
‘Earthy’ is often synonymous.
Canned/cooked – cabbage, asparagus, etc.
Dried – hay, tobacco, tea, leather saddles.
‘Rotten’ – ‘sulfurous’.
4 Balsamic (The term resinous is often a synonym).
5 Fruity Citrus – (yellow fruit) lemon, limes, grapefruit; (orange) orange.
Berry – ‘grapey’, strawberry, blackcurrants, gooseberries, cherries, etc.
Tree fruit – apricot, plum, etc.
Tropical fruit – kiwi, lychées, passion fruit, banana, etc.
Dried and other fruit – figs, prunes, raisins, etc.
6 Animal The originating animal is quoted, e.g. ‘goaty’.
7 Empyreumatic ‘Toasted’, as from charred barrels used in ageing. Caramel, ‘smoky’.
8 Chemical Kerosene, bottle-age
(a catch-all term, with sub-divisions relating to the petroleum industry, creosote and phenols).
9 Spicy Cloves, pepper, liquorice but specific note often not quoted.
10 Ethereal Buttery, sweetish, caramel-like, cake-like, vanilla.

Notably the main characterizing flavour descriptions in Tables 5.5 and 5.6 command a considerable degree of unanimity amongst wine writers/experts. Emphasis on one or more of the characteristics may be different, as can be expected on account of the different area origins of the same variety. Flavours will differ according to the degree of ripeness of the grapes when picked and of course, on the mode of vinification and the length and type of ageing. However, the flavour descriptions given by wine experts/writers are often not clearly defined in respect of grape ripeness and particularly of the type/length of ageing. It is also sometimes difficult to secure wines from one truly single variety in the commercial market, especially in Europe. Many wines in France (except Alsace), though conforming to Appellation Contrôlée requirements, are in fact blends of different varieties, often necessarily determined by the climatic conditions in a given vintage year. A particular label name may mean a somewhat different blend composition in different years, e.g. Châteauneuf-du-Pape, though the overall aim is to achieve a degree of uniformity over the years. Barr (1988) has discussed this issue in detail, though his observations may not command universal agreement in the wine-producing fraternity. This practice of variety blending is less common in the newer wine industries of the USA, in Australia and New Zealand, where wines made from a single variety are often marketed and labelled accordingly; for example, Cabernet Sauvignon or Chardonnay. Single varietal grape wines may be possible due to more predictable climatic conditions. Nevertheless, grapes may still be drawn from a very wide growing area. A similar situation arises in the roasted and instant coffee industries, where brand names are predominant, with products based on blends of different types of coffee beans; the exact composition will vary from time to time according to availability and cost. However, the marketing of single 100% coffee types, such as Kenya, Costa Rica and Colombia, has become more commonplace.

Table 5.4 Relationship of flavour/odour terms in wines to specific chemical compounds.

1 Floral Terpenes: linalool (rose), α-terpineol (lily-of-the-valley), cis-Rose oxide (geranium), etc.
Phenyl alcohol.
Complex ketones such as β-ionone (violets), β-damascenone
(Z isomer), undecanone.
2 Woody Higher unsaturated alkyl aldehydes [e.g. (E)-nonenal].
Oak lactones (oaky), see Chapter 4, Table 4.16.
3 Rustic/Vegetal Particular alkyl aldehydes and alcohols; thus 1-hexanal and 1-hexanol, and trans-2-hexenal in high dilution (‘grassiness’).
The methoxypyrazines, in particular 2-methoxy-3-isopropyl- in low dilutions (potatoes, peas; also green bell peppers).
Some thio-compounds (cooked vegetable), benzenemethanethiol (gunflint, smoke), 2-methyl-3-furanthiol (cooked meat), 2-furanemethanethiol (roast coffee), see Chapter 4, Table 4.28.
Complex lactones (oak lactones – ‘coconutty’) see Chapter 4, 4.26.
4 Balsamic Alkyl hydrocarbons o-cresol (‘medicinal’).
5 Fruity Certain fruity (mainly tree and bush fruit) flavours.
Alkyl ethyl and other alkyl esters from C4–C8 carboxylic acids, and benzoic acid; acetates of higher alkyl alcohols (see Chapter 4, Tables 4.11 and 4.12).
Thio-ketones, 4-methyl-4-mercapto-pentan-2-one
(tropical fruit – kiwi).
Branched esters ethyl 2-, 3- and 4-methylpentanoate and cyclic ester ethyl cyclohexanoate (sweet, fruity).
6 Animal Butanoic acid (‘goaty’).
7 Empyreumatic –
8 Chemical TND 1,1,6-trimethyl-1,2-dihydronaphtalene (kerosine, bottle age).
Phenols: 2-ethyl- and 2-ethyl-guaiacol (‘smoky/phenolic’).
p-cresol (‘tarry/smoky’).
guaiacol (‘smoky/woody’).
9 Spicy Isoeugenol (cloves).
Sotolon, or 3-hydroxy-4,5-dimethyl-2(H)-furan-2-one (‘spicy’).
Rotundone (peppery).
10 Ethereal Diacetyl or butan-1,2-dione (‘buttery’).
Vanillin (‘sweet, creamy’ vanilla).
HDMF or 4-hydroxy-2,5-dimethyl-2(H)-furan-3-one
(‘sweetish-caramel’).

Table 5.5 Grape varieties used for red wines, with described flavour characteristics relating to odour/aroma, for further information (P, S, T) see text.

image
image
image

Wines from a specific area/grape variety(ies) for a certain year are, of course, continuously assessed by producers, merchants and wine experts/writers and judged for quality, value for money, etc. Wines are necessarily identified by their producer, e.g. Château or Domaine in France, or by suppliers. Perhaps the most important assessment is the identification of the quality attributes of wines made in (exceptionally) ‘good years’ for a particular wine on a regional or more local basis. It can refer back to many years from the present. There is usually a fairly good consensus on these dates, at any given time; thus, for clarets from Bordeaux the outstanding years were 1945, 1961, 1982 and possibly 2000, whilst 1968 was considered a disaster (due to a very wet August). The growers proclaimed the product of the 2000 harvest to be the best for a decade, or even a century (Times report, 2 February 2001), and this early assessment seems to be borne out to an extent, as it is currently described as a vintage of ‘great consistency and balance’ (Robinson, website), although this may not be the terms to describe the best vintage for a century! The vintage of 2005 is described as ‘textbook perfection in respects other than price’. Similarly, tables are published showing how a particular type of wine is developing with age and whether a particular vintage year has reached the maximum quality. All such information can only be obtained by studying the publications of reputable wine writers, e.g. Clarke & Rand (2001), Hugh Johnson’s Pocket Wine Book (each year) or Jancis Robinson (website). Nevertheless, fashions of particular styles of wine change amongst the general public, if not also amongst the experts.

Table 5.6 Grape varieties for white wines with described odour/aroma characteristics.

image
image

The next two sections reflect the kinds of difference of quality and flavour judgements given by wine experts when assessing wines from different sources, albeit the same variety of grape was used for vinification.

5.5.3 Variants in Cabernet Sauvignon wine flavour

A wine from a 100% grape variety such as a Cabernet Sauvignon, using the harvesting/ vinification processes typical for the different countries in which the grapes were grown, will inevitably give a range of different wines. However, one particular flavour characteristic, ‘blackcurranty’ (Table 5.5), appears to be present regardless of origin and thus in general, varietal. A woody flavour characteristic; ‘cigar box’, ‘pencil shavings’ or ‘cedar’ arises when the wine has been aged in-cask, together with vanilla. The presence or otherwise of the green bell pepper herbaceous note from 3 methoxy-2-isopropylpyrazine is evident in the descriptions.

Some of the differences noted in wines based on Cabernet Sauvignon grapes are described below. Thus Ribéreau-Gayon (from Peynaud, 1986) comments that:

Cabernet Sauvignon grapes grown in soils that are too fertile, or grown in climates that are too hot, give wines that have a rather crude and more vegetal aromatic character, evoking a bruised leaf rather than a fruitiness.

Ribéreau-Gayon (1986)

(This may, however, be due, as he also mentions, to the use of slightly unripe grapes.) He itemizes such differences by characterization, as ‘cloves’ from a Californian winery (Guadeloupe Valley), ‘liquorice’; Rioja, Spain, ‘seaweed’; France (Languedoc); ‘industrial fumes’; Chile (Santiago); ‘soot’, Hungary (Egar); and states that ‘some closely resemble the Médocean model, but mostly have defects paraded as a virtue’.

However, currently the Australians are seen to be at the cutting edge of New World wine. Formal wine training:

… is at a very high level and the reputation of Australian wines is at a pinnacle. With exports year-on-year increasing. Despite being a relatively small producer, it certainly exerts influence in world wine-marketing and receives positive write-ups for many of its wines.

(Robinson, 1995)

The current water shortages limit the irrigation of vines in hotter regions, normally producing consistent quality grapes and the resulting production may well become affected.

Descriptions of differences noted by other wine writers are similarly indicative, though not so sarcastic and are quoted below, including reference also to the basic taste. Thus Clarke describes a Cabernet Sauvignon wine from the Médoc as:

… a dark tannic wine with a strong initial acid attack, and a stark pure black currant fruit. When aged in new oak, its black currant fruit combined with a cedar cigar-like perfume is stunning.

Clarke (1989)

This description is similar to that given by Robinson (1995), who goes on to describe wines made in Eastern Europe as:

… less refined than Bordeaux counterparts, often recalling red fruits rather than black, and rarely showing any oak influence (other than occasionally, use of heavy-handed oak chippings).

Robinson (1995)

McQuitty (1998) refers also to a ‘grassy, herbaceous quality, reminiscent of green pepper or green olive’, but also to ‘gorgeous, smoky, cedar scents … after two years of oak ageing’, in Cabernet Sauvignon wines from Bordeaux, though usually blended with some Merlot and Cabernet Franc grape wines. Cheap Chilean Cabernet Sauvignon-dominated wines are described as ‘relatively ripe, cassis-sweet and glycerine rich’, whilst New World Cabernets are ‘bigger, bolder and jammier than their European equivalents’. In particular, Californian Cabernet reeks of ‘ripe, minty, fruit’, and ‘Australian versions all have eucalyptus well to the fore’, along with ‘a burly, peppery quality from the Shiraz (grape) with which they are often blended’.

5.5.4 Variants of Chardonnay wine flavour

Chardonnay is another internationally planted grape variety, originally associated primarily with the Burgundy area. Its most quoted aroma/flavour characteristics are given in Table 5.6 where it can be seen to have a number of fruity characteristics (apple, peach, melon). Many Chardonnay wines are barrel aged, so that other additional characteristics can be perceived, e.g. vanilla, together with ‘buttery’ as a descriptor, also ‘toasted’, ‘nutty’ from the oak lactones. Again, overall flavour characteristics do depend upon where the vine is grown and how the grapes are vinified. The grape is regarded as ‘amiable’ since the vines are easy to grow. Both Clarke (1989) and Robinson (1995) emphasize the flavour significance of ageing in oak barrels where practiced, together with fermentation also in oak barrels. The Chablis, Côte d’Or, especially Côte de Beaune (e.g. Meursault, Montrachet) areas in France represent the peaks of vinifying performance and are thus described as ‘luscious, creamy honeyed yet totally dry, the rich ripe fruit intertwined with the scents of new oak’ of Côte d’Or wines, and the ‘gentle, nutty richness, steely acidity’ of Chablis (Clarke, 1989). Máconnais Chardonnay wines are claimed to be somewhat different, being produced in stainless steel tanks, with little ageing and the wines are meant to be drunk young, ‘having a slightly appley flavour, as well as something fat and yeasty’ (Clarke, 1989).

McQuitty (1997 and 1998) distinguishes between Chardonnays from the more northerly and southerly regions of the Burgundy area. Thus wines from the areas in the cool northern climes are described as:

… distinctly herbaceous, leafy in style, green apple/lemon in the mix; Chablis, vegetal richness has an almost cheesy aspect, underpinned by a steely lemon acidity due to its heavy limestone soil.

McQuitty (1997)

Further south, Chardonnay produces ‘those ethereal, creamy, buttery peach and pineapple scents. With judicious oaking, these wines yield a fine, nutty quality’. Robinson (1995) refers to Chablis as tasting of ‘wet stones, with some suggestion of very green fruit’, and …‘after eight years in-bottle, they can develop much more complicated, often deliciously honey-like flavours’. The Côte d’Or wines she describes as ‘hazelnuts, liquorice, butter and spice’.

Nowadays, Chardonnay is planted in vineyards all over the world. Chardonnay wines are produced in a number of other areas in France, e.g. the Loire and Languedoc (as a Vin de Pays). Chardonnay grapes have produced very successful wines in California, e.g. ‘positive fruit salad of flavours with figs, melon, peaches and lychées all fighting for prominence, and some buttery oak to round it out’. ‘Best are drier, steelier, with a slightly smoky, toasty taste from the partially charred oak barrels they mature in’. Australian Chardonnays have conquered the market and are described as having ‘a simple fruitiness, perked up with (usually added) acid and often oak chips’. ‘Not necessarily good for keeping inbottle, unless sufficiently acid, leaving a lemon-peel/citrus flavour’. New Zealand Chardonnay wines reflect the cooler climate and thus more acid, with one example described as ‘a herby base with sweet spicy, hazelnut-scented oak’ (McQuitty, 1998). Chardonnay has been planted in Italy (Alto Adige, Friuli) to give wines with a ‘characteristic cream-like taste with a little nutmeg’.

5.5.5 Flavour description of some other commercial wines

At times, the flavour descriptions of certain specific commercial wines by wine writers, particularly during the 1980s and 1990s, have been quite disparaging. It has to be recognized also that certain wine types become popular for a period (to the UK consumer) and then become unfashionable. Quality and flavour descriptions are bound to be subjective. Certain wine types and vintages can clearly be hyped up for sound commercial reasons, as already described. Nevertheless, opinions that have been expressed may be of interest to the wine chemist.

For example the Liebfraumilch and Blue Nun brand from the Rheinhessen and Pfalz regions of Germany, both essentially branded wines produced in large quantity (for export only), were particularly subject to adverse criticism. Thus, Clarke (1986) wrote ‘should be designated just “tafel wein”, that is, low in alcohol, low in acid, from grapes which are not ripe, sugared up (before fermentation)’.

Blue Nun was re-launched in 1997, but McQuitty (The Times, 15 November 1997) commented that it had a ‘weird, spicy scent and very thin palate. Still a lot of sugar there’. Jancis Robinson refers to Liebfraumilch as ‘almost any medium dry, vaguely aromatic blend can qualify; Niersteiner Gutes Domtal made up of over-priced Müller-Thurgau grown miles away’. Goolden (1980) refers to Liebfraumilch with its ‘tartness, hints of stale milk and sulphur’.

Other branded wines similarly disparaged by McQuitty (1997) were the French Piat d’Or (at one time heavily TV advertised with amusing commercials); and Vin du Pays d’Oc (white) described as ‘oxidised and sulphury’. The grape juice from which ‘this wine has been made must have been left open to the air. Very acid.’ And the red. ‘nasty perfume’. Mateus Rosé, the once very popular Portuguese wine, did not do much better, being described as ‘gross, sweet with a jammy, dirty flavour, as if it had been made with the sediment of old Victoria plums’. Goolden (1980) refers to its former popularity but gives a reasonable verdict. Mouton Cadet (red 1995 vintage) was described (McQuitty, 1997) as ‘dry, starkly grassy, thin and unpleasant’, and the white ‘sulphurous, rubbery, disagreeable’. McQuitty (1997–98) also wrote some negative comments about some Paul Masson Californian wines and some wines from Jacob’s Creek, Australia.

Underlying these comments is the belief that good quality means the absence of any mouth-cloying and sickly characteristics, as is imparted by sugar and certain milk flavours but there should be some slight bitterness and acidity present. The concept of dryness is very powerful also with Martinis, Sherries and some other drinks, especially those normally to be drunk as aperitifs and along with savoury meals. Sweet wines are only considered acceptable after the sweet has been served.

Wine writers are bound to be accused of non-scientific assessments. However, their rôle is not intended to be scientific. As described above, wine writers have a lot of knowledge about the taste of wines and generally assess the wines in the light of their experience and expectation, and they do not usually hide their personal opinion of the wine. However, some of the science of taste and smell does create misunderstandings, although these also occur within the scientific community (see the earlier discussion on the tongue map).

The apparent misunderstanding of the rôle of the tongue in tasting is not, however, unusual. Goolden (1980) referring to a Muscadet: ‘Get past the nose (sulphur) and into the taste, and in a good version (generally “sur lie”) you can pick up the faint waft of flowers. There’s a definite citrussy edge blended with the flavour of unripe melon’. Again, Clarke:

Interrogate your taste buds more thoroughly and they give you the flavour of hay. But the Muscadet tradition is to leave the new wine undisturbed on the lees until bottling. This does two things for the taste; the wine picks up yeasty, salty flavours, and keeps its fresh prickly taste, because some of the CO2 from fermentation is delayed.

(Clarke, 1986)

Clarke & Rand (2001), in the glossary to their valuable encyclopaedic text on grape varieties, somewhat unhelpfully define: ‘Flavour or aroma compounds – substances in wine that can be smelled or tasted’. Use, however, has occasionally been made of Jackson (1994, 2008) with reference to Sotolon in Vin Jaune but it is not clear what ‘glucosyl-glucose compounds’ are. Similarly, Robinson (1995) makes erudite technical reference to 2-methoxy-3-isobutyl pyrazine, but surprisingly refers to ‘fruit’ as the youthful combination of flavour (aroma) and body coming from the grapes rather than wine-making or ageing and ‘esters’ formed during fermentation or ageing, often intensely aromatic (‘… nail polish remover smells strongly esterified’). In fact, it is the other way round, ‘fruity’ is from largely non-aromatic alkyl esters, whilst aromatic fruity or ‘grapey’ comes from terpenes in the grapes. Descriptions of flavour can lack scientific clarity. Thus (Clarke & Rand, 2001) ‘oaked Australian Sémillon is different, being picked riper, has richer fruit flavours of greengages, apricots and mangoes, all mixed with the custardy vanilla of the oak’; though presumably unoaked Australian would have the same fruit flavours (secondary) and oaking is only adding vanilla (tertiary aroma).

5.5.6 Off-odours and taints

A number of olfactory sensations in wines are regarded as undesirable when they occur, as indicated in Table 5.7. An off-odour is an atypical flavour often associated with deterioration or transformation of the product (British Standard, 1992), for example, an excess of acetic acid caused by wine spoilage. A taint is a taste or odour foreign to the product (British Standard, 1992), for example the contamination of the wine with the musty odour attributed to 2,4,6-trichloroanisole, a compound associated with faulty corks. The quality of wine in recent decades has improved and faulty wines with off-odours or taints are relatively uncommon. In part, scientific research has given a much better understanding of the causes and formation of off-odours, and the industrial response of improved grape and wine handling (avoiding oxidation, improved fermentation, scrupulous hygiene, etc.) has eradicated many wine faults.

However, changes in winemaking culture as a result of changes in consumer preference for the required style of wine has lead to an increase in some off-flavour formation, for example volatile ethyl phenols. Wines can also suffer from unsightly hazes but again scientific understanding of these faults has led to hazy wines being virtually a problem of the past. Some of the more common taint problems are briefly discussed below, and for further information on taints and measures to prevent taint as much as possible refer to the review of Sefton & Simpson (2005), and Ribéreau-Gayon et al. (2006).

Cork taint

Cork taint is a serious economic problem, which can also have a negative effect on the reputation of a producer. It is difficult to ascertain the size of the problem, although estimates of affected bottles vary from 1–5%, (see Sefton & Simpson, 2005). Despite the developments in instrumental analytical methods, cork taint remains difficult to analyse and even the sensory assessment is not always easy, especially in wines that have been aged in wood.

Cork taint odour is generally attributed to 2,4,6-trichloroanisole, a stable chemical compound in wine. Other anisoles that give rise to cork taint odours include 2,3,4,6-tetrachloroanisole, 2,4,6-tribromoanisole and pentachloroanisole (Grainger, 2009). The incidence of bromoanisole taint in wines appears to be increasing. However, the identification of the musty smell may be a cause of confusion and occasionally a musty smelling wine is wrongly attributed to a cork problem. For example further investigation shows that the musty smell is due to 1-octen-3-one (mushroom odour), which can be present in wine from grapes affected by powdery mildew, or the musty smell may be geosmin (earthy aroma), also attributed to mouldy grapes. It is estimated that about 15% of wines labeled with cork taint have musty smells due to another cause (see the review from Sefton & Simpson, 2005). Strong earthy aromas have also been detected in wine resulting due to 2-methoxy-3-iso-propyl-pyrazine, derived from mouldy barrels.

Cork taint is transferred from faulty corks into the wine after bottling, although incidences of this taint have been reported that could not be traced back to corks. Much work has been done to try and eradicate this cork problem, however, the musty smell as a result of a faulty cork remains a problem, in part because the sensory threshold is very low (Land, 1989). Land quotes that 50% of the population can detect 1 mg of 2,4,6-trichloroanisole in 5 million gallons of water. Interestingly, ethanol has a significant effect on threshold values, in water Land quoted 3 × 10−8 ppm (0.03 mg L−1) whereas in wine of about 11% ethanol the threshold is 1 × 10−5 ppm, a 333-fold increase in concentration.

Thresholds for detection are reported to be between 1.4 and 4.6 ng L−1, whilst recognition concentrations tend to be a little higher (see the review by Sefton & Simpson, 2005). Consumer rejection thresholds reported are between 3.1 and 3.7 ng L−1. However, this appears to depend on the methodology used and the experience of the consumer, and individual thresholds for experienced panelists have been reported to vary by a factor of 100, whilst inexperienced panelists varied by a factor of 10 000. One can only assume that human sensitivity for this compound varies and training will help with the recognition of the taint.

Table 5.7 Off-odours in wine.

image
image

aQuoted by Sefton & Simpson (2005).

The presence of 2,4,6-trichloroanisole leaves the wine dull, devoid of fruity smells, often with a perceptible nasty chemically musty smell. Instrumentally, it is not easy to pick up the low concentrations of this compound able to spoil the flavour of the wine.

Recently different methoxypyrazines have been reported as taint compounds in wines derived from cork. This taint has been described as musty, earthy or mouldy. The taint compounds have been identified as 2-methoxy-3,5-dimethylpyrazine and 2-methoxy-3-isopropylpyrazine. (see the review by Sefton & Simpson, 2005). Sensory descriptors reported are fungal must, aldehydic, coffee, acrid and nutty taint. In water the sensory threshold has been determined as earthy at 0.4 ng L−1. There are not much quantitative data on these musty taints in wine, presumably it is not easy to analyse and identify this taint which the human nose can detect at very low concentrations.

Pratt et al. (2009) reported a study on mousiness in wine using DNA fingerprinting techniques, in order to match the taint with the microbiological organism responsible for the taint. They found strong links between the microbial composition and some of the most common taints in wine, trichloroanisole was related to Penicillium variabile in most samples, but the accumulation of 2-methoxy-3,5-dimethylpyrazine in cork was complex and it is assumed that a number of micro-organisms are involved. These authors also discussed the possible degradation of taint compounds by micro-organisms, which may ever help to reduce the risk of wines becoming contaminated with these chemically stable taints.

Mousiness

A relatively rare but very unpleasant off-flavour is mousiness in wine. It has been described as a ‘peculiarly disagreeable flavour in wine, which is closely resembling the smell of a residence of mice’ as quoted in a recent review on mousy off-flavour (Snowdon et al., 2006); some of their main points are summarized later in this section. Currently there is no method to remove the mousy off-flavour from the unpalatable wine. Its sensory perception is usually delayed on the palate and it cannot be perceived by sniffing, however, once perceived it can be very persistent. Human sensitivity varies and is thought to be genetically determined and ranges from very sensitive to anosmic. This variability must make it more difficult to deal with this off-flavour in the wine industry and wine appreciating circles. Since the off-flavour is very persistent and is reported to last up to ten minutes, various other ways for testing have been developed. One common one is to rub some wine on the back of the hand and sniff close to the skin. This method can easily be used by the amateur wine drinker.

Three chemical compounds have been identified to give a mousy off-flavour to wine: 2-ethylhydropyridine, 2-acetylhydropyridine and 2-acetylpyrroline. Contaminated wines tend to contain more than one of these compounds, although there is no scientific information regarding the individual contribution to mousiness of these compounds. The formation of these mousy off-flavours is under microbiological control and can be formed in wines infected with lactic acid bacteria or Dekkera/Brettanomyces yeast. The conditions leading to the formation of mousy off-flavour are not yet understood.

Ethylphenols

Raised levels of 4-ethylphenols in wines are associated with undesirable aromas describes as phenolic, leather, horse sweat, stable, varnish, etc. This problem, reviewed by Suarez et al. (2007) has become more prominent in aged red wines, in particular due to the current trend requiring ageing in wooden barrels, since it appears that certain organisms can thrive in old wood. Wines are subjected to ageing in wood without any physical clarification processes, which is thought to enhance the quality of the wine, but this also poses an increased risk of spoilage by slow growing yeasts such as Brettanomyces bruxellensis, Brettanomyces anomalus and Saccharomyces baiili, as well as some lactic acid bacteria, all capable of producing undesirable off-flavours. These organisms appear to be able to grow in the presence of ethanol and require minimal nutrients.

The sensory threshold for 4-ethylphenol is low (230 μg L−1, quoted by Suarez et al., 2007) and typical sensory descriptors are phenolic, leather, horse sweat, stable and varnish. The related 4-ethylguiacol gives is typically described as smoked or bacon and has a lower sensory threshold (47 μg L−1, quoted by Suarez et al., 2007), however, its negative impact on wines is thought to be less. Wines identified as having a high, medium and low typical character due to the growth of Brettanomyces have been reported to contain 3.0, 1.74 and 0.68 mg L−1 4-ethylphenol. Variations in sensory threshold have been reported to depend on the type of wine. For further description of formation see Section 7.4.9.

5.6 Wine and food flavour

When wines are consumed with food, wine and food flavours are intimately mixed in a lunch or dinner and usually aligned with specific courses in the menu. Many food writers like to recommend also the wines that should accompany specific recipes of food. There have always been traditional guiding principles in this choice, in which the most obvious is the recommended drinking of full-bodied red wines with roasted red meat. Similarly, dry white wines should accompany white meats and fish; whilst sweeter wines should be provided with the dessert. In fact, the French preference for eating the cheese and biscuits before the ‘pudding’ lies in the idea of a suitable food to accompany remaining (red) wine not consumed during the main course. When only one wine is used during the meal, modern-day choice often seems to be red wine in preference to white, whatever the food.

These principles can be summarized by the concept of similarity in strength of flavour, so that a strong flavoured wine should accompany strong flavoured food and not be allowed to overwhelm a delicate tasting dish. More sophisticated relationships between food and drink, including non-flavour factors such as political ones, are described by food and wine writers and their chefs or sommeliers, in particular Simon (1946) and Johnson (1979). They feature in menus of public and State occasions, as exampled below:

State Banquet at Buckingham Palace in honour of the French President 21 March 1939
The Food The Wines
Consommé Quennelles aux trois couleurs Sherry, 1865
(soup of fish balls with three colours) Madeira sercial, 1834
Filet de truite saumonée roi Piesporter Goldtropfchen, 1924
Georges VI (Middle Mosel, Riesling)
Deidesheimer Kieselberg, 1921
(Rheinpfalz, dry white wine)
Rouennais à la gelée Reine Perrier-Jouet, 1919
Elizabeth; Garniture Buzanay (Champagne)
Mignonette d’Agneau Royale; Château Haut-Brion, 1904
Petits pois à la française; Pommes nouvelles, rissolées au beurre (Red, Graves)
Poussin Mercy-le-haut; Salade
Elysée
Asperges vertes; Sauce maltais
Bombe L’Entente Cordiale; Château Yquem, 1921
Corbeille Lorraine (Sauternes, sweet white)
Cassolette Basillac Port (Royal Tawny, 1912) and Brandy (1815)
(Adapted from Simon, 1946)

Of interest here is the expected appearance of French wines from prestigious vineyards and their age, i.e. some 35 and 18 years respectively. The German vineyards from which wines featured in this banquet are equally prestigious to this day, see also Section 6.3.2 on Tawny Port.

Modern recommendations for drinking particular wines with certain foods have been made characteristically by Johnson (1979) in great detail, a small selection of which is given in Table 5.8.

Adapted from Johnson (1979).

Popular wine books, web sites and wine retail outlets, even sometimes the back label of the bottle will all give information regarding what wines to consume with selected foods, usually based on the general principles discussed earlier, or occasionally the image the wine company likes to present about its wines. In sensory research there is a great awareness regarding interaction of the wine samples, the time required between samples tasting for assessors to judge all samples under the same conditions and avoidance of a tired palate. A recent research paper researched consumer preference for Shiraz wines and how it was influenced by tasting different cheddar cheese samples just before wine tasting (Bastian et al., 2010). They reported that eating Cheddar cheese just before wine tasting reduced the flavour length and astringency of the wine, whilst the tannins were perceived as silkier. Possibly, this can be explained by the residual fat from the cheese in the mouth interacting with the wine volatiles and phenols. The most preferred wine and cheese combinations tended to contain the wine and cheese samples getting the highest quality and ‘like’ scores. In all pairs the wines had a stronger flavour than the accompanying cheese.

Table 5.8 Modern recommendations for wine and food.

First Course
Asparagus – white Burgundy or Chardonnay. Tavel Rosé
Consommé – medium dry Sherry, Madeira
Croque Monsieur – young red, e.g. Beaujolais
Fish
Fish fingers (with tomato sauce) – white Burgundy, Chassagne-Montrachet
Coquilles St. Jacques – white German wines with cream sauces, otherwise Gewèrztraminer or Hermitage Blanc
Plaice (fried or grilled) – white Burgundy
Meat
Roast beef – fine red wine
Duck – rather sweet Rhine wine (fat counter balance)
Steak (beef)
Filet – red of any kind
T-bone – Barolo, Hermitage, Australian Cabernet
Ham – fairly young red Burgundy
Cheese
English (strong, acid) – Ruby, Tawny or Vintage Port
– red wine, e.g. Châteauneuf-du-Pape
Desserts
Baked Alaska – Sweet Champagne or Asti Spumante
Christmas Pudding
Fruit salad – None
Strawberries and cream – Sauternes or Vouvray

5.7 Aroma indices and statistical methods

Aroma indices described in the literature are essentially based on the ratio shown in Equation 5.2, which can be determined for each constituent. Such indices have also been described as Odour Activity Values in studies on roast coffee brews (Grosch, 2001), whilst so-called Flavour Units have been used in beer flavour studies (Meilgaard, 1982; Meilgaard & Peppard, 1986, in greater detail).

(5.2)    Amount of constituent (say in ppm) / Threshold level of constituent (say also in ppm)

As they are based upon perceived intensities, they are subject to Stevens’ Law (Chapter 4, at the end of Section 4.1.3) so that there are theoretical shortcomings to the application of these concepts. The difficulty associated with trying to interpret flavour units may be the reason why few scientific papers have been published using flavour units. Considerable careful work is needed to establish reliable threshold levels of all the constituent compounds.

5.7.1 Flavour unit concept

As used by Meilgaard, Equation 5.2 is expressed in terms of a difference threshold (as opposed to absolute values) as described in Chapter 4, which depend upon the amount of the same compound already present in the beer.

In determining threshold levels in practice, Meilgaard & Peppard (1986) stress a number of important points: (1) due to the wide variation in sensitivity of perception amongst individual people, group-averaged data (i.e. from trained panels of some 12–15 members) is preferred; (2) due to the possible aroma perception effect of even very small amounts of some contaminant compounds (e.g. ppt) it is essential with most reference flavour substances (e.g. butadione) to prepare them in a very highly purified form, involving multiple distillations and absorption techniques, before use in tasting.

Meilgaard & Peppard (1986) recognized the limitations imposed by Stevens’ equation but quoting from the findings of Teghsoonians (1973), in the restricted range of FU to 0.5–5.0 units in beer samples, n is around 1. The errors will then rarely exceed about ±20%, except in the case of ethanol, which causes direct solvents in the mouth.

Meilgaard (1982) regarded this concept and the resulting technique as useful in producing or giving best estimates of the flavour effect of changing the composition of beers in respect of particular constituents by changing relevant brewing practices.

The flavour significance of a particular substance in a given beer is assessed by knowledge of its threshold value and flavour unit in a so-called null beer, which had a composition not too different from that of the sample under test. The composition of the sample beer has to be known for all the compounds of flavour (odour/taste) interest and, of course, that of the null beer. This technique is used primarily for the prediction of flavour differences, from composition data.

For many substances, threshold values are known for a given null beer, as determined by numerous co-operative studies organized by the American Brewing Institute; thus purified isoamyl acetate (estery/fruity – banana) has a standardized difference threshold of 0.5–1.7 mg L−1 in a beer already containing 1–3 mg L−1. In a particular test situation, isoamyl acetate (actually a 25:75 mixture of 2-methyl-butyl and 3-methyl-butyl) was shown to have a threshold of 1.2 mg L−1 in a null beer already containing 1.8 mg L−1. In a sample beer containing 2.4 mg L−1, it can then be calculated that sample beer had 2.4/1.2 = 2 FU of this component, whilst the null beer has 1.8/1.2 = 1.5 FU. The difference between the two is 0.5 FU, so that the sample beer was considered not likely to show any additional ‘banana-like ester’ impact over that in the null sample. There is also a small amount of a 2-methylpropyl acetate ester (with similar odour characteristic), in each (about 0.1 mg L−1) but inclusion of this figure would not alter the conclusion. If the ‘banana’ esters content is added to the fusel oil content (i.e. C3–C5 alcohols + 2-phenylethanol) of each, there is, however, now a marked difference of 1 FU between the sample and null beer.

In relating flavour-to-flavour units, Meilgaard (1982) set out the following groups:

(1) Primary flavour constituents (‘primary’ used here in the sense of effective intensity, not origin). These are present at above 2 FU, so that removal of any such constituent in a given beverage would produce a decisive change of flavour.

(2) Secondary flavour constituents are those that are present between 0.5 and 2 FU; removal of any one constituent from this group would produce a minor change in flavour.

(3) Tertiary flavour constituents occur at levels of 0.1–0.5 FU where they cannot be perceived individually, so that removal of any one would not produce a perceptible flavour change.

(4) Background flavour constituents are those that are present below 0.1 FU, often the majority of constituents present.

In using this technique, Meilgaard (1982) also considered the question of interactions between different constituents in respect of different threshold values, by introducing an interaction factor (d int) where combinations of individual constituents were being examined. Where flavour effects are additive, d int = 1, but where synergism exists, d int will be >1.00; and values less than 1.00 are possible. In general, he found that where constituents had similar flavours, such as the banana esters and the apple esters (ethyl hexanoate and ethyl octanoate) there was little apparent synergism. For octanoic acid and ethyl acetate, with such distinctive flavours, suppression was occurring (i.e. d int = 0.54).

Only three components in beers, hop bitters, ethanol and carbon dioxide, were considered likely to produce FU values of about 2.0 in general. Off-flavour substances when present will come into this category. It is evident, however, that if we would have 20 similarly flavoured but different chemical compounds (thus ‘fruity esters’), each with an individual level near 0.1 FU, their overall FU would be 2 or more.

Meilgaard & Peppard (1986) recognized the rough-and-ready nature of the calculations involved and stated that it could not be expected to cope with the more complex aspects of flavour, such as the effect of changing yeast strain or of oxidation, which affect many compounds and groups of compounds simultaneously. This work is of potential interest in the study of wine flavour, where the composition of the various groups of substances described in Chapter 4 shows that very few individual aroma compounds are likely to show high flavour unit values.

5.7.2 Odour activity unit

Grosch (2001) showed some very high ratios of the Aroma Index, which he called Odour Activity Value, based upon threshold values on an absolute basis in pure water.

The concentrations of aroma substances in coffee brews are often much higher than in wines, e.g. thus furaneol at 7.2 mg L−1 and 3-methylbutanal at 0.57. Some very high individual OAVs have been recorded, particularly amongst sulfur compounds, like 3-methyl-2-butan-1-thiol at 2000-fold. As in wine, β-damascenone is a recently discovered component at a low content level, though at some 2130 times the determined threshold in water. These two substances will clearly play an important rôle in the flavour of the brew. OAVs were not, however, used directly to determine which volatile compounds were determining flavour.

Grosch (2001) has described laboratory procedures to determine the really significant contributors in coffee brews. Roasted coffee has even more known volatile compounds than wines, over 800, including some 80 pyrazines generated by roasting. There is perhaps a lesser complexity, in that attention can be primarily directed to one roasting level (medium), two species, some four varieties/cultivars and two methods of green bean preparation, rather than the wide range of wines in the market place. Grosch and his colleagues (2001) examined distillates (by the Likens–Nickerson method, using ethyl ether or a low temperature extractant) and determined that there were only some 28 really important odoriferous substances present. They then prepared model aroma distillates and compared the aroma intensity with an actual coffee distillate (from medium roasted Colombian arabica coffee) by flavour panel testing. Omission experiments (i.e. by omitting one or more components from the model) together with other techniques enabled fine honing into a model aroma closely comparable with an authentic coffee aroma (both from a brew and the roasted coffee itself). Quantitative data on the content of each of these 28 compounds, together with their odour threshold values in water, was presented. All these compounds were tabulated according to six of the odour groups described earlier in this chapter. It is of interest that a number of these compounds are also the ones described as present in wines – the various furanones, alkanals, and even isobutyl-methoxy-pyrazine. Like wines, some of these compounds come through from the raw bean or grape but in coffee most are generated by the roasting process and in wine by the fermentation.

Interestingly, the earlier but reliable quantitative GC-MS work of Silwar et al. (1987) on roasted coffee volatile compounds showed that about 800 mg of coffee distillate was obtainable from 1 kg of a medium roast arabica coffee. This distillate was made up of some 38–45% of furans of all kinds (including sulfur derivatives), 25–30% pyrazines and only 3–5% aliphatic compounds, in the many hundreds of separately identified compounds; but no connection was attempted as to any relationship with odour characteristics or activity values. Grosch (2001) in his investigations shows some 500 mg distillate, containing 28 different compounds, screened out as being the important odorants for roast coffee, of which certainly 28% by weight can be seen to be furans of different kinds (including furanones and thiols). As already mentioned, most of these compounds had OAVs determined in an actual brew from the coffee of well over ten, and several compounds had very high OAVs. Grosch used absolute values in pure water for thresholds; no difference values in coffee brews appear available. Compounds such as the many thiazoles and oxazoles were not seen to be important, as also all of the simpler pyrazines, identified as being present and thought very important in many earlier studies.

A review by Francis & Newton (2005) considered recent publications with reliable identifications of volatile aroma compounds in wine, including quantification and threshold values. Based on this literature survey they compiled a list with compounds in wines all with OAV greater than one, consisting of the compounds listed in Table 5.9.

Of course, interactions between compounds in wines as well as inherent difficulties in determining accurate threshold values will influence these data. Only a subset of compounds may be present in wines but these compounds may typically contribute to the aroma of a wine.

Beverages, including wines and spirits, have been happy hunting grounds for chemists wishing to flex their analytical muscle to the full, whereas the odour significance of many of these compounds has been found subsequently to be minimal. No comparable work in narrowing down the really significant contributors to the flavour of each important wine, such as those from Cabernet Sauvignon and Chardonnay grapes, appears to have been published.

5.7.3 Multivariate and other statistical procedures

As has been noted in Section 5.1, the number of variables involved in assessing the flavour of wine can be very large, both in the elements describing the flavour of a wine and in the contributing factors to that flavour from grape variety, growing location and vinification practices. By the use of so-called multivariable statistical procedures and techniques, connections can be made between such sets of variables, see the review by Noble & Ebeler (2002). These procedures have been used in the last few decades for a number of different beverages. There have recently been a number related to wines; for example, Vernin et al. (1993) have used statistical techniques (PCA) to classify Bandol Mourvèdre wines from various vintage years (1986–1988) on the basis of GC data from head-space analysis of some 40 different aroma compounds. In Vernin’s work, the marked significance of such compounds as ethyl propanoate, isobutyl acetate, the methyl butanols, ethyl octanoate, diethyl succinnate is of interest. Multiple regression analysis is frequently the method of choice for prediction of specific flavour notes from measurements of volatile compound content. Such a technique has been described for the volatile compounds in tea by Togoni (1998); thus, the intensity of a ‘sweet, floral note’ could be predicted from compound concentrations,

= - 0.0591[pentanal] - 0.671[2-heptanone] + 0.562[linalool]

+ 0.693[2-phenyl ethanol] + 0.0713[Jasmine lactone] - 0.134

for which reference to the original publication is necessary.

Table 5.9 List of compounds in wines with an OAV greater than 1.

No. Chemical groups Compounds
7 Ethyl esters Ethyl isobutyrate, Eethyl 2-methylbutyrate, ethyl isovalerate, ethylhexanoate, ethyl octanoate, ethyl decanoate
3 Acetates Isoamyl acetate, phenylethyl acetate, ethyl acetate
2 Cinnamic esters Ethyl dihydrocinnamate, trans-ethyl cinnamate
8 Acids Isobutyric acid, isovaleric acid, acetic acid, butyric acid, propanoic acid, hexanoic acid, octanoic acid, decanoic acid
6 Alcohols Isobutanol, isoamyl aclohol, 2-phenylethyl alcohol, methionol, 1-hexanol, (Z)-3-hexenol
4 Monoterpenes Linalool, geraniol, cis-rose oxide, wine lactone
6 Phenols Guiacol,4-ethylguiacol, eugenol, 4-vinylguiacol, 4-ethyl phenol, vannillin
7 Lactones cis-Oak lactone, γ-nonalactone, γ-decalactone, γ-dodecalactone, 4-hydroxy-2,5-dimethyl-3-(2H)-furanone, (Z)-6-dodecanoic acid-γ-lactone, sotolon
2 Norisoprenoids β-Damascenone, β-ionone
7 Sulfur compounds 3-Mercaptohexyl acetate, 4-mercaptomethyl pentan-2-one, 3-mercaptohexanol, 2-methyl-3-furanthiol, 3-methyl thio-1-propanol, benzenemethanethiol, dimethyl sulfide
6 Other 2,3-Butanedione, acetoin, 3-isobutyl-2-methoxypyrazine, acetaldehyde, phenyl acetaldehyde, 1,1-dioxyethane

Francis & Newton (2005).

The use of statistics linking sensory data from trained taste panels and chemical analytical data from gas chromatography olfactometry are now increasingly linked together using statistical methods such regression analysis, partial least squares or principal components analysis trying to model the data in order to explain the the differences between the wines in chemical terms. Often the end result of the analyses is between two and four axis on which the data can be pictured showing the largest amount of variation between the samples. A predictive model can be the end result. Usually teams of scientists with an array of different skills need to work together, to ensure accurate chemical and sensory analytical data are collected on a well selected sample set. However, these methods have their limitations. The wines used for such an experiment need to have a wide range of sensory different properties and the resulting statistical model is only valid for wines with properties comparable to the ones in the original data set. However, without statistical methods it would be extremely difficult to draw any conclusions from such large and complicated data sets, and the application of statistical methods can lead to being able to unmask interesting information. Numerous studies have been done, usually confirming the significance of a relatively small number of volatile aroma compounds in their contribution to wine aroma of the sample set, some are briefly discussed by Francis & Newton (2005).

Table 5.10 Compounds contributing significantly to quality of 25 Spanish red winesa.

Compounds contributing fruity or sweet characteristics Off-flavours contributing negatively to quality Compounds which possibly suppress the fruity character of wines
Propyl acetate 3,5-Dimethyl- Metionol
2- methoxypyrazine
2,3-Butanedione 4-Ethylphenol Metional
Isobutyl acetate 2,4,6-Trichloroanisol (Z)-2-Nonenal
Ethyl butyrate 3-Ethylphenol (E,E)-2,4-Decadienal
Ethyl 2-methylbutyrate 4-Ethylguiacol 3-Isopropyl-2-methoxypyrazine
2,3-Pentanedione o-Cresol Acetic acid
Ethyl 3-methylbutyrate 2-Methylisobomeol
Isoamyl acetate
Ethyl 2-methylpentanoate
Ethyl 3-methylpentanoate
Ethyl 4-methylpentanoate
Ethyl hexanoate
β-Damascenone
2,5-Dimethyl-4-hydroxy-3(2H)-furanone (Furaneol)

a Information from Ferreira et al. (2009).

One such study for example (Ferreira et al., 2009) used data obtained on 25 Spanish red wines analysed by descriptive analyses using a trained panel and data collected by quantitative gas chromatography olfactometry. Using correlation methods and partial least squares regression models, they found that the regression model containing three vectors explained 78% of the variation between the wines. Mainly 15 fruity and sweet smelling aroma compounds significantly contributed to the quality of this data set (Table 5.10). The presence of any off-flavour had a significant negative effect on the overall quality. However, the presence of nine compounds with negative sensory attributes in very low concentrations also could significantly affect the perceived quality of a wine. The authors concluded that the presence of even very low concentrations of odorants with negative sensory attributes may strongly suppress the fruity characteristics of wines. Hence is may not just be the presence of positively contributing aroma volatiles, but the absence of any off-odours that exert important sensory quality attributes to wines.

Bibliography

Barr, A. (1988) Wine Snobbery. Faber and Faber, London.

Belitz, H.D. & Grosch, W. (1986) Food Chemistry, 2nd edn. Springer Verlag, Berlin.

Clarke, O. & Rand, M. (2001) Grapes and Wines, Webster, London.

Goolden, J. (1990) Wine Tasting, loc. cit.

Grainger, K. (2009) Wine Quality Tasting and Selection, Blackwell Publishing Ltd., Oxford.

Jackson, R.S. (2008) Wine Science, Principles, Practice, Perception. 2nd edn. Academic Press, San Diego.

Jackson, R.S. (2009) Wine Tasting: A Professional Handbook. 2nd edn. Academic Press, San Diego.

Labows, J.N. & Cagan, R.H. (1993) Complexity of flavour recognition and transduction. In: Food Flavour and Safety; Molecular Analysis and Design (eds. A.M. Spanier, H. Okai & M. Tamura), pp. 9–27. American Chemical Society, Washington.

O’Mahony, M. (1985) Sensory Evaluation of Foods (Food Science and Technology 16). Marcel Dekker Inc., New York.

Rapp, A. & Marais, J. (1993) Changes in aroma substances during storage. In: Shelf-Life Studies of Foods and Beverages (ed. G. Charalambous), Developments in Food Science, 33, pp. 891–922. Elsevier, Amsterdam.

Ribéreau-Gayon, P., Glories, Y., Maujean, A. & Dubourdieu, D. (2006) Handbook of Enology, Volume I, Wine and Wine Making. John Wiley & Sons, Ltd, Chichester.

Robinson, J. (1993, 1995) Jancis Robinson’s Wine Course. BBC Books, London.

Teranishi, R., Buttery, R.G., Stern, D.J. & Takeoka, G. (1991) Use of odour thresholds in aroma research. Lebensmittel Wissenschaft und Technologie, 24, 1–5.

Vernin, G., Pascal-Mousselard, H., Metzger, J.L. & Parkonyi, C. (1993) Classification of wines originating from AOC bandol using multi-variate statistical methods. In: Shelf-Life Studies of Foods and Beverages (ed. G. Charalambous), pp. 975–990. Elsevier, Amsterdam.

References

Amoore, J.E. (1977) Specific anosmia and the concept of primary odours. Chemical Senses and Flavour, 2, 267–281.

Barr, A. (1988) Wine Snobbery. Faber and Faber, London.

Bartoshuk, L. (1993) Genetic and pathological taste variation: what can we learn from animal models and human disease? In: The Molecular Basis of Smell and Taste (eds. D. Chadwick, J. March & J. Goode), pp. 251–267. John Wiley & Sons, Ltd, Chichester.

Bastian, S.E.P., Collins, C. & Johnson, T.E. (2010) Understanding consumer preferences for Shiraz wine and Cheddar cheese pairings. Journal of Food Quality and preference (IN PRESS)

British Standard Glossary of Terms relating to sensory analysis (1992). BS.5098, British Standards Institution, Chiswick, London.

Broadbent, M. (1979) Wine Tasting. Christies’ Wine Publications, London.

Buck, L. & Axel, R. (1991) A novel multigene family may encode odorant receptors: A molecular basis for odor recognition. Cell, 65, 175–187.

Clarke, O. (1986, 1989, 1998, 2001) Wine Fact Finder. Websters, London.

Clarke, O. (2003) Introducing Wine. Websters, London.

Clarke, O. & Rand, M. (2001) Grapes and Wines. Websters, London

Delwiche, J. (2004). The impact of perceptual interactions on perceived flavor. Food Quality and Preference, 15, 137–146.

Ferreira, V., San Juan, F., Escudero, A., Cullere, L., Fernandez-Zurbano, P., Saenz-Navajas, M.P. & Cacho, J. (2009) Modeling quality of premium Spanish red wines from gas chromatography – Olfactometry data. Journal of Agriculture and Food Chemistry, 57, 7490–7498.

Francis, I.L. & Newton, J.L. (2005) Determining wine aroma from compositional data. Australian Journal of Grape and Wine Research, 11, 114–126.

Genovese, A., Piombino, P., Gambuti, A. & Moio, L. (2009) Simulation of retronasal aroma of white and red wine in a model mouth system. Investigating the influence of saliva on volatile compound concentrations. Food Chemistry, 114, 100–107.

Grosch, W. (2001) Volatile compounds. In: Coffee: Recent Developments (eds. R.J. Clarke and O.G. Vitztham), pp. 68–89. Blackwell Publishing Ltd., Oxford.

Guti, H., Blehr, K. & Fritzler, R. (2001) Descriptors for structive-property correlation studies of odorants. In: Aroma Active Compounds in Foods (eds. M. Gundert & K.-H. Engel) ACS Symposis Series No. 794, pp. 93–108. American Chemical Society, Washington.

Heath, B. (1988) The physiology of flavour, taste and aroma perception. In: Coffee Vol. 3 Physiology (eds. R.J. Clarke & R. Macrae), pp. 141–170. Elsevier Applied Science, London.

Hughson, A.L. & Boakes, R.A. (2002) The knowing nose; the rôle of knowledge in wine expertise. Journal of Food Quality and Preference, 13, 463–472.

Isci, B., Yildirim, H.K. & Altindisli, A. (2009) A review of the authentication of wine origin by molecular markers. Journal of the Institute of Brewing, 115, 259–264.

Jackson, R.S. (2008) Wine Science, Principles, Practice, Perception. 2nd edn. Academic Press, San Diego.

Johnson, H. (1979) Pocket Wine Book, Mitchell Beazley, London.

Land, D.G. (1989) Taints – causes and prevention. In: Distilled Beverage Flavour (eds. J.R. Piggott & A. Paterson), pp. 17–32. Ellis Horwood, Ltd., Chichester.

Maruniak, J.A. & Mackay-Sim, A. (1985) The sense of smell. In: Sensory Analysis of Foods (ed. J.R. Piggott), pp. 23–57. Elsevier Applied Science, Barking.

McQuitty, J. (1997–98) Saturday articles on wine in The Times newspaper, London.

Meilgaard, M.C. (1982) Prediction of flavour differences between beers from their chemical composition. Journal of Agricultural and Food Chemistry, 30, 1009–1017.

Meilgaard, M.C. & Peppard, T.L. (1986) The flavour of beer. In: Developments in Food Science 33, Food Flavours, Part B, The Flavour of Beverages (eds. I.D. Morton & A.J. Macleod), loc. cit., pp. 99–170.

Moulton, D.G. (1982) Sensory basis and perception of flavour. In: Food Flavours, Part A, Introduction (eds. I.D. Morton & A.J. Macleod), pp. 1–13. Elsevier, Amsterdam.

Noble, A.C., Williams, A.A. & Langron, S.P. (1983) Descriptive analysis and quality of Bordeaux wines. In: Sensory Quality in Foods and Beverages (eds. A.A. Williams & R.K. Atkin), pp. 324–334. Ellis Horwood, Ltd., Chichester.

Noble, A.C., Arnold, R.A., Buechsenstein, J., Leach, E.J., Schmidt, J.O. & Stern, P.M. (1987) Modification of a standardized system of wine aroma terminology. American Journal of Enology and Viticulture, 36, 143–146.

Noble, A.C. & Ebeler, S.E. (2002) Use of multivariate statistics in understanding wine flavor. Food Reviews International, 18, 1–21.

Palaskova, P., Herszage, J. & Ebeler, S.E. (2008) Wine flavor: chemistry in a glass. Chemical Society Reviews, 37, 2478–2489.

Peynaud, E. (1986) The Taste of Wine. Macdonald, Orbis, London.

Piggott, J. (ed.) (1984) Sensory Analysis of Foods. Elsevier Applied Science, Barking.

Piggott, J. (ed.) (1986) Statistical Procedures in Food Research. Elsevier Applied Science, Barking.

Pratt, C., Ruiz-Rueda, O., Trias, R., Anticó, D.C., Capone, D., Sefton, M. & Bañeras, L. (2009) Molecular fingerprinting by PCR-denaturing gradient gel electrophoresis reveals differences in levels of microbial diversity for musty-earthy tainted corks. Applied and Environmental Microbiology, 75(7), 1922–1931.

Ribéreau-Gayon, P. (1978) Wine flavour. In: The Flavour of Foods and Beverages (eds. G. Charalambous & G.E. Inglett), pp. 355–380. Academic Press, New York.

Ribéreau-Gayon, P., Glories, Y., Maujean, A. & Dubourdieu, D. (2006) Handbook of Enology, Volume I, Wine and Wine Making. John Wiley & Sons, Ltd, Chichester.

Robinson, J. (1993, 1995) Jancis Robinson’s Wine Course. BBC Books, London.

Sefton, M.A. & Simpson, R.F. (2005) Compounds causing cork taint and the factors affecting their transfer from natural oak closures to wine – a review. Australian Journal of Grape and Wine Research, 11, 226–240.

Silwar, R., Kempschoerer, M. & Tressl, R. (1987) Gas-chromatographic-mass spectrometry-quantitative determination of steam-volatile aroma constituents. Chemistry Microbiolology Technology Lebensm., 10, 176–87 [in German].

Snowdon, E.M., Bowyer, M.C., Grbin, P.R. & Bowyer, P.K. (2006) Mousy off-flavour: a review. Journal of Agriculture and Food Chemistry, 54, 6464–6474.

Stevens, J.C., Cain, W.S. & Burke, R.J. (1988) Variability of olfactory thresholds. Chemical Senses, 13, 643–653.

Suarez, R., Suarez-Lepe, J.A., Morata, A. & Calderon, F. (2007) The production of ethylphenols in wine by yeasts of the genera Brettanomyces and Dekkera: A review. Food Chemistry, 102, 10–21.

Togoni, A. (1998) Tea flavour. American Chemical Society Symposium Series (No. 173). American Chemical Society, Washington.