Margaret A. Cliff

Sensory evaluation as a selection tool in apple breeding

Several conventional sensory methods were adapted to provide aprocedure that is suitable for screening apple (Malus × domestica) breeding selections for dessert quality. Trained judges were presented with randomized coded samples (apple slices) and asked to rate them on 0 to 9 bipolar hedonic (liking) scales for texture and flavour, and 0 to 9 unipolar intensity scales for skin toughness, crispness, hardness, juiciness, aroma, sweetness and sourness. Appearance liking was rated on coded samples of five whole apples, using the 0 to 9 hedonic scale. A minimum panel size of eleven judges was generally sufficient to obtain statistical discrimination of one point on the 0 to 9 scales. Panel mean scores for breeding selections relative to standards have been consistent from panel to panel and year to year. A subset of selections and cultivars was rated for appearance and taste by consumers in blind taste tests. In-house panel findings were comparable to consumer ratings for taste and appearance liking, with a few exceptions in appearance. Panel mean scores for texture and flavour liking were regressed on texture and flavour components. Crispness accounted for about 90% of the variation in texture liking. Juiciness, aroma, sweetness and sourness were all important to flavour liking, but their relative importance changed from year to year and in total accounted for only about 60% of variation in flavour liking. Perceived sweetness and sourness were better predictors of liking than analytical measurements of soluble solids and titratable acidity. Formal sensory evaluation can be used successfully for screening breeding selections, and may provide more reliable data than the opinions of only one or two people.

Sensitization and desensitization to capsaicin and menthol in the oral cavity: Interactions and individual differences

It was reported in a recent study that, like capsaicin, menthol is capable of producing a desensitization to sensory irritation in the oral cavity. Whereas capsaicin is known to be able to cross-desensitize with other chemical irritants, no such information exists for menthol. To address this question, the first experiment was designed to reveal whether cross-desensitization would occur between menthol and capsaicin. After a pretest on the tongue tip in which subjects rated the intensity of irritation and cold produced by 3.5 ppm capsaicin or 0.3% l-menthol, five samples of the same stimuli were sipped and swished at 1-min intervals for 5 min. Fifteen minutes later subjects were tested on the tongue tip with either capsaicin or menthol. The results 1) confirmed self-desensitization for both chemicals, 2) demonstrated cross-desensitization of menthol by capsaicin, and 3) revealed cross-sensitization of capsaicin by menthol. This series of outcomes suggests that menthol produces much of its sensory irritation via capsaicin-sensitive pathways, but that it excites and/or desensitizes those pathways via different mechanisms than does capsaicin. Analysis of the individual data revealed large differences in sensitization, and desensitization that were significantly correlated across chemicals, which suggests the possibility that the perceptual response to repeated exposures to irritants may be idiosyncratic. Contrary to earlier findings, the first experiment also revealed apparent self- and cross-desensitization of the menthol sensation of coolness. The latter outcome was investigated in a second experiment in which the effect of capsaicin desensitization on the perception of physical as well as chemical (menthol) cooling was measured when the stimuli were presented as oral rinses. No desensitization was found for either form of stimulation, which implied the apparent desensitization of coolness in Experiment I may have been due to the difficulty of discriminating sensations of cold from sensations of chemical irritation. The overall findings are discussed in terms of the complex sensory and perceptual interactions that take place within the chemesthetic modality.

Vinification effects on the sensory, colour and GC profiles of Pinot noir wines from British Columbia

Abstract Analytical and sensory profiles were obtained for Pinot noir wines fermented with two yeasts (Lalvin EC-1118, P; Lalvin L-2056, L) and four vinification techniques: high (T), ambient, cold and modified-cold temperature (M) fermentations. Wines were analysed for titratable acidity, pH, ethanol, colour, total phenolics, flavonols, tartaric esters, anthocyanins and volatile constituents. Sensory descriptive analysis was conducted using 12 judges to evaluate colour, aroma, flavour, body and overall quality. Wines from the M vinification had the most red colour, body, tropical fruit and spicy aromas and had the least vegetal character. Wines from the T vinification were the most vegetal. Increased ester concentration (mainly acetates) and total volatile compounds were found in wines from the M vinification with yeast P in comparison to the other vinification methods and yeast L. Principal component analyses distinguished four distinct groups of wines based on 33 main volatile constituents. Partial least square regressions revealed strong relationships between sensory and instrumental colour measurements in terms of red colour intensity ( r =0.88) and red hue ( r =0.87).

Canadian terroir: characterization of Riesling wines from the Niagara Peninsula

Abstract The Niagara Peninsula (Ontario, Canada) is thought to have distinct viticultural areas (‘bench’, ‘plains’, ‘lake’), but the uniqueness of wines from these regions has not been documented. Therefore this research was undertaken to document the nature and magnitude of the terroir differences using 14 commercial Riesling wines from producers who utilized grapes from the ‘Beamsville Bench’ (‘bench’) and the ‘Niagara Plains’ (‘plains’) regions. The sensory characteristics of the wines were profiled using descriptive analysis (DA) coupled with univariate and multivariate statistics. Six trained judges evaluated 10 aroma (pineapple, melon, grapefruit, lemon/lime, apple, peach/apricot, rose/floral, honey, diesel/petrol, mineral/flint), six flavour (pineapple, melon, grapefruit, lemon/lime, peach/apricot, honey), two taste (acidity, sweetness), and three other (alcohol, body, finish) sensory attributes. Data were analyzed using analysis of variance, step-wise discriminant analysis (SDA) and principle component analysis (PCA). Wines from the ‘bench’ had significantly higher grapefruit, pineapple, melon and lemon/lime aromas, more acidity and greater lemon/lime flavour than the ‘plains’ wines. ‘Plains’ wines were more diesel/petrol-like in character. A PCA of the data, using the most relevant sensory descriptors (as determined by SDA), indicated that 68.7% of the variability could be explained in the first two dimensions, with factor one accounting for the lemon/lime, grapefruit, melon and acidity of the wines and factor two representing the mineral/flint character.

Sensory and compositional profiles of British Columbia Chardonnay and Pinot noir wines

Abstract Sensory and compositional analyses were used to profile (i) 1993 Chardonnay wines ( n = 16), (ii) 1988–1993 Chardonnay wines ( n = 25) and (iii) 1992–1993 Pinot noir wines ( n = 14) from British Columbia (B.C.). All wines were analyzed for titratable acidity, pH, absorbance, phenol and alcohol content. In a series of three experiments, 10 judges evaluated colour, aroma and flavour attributes of all wines in duplicate. Principal component analysis of the sensory data for 1993 Chardonnay wines indicated a diversity of styles including wines which were both oaky and buttery to those which were intensely fruity. Wines with oaky character had greater complexity and deeper colour, while those with fruity character were sweeter and more perfumed. Discriminant analysis indicated that, in general, the 1992–1993 vintages were more fruity than the 1988–1991 vintages which were more yellow and had higher oak and diacetyl characters. Discriminant analysis also showed specific winery “styles”. While some wineries produced fruity, floral wines and others produced wines which were more yellow, oaky, buttery and astringent. Principal component analysis of the sensory data for the Pinot noir wines indicated four major styles of B.C. wines: those which were complex with high fruit and berry flavours, those which were intensely red, oaky, astringent and spicy, those which were herbaceous, and those which lacked fruit/berry character, astringency and body.

Development and evaluation of multiple regression models for prediction of sweet cherry liking

Abstract In two experiments, multiple regression models were developed and evaluated to identify the relevant sensory attributes for cherry liking. In Experiment 1, 16 judges evaluated 18 cherry varieties for seven visual characteristics (colour intensity, uniformity-of-colour, speckles, size, stem length, external firmness and ‘visual’ liking) and seven flavour/texture characteristics (flesh firmness, flesh colour intensity, juiciness, sweetness, sourness, flavour intensity and ‘flavour/ texture’ liking). Stepwise multiple regression was used to develop the most appropriate statistical models for prediction of visual and flavour/texture liking based on visual and flavour/texture characteristics, respectively. Both models were simple and easily understandable with two sensory variables. The best model for visual liking required only size and uniformity-of-colour variables; whereas, the best model for flavour/texture liking required sweetness and flavour intensity variables. In Experiment 2, 18 judges evaluated 30 sweet cherry cultivars, using the same methodology, to create a validation data set. Correlation coefficients ( R ) and prediction standard errors (PSEs) between the observed (Experiment 2) and predicted (Experiment 1) liking scores were used to evaluate the prediction equations. The prediction equation for flavour/texture liking was most satisfactory ( R = 0.85, PSE = 0.61). A new equation developed from the validation data confirmed the importance of sweetness and flavour intensity. In contrast, the prediction equation for visual liking was less satisfactory ( R = 0.56) and a new equation developed from the validation data set confirmed only size as an important variable.

Determination of odour detection thresholds for acetic acid and ethyl acetate in ice wine

Abstract Collectively acetic acid and ethyl acetate are responsible for ‘volatile acidity’ (VA) in wine. The detection limit or threshold for these compounds is well documented in table wine but not for ice wine. Knowledge of the ice wine thresholds is important for understanding perception limits and setting legal standards, particularly for a product with high intrinsic concentrations. Thresholds were determined for each compound using seventeen subjects and an ascending series of paired comparison tests, consisting of five concentrations. The detection threshold, the concentration at which there is 75% correct detection, was determined using least-squares linear regression. The correlation coefficients for the linear regressions for acetic acid and ethyl acetate were r = 0.996 (p < 0.001) and r = 0.972 (p < 0.001), respectively. The thresholds for acetic acid and ethyl acetate were 3.185 and 0.198 g/l, respectively. The threshold for acetic acid was approximately three times that found in table wines, but was relatively unchanged for ethyl acetate. This work supports the need for the legal limit for VA to be higher in ice wine, but also suggests that the legal requirements for the two compounds should be specified independently, not together.

A proposed approach for evaluating expert wine judge performance using descriptive statistics

Descriptive statistics were used to evaluate the performance of five expert wine judges using a 20‐point quality scorecard in a wine competition environment. For each judge, frequency distributions were used to assess scorecard usage, and deviation font‐means were calculated as an index of consistency with the group. Differences in curve shapes were described using means, skewness and kurtosis values. Four of the five judges showed unimodal frequency distributions that were slightly skewed to lower scores. This information was believed to be useful for identifying expert judges whose use of a 20‐point quality scorecard and perspective of wine quality are consistent with the group.

Use of principal component analysis for the evaluation of judge performance at wine competitions

Abstract Principal component analysis (PCA) was used to evaluate the performance of five wine judges using a 20‐point quality scoring system. Data were collected for 97 wines as part of a wine competition. Principal component plots were used to (1) identify underlying mechanisms that judges used to differentiate wine quality, and (2) evaluate the relationships among the quality components. Judges’ ‘styles’ differed according to their use, or lack of use, of the individual quality components (appearance, aroma, flavour, acidity, finish, body, balance and overall character) in making a final quality assessment of the wines. In general, judges showed individual patterns of separating the wines according to the presence or absence of aroma and flavour and overall character. These component scores were highly correlated with each other and with the final wine quality score. A comparison of the PCA plots with eggshell plots, an established methodology, indicated that judges who used a systematic approach to qua...

The evaluation of judges at wine competitions: the application of Eggshell plots

Abstract Eggshell plots were used to evaluate wine judge performance at two annual wine competitions. Judge curves were examined on a daily basis to identify similarities and differences among the judges. These techniques are believed to be an important complement to existing techniques to examine judge performance and assist competition co‐ordinators in future judge selection.