Citation
Flavor of fresh market tomato (Lycopersicon esculentum Mill.) as influenced by harvest maturity and storage temperature

Material Information

Title:
Flavor of fresh market tomato (Lycopersicon esculentum Mill.) as influenced by harvest maturity and storage temperature
Creator:
Maul, Fernando, 1973-
Publication Date:
Language:
English
Physical Description:
viii, 190 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Acidity ( jstor )
Electronics ( jstor )
Flavors ( jstor )
Fruits ( jstor )
Low temperature ( jstor )
Pericarp ( jstor )
pH ( jstor )
Ripening ( jstor )
Sensors ( jstor )
Tomatoes ( jstor )
Dissertations, Academic -- Horticultural Science -- UF ( lcsh )
Horticultural Science thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 178-189).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Fernando Maul.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
030361497 ( ALEPH )
41954008 ( OCLC )

Downloads

This item has the following downloads:


Full Text










FLAVOR OF FRESH MARKET TOMATO (Lycopersicon esculentum Mill.) AS INFLUENCED BY HARVEST MATURITY AND STORAGE TEMPERATURE














By

FERNANDO MAUL













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1999































To my family in Guatemala

To Vicky














ACKNOWLEDGMENTS

First and foremost, my most sincere gratitude and admiration to Dr. Steven A. Sargent for his guidance, support and enthusiasm as the chairman of my advisory committee. To the members of my advisory committee, Dr. Elizabeth A. Baldwin, Dr. Charles A. Sims, Dr. Donald J. Huber and Dr. Murat O. Balaban, my deepest gratitude for allowing me and my seemingly endless endeavors into your laboratories.

I am very grateful to Diego A. Luzuriaga, Margery Einstein, Dr. J. W. Scott, and Mr. Kenneth Shuler for their valuable contributions and help. I would like to thank the members of my trained descriptive sensory panel, whose motivation and attendance helped me enormously.

I am thankful to Abbie J. Fox, James Lee, Holly Sisson and Bob Golaszewski. for all their help and to Dr. H. Klee for allowing storage of my samples in his limited freezer space. Very special thanks to my friends in the Postharvest laboratory Pornchai P., Cleisa Cartaxo, Domingos Almeida, Ernesto Brovelli, Celso Moretti, and Shahab Hanif-Khan. Finally, I would like to express my overwhelming gratitude towards my family and my wife, who have unconditionally stood by me throughout my graduate studies at UF.











111















TABLE OF CONTENTS



Page

ACKNOWLEDGMENTS ................................................................... iii

LIST OF ABBREVIATIONS ............................................................. vi

ABSTRACT ................................................................................. vii

CHAPTERS

1 INTRODUCTION ......................................................... 1

2 LITERATURE REVIEW.................................................. 5

Tomato Fruit Ripening .................................................... 5
Ripe Tomato Composition............................................. 7
Ripe Tomato Flavor ....................... ................................ 10
Lipid-derived Volatile Compounds.................... ............ 12
Carotenoid-derived Volatile Compounds ............................ 13
Sensory Flavor Perception ................................................14
Sensory Analysis ...................................................... . 16
Electronic Aroma Sensing Technology.................................. 20
Harvest Maturity and Tomato Flavor .................................... 22
Postharvest Treatments and Tomato Flavor ............................. 23
Storage Temperature Management ................................... 24
Research Objectives ........................................................ 26


3 RIPENESS STAGE AT HARVEST AFFECTS CHEMICAL
COMPOSITION AND COLOR QUALITY OF
TABLE-RIPE TOMATOES ......................................... 27


Introduction ................................................................. 27
Materials and Methods ....................................................28
Results and Discussion ............................................. .... 31
Conclusions............................................................... 42



iv











Page

4 POTENTIAL FOR NONDESTRUCTIVE QUALITY
SCREENING OF TOMATOES WITH ETHYLENE OR
ELECTRONIC NOSE SENSOR ARRAYS ........................ 45

Introduction ................................................. ............. 45
Materials and Methods .................................................... 47
Results and Discussion ...................................................51
Conclusions ........................................................... .... 60


5 AROMA VOLATILE PROFILES AND RIPE TOMATO
FLAVOR ARE INFLUENCED BY PHYSIOLOGICAL
MATURITY AT HARVEST ............................ ........... 67

Introduction ............................................................. 67
Materials and Methods .................................................... 68
Results and Discussion .................................................... 77
Conclusions ................................................................. 105


6 HARVEST MATURITY AND STORAGE TEMPERATURE
AFFECT TOMATO VOLATILE PRODUCTION AND
FLAVOR ............................................................. ..... 118

Introduction ................... . ......................................... 118
Materials and Methods ................................................... 120
Results and Discussion.................................................... 124
Conclusions ....................................................... ...... 155


7 CONCLUSIONS......................................................... 174


REFERENCES .............................................. ............................. 178

BIOGRAPHICAL SKETCH ...................................................................... 190











V














LIST OF ABBREVIATIONS



CI Chilling injury DCIP 2,6-dichlorophenolindolphenol DNPH Dinitrophenylhydrazine EN Electronic nose GC Gas chromatography HPLC High performance liquid chromatography IG Immature-green stage MD Mahalanobis distance MG Mature-green stage MVDA Multivariate discriminant analysis M1 Immature-green M2 Partially mature-green M3 Mature-green M4 Advanced mature-green OUV Odor unit value QDA Quantitative descriptive analysis SSC Soluble solids content TA Titratable acidity T301 Sensor type-301



vi














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

FLAVOR OF FRESH MARKET TOMATO (Lycopersicon esculentum Mill.) AS INFLUENCED BY HARVEST MATURITY AND STORAGE TEMPERATURE By

Fernando Maul

May 1999

Chairman: Dr. Steven A. Sargent
Major Department: Horticultural Sciences

The effects of harvest maturity and storage temperature on fresh tomato (Lycopersicon esculentum Mill.) chemical composition and sensory quality at ripe stage were investigated. Tomatoes from seven commercial cultivars harvested at green stage were exposed to exogenous ethylene (100 lpL/L) to accelerate the onset of ripening for 1 to 7 days until they attained breaker stage (<10% red color). There was a strong relationship (r2 = 0.84) between green-tomato maturity at harvest and ethylene exposure time to attain breaker stage. Immature-green (MI) tomatoes had to be treated with ethylene for >3 days compared to mature-green (M4) tomatoes which were treated for 1 day. Ripe tomato color parameters (L* and a* values) consistently increased with increasing ethylene requirement, indicating development of less intense-red coloration with immature-harvested fruit.




vii








Untrained sensory panelists found significant differences in ripe tomato flavor between fruits exposed to ethylene for 1, 3 or 5 days to attain breaker stage. Descriptive sensory panelists determined that at ripe stage green-harvested tomatoes exposed to <3 days of ethylene treatment were comparable in ripe aroma, sweetness and tomato flavor to tomatoes harvested at light-red stage. In contrast, tomatoes that reached breaker stage after 3 days of ethylene treatment were rated highest in sourness and green/grassy flavor, and lowest in ripe aroma, sweetness and tomato flavor. Ripe tomatoes with inferior sensory qualities showed significant changes in several aroma volatile compounds and chemical composition parameters. An ethylene exposure threshold of 3 days could be immediately utilized commercially as a non-destructive means for segregating immaturegreen from mature-green tomatoes, thus ensuring consistently high-flavor tomatoes.

Ripe tomatoes stored at 200C for 12 days were rated significantly higher for ripe aroma, tomato flavor and sweetness compared to those stored at 100 or 50C. High temperature pre-treatments (380C for 2 days) were not effective in alleviating flavor changes induced during storage at 50C for 7 days.

The electronic nose sensor array successfully classified intact green tomatoes into immature and mature stages, and distinguished between tomatoes harvested at different ripeness stages or stored at various temperatures below 200C. With increased detection speeds, electronic noses have potential for screening green tomatoes in commercial operations.









viii














CHAPTER 1
INTRODUCTION



Tomatoes (Lycopersicon esculentum Mill) are edible members of the Solanaceous family, they originated in Central and South America, and were first introduced to Europe by Spanish explorers returning from America. Because Europeans believed tomatoes were poisonous, consumption and popularity did not increase until the eighteenth century (Petro-Turza, 1987). Worldwide, tomatoes are among the ten most important fruits and vegetables in consumption. In a recent survey, tomatoes were ranked 16th in overall nutrient content; however, based on their nutritional contribution to the diet of most Americans was considered, tomatoes were ranked 1st (Goodenough, 1990).

Much popular concern has been expressed in the last two decades regarding the poor quality of fresh market tomatoes available in consumer markets (Harris, 1973; Cerra, 1975; Bisogni and Armbruster, 1976). USDA consumer surveys indicated a higher level of dissatisfaction with fresh tomatoes than with any of the 32 other products included in the surveys. The source of dissatisfaction was price, taste and ripeness (Resureccion and Shewfelt, 1985). Researchers have proposed several reasons for the inferior flavor in fresh market tomatoes. Most notably, commercial breeding programs have emphasized disease resistance, productivity and fruit firmness in selections at the expense of flavor and texture qualities (Baldwin et al., 1992). In addition, the lack of proper commercial harvesting and postharvest handling procedures for fresh market tomatoes has been shown to play an important role in affecting tomato flavor (Kader et


1






2


al., 1978; Sargent et al., 1997). Postharvest abuses, such as harvesting immature fruit, mechanical injury during sorting and packing, and improper temperature management have been related to altered aroma volatile profiles and altered flavor perception (Sargent et al., 1997; Moretti et al., 1997). Evidence of the adverse effects of low temperature storage on tomato flavor has been published previously (Kader et al., 1978; Stern et al., 1994; McDonald et al., 1996). In addition, the effects of storing tomatoes at currently recommended temperatures on their flavor and aroma has not been thoroughly addressed.

Throughout tomato production and marketing, efforts are made to maintain optimal visual quality (uniform color, absence of decay, proper firmness, etc.) to attract customers. As a consequence, disorders not readily detectable during sorting operations have received less attention. Visual appearance is a critical factor driving the initial purchase, but subsequent purchases are influenced greatly by eating quality (flavor, aroma and mouthfeel). Overall fresh tomato quality is determined not only by fruit appearance and firmness, but also flavor, aroma, and nutritive value. In the U.S., produce quality is often a more important factor determining consumer purchases than price (Schwartz, 1995).

Most fresh market tomatoes sold in U.S. supermarkets are harvested before they are "table-ripe" because retailing ripe tomato fruit is not practical within the current long distance handling and marketing system (Kader et al., 1978). In Florida, over 85% of the fresh market fruit are harvested green for transport to distant markets. Such factors as harvesting immature-green fruit, occurrence of mechanical injuries and storage below safe temperatures can reduce the overall quality from the shipping point to the consumer's table. The winter vegetable industry in the U.S. has seen drastic competition






3

and decline in sales due to imports from Mexico and Europe. Following the North American Free Trade Agreement (NAFTA), imports of Mexican produce during the winter months have more than doubled (Florida Tomato Committee, 1998). In contrast, fresh market tomato acreage in Florida has declined by approximately 25% between the 1993-94 and 1997-98 seasons. It is believed that the only solution for Florida growers is to provide consistent high quality tomatoes based on their green-harvest system.

External indicators of green tomato maturity, such as shape, color, surface appearance, and stem scar condition (Kader and Morris, 1976) are subjective and often impractical during commercial harvesting operations (Brecht et al., 1991). In fact, Florida tomatoes picked at green stage are harvested and sorted based on minimum size requirements set by the Florida Tomato Committee (Florida Tomato Committee, 1989). Fresh tomato maturity distribution data, collected from commercial green harvest operations, revealed that in average 49% of green-harvested fruits were at immaturegreen stage (Ml) with larger fruit sizes consistently showing lower immature fruit proportions (Chomchalow, 1991). The inability to accurately distinguish between the maturity classes (MI-M4) at harvest has obscured the results of sensory quality studies in the past (Watada and Aulenbach, 1979). Nonetheless, negative effects from immaturegreen harvest stages on ripe tomato chemical composition, overall sensory acceptance, and aroma volatile profiles have been suggested (Kader et al., 1978; Maul et al., 1997a).

During tomato fruit ripening, a series of quantitative and qualitative changes take place in tomato aroma volatile profiles. Organic acids, soluble sugars, amino acids, pigments, and over 400 aroma compounds contribute to characteristic tomato flavor (Petro-Turza, 1987). Due to the diversity of biosynthetic pathways contributing to the






4


formation of volatile compounds, tomato aroma could be a good indicator of fruit injury as a result of harvest and postharvest handling treatments.

New technologies such as machine vision, magnetic resonance imaging, acoustic impulse transmission, chlorophyll fluorescence, and electronic aroma sensing are powerful tools that could be used to define and predict fresh produce quality beyond traditional visual quality parameters. "Electronic nose" (EN) sensor arrays are used to classify specific samples of interest based on their headspace volatiles. The electronic nose consists of a series of non-specific chemical sensors, each of which shows characteristic responses to the volatile chemicals within the headspace over a sample (Anonymous, 1996).

The present study was conducted to document the effects of harvest maturity of green-harvested tomatoes on their ripe sensory quality. Sensory panelists were trained to describe flavor and aroma differences between table-ripe tomatoes harvested at green stage compared to tomatoes harvested at later ripening stages. In addition, nondestructive analysis using electronic nose sensor arrays and predictive models based on ethylene treatment requirements were explored to screen immature-green tomatoes. The effects of different storage temperatures on ripe tomato flavor were also investigated as potential factors contributing to inferior quality of fresh-market tomatoes.















CHAPTER 2
LITERATURE REVIEW



Tomato Fruit Ripening


After fruits and vegetables attain physiological maturity they undergo a distinct ripening phase characterized by a series of changes under genetic regulation. Ripening involves the synthesis of new proteins, nucleic acids, and carbohydrates (Mitcham et al., 1989), chlorophyll degradation, starch hydrolysis, formation of pigments, and synthesis of aroma volatiles (Biale and Young, 1991). All of ripening related metabolic pathways are very likely to be under hormonal control for both initiation and coordination (Tucker, 1993). Tomato ripening involves the autolysis of cell wall pectins, the synthesis of lycopene and other carotenoid pigments, changes in sugar and acid contents, and production of volatiles associated with flavor and aroma (Hobson and Davies, 1971; Gray et al., 1992; Petro-Turza, 1987). During ripening, sugar levels increase, possibly due to the hydrolysis of starch reserves in the fruit (Whiting, 1970). In fruits still attached to the plant, import of assimilates may also play an important role in sugar accumulation. Tomatoes utilize both sugars and organic acids as substrates for respiratory activities (Ulrich, 1970).

In climacteric fruits, such as tomatoes, the initiation of ripening is associated with a burst in ethylene production, accompanied by a large increase in respiratory activity



5






6


prior to or during ripening (Rhodes, 1980). Non-climacteric fruits, on the other hand, only show gradual decrease in respiration as they approach senescence. Climacteric ripening behavior, also common to bananas, mangoes, avocados, etc. (Hardenburg et al., 1986), allows for harvest at an "under-ripe" stage, because remaining ripening events can occur postharvest.

Ethylene production has been reported to increase at the mature-green stage, before the first visual symptoms of ripening (Brecht, 1985). Ethylene biosynthesis during ripening is autocatalytic (system II ethylene). Small amounts of exogenous ethylene stimulate a massive increase in the fruit's ethylene production (Yang and Hoffman, 1984) which, in turn may accelerate the onset of ripening. Ethylene is believed to coordinately induce the expression of a large number of genes which encode for regulatory enzymes in the ripening processes (Theologis et al., 1993). Ethylene biosynthesis starts with the conversion of methionine to S-adenosyl-methionine (SAM), followed by conversion of SAM to 1-aminocyclopropane-l-carboxylic acid (ACC) by ACC synthase, and finally conversion of ACC to ethylene by ACC oxidase (Yang, 1985).

Exogenous ethylene treatment is a commercial practice used to accelerate the ripening of green tomatoes and has been utilized by industry for the last five decades. Under laboratory conditions, green tomatoes have been shown to respond to ethylene concentrations as low as 0.5 tL/L, conversely, concentrations greater than 10 pL/L would not result in an enhanced ripening response. Nonetheless, most commercial tomato ripening operations utilize concentrations ranging between 50 and 500giL/L to assure response. Green tomato ripening response under ethylene treatment has been shown to decrease with increasing time between harvest and ethylene treatment






7


(Chomchalow, 1991). In addition, tomato response to ethylene has been related to physiological maturity at harvest (Kader et al., 1977; 1978). Green tomato ethylene requirement was inversely related to maturity at harvest. In fact, Kader et al. (1978) considered tomatoes requiring more than 6 days of ethylene treatment as immature-green at the time of harvest. Once autocatalytic ethylene production has commenced, exogenous ethylene exposure will not have a significant effect on subsequent ripening rates.

There is evidence that tomato ripening does not occur uniformly throughout the fruit. The initiation of ripening-related processes first occur in the locule tissue of mature-green tomatoes (Grierson et al., 1985) and could, in turn, stimulate the ripening of adjacent tissues (pericarp and columnella). The classification system for physiological maturity of green-stage tomatoes proposed by Kader and Morris (1976) relied on locule tissue development to determine maturity classes. Maturity class distribution was defined as immature-green (MI), partially mature-green (M2), mature-green (M3) and advanced mature-green (M4) based on locule tissue development. However, if ethylene requirement is to be employed as a screening tool for immature tomatoes, further research is needed to clarify the relationship between ethylene treatment requirement and physiological maturity.




Ripe Tomato Composition


Environmental factors such as water availability, soil fertility, and cultural practices have been shown to influence sugar/acid ratios in fresh market tomatoes






8


(Stevens, 1985). In general, ripe tomatoes are composed of approximately 93%-94% water, where the most prevalent dissolved organic constituent is sugar. Fructose and glucose may account for up to 50% of the fruit's dry matter and over 95 % of total sugars (Stevens, 1985). Sucrose concentrations are extremely low, probably due to acid invertase activity (Klann et al., 1996). Fructose could be considered more important for sweetness perception than glucose, due to the lesser sweetening power of the latter hexose (Stevens et al., 1977). In fact, Koehler and Kays (1991) proposed calculating "sucrose equivalents" by multiplying fructose and glucose concentrations times 1.73 and

0.74, respectively, to better represent individual sugar relative sweetening potential.

The major organic acid constituent in ripe tomato fruit is citric, which, may account for 60-90% of the total acid content. (Davies and Cocking, 1965) (Table 2.1). The second major organic acid in ripe tomatoes is malic. Malic acid's higher acidity potential determined to be approximately 14% higher than citric acid's during model solution sensory studies (Gardner, 1966) emphasizes its contribution to ripe tomato acidity.


Table 2.1. Typical compositional parameters for ripe fresh market tomatoes (expressed on a % fresh-weight basis)
Vitamin C Dry Total Reducing Sucrose Total Malic Citric (mg/100g Matter Carbohydrates Sugars Soluble Acid Acid fresh wt.) Solids 20 6.5 4.7 3.0 0.1 4.5 0.1 0.2 Adapted from Hobson and Grierson, 1993.



Sugar/acid ratios have been traditionally used as indicators of fruit quality and flavor potential. However, during tomato sensory studies, the ratios between soluble solids content and titratable acidity did not relate to tomato acceptability (Malundo et al.,






9


1995; Resureccion and Shewfelt, 1985). Furthermore, if both acid and sugar levels are low, it is possible to have poor flavor quality even though sugar/acid ratios remain desirable (Stevens, 1985). Malundo et al. (1995) challenged the assumption that increasing sugar or acid concentrations improved tomato flavor. In fact, it was demonstrated that only at certain acidity levels (-0.80% citric acid) did increased sugar concentrations lead to better acceptability ratings. Watada and Aulenbach (1979) reported "vine ripe"-harvested tomatoes were judged to be sweeter than green-harvested ones, even though there were no significant differences in dry matter content. This demonstrated the importance of other compounds, besides sugars, in sweetness perception.

Vitamin C content (ascorbic acid) has been shown to be affected by ripeness stage at harvest (Watada et al., 1976). During ripening, vitamin C contents increase from the green stage to reach its maximum prior to the ripe stage (Malewski and Markakis, 1971). Similarly, lycopene and 1-carotene levels consistently increase until the ripe stage.

Locule tissue may constitute between 14 and 34% of fresh tomato mass (Stevens et al., 1977), and, when compared to pericarp tissue, locule tissue contains considerably higher vitamin C, citric and malic acid contents, contrasted by lower sugars (Brecht et al., 1976; Stevens et al., 1977). Sharp contrasts in chemical composition, ethylene sensitivity, enzymatic activity, and contribution to overall fruit flavor, between locule and pericarp tissues suggest their separate analysis may prove to be useful during ripening and flavor biosynthesis studies.

One of the most important processes occurring during fruit ripening is the enhanced production of aroma volatiles (Pesis, 1994). The most important aroma






10


compounds in fruits are usually alcohols, aldehydes, esters, and other hydrocarbons (Nursten, 1970). Flavor volatiles, usually present at low levels, are considered to provide peculiar flavor and aroma notes, or descriptors, to different fruits (Buttery et al., 1990). Some vegetables, such as onions (Brodnitz et al., 1969) and peppers (Buttery et al., 1969) have a characteristic flavor attributed to a single or small group of volatile compounds. However, for most fruits and vegetables, flavor and aroma are determined by a complex combination of numerous volatile compounds (Bartley et al., 1985; Perez et al., 1992; Baldwin et al., 1991ac; Rizzolo et al., 1995).



Ripe Tomato Flavor


Natural aromas and flavors present in fruits and vegetables are almost always complex mixtures, comprising tens, and more often, hundreds of chemical constituents (Dodds et al., 1992). To date, over 6,000 flavor-active compounds have been identified, and the list will most likely increase as newer, more sensitive instruments and analytical techniques are developed. The characteristic flavor of tomato fruits arises from the complex interaction between organic acids, soluble sugars, amino acids, minerals, and aroma volatiles (Baldwin et al., 1991a). Approximately 400 different volatile compounds have been identified in ripe tomato fruits (Petro-Turza, 1987) with fewer than 30 volatile compounds considered important based on their odor and flavor threshold values (Buttery et al., 1990). From the list of ripe tomato aroma volatile compounds there, is limited quantitative evidence of individual compound contribution to overall tomato aroma (Goodenough, 1990). Cis-3-hexenal, 3-ionone, hexanal, 3-damascenone, 1-penten-3one, 3-methylbutanol, trans-2-hexenal, 2-isobutylthiazole, 1-nitro-2-phenylethane, trans-









2-heptenal, phenylacetaldehyde, 6-methyl-5-hepten-one, and cis-3-hexenol are among the most important volatile compounds present in ripe tomatoes (Buttery, 1993).

Characterization of volatile production and enzymatic activity by individual fruit tissues (locule and pericarp tissue) has demonstrated quantitative aroma volatile production by individual tomato tissues (Buttery et al., 1988). Levels of important tomato aldehydes and alcohols such as hexanal, cis-3-hexenal, trans-2-hexenal, and cis3-hexenol were two-fold higher in pericarp tissue compared to locule gel, whereas important ketone volatiles such as 6-methyl-5-hepten-2-one and geranylacetone were slightly higher in the locule tissue (Buttery et al., 1988). The important contribution of locule tissue fluidity for immediate access to taste receptors may influence flavor perception during consumption (Stevens et al., 1977).

One of the most rapidly growing areas in fruit flavor research is the characterization of flavor precursors (Rouseff and Leahy, 1995). Many plant nutrients such as free amino acids (Perez et al., 1992), lipids and carotenoid pigments (Stevens, 1985), and bound glycosides (Krammer et al., 1995) can act as flavor volatile precursors. In order to understand ripe tomato aroma, an understanding of precursor metabolism during ripening is necessary (Perez et al., 1992). Aroma volatile production studies could help elucidate important storage and handling factors affecting fruit flavor and odor preservation (Fellman et al., 1993).

Tomato aroma characterization studies would generally consider cultivars with higher volatile compound production as having potential for better flavor. In a study that compared sensory attributes for six commercial tomato cultivars, it was concluded that






12


higher levels of sugars, medium acidity and higher concentrations of important volatiles concurred with higher sensory acceptability (Baldwin et al., 1991b).

Aroma volatile production from tomato fruits has been reported to undergo considerable qualitative and quantitative changes during ripening. Aldehyde concentrations increased 9-fold from green to red ripeness stages while ketone volatiles increased approximately 2-fold during ripening (Baldwin et al., 1991c).



Lipid-derived Volatile Compounds


The characteristic aroma of tomatoes, apples, cucumbers, bell peppers, and bananas comes from aldehydes, ketones, and alcohols (Petro-Turza, 1987; Luning, 1995; Pesis, 1994). The formation of aroma and flavor volatile compounds from membrane lipids and free fatty acids is usually associated with free radical initiation or lipoxygenase-mediated oxidation (Shewfelt and Purvis, 1995). The presence of lipoxygenases and hydroperoxide lyases in microsomal membranes suggested that aldehydes, ketones, alcohols, and perhaps other volatile compounds could be produced at a membrane site (Riley et al., 1996). When linoleic acid substrate was utilized, lipoxygenase (LOX) isolated from algae produced aromas described as apple-like, green, cucumber, and mango (Kuo et al., 1996).

In tomato, LOX forms specific 13(s)- or 9(s)-hydroperoxides from the degradation of linoleic or linolenic acids (Riley et al., 1996). Hydroperoxide lyase (HPL) cleaves those position-specific hydroperoxides produced by LOX yielding aldehydes. Hexanal and cis-3-hexenal arise from cleavage of 13(s)-hydroperoxides of linoleic and linolenic acids, respectively (Gardner, 1995). During tomato ripening, Buttery (1993)






13


showed concentrations of cis-3-hexenal and hexanal increased 4 to 5-fold between green and breaker stage, and 5 to 10-fold between breaker and red stage. Cis-3-hexenal is highly unstable (Petro-Turza, 1987) and will be rapidly isomerized to trans-2-hexenal due to isomerase activity. Aldehyde compounds may, in turn, be converted to alcohols such as cis-3-hexenol due to the activity of alcohol oxidoreductases.

Evidence relating other carboxylic acids in the biosynthesis of aldehydes (C3-C5) is not yet available (De Pooter et al., 1987). Aroma compounds produced by LOX are desirable but, could also give rise to off-flavors in some cases. There is evidence that HPL-derived aldehydes play a role as antifungal agents in fruits (Gardner, 1995; Mattheis and Roberts, 1993). Among tomato lipid-derived volatile compounds, cis-3-hexenal, hexanal, 1-penten-3-one, trans-2-hexenal, trans-2-heptenal, and cis-3-hexenol have been proposed as important based on compound concentrations required for sensory response and their concentrations present in tomatoes (Buttery et al., 1990; Tandon, 1998).



Carotenoid-derived Volatile Compounds


Plant breeders have extensively used high pigment mutants for developing improved fruit color in new tomato cultivars (Wann et al., 1985). Lycopene, being one of the major pigments in tomato fruits and is an important flavor precursor for ketone volatiles (Petro-Turza, 1987). Variations in fruit color have been related to changes in flavor (Stevens, 1985). Some of the important tomato volatiles, such as 6-methyl-5hepten-2-one and geranylacetone, are produced from the oxidative decomposition of lycopene, while, 8-ionone is a product of 8-carotene decomposition (Buttery et al., 1990).






14

There is evidence that 6-methyl-5-hepten-2-one may have a different biosynthetic pathway than geranylacetone (Petro-Turza, 1987).

Another important enzyme in volatile biosynthesis is alcohol dehydrogenase (ADH). ADH catalyzes the inter-conversion of acetaldehyde to ethanol (Massantini et al., 1995) or aldehydes to alcohols such as cis-3-hexenal to cis-3-hexenol (Speirs et al., 1998). ADH activity in plant tissues has been attributed to several functions including survival during periods of hypoxia, protection from chilling stress, and the biogenesis of flavor ester volatiles (Mitchell and Jelenkovic, 1995). Ester volatiles are an important group of aroma compounds which augment at the time of ripening of both climacteric and non-climacteric fruits (Pesis, 1995). In addition, numerous volatile compounds may be derived from deamination-decarboxylation of amino acids. Notably, 2isobutylthiazole has been related to thiamine as a probable sulfur precursor (Petro-Turza, 1987).



Sensory Flavor Perception


Fruit flavor is without a doubt one of the most important parameters to consider in fruit quality. In general, humans are able to perceive flavors from the interaction of compounds with the sense of taste and smell. There are basically four different tastes our senses can distinguish: sweetness, sourness, bitterness, and saltiness. Taste buds on the tongue are the major organs for the perception of taste. However, the presence of taste buds on the epiglottis and of stratified squamous epithelial cells lining the pharynx and larynx point to the relevance of other organs in taste perception (Nagodawithana, 1994).






15


In addition to the basic sensations of sweetness, sourness, bitterness, and saltiness that are perceived by taste buds present in the tongue, the aroma given off by food as it travels the retro-nasal route, influences the perception of flavor (Maruniak, 1988). In humans, the flavor sensation is comprised of at least three separate sensory systems: the gustation, olfaction and trigeminal senses (Dodds et al., 1992). Gustation refers to sensations perceived by taste receptors present in the tongue (Plattig, 1988), while the trigeminal sense refers to the perception of noxious stimuli within the nasal cavity by the trigeminal nerve (Cain and Murphy, 1980). However, of these sensory systems, olfaction plays the dominant role in flavor sensation. The olfactory sensory tissue is a section of thin epithelium located high in the nose and contains millions of olfactory receptor neurons. The specificity of the olfactory sensing system in higher animals is derived from a combination of receptor cells with partially overlapping sensitivities, resulting in the human nose being able to detect complex odors in the parts-per-billion range (Hodgins and Simmonds, 1995).

In the human olfactory system, the sensitivity to odor intensities is high and generally follows a logarithmic relationship between odor sensitivity and increases in odorant concentration. In addition to sensitivity, the olfactory system has an excellent capacity of discriminating similar complex odors (Gardner and Hines, 1997). The intake of a complex mixture of odorant molecules stimulates primary neurons within the olfactory epithelium. Each secondary neuron integrates the signals of 10,000 to 20,000 primary neurons into a single signal that is fed to the brain, where the information is processed (Shurmer, 1990). The olfactory bulb is a bulbous tissue in the brain that performs the higher level processing of all the olfactory signals (Dodds et al., 1992) and






16

identifies chemical odorant mixtures as a unitary stimulus (Shurmer, 1990). The brain is trained to recognize odor patterns corresponding to odor descriptions saved in a person's memory (Tomlinson et al., 1995).

Stimuli released from flavor substances are perceived in the mouth and also in air delivered backwards to the nasal cavity while chewing or swallowing (Land, 1994). During swallowing, the odor-containing air accumulates in the nose because there is no expiratory airflow. Thus, the flavor quality and intensity perceived is influenced by the time period the food stays in the mouth prior to swallowing. Flavor perception has been shown to reach its full flavor potential within 10 seconds of chewing or swallowing (Land, 1994). In addition, whenever an aroma stimulus in the mouth reaches the olfactory receptor neurons via the retro-nasal route, it is perceived as a gustation rather than an olfaction sensation (Burdach et al., 1984). This event is referred to as the "taste-smell illusion".

There is a close relationship between aroma and flavor. The compound concentrations needed for human flavor and odor perception (odor and flavor thresholds) are almost identical (Ahmed, 1978). The importance of flavor volatile compounds present in fruit tissues have been determined by odor unit values (OUV) calculated in flavor characterization studies. OUV's relate the odor thresholds for the different aroma compounds with their actual concentration present in the fruit tissue (Buttery et al., 1990). However, OUV information cannot replace sensory evaluation since odor thresholds may vary between individuals by up to 2 orders of magnitude, while complex volatile mixtures may elicit different qualities than individual compounds. Laing et al.





17


(1984) showed that the perceived intensities of volatile compound mixtures were usually lower than the sum of the individual components.



Sensory Analysis


Taste and smell are the primary instruments used by various industries for evaluating the quality of a wide range of food products. No two humans are able to perceive an aroma in identical fashion, nor can they share common descriptors because of a series of cultural and associative memory effects (Strassburger, 1996). Therefore, sensory panel evaluation aims to provide objective information derived from the evaluation of subjective qualities present in a food product. Most classifications of odors used during sensory analysis are based on subjective perception and use common names, e.g. green, grassy, musty, floral, etc. Analytical chemists criticize the subjectivity and poor reproducibility of sensory test results, especially descriptive, whereas sensory scientists are often unable to correlate analytical data in a meaningful way with sensory qualities (Marsili, 1995). Nonetheless, sensory evaluation is the only tool readily available to assess the quality and acceptability of fruits and vegetables. Different sensory descriptors used by several researchers have helped to demonstrate the relationships between sensory evaluation and chemical composition of tomato fruits (Table 2.2). In addition, during odor threshold studies for important volatile compounds, sensory panels have described the attributes of selected volatile compounds in aqueous solutions (Buttery et al., 1990) and bland tomato homogenate (Tandon, 1998) (Table 2.3).






18


Table 2.2. Sensory panel descriptors utilized by researchers during fresh-market tomato quality studies.
Reference Flavor Descriptors Aroma Descriptors
Bisogni and Armbruster, Overall quality, overall flavor, Not evaluated
1976 sweetness, acidity
Kader et al., 1977 Sweetness, sourness, off- Vine-like or green, fruity or flavor, tomato-like, overall floral, overall intensity intensity
Kader et al., 1978 Sweetness, sourness, off-flavor Fruity-floral, overall intensity
Watada and Aulenbach, 1979 Sweetness, acidity, saltiness, Not evaluated grassiness, stemminess, fruityfloral flavor, mustiness,
bitterness, astringency
Resureccion and Shewfelt, Sweetness, acidity, tomato- Not evaluated
1985 like, off-flavor, overall intensity, juiciness
Kavanagh et al., 1986 Tomato flavor, sweetness, Not evaluated acidity/sourness


Table 2.3. Sensory descriptors and odor thresholds for important aroma volatile compounds found in ripe tomatoes.
Odor thresholds Y
Aroma volatile Reported Flavor and Aroma aqueous bland tomato
Compounds Descriptors z solutions homogenate
Cis-3-hexenal Green aroma, pleasant "fresh"
green aroma, and desirable mouth- 0.25 469 feel properties.
Trans-2-hexenal Lesser green character than cis-3- 17 592 hexenal. Not like tomato.
2-Isobutylthiazole Green leaf aroma, spoiled vine-like, 2 102 horseradish type flavor.
4.5 2427
Hexanal Green flavor

6-Methyl-5-hepten-2- Fruit-like aroma 50 522
one
zData from Petro-Turza, 1987; Kazeniac and Hall, 1970; Buttery et al., 1990; Tandon, 1998. Y Odor thresholds expressed in nL/L.



Discrimination sensory testing is one of the most useful analytical tools in sensory

work. Discrimination tests, such as triangle and difference from control tests, have been

designed to identify perceivable sensory differences between two or more products of





19

interest. Only on the basis of a perceived difference between products can a sensory scientist justify proceeding to a descriptive sensory test in order to identify the basis for such differences (Stone et al., 1974). Difference/discrimination triangle tests were conducted to assess flavor differences in tomatoes picked at different ripeness stages, revealing the relative ease of untrained panelists to differentiate these fruits (Kader et al., 1977).

Descriptive sensory analysis is defined "as a process of describing the perceived sensory characteristics of a product, usually in the order of occurrence" (Stone and Sidel, 1985). Descriptive sensory analysis involves the detection and the description of both qualitative and quantitative sensory aspects of a food product. All descriptive evaluations should be based only on perceived intensities and should be free of hedonic responses. The absolute scale values are less important than the relative differences among products to provide valuable information (Moskowitz, 1988). The two most prevalent methodologies currently employed in descriptive sensory analysis are the Quantitative Descriptive Analysis (QDA@) and the Spectnun methods.

The QDA method relies heavily on statistical analysis to determine appropriate terms, procedures and panelists. Training requires the use of product and ingredient references to stimulate the generation of terminology (descriptor terms). The panel leader acts as a facilitator rather than as an instructor, refrains from influencing the group and is not an active participant in the panel (Zook and Pearce, 1988). Panelists are screened for normal odor and taste perception using actual products from the category. In contrast, the Spectrum method is based on extensive use of reference points. Appropriate reference points reduce panel variability and allow for comparison of data over time and products






20


(Lawless and Heymann, 1998). Reference points are used to precisely calibrate the panelists in the same way as pH buffers calibrate a pH meter. Both the QDA and Spectrum methods were created and successfully employed in prepared food studies where ingredients and variability between samples may be closely controlled. In the case of sensory evaluation of fruits and vegetables, the variability in chemical composition between samples and the potential for sensory quality changes over a short period of time during sample preparation and storage make these popular descriptive sensory methods less suitable. Instead, fruit and vegetable flavor research usually relies on hybrid sensory methods that incorporate the strengths of formal methods while still allowing for flexibility during training, sample preparation, and sensory sessions.




Electronic Aroma Sensing Technology


An electronic nose is a sensor-based instrument designed to respond to volatile chemicals present in the headspace over a sample of interest. The electronic nose is comprised of a series of non-specific semiconductor chemical sensors manufactured from materials such as metal oxides, conducting polymers, quartz microbalances and fiber optics. Such chemical sensors are useful in aroma discrimination due to their altered electrical resistance properties as a result of the interaction with volatile compound molecules (Anon., 1996). Most semiconductor materials used for odor sensing in electronic noses experience a decrease in electrical resistance properties as volatile compound concentrations adsorbed to the sensor increase. The array of non-specific sensors allows an electronic nose to detect a wide range of volatile compounds (Benady






21


et al., 1995). Electronic noses not only offer the possibility of objective analysis of volatile chemicals responsible for the aroma of foods, but also detection of volatile chemicals odorless to humans (Anon., 1996).

Electronic noses have numerous similarities in architecture and other properties, with biological olfaction systems such as odor delivery, nonspecific sensor/receptor response, sensor/receptor preprocessing and content addressable memory (Pearce, 1997a). In addition, similar limiting factors such as sensor/receptor drift, degeneration and poisoning, and limited sensor/receptor sensitivity, are shared by both systems (Pearce, 1997b). Polymer-based sensor arrays are able to detect organic compounds with molecular weights from 30 to 300 Daltons, values similar to what biological noses perceive (Marsili, 1995).

Following odor sensing, the output from the sensor array must be analyzed. A pattern recognition procedure seeks identifying characteristics in the sensor outputs to be used in grouping algorithms for samples belonging to the same treatment. This pattern recognition approach to analyze sensor outputs closely mimics the way the human brain discriminates sensory signals and relates them to similar signals stored in our memory. The levels of aromatic compounds which accumulate as climacteric and non-climacteric fruits and vegetables ripen can be used to predict flavor quality (Benady et al., 1995). Use of an electronic nose in harvest and postharvest operations to select inferior quality and overripe fruits and vegetables has been explored using packaged blueberries (Simon et al., 1996) and melons (Benady et al., 1995). Due to the costly operation of analytical instrumentation, such as gas chromatography and mass spectroscopy, electronic noses may become useful alternative tools in commercial food handling and quality control






22

operations. Thus, portable aroma sensing instruments using electronic nose technology could be useful to predict maturity during field harvesting operations for a number of fruit and vegetable crops (Benady et al., 1995).



Harvest Maturity and Tomato Flavor


Physiological maturity at the time of harvest is critical for normal fruit ripening to proceed (Brecht, 1987). Immature-green fruits often do not express their full flavor potential (Kader et al., 1977). Chemical composition analyses have shown that greenharvested tomatoes often result in ripe fruits with lower pH, higher ascorbic acid and total sugar levels when compared to partially ripe-harvested ones (Al-Shaibani and Greig, 1979). Accumulation of free amino acids (Perez et al., 1992), lipids, lycopene and carotenoid pigments (Stevens, 1985), and bound glycosides (Krammer et al., 1995) acting as flavor precursors could explain flavor differences between green-harvested tomatoes and those harvested with color, when analyzed at ripe stage (Watada and Aulenbach, 1979). Lycopene concentrations have been shown to be higher in tomatoes harvested at "vine-ripe" stage compared to those ripened postharvest (Mencarelli and Saltveit, 1988). Numerous studies have focused on the relation between ripe tomato quality and chemical composition from green-harvested tomatoes (Bisogni and Armbruster, 1976; Kader et al., 1977; Kavannagh et al., 1986; Watada and Aulenbach, 1979) with relative success correlating sensory parameters to compositional quality. However, without a thorough understanding of the factors influencing tomato flavor, significant correlations between sensory attributes and chemical composition parameters cannot be interpreted using a cause and effect relationship.






23

There is evidence that green-harvested tomatoes supplied with exogenous ethylene (100 pL/L C2H4 for 1 day) resulted in significantly altered flavor volatile profiles (McDonald et al., 1996). Ethylene treatment resulted in tomatoes with significantly reduced levels of hexanal, 6-methyl-5-hepten-2-one, geranylacetone, methanol, 2-isobutylthiazole, and 1-nitro-2-phenylethane. Nonetheless, the effects of harvest maturity on ripe tomato flavor quality and the contribution of volatile profiles to inferior sensory characteristics have not been clearly established.



Postharvest Treatments and Tomato Flavor.


It has been documented in the past how postharvest treatments can affect production of aroma volatiles, thus influencing fruit flavor (Willaert et al., 1983). The senescence of fruits during extended storage periods results in a decreased formation of volatile compounds, probably from reduced enzymatic activity or loss of flavor precursors (Yahia et al., 1992). In 'Anjou' pears, exposure to chilling temperatures leads to an increased unsaturation of linolenic and linoleic acids (Gerasopoulos and Richardson, 1995), thus increasing substrate levels for lipoxygenase activity. The accessibility of substrates for hydroperoxide lyase and other key enzymes during fruit ripening could represent a limiting factor in the biosynthesis of flavor volatiles (Riley, 1996). The effects of storage temperature and physiological maturity, among others, have shown significant effects on ripe tomato flavor (Stem et al., 1994; Buttery et al., 1987; Kader et al., 1977).





24

Storage Temperature Management


Chilling injury is a physiological disorder resulting from the exposure of susceptible plant tissues to temperatures above 0oC, but usually below 150C (Crooks and Ludford, 1984). The major symptoms of chilling injury (CI) in tomato are: increased susceptibility to Alternaria alternata (McColloch and Worthington, 1952), delayed, partial, or uneven ripening, pitting (Dodds et al., 1991), and enhanced fruit softening (Chomchalow, 1991). Tomato sensitivity to CI appears to be influenced by ripening, probably due in part to dramatic changes in membrane lipid composition as affected by the exposure to chilling temperatures (Whitaker, 1994; McDonald et al., 1996; 1998). Mature-green (Autio and Bramlage, 1986) and breaker-stage (Hobson, 1981) tomatoes appear to be the most sensitive to CI. The extent of CI is dependent upon the interaction between chilling temperature, length of exposure time, and varietal susceptibility (Chomchalow, 1991). Chilling temperatures used for storage of breaker stage tomatoes (20C for 14 days) significantly affected the aroma volatile profiles at ripe stage (McDonald et al., 1996; Buttery et al., 1987). Reduced levels of hexanal, 6-methyl-5hepten-2-one, geranylacetone, methanol, 2-isobutylthiazole, and 1-nitro-2-phenylethane were attributed to chilling temperature storage. Hobson (1987) recommended 8.50C as the minimum storage temperature to be tolerated by breaker-stage tomato fruits, as a 100C storage temperature appeared to cause only minor changes in fruit composition (Hobson, 1981). Even though Morris (1954) reported CI in tomato fruits after 10 days at 100C, the storage of partially-ripe tomatoes at this temperature has become an increasingly common practice in Europe (Hobson, 1987). In the U.S. it is common for tomato handlers to store and ship green or partially ripe tomatoes at 100C for extended






25

periods of time to maximize marketing flexibility (S. A. Sargent, personal communication). However, Stern et al. (1994) found tomatoes harvested at breaker stage and ripened at 150C contained higher concentrations of volatile compounds compared to those stored at 200, 100 or 50C. Nevertheless, the effect of postharvest storage at or below 150C, often recommended for commercial operations, needs to be addressed.

Alleviation of CI symptoms has been demonstrated by intermittent warming cycles (Artes et al., 1998) or high temperature treatment prior to storage at low temperatures (McDonald et al., 1998). Sabehat et al. (1996) documented a consistent relationship between heat-shock protein accumulation and persistence with tolerance to CI symptoms in tomatoes, thus suggesting a prominent role of heat-shock proteins in tomato fruit tissue acclimation and tolerance to low temperature storage. Short exposures to temperatures above 350C have been shown to reversibly inhibit tomato ripening (Lurie and Klein 1991), ethylene biosynthesis (Biggs et al., 1988) and lycopene accumulation (Ogura et al., 1975). Levels of most aroma volatile compounds in heat treated/chilled fruit were intermediate between chilled and non-chilled controls (McDonald et al., 1998). Even though there is strong evidence regarding the effectiveness that high temperature pre-treatments (380C for 2 days) have on visual CI symptom alleviation (Sabehat et al., 1996; Lurie et al., 1996; Artes et al., 1998), there is no information regarding potential benefits of high temperature pre-treatments on chilling-induced flavor alteration.






26

Research Objectives



The objectives of this research project were to:


1. Investigate the effects of green tomato harvest maturity on flavor and aroma quality at

ripe stage utilizing discrimination sensory tests, GC aroma volatile profiles, and

chemical composition analyses.

2. Explore the potential use of exogenous ethylene treatment or electronic nose sensor

arrays as nondestructive screening tools for green-tomato physiological maturity at

harvest.

3. Identify ripe flavor and aroma quality differences between fresh market tomatoes

harvested at green-stage or partially ripe stages utilizing a trained descriptive sensory

panel, GC aroma volatile profiles, and chemical composition analyses.

4. Determine the effects of postharvest storage temperatures below 200C on ripe tomato

flavor and aroma quality, and, explore the effectiveness of high-temperature pretreatments (380C for 2 days) as a means to alleviate chilling-induced flavor and

aroma quality alterations.














CHAPTER 3
RIPENESS STAGE AT HARVEST AFFECTS CHEMICAL COMPOSITION AND
COLOR QUALITY OF TABLE-RIPE TOMATOES


Introduction


Dissatisfaction with fresh market tomatoes has been attributed to inferior flavor and aroma, possibly affected by cultivar, cultural practices, growing conditions, maturity stage at harvest, and inadequate or inappropriate postharvest handling practices (Stevens, 1985; Kader et al., 1977; Baldwin et al., 1991c). Traditionally, tomatoes are grown in Florida during winter months and shipped to distant markets. Due to the lack of accurate visible indicators of physiological maturity, commercial green harvest operations must rely almost entirely on fruit size as harvest index. Physiological maturity of greenharvested tomatoes has been circumstantially related to ripe, fresh tomato flavor and quality (Kader et al., 1977; Watada and Aulenbach, 1979; Al-Shaibani and Greig, 1979).

It has been proposed that immature-harvested tomato fruits will never ripen properly (Brecht et al., 1991) or develop their full flavor potential (Kader et al., 1978). Numerous researchers have found increasing flavor desirability with increased maturity at harvest (Watada and Aulenbach, 1979; Kader et al., 1977). Nonetheless, chemical composition data collected by these authors did not always concur with their sensory panel results. Recent studies (Hobson and Bedford, 1989; Malundo et al., 1995) have determined that compositional parameters, such as acidity and sugar content, are closely



27





28


related to overall tomato acceptability. Tomatoes picked at the red-ripe stage were described as sweeter and with stronger fruity-floral character compared to those harvested at green stage (Watada and Aulenbach, 1979).

The following studies were conducted to document compositional differences in table-ripe tomato fruit as related to harvest date, ripeness stage and physiological maturity at harvest. Also investigated were the effects of the exposure of green tomatoes to ethylene (to initiate ripening) on compositional quality.



Materials and Methods



Tomato fruits from three tomato cultivars were harvested at various ripeness stages from experimental research plots at Collier Farms, Naples, FL. During the first experiment conducted in December 1995, 'Agriset-761' (Agrisales Inc., Plant City, FL) and 'CPT-5' (Collier Farms, Naples, FL) tomatoes were harvested at green (stage 1, 0% red color), breaker (stage 2, <10% red color), light red (stage 5, <90% red color) and red (stage 6, >90% red color) ripeness stages (USDA, 1976). In a second experiment conducted in January 1996, 'Agriset-761' and 'CPT-5' tomatoes were harvested at green, breaker and pink ripeness (stage 4, 30-60% red color) stages. Immediately after harvest, tomatoes were transported to Gainesville, FL, washed with chlorinated water and towel dried, and stored at 200C and 85-90% RH for subsequent ripening.

In a third experiment conducted in March 1996, 'Agriset-761' and 'CPT-5' tomatoes were harvested at green stage and transported the same day to Gainesville, FL for subsequent ethylene gassing treatment. Tomatoes were washed with chlorinated





29


water and towel dried prior to ethylene treatment inside sealed chambers at 200C where a humidified 100 gL/L ethylene/air mixture was constantly supplied using a flow-through system. The gassing chambers were opened daily to remove any fruits that had initiated ripening evidenced by red color on the surface (attained breaker stage). Ethylene treatment was terminated after 7 days, at which time all tomatoes had initiated ripening. Following ethylene treatment, tomatoes were placed in air at 200C and 85-90% RH for subsequent ripening.

In all three experiments, tomatoes were analyzed once they attained the table-ripe stage, defined as red stage coloration and significant fruit softening (3-4 mm deformation when a constant 9.8N force was applied to the equator for 5 sec) (Gull, 1980). Average number of days to table-ripe stage, fruit deformation and fruit equatorial diameter data were recorded. In addition, fruit color was measured on four equidistant locations around the fruit equator using a tritimulus colorimeter (Minolta CR-300, Ramsey, NJ). Color data were expressed in terms of lightness coefficient (L*), a* and b* values, from which hue angle (tan-' b*/a*) and chroma values (a*2 + b*2)1/2 were calculated.

Upon reaching table-ripe stage, fruit samples from different ripeness stages at harvest and ethylene exposure times were collected for chemical composition analyses. The chemical composition assays conducted (n=15 fruits/treatment) included pH, titratable acidity (expressed as % citric acid), soluble solids (oBrix), vitamin C (mg ascorbic acid/100 g fresh weight), and total sugars (% of fruit fresh weight).

Tomato homogenates from each of the harvest maturity treatments were centrifuged at 18,000 X gn and 50C temperature. The supernatant was filtered using cheesecloth, stored inside scintillation vials and frozen at -200C for later analysis.





30

Titratable acidity, expressed as % citric acid, was determined by titrating 1.5 g of tomato supernatant to 8.2 pH with a 0.1 N NaOH solution using an automatic titrimeter (Fisher Scientific, Pittsburgh, PA). Soluble solids content, expressed as oBrix, was measured using a tabletop digital refractometer (Abbe Mark II, Reichart-Jung, Buffalo, NY) and pH measurements were conducted using a digital pH-meter (Coming model 140).

For the vitamin C assays, tomato homogenates (n=4) from each treatment (2 g/sample) were stabilized using 20-mL of acid mixture (6% HPO3 containing 2 N acetic acid in water) prior to centrifugation. Vitamin C content was analyzed using a dinitrophenylhydrazine (DNPH) method adapted from Terada et al. (1978). Stabilized homogenate samples (1 mL/sample) were mixed with 50 p.L of a 0.2% water solution of 2,6-dichlorophenolindolphenol (DCIP), 1 mL of 2% thiourea solution in water and 0.5 mL of DNPH. Samples were incubated at 600C for 3 hours, then cooled on ice prior to adding 2.5 mL of 90% sulfuric acid (H2SO4) to each sample. Absorbance was read at 540 nm using a spectrophotometer (Beckman DU-20, Irvine, CA). Vitamin C concentrations (mg ascorbic acid/100g fresh weight) were determined from a standard curve using ascorbic acid standards.

Total soluble sugar contents were analyzed using a spectrophotometric method adapted from Dubois et al. (1956). Each tomato homogenate sample (n=4/treatment) was diluted with distilled water (1:500 mL), then a 0.5 mL sample of the dilute homogenate was combined with 0.5 mL of 5% phenol (w/v) and mixed thoroughly. Concentrated sulfuric acid (2.5 mL/sample) was added and then samples were cooled to room temperature prior to spectrophotometer readings. Absorbance was read at 490 nm using a






31


spectrophotometer (Beckman DU-20, Irvine, CA). Total sugar content present in tomato samples was determined from a standard curve derived from sucrose standards.



Results and Discussion




'Agriset-761'. Green tomatoes harvested in December 1995 required an average of 13.6 days to attain table-ripe stage during storage at 200C, while those harvested in January 1996 required 15.2 days (Table 3.1). Breaker-harvested fruits showed a similar trend, where those harvested in January required longer time to attain table-ripe stage, compared to those harvested in December (10.7 and 9.9 days, respectively). In contrast, pink fruit harvested in January required less time to ripen than those harvested at light red stage in December (5.9 and 6.8 days, respectively). Equatorial diameter showed no significant differences between fruit from the different ripeness stages at either harvest date. However, fruit harvested in January were of slightly greater diameter than those harvested in December.

Chemical composition analyses for 'Agriset-761' showed no significant treatment effects between the four ripeness stages harvested in December. However, tomatoes from the January harvest picked at breaker or pink stages were found to have significantly lower pH compared to those harvested at green stage (4.35, 4.33, and 4.41, respectively) (Table 3.1). Comparing pH values for the three harvest dates revealed that pH values were comparable regardless of ripeness stage at harvest. Titratable acidity was significantly higher in tomatoes harvested at pink stage (January) compared to those harvested at earlier ripeness stages. Meanwhile, no significant differences in titratable






32

acidity were found for tomatoes harvested in December. Soluble solids content decreased from green to breaker harvest, but increased at later ripeness stages at harvest. A positive relationship between titratable acidity and soluble solids content was evidenced in the December harvest; light-red harvested tomatoes were highest in titratable acidity and soluble solids content, breaker-harvested tomatoes were lowest in both parameters, while red- and green-harvested fruit were intermediate for both parameters (Table 3.1). Ripe tomato vitamin C content for tomatoes harvested in January showed no significant differences between ripeness stages. Nonetheless, greater average vitamin C contents were found for pink-harvested fruit (16.4 mg/100 g FW), compared to those harvested at either green or breaker stage (10.8 and 10.3 mg/100 g FW, respectively) (Figure 3.1).



Table 3.1. Compositional quality parameters for table-ripe 'Agriset-761' tomatoes harvested at several ripeness stages and two harvest dates. Compositional December 1995 January 1996 Parameters z Green Breaker Light Red Green Breaker Pink Red
Firmness Y 2.81 b 3.86 a 4.01 a 3.90 a 4.03 a 3.91 a 4.01 a Fruit Diameter (mm) 76.8 a 77.2 a 79.5 a 78.2 a 87.1 a 86.8 a 84.0 a Days harvest to ripe stage 13.6 a 9.91 b 6.8 c 3.4 d 15.2 a 10.7 b 5.9 c pH 4.40 a 4.41 a 4.36 a 4.42 a 4.41 a 4.33 b 4.32 b Titratable Acidityy 0.89 a 0.86 a 0.97 a 0.90 a 0.79 b 0.79 b 0.85 a Soluble Solids (Brix) 3.80 a 3.40 a 4.07 a 4.00 a 3.88 a 3.57 b 3.79 a
'Within each harvest date (n = 15 fruit/treatment), values followed by different letters between ripeness stages are significantly different at the 5% level according to Duncan's Multiple Range Test.
YFruit firmness (deformation in mm) induced by a constant 9.8N force applied for 5 sec. x Titratable acidity expressed as % of citric acid.


In the third experiment, where 'Agriset-761' tomatoes harvested green were subsequently treated with exogenous ethylene, fruit diameter varied with the time of ethylene treatment required to reach breaker stage. Even though tomatoes exposed for 1day ethylene treatment had significantly greater diameter compared to longer ethylene





33

exposure times, there was no consistent trend relating fruit diameter and ethylene gassing exposure time (Table 3.2). The length of ethylene exposure to attain breaker stage appeared to have a significant effect on subsequent ripening rate, expressed as the number of days to attain table-ripe stage from breaker stage. Tomatoes exposed for 4 and 5 days of ethylene treatment had significantly shorter average ripening times to table-ripe stage (7.6 and 7.9 days) compared to the rest of ethylene exposure times.

Chemical composition parameters at table-ripe stage were also affected by the ethylene exposure time from green-harvested tomatoes. Soluble solids contents were significantly lower in fruit exposed for 5 days of ethylene treatment, however, fruit exposed for 6 or 7 days did not show significant soluble solids content differences when compared to fruits exposed for less than 5 days of ethylene treatment (Table 3.2). Titratable acidity and pH values showed no significant differences between treatments, however, tomatoes exposed for 5 days of ethylene did have lower acidity compared to the rest of the treatments. Concurring with soluble solids results, total sugars in fruit exposed for 5 days of ethylene were lower compared to the remaining treatments. No significant differences in vitamin C content were found between tomatoes with different ethylene exposure times. However, it is important to note that fruits exposed for 6 and 7 days of ethylene treatment had the lowest vitamin C contents (Figure 3.2).

From these experiments it was observed that the ripening rate for 'Agriset-761' tomatoes was affected by harvest date. Light-red stage tomatoes harvested in December required longer ripening times than pink stage fruits harvested in January. Breaker stage tomatoes required about 1 additional day to ripen between December and January harvest dates (9.9 and 10.7 days, respectively). A similar response was obtained from green






34


stage tomatoes, where ripening time differential for the two harvest dates approached 2

days (13.6 and 15.2 days for December and January, respectively). Green-harvested fruit

treated with exogenous ethylene (March 1996) required ripening times similar to those

tomatoes harvested at breaker stage in the two previous experiments. Similar ripening

times suggested that ethylene exposure accelerates the onset of ripening, however, once

breaker stage has been achieved ripening rates were practically unaffected.



Table 3.2. Compositional quality parameters for table-ripe 'Agriset-761' tomatoes harvested in March 1996 at green stage and supplied exogenous ethylene (100 liL/L) until breaker stage attained.
Compositional Ethylene gassing exposure times Parameters 1 day 2 day 3 day 4 day 5 day 6 day 7 day Firmness Y 3.70 c 4.52 ab 4.48 ab 4.36 b 4.70 ab 5.14 a 4.81 ab Fruit diameter (mm) 89.7 a 77.4 bc 80.7 bc 80.3 bc 82.6 ab 73.0 c 79.0 bc Days harvest to ripe stage 9.67 a 9.78 a 9.1 ab 7.6 c 7.9 bc 9.67 a 10.2 a Soluble solids (OBrix) 3.95ab 4.04 a 4.10 a 4.01 a 3.35 b 3.93ab 4.35 a pH 4.30 a 4.30 a 4.29 a 4.27 a 4.27 a 4.22 a 4.26 a Titratable acidity x 0.76 a 0.79 a 0.76 a 0.80 a 0.69 a 0.77 a 0.77 a Total sugars (% fresh wt.) 1.64ab 1.70 ab 1.80 a 1.69 ab 1.43 b 1.61ab 1.68 ab
zValues followed by different letters between ethylene exposure times (n = 15 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. Y Fruit firmness (deformation in mm) induced by a constant 9.8N force applied for 5 sec. x Titratable acidity expressed as % of citric acid.


Although soluble solids content varied between harvest stages and harvest dates

there was no consistent pattern. Titratable acidity, however, seemed to be affected by

ripeness stage at harvest. In the first two experiments, a consistent trend of increasing

acidity with increasing ripeness was evident. This observation concurs with Kavanagh et

al. (1986), who reported lower acidity with decreasing maturity at harvest. In this study,

titratable acidity was not influenced by green tomato ethylene exposure time. However, it

was generally observed that there were lower pH values in green-harvested tomatoes






35

treated with ethylene in the third experiment when compared to those harvested at the same stage and not treated with ethylene in previous experiments.

Fruit color was significantly affected by ripeness stage at harvest for 'Agriset761' tomatoes. For both harvest dates, significantly lower hue angles were documented for fruit harvested at breaker or green stages compared to those harvested at later ripeness stages (Table 3.3). In addition, a* values were significantly lower for tomatoes harvested at red stage (December) compared to the other ripeness stages. Although not significant, a consistent trend of reduced L* values from later ripeness stages at harvest, was evidenced during both harvest dates.


Table 3.3. Color quality parameters for table-ripe 'Agriset-761' tomatoes harvested at different ripeness stages on two harvest dates.
Color December 1995 January 1996
Parameters z Green Breaker Light Red Green Breaker Pink Red
L* value 42.23 a 42.97 a 41.63 a 41.54 a 43.69 a 43.45 a 42.95 a a* value 27.47 a 28.07 a 26.11 ab 23.6 b 22.36 a 20.87 a 21.71 a b* value 25.51 a 25.75 a 27.29 a 26.56 a 22.41 b 20.32 c 23.89 a Hue angle Y 42.88 b 42.50 b 46.35 a 48.41 a 45.11 b 44.27 b 47.77 a Chroma x 37.51 a 38.11 a 37.80 a 35.57 a 31.7 a 29.2 b 32.30 a
zWithin each harvest date (n = 15 fruit/treatment), values followed by different letters between ripeness stages are significantly different at the 5% level according to Duncan's Multiple Range Test.
YHue angle = Tan' (b*/a*)
xChroma = (a*2 + b*2)1/2


L* values were significantly lower for fruit exposed for I to 5 days of ethylene compared to those exposed for more than 5 days (Table 3.4). Significant differences in average a* values showed that tomatoes exposed for 2 days of ethylene treatment were lowest (least red: a* = 22.56); fruits from all treatments were significantly lower than those exposed for 7 days (a* = 27.49). Tomatoes exposed for 1 and 2 days of ethylene treatment were lowest in b* and chroma values when compared to those tomatoes






36


exposed for over 5 days. In addition, tomatoes exposed for 1 day ethylene had significantly lower hue angles compared to those exposed for 7 days (43.48 and 47.61, respectively) (Table 3.4).



Table 3.4. Color quality parameters for table-ripe 'Agriset-761' tomatoes harvested at green stage and supplied exogenous ethylene (100 pL/L) until breaker stage was attained.
Color Ethylene gassing exposure time
Parameters 1 day 2 day 3 day 4 day 5 day 6 day 7 day L* value 43.5 b 42.05 b 41.88 b 41.83 b 43.04 b 43.79 ab 45.65 a a* value 23.2 bc 22.56 c 23.81 bc 24.1 bc 25.18 b 25.1 b 27.49 a b* value 22.0 d 22.56 d 23.42 ed 24.1 bcd 25.4 be 26.17 b 30.13 a Hue Angle y 43.48 b 45.04ab 44.45 ab 44.9 ab 45.3 ab 46.27 ab 47.61 a Chroma x 31.98 d 31.94 d 33.44 bc 34.1 bcd 35.8 ab 36.28 b 40.78 a zValues followed by different letters between ethylene exposure times (n = 15 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. Y Hue angle = Tan"' (b*/a*)
xChroma = (a*2 + b*2)1/2


'CPT-5'. For 'CPT-5' tomatoes, significant differences in fruit deformation at table-ripe stage were documented, as was the case for 'Agriset-761' harvested in December. In contrast, fruit harvested in January showed no significant differences in fruit firmness, thus reflecting a greater degree of uniformity criteria for determination of the in table-ripe stage. Fruit diameter was considerably greater in 'CPT-5' fruit harvested in January compared to those in December, however, there were no significant differences in fruit diameter for either harvest date (Table 3.5). The ripening times for fruit harvested at different ripeness stages was also influenced by harvest date, where fruit harvested in January generally required longer ripening times (except for pink stage).

The effects of ripeness stage on pH values were contradicting between harvest dates. In December, fruits harvested at red stage had significantly higher pH compared to






37


those harvested at green stage (4.41 and 4.32), while pH values for breaker and light redharvested fruit were intermediate (4.34 and 4.36). Fruits harvested at breaker or pink stage in January were significantly lower in pH compared to green-harvested ones (4.35, 4.33 and 4.41, respectively) (Table 3.5). Titratable acidity content for green-harvested fruit was significantly lower than breaker-harvested (December), while there were no significant differences between tomatoes harvested in January. For tomatoes harvested in January, vitamin C content showed an increasing trend with increasing ripeness at harvest, although, no significant differences were found (Figure 3.1).


Table 3.5. Compositional quality parameters for table-ripe 'CPT-5' tomatoes harvested at several ripeness stages and two harvest dates.
Compositional December 1995 January 1996
Parameters z Green Breaker Light Red Green Breaker Pink Red
Firmness Y 2.60 b 3.29 ab 3.16 ab 3.37 a 3.88 a 3.74 a 4.09 a Fruit Diameter (mm) 81.2 a 76.1 a 84.8 a 79.3 a 94.3 a 88.5 a 90.0 a Days harvest to red stage 12.4 a 10.8 a 8.0 b 5.8 c 15.1 a 11.4 b 7.1 c pH 4.32 b 4.34 ab 4.36 ab 4.41 a 4.41 a 4.35 b 4.33 b Titratable Acidity x 0.76 b 0.97 a 0.87 ab 0.84 ab 0.82 a 0.80 a 0.82 a Soluble Solids (oBrix) 4.02 a 3.71 a 3.92 a 3.95 a 3.92 a 3.79 a 3.81 a
zWithin each harvest date (n = 15 fruit/treatment), values followed by different letters between ripeness stages are significantly different at the 5% level according to Duncan's Multiple Range Test.
YFruit deformation (in mm) induced by a constant 9.8N force applied for 5 sec. x Titratable acidity expressed as % of citric acid.


In 'CPT-5' tomatoes, fruit diameter varied significantly with different ethylene exposure times. Fruit which reached breaker stage after 2 days of ethylene had the largest diameter (84.4 mm) while fruit exposed for 7 days had the smallest average diameter (72.3 mm) (Table 3.6). Diameter data from the remaining ethylene treatments showed no relationship to ethylene exposure time. The number of days from breaker to table-ripe stages were not influenced significantly by ethylene exposure time to attain






38

breaker stage, which contradicts the data for 'Agriset-761'. Soluble solids generally

followed an increasing trend with number of days of ethylene treatment. Tomatoes

exposed for 1 and 2 days of ethylene of treatment had the lowest soluble solids contents

(3.48 and 3.490 Brix, respectively) while those exposed for 6 and 7 days had the highest

(4.18 and 4.600 Brix) (Table 3.6). The highest pH value was obtained from fruit that

required 7 days of ethylene treatment while the lowest from those that required only 2

days of ethylene treatment. Tomatoes exposed for 7 days of ethylene were not only

highest in pH but also highest in titratable acidity (0.83% citric acid) and total sugars

(1.94 %). Vitamin C content showed no significant differences between ethylene

treatments, however, a consistent trend of decreasing vitamin C with increasing ethylene

gassing exposure time was evident (Figure 3.2).


Table 3.6. Compositional quality parameters for table-ripe 'CPT-5' tomatoes harvested at green stage and supplied exogenous ethylene (100 pL/L) until breaker stage was attained.
Compositional Ethylene gassing exposure time
Parameters z 1 day 2 day 3 day 4 day 5 day 6 day 7 day Deformation Y 4.50 a 4.56 a 4.56 a 4.40 a 4.30 a 4.80 a 4.19 a Fruit Diameter (mm) 83.3 ab 84.4 a 80.7 ab 76.6 bc 80.5 ab 79.3 ab 72.3 c Days to table-ripe 13 a 12.4 a 12.8 a 11.7 a 10.3 a 12.0 a 11.0 a stage
Soluble solids (oBrix) 3.48 b 3.49 b 4.02 ab 3.60 b 3.70 ab 4.18 ab 4.60 a pH 4.28 abc 4.22 b 4.30 abc 4.34 ab 4.24 bc 4.28 abc 4.38 a Titratable acidity x 0.74 a 0.67 a 0.80 a 0.70 a 0.79 a 0.77 a 0.83 a Total sugars 1.45 b 1.79 ab 1.68 ab 1.59 ab 1.76 ab 1.75 ab 1.94 a (% fresh wt.)
zValues followed by different letters between ethylene exposure times (n = 15 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. Y Fruit deformation (in mm) induced by a constant 9.8N force applied for 5 sec. x Titratable acidity expressed as % of citric acid.


'CPT-5' tomatoes ripened darker (decreased L* value) with increasing ripeness

stage at harvest (Table 3.7). This was also observed in 'Agriset-761', particularly in

light-red (December) and pink (January) stage harvested fruits. In general, hue angles






39

were higher for tomatoes harvested in January, however, for both harvests there was a trend of increasing hue angles (less red) with increasing ripeness stage at harvest. During the January harvest, chroma and b* values were significantly lower in fruit harvested at green or breaker stages (36.65 and 36.25, respectively) compared to those harvested at light-red or red ripeness stages (39.18 and 38.77, respectively) (Table 3.7).



Table 3.7. Color quality parameters for table-ripe 'CPT-5' tomatoes harvested at several riness stages and two harvest dates.
Color December 1995 January 1996
Parameters z Green Breaker Light Red Green Breaker Pink Red
L* value 42.61 a 42.65 a 40.90 b 40.53 b 45.30 a 44.36 a 41.89 b a* value 26.99 a 26.98 a 28.40 a 28.21 a 22.54 a 21.27 a 21.62 a b* value 24.77 b 24.18 b 26.98 a 26.58 a 21.89 ab 20.67 b 23.02 a Hue angle 42.56 a 41.89 a 43.56 a 43.29 a 44.18 b 44.23 b 46.80 a Chroma x 36.65 b 36.25 b 39.18 a 38.77 a 31.44 a 29.69 b 31.64 a
zWithin each harvest date (n = 15 fruit/treatment), values followed by different letters between ripeness stages are significantly different at the 5% level. YHue angle = Tan"' (b*/a*)
xChroma = (a*2 + b*2)1/2


'CPT-5' tomatoes exposed for 7 days of ethylene treatment ripened with significantly higher lightness coefficient (L* = 48.06) compared to tomatoes exposed for fewer days of ethylene treatment (Table 3.8). A similar increasing trend was evident for a* and chroma values. Tomatoes exposed for I or 2 days of ethylene treatment had significantly lower a* values (less red) (22.90 and 22.42, respectively) compared to those exposed for 6 or 7 days of treatment (25.84 and 25.99, respectively). Meanwhile, chroma and b* values were significantly higher in fruit exposed for 6 or 7 days of ethylene treatment (37.20 and 39.60, respectively) compared to the rest of the treatments. Hue angles showed no significant differences between treatments.






40



Table 3.8. Color quality parameters for table-ripe 'CPT-5' tomatoes harvested at green stage and supplied exogenous ethylene (100 pL/L) until breaker stage was attained.
Color Ethylene gassing exposure time
Parameters 1 day 2 day 3 day 4 day 5 day 6 day 7 day L* value 43.50 b 42.38 b 42.60 b 42.16 b 43.63 b 43.42 b 48.06 a a* value 22.90 b 22.42 b 23.99 ab 24.07 ab 24.32 ab 25.84 a 25.99 a b* value 25.20 bc 23.08 c 25.09 bc 25.24 bc 25.27 bc 26.66 b 29.87 a Hue Angle Y 47.70 a 45.69 a 46.33 a 46.35 a 46.15 a 45.81 a 49.02 a Chroma x 34.10 ed 32.24 d 34.78 cd 34.89 cd 35.09 cd 37.20 b 39.60 a z Values followed by different letters between ethylene exposure times (n = 15 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. Y Hue angle = Tan-' (b*/a*)
xChroma = (a*2 + b*21/2


For 'CPT-5' tomatoes there was no clear relationship between physical fruit measurements (firmness and diameter) and ripeness stage at harvest. Fruit size, a possible indicator of physiological maturity at harvest, proved to be a poor indicator of maturity. Similar to 'Agriset-761', there were considerable differences in the length of time required to attain ripe stage between harvest dates. Nonetheless, 'CPT-5' tomatoes in general required longer periods of time to ripen compared to 'Agriset-761' tomatoes harvested during the same dates and ripeness stages. Green-harvested tomatoes required longer times to ripen during the January experiment. Such observation is relevant since this could ultimately indicate a lesser degree of physiological maturity at harvest. Watada and Aulenbach (1979) concluded that green-harvested tomatoes exposed for more than 15 days to ripen at room temperature (230C) would probably result in inferior quality at ripe stage.

Soluble solids content varied widely between ripeness stages and harvest dates. Nonetheless, green-harvested tomatoes exposed for 6 and 7 days of ethylene had consistently higher soluble solids content compared to the other treatments, coinciding with an earlier report that, upon ripening, immature-green harvested had higher soluble






41

solids content at ripe stage (Kavanagh et al., 1986). In addition, pH values were considerably higher in tomato fruit exposed for extended ethylene treatment. Soluble solids content / titratable acidity ratios (SSC/TA) were considerably higher for greenharvested, ethylene-treated tomatoes from both 'Agriset-761' and 'CPT-5' cultivars (data not shown). However, except for 'Agriset-761' harvested in December 1995, greenharvested tomatoes without ethylene treatment also showed considerably higher SSC/TA ratios. This supports the contention that ethylene treatment doesn't influence sugar/acid ratios or flavor directly (Kader et al., 1978).

Compositional studies combined with sensorial analyses have demonstrated significant differences in sweetness, sourness and off-flavor perception without significant differences in soluble solids or sugar contents (Bisogni and Armbruster, 1976; Kader et al., 1977; Watada and Aulenbach, 1979). In contrast, significant differences in titratable acidity and pH do not necessarily reflect significant sensory quality changes. Changes in acidity of lesser than 0.1% citric acid or of less than 0.2 pH points were not perceived during sensory studies with model solutions (Gould, 1978).

Vitamin C contents for 'CPT-5' tomatoes were slightly greater than those found in 'Agriset-761' tomatoes regardless of ripeness stage or harvest dates. A trend of increasing vitamin C concentrations with increasing ripeness at harvest was evident for both cultivars. Except for tomatoes exposed for more than 5 days (CPT-5) and 6 days (Agriset-761) ethylene, green-harvested tomatoes with supplemental ethylene treatment had consistently higher vitamin C contents when compared to those fruits not treated with ethylene. This agreed with a previous study (Watada et al., 1976), in which greater variability in vitamin C content was due to cultivar differences rather than harvest






42

maturity. In addition, green-harvested tomatoes treated with ethylene were shown to have higher vitamin C contents when compared to those that were not treated.



Conclusions



Ripeness stage at harvest and harvest dates influenced compositional parameters and fruit color at table-ripe stage. However, no consistent relationship could be determined at table-ripe stage between ripeness stage at harvest and compositional quality. Soluble solids contents were highest in 'CPT-5' tomatoes harvested at green stage compared to later ripeness stages. Furthermore, green-harvested 'CPT-5' tomatoes that required 7 days of ethylene treatment to attain breaker stage had the highest soluble solids, pH, titratable acidity and total sugars compared to fruit exposed for shorter ethylene treatments. On the other hand, green-harvested 'Agriset-761' that required 7 days of ethylene treatment only had higher soluble solids compared to shorter ethylene exposure times.

Results for 'Agriset-761' and 'CPT-5' tomatoes suggest ripe fruit color was significantly affected by ripeness stage at harvest and harvest date. In general, lightness coefficients (L*) were lower, regardless of ripeness stage, for tomatoes harvested during December 1995 compared to subsequent harvests. Tomatoes harvested beyond the pink ripeness stage developed a darker red color compared to green or breaker-harvested fruit.






43







50
- Green
S Breaker
40 - Pink

4

S30



o 20



10



0
Agriset-761 CPT-5

Tomato Cultivars





Figure 3.1. Vitamin C content (mg/100 g fresh weight) from table-ripe 'Agriset-761' and 'CPT-5' tomatoes harvested at different ripeness stages in January 1996. Deviation bars represent the Duncan's critical range for significant differences at the 5% level according to Duncan's Multiple Range Test.






44








40

36 1 day E 2 days 32 3 days I-1 4 days 28 m 5 days S6 days 24 _7 days
24

4 20
o
o 16

6 12

8

4

0
Agriset-761 CPT-5 Tomato Cultivars




Figure 3.2. Vitamin C content (rg/ 100g fresh weight) from table-ripe 'Agriset-761' and 'CPT-5' tomatoes harvested at green stage and supplied exogenous ethylene (100 gL/L) for 1 to 7 days until breaker stage was attained. Deviation bars represent the Duncan's critical range for significant differences at the 5% level according to Duncan's Multiple Range Test.















CHAPTER 4
POTENTIAL FOR NONDESTRUCTIVE QUALITY SCREENING OF TOMATOES WITH ETHYLENE OR ELECTRONIC NOSE SENSOR ARRAYS



Introduction



Throughout tomato (Lycopersicon esculentum Mill.) production and marketing, efforts are made to maintain optimal visual quality (uniform color, absence of decay, proper firmness, etc.) to attract customers. As a consequence, disorders not readily detectable during sorting operations have received less attention. Visual appearance is a critical factor driving the initial purchase, but subsequent purchases are influenced greatly by eating quality (flavor, aroma and mouthfeel). Overall fresh tomato quality is determined not only by fruit appearance and firmness, but also flavor, aroma, and nutritive value.

Most fresh market tomatoes sold in U.S. supermarkets are harvested before they are "table-ripe" because retailing ripe tomato fruit is not practical within the current long distance handling and marketing system (Kader et al., 1978). In Florida, over 85% of the fresh market fruit are green-harvested for transport to distant markets. Such factors as harvesting immature-green fruit, incurrence of mechanical injuries, and storage at less than ideal temperatures contribute to quality loss from the shipping point to the consumer's table.




45






46

Recent literature has documented the effects of postharvest mismanagement on tomato quality. A negative effect from immature-green harvest was found after ripening regarding chemical composition, overall sensory acceptance, and aroma volatile profiles (Kader et al., 1978; Maul et al., 1997a). Internal bruising, a disruption of regular ripening of internal tomato tissues, occurs readily in breaker stage fruit (Sargent et al., 1989; 1992), and results in altered chemical composition, aroma volatile profiles, and sensory quality (Moretti et al., 1997; Sargent et al., 1997). Finally, low temperature storage (20C for 14 days) was shown to suppress the levels of six important aroma compounds in ripe tomato fruit (McDonald et al., 1996; Buttery et al., 1987). Distinctive external visual symptoms resulting from these treatments were never observed.

"Electronic nose" (EN) sensor arrays are used to classify specific samples of interest based on their headspace volatiles. The electronic nose consists of a series of non-specific chemical sensors, each of which shows characteristic responses to the volatile chemicals within the headspace over a sample (Anon., 1996). The electrical resistance of chemical sensors changes as the concentration of volatile compounds present in the headspace increases. These changes in electrical resistance (output) are then analyzed by a pattern recognition procedure, such as multivariate discriminant analysis (MVDA). MVDA identifies classification function(s) that maximize the dissimilarities between treatments, thus increasing the probability for accurate classification/prediction of unknown samples. The EN can be trained by a set of classification functions (canonical functions) to discriminate samples of unknown aroma quality characteristics.





47

The objectives of this study were to 1) determine the relationship between greenharvested tomato physiological maturity and ethylene treatment exposure time required for ripening initiation, then develop a predictive model; and 2) assess the ability of an electronic nose sensor array to accurately discriminate between intact tomato fruit harvested at different ripeness stages, or subjected to postharvest treatments (low temperature storage and impact bruising) with no apparent visual symptoms.



Materials and Methods




Experiment 1. Green tomato harvests were conducted on nine different occasions during the 1996 and 1997 seasons at commercial farms in seven different locations throughout the state of Florida. (Table 1). Tomatoes were harvested green (stage 1, 0% red color, USDA, 1976) following commercial harvesting guidelines, then transported to Gainesville on the day of harvest. Fruit were sorted for defects and then placed inside sealed chambers where a humidified ethylene/air mixture (100 pL/L C2H4) was administered using a flow-through system. Prior to ethylene treatment, a random sample (30 to 100 tomatoes/harvest) was set aside to determine the physiological maturity based on locule tissue development (M1-M4) according to Kader and Morris (1976). Immaturegreen tomatoes (MI+M2) ranged from no gel formation in any locule and cut seeds upon slicing (MI) to gel formation in one or two locules only (M2). Mature-green fruit ranged from gel formation in all locules with no red coloration (M3) to pink-red gel coloration

(M4). After 1, 3 and 5 days of ethylene treatment, tomatoes which had attained breaker






48

stage were removed from treatment and placed at 200C and 95% RH for subsequent ripening until the table-ripe stage.

In order to develop a regression model describing the relationship between harvest maturity and ethylene gassing exposure time, it was necessary to assume that immature-green fruit would require longer ethylene exposure to initiate ripening (Kader et al., 1978). Harvest maturity and ethylene exposure time distributions (% of population) were paired into three groups. Percent MI (immature-green) was compared to % fruit exposed to 5 or more days of ethylene gassing, % M2 (partially mature-green) and M3 (mature-green) were combined and compared to % of fruit exposed to 3 days of ethylene gassing. Finally, % M4 (advanced mature-green) fruit was compared to % fruit exposed to 1-day ethylene treatment. The statistical relationship between the resulting comparisons was estimated with a third-order polynomial regression equation, which yielded the best fit to the experimental data (R2) (STATISTICA v4.5, Statsoft Corp., 1994).

Experiment 2. For electronic nose (EN) analyses, 'Solimar' tomatoes (Asgrow Seed Co., Kalamazoo, MI) were harvested at green (stage 1), breaker (stage 2, <10% red color), turning (stage 3, 10-30% red color), light-red (stage 5, <90% red color), and red (stage 6, 100% red color) ripeness stages from a commercial field in Del Ray Beach, FL, on two separate occasions. Tomato fruit were transported the same day of harvest to Gainesville, FL, for sorting, postharvest treatments and electronic aroma sensing. A total of four separate experiments were conducted with the EN sensor array, each of them intended to explore the capability of electronic aroma sensing as a nondestructive means





49


of detecting physiological maturity, ripeness stages, low temperature storage and internal bruising.

Green-harvested tomato fruit (107.5 � 13.5 g) were individually placed (n=36) inside the sampling vessel of the EN sensor array for electronic aroma sensing. Immediately after electronic aroma sensing, each green tomato was sliced equatorially to assess locule tissue development and the fruits were classified into two groups (immature- and mature-green) based on their locule tissue (M1-M4) maturity distribution (Kader and Morris, 1976). Visual maturity classification was contrasted to EN classification during the pattern recognition analysis.

In a subsequent experiment, tomatoes (105.43 � 14.87 g) harvested at four ripeness stages (breaker, turning, light red, red) were placed inside the sampling vessel of the electronic nose sensor array for electronic aroma sensing of individual fruit (n=24).

The EN sensor array's capability to identify ripe tomatoes stored at low temperature prior to or during ripening (breaker or light-red stage) was explored utilizing two experiments. Green-harvested tomatoes were treated with ethylene (100PL/L) to initiate ripening, tomatoes exposed to 3 days of treatment were considered mature-green, while those exposed to more than 3 days of ethylene were considered immature-green at harvest. Following ethylene treatment, MG and IG tomatoes were placed at 50C for 7 days. In addition, tomatoes harvested at light-red stage were also stored at 50C for 7 days. Following the low temperature treatment, tomatoes were stored at 200C until they reached the table-ripe stage, prior to electronic aroma sensing. A second group of tomatoes harvested from the same farm plots was allowed to ripen continuously at 200C and used as controls. Fruits from both the low temperature (50C) and room temperature





50


(200C) treatments were at table-ripe stage (as described in Chapter 3) when analyzed by the EN sensor array.

Finally, breaker stage tomatoes (107.3 � 3.9 g) were dropped twice, on opposite sides, from a 40 cm height on to a flat solid surface to initiate internal bruising symptoms (cloudy, viscous locule tissue) according to Moretti et al. (1997). A separate group of breaker stage fruit were not bruised and used as controls. Both bruised and non-bruised tomatoes were allowed to ripen for 10 days at 200C (until the red stage) before electronic aroma sensing. Following the EN analysis, fruit were sliced equatorially to assess the presence of internal bruising symptoms where tomatoes had received impacts.

During electronic aroma sensing, intact tomatoes were placed individually inside the sampling vessel of an e-NOSE 4000 electronic nose (Neotronics Scientific Inc., Flowery Branch, GA). The EN analysis consisted of a three-step operation controlled by a PC connected to the e-NOSE 4000. First, the sampling vessel was purged with compressed air for 2 min to eliminate any extraneous odors present in the vessel. Second, the sampling head, which contained twelve conducting polymer sensors, was purged with compressed air for 4 min to eliminate any volatile compounds bound to the polymer sensors. During the purging of the head, the headspace inside the sampling vessel equilibrated with the volatile compounds given off by the tomato fruit. Finally, the conducting polymer sensors were lowered into the sampling vessel for 4 min to expose them to volatiles present in the headspace over the sample. The total time for the electronic aroma sensing procedure for each fruit was 10 minutes.

The change in electrical resistance output (i.e. sensor response) from the electronic nose was stored in a data base and later analyzed using multivariate






51

discriminant analysis (MVDA) from STATISTICA (v4.5, Statsoft Inc., 1994). This multivariate statistical procedure created linear discriminant functions (canonical functions) that maximized the differences between postharvest treatments and their controls, thus improving the probability for accurate classification based on En sensor outputs (cross-validation). Probabilities (posterior) for accurate classification based on EN outputs and Mahalanobis distances (distance between clusters or groupings, from a canonical analysis, adjusted for probability) were computed to compare the extent of differences across treatments. For graphical representation of treatment groupings based on canonical function(s) scores, histograms and canonical plots were utilized.



Results and Discussion




Experiment 1. There was considerable variation in the distribution of tomatoes classified as immature-green (MI), partially mature-green (M2), mature-green (M3) and advanced mature-green (M4) for the nine harvest dates and seven commercial tomato cultivars (Table 4.1). However, the distribution of maturity classes at harvest was highly correlated to the number of days ethylene treatment required to attain breaker stage (R2 = 0.84) (Figure 4.1). In general, MI fruit required longer ethylene gassing exposure (5days) whereas M4 fruit required the shortest treatment period (1-day).

The strong relationship between tomato maturity at harvest (MI-M4) and ethylene treatment required to initiate ripening demonstrated that ethylene gassing could be utilized as a nondestructive tool to indirectly assess the physiological maturity of a population of green tomato fruit at harvest. The proportion of immature-green (MI) fruit






52

(17% to 53.7%) correlated to the proportion of fruit exposed to 5 or more days of

ethylene treatment (6% to 68.6%). Chomchalow (1991) reported an average of 41%

immature-green tomatoes during several commercial harvests; such a proportion is

comparable to the 31.3% average found in this study. The onset of ethylene production

occurs at the mature-green stage (M3-M4) (Brecht et al., 1991), therefore, exogenous

ethylene treatment beyond physiological maturity might not further accelerate the onset

of ripening (breaker stage).


Table 4.1. Physiological maturity class at harvest (M1-M4) and response time to ethylene treatment required to attain breaker stage (in days) for seven commercial tomato cultivars harvested from different locations in the state of Florida.
Maturity at Harvest Ethylene Gassing
Harvest Tomato Growing Distribution (%) Exposure time (%)
Date Cultivar Area M4 M3 M2 M1 1- 3- 5day days days
4/25/96 Agriset-761 Naples 10.0 38.0 27.0 25.0 16.6 59.7 23.7 4/25/96 CPT-5 Naples 26.5 25.9 27.1 19.5 18.7 61.6 19.7 4/25/96 BHN-102 Naples 16.0 30.0 37.0 17.0 16.3 61.3 22.4 3/24/97 Solimar Del Ray 4.2 36.5 11.4 47.9 4.8 26.6 68.6 Beach
5/15/97 Solar Set Bradenton 8.9 10.5 26.9 53.7 8.6 53.0 38.4 6/16/97 Agriset-761 Gainesville 12.0 32.0 36.0 20.0 16.4 61.7 21.9 7/30/97 Mountain Quincy 6.9 17.2 34.5 41.4 23.9 70.1 6.0
Spring
9/26/97 BHN-189 Quincy 10.4 6.9 51.7 31.0 18.0 60.7 21.3 11/17/97 Solar Set Gainesville 6.6 16.7 50.0 26.7 3.6 64.3 32.1




Chomchalow (1991) reported that the proportion of immature-green tomatoes was

inversely related to fruit size at harvest; however, no consistent trends relating fruit mass

or diameter were found in this study (data not shown). This lack of dependable visual

indicators of physiological maturity for green tomatoes has also been demonstrated by

Kader et al. (1977). Advancements in automated sorting equipment permit the






53

determination of physical properties such as mass, volume or density on a commercial scale. The availability of this regression equation relating physiological maturity and ethylene exposure time could be readily implemented by commercial tomato growers and handlers to minimize the proportion of immature-green tomatoes being harvested and reduce ethylene treatment costs by improving the accuracy of gassing schedules.



Experiment 2. Electronic aroma sensing performed by the EN provided enough information for the MVDA pattern recognition procedure to find highly significant differences (P<0.00001) between intact mature- and immature-green 'Solimar' tomatoes. MVDA was used to assess the EN sensor array's capability to classify green-harvested tomatoes according to their physiological maturity based on volatile production. 'Soliniar' tomatoes were sliced immediately after EN analysis in order to visually classify them as immature- (MI+M2) and mature-green (M3+M4) based on their locule tissue development. The relationship between volatile compound production and locule tissue development was described by a single linear function (canonical function) from the MVDA. The distribution of scores from this canonical function was represented using a histogram where immature-green fruit grouped distinctly from mature-green ones (Fig. 4.2). The Mahalanobis distance (MD) of 4.37 units helped contrast the degree of difference between physiological maturity groupings, as perceived by the EN, with those between groupings from other experiments. MVDA from EN sensor outputs was capable of classifying green 'Solimar' tomatoes into immature (MI+M2) and mature (M3+M4) with accuracy averaging 98.5% and 96.7%, respectively. (Table 4.2).






54

The volatile compounds that contribute to the EN discriminating capabilities are not clearly understood from the sensor output data. Ethylene production has been reported to initiate at the mature-green stage (M3-M4) (Brecht, 1987), therefore the ability of EN sensors to detect ethylene would be a major factor to green tomato maturity discrimination. Production of numerous volatile compounds commences at the green stage, with lipid-derived aldehydes being prevalent (Baldwin at al., 1991c). In fact, lipidderived volatile production (hexanal, cis-3-hexenal, cis-3-hexenol and trans-2-hexenal) has been reported to be closely associated with ethylene production (Baldwin et al., 1991c). The contribution of ketone compounds to green-tomato volatile profiles is less significant, however, production of 6-methyl-5-hepten-2-one increased considerably between green and breaker stages (Buttery, 1993).

The MVDA from EN sensor outputs also identified very highly significant differences (P < 0.0033) between volatile compound production from tomatoes at different ripeness stages (breaker, turning, light red and red), then classified them into four distinct clusters. A two-dimensional canonical plot was utilized to visualize separation between the four groupings (Figure 4.3). The first two canonical functions from the MVDA were capable of explaining 95.4% of the differences in volatile compound production between ripeness stages. MVDA classified tomato ripeness stages based on EN sensor outputs with accuracy of 92.3% for breaker, 93.2% for turning, 98.2% for light-red, and 100% for red stage fruit. (Table 4.2). The MD between groupings demonstrated that the lowest degree of differences in volatile production occurred between breaker and turning stages (3.24 units), meanwhile, the greatest differences were between breaker and red stages (10.72 units).





55



Table 4.2. MVDA classification accuracy based on EN sensor outputs for 'Solimar' tomatoes at different ripeness stages. Green-harvested tomatoes were divided into immature (MI+M2) and mature (M3+M4) based on their locule tissue development.
Probability for correct classification Z
Maturity/Ripeness Number of Average Probability Probability Range
stage Tomatoes
Immature-green 22 98.5 80.3-100 Mature-green 14 96.7 57.9-100 Breaker 6 92.3 70.1-99.9 Turning 6 93.2 84.2-99.9 Light Red 6 98.2 89.6-100 Red 6 100 100-100 zClassification probabilities were based on the tomato samples analyzed.



Multivariate discriminant analysis successfully classified 'Solimar' tomatoes from different ripeness stages based on volatile profiles perceived by the EN sensor array (cross-validation of visual classifications). Such results show the potential of EN technology for nondestructive quality screening of fresh market tomatoes. Although the production of C02, ethylene and aroma volatiles of green tomatoes is significantly lower than that of red-ripe tomatoes (Baldwin et al., 1991a), the EN sensor array accurately discriminated immature from mature-green tomatoes. The high sensitivity to low volatile compound concentrations could be explained by the exponential reduction in electrical conductivity in the gas sensors as volatile compound concentrations increase (Benady et al., 1995).

During tomato ripening, considerable changes occur in chemical composition and visual appearance. The production of aroma volatiles undergoes considerable quantitative and qualitative changes throughout tomato ripening. Alcohols (ethanol, methanol, cis-3-hexenol) have been reported to show slight increases in concentration, whereas aldehydes (cis-3-hexenal, trans-2-hexenal, and hexanal) and ketones (acetone, 1-





56

penten-3-one, 6-methyl-5-hepten-2-one, and geranylacetone) nearly doubled their concentrations from the breaker through red stages (Baldwin et al., 1991a). Studies in aroma and flavor characterization of tomato fruit consider cultivars with high aroma volatile profiles as potentially having better aroma and flavor quality (Baldwin et al., 1991b). Tomato fruit show increased production of CO2 and ethylene at the M3-M4 stages and possibly upon bruising (MacLeod et al., 1976). Levels of ethylene produced by the fruit could also be affected by storage temperature (Brown et al., 1989). Ethylene production in combination with aroma volatile changes may have influenced the electronic nose sensor responses.

The EN analysis identified significant differences between ripe 'Solimar' tomatoes stored at 50 for 7 days at latter ripeness stage (light-red) and controls stored at 20oC (P<0.0014). The MD between storage temperature groupings (17.26 units) illustrated the large extent of the differences in volatile compound production found between the ripe tomatoes (Table 4.3). MVDA classification from EN sensor outputs had 100% accuracy when low temperature exposure occurred when the fruit were partiallyripe (Table 4.4). A histogram plot showed the distribution of canonical function scores grouped low temperature-stored tomatoes away from room temperature-stored tomatoes (Figure 4.4).

Even though the EN nondestructive analysis successfully classified ripe tomatoes stored at low temperature during early ripeness stages (breaker) the extent of differences between treatments was lower compared to those exposed to low temperature at late ripeness stages (light-red) (Figures 4.4 and 4.5). MVDA showed greater separation between clusters as a result of storage temperature than the separation due to






57


physiological maturity at harvest. The MD was 4.37 units between mature-green (MG)

and immature-green (IG) fruit, meanwhile, MD's between 200C and 50C storage

temperatures were 9.32 and 8.60 units for MG and IG fruit, respectively (Table 4.3). EN

polymer sensors showed significant output changes following exposure to the headspace

volatiles being produced by intact fruit from the different temperature treatments, notably

IG-harvested fruit that were subsequently exposed to low temperature storage induced

significantly higher outputs from all twelve EN sensors (data not shown). Classification

accuracy based on EN sensor outputs ranged between 96.2% and 100% for tomatoes

stored at low temperature during early ripeness stages.


Table 4.3. Mahalanobis distances between classification groupings for 'Solimar' tomatoes based on MVDA from EN sensor outputs. The distances between groupings were highly significant for all treatments (P < 0.01)
Between Grouping Centroids
From To Mahalanobis Distance z
Immature-green Mature-green 4.37 units
Breaker Turning 3.25 units Breaker Light red 5.35 units Breaker Red 10.72 units Turning Light red 4.68 units Turning Red 9.27 units Light-red Red 9.59 units
Light-red + 50C Light-red + 200C 17.26 units
Mature-green + 50C Mature-green + 200C 9.32 units
Immature-green + 5oC Immature-green + 200C 8.60 units
Internal-Bruised Non-bruised 9.15 units
Mahalanobis distances represent the separation between classification clusters from the MVDA adjusted for probability and were determined based on the tomato samples analyzed.


Differences in EN sensor outputs in response to volatile compounds produced by

tomatoes stored at low temperature (7 days at 50C) could be due, in part, to reduced

levels of important aroma volatile compounds. McDonald et al. (1996) found reduced

concentrations of hexanal, 6-methyl-5-hepten-2-one, geranylacetone, methanol, 2-






58

isobutylthiazole, and 1-nitro-2-phenylethane in ripe tomatoes that were held at 20C for 14 days at breaker stage. The sensitivity of tomato fruits to chilling injury has been shown to change during fruit ripening. Based on the expression of visual chilling injury symptoms, mature-green (Autio and Bramlage, 1986) and breaker-stage tomatoes (Hobson, 1987) were considered most sensitive to low temperatures. Under the current marketing system, green-harvested tomatoes are frequently exposed to temperatures below the chilling injury threshold (12oC) before the ripe stage (S.A. Sargent, personal communication). However, as evidenced by greater separation between EN classification groupings (higher MD), greater differences in volatile profiles occurred when tomatoes were exposed to low temperature at latter rather than earlier ripeness stages. This suggests that changes in volatile production induced by low temperature were partially reversed following subsequent ripening at 200C. Furthermore, sensory panelists found greater flavor differences in tomatoes exposed to low temperatures at latter ripeness stages, than those exposed to low temperature at early ripeness stages (breaker) (Sargent et al., 1997).

The volatile profile differences between bruised and non-bruised tomatoes, as perceived by the EN sensors, were also significant (P<0.0136). Based on canonical function scores, internal-bruised tomatoes were segregated from non-bruised fruits by 9.15 MD units (Figure 4.6). Incidence of internal bruising in ripe tomatoes, depicted by cloudy viscous locule tissue, was corroborated following EN analysis on tomatoes that were dropped. MVDA classification of bruised based on EN sensor outputs was highly accurate (99.6-97.9%).






59

Table 4.4. MVDA classification accuracy based on EN sensor outputs for 'Solimar' tomatoes following different postharvest treatments.
Probability for correct classification z
Postharvest Treatment Number of Average Probability Probability Range Tomatoes
Internal-Bruised 10 99.6 96.5-100 Non-Bruised 8 97.9 93.2-99.8 Light-red + 50C 6 100 100-100 Light-red + 200C 6 100 100-100 Mature-green + 50C 6 97.8 90.6-99.9 Mature-green + 200C 6 99.9 99.9-100 Immature-green + 5oC 6 99.9 99.9-100 Immature-green + 200C 6 99.7 99.4-99.9 Classification probabilities were based on the tomato samples analyzed.



The EN results concur with the findings of recent sensory and metabolite analysis studies. The incidence of internal bruising, a physiological disorder without apparent external visual symptoms, has been shown to stimulate increased ethylene and CO2 evolution, and to decrease titratable acidity (MacLeod 1976). Significant reductions in ascorbic acid and total carotenoid pigments in bruised locule tissue were contrasted by increased polygalacturonase activity and electrolyte efflux in bruised pericarp tissue (Moretti et al., 1998). In a recent study, sensory panelists were able to successfully distinguish between internal-bruised and non-bruised tomatoes, confirming significant changes in the aroma profiles and chemical composition from the locular gel and pericarp tissues as a consequence of internal bruising (Sargent et al., 1997). Possible changes in CO2, ethylene production and aroma profiles induced in internal-bruised tomatoes could explain the differences detected during electronic nose and sensory analyses.






60

Conclusions




The polynomial regression equation developed in this study relates harvest maturity and ripening response time under ethylene treatment (Figure 1). It could be implemented by commercial tomato growers to minimize the proportions immature-green fruit harvested and by packer/shippers to optimize harvesting and ethylene gassing schedules.

Results from these tests showed the potential use of electronic volatile sensing technology in non-destructive quality screening of fresh market tomatoes. The 400+ aroma compounds found in tomatoes are derived from lipids, amino acids, carotenoid pigments, and lignin-related compounds. The great potential for non-destructive quality screening of tomatoes based on volatile sensing lies in the diversity of biosynthetic pathways contributing to the formation of C02, ethylene and aroma volatile compounds. Tomato volatiles could perhaps tell the story of tomato abuse when no visual symptoms are available and, therefore, help fresh tomatoes reach consumers with consistently higher flavor quality.






61










o 90
S80 -y = 0.0002x3 - 0.0188x2 + 1.6007x - 3.1828 R = 0.8374
70
" 60
o 50

S40 * Days C2H4
30 - Poly. (Days C2H4)
20
9 10
o
0 20 40 60 80
Distribution of Maturity Classes (M1-M4)(%)






Figure 4.1. Polynomial regression equation describing the relationship between physiological maturity class distribution (%) at harvest (M1-M4) and the distribution of fruit (%) based on the length of ethylene treatment exposure time to attain breaker stage by green-harvested tomatoes from seven different commercial cultivars.






62

















W- Mature-green
5 - Expected mature-gr.
E Immature-green
4 Expected immature-gr




0







-4 -2 0 2 4 6 Canonical Score Distribution

Figure 4.2. Comparison between immature-green and mature-green tomato fruit based on electronic aroma sensing. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of immature- and mature-green tomato fruit. Classification of immature- and mature-green tomatoes was very highly significant (P < 0.00001).






63
















5
4 3

N O
.21
. 1 o o


- -1
0
5 -2 0 Turning
o -3 13 Light-red E3 0
-4 O Red
-5 A Breaker

-8 -6 -4 -2 0 2 4 6 8 10 Canonical Function 1




Figure 4.3. Comparison between breaker, turning, light-red and red ripeness stages for tomatoes based on electronic aroma sensing. Canonical plot represents the classification of canonical scores obtained after EN sensor outputs analyzed using the first 2 canonical functions. The ellipses around ripeness stage groupings represent the 95% confidence intervals. Classification of tomato ripeness stages was highly significant (P < 0.0033).






64
















2.0 / Low Temperature (5C) ..------.... Expected Low Temp.
I-- Room Temp. (20C)
1. - Expected Room Temp.
[1.5

1.0





0.5
-12 -8 -4 0 4 8 Canonical Score Distribution



Figure 4.4. Comparison between low temperature-stored (50C) and room temperaturestored (200C) light-red 'Solimar' tomato fruit based on electronic aroma sensing. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of low temperature- and room temperature-stored tomato fruit. Classification of low temperature- and room temperature-stored tomatoes was very highly significant (P <
0.0014).






65















7 6 5
4
3 M6

2 OO O
_0

'a -2 MG + 200C U-3 MG + 50C
-4 IG + 200C -5 A IG + 50C
-6
-7
-14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 Canonical Function I



Figure 4.5. Comparison between ripe 'Solimar' tomatoes exposed to low temperature (50C for 7 days) at early stages of ripening (breaker). Based on their ethylene treatment exposure time, tomatoes were divided into immature-green (IG) or mature-green (MG) at harvest. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of tomatoes stored at 50C or 200C. Classification of low temperature-stored tomatoes was significant (P < 0.0136).






66
















5 M Internal-bruised
------- Expected Int-bruised
4I- Non-bruised
S- Expected Non-bruised

F- 3







-8 -4 0 4 8 Canonical Score Distribution



Figure 4.6. Comparison between internal-bruised and non-bruised tomatoes based on electronic aroma sensing. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of immature- and mature-green tomato fruit. Classification of internal-bruised and non-bruised tomatoes was very highly significant (P < 0.00001).















CHAPTER 5
AROMA VOLATILE PROFILES AND RIPE TOMATO FLAVOR ARE
INFLUENCED BY PHYSIOLOGICAL MATURITY AT HARVEST




Introduction



Consumer dissatisfaction with fresh tomato flavor promoted extensive research during the late 1970's and early 1980's. Tomatoes harvested at early stages of maturity were perceived as less sweet, more sour, and as having less tomato-like flavor compared to those harvested more mature (Kader et al., 1977). Furthermore, vine-ripe fruits were considered sweeter by sensory panels without significant differences in soluble solids content or dry matter (Watada and Aulenbach, 1979). This observation highlighted the possible relevance of aroma compounds on tomato flavor perception. Physiological maturity of green-harvested tomatoes has been circumstantially related to ripe, fresh tomato flavor and quality by numerous researchers (Kader et al., 1977; Watada and Aulenbach, 1979; Al-Shaibani and Greig, 1979). Nonetheless, chemical composition data collected by these authors did not always concur with sensory panel results.

Over 85% of Florida's fresh market tomatoes are harvested commercially at the green stage and are exposed to exogenous ethylene treatment to accelerate the onset of ripening processes. Difficulty in assessing harvest maturity leads to 25-40% of tomatoes being harvested at immature-green stage (Chomchalow, 1991), leading to compromised



67






68

quality when ripe. Without an accurate and dependable non-destructive method of screening inferior quality immature fruits in green-harvest tomato operations, growers cannot provide the consistent quality that the markets demand. The development of locule tissue has been proposed as an accurate way to separate immature-green (M1+M2) from mature-green tomato fruits (M3+M4) (Kader and Morris, 1976). Data collected in our laboratory over the past two years showed that there is a strong direct relationship between immature-green fruit (M1-M2) and extended exposure to ethylene gas required to initiate ripening.

The objectives for these experiments were: 1) to document changes in important volatile compounds from different fruit tissues as affected by fruit maturity at harvest; 2) to explore the potential use of electronic nose volatile sensing technology (EN) as a nondestructive tool to screen ripe tomato fruits harvested at immature- and mature-green stages; and 3) to identify and quantify tomato flavor and aroma differences utilizing difference discrimination tests and descriptive sensory panels.



Materials and Methods


In a series of experiments aimed at determining the effects of physiological maturity of green-harvested tomatoes on their quality at table-ripe stage, information was gathered from several commercial cultivars grown in different parts of the state of Florida. 'Agriset-761' (Agrisales Inc., Plant City, FL) and 'CPT-5' (Collier Farms, Naples, FL) tomatoes were harvested from experimental plots in Naples, FL; 'Solimar' (Asgrow Seed Co., Kalamazoo, MI) tomatoes harvested on two separate occasions from commercial plots in Palm Beach county, FL; and 'BHN-189' (BHN Research Inc.,






69


Naples, FL) tomatoes harvested from a commercial farm in Quincy, FL. Following all harvests, tomatoes were transported to Gainesville, FL for postharvest treatments and subsequent ripening. Upon arrival, green tomatoes were sorted for defects, washed with chlorinated water, and towel dried. Prior to ethylene treatment, a random sample of fruits (n=50-100) were sliced equatorially to assess the physiological maturity distribution of the green tomato population. The physiological maturity stages were rated based on locule tissue development (M1-M4) following the internal maturity standards for tomatoes proposed by Kader and Morris (1976). Tomatoes were stored at 200C and gassed in bulk with a humidified, 100 .L/L ethylene/air mixture in sealed chambers connected to a flow-through system. The flow rate calculation was based on maturegreen tomato respiratory rates for 20oC (mg CO2/kg fresh weight/hour) (Hardenburg et al., 1986) and total fruit mass (kg) inside each chamber.

After 1, 3 and 5 days of ethylene treatment, tomatoes that attained breaker stage were removed from the gassing chamber and placed at 200C and 95% RH for subsequent ripening. In order to compare tomatoes harvested at latter ripeness stages with those harvested at green stage using descriptive sensory panels, 'BHN-189' and 'Solimar' tomatoes were harvested at light-red stage (stage 5) from the same plots where green fruit were harvested approximately 8 days before. Tomatoes harvested at light-red stage were subsequently ripened along with the green-harvested tomatoes.

Tomatoes were considered table-ripe upon attaining red stage (stage 6, >90% red coloration) and noticeable firmness loss (3-4 mm deformation threshold) measured by applying a constant 9.8 N force for 5 seconds on the fruit equator (Gull et al., 1980). Table-ripe tomatoes were removed from storage and utilized for compositional, aroma






70

volatile, electronic nose and sensory analyses. In addition to the number of days required to attain table-ripe stage following ethylene treatment, physical traits characterizing tomatoes with different ethylene gassing exposure times were documented by recording fruit mass (g), fruit equatorial diameter (mm) and locule tissue content (% fresh weight) for 'Agriset-761' and 'CPT-5' tomatoes. During the first harvest of 'Solimar' tomatoes, the compositional differences between individual fruit tissues (locule, pericarp and whole fruit) were documented by excising the tissues and conducting compositional and GC analyses (Experiment 1).

Samples from table-ripe tomatoes that required different ethylene treatments (1, 3, and 5 days) and tomatoes harvested at light-red stage ('BHN-189' and 'Solimar') were collected for compositional analyses. Six composite samples (3 fruit/sample) from each treatment and fruit tissue were homogenized and centrifuged at 18,000 X gn and 50C temperature. For vitamin C assays, homogenates (2 g/sample) were stabilized using 20mL of acid mixture (6% HPO3 containing 2N acetic acid in water) prior to centrifugation. The supernatant was filtered using cheesecloth, stored inside scintillation vials and frozen at -200C for later analysis. Titratable acidity, expressed as % citric acid, was determined by titrating 4 g of tomato supernatant to 8.2 pH with a 0.1 N NaOH solution using an automatic titrimeter (Fisher Scientific, Pittsburgh, PA). Soluble solids content, expressed as oBrix, was measured using a tabletop digital refractometer (Abbe Mark II, ReichartJung, Buffalo, NY) and pH measurements were conducted using a digital pH-meter (Coming model 140).

Tomato fruit lycopene content was determined using a colorimetric method adapted from Umiel and Gabelman (1971). Four 10-g samples of tomato homogenate






71


were mixed with 30 mL of acetone inside 100-mL glass vials covered with aluminum foil to minimize light-induced lycopene breakdown. After mixing for 60 sec using a Polytron blender, the samples were vacuum filtered (Whatman # 4 filter paper) into 500 mL sidearm Erlenmeyer flasks containing 45 mL of hexane. The hexane-lycopene phase was separated from the acetone through a series of deionized water washes using separatory funnels. Lycopene absorbance was read at 503 nm using a spectrophotometer (Beckman DU-20, Irvine, CA). Lycopene concentrations were determined using a standard curve derived from pure lycopene standards.

Vitamin C and total soluble sugars were analyzed using spectrophotometric methods adapted from Terada et al. (1978) and Dubois et al. (1956), respectively, as described in Chapter 3. Individual sugar analysis (glucose and fructose) was performed for 'BHN-189' and 'Solimar' samples analyzed during the descriptive sensory panels using an adaptation of the high performance liquid chromatography (HPLC) method described by Baldwin et al (1991c). Approximately 20 g of tomato homogenate were extracted using 35 mL of 80% ethanol/deionized water solution. The homogenate/ethanol mixture was boiled for 15 min, then cooled prior to filtration (Whatman # 4 filter paper). The filtered solution was brought up to a 50-mL volume with 80% ethanol inside a volumetric flask. Ten mL of the filtered solution were then passed through a C-18 Sep-pak (Waters/Millipore, Milford, MA) then filtered through a 0.45 jpm Millipore filter. Sugars were analyzed using HPLC refractive index detector with a Waters Sugar Pak column and a 104 M ethylenediaminetetraacetic acid disodium calcium salt (CaEDTA) mobile phase (0.5 mL-min-1 flow rate at 900C). Glucose and fructose concentrations were converted to sucrose equivalents (Koehler and Kays, 1991),






72

where values were multiplied by 0.74 and 1.73, respectively to better represent individual hexose sweetness perception potential.

Tomato volatile compounds were identified and quantified using a gas chromatography (GC) headspace analysis technique (Baldwin et al., 1991b). Four samples of homogenate (40 mL) from each tomato treatment were combined with 10-mL of saturated CaC12 solution, blended for 10 seconds, immediately frozen using liquid nitrogen and stored at -800C. The saturated CaCl2 solution was added to the tomato samples to help reduce enzymatic changes that might induce quantitative and qualitative changes in the tomato sample's volatile profile following tissue maceration (Buttery and Ling, 1993). For GC analysis, each tomato sample was thawed under running tap water and a 2-mL sample placed inside a 6-mL vial sealed with a crimp-top and Teflon/silicone septum. Vials were heated rapidly to 80C and incubated for 15 minutes before injection to a Perkin Elmer HS-6 headspace sampler heating block. The analysis was carried out using a Perkin Elmer Model 8500 gas chromatograph equipped with a 0.53 mm X 30 m polar stabilwax capillary column (1.0-jm film thickness, Restek Corp., Bellefonte, Pa.) and a flame ionization detector. Column oven temperature was held at 400C for 6 min, then raised to 1800C at a rate of 60C/min. The resulting GC peak heights for 16 important aroma volatile compounds (acetaldehyde, acetone, methanol, ethanol, 1-penten-3-one, hexanal, cis-3-hexenal, 2+3-methylbutanol, trans-2-hexenal, trans-2-heptenal, 6-methyl5-hepten-2-one, cis-3-hexenol, 1-nitro-2-phenylethane, geranylacetone, 2isobutylthiazole, and 3-ionone) as reported in previous research (McDonald et al., 1996; Baldwin et al., 1991ab, Buttery et al., 1988; Petro-Turza, 1987) were quantified in tL/L






73

utilizing standard curves determined by enrichment of bland tomato homogenate with authentic volatile compound standards (Baldwin et al., 1991abc).

In addition to GC and compositional analyses, an electronic nose (EN) sensor array was utilized to discriminate treatments based on their volatile production. The EN consisted of a sampling head equipped with 12 polymer sensors, glass sampling vessel and purging valves (e-NOSE-4000, Neotronics Scientific Inc., Flowery Branch, GA). Each individual polymer sensor changed its electrical conductivity upon exposure to volatile compounds present inside the headspace of the sampling vessel. A computer recorded the sensor outputs over time.

EN analysis consisted of placing individual 20 mL samples of fruit homogenate inside the sampling vessel (n=6 homogenate samples/treatment) and sealing it against the sensor head. Approximately 20 g from each frozen tomato sample was placed inside 113mL plastic cups, lidded and thawed in a 25oC water bath. Immediately upon thawing, the lid was removed and the sample cup placed inside the glass vessel of an electronic nose (e-NOSE 4000, Neotronics Scientific, Flowery Branch, GA). EN analysis and sensor data acquisition was controlled by personal computer. EN sampling began with a 2-min purge of the glass vessel using compressed air. Next, the sampling head, containing twelve polymer sensors (manufacturer ID numbers: T301, T298, T297, T283, T278, T264, T263, T262, T261, T260, T259, and T258), was purged with compressed air for 4 min to eliminate any volatile compounds in contact with the polymer sensors, while volatile compounds from the tomato sample equilibrated inside the sealed sampling vessel. Both sampling vessel and sensor head were purged using compressed air at a 400 mL/min flow rate. Finally, the sensor head was lowered automatically into the sampling





74


vessel to expose the polymer sensors to the volatile compounds produced by the sample in the headspace for an additional 4 min. EN analysis was carried out at room temperature (ca. 250C) and relative humidity inside the sampling vessel was also recorded. The EN analysis lasted a total of 10 min per sample.

For preliminary sensory analyses, a "difference from control" test was chosen because of its ability to identify overall differences between samples (when compared against a control), while allowing panelists to rate the extent of those differences by including a hidden control sample within the treatments. Samples from ripe 'Agriset-761' and 'CPT-5' tomatoes (Experiment 2) were presented to a group of 27 untrained panelists on two separate panel sessions, one session for each cultivar. Samples from tomatoes exposed to 1-day ethylene gassing to attain breaker stage were chosen as controls, and compared to tomatoes exposed to 3 and 5 days ethylene treatment. Panelists were asked to rate the degree of difference they perceived between the control sample and three other samples (1-day "hidden control", 3 and 5 days ethylene exposure time). A 12-point scale with verbal descriptors on either end was used to rate the extent of difference in the sensory test ballots (1 = no difference and 12 = extremely different). All samples were presented in plastic cups labeled with random numbers and presented in random order to panelists to avoid biased results. Following the sensory test sessions, the results from the ballots were computed, and in both cases, some degree of panelist screening was required when panelists rated the hidden control (1-day ethylene) extremely different than the control sample (same treatment).

Based on the results from the "difference from control" sensory tests, a trained descriptive sensory panel was assembled. To organize and train a descriptive sensory






75

panel, a group of 20 volunteers showing no dislikes for tomatoes were screened for proper sensory perception with the use of citric acid and sucrose solutions of varying concentrations. Over a period of three months, the group was reduced to 16 panelists, 10 males and 6 females comprising ages between 20 and 65 years old, and were trained to describe flavor and aroma attributes from fresh market tomatoes.

Initially panelists were screened for proper sensory perception using sucrose, citric acid and sucrose/citric acid solutions of varying concentrations. During the initial training sessions, panelists were presented with a variety of tomato samples representing effects of ripeness stage, storage temperature and cultivar on characteristic tomato flavor. After having familiarized panelists with a wide range of tomato samples during the first five training sessions, the panel leader compiled a descriptor list from published literature on tomato flavor to aid panelists in verbalizing flavor and aroma characters perceived in the samples. The panel reached a consensus on five flavor attributes (typical tomato, sweetness, sourness, green/grassy and off-flavor) and two aroma attributes (ripe tomato and off-odor). Descriptor intensity was rated using a 150-mm, unstructured line scale with a low intensity on the zero (left) side and high intensity on the 150 (right) side as anchor terms.

Approximately 20 minutes before sensory analysis, 'BHN-189' and 'Solimar' whole tomato samples (Experiment 3) were chopped into a coarse puree using 8 to 10 pulses of a food processor (M. Einstein, personal communication). Two tablespoons (ca 40-50 g) of tomato puree were placed in 113-mL plastic cups, sealed with lids and labeled with a two-digit random number. Evaluations were conducted in individual booths with dim lighting and samples were presented in random order. Panelists were






76

instructed to open the lid from each tomato sample cup to rate the aroma descriptors, then to proceed with the flavor descriptors. Water and unsalted crackers were provided for panelists to rinse their palates between samples. In any given session, panelists were asked to rate the flavor attributes of 4 to 6 tomato samples. During sensory analysis, four 40-mL samples of tomato homogenate for GC and EN analyses were combined with 10mL of a saturated CaCl2 solution, blended at high speed for 10 seconds and flash-frozen with liquid nitrogen. CaCl2 was added to reduce enzymatic activity that could contribute to further volatile changes following tissue maceration and subsequent storage at -800C (Buttery and Ling, 1993).

Descriptive sensory panel scores for tomato flavor and aroma descriptors were analyzed as complete block design with panelists as blocks and maturity at harvest as treatments using GLM procedure of SAS (v6.12, SAS Institute, Cary, NC). All compositional, HPLC, and GC data were analyzed using multiple analysis of variance (MANOVA), Duncan's Multiple Range Test for means separation using SAS. Meanwhile, EN sensor outputs were analyzed using multivariate discriminant analysis (MVDA) with STATISTICA ( v4.5, Statsoft, Inc., Tulsa, OK). The differences between volatile profiles found in ripe tomato samples were identified and visualized through MVDA, which created two-dimensional canonical plots where descriptive linear functions (canonical functions) classified the different tomato samples based on the pattern of outputs from the polymer sensors In certain cases a forward stepwise procedure was utilized to optimize the number of variables considered for the descriptive linear functions, thus simplifying data collection and improving statistical significance. Relationships between instrumental and sensory parameters were explored through the






77


use of correlation matrices (PROC CORR) and stepwise regressions (PROC STEPWISE) using SAS.



Results and Discussion




'Solimar' (Experiment 1). The effects of maturity stage at harvest on tomato fruit tissue (locule and pericarp) aroma volatile profiles and chemical composition were documented in table-ripe 'Solimar' fruit. Even though no significant differences in pH were found for either locule or pericarp tissues, increasing pH values with increasing ethylene exposure time were documented for both tissues. Increasing pH values became significant in whole tomato samples exposed to 5 days ethylene when compared to those exposed to 1 day (4.24 and 4.20, respectively) (Table 5.1). No relationship between pH values and titratable acidity was evident, as pH values increased, titratable acidity either decreased (locule) or increased (whole) with increasing ethylene exposure time. Soluble solids were significantly higher in pericarp and whole tomato samples exposed to 5 days ethylene treatment (4.280 and 4.280Brix, respectively) compared to those exposed to 1 and 3 days (whole fruit) ethylene treatment (3.950, 4.100 and 4.120 Brix for pericarp and whole fruit samples respectively). No significant differences in vitamin C content were found between ethylene exposure times for any of the tomato tissues. However, reduced vitamin C content with increasing ethylene exposure time as observed in locule and whole fruit samples (Table 5.1).






78

Table 5.1. Table-ripe chemical composition parameters for green-harvested 'Solimar' tomatoes that required 1, 3 or 5 days of ethylene treatment to attain breaker stage.
Compositional Parameters z Ethylene Gassing Exposure time
1-Day 3-Days 5-Days Whole Fruit
pH 4.20 b 4.21 ab 4.24 a Soluble solids content (oBrix) 4.10 b 4.12 b 4.28 a Titratable acidity (% citric acid) 0.78 a 0.84 a 0.90 a Vitamin C (mg/100g fresh wt.) 13.8 a 13.6 a 12.1 a Locule Tissue
pH 4.41 a 4.38 a 4.44 a Soluble solids content (OBrix) 3.90 a 3.72 a 3.98 a Titratable acidity (% Citric acid) 1.00 a 0.97 a 0.94 a Vitamin C (mg/100g fresh wt.) 24.6 a 18.6 a 16.8 a Pericarp Tissue
pH 4.17 a 4.18 a 4.19 a Soluble solids content (oBrix) 3.95 b 4.12 ab 4.28 a Titratable acidity (% Citric acid) 0.72 a 0.74 a 0.71 a Vitamin C (mg/100g fresh wt.) 12.7 a 18.6 a 16.8 a
z Means for parameters with different letters within rows were significantly different at the 5% level according to Duncan's Multiple Range Test. (n=6 composite samples/trt).


Aroma volatile compound concentrations were generally lower in table-ripe

'Solimar' tomatoes that required extended ethylene treatments to attain breaker stage.

Five of 16 aroma volatile compounds showed significant differences between treatments.

Table-ripe tomatoes that required 3 or 5 days ethylene treatment had significantly lower

levels of 1-penten-3-one, cis-3-hexenal, 6-methyl-5-hepten-2-one, 2-isobutythiazole and

geranylacetone. In addition, tomatoes exposed to 5 days of ethylene treatment had

significantly lower levels ofcis-3-hexenal and geranylacetone compared to those exposed

to 3 days of treatment (Table 5.2).

Differences in aroma volatile compound concentrations for pericarp tissue

samples from 'Solimar' tomatoes followed a similar trend to those observed for whole

fruit homogenate samples. Pericarp samples from tomatoes exposed to 5 days of

ethylene treatment had significantly lower concentrations of 1-penten-3-one, trans-2-






79


heptenal, 6-methyl-5-hepten-2-one, 2-isobutylthiazole, geranylacetone and 13-ionone

compared to those exposed to I or 3 days of ethylene treatment (Table 5.3). Pericarp

tissue samples from tomatoes exposed to I day of ethylene treatment had significantly

higher concentrations of cis-3-hexenal compared to pericarp samples from tomatoes

exposed to 3 or 5 days of ethylene treatment. Concentrations of 2+3 methylbutanol were

highest in pericarp tissue samples from tomatoes exposed to 3 days of ethylene treatment

(Table 5.3).


Table 5.2. Aroma volatile compound concentrations for whole table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage.
Aroma Volatile Compounds Z Ethylene gassing exposure time
1 day 3 days 5 days Acetaldehyde 12.64 a 13.54 a 12.04 a Acetone 0.45 a 0.48 a 0.41 a
Methanol 310.82 a 304.13 a 307.41 a Ethanol 16.80 a 15.98 a 16.67 a 1-Penten-3-one 0.27 a 0.19 b 0.17 b Hexanal 12.33 a 8.22 a 9.19 a Cis-3-hexenal 9.04 a 6.18 b 4.56 c 2+3-Methylbutanol 2.16 a 2.26 a 1.79 a Trans-2-hexenal 8.59 a 7.51 a 7.03 a Trans-2-heptenal 0.04 a 0.03 a 0.03 a 6-Methyl-5-hepten-2-one 0.83 a 0.55 b 0.42 b Cis-3-hexenol 0.05 a 0.04 a 0.05 a 2-Isobutylthiazole 0.09 a 0.06 b 0.05 b 1-Nitro-2-phenylethane 0.06 a 0.06 a 0.05 a Geranylacetone 6.90 a 4.26 b 2.25 c 1-ionone 0.12 a 0.09 a 0.06 a
Total 381.18 363.58 362.65
z Means for aroma volatile compounds (pLL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. Y Total aroma volatile concentrations (piL/ L) based on the sum of the 16 compounds quantified.


Aroma volatile compound concentrations in table-ripe 'Solimar' locule tissue

samples showed significant differences in 5 of 16 compounds quantified. Locule tissue

samples from tomatoes exposed to 1 day of ethylene treatment had significantly higher






80


levels of 1-penten-3-one, 2-isobutylthiazole, 1-nitro-2-phenylethane, and 3-ionone.

Conversely, cis-3-hexenol levels were significantly higher for locule tissue samples from

tomatoes exposed to 5 days of ethylene treatment (Table 5.4).


Table 5.3. Aroma volatile compound concentrations for pericarp tissue (including columnella) from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage.
Ethylene gassing exposure time
Aroma Volatile Compounds z 1 day 3 days 5 days Acetaldehyde 14.53 a 16.27 a 11.97 a Acetone 0.45 a 0.47 a 0.38 a
Methanol 340.43 a 379.90 a 375.78 a Ethanol 22.84 a 18.06 a 17.99 a 1-Penten-3-one 0.21 a 0.18 a 0.14 b Hexanal 19.43 a 14.09 a 14.90 a Cis-3-hexenal 10.96 a 7.57 b 6.45 b 2+3-Methylbutanol 2.60 b 3.31 a 2.18 b Trans-2-hexenal 9.78 a 8.66 a 6.98 a Trans-2-heptanal 0.05 a 0.04 a 0.03 b Cis-3-hexenol 0.10 a 0.09 a 0.07 a 6-Methyl-5-hepten-2-one 0.95 a 0.76 a 0.50 b 2-Isobutylthiazole 0.11 a 0.09 a 0.05 b 1 -Nitro-2-phenylethane 0.06 a 0.06 a 0.06 a Geranylacetone 7.55 a 5.92 a 3.66 b 1-ionone 0.13 a 0.11 a 0.05 b
Total Y 430.64 455.56 441.19
Means for aroma volatile compounds (lpL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. Y Total aroma volatile concentrations (jiL/ L) based on the sum of the 16 compounds quantified.


In general, table-ripe 'Solimar' tomatoes harvested at immature-green stage

(exposed to >3 days of ethylene treatment) had significantly lower production of the

aroma volatile compounds quantified. Pericarp tissue samples produced approximately a

20% higher concentration of aroma volatile compounds compared to whole fruit samples

(443 and 369 pL/L, respectively). Buttery et al. (1988) found similar results and

suggested that higher volatile production found in pericarp tissue could be due in part to






81


tissue damage during separation. Tissue wounding would induce lipoxygenase activity,

an important enzyme in the biosynthetic pathway of aldehydes and alcohols from

membrane fatty acids (Hilderbrand, 1989). On the other hand, aroma volatile production

in locule tissue samples was 43% lower than whole tomato samples (203.30 gL/L and

369 pL/L, respectively). In fresh market tomatoes, locule tissue constitutes 14.4% to

34% of the fresh weight (Stevens et al., 1977), thus showing a lesser contribution of

locule tissue to whole fruit aroma volatile production compared to pericarp tissue.


Table 5.4. Aroma volatile compound concentrations for locule tissue from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage.
Aroma Volatile Compounds z Ethylene gassing exposure time
1 day 3 days 5 days Acetaldehyde 19.95 a 17.43 a 16.45 a Acetone 0.49 a 0.48 a 0.47 a
Methanol 146.63 a 144.74 a 136.33 a Ethanol 18.84 a 19.33 a 18.78 a 1-Penten-3-one 0.18 a 0.12 b 0.14 b Hexanal 4.10 a 3.02 a 4.78 a Cis-3-hexenal 6.53 a 4.89 a 5,59 a 2+3-Methylbutanol 2.38 a 2.66 a 2.95 a Trans-2-hexenal 6.83 a 5.99 a 7.55 a Trans-2-heptenal 0.03 a 0.03 a 0.03 a 6-Methyl-5-hepten-2-one 0.64 a 0.52 a 0.56 a Cis-3-hexenol 0.07 b 0.09 ab 0.14 a 2-Isobutylthiazole 0.11 a 0.63 b 0.07 b 1-Nitro-2-phenylethane 0.07 a 0.06 b 0.06 b Geranylacetone 2.98 a 2.69 a 2.72 a f3-ionone 0.22 a 0.17 ab 0.08 b Total compound cone. Y 210.05 202.28 196.68
z Means for aroma volatile compounds (LL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. Y Total aroma volatile concentrations (pL/ L) based on the sum of the 16 compounds quantified.



Significant reductions in the concentration of p-ionone, cis-3-hexenal, and 1penten-3-one found in table-ripe tomatoes exposed to extended ethylene treatment






82


deserve special attention because of their low odor thresholds (Buttery et al., 1988). The odor unit values (OUV) represent a ratio between the concentration of an aroma compound present in a sample and the threshold concentration required for its sensory detection. A high OUV for a volatile compound would imply that small changes in its concentration would have greater impact during sensory evaluation, than similar magnitude changes in a volatile compound with low OUV. The levels of 3-ionone, a volatile compound with one of the highest OUV among tomato compounds, dropped by as much as 70% in locule tissue samples which required extended ethylene treatment (5 days). In addition, geranylacetone levels dropped at least 50% in pericarp and whole tomato samples exposed to 1 day of ethylene treatment compared to those exposed to 5 days.

Even though the levels of hexanal respectively dropped 33.4% and 27.5% in respective whole fruit and pericarp samples exposed to 3 days of ethylene gassing, those differences were not significant compared to 1 day treatment samples. It is important to note that methanol and hexanal concentrations were approximately 100% and 300% higher in the pericarp homogenate than in the locule tissue homogenate. Higher concentrations of hexanal present in the pericarp tissue could suggest its higher contribution to the perception of green/grassy flavor (Petro-Turza, 1987; Tandon, 1998) in whole tomato fruit. Cis-3-hexenol levels decreased significantly in pericarp samples that required extended ethylene treatment. Conversely, cis-3-hexenol concentrations increased with increasing ethylene exposure time in locule tissue samples. The possibility of opposite trends in the production of volatile compounds from different tomato tissues could explain lesser extent of volatile concentration difference found in whole tomato






83

samples compared to pericarp samples. Three important aroma volatile compounds showed significant reductions in all three tomato tissues analyzed, 1-penten-3-one, (3ionone, and 2-isobutylthiazole, the latter deserving special attention since it is a compound unique to tomato aroma (Petro-Turza, 1987).

Reported aroma volatile concentrations from different tomato tissues (Buttery et al., 1988) do not entirely concur with our present results. Contrary to previous reports on the contribution of excised tissues to whole fruit volatile profiles, the concentrations of 6methyl-5-hepten-2-one and geranylacetone were lower in locule tissue samples from tomatoes exposed to 1 day of ethylene treatment when compared to pericarp tissue samples. However, locule tissue samples that required extended ethylene treatment produced slightly higher levels of both compounds than pericarp samples. The extent of these discrepancies might be related to cultivar variability as was shown by Baldwin et al. (1991b) and by maturity at harvest as suggested in this study. Nevertheless, the absolute levels of P-ionone, geranylacetone, 2+3 methylbutanol and 6-methyl-5-hepten-2-one were considerably higher in our results compared to those reported by Buttery et al. (1988).



'Agriset-761' and 'CPT-5' (Experiment 2). Table-ripe 'Agriset-761' and 'CPT-5' tomatoes were utilized to document compositional and aroma volatile concentration changes, as related to ethylene exposure time (physiological maturity), and their possible effects on sensory quality. 'Agriset-761' tomatoes that required 1 day of ethylene treatment had significantly lower mass than those exposed to 3 or 5 days (195.5, 216.8 and 230.6 g, respectively). On the other hand, 'CPT-5' fruit exposed to 3 days ethylene





84


gassing had significantly greater mass when compared to either 1 day or 5 days fruit (242.7, 195.2 and 204.9 g, respectively). Locule tissue content (% fresh weight) showed no significant differences between treatments for either cultivar, however, 'Agriset-761' tomatoes had slightly lower locule tissue contents when compared to 'CPT-5' fruits (19.8% and 21.1%, respectively) (data not shown). Both cultivars required approximately 10.5 days from breaker (following ethylene treatment) to table-ripe stage.

During "difference from control" sensory tests, untrained panelists detected significant flavor differences between table-ripe tomato samples with different ethylene exposure time from both cultivars (Table 5.5). Panelists found lesser differences between samples exposed to 1 day of ethylene treatment (control and hidden controls), meanwhile the flavor differences between the control and 3 or 5 day ethylene treatment samples were significantly greater when compared to 1-day tomatoes.



Table 5.5. "Difference from control" sensory tests conducted on table-ripe 'Agriset-761' and 'CPT-5' tomatoes. Flavor ratings were based on a 12-point scale with verbal descriptors rang ng from 1 (no different) to 12 (extremely different) from control. Tomato cultivars Ethylene Gassing Exposure time Pr > Fz 1-Day 3-Days 5-Days 'Agriset- 761' 0.0008 3.85 a 6.65 b 6.80 b 'CPT-5' 0.0001 3.00 a 5.90 b 7.35 b z Means for tomato fruit sensory scores with different letters within rows were significantly different at the 1% level according to Duncan's Multiple Range Test. (n=27 untrained panelists).


A close relationship has been reported between physiological maturity and ripening response under ethylene treatment, where green-harvested tomatoes that required 6 or more days of ethylene treatment to initiate ripening were likely immature at harvest (Kader at al., 1977; Kavanagh et al., 1986). In this study, sensory tests demonstrated that untrained panelists could detect overall flavor differences between






85


table-ripe tomatoes as related to their ethylene exposure time. There was a trend of increasing degree of difference from the control with increasing ethylene exposure time to initiate ripening. Previous research (Kader et al., 1977; Watada and Aulenbach, 1979) has also alluded to sensory differences related to fruit maturity at harvest.

Several panelists in this study noted "unpleasant", "metallic", "strange" or "lingering" off-flavors in tomato samples exposed to extended ethylene treatments. Hayase et al. (1984) reported tomatoes picked at green stage and ripened postharvest occasionally presented off-flavors, substantially weakening their characteristic tomato flavor. Significant differences in pH, titratable acidity, total sugars and vitamin C content were found for 'Agriset-761' fruits. In general, significantly lower pH values, found in 1 day ethylene samples, concurred with significantly higher titratable acidity content. Fruit exposed to 5 days ethylene had significantly higher pH values, while significantly lower total sugars, compared to those exposed to 1 and 3 days ethylene exposure time. Vitamin C content was significantly lower in 1 day fruit when compared to those exposed to 3 days or 5 days of ethylene treatment (9.96, 14.10 and 14.30 mg/100g fresh weight, respectively) (Table 5.6). In contrast, 'CPT-5' fruit showed significant differences only in titratable acidity, where fruit exposed to 3 days ethylene had significantly higher acidity than those exposed to 1 or 5 of ethylene treatment (0.86%, 1.04% and 0.94 % citric acid, respectively) (Table 5.6). The magnitude of pH and acidity variations required for sensory perception have been reported by Gould (1978), who concluded that changes in pH greater than 0.16 units and of 0.1% in titratable acidity were required for sensory detection. Therefore, flavor differences detected by untrained panelists may have been






86


attributed in part to sugar and/or acid concentration changes related to harvest maturity

for both cultivars.



Table 5.6. Table-ripe tomato chemical composition for 'Agriset-761' and 'CPT-5' tomatoes exposed to 1, 3 or 5 days of ethylene treatment to attain breaker stage. Compositional Parameters Z Ethylene Gassing Exposure time 1-Day 3-Days 5-Days
Agriset-761
pH 4.23 b 4.28 ab 4.34 a Soluble Solids Content (oBrix) 3.85 a 4.02 a 3.45 a Titratable Acidity (% Citric acid) 0.91 a 0.81 ab 0.73 b Total Sugars (% fresh weight) 1.93 ab 2.29 a 1.69 b Ascorbic Acid (mg/ 100g fresh weight) 9.96 b 14.10 a 14.30 a CPT-5
pH 4.25 a 4.28 a 4.23 a Soluble Solids (OBrix) 3.90 a 3.90 a 4.35 a Titratable Acidity (% Citric acid) 1.04 a 0.86 b 0.94 ab Total Sugars (% fresh weight) 10.69 b 13.01 ab 14.33 a Ascorbic Acid (mg/100g fresh weight) 1.80 a 2.13 a 2.35 a
Means for parameters with different letters within rows were significantly different at the 5% level according to Duncan's Multiple Range Test. (n=6 composite samples/trt).


Aroma volatile analysis by gas chromatography helped to identify additional

factors contributing to the significant differences found during sensory panels. The

concentrations of five aroma volatile compounds from table-ripe 'Agriset-761' tomatoes

were significantly different between ethylene gassing exposure times. Significantly

higher concentrations of acetone, hexanal and 2+3-methylbutanol were found in tomatoes

exposed to extended ethylene treatment (5-days), whereas, 2-isobutylthiazole and 13ionone decreased significantly compared to other treatments (Table 5.7). Table-ripe

'CPT-5' tomatoes had significant differences in 9 of 15 aroma volatile compounds

quantified. Tomatoes exposed to extended ethylene treatments showed increased

production of hexanal, trans-2-heptenal, 6-methyl-5-hepten-2-one, 2-isobutylthiazole, 1nitro-2-phenylethane and geranylacetone. Whereas, increased production of ethanol,






87


2+3-methylbutanol and 1-ionone was observed in tomatoes exposed to 3 days of ethylene

treatment (Table 5.8).


Table 5.7. Aroma volatile compound concentrations for table-ripe 'Agriset-761' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage (n=4 samples/treatment).
Aroma Volatile Compoundsz Ethylene Gassing Exposure time
1 day 3 days 5 days Acetone 0.82 b 0.77 b 1.69 a Methanol 297.01 a 285.35 a 268.32 a Ethanol 52.07 a 55.77 a 84.65 a 1-Penten-3-one 0.30 a 0.24 a 0.25 a Hexanal 13.24 ab 12.24 b 18.41 a Cis-3-hexenal 11.06 a 8.20 a 11.03 a 2+3-Methylbutanol 2.14 b 2.18 b 3.54 a Trans-2-hexenal 8.32 a 6.56 a 9.23 a Trans-2-heptenal 0.05 a 0.05 a 0.05 a 6-Methyl-5-hepten-2-one 0.85 a 0.68 a 0.83 a Cis-3-hexenol 0.05 a 0.05 a 0.05 a 2-Isobutylthiazole 0.14 a 0.13 a 0.10 b 1-Nitro-2-phenylethane 0.20 a 0.20 a 0.20 a Geranylacetone 6.54 a 8.14 a 5.39 a 3-ionone 0.17 ab 0.45 a 0.12 b Total Y 392.96 381.01 403.86
Means for aroma volatile compounds (pL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. Y Total aroma volatile concentrations (pL/ L) based on the sum of the 16 compounds quantified.


For both cultivars, concentrations of 2+3-methylbutanol, geranylacetone and Pionone were significantly affected by physiological maturity at harvest. In general,

volatile compound concentrations increased in fruit exposed to longer ethylene treatments

to initiate ripening. This finding contradicts the results observed previously for 'Solimar'

tomatoes with different ethylene exposure times. McDonald et al. (1996) reported

reductions in six aroma volatile compounds from tomatoes exposed to 1 day of ethylene

treatment, when compared to non-treated controls, changes in the concentrations of

hexanal, 6-methyl-5-hepten-2-one, geranylacetone and 2+3-methylbutanol were also






88


observed for 'Agriset-761' and 'CPT-5' samples. Significant reductions in 6-methyl-5hepten-2-one, 2-isobutylthiazole and geranylacetone found in table-ripe 'Solimar'

tomatoes exposed to extended ethylene treatment further supports the relevant effects of

harvest maturity on ripe tomato flavor/aroma.


Table 5.8. Aroma volatile compound concentrations for table-ripe 'CPT-5' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage (n=4 samples/treatment).
Aroma Volatile Compoundsz Ethylene Gassing Exposure time
1 day 3 days 5 days Acetone 1.01 a 1.72 a 1.23 a Methanol 268.50 a 301.65 a 282.44 a Ethanol 70.20 ab 100.48 a 38.53 b 1-penten-3-one 0.28 a 0.28 a 0.33 a Hexanal 17.66 b 16.84 b 25.34 a Cis-3-hexenal 8.80 a 11.30 a 9.51 a 2+3-methylbutanol 2.29 b 3.15 a 2.80 b Trans-2-hexenal 10.41 a 8.83 a 10.23 a Trans-2-heptenal 0.06 ab 0.05 b 0.07 a 6-methyl-5-hepten-2-one 0.93 ab 0.75 b 1.11 a Cis-3-hexenol 0.06 a 0.06 a 0.05 a 2-isobutylthiazole 0.13 ab 0.11 b 0.14 a 1 -nitro-phenylethane 0.22 b 0.21 b 0.32 a Geranylacetone 6.64 b 7.83 ab 8.87 a
-ionone 0.13 b 0.22 a 0.17 ab Total Y 387.32 453.48 381.14
Means for aroma volatile compounds (pL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. Y Total aroma volatile concentrations (LL/ L) based on the sum of the 16 compounds quantified.


Preliminary "difference from control" sensory tests confirmed the perception of

significant differences between table-ripe tomatoes with varying ethylene exposure times

to reach breaker stage. It is probable that reported reductions in hexanal, 6-methyl-5heptene-2-one, geranylacetone, I-penten-3-one, methanol, 2+3 methylbutanol, and 1nitro-2-phenylethane by McDonald et al. (1996) were not entirely a direct effect of

ethylene treatment but rather a consequence of tomato maturity at harvest.





89


'BHN-189' and 'Solimar' (Experiment 3). Further information on the character/magnitude of flavor differences due to harvest maturity was obtained during descriptive sensory panels. Flavor attributes for table-ripe 'BHN-189' and 'Solimar' tomatoes picked at green stage and treated with ethylene were compared to those in tomatoes harvested at light-red stage. Table-ripe 'BHN-189' tomatoes that required 3 and 5 days of ethylene treatment were considered lower in ripe aroma, sweetness, tomato flavor, while higher in sourness (5 days) and green-grassy flavor compared to ripeharvested (light-red, stage 5) or 1 day ethylene treatment tomatoes. Ripe-harvested 'BHN-189' fruit were not considered significantly different than those exposed to 1 day ethylene treatment in any flavor descriptor except for off-flavor, where ripe-harvested fruit were rated slightly higher (Figure 5.1).

Table-ripe 'BHN-189' tomatoes were stored at 200C for an additional 14 days (table-ripe + 14 days) to re-assess flavor differences between harvest maturities in overripe tomatoes. At this time, ripe-harvested 'BHN-189' tomatoes were considered higher in ripe aroma (significantly), off-odors, sweetness, and tomato flavor compared to green-harvested tomatoes regardless of ethylene exposure time (Figure 5.2). Only slight sensory differences between ethylene treatments were documented in overripe tomatoes.

For 'Solimar' tomatoes, slightly different sensory ratings between harvest maturities at table-ripe ripe stage were documented. Tomatoes that required 1 and 3 days of ethylene treatment were considered significantly higher in ripe aroma, while significantly lower in sourness, compared to 5-day ethylene treatment or ripe-harvested tomatoes (light-red, stage 5) (Figure 5.3). Similar to 'BHN-189', overripe 'Solimar' tomatoes (table-ripe + 14 days) harvested at light-red stage were rated higher in ripe






90


aroma, sweetness and tomato flavor when compared to the tomatoes picked at green

stage, regardless of ethylene exposure time. In addition, tomatoes exposed to 1 and 3

days of ethylene treatment, although not significant, had considerably higher ratings for

off-odors and off-flavors compared to those exposed to 5 days of ethylene treatment or

harvested at light red stage (Figure 5.4).

Compositional analyses from table-ripe 'BHN-189' tomatoes showed slight

reductions in pH and increased soluble solids with increasing ethylene exposure time.

Ripe-harvested tomatoes had higher pH than the 3- and 5-day ethylene treatments, and

lower or equal soluble solids than the 1- or 3-day ethylene treatments (Table 5.9).

Titratable acidity was highest in ripe-harvested tomatoes, whereas lowest in tomatoes

exposed to I day of ethylene treatment. After the 14-day storage period, soluble solids

and titratable acidity decreased considerably in all treatments. Conversely, pH values

were lower in ripe-harvested and 1-day ethylene exposure time tomatoes, while higher in

3-day and 5-day ethylene treatments (Table 5.9).



Table 5.9. Table-ripe and overripe chemical composition parameters for 'BHN-189' tomatoes harvested at light-red or green stages.
Chemical Composition Parameters z Ripe Green stage ethylene exposure time harvest 1-Day 3-Days 5-Days Table-ripe stage
pH 4.32 4.34 4.31 4.27
Soluble solids content (OBrix) 3.8 3.7 3.8 4.0 Titratable acidity (% citric acid) 1.15 1.07 1.13 1.13 Sucrose equivalents 3.30 4.33 4.35 4.30 Lycopene content (ptg/g fresh weight) 26 24 28 16 Overripe (table-ripe + 14 days)
pH 4.30 4.25 4.38 4.35 Soluble solids content (OBrix) 3.6 3.3 3.6 3.7 Titratable acidity (% citric acid) 0.84 0.92 0.79 0.72 Sucrose equivalents 4.17 3.79 4.20 3.63
zMeans for compositional parameters represent the average of 2 composite samples/ treatment (n=8 fruit/sample).






91


Sucrose equivalents were lower in ripe-harvested tomatoes when compared to

green-harvested tomatoes regardless of ethylene exposure time. Nonetheless, sugar

levels increased in ripe-harvested tomatoes while they decreased in green-harvested ones

at overripe stage. Lycopene content was considerably lower in tomatoes exposed to

extended ethylene treatment compared to remaining treatments. Table-ripe 'Solimar'

tomatoes had increasing pH and soluble solids contents with increasing ethylene

exposure time, while, ripe-harvested tomatoes had the lowest pH (4.2), lowest titratable

acidity (0.92%) and next to lowest soluble solids content (Table 5.10).



Table 5.10. Table-ripe and overripe stage chemical composition parameters for 'Solimar' tomatoes harvested at light-red or green stages.
Chemical Composition Parameters z Ripe Green stage ethylene exposure time harvest 1-Day 3-Days 5-Days Table-ripe stage
pH 4.20 4.27 4.32 4.35
Soluble solids content (oBrix) 3.9 3.8 4.1 4.4 Titratable acidity (% citric acid) 0.92 1.19 1.19 1.08 Sucrose equivalents 4.39 5.20 5.55 4.17 Lycopene content (tg/g fresh weight) 19 14 15 12 Overripe (table-ripe + 14 days)
pH 4.28 4.25 4.27 4.21 Soluble solids content (OBrix) 3.4 3.8 4.2 3.9 Titratable acidity (% citric acid) 0.67 0.72 0.71 0.74 Sucrose equivalents 5.60 5.90 5.75 3.60
Means for compositional parameters represent the average of 2 composite samples/ treatment (n=8 fruit/sample).


After 14 days of storage in 'Solimar' tomatoes, pH values decreased in all

treatments except for ripe-harvested tomatoes, whereas, titratable acidity was

considerably lower in all treatments at the overripe stage. Soluble solids remained lowest

for ripe-harvested tomatoes, while sucrose equivalents were consistently lowest in

tomatoes exposed to extended ethylene treatments. At overripe stage, ripe-harvested






92


tomatoes had comparable sugar levels to those in tomatoes exposed to I and 3 days of

ethylene treatment. Lycopene content was considerably higher in tomatoes ripeharvested tomatoes, while tomatoes exposed to 5 days of ethylene treatment had the

lowest contents.


Table 5.11. Aroma volatile compound concentrations for table-ripe 'BHN-189' harvested at light-red or green stages.
Aroma volatile compounds Ripe Green stage ethylene exposure time harvest I day 3 day 5 day Acetaldehyde 17.47 a 12.97 a 11.25 a 16.04 a Acetone 0.31 c 0.39 abc 0.37 ab 0.43 a
Methanol 271.77 a 263.98 a 250.59 a 264.93 a Ethanol 20.85 a 21.47 a 21.62 a 23.00 a l-Penten-3-one 0.08 a 0.06 a 0.04 a 0.06 a Hexanal 1.78 a 2.66 a 1.14 a 2.34 a Cis-3-hexenal 4.35 a 2.10 b 1.85 b 2.51 b 2+3-Methylbutanol 0.81 c 2.04 b 2.20 b 2.79 a Trans-2-hexenal 3.29 a 2.83 a 2.71 a 2.99 a Trans-2-heptenal 0.11 b 0.18 a 0.18 a 0.16 a 6-Methyl-5-hepten-2-one 0.97 a 0.83 a 0.71 a 0.67 a Cis-3-hexenol 0.81 a 0.53 b 0.47 b 0.47 b 2-Isobutylthiazole 0.06 b 0.08 a 0.09 a 0.09 a I-Nitro-2-phenylethane 0.08 a 0.06 a 0.06 a 0.06 a Geranylacetone 2.01 b 3.02 a 2.46 b 2.26 b 1-ionone nd Y nd nd 0.16 a
Total' 324.76 313.19 295.74 321.94
Means for aroma volatile compounds (pL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. Y Volatile compound concentrations not detected. x Total aroma volatile concentrations (pL/ L) based on the sum of the 16 compounds quantified.


Aroma volatile analysis for table-ripe 'BHN-189' tomatoes sampled during

descriptive sensory panels revealed that 7 of 16 volatile compounds had significant

concentration differences due to harvest treatments. Notably, ripe-harvested tomatoes

had the lowest concentrations of acetone, 2+3-methylbutanol, trans-2-heptenal, 2isobutylthiazole and geranylacetone, and the highest concentrations of cis-3-hexenal and

cis-3-hexenol (Table 5.11). Tomatoes picked at green stage, regardless of ethylene




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ELU36IDA0_6LDW50 INGEST_TIME 2015-03-05T21:37:51Z PACKAGE AA00028850_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

FLAVOR OF FRESH MARKET TOMATO (Lycopersicon esculentum Mill) AS INFLUENCED BY HARVEST MATURITY AND STORAGE TEMPERATURE By FERNANDO MAUL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1999

PAGE 2

To my family in Guatemala To Vicky

PAGE 3

ACKNOWLEDGMENTS First and foremost, my most sincere gratitude and admiration to Dr. Steven A. Sargent for his guidance, support and enthusiasm as the chairman of my advisory committee. To the members of my advisory committee, Dr. Elizabeth A. Baldwin, Dr. Charles A. Sims, Dr. Donald J. Huber and Dr. Murat O. Balaban, my deepest gratitude for allowing me and my seemingly endless endeavors into your laboratories. I am very grateful to Diego A. Luzuriaga, Margery Einstein, Dr. J. W. Scott, and Mr. Kenneth Shuler for their valuable contributions and help. I would like to thank the members of my trained descriptive sensory panel, whose motivation and attendance helped me enormously. I am thankful to Abbie J. Fox, James Lee, Holly Sisson and Bob Golaszewski. for all their help and to Dr. H. Klee for allowing storage of my samples in his limited freezer space. Very special thanks to my friends in the Postharvest laboratory Pornchai P., Cleisa Cartaxo, Domingo s Almeida, Ernesto Brovelli, Celso Moretti, and Shahab Hanif-Khan. Finally, I would like to express my ovemhelrning gratitude towards my family and my wife, who have unconditionally stood by me throughout my graduate studies at UF. iii

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGMENTS iii LIST OF ABBREVIATIONS vi ABSTRACT vii CHAPTERS 1 INTRODUCTION 1 2 LITERATURE REVIEW 5 Tomato Fruit Ripening 5 Ripe Tomato Composition 7 Ripe Tomato Flavor 10 Lipid-derived Volatile Compounds 12 Carotenoid-derived Volatile Compounds 13 Sensory Flavor Perception 14 Sensory Analysis 16 Electronic Aroma Sensing Technology 20 Harvest Maturity and Tomato Flavor 22 Postharvest Treatments and Tomato Flavor 23 Storage Temperature Management 24 Research Objectives 26 3 RIPENESS STAGE AT HARVEST AFFECTS CHEMICAL COMPOSITION AND COLOR QUALITY OF TABLE-RIPE TOMATOES 27 Introduction 27 Materials and Methods 28 Results and Discussion 31 Conclusions 42 iv

PAGE 5

Page 4 POTENTIAL FOR NONDESTRUCTIVE QUALITY SCREENING OF TOMATOES WITH ETHYLENE OR ELECTRONIC NOSE SENSOR ARRAYS 45 Introduction 45 Materials and Methods 47 Results and Discussion 51 Conclusions 60 5 AROMA VOLATILE PROFILES AND RIPE TOMATO FLAVOR ARE INFLUENCED BY PHYSIOLOGICAL MATURITY AT HARVEST 67 Introduction 67 Materials and Methods 68 Results and Discussion 77 Conclusions 105 6 HARVEST MATURITY AND STORAGE TEMPERATURE AFFECT TOMATO VOLATILE PRODUCTION AND FLAVOR 118 Introduction 118 Materials and Methods 120 Results and Discussion 124 Conclusions 155 7 CONCLUSIONS 174 REFERENCES 178 BIOGRAPHICAL SKETCH 190 v

PAGE 6

LIST OF ABBREVIATIONS CI Chilling injury DCIP 2,6-dichloropheno lindo lpheno 1 DNPH Dinitrophenylhydrazine EN Electronic nose GC Gas chromatography HPLC High performance liquid chromatography IG Immature-green stage MD Mahalanobis distance MG Mature-green stage MVDA Multivariate discriminant analysis Ml Immature-green M2 Partially mature-green M3 Mature-green M4 Advanced mature-green ouv Odor unit value QDA Quantitative descriptive analysis SSC Soluble solids content TA Titratable acidity T301 Sensor type-301 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FLAVOR OF FRESH MARKET TOMATO (Lycopersicon esculentum Mill.) AS INFLUENCED BY HARVEST MATURITY AND STORAGE TEMPERATURE By Fernando Maul May 1999 Chairman: Dr. Steven A. Sargent Major Department: Horticultural Sciences The effects of harvest maturity and storage temperature on fresh tomato {Lycopersicon esculentum Mill.) chemical composition and sensory quality at ripe stage were investigated. Tomatoes from seven commercial cultivars harvested at green stage were exposed to exogenous ethylene (100 uL/L) to accelerate the onset of ripening for 1 to 7 days until they attained breaker stage (<10% red color). There was a strong relationship (r 2 = 0.84) between green-tomato maturity at harvest and ethylene exposure time to attain breaker stage. Immature-green (Ml) tomatoes had to be treated with ethylene for >3 days compared to mature-green (M4) tomatoes which were treated for 1 day. Ripe tomato color parameters (L* and a* values) consistently increased with increasing ethylene requirement, indicating development of less intense-red coloration with immature-harvested fruit. vii

PAGE 8

Untrained sensory panelists found significant differences in ripe tomato flavor between fruits exposed to ethylene for 1, 3 or 5 days to attain breaker stage. Descriptive sensory panelists determined that at ripe stage green-harvested tomatoes exposed to <3 days of ethylene treatment were comparable in ripe aroma, sweetness and tomato flavor to tomatoes harvested at light-red stage. In contrast, tomatoes that reached breaker stage after 3 days of ethylene treatment were rated highest in sourness and green/grassy flavor, and lowest in ripe aroma, sweetness and tomato flavor. Ripe tomatoes with inferior sensory qualities showed significant changes in several aroma volatile compounds and chemical composition parameters. An ethylene exposure threshold of 3 days could be immediately utilized commercially as a non-destructive means for segregating immaturegreen from mature-green tomatoes, thus ensuring consistently high-flavor tomatoes. Ripe tomatoes stored at 20°C for 12 days were rated significantly higher for ripe aroma, tomato flavor and sweetness compared to those stored at 10° or 5°C. High temperature pre-treatments (38°C for 2 days) were not effective in alleviating flavor changes induced during storage at 5°C for 7 days. The electronic nose sensor array successfully classified intact green tomatoes into immature and mature stages, and distinguished between tomatoes harvested at different ripeness stages or stored at various temperatures below 20°C. With increased detection speeds, electronic noses have potential for screening green tomatoes in commercial operations. viii

PAGE 9

CHAPTER 1 INTRODUCTION Tomatoes (Lycopersicon esculentum MilL) are edible members of the Solanaceous family, they originated in Central and South America, and were first introduced to Europe by Spanish explorers returning from America. Because Europeans believed tomatoes were poisonous, consumption and popularity did not increase until the eighteenth century (Petro-Turza, 1987). Worldwide, tomatoes are among the ten most important fruits and vegetables in consumption. In a recent survey, tomatoes were ranked 16 th in overall nutrient content; however, based on their nutritional contribution to the diet of most Americans was considered, tomatoes were ranked 1 st (Goodenough, 1990). Much popular concern has been expressed in the last two decades regarding the poor quality of fresh market tomatoes available in consumer markets (Harris, 1973; Cerra, 1975; Bisogni and Armbruster, 1976). USDA consumer surveys indicated a higher level of dissatisfaction with fresh tomatoes than with any of the 32 other products included in the surveys. The source of dissatisfaction was price, taste and ripeness (Resureccion and Shewfelt, 1985). Researchers have proposed several reasons for the inferior flavor in fresh market tomatoes. Most notably, commercial breeding programs have emphasized disease resistance, productivity and fruit firmness in selections at the expense of flavor and texture qualities (Baldwin et al, 1992). In addition, the lack of proper commercial harvesting and postharvest handling procedures for fresh market tomatoes has been shown to play an important role in affecting tomato flavor (Kader et 1

PAGE 10

2 al., 1978; Sargent et al., 1997). Postharvest abuses, such as harvesting immature fruit, mechanical injury during sorting and packing, and improper temperature management have been related to altered aroma volatile profiles and altered flavor perception (Sargent et al., 1997; Moretti et al., 1997). Evidence of the adverse effects of low temperature storage on tomato flavor has been published previously (Kader et al., 1978; Stern et al., 1994; McDonald et al., 1996). In addition, the effects of storing tomatoes at currently recommended temperatures on their flavor and aroma has not been thoroughly addressed. Throughout tomato production and marketing, efforts are made to maintain optimal visual quality (uniform color, absence of decay, proper firmness, etc.) to attract customers. As a consequence, disorders not readily detectable during sorting operations have received less attention. Visual appearance is a critical factor driving the initial purchase, but subsequent purchases are influenced greatly by eating quality (flavor, aroma and mouthfeel). Overall fresh tomato quality is determined not only by fruit appearance and firmness, but also flavor, aroma, and nutritive value. In the U.S., produce quality is often a more important factor determining consumer purchases than price (Schwartz, 1995). Most fresh market tomatoes sold in U.S. supermarkets are harvested before they are "table-ripe" because retailing ripe tomato fruit is not practical within the current long distance handling and marketing system (Kader et al., 1978). In Florida, over 85% of the fresh market fruit are harvested green for transport to distant markets. Such factors as harvesting immature-green fruit, occurrence of mechanical injuries and storage below safe temperatures can reduce the overall quality from the shipping point to the consumer's table. The winter vegetable industry in the U.S. has seen drastic competition

PAGE 11

3 and decline in sales due to imports from Mexico and Europe. Following the North American Free Trade Agreement (NAFTA), imports of Mexican produce during the winter months have more than doubled (Florida Tomato Committee, 1998). In contrast, fresh market tomato acreage in Florida has declined by approximately 25% between the 1993-94 and 1997-98 seasons. It is believed that the only solution for Florida growers is to provide consistent high quality tomatoes based on their green-harvest system. External indicators of green tomato maturity, such as shape, color, surface appearance, and stem scar condition (Kader and Morris, 1976) are subjective and often impractical during commercial harvesting operations (Brecht et al., 1991). In fact, Florida tomatoes picked at green stage are harvested and sorted based on minimum size requirements set by the Florida Tomato Committee (Florida Tomato Committee, 1989). Fresh tomato maturity distribution data, collected from commercial green harvest operations, revealed that in average 49% of green-harvested fruits were at immaturegreen stage (Ml) with larger fruit sizes consistently showing lower immature fruit proportions (Chomchalow, 1991). The inability to accurately distinguish between the maturity classes (M1-M4) at harvest has obscured the results of sensory quality studies in the past (Watada and Aulenbach, 1979). Nonetheless, negative effects from immaturegreen harvest stages on ripe tomato chemical composition, overall sensory acceptance, and aroma volatile profiles have been suggested (Kader et al., 1978; Maul et al., 1997a). During tomato fruit ripening, a series of quantitative and qualitative changes take place in tomato aroma volatile profiles. Organic acids, soluble sugars, amino acids, pigments, and over 400 aroma compounds contribute to characteristic tomato flavor (Petro-Turza, 1987). Due to the diversity of bio synthetic pathways contributing to the

PAGE 12

4 formation of volatile compounds, tomato aroma could be a good indicator of fruit injury as a result of harvest and postharvest handling treatments. New technologies such as machine vision, magnetic resonance imaging, acoustic impulse transmission, chlorophyll fluorescence, and electronic aroma sensing are powerful tools that could be used to define and predict fresh produce quality beyond traditional visual quality parameters. "Electronic nose" (EN) sensor arrays are used to classify specific samples of interest based on their headspace volatiles. The electronic nose consists of a series of non-specific chemical sensors, each of which shows characteristic responses to the volatile chemicals within the headspace over a sample (Anonymous, 1996). The present study was conducted to document the effects of harvest maturity of green-harvested tomatoes on their ripe sensory quality. Sensory panelists were trained to describe flavor and aroma differences between table-ripe tomatoes harvested at green stage compared to tomatoes harvested at later ripening stages. In addition, nondestructive analysis using electronic nose sensor arrays and predictive models based on ethylene treatment requirements were explored to screen immature-green tomatoes. The effects of different storage temperatures on ripe tomato flavor were also investigated as potential factors contributing to inferior quality of fresh-market tomatoes.

PAGE 13

CHAPTER 2 LITERATURE REVIEW Tomato Fruit Ripening After fruits and vegetables attain physiological maturity they undergo a distinct ripening phase characterized by a series of changes under genetic regulation. Ripening involves the synthesis of new proteins, nucleic acids, and carbohydrates (Mitcham et al, 1989), chlorophyll degradation, starch hydrolysis, formation of pigments, and synthesis of aroma volatiles (Biale and Young, 1991). All of ripening related metabolic pathways are very likely to be under hormonal control for both initiation and coordination (Tucker, 1993). Tomato ripening involves the autolysis of cell wall pectins, the synthesis of lycopene and other carotenoid pigments, changes in sugar and acid contents, and production of volatiles associated with flavor and aroma (Hobson and Davies, 1971 ; Gray et al., 1992; Petro-Turza, 1987). During ripening, sugar levels increase, possibly due to the hydrolysis of starch reserves in the fruit (Whiting, 1970). In fruits still attached to the plant, import of assimilates may also play an important role in sugar accumulation. Tomatoes utilize both sugars and organic acids as substrates for respiratory activities (Ulrich, 1970). In climacteric fruits, such as tomatoes, the initiation of ripening is associated with a burst in ethylene production, accompanied by a large increase in respiratory activity 5

PAGE 14

6 prior to or during ripening (Rhodes, 1980). Non-climacteric fruits, on the other hand, only show gradual decrease in respiration as they approach senescence. Climacteric ripening behavior, also common to bananas, mangoes, avocados, etc. (Hardenburg et al., 1986), allows for harvest at an "under-ripe" stage, because remaining ripening events can occur postharvest. Ethylene production has been reported to increase at the mature-green stage, before the first visual symptoms of ripening (Brecht, 1985). Ethylene biosynthesis during ripening is autocatalytic (system II ethylene). Small amounts of exogenous ethylene stimulate a massive increase in the fruit's ethylene production (Yang and Hoffman, 1984) which, in turn may accelerate the onset of ripening. Ethylene is believed to coordinately induce the expression of a large number of genes which encode for regulatory enzymes in the ripening processes (Theologis et al., 1993). Ethylene biosynthesis starts with the conversion of methionine to S-adenosyl-methionine (SAM), followed by conversion of SAM to 1-aminocyclopropane-l-carboxylic acid (ACC) by ACC synthase, and finally conversion of ACC to ethylene by ACC oxidase (Yang, 1985). Exogenous ethylene treatment is a commercial practice used to accelerate the ripening of green tomatoes and has been utilized by industry for the last five decades. Under laboratory conditions, green tomatoes have been shown to respond to ethylene concentrations as low as 0.5 uL/L, conversely, concentrations greater than 10 uL/L would not result in an enhanced ripening response. Nonetheless, most commercial tomato ripening operations utilize concentrations ranging between 50 and 500uL/L to assure response. Green tomato ripening response under ethylene treatment has been shown to decrease with increasing time between harvest and ethylene treatment

PAGE 15

7 (Chomchalow, 1991). In addition, tomato response to ethylene has been related to physiological maturity at harvest (Kader et al., 1977; 1978). Green tomato ethylene requirement was inversely related to maturity at harvest. In fact, Kader et al. (1978) considered tomatoes requiring more than 6 days of ethylene treatment as immature-green at the time of harvest. Once autocatalytic ethylene production has commenced, exogenous ethylene exposure will not have a significant effect on subsequent ripening rates. There is evidence that tomato ripening does not occur uniformly throughout the fruit. The initiation of ripening-related processes first occur in the locule tissue of mature-green tomatoes (Grierson et al., 1985) and could, in turn, stimulate the ripening of adjacent tissues (pericarp and columnella). The classification system for physiological maturity of green-stage tomatoes proposed by Kader and Morris (1976) relied on locule tissue development to determine maturity classes. Maturity class distribution was defined as immature-green (Ml), partially mature-green (M2), mature-green (M3) and advanced mature-green (M4) based on locule tissue development. However, if ethylene requirement is to be employed as a screening tool for immature tomatoes, further research is needed to clarify the relationship between ethylene treatment requirement and physiological maturity. Ripe Tomato Composition Environmental factors such as water availability, soil fertility, and cultural practices have been shown to influence sugar/acid ratios in fresh market tomatoes

PAGE 16

8 (Stevens, 1985). In general, ripe tomatoes are composed of approximately 93%-94% water, where the most prevalent dissolved organic constituent is sugar. Fructose and glucose may account for up to 50% of the fruit's dry matter and over 95 % of total sugars (Stevens, 1985). Sucrose concentrations are extremely low, probably due to acid invertase activity (Klann et al., 1996). Fructose could be considered more important for sweetness perception than glucose, due to the lesser sweetening power of the latter hexose (Stevens et al, 1977). In fact, Koehler and Kays (1991) proposed calculating "sucrose equivalents" by multiplying fructose and glucose concentrations times 1.73 and 0.74, respectively, to better represent individual sugar relative sweetening potential. The major organic acid constituent in ripe tomato fruit is citric, which, may account for 60-90% of the total acid content. (Davies and Cocking, 1965) (Table 2.1). The second major organic acid in ripe tomatoes is malic. Malic acid's higher acidity potential determined to be approximately 14% higher than citric acid's during model solution sensory studies (Gardner, 1966) emphasizes its contribution to ripe tomato acidity. Table 2.1. Typical compositional parameters for ripe fresh market tomatoes (expressed on a % fresh-weight basis) Vitamin C (mg/lOOg fresh wt.) Dry Matter Total Carbohydrates Reducing Sugars Sucrose Total Soluble Solids Malic Acid Citric Acid 20 6.5 4.7 3.0 0.1 4.5 0.1 0.2 Adapted from Hobson and Grierson, 1993. Sugar/acid ratios have been traditionally used as indicators of fruit quality and flavor potential. However, during tomato sensory studies, the ratios between soluble solids content and titratable acidity did not relate to tomato acceptability (Malundo et al,

PAGE 17

9 1995; Resureccion and Shewfelt, 1985). Furthermore, if both acid and sugar levels are low, it is possible to have poor flavor quality even though sugar/acid ratios remain desirable (Stevens, 1985). Malundo et al. (1995) challenged the assumption that increasing sugar or acid concentrations improved tomato flavor. In fact, it was demonstrated that only at certain acidity levels (-0.80% citric acid) did increased sugar concentrations lead to better acceptability ratings. Watada and Aulenbach (1979) reported "vine ripe"-harvested tomatoes were judged to be sweeter than green-harvested ones, even though there were no significant differences in dry matter content. This demonstrated the importance of other compounds, besides sugars, in sweetness perception. Vitamin C content (ascorbic acid) has been shown to be affected by ripeness stage at harvest (Watada et al., 1976). During ripening, vitamin C contents increase from the green stage to reach its maximum prior to the ripe stage (Malewski and Markakis, 1971). Similarly, lycopene and P-carotene levels consistently increase until the ripe stage. Locule tissue may constitute between 14 and 34% of fresh tomato mass (Stevens et al., 1977), and, when compared to pericarp tissue, locule tissue contains considerably higher vitamin C, citric and malic acid contents, contrasted by lower sugars (Brecht et al., 1976; Stevens et al, 1977). Sharp contrasts in chemical composition, ethylene sensitivity, enzymatic activity, and contribution to overall fruit flavor, between locule and pericarp tissues suggest their separate analysis may prove to be useful during ripening and flavor biosynthesis studies. One of the most important processes occurring during fruit ripening is the enhanced production of aroma volatiles (Pesis, 1994). The most important aroma

PAGE 18

10 compounds in fruits are usually alcohols, aldehydes, esters, and other hydrocarbons (Nursten, 1970). Flavor volatiles, usually present at low levels, are considered to provide peculiar flavor and aroma notes, or descriptors, to different fruits (Buttery et al., 1990). Some vegetables, such as onions (Brodnitz et al., 1969) and peppers (Buttery et al., 1969) have a characteristic flavor attributed to a single or small group of volatile compounds. However, for most fruits and vegetables, flavor and aroma are determined by a complex combination of numerous volatile compounds (Bartley et al, 1985; Perez et al., 1992; Baldwin et al., 199 lac; Rizzolo et al, 1995). Ripe Tomato Flavor Natural aromas and flavors present in fruits and vegetables are almost always complex mixtures, comprising tens, and more often, hundreds of chemical constituents (Dodds et al, 1992). To date, over 6,000 flavor-active compounds have been identified, and the list will most likely increase as newer, more sensitive instruments and analytical techniques are developed. The characteristic flavor of tomato fruits arises from the complex interaction between organic acids, soluble sugars, amino acids, minerals, and aroma volatiles (Baldwin et al., 1991a). Approximately 400 different volatile compounds have been identified in ripe tomato fruits (Petro-Turza, 1987) with fewer than 30 volatile compounds considered important based on their odor and flavor threshold values (Buttery et al., 1990). From the list of ripe tomato aroma volatile compounds there, is limited quantitative evidence of individual compound contribution to overall tomato aroma (Goodenough, 1990). C/s-3-hexenal, (3-ionone, hexanal, P-damascenone, l-penten-3one, 3-methylbutanol, /raws-2-hexenal, 2-isobutylthiazole, l-nitro-2-phenylethane, trans-

PAGE 19

11 2heptenaL phenylacetaldehyde, 6-methyl-5-hepten-one, and cw-3-hexenol are among the most important volatile compounds present in ripe tomatoes (Buttery, 1993). Characterization of volatile production and enzymatic activity by individual fruit tissues (locule and pericarp tissue) has demonstrated quantitative aroma volatile production by individual tomato tissues (Buttery et al., 1988). Levels of important tomato aldehydes and alcohols such as hexanal, cw-3-hexenal, frww-2-hexenal, and cis3hexenol were two-fold higher in pericarp tissue compared to locule gel, whereas important ketone volatiles such as 6-methyl-5-hepten-2-one and geranylacetone were slightly higher in the locule tissue (Buttery et al., 1988). The important contribution of locule tissue fluidity for immediate access to taste receptors may influence flavor perception during consumption (Stevens et al., 1977). One of the most rapidly growing areas in fruit flavor research is the characterization of flavor precursors (Rouseff and Leahy, 1995). Many plant nutrients such as free amino acids (Perez et al., 1992), lipids and carotenoid pigments (Stevens, 1985), and bound glycosides (Krammer et al., 1995) can act as flavor volatile precursors. In order to understand ripe tomato aroma, an understanding of precursor metabolism during ripening is necessary (Perez et al, 1992). Aroma volatile production studies could help elucidate important storage and handling factors affecting fruit flavor and odor preservation (Fellman et al., 1993). Tomato aroma characterization studies would generally consider cultivars with higher volatile compound production as having potential for better flavor. In a study that compared sensory attributes for six commercial tomato cultivars, it was concluded that

PAGE 20

12 higher levels of sugars, medium acidity and higher concentrations of important volatiles concurred with higher sensory acceptability (Baldwin et al., 1991b). Aroma volatile production from tomato fruits has been reported to undergo considerable qualitative and quantitative changes during ripening. Aldehyde concentrations increased 9-fold from green to red ripeness stages while ketone volatiles increased approximately 2-fold during ripening (Baldwin et al., 1991c). Lipid-derived Volatile Compounds The characteristic aroma of tomatoes, apples, cucumbers, bell peppers, and bananas comes from aldehydes, ketones, and alcohols (Petro-Turza, 1987; Luning, 1995; Pesis, 1994). The formation of aroma and flavor volatile compounds from membrane lipids and free fatty acids is usually associated with free radical initiation or lipoxygenase-mediated oxidation (Shewfelt and Purvis, 1995). The presence of lipoxygenases and hydroperoxide lyases in microsomal membranes suggested that aldehydes, ketones, alcohols, and perhaps other volatile compounds could be produced at a membrane site (Riley et al., 1996). When linoleic acid substrate was utilized, lipoxygenase (LOX) isolated from algae produced aromas described as apple-like, green, cucumber, and mango (Kuo et al., 1996). In tomato, LOX forms specific 13(s)or 9(s)-hydroperoxides from the degradation of linoleic or linolenic acids (Riley et al., 1996). Hydroperoxide lyase (HPL) cleaves those position-specific hydroperoxides produced by LOX yielding aldehydes. Hexanal and cw-3-hexenal arise from cleavage of 13(s)-hydroperoxides of linoleic and linolenic acids, respectively (Gardner, 1995). During tomato ripening, Buttery (1993)

PAGE 21

13 showed concentrations of c/s-3-hexenal and hexanal increased 4 to 5-fold between green and breaker stage, and 5 to 10fold between breaker and red stage. Cw-3-hexenal is highly unstable (Petro-Turza, 1987) and will be rapidly isomerized to fra/7s-2-hexenal due to isomerase activity. Aldehyde compounds may, in turn, be converted to alcohols such as cw-3-hexenol due to the activity of alcohol oxidoreductases. Evidence relating other carboxylic acids in the biosynthesis of aldehydes (C3-C5) is not yet available (De Pooter et al., 1987). Aroma compounds produced by LOX are desirable but, could also give rise to off-flavors in some cases. There is evidence that HPL-derived aldehydes play a role as antifungal agents in fruits (Gardner, 1995; Mattheis and Roberts, 1993). Among tomato lipid-derived volatile compounds, c/'s-3-hexenal, hexanal, l-penten-3-one, fraw.s-2-hexenal, /rans-2-heptenal, and cw-3-hexenol have been proposed as important based on compound concentrations required for sensory response and their concentrations present in tomatoes (Buttery et al., 1990; Tandon, 1998). Carotenoid-derived Volatile Compounds Plant breeders have extensively used high pigment mutants for developing improved fruit color in new tomato cultivars (Warm et al., 1985). Lycopene, being one of the major pigments in tomato fruits and is an important flavor precursor for ketone volatiles (Petro-Turza, 1987). Variations in fruit color have been related to changes in flavor (Stevens, 1985). Some of the important tomato volatiles, such as 6-methyl-5hepten-2-one and geranylacetone, are produced from the oxidative decomposition of lycopene, while, P-ionone is a product of B-carotene decomposition (Buttery et al., 1990).

PAGE 22

14 There is evidence that 6-methyl-5-hepten-2-one may have a different biosynthetic pathway than geranylacetone (Petro-Turza, 1987). Another important enzyme in volatile biosynthesis is alcohol dehydrogenase (ADH). ADH catalyzes the inter-conversion of acetaldehyde to ethanol (Massantini et al., 1995) or aldehydes to alcohols such as c/.s-3-hexenal to c/s-3-hexenol (Speirs et al., 1998). ADH activity in plant tissues has been attributed to several functions including survival during periods of hypoxia, protection from chilling stress, and the biogenesis of flavor ester volatiles (Mitchell and Jelenkovic, 1995). Ester volatiles are an important group of aroma compounds which augment at the time of ripening of both climacteric and non-climacteric fruits (Pesis, 1995). In addition, numerous volatile compounds may be derived from deamination-decarboxylation of amino acids. Notably, 2isobutylthiazole has been related to thiamine as a probable sulfur precursor (Petro-Turza, 1987). Sensory Flavor Perception Fruit flavor is without a doubt one of the most important parameters to consider in fruit quality. In general, humans are able to perceive flavors from the interaction of compounds with the sense of taste and smell. There are basically four different tastes our senses can distinguish: sweetness, sourness, bitterness, and saltiness. Taste buds on the tongue are the major organs for the perception of taste. However, the presence of taste buds on the epiglottis and of stratified squamous epithelial cells lining the pharynx and larynx point to the relevance of other organs in taste perception (Nagodawithana, 1994).

PAGE 23

15 In addition to the basic sensations of sweetness, sourness, bitterness, and saltiness that are perceived by taste buds present in the tongue, the aroma given off by food as it travels the retro-nasal route, influences the perception of flavor (Maruniak, 1988). In humans, the flavor sensation is comprised of at least three separate sensory systems: the gustation, olfaction and trigeminal senses (Dodds et al., 1992). Gustation refers to sensations perceived by taste receptors present in the tongue (Plattig, 1988), while the trigeminal sense refers to the perception of noxious stimuli within the nasal cavity by the trigeminal nerve (Cain and Murphy, 1980). However, of these sensory systems, olfaction plays the dominant role in flavor sensation. The olfactory sensory tissue is a section of thin epithelium located high in the nose and contains millions of olfactory receptor neurons. The specificity of the olfactory sensing system in higher animals is derived from a combination of receptor cells with partially overlapping sensitivities, resulting in the human nose being able to detect complex odors in the parts-per-billion range (Hodgins and Simmonds, 1995). In the human olfactory system, the sensitivity to odor intensities is high and generally follows a logarithmic relationship between odor sensitivity and increases in odorant concentration. In addition to sensitivity, the olfactory system has an excellent capacity of discriminating similar complex odors (Gardner and Hines, 1997). The intake of a complex mixture of odorant molecules stimulates primary neurons within the olfactory epithelium. Each secondary neuron integrates the signals of 10,000 to 20,000 primary neurons into a single signal that is fed to the brain, where the information is processed (Shurmer, 1990). The olfactory bulb is a bulbous tissue in the brain that performs the higher level processing of all the olfactory signals (Dodds et al., 1992) and

PAGE 24

16 identifies chemical odorant mixtures as a unitary stimulus (Shurmer, 1990). The brain is trained to recognize odor patterns corresponding to odor descriptions saved in a person's memory (Tomlinson et al., 1995). Stimuli released from flavor substances are perceived in the mouth and also in air delivered backwards to the nasal cavity while chewing or swallowing (Land, 1994). During swallowing, the odor-containing air accumulates in the nose because there is no expiratory airflow. Thus, the flavor quality and intensity perceived is influenced by the time period the food stays in the mouth prior to swallowing. Flavor perception has been shown to reach its full flavor potential within 10 seconds of chewing or swallowing (Land, 1994). In addition, whenever an aroma stimulus in the mouth reaches the olfactory receptor neurons via the retro-nasal route, it is perceived as a gustation rather than an olfaction sensation (Burdach et al, 1984). This event is referred to as the "taste-smell illusion". There is a close relationship between aroma and flavor. The compound concentrations needed for human flavor and odor perception (odor and flavor thresholds) are almost identical (Ahmed, 1978). The importance of flavor volatile compounds present in fruit tissues have been determined by odor unit values (OUV) calculated in flavor characterization studies. OUV's relate the odor thresholds for the different aroma compounds with their actual concentration present in the fruit tissue (Buttery et al., 1990). However, OUV information cannot replace sensory evaluation since odor thresholds may vary between individuals by up to 2 orders of magnitude, while complex volatile mixtures may elicit different qualities than individual compounds. Laing et al.

PAGE 25

17 (1984) showed that the perceived intensities of volatile compound mixtures were usually lower than the sum of the individual components. Sensory Analysis Taste and smell are the primary instruments used by various industries for evaluating the quality of a wide range of food products. No two humans are able to perceive an aroma in identical fashion, nor can they share common descriptors because of a series of cultural and associative memory effects (Strassburger, 1996). Therefore, sensory panel evaluation aims to provide objective information derived from the evaluation of subjective qualities present in a food product. Most classifications of odors used during sensory analysis are based on subjective perception and use common names, e.g. green, grassy, musty, floral, etc. Analytical chemists criticize the subjectivity and poor reproducibility of sensory test results, especially descriptive, whereas sensory scientists are often unable to correlate analytical data in a meaningful way with sensory qualities (Marsili, 1995). Nonetheless, sensory evaluation is the only tool readily available to assess the quality and acceptability of fruits and vegetables. Different sensory descriptors used by several researchers have helped to demonstrate the relationships between sensory evaluation and chemical composition of tomato fruits (Table 2.2). In addition, during odor threshold studies for important volatile compounds, sensory panels have described the attributes of selected volatile compounds in aqueous solutions (Buttery et al., 1990) and bland tomato homogenate (Tandon, 1998) (Table 2.3).

PAGE 26

18 Table 2.2. Sensory panel descriptors utilized by researchers during fresh-market tomato quality studies. Reference Flavor Descriptors Aroma Descriptors Bisogni and Armbruster, 1976 Overall quality, overall flavor, sweetness, acidity Not evaluated Kaderetal., 1977 Sweetness, sourness, offflavor tomato-like overall intensity Vine-like or green, fruity or floral, overall intensity Kaderetal., 1978 Sweetness, sourness, off-flavor Fruity-floral, overall intensity Watada and Aulenbach, 1979 Sweetness, acidity, saltiness, grassiness, stemminess, fruityfloral flavor, mustiness, bitterness, astringency Not evaluated Resureccion and Shewfelt, 1985 Sweetness, acidity, tomatolike, off-flavor, overall intensity, juiciness Not evaluated Kavanagh etal., 1986 Tomato flavor, sweetness, acidity/sourness Not evaluated Table 2.3. Sensory descriptors and odor thresholds for important aroma volatile Aroma volatile Compounds Reported Flavor and Aroma Descriptors z Odor thresholds y aqueous solutions bland tomato homogenate C/s-3-hexenal Green aroma, pleasant "fresh" green aroma, and desirable mouthfeel properties. 0.25 469 7>a/w-2-hexenal Lesser green character than cis-3hexenal. Not like tomato. 17 592 2-Isobutylthiazole Green leaf aroma, spoiled vine-like, horseradish type flavor. 2 102 Hexanal Green flavor 4.5 2427 6-Methyl-5-hepten-2one Fruit-like aroma 50 522 z Data from Petro-Turza, 1987; Kazeniac and Hall, 1970; Buttery et al., 1990; Tandon, 1998. y Odor thresholds expressed in nL/L. Discrimination sensory testing is one of the most useful analytical tools in sensory work. Discrimination tests, such as triangle and difference from control tests, have been designed to identify perceivable sensory differences between two or more products of

PAGE 27

19 interest. Only on the basis of a perceived difference between products can a sensory scientist justify proceeding to a descriptive sensory test in order to identify the basis for such differences (Stone et al., 1974). Difference/discrimination triangle tests were conducted to assess flavor differences in tomatoes picked at different ripeness stages, revealing the relative ease of untrained panelists to differentiate these fruits (Kader et al., 1977). Descriptive sensory analysis is defined "as a process of describing the perceived sensory characteristics of a product, usually in the order of occurrence" (Stone and Sidel, 1985). Descriptive sensory analysis involves the detection and the description of both qualitative and quantitative sensory aspects of a food product. All descriptive evaluations should be based only on perceived intensities and should be free of hedonic responses. The absolute scale values are less important than the relative differences among products to provide valuable information (Moskowitz, 1988). The two most prevalent methodologies currently employed in descriptive sensory analysis are the Quantitative Descriptive Analysis (QDA®) and the Spectrum® methods. The QDA method relies heavily on statistical analysis to determine appropriate terms, procedures and panelists. Training requires the use of product and ingredient references to stimulate the generation of terminology (descriptor terms). The panel leader acts as a facilitator rather than as an instructor, refrains from influencing the group and is not an active participant in the panel (Zook and Pearce, 1988). Panelists are screened for normal odor and taste perception using actual products from the category. In contrast, the Spectrum method is based on extensive use of reference points. Appropriate reference points reduce panel variability and allow for comparison of data over time and products

PAGE 28

20 (Lawless and Heymann, 1998). Reference points are used to precisely calibrate the panelists in the same way as pH buffers calibrate a pH meter. Both the QDA and Spectrum methods were created and successfully employed in prepared food studies where ingredients and variability between samples may be closely controlled. In the case of sensory evaluation of fruits and vegetables, the variability in chemical composition between samples and the potential for sensory quality changes over a short period of time during sample preparation and storage make these popular descriptive sensory methods less suitable. Instead, fruit and vegetable flavor research usually relies on hybrid sensory methods that incorporate the strengths of formal methods while still allowing for flexibility during training, sample preparation, and sensory sessions. Electronic Aroma Sensing Technology An electronic nose is a sensor-based instrument designed to respond to volatile chemicals present in the headspace over a sample of interest. The electronic nose is comprised of a series of non-specific semiconductor chemical sensors manufactured from materials such as metal oxides, conducting polymers, quartz microbalances and fiber optics. Such chemical sensors are useful in aroma discrimination due to their altered electrical resistance properties as a result of the interaction with volatile compound molecules (Anon., 1996). Most semiconductor materials used for odor sensing in electronic noses experience a decrease in electrical resistance properties as volatile compound concentrations adsorbed to the sensor increase. The array of non-specific sensors allows an electronic nose to detect a wide range of volatile compounds (Benady

PAGE 29

21 et al., 1995). Electronic noses not only offer the possibility of objective analysis of volatile chemicals responsible for the aroma of foods, but also detection of volatile chemicals odorless to humans (Anon., 1996). Electronic noses have numerous similarities in architecture and other properties, with biological olfaction systems such as odor delivery, nonspecific sensor/receptor response, sensor/receptor preprocessing and content addressable memory (Pearce, 1997a). In addition, similar limiting factors such as sensor/receptor drift, degeneration and poisoning, and limited sensor/receptor sensitivity, are shared by both systems (Pearce, 1997b). Polymer-based sensor arrays are able to detect organic compounds with molecular weights from 30 to 300 Daltons, values similar to what biological noses perceive (Marsili, 1995). Following odor sensing, the output from the sensor array must be analyzed. A pattern recognition procedure seeks identifying characteristics in the sensor outputs to be used in grouping algorithms for samples belonging to the same treatment. This pattern recognition approach to analyze sensor outputs closely mimics the way the human brain discriminates sensory signals and relates them to similar signals stored in our memory. The levels of aromatic compounds which accumulate as climacteric and non-climacteric fruits and vegetables ripen can be used to predict flavor quality (Benady et al., 1995). Use of an electronic nose in harvest and postharvest operations to select inferior quality and overripe fruits and vegetables has been explored using packaged blueberries (Simon et al., 1996) and melons (Benady et al., 1995). Due to the costly operation of analytical instrumentation, such as gas chromatography and mass spectroscopy, electronic noses may become useful alternative tools in commercial food handling and quality control

PAGE 30

22 operations. Thus, portable aroma sensing instruments using electronic nose technology could be useful to predict maturity during field harvesting operations for a number of fruit and vegetable crops (Benady et al, 1995). Harvest Maturity and Tomato Flavor Physiological maturity at the time of harvest is critical for normal fruit ripening to proceed (Brecht, 1987). Immature-green fruits often do not express their full flavor potential (Kader et al., 1977). Chemical composition analyses have shown that greenharvested tomatoes often result in ripe fruits with lower pH, higher ascorbic acid and total sugar levels when compared to partially ripe-harvested ones (Al-Shaibani and Greig, 1979). Accumulation of free amino acids (Perez et al., 1992), lipids, lycopene and carotenoid pigments (Stevens, 1985), and bound glycosides (Krammer et al., 1995) acting as flavor precursors could explain flavor differences between green-harvested tomatoes and those harvested with color, when analyzed at ripe stage (Watada and Aulenbach, 1979). Lycopene concentrations have been shown to be higher in tomatoes harvested at "vine-ripe" stage compared to those ripened postharvest (Mencarelli and Saltveit, 1988). Numerous studies have focused on the relation between ripe tomato quality and chemical composition from green-harvested tomatoes (Bisogni and Armbruster, 1976; Kader et al., 1977; Kavannagh et al., 1986; Watada and Aulenbach, 1979) with relative success correlating sensory parameters to compositional quality. However, without a thorough understanding of the factors influencing tomato flavor, significant correlations between sensory attributes and chemical composition parameters cannot be interpreted using a cause and effect relationship.

PAGE 31

23 There is evidence that green-harvested tomatoes supplied with exogenous ethylene (100 uL/L C2H4 for 1 day) resulted in significantly altered flavor volatile profiles (McDonald et al., 1996). Ethylene treatment resulted in tomatoes with significantly reduced levels of hexanal, 6-methyl-5-hepten-2-one, geranylacetone, methanol, 2-isobutylthiazole, and l-nitro-2-phenylethane. Nonetheless, the effects of harvest maturity on ripe tomato flavor quality and the contribution of volatile profiles to inferior sensory characteristics have not been clearly established. Postharvest Treatments and Tomato Flavor. It has been documented in the past how postharvest treatments can affect production of aroma volatiles, thus influencing fruit flavor (Willaert et al., 1983). The senescence of fruits during extended storage periods results in a decreased formation of volatile compounds, probably from reduced enzymatic activity or loss of flavor precursors (Yahia et al., 1992). In 'Anjou' pears, exposure to chilling temperatures leads to an increased unsaturation of linolenic and linoleic acids (Gerasopoulos and Richardson, 1995), thus increasing substrate levels for lipoxygenase activity. The accessibility of substrates for hydroperoxide lyase and other key enzymes during fruit ripening could represent a limiting factor in the biosynthesis of flavor volatiles (Riley, 1996). The effects of storage temperature and physiological maturity, among others, have shown significant effects on ripe tomato flavor (Stern et al., 1994; Buttery et al., 1987;Kaderetal., 1977).

PAGE 32

24 Storage Temperature Management Chilling injury is a physiological disorder resulting from the exposure of susceptible plant tissues to temperatures above 0°C, but usually below 15°C (Crooks and Ludford, 1984). The major symptoms of chilling injury (CI) in tomato are: increased susceptibility to Alternaria alternata (McColloch and Worthington, 1952), delayed, partial, or uneven ripening, pitting (Dodds et al, 1991), and enhanced fruit softening (Chomchalow, 1991). Tomato sensitivity to CI appears to be influenced by ripening, probably due in part to dramatic changes in membrane lipid composition as affected by the exposure to chilling temperatures (Whitaker, 1994; McDonald et al., 1996; 1998). Mature-green (Autio and Bramlage, 1986) and breaker-stage (Hobson, 1981) tomatoes appear to be the most sensitive to CI. The extent of CI is dependent upon the interaction between chilling temperature, length of exposure time, and varietal susceptibility (Chomchalow, 1991). Chilling temperatures used for storage of breaker stage tomatoes (2°C for 14 days) significantly affected the aroma volatile profiles at ripe stage (McDonald et al., 1996; Buttery et al., 1987). Reduced levels of hexanal, 6-methyl-5hepten-2-one, geranylacetone, methanol, 2-isobutylthiazole, and l-nitro-2-phenylethane were attributed to chilling temperature storage. Hobson (1987) recommended 8.5°C as the minimum storage temperature to be tolerated by breaker-stage tomato fruits, as a 10°C storage temperature appeared to cause only minor changes in fruit composition (Hobson, 1981). Even though Morris (1954) reported CI in tomato fruits after 10 days at 10°C, the storage of partially-ripe tomatoes at this temperature has become an increasingly common practice in Europe (Hobson, 1987). In the U.S. it is common for tomato handlers to store and ship green or partially ripe tomatoes at 10°C for extended

PAGE 33

25 periods of time to maximize marketing flexibility (S. A. Sargent, personal communication). However, Stern et al. (1994) found tomatoes harvested at breaker stage and ripened at 15°C contained higher concentrations of volatile compounds compared to those stored at 20°, 10° or 5°C. Nevertheless, the effect of postharvest storage at or below 15°C, often recommended for commercial operations, needs to be addressed. Alleviation of CI symptoms has been demonstrated by intermittent warming cycles (Artes et al., 1998) or high temperature treatment prior to storage at low temperatures (McDonald et al, 1998). Sabehat et al. (1996) documented a consistent relationship between heat-shock protein accumulation and persistence with tolerance to CI symptoms in tomatoes, thus suggesting a prominent role of heat-shock proteins in tomato fruit tissue acclimation and tolerance to low temperature storage. Short exposures to temperatures above 35°C have been shown to reversibly inhibit tomato ripening (Lurie and Klein 1991), ethylene biosynthesis (Biggs et al., 1988) and lycopene accumulation (Ogura et al, 1975). Levels of most aroma volatile compounds in heat treated/chilled fruit were intermediate between chilled and non-chilled controls (McDonald et al., 1998). Even though there is strong evidence regarding the effectiveness that high temperature pre-treatments (38°C for 2 days) have on visual CI symptom alleviation (Sabehat et al., 1996; Lurie et al., 1996; Artes et al., 1998), there is no information regarding potential benefits of high temperature pre-treatments on chilling-induced flavor alteration.

PAGE 34

26 Research Objectives The objectives of this research project were to: 1 . Investigate the effects of green tomato harvest maturity on flavor and aroma quality at ripe stage utilizing discrimination sensory tests, GC aroma volatile profiles, and chemical composition analyses. 2. Explore the potential use of exogenous ethylene treatment or electronic nose sensor arrays as nondestructive screening tools for green-tomato physiological maturity at harvest. 3. Identify ripe flavor and aroma quality differences between fresh market tomatoes harvested at green-stage or partially ripe stages utilizing a trained descriptive sensory panel, GC aroma volatile profiles, and chemical composition analyses. 4. Determine the effects of postharvest storage temperatures below 20°C on ripe tomato flavor and aroma quality, and, explore the effectiveness of high-temperature pretreatments (38°C for 2 days) as a means to alleviate chilling-induced flavor and aroma quality alterations.

PAGE 35

CHAPTER 3 RIPENESS STAGE AT HARVEST AFFECTS CHEMICAL COMPOSITION AND COLOR QUALITY OF TABLE-RIPE TOMATOES Introduction Dissatisfaction with fresh market tomatoes has been attributed to inferior flavor and aroma, possibly affected by cultivar, cultural practices, growing conditions, maturity stage at harvest, and inadequate or inappropriate postharvest handling practices (Stevens, 1985; Kader et al, 1977; Baldwin et al., 1991c). Traditionally, tomatoes are grown in Florida during winter months and shipped to distant markets. Due to the lack of accurate visible indicators of physiological maturity, commercial green harvest operations must rely almost entirely on fruit size as harvest index. Physiological maturity of greenharvested tomatoes has been circumstantially related to ripe, fresh tomato flavor and quality (Kader et al., 1977; Watada and Aulenbach, 1979; Al-Shaibani and Greig, 1979). It has been proposed that immature-harvested tomato fruits will never ripen properly (Brecht et al., 1991) or develop their full flavor potential (Kader et al., 1978). Numerous researchers have found increasing flavor desirability with increased maturity at harvest (Watada and Aulenbach, 1979; Kader et al, 1977). Nonetheless, chemical composition data collected by these authors did not always concur with their sensory panel results. Recent studies (Hobson and Bedford, 1989; Malundo et al., 1995) have determined that compositional parameters, such as acidity and sugar content, are closely 27

PAGE 36

28 related to overall tomato acceptability. Tomatoes picked at the red-ripe stage were described as sweeter and with stronger fruity-floral character compared to those harvested at green stage (Watada and Aulenbach, 1979). The following studies were conducted to document compositional differences in table-ripe tomato fruit as related to harvest date, ripeness stage and physiological maturity at harvest. Also investigated were the effects of the exposure of green tomatoes to ethylene (to initiate ripening) on compositional quality. Materials and Methods Tomato fruits from three tomato cultivars were harvested at various ripeness stages from experimental research plots at Collier Farms, Naples, FL. During the first experiment conducted in December 1995, 'Agriset-761' (Agrisales Inc., Plant City, FL) and 'CPT-5' (Collier Farms, Naples, FL) tomatoes were harvested at green (stage 1, 0% red color), breaker (stage 2, <10% red color), light red (stage 5, <90% red color) and red (stage 6, >90% red color) ripeness stages (USD A, 1976). In a second experiment conducted in January 1996, 'Agriset-761' and 'CPT-5' tomatoes were harvested at green, breaker and pink ripeness (stage 4, 30-60% red color) stages. Immediately after harvest, tomatoes were transported to Gainesville, FL, washed with chlorinated water and towel dried, and stored at 20°C and 85-90% RH for subsequent ripening. In a third experiment conducted in March 1996, 'Agriset-761' and 'CPT-5' tomatoes were harvested at green stage and transported the same day to Gainesville, FL for subsequent ethylene gassing treatment. Tomatoes were washed with chlorinated

PAGE 37

29 water and towel dried prior to ethylene treatment inside sealed chambers at 20°C where a humidified 100 uL/L ethylene/air mixture was constantly supplied using a flow-through system. The gassing chambers were opened daily to remove any fruits that had initiated ripening evidenced by red color on the surface (attained breaker stage). Ethylene treatment was terminated after 7 days, at which time all tomatoes had initiated ripening. Following ethylene treatment, tomatoes were placed in air at 20°C and 85-90% RH for subsequent ripening. In all three experiments, tomatoes were analyzed once they attained the table-ripe stage, defined as red stage coloration and significant fruit softening (3-4 mm deformation when a constant 9.8N force was applied to the equator for 5 sec) (Gull, 1980). Average number of days to table-ripe stage, fruit deformation and fruit equatorial diameter data were recorded. In addition, fruit color was measured on four equidistant locations around the fruit equator using a tritimulus colorimeter (Minolta CR-300, Ramsey, NJ). Color data were expressed in terms of lightness coefficient (L*), a* and b* values, from which hue angle (tan" 1 b*/a*) and chroma values (a* 2 + b* 2 ) m were calculated. Upon reaching table-ripe stage, fruit samples from different ripeness stages at harvest and ethylene exposure times were collected for chemical composition analyses. The chemical composition assays conducted (n=15 fruits/treatment) included pH, titratable acidity (expressed as % citric acid), soluble solids (°Brix), vitamin C (mg ascorbic acid/100 g fresh weight), and total sugars (% of fruit fresh weight). Tomato homogenates from each of the harvest maturity treatments were centrifuged at 18,000 X g n and 5°C temperature. The supernatant was filtered using cheesecloth, stored inside scintillation vials and frozen at -20°C for later analysis.

PAGE 38

30 Titratable acidity, expressed as % citric acid, was determined by titrating 1.5 g of tomato supernatant to 8.2 pH with a 0. 1 N NaOH solution using an automatic titrimeter (Fisher Scientific, Pittsburgh, PA). Soluble solids content, expressed as °Brix, was measured using a tabletop digital refractometer (Abbe Mark II, Reichart-Jung, Buffalo, NY) and pH measurements were conducted using a digital pH-meter (Corning model 140). For the vitamin C assays, tomato homogenates (n=4) from each treatment (2 g/sample) were stabilized using 20-mL of acid mixture (6% HPO3 containing 2 N acetic acid in water) prior to centrifugation. Vitamin C content was analyzed using a dinitrophenylhydrazine (DNPH) method adapted from Terada et al. (1978). Stabilized homogenate samples (1 mL/sample) were mixed with 50 uL of a 0.2% water solution of 2,6-dichlorophenolindolphenol (DCIP), 1 mL of 2% thiourea solution in water and 0.5 mL of DNPH. Samples were incubated at 60°C for 3 hours, then cooled on ice prior to adding 2.5 mL of 90% sulfuric acid (H 2 S0 4 ) to each sample. Absorbance was read at 540 nm using a spectrophotometer (Beckman DU-20, Irvine, CA). Vitamin C concentrations (mg ascorbic acid/lOOg fresh weight) were determined from a standard curve using ascorbic acid standards. Total soluble sugar contents were analyzed using a spectrophotometric method adapted from Dubois et al. (1956). Each tomato homogenate sample (n=4/treatment) was diluted with distilled water (1:500 mL), then a 0.5 mL sample of the dilute homogenate was combined with 0.5 mL of 5% phenol (w/v) and mixed thoroughly. Concentrated sulfuric acid (2.5 mL/sample) was added and then samples were cooled to room temperature prior to spectrophotometer readings. Absorbance was read at 490 nm using a

PAGE 39

31 spectrophotometer (Beckman DU-20, Irvine, CA). Total sugar content present in tomato samples was determined from a standard curve derived from sucrose standards. Results and Discussion 'Agriset-76r . Green tomatoes harvested in December 1995 required an average of 13.6 days to attain table-ripe stage during storage at 20°C, while those harvested in January 1996 required 15.2 days (Table 3.1). Breaker-harvested fruits showed a similar trend, where those harvested in January required longer time to attain table-ripe stage, compared to those harvested in December (10.7 and 9.9 days, respectively). In contrast, pink fruit harvested in January required less time to ripen than those harvested at light red stage in December (5.9 and 6.8 days, respectively). Equatorial diameter showed no significant differences between fruit from the different ripeness stages at either harvest date. However, fruit harvested in January were of slightly greater diameter than those harvested in December. Chemical composition analyses for 'Agriset-761' showed no significant treatment effects between the four ripeness stages harvested in December. However, tomatoes from the January harvest picked at breaker or pink stages were found to have significantly lower pH compared to those harvested at green stage (4.35, 4.33, and 4.41, respectively) (Table 3.1). Comparing pH values for the three harvest dates revealed that pH values were comparable regardless of ripeness stage at harvest. Titratable acidity was significantly higher in tomatoes harvested at pink stage (January) compared to those harvested at earlier ripeness stages. Meanwhile, no significant differences in titratable

PAGE 40

32 acidity were found for tomatoes harvested in December. Soluble solids content decreased from green to breaker harvest, but increased at later ripeness stages at harvest. A positive relationship between titratable acidity and soluble solids content was evidenced in the December harvest; light-red harvested tomatoes were highest in titratable acidity and soluble solids content, breaker-harvested tomatoes were lowest in both parameters, while redand green-harvested fruit were intermediate for both parameters (Table 3.1). Ripe tomato vitamin C content for tomatoes harvested in January showed no significant differences between ripeness stages. Nonetheless, greater average vitamin C contents were found for pink-harvested fruit (16.4 mg/100 g FW), compared to those harvested at either green or breaker stage (10.8 and 10.3 mg/100 g FW, respectively) (Figure 3.1). Table 3.1. Compositional quality parameters for table-ripe 'Agriset-761' tomatoes harvested at several ripeness stages and two harvest dates. Compositional Parameters z December 1995 January 1996 Green Breaker Light Red Red Green Breaker Pink Firmness y 2.81 b 3.86 a 4.01 a 3.90 a 4.03 a 3.91 a 4.01 a Fruit Diameter (mm) 76.8 a 77.2 a 79.5 a 78.2 a 87.1 a 86.8 a 84.0 a Days harvest to ripe stage 13.6 a 9.91 b 6.8 c 3.4 d 15.2 a 10.7 b 5.9 c PH 4.40 a 4.41 a 4.36 a 4.42 a 4.41 a 4.33 b 4.32 b Titratable Acidity y 0.89 a 0.86 a 0.97 a 0.90 a 0.79 b 0.79 b 0.85 a Soluble Solids (°Brix) 3.80 a 3.40 a 4.07 a 4.00 a 3.88 a 3.57 b 3.79 a z Within each harvest date (n = 1 5 fruit/treatment), values 1 'ollowed 5y different letters between ripeness stages are significantly different at the 5% level according to Duncan's Multiple Range Test. y Fruit firmness (deformation in mm) induced by a constant 9.8N force applied for 5 sec. "Titratable acidity expressed as % of citric acid. In the third experiment, where 'Agriset-761' tomatoes harvested green were subsequently treated with exogenous ethylene, fruit diameter varied with the time of ethylene treatment required to reach breaker stage. Even though tomatoes exposed for 1day ethylene treatment had significantly greater diameter compared to longer ethylene

PAGE 41

33 exposure times, there was no consistent trend relating fruit diameter and ethylene gassing exposure time (Table 3.2). The length of ethylene exposure to attain breaker stage appeared to have a significant effect on subsequent ripening rate, expressed as the number of days to attain table-ripe stage from breaker stage. Tomatoes exposed for 4 and 5 days of ethylene treatment had significantly shorter average ripening times to table-ripe stage (7.6 and 7.9 days) compared to the rest of ethylene exposure times. Chemical composition parameters at table-ripe stage were also affected by the ethylene exposure time from green-harvested tomatoes. Soluble solids contents were significantly lower in fruit exposed for 5 days of ethylene treatment, however, fruit exposed for 6 or 7 days did not show significant soluble solids content differences when compared to fruits exposed for less than 5 days of ethylene treatment (Table 3.2). Titratable acidity and pH values showed no significant differences between treatments, however, tomatoes exposed for 5 days of ethylene did have lower acidity compared to the rest of the treatments. Concurring with soluble solids results, total sugars in fruit exposed for 5 days of ethylene were lower compared to the remaining treatments. No significant differences in vitamin C content were found between tomatoes with different ethylene exposure times. However, it is important to note that fruits exposed for 6 and 7 days of ethylene treatment had the lowest vitamin C contents (Figure 3.2). From these experiments it was observed that the ripening rate for 'Agriset-761' tomatoes was affected by harvest date. Light-red stage tomatoes harvested in December required longer ripening times than pink stage fruits harvested in January. Breaker stage tomatoes required about 1 additional day to ripen between December and January harvest dates (9.9 and 10.7 days, respectively). A similar response was obtained from green

PAGE 42

34 stage tomatoes, where ripening time differential for the two harvest dates approached 2 days (13.6 and 15.2 days for December and January, respectively). Green-harvested fruit treated with exogenous ethylene (March 1996) required ripening times similar to those tomatoes harvested at breaker stage in the two previous experiments. Similar ripening times suggested that ethylene exposure accelerates the onset of ripening, however, once breaker stage has been achieved ripening rates were practically unaffected. Table 3.2. Compositional quality parameters for table-ripe 'Agriset-761' tomatoes harvested in March 1996 at green stage and supplied exogenous ethylene (100 uL/L) Compositional Parameters 2 Ethylene gassing exposure times 1 day 2 day 3 day 4 day 5 day 6 day 7 day Firmness y 3.70 c 4.52 ab 4.48 ab 4.36 b 4.70 ab 5.14a 4.81 ab Fruit diameter (mm) 89.7 a 77.4 be 80.7 be 80.3 be 82.6 ab 73.0 c 79.0 be Days harvest to ripe stage 9.67 a 9.78 a 9.1 ab 7.6 c 7.9 be 9.67 a 10.2 a Soluble solids (°Brix) 3.95ab 4.04 a 4.10a 4.01 a 3.35 b 3.93ab 4.35 a PH 4.30 a 4.30 a 4.29 a 4.27 a 4.27 a 4.22 a 4.26 a Titratable acidity " 0.76 a 0.79 a 0.76 a 0.80 a 0.69 a 0.77 a 0.77 a Total sugars (% fresh wt.) 1.64ab 1.70 ab 1.80 a 1.69 ab 1.43 b 1.61ab 1.68 ab 2 Values followed by different letters between ethylene exposure times (n = 15 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. y Fruit firmness (deformation in mm) induced by a constant 9.8N force applied for 5 sec. "Titratable acidity expressed as % of citric acid. Although soluble solids content varied between harvest stages and harvest dates there was no consistent pattern. Titratable acidity, however, seemed to be affected by ripeness stage at harvest. In the first two experiments, a consistent trend of increasing acidity with increasing ripeness was evident. This observation concurs with Kavanagh et al. (1986), who reported lower acidity with decreasing maturity at harvest. In this study, titratable acidity was not influenced by green tomato ethylene exposure time. However, it was generally observed that there were lower pH values in green-harvested tomatoes

PAGE 43

35 treated with ethylene in the third experiment when compared to those harvested at the same stage and not treated with ethylene in previous experiments. Fruit color was significantly affected by ripeness stage at harvest for 'Agriset761' tomatoes. For both harvest dates, significantly lower hue angles were documented for fruit harvested at breaker or green stages compared to those harvested at later ripeness stages (Table 3.3). In addition, a* values were significantly lower for tomatoes harvested at red stage (December) compared to the other ripeness stages. Although not significant, a consistent trend of reduced L* values from later ripeness stages at harvest, was evidenced during both harvest dates. Table 3.3. Color quality parameters for table-ripe 'Agriset-761' tomatoes harvested at different ripeness stages on two harvest dates. Color Parameters z December 1995 January 1996 Green Breaker Light Red Red Green Breaker Pink L* value 42.23 a 42.97 a 41.63 a 41.54 a 43.69 a 43.45 a 42.95 a a* value 27.47 a 28.07 a 26.11 ab 23.6 b 22.36 a 20.87 a 21.71 a b* value 25.51 a 25.75 a 27.29 a 26.56 a 22.41 b 20.32 c 23.89 a Hue angle y 42.88 b 42.50 b 46.35 a 48.41 a 45.11 b 44.27 b 47.77 a Chroma " 37.51 a 38.11 a 37.80 a 35.57 a 31.7a 29.2 b 32.30 a z Within each harvest date (n = 15 fruit/treatment), values followed by different letters between ripeness stages are significantly different at the 5% level according to Duncan's Multiple Range Test. y Hue angle = Tan" 1 (b*/a*) "Chroma = (a* 2 + b* 2 ) 1 ' 2 L* values were significantly lower for fruit exposed for 1 to 5 days of ethylene compared to those exposed for more than 5 days (Table 3.4). Significant differences in average a* values showed that tomatoes exposed for 2 days of ethylene treatment were lowest (least red: a* = 22.56); fruits from all treatments were significantly lower than those exposed for 7 days (a* = 27.49). Tomatoes exposed for 1 and 2 days of ethylene treatment were lowest in b* and chroma values when compared to those tomatoes

PAGE 44

36 exposed for over 5 days. In addition, tomatoes exposed for 1 day ethylene had significantly lower hue angles compared to those exposed for 7 days (43.48 and 47.61, respectively) (Table 3.4). Table 3.4. Color quality parameters for table-ripe 'Agriset-761' tomatoes harvested at Color Parameters z Ethylene gassing exposure time 1 day 2 day 3 day 4 day 5 day 6 day 7 day L* value 43.5 b 42.05 b 41.88 b 41.83 b 43.04 b 43.79 ab 45.65 a a* value 23.2 be 22.56 c 23.81 be 24.1 be 25.18 b 25.1 b 27.49 a b* value 22.0 d 22.56 d 23.42 cd 24.1 bed 25.4 be 26.17 b 30.13 a Hue Angle y 43.48 b 45.04ab 44.45 ab 44.9 ab 45.3 ab 46.27 ab 47.61 a Chroma * 31.98 d 31.94 d 33.44 be 34.1 bed 35.8 ab 36.28 b 40.78 a z Values followed by different letters between ethylene exposure times (n = 15 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. y Hue angle = Tan" 1 (b*/a*) "Chroma = (a* 2 + b* 2 ) 1/2 'CPT-5'. For 'CPT-5' tomatoes, significant differences in fruit deformation at table-ripe stage were documented, as was the case for 'Agriset-761' harvested in December. In contrast, fruit harvested in January showed no significant differences in fruit firmness, thus reflecting a greater degree of uniformity criteria for determination of the in table-ripe stage. Fruit diameter was considerably greater in 'CPT-5' fruit harvested in January compared to those in December, however, there were no significant differences in fruit diameter for either harvest date (Table 3.5). The ripening times for fruit harvested at different ripeness stages was also influenced by harvest date, where fruit harvested in January generally required longer ripening times (except for pink stage). The effects of ripeness stage on pH values were contradicting between harvest dates. In December, fruits harvested at red stage had significantly higher pH compared to

PAGE 45

37 those harvested at green stage (4.41 and 4.32), while pH values for breaker and light redharvested fruit were intermediate (4.34 and 4.36). Fruits harvested at breaker or pink stage in January were significantly lower in pH compared to green-harvested ones (4.35, 4.33 and 4.41, respectively) (Table 3.5). Titratable acidity content for green-harvested fruit was significantly lower than breaker-harvested (December), while there were no significant differences between tomatoes harvested in January. For tomatoes harvested in January, vitamin C content showed an increasing trend with increasing ripeness at harvest, although, no significant differences were found (Figure 3.1). Table 3.5. Compositional quality parameters for table-ripe 'CPT-5' tomatoes harvested at Compositional Parameters z December 1995 January 1996 Green Breaker Light Red Red Green Breaker Pink Firmness y 2.60 b 3.29 ab 3.16 ab 3.37 a 3.88 a 3.74 a 4.09 a Fruit Diameter (mm) 81.2a 76.1 a 84.8 a 79.3 a 94.3 a 88.5 a 90.0 a Days harvest to red stage 12.4 a 10.8 a 8.0 b 5.8 c 15.1 a 11.4b 7.1 c PH 4.32 b 4.34 ab 4.36 ab 4.41 a 4.41 a 4.35 b 4.33 b Titratable Acidity x 0.76 b 0.97 a 0.87 ab 0.84 ab 0.82 a 0.80 a 0.82 a Soluble Solids (°Brix) 4.02 a 3.71 a 3.92 a 3.95 a 3.92 a 3.79 a 3.81 a z Within each harvest date (n = 1 5 fruit/treatment), values followed by ripeness stages are significantly different at the 5% level according to Test. y Fruit deformation (in mm) induced by a constant 9.8N force applied for 5 sec "Titratable acidity expressed as % of citric acid. different letters between Duncan's Multiple Range In 'CPT-5' tomatoes, fruit diameter varied significantly with different ethylene exposure times. Fruit which reached breaker stage after 2 days of ethylene had the largest diameter (84.4 mm) while fruit exposed for 7 days had the smallest average diameter (72.3 mm) (Table 3.6). Diameter data from the remaining ethylene treatments showed no relationship to ethylene exposure time. The number of days from breaker to table-ripe stages were not influenced significantly by ethylene exposure time to attain

PAGE 46

38 breaker stage, which contradicts the data for 'Agriset-761'. Soluble solids generally followed an increasing trend with number of days of ethylene treatment. Tomatoes exposed for 1 and 2 days of ethylene of treatment had the lowest soluble solids contents (3.48 and 3.49° Brix, respectively) while those exposed for 6 and 7 days had the highest (4.18 and 4.60° Brix) (Table 3.6). The highest pH value was obtained from fruit that required 7 days of ethylene treatment while the lowest from those that required only 2 days of ethylene treatment. Tomatoes exposed for 7 days of ethylene were not only highest in pH but also highest in titratable acidity (0.83% citric acid) and total sugars (1.94 %). Vitamin C content showed no significant differences between ethylene treatments, however, a consistent trend of decreasing vitamin C with increasing ethylene gassing exposure time was evident (Figure 3.2). Table 3.6. Compositional quality parameters for table-ripe 'CPT-5' tomatoes harvested at Compositional Parameters z ethylene gassing exposure time 1 day 2 day 3 day 4 day 5 day 6 day 7 day Deformation y 4.50 a 4.56 a 4.56 a 4.40 a 4.30 a 4.80 a 4.19a Fruit Diameter (mm) 83.3 ab 84.4 a 80.7 ab 76.6 be 80.5 ab 79.3 ab 72.3 c Days to table-ripe stage 13a 12.4 a 12.8 a 11.7a 10.3 a 12.0 a 11.0a Soluble solids (°Brix) 3.48 b 3.49 b 4.02 ab 3.60 b 3.70 ab 4.18 ab 4.60 a PH 4.28 abc 4.22 b 4.30 abc 4.34 ab 4.24 be 4.28 abc 4.38 a Titratable acidity x 0.74 a 0.67 a 0.80 a 0.70 a 0.79 a 0.77 a 0.83 a Total sugars (% fresh wt.) 1.45 b 1.79 ab 1.68 ab 1.59 ab 1.76 ab 1.75 ab 1.94 a significantly different at the 5% level according to Duncan's Multiple Range Test. y Fruit deformation (in mm) induced by a constant 9.8N force applied for 5 sec. "Titratable acidity expressed as % of citric acid. 'CPT-5' tomatoes ripened darker (decreased L* value) with increasing ripeness stage at harvest (Table 3.7). This was also observed in 'Agriset-761', particularly in light-red (December) and pink (January) stage harvested fruits. In general, hue angles

PAGE 47

39 were higher for tomatoes harvested in January, however, for both harvests there was a trend of increasing hue angles (less red) with increasing ripeness stage at harvest. During the January harvest, chroma and b* values were significantly lower in fruit harvested at green or breaker stages (36.65 and 36.25, respectively) compared to those harvested at light-red or red ripeness stages (39.18 and 38.77, respectively) (Table 3.7). Table 3.7. Color quality parameters for table-ripe 'CPT-5' tomatoes harvested at several Color Parameters z Decern t>er 1995 January 1996 Green Breaker Light Red Red Green Breaker Pink L* value 42.61 a 42.65 a 40.90 b 40.53 b 45.30 a 44.36 a 41.89 b a* value 26.99 a 26.98 a 28.40 a 28.21 a 22.54 a 21.27 a 21.62 a b* value 24.77 b 24.18 b 26.98 a 26.58 a 21.89 ab 20.67 b 23.02 a Hue angle y 42.56 a 41.89 a 43.56 a 43.29 a 44.18 b 44.23 b 46.80 a Chroma x 36.65 b 36.25 b 39.18 a 38.77 a 31.44 a 29.69 b 31.64 a z Within each harvest date (n = 15 fruit/treatment), values followed by different letters between ripeness stages are significantly different at the 5% level. y Hue angle = Tan" 1 (b*/a*) "Chroma = (a* 2 + b* 2 ) 1/2 'CPT-5' tomatoes exposed for 7 days of ethylene treatment ripened with significantly higher lightness coefficient (L* = 48.06) compared to tomatoes exposed for fewer days of ethylene treatment (Table 3.8). A similar increasing trend was evident for a* and chroma values. Tomatoes exposed for 1 or 2 days of ethylene treatment had significantly lower a* values (less red) (22.90 and 22.42, respectively) compared to those exposed for 6 or 7 days of treatment (25.84 and 25.99, respectively). Meanwhile, chroma and b* values were significantly higher in fruit exposed for 6 or 7 days of ethylene treatment (37.20 and 39.60, respectively) compared to the rest of the treatments. Hue angles showed no significant differences between treatments.

PAGE 48

40 Table 3.8. Color quality parameters for table-ripe 'CPT-5' tomatoes harvested at green Color Parameters z Ethylene gassing exposure time 1 day 2 day 3 day 4 day 5 day 6 day 7 day L* value 43.50 b 42.38 b 42.60 b 42.16 b 43.63 b 43.42 b 48.06 a a* value 22.90 b 22.42 b 23.99 ab 24.07 ab 24.32 ab 25.84 a 25.99 a b* value 25.20 be 23.08 c 25.09 be 25.24 be 25.27 be 26.66 b 29.87 a Hue Angle y 47.70 a 45.69 a 46.33 a 46.35 a 46.15 a 45.81 a 49.02 a Chroma * 34.10 cd 32.24 d 34.78 cd 34.89 cd 35.09 cd 37.20 b 39.60 a z Values followed by different letters between ethylene exposure times (n = 1 5 fruit/treatment) are significantly different at the 5% level according to Duncan's Multiple Range Test. y Hue angle = Tan" 1 (b*/a*) "Chroma = (a* 2 + b* 2 ) 1/2 For 'CPT-5' tomatoes there was no clear relationship between physical fruit measurements (firmness and diameter) and ripeness stage at harvest. Fruit size, a possible indicator of physiological maturity at harvest, proved to be a poor indicator of maturity. Similar to 'Agriset-761', there were considerable differences in the length of time required to attain ripe stage between harvest dates. Nonetheless, 'CPT-5' tomatoes in general required longer periods of time to ripen compared to 'Agriset-761' tomatoes harvested during the same dates and ripeness stages. Green-harvested tomatoes required longer times to ripen during the January experiment. Such observation is relevant since this could ultimately indicate a lesser degree of physiological maturity at harvest. Watada and Aulenbach (1979) concluded that green-harvested tomatoes exposed for more than 15 days to ripen at room temperature (23 °C) would probably result in inferior quality at ripe stage. Soluble solids content varied widely between ripeness stages and harvest dates. Nonetheless, green-harvested tomatoes exposed for 6 and 7 days of ethylene had consistently higher soluble solids content compared to the other treatments, coinciding with an earlier report that, upon ripening, immature-green harvested had higher soluble

PAGE 49

41 solids content at ripe stage (Kavanagh et al., 1986). In addition, pH values were considerably higher in tomato fruit exposed for extended ethylene treatment. Soluble solids content / titratable acidity ratios (SSC/TA) were considerably higher for greenharvested, ethylene-treated tomatoes from both 'Agriset-761' and 'CPT-5' cultivars (data not shown). However, except for 'Agriset-761' harvested in December 1995, greenharvested tomatoes without ethylene treatment also showed considerably higher SSC/TA ratios. This supports the contention that ethylene treatment doesn't influence sugar/acid ratios or flavor directly (Kader et al., 1978). Compositional studies combined with sensorial analyses have demonstrated significant differences in sweetness, sourness and off-flavor perception without significant differences in soluble solids or sugar contents (Bisogni and Armbruster, 1976; Kader et al., 1977; Watada and Aulenbach, 1979). In contrast, significant differences in titratable acidity and pH do not necessarily reflect significant sensory quality changes. Changes in acidity of lesser than 0.1% citric acid or of less than 0.2 pH points were not perceived during sensory studies with model solutions (Gould, 1978). Vitamin C contents for 'CPT-5' tomatoes were slightly greater than those found in 'Agriset-761' tomatoes regardless of ripeness stage or harvest dates. A trend of increasing vitamin C concentrations with increasing ripeness at harvest was evident for both cultivars. Except for tomatoes exposed for more than 5 days (CPT-5) and 6 days (Agriset-761) ethylene, green-harvested tomatoes with supplemental ethylene treatment had consistently higher vitamin C contents when compared to those fruits not treated with ethylene. This agreed with a previous study (Watada et al., 1976), in which greater variability in vitamin C content was due to cultivar differences rather than harvest

PAGE 50

42 maturity. In addition, green-harvested tomatoes treated with ethylene were shown to have higher vitamin C contents when compared to those that were not treated. Conclusions Ripeness stage at harvest and harvest dates influenced compositional parameters and fruit color at table-ripe stage. However, no consistent relationship could be determined at table-ripe stage between ripeness stage at harvest and compositional quality. Soluble solids contents were highest in 'CPT-5' tomatoes harvested at green stage compared to later ripeness stages. Furthermore, green-harvested 'CPT-5' tomatoes that required 7 days of ethylene treatment to attain breaker stage had the highest soluble solids, pH, titratable acidity and total sugars compared to fruit exposed for shorter ethylene treatments. On the other hand, green-harvested 'Agriset-761' that required 7 days of ethylene treatment only had higher soluble solids compared to shorter ethylene exposure times. Results for 'Agriset-761' and 'CPT-5' tomatoes suggest ripe fruit color was significantly affected by ripeness stage at harvest and harvest date. In general, lightness coefficients (L*) were lower, regardless of ripeness stage, for tomatoes harvested during December 1995 compared to subsequent harvests. Tomatoes harvested beyond the pink ripeness stage developed a darker red color compared to green or breaker-harvested fruit.

PAGE 51

43 ah '53 on
PAGE 52

44 5) tfa 20 16 36 i 1 day I I 2 days 32 H 3 days i i 4 days 28 H 5 days MH 6 days 24 1 7 days 12 8 4 0 TT Agriset-761 CPT-5 Tomato Cultivars Figure 3.2. Vitamin C content (mg/ lOOg fresh weight) from table-ripe 'Agriset-761 ' and 'CPT-5' tomatoes harvested at green stage and supplied exogenous ethylene (100 uL/L) for 1 to 7 days until breaker stage was attained. Deviation bars represent the Duncan's critical range for significant differences at the 5% level according to Duncan's Multiple Range Test.

PAGE 53

CHAPTER 4 POTENTIAL FOR NONDESTRUCTIVE QUALITY SCREENING OF TOMATOES WITH ETHYLENE OR ELECTRONIC NOSE SENSOR ARRAYS Introduction Throughout tomato (Lycopersicon esculentum Mill.) production and marketing, efforts are made to maintain optimal visual quality (uniform color, absence of decay, proper firmness, etc.) to attract customers. As a consequence, disorders not readily detectable during sorting operations have received less attention. Visual appearance is a critical factor driving the initial purchase, but subsequent purchases are influenced greatly by eating quality (flavor, aroma and mouthfeel). Overall fresh tomato quality is determined not only by fruit appearance and firmness, but also flavor, aroma, and nutritive value. Most fresh market tomatoes sold in U.S. supermarkets are harvested before they are "table-ripe" because retailing ripe tomato fruit is not practical within the current long distance handling and marketing system (Kader et al, 1978). In Florida, over 85% of the fresh market fruit are green-harvested for transport to distant markets. Such factors as harvesting immature-green fruit, incurrence of mechanical injuries, and storage at less than ideal temperatures contribute to quality loss from the shipping point to the consumer's table. 45

PAGE 54

46 Recent literature has documented the effects of postharvest mismanagement on tomato quality. A negative effect from immature-green harvest was found after ripening regarding chemical composition, overall sensory acceptance, and aroma volatile profiles (Kader et al., 1978; Maul et al., 1997a). Internal bruising, a disruption of regular ripening of internal tomato tissues, occurs readily in breaker stage fruit (Sargent et al., 1989; 1992), and results in altered chemical composition, aroma volatile profiles, and sensory quality (Moretti et al, 1997; Sargent et al., 1997). Finally, low temperature storage (2°C for 14 days) was shown to suppress the levels of six important aroma compounds in ripe tomato fruit (McDonald et al., 1996; Buttery et al., 1987). Distinctive external visual symptoms resulting from these treatments were never observed. "Electronic nose" (EN) sensor arrays are used to classify specific samples of interest based on their headspace volatiles. The electronic nose consists of a series of non-specific chemical sensors, each of which shows characteristic responses to the volatile chemicals within the headspace over a sample (Anon., 1996). The electrical resistance of chemical sensors changes as the concentration of volatile compounds present in the headspace increases. These changes in electrical resistance (output) are then analyzed by a pattern recognition procedure, such as multivariate discriminant analysis (MVDA). MVDA identifies classification function(s) that maximize the dissimilarities between treatments, thus increasing the probability for accurate classification/prediction of unknown samples. The EN can be trained by a set of classification functions (canonical functions) to discriminate samples of unknown aroma quality characteristics.

PAGE 55

47 The objectives of this study were to 1) determine the relationship between greenharvested tomato physiological maturity and ethylene treatment exposure time required for ripening initiation, then develop a predictive model; and 2) assess the ability of an electronic nose sensor array to accurately discriminate between intact tomato fruit harvested at different ripeness stages, or subjected to postharvest treatments (low temperature storage and impact bruising) with no apparent visual symptoms. Materials and Methods Experiment 1. Green tomato harvests were conducted on nine different occasions during the 1996 and 1997 seasons at commercial farms in seven different locations throughout the state of Florida. (Table 1). Tomatoes were harvested green (stage 1, 0% red color, USD A, 1976) following commercial harvesting guidelines, then transported to Gainesville on the day of harvest. Fruit were sorted for defects and then placed inside sealed chambers where a humidified ethylene/air mixture (100 uL/L C2H4) was administered using a flow-through system. Prior to ethylene treatment, a random sample (30 to 100 tomatoes/harvest) was set aside to determine the physiological maturity based on locule tissue development (M1-M4) according to Kader and Morris (1976). Immaturegreen tomatoes (M1+M2) ranged from no gel formation in any locule and cut seeds upon slicing (Ml) to gel formation in one or two locules only (M2). Mature-green fruit ranged from gel formation in all locules with no red coloration (M3) to pink-red gel coloration (M4). After 1, 3 and 5 days of ethylene treatment, tomatoes which had attained breaker

PAGE 56

48 stage were removed from treatment and placed at 20°C and 95% RH for subsequent ripening until the table-ripe stage. In order to develop a regression model describing the relationship between harvest maturity and ethylene gassing exposure time, it was necessary to assume that immature-green fruit would require longer ethylene exposure to initiate ripening (Kader et al., 1978). Harvest maturity and ethylene exposure time distributions (% of population) were paired into three groups. Percent Ml (immature-green) was compared to % fruit exposed to 5 or more days of ethylene gassing, % M2 (partially mature-green) and M3 (mature-green) were combined and compared to % of fruit exposed to 3 days of ethylene gassing. Finally, % M4 (advanced mature-green) fruit was compared to % fruit exposed to 1-day ethylene treatment. The statistical relationship between the resulting comparisons was estimated with a third-order polynomial regression equation, which yielded the best fit to the experimental data (R 2 ) (STATISTTCA v4.5, Statsoft Corp., 1994). Experiment 2. For electronic nose (EN) analyses, 'Solimar' tomatoes (Asgrow Seed Co., Kalamazoo, MI) were harvested at green (stage 1), breaker (stage 2, <10% red color), turning (stage 3, 10-30% red color), light-red (stage 5, <90% red color), and red (stage 6, 100% red color) ripeness stages from a commercial field in Del Ray Beach, FL, on two separate occasions. Tomato fruit were transported the same day of harvest to Gainesville, FL, for sorting, postharvest treatments and electronic aroma sensing. A total of four separate experiments were conducted with the EN sensor array, each of them intended to explore the capability of electronic aroma sensing as a nondestructive means

PAGE 57

49 of detecting physiological maturity, ripeness stages, low temperature storage and internal bruising. Green-harvested tomato fruit (107.5 ± 13.5 g) were individually placed (n=36) inside the sampling vessel of the EN sensor array for electronic aroma sensing. Immediately after electronic aroma sensing, each green tomato was sliced equatorially to assess locule tissue development and the fruits were classified into two groups (immatureand mature-green) based on their locule tissue (M1-M4) maturity distribution (Kader and Morris, 1976). Visual maturity classification was contrasted to EN classification during the pattern recognition analysis. In a subsequent experiment, tomatoes (105.43 ± 14.87 g) harvested at four ripeness stages (breaker, turning, light red, red) were placed inside the sampling vessel of the electronic nose sensor array for electronic aroma sensing of individual fruit (n=24). The EN sensor array's capability to identify ripe tomatoes stored at low temperature prior to or during ripening (breaker or light-red stage) was explored utilizing two experiments. Greenharvested tomatoes were treated with ethylene (lOOuL/L) to initiate ripening, tomatoes exposed to 3 days of treatment were considered mature-green, while those exposed to more than 3 days of ethylene were considered immature-green at harvest. Following ethylene treatment, MG and IG tomatoes were placed at 5°C for 7 days. In addition, tomatoes harvested at light-red stage were also stored at 5°C for 7 days. Following the low temperature treatment, tomatoes were stored at 20°C until they reached the table-ripe stage, prior to electronic aroma sensing. A second group of tomatoes harvested from the same farm plots was allowed to ripen continuously at 20°C and used as controls. Fruits from both the low temperature (5°C) and room temperature

PAGE 58

50 (20°C) treatments were at table-ripe stage (as described in Chapter 3) when analyzed by the EN sensor array. Finally, breaker stage tomatoes (107.3 ± 3.9 g) were dropped twice, on opposite sides, from a 40 cm height on to a flat solid surface to initiate internal bruising symptoms (cloudy, viscous locule tissue) according to Moretti et al. (1997). A separate group of breaker stage fruit were not bruised and used as controls. Both bruised and non-bruised tomatoes were allowed to ripen for 10 days at 20°C (until the red stage) before electronic aroma sensing. Following the EN analysis, fruit were sliced equatorially to assess the presence of internal bruising symptoms where tomatoes had received impacts. During electronic aroma sensing, intact tomatoes were placed individually inside the sampling vessel of an e-NOSE 4000 electronic nose (Neotronics Scientific Inc., Flowery Branch, GA). The EN analysis consisted of a three-step operation controlled by a PC connected to the e-NOSE 4000. First, the sampling vessel was purged with compressed air for 2 min to eliminate any extraneous odors present in the vessel. Second, the sampling head, which contained twelve conducting polymer sensors, was purged with compressed air for 4 min to eliminate any volatile compounds bound to the polymer sensors. During the purging of the head, the headspace inside the sampling vessel equilibrated with the volatile compounds given off by the tomato fruit. Finally, the conducting polymer sensors were lowered into the sampling vessel for 4 min to expose them to volatiles present in the headspace over the sample. The total time for the electronic aroma sensing procedure for each fruit was 10 minutes. The change in electrical resistance output (i.e. sensor response) from the electronic nose was stored in a data base and later analyzed using multivariate

PAGE 59

51 discriminant analysis (MVDA) from STATISTIC A (v4.5, Statsoft Inc., 1994). This multivariate statistical procedure created linear discriminant functions (canonical functions) that maximized the differences between postharvest treatments and their controls, thus improving the probability for accurate classification based on En sensor outputs (cross-validation). Probabilities (posterior) for accurate classification based on EN outputs and Mahalanobis distances (distance between clusters or groupings, from a canonical analysis, adjusted for probability) were computed to compare the extent of differences across treatments. For graphical representation of treatment groupings based on canonical functions) scores, histograms and canonical plots were utilized. Results and Discussion Experiment 1. There was considerable variation in the distribution of tomatoes classified as immature-green (Ml), partially mature-green (M2), mature-green (M3) and advanced mature-green (M4) for the nine harvest dates and seven commercial tomato cultivars (Table 4.1). However, the distribution of maturity classes at harvest was highly correlated to the number of days ethylene treatment required to attain breaker stage (R 2 = 0.84) (Figure 4.1). In general, Ml fruit required longer ethylene gassing exposure (5days) whereas M4 fruit required the shortest treatment period (1-day). The strong relationship between tomato maturity at harvest (M1-M4) and ethylene treatment required to initiate ripening demonstrated that ethylene gassing could be utilized as a nondestructive tool to indirectly assess the physiological maturity of a population of green tomato fruit at harvest. The proportion of immature-green (Ml) fruit

PAGE 60

52 (17% to 53.7%) correlated to the proportion of fruit exposed to 5 or more days of ethylene treatment (6% to 68.6%). Chomchalow (1991) reported an average of 41% immature-green tomatoes during several commercial harvests; such a proportion is comparable to the 31.3% average found in this study. The onset of ethylene production occurs at the mature-green stage (M3-M4) (Brecht et al., 1991), therefore, exogenous ethylene treatment beyond physiological maturity might not further accelerate the onset of ripening (breaker stage). Table 4.1. Physiological maturity class at harvest (M1-M4) and response time to ethylene treatment required to attain breaker stage (in days) for seven commercial tomato cultivars harvested from different locations in the state of Florida. Harvest Tomato Growing Maturity at Harvest Distribution (%) Ethylene Gassing Exposure time (%) Date Cultivar Area M4 M3 M2 Ml 1day 3days 5days 4/25/96 Agriset-761 Naples 10.0 38.0 27.0 25.0 16.6 59.7 23.7 4/25/96 CPT-5 Naples 26.5 25.9 27.1 19.5 18.7 61.6 19.7 4/25/96 BHN-102 Naples 16.0 30.0 37.0 17.0 16.3 61.3 22.4 3/24/97 Solimar Del Ray Beach 4.2 36.5 11.4 47.9 4.8 26.6 68.6 5/15/97 Solar Set Bradenton 8.9 10.5 26.9 53.7 8.6 53.0 38.4 6/16/97 Agriset-761 Gainesville 12.0 32.0 36.0 20.0 16.4 61.7 21.9 7/30/97 Mountain Spring Quincy 6.9 17.2 34.5 41.4 23.9 70.1 6.0 9/26/97 BHN-189 Quincy 10.4 6.9 51.7 31.0 18.0 60.7 21.3 11/17/97 Solar Set Gainesville 6.6 16.7 50.0 26.7 3.6 64.3 32.1 Chomchalow (1991) reported that the proportion of immature-green tomatoes was inversely related to fruit size at harvest; however, no consistent trends relating fruit mass or diameter were found in this study (data not shown). This lack of dependable visual indicators of physiological maturity for green tomatoes has also been demonstrated by Kader et al. (1977). Advancements in automated sorting equipment permit the

PAGE 61

53 determination of physical properties such as mass, volume or density on a commercial scale. The availability of this regression equation relating physiological maturity and ethylene exposure time could be readily implemented by commercial tomato growers and handlers to minimize the proportion of immature-green tomatoes being harvested and reduce ethylene treatment costs by improving the accuracy of gassing schedules. Experiment 2. Electronic aroma sensing performed by the EN provided enough information for the MVDA pattern recognition procedure to find highly significant differences (P<0.00001) between intact matureand immature-green 'Solimar' tomatoes. MVDA was used to assess the EN sensor array's capability to classify green-harvested tomatoes according to their physiological maturity based on volatile production. 'Solimar' tomatoes were sliced immediately after EN analysis in order to visually classify them as immature(M1+M2) and mature-green (M3+M4) based on their locule tissue development. The relationship between volatile compound production and locule tissue development was described by a single linear function (canonical function) from the MVDA. The distribution of scores from this canonical function was represented using a histogram where immature-green fruit grouped distinctly from mature-green ones (Fig. 4.2). The Mahalanobis distance (MD) of 4.37 units helped contrast the degree of difference between physiological maturity groupings, as perceived by the EN, with those between groupings from other experiments. MVDA from EN sensor outputs was capable of classifying green 'Solimar' tomatoes into immature (M1+M2) and mature (M3+M4) with accuracy averaging 98.5% and 96.7%, respectively. (Table 4.2).

PAGE 62

54 The volatile compounds that contribute to the EN discriminating capabilities are not clearly understood from the sensor output data. Ethylene production has been reported to initiate at the mature-green stage (M3-M4) (Brecht, 1987), therefore the ability of EN sensors to detect ethylene would be a major factor to green tomato maturity discrimination. Production of numerous volatile compounds commences at the green stage, with lipid-derived aldehydes being prevalent (Baldwin at al., 1991c). In fact, lipidderived volatile production (hexanal, cw-3-hexenal, cw-3-hexenol and /ra>M-2-hexenal) has been reported to be closely associated with ethylene production (Baldwin et al, 1991c). The contribution of ketone compounds to green-tomato volatile profiles is less significant, however, production of 6-methyl-5-hepten-2-one increased considerably between green and breaker stages (Buttery, 1993). The MVDA from EN sensor outputs also identified very highly significant differences (P < 0.0033) between volatile compound production from tomatoes at different ripeness stages (breaker, turning, light red and red), then classified them into four distinct clusters. A two-dimensional canonical plot was utilized to visualize separation between the four groupings (Figure 4.3). The first two canonical functions from the MVDA were capable of explaining 95.4% of the differences in volatile compound production between ripeness stages. MVDA classified tomato ripeness stages based on EN sensor outputs with accuracy of 92.3% for breaker, 93.2% for turning, 98.2% for light-red, and 100% for red stage fruit. (Table 4.2). The MD between groupings demonstrated that the lowest degree of differences in volatile production occurred between breaker and turning stages (3.24 units), meanwhile, the greatest differences were between breaker and red stages (10.72 units).

PAGE 63

55 Table 4.2. MVDA classification accuracy based on EN sensor outputs for 'Solimar' tomatoes at different ripeness stages. Green-harvested tomatoes were divided into i mmature (M1+M2) and mature (M3+M4) based on their locule tissue development. Probability for correct classification z Maturity/Ripeness Number of Average Probability Probability Range stage Tomatoes Immature-green 22 98.5 80.3-100 Mature-green 14 96.7 57.9-100 Breaker 6 92.3 70.1-99.9 Turning 6 93.2 84.2-99.9 Light Red 6 98.2 89.6-100 Red 6 100 100-100 z Classification probabilities were based on the tomato samples analyzed. Multivariate discriminant analysis successfully classified 'Solimar' tomatoes from different ripeness stages based on volatile profiles perceived by the EN sensor array (cross-validation of visual classifications). Such results show the potential of EN technology for nondestructive quality screening of fresh market tomatoes. Although the production of CO2, ethylene and aroma volatiles of green tomatoes is significantly lower than that of red-ripe tomatoes (Baldwin et al., 1991a), the EN sensor array accurately discriminated immature from mature-green tomatoes. The high sensitivity to low volatile compound concentrations could be explained by the exponential reduction in electrical conductivity in the gas sensors as volatile compound concentrations increase (Benady et al, 1995). During tomato ripening, considerable changes occur in chemical composition and visual appearance. The production of aroma volatiles undergoes considerable quantitative and qualitative changes throughout tomato ripening. Alcohols (ethanol, methanol, cz's-3-hexenoi) have been reported to show slight increases in concentration, whereas aldehydes (cw-3-hexenal, trans-2-hexenal, and hexanal) and ketones (acetone, 1-

PAGE 64

56 penten-3-one, 6-methyl-5-hepten-2-one, and geranylacetone) nearly doubled their concentrations from the breaker through red stages (Baldwin et al., 1991a). Studies in aroma and flavor characterization of tomato fruit consider cultivars with high aroma volatile profiles as potentially having better aroma and flavor quality (Baldwin et al., 1991b). Tomato fruit show increased production of CO2 and ethylene at the M3-M4 stages and possibly upon bruising (MacLeod et al., 1976). Levels of ethylene produced by the fruit could also be affected by storage temperature (Brown et al., 1989). Ethylene production in combination with aroma volatile changes may have influenced the electronic nose sensor responses. The EN analysis identified significant differences between ripe 'Solimar' tomatoes stored at 5° for 7 days at latter ripeness stage (light-red) and controls stored at 20°C (PO.0014). The MD between storage temperature groupings (17.26 units) illustrated the large extent of the differences in volatile compound production found between the ripe tomatoes (Table 4.3). MVDA classification from EN sensor outputs had 100% accuracy when low temperature exposure occurred when the fruit were partiallyripe (Table 4.4). A histogram plot showed the distribution of canonical function scores grouped low temperature-stored tomatoes away from room temperature-stored tomatoes (Figure 4.4). Even though the EN nondestructive analysis successfully classified ripe tomatoes stored at low temperature during early ripeness stages (breaker) the extent of differences between treatments was lower compared to those exposed to low temperature at late ripeness stages (light-red) (Figures 4.4 and 4.5). MVDA showed greater separation between clusters as a result of storage temperature than the separation due to

PAGE 65

57 physiological maturity at harvest. The MD was 4.37 units between mature-green (MG) and immature-green (IG) fruit, meanwhile, MD's between 20°C and 5°C storage temperatures were 9.32 and 8.60 units for MG and IG fruit, respectively (Table 4.3). EN polymer sensors showed significant output changes following exposure to the headspace volatiles being produced by intact fruit from the different temperature treatments, notably IG-harvested fruit that were subsequently exposed to low temperature storage induced significantly higher outputs from all twelve EN sensors (data not shown). Classification accuracy based on EN sensor outputs ranged between 96.2% and 100% for tomatoes stored at low temperature during early ripeness stages. Table 4.3. Mahalanobis distances between classification groupings for 'Solimar' tomatoes based on MVDA from EN sensor outputs. The distances between groupings were highly significant for all treatments (P < 0.01) Between Grouping Centroids Mahalanobis Distance z From To Immature-green Mature-green 4.37 units Breaker Turning 3.25 units Breaker Light red 5.35 units Breaker Red 10.72 units Turning Light red 4.68 units Turning Red 9.27 units Light-red Red 9.59 units Light-red + 5°C Light-red + 20°C 17.26 units Mature-green + 5°C Mature-green + 20°C 9.32 units Immature-green + 5°C Immature-green + 20°C 8.60 units Internal-Bruised Non-bruised 9.15 units z Mahalanobis distances represent the separation between classification clusters from the MVDA adjusted for probability and were determined based on the tomato samples analyzed. Differences in EN sensor outputs in response to volatile compounds produced by tomatoes stored at low temperature (7 days at 5°C) could be due, in part, to reduced levels of important aroma volatile compounds. McDonald et al. (1996) found reduced concentrations of hexanal, 6-methyl-5-hepten-2-one, geranylacetone, methanol, 2-

PAGE 66

58 isobutylthiazole, and l-nitro-2-phenylethane in ripe tomatoes that were held at 2°C for 14 days at breaker stage. The sensitivity of tomato fruits to chilling injury has been shown to change during fruit ripening. Based on the expression of visual chilling injury symptoms, mature-green (Autio and Bramlage, 1986) and breaker-stage tomatoes (Hobson, 1987) were considered most sensitive to low temperatures. Under the current marketing system, green-harvested tomatoes are frequently exposed to temperatures below the chilling injury threshold (12°C) before the ripe stage (S.A. Sargent, personal communication). However, as evidenced by greater separation between EN classification groupings (higher MD), greater differences in volatile profiles occurred when tomatoes were exposed to low temperature at latter rather than earlier ripeness stages. This suggests that changes in volatile production induced by low temperature were partially reversed following subsequent ripening at 20°C. Furthermore, sensory panelists found greater flavor differences in tomatoes exposed to low temperatures at latter ripeness stages, than those exposed to low temperature at early ripeness stages (breaker) (Sargent etal., 1997). The volatile profile differences between bruised and non-bruised tomatoes, as perceived by the EN sensors, were also significant (PO.0136). Based on canonical function scores, internal-bruised tomatoes were segregated from non-bruised fruits by 9.15 MD units (Figure 4.6). Incidence of internal bruising in ripe tomatoes, depicted by cloudy viscous locule tissue, was corroborated following EN analysis on tomatoes that were dropped. MVDA classification of bruised based on EN sensor outputs was highly accurate (99.6-97.9%).

PAGE 67

59 Table 4.4. MVDA classification accuracy based on EN sensor outputs for 'Solimar' Probability for correct classification z Postharvest Treatment Number of Tomatoes Average Probability Probability Range Internal-Bruised 10 99.6 96.5-100 Non-Bruised 8 97.9 93.2-99.8 Light-red + 5°C 6 100 100-100 Light-red + 20°C 6 100 100-100 Mature-green + 5°C 6 97.8 90.6-99.9 Mature-green + 20°C 6 99.9 99.9-100 Immature-green + 5°C 6 99.9 99.9-100 Immature-green + 20°C 6 99.7 99.4-99.9 Classification probabilities were based on the tomato samples analyzed. The EN results concur with the findings of recent sensory and metabolite analysis studies. The incidence of internal bruising, a physiological disorder without apparent external visual symptoms, has been shown to stimulate increased ethylene and C0 2 evolution, and to decrease titratable acidity (MacLeod 1976). Significant reductions in ascorbic acid and total carotenoid pigments in bruised locule tissue were contrasted by increased polygalacturonase activity and electrolyte efflux in bruised pericarp tissue (Moretti et al., 1998). In a recent study, sensory panelists were able to successfully distinguish between internal-bruised and non-bruised tomatoes, confirming significant changes in the aroma profiles and chemical composition from the locular gel and pericarp tissues as a consequence of internal bruising (Sargent et al., 1997). Possible changes in C0 2 , ethylene production and aroma profiles induced in internal-bruised tomatoes could explain the differences detected during electronic nose and sensory analyses.

PAGE 68

60 Conclusions The polynomial regression equation developed in this study relates harvest maturity and ripening response time under ethylene treatment (Figure 1). It could be implemented by commercial tomato growers to minimize the proportions immature-green fruit harvested and by packer/shippers to optimize harvesting and ethylene gassing schedules. Results from these tests showed the potential use of electronic volatile sensing technology in non-destructive quality screening of fresh market tomatoes. The 400+ aroma compounds found in tomatoes are derived from lipids, amino acids, carotenoid pigments, and lignin-related compounds. The great potential for non-destructive quality screening of tomatoes based on volatile sensing lies in the diversity of biosynthetic pathways contributing to the formation of C0 2 , ethylene and aroma volatile compounds. Tomato volatiles could perhaps tell the story of tomato abuse when no visual symptoms are available and, therefore, help fresh tomatoes reach consumers with consistently higher flavor quality.

PAGE 69

61 Figure 4.1. Polynomial regression equation describing the relationship between physiological maturity class distribution (%) at harvest (M1-M4) and the distribution of fruit (%) based on the length of ethylene treatment exposure time to attain breaker stage by green-harvested tomatoes from seven different commercial cultivars.

PAGE 70

62 0 2 Canonical Score Distribution Figure 4.2. Comparison between immature-green and mature-green tomato fruit based on electronic aroma sensing. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of immatureand mature-green tomato fruit. Classification of immatureand mature-green tomatoes was very highly significant (P < 0.00001).

PAGE 71

63 Figure 4.3. Comparison between breaker, turning, light-red and red ripeness stages for tomatoes based on electronic aroma sensing. Canonical plot represents the classification of canonical scores obtained after EN sensor outputs analyzed using the first 2 canonical functions. The ellipses around ripeness stage groupings represent the 95% confidence intervals. Classification of tomato ripeness stages was highly significant (P < 0.0033).

PAGE 72

64 u o o f— I i Low Temperature (5C) Expected Low Temp. Room Temp. (20C) Expected Room Temp. -4 0 Canonical Score Distribution Figure 4.4. Comparison between low temperature-stored (5°C) and room temperaturestored (20°C) light-red 'Solimar' tomato fruit based on electronic aroma sensing. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of low temperatureand room temperature-stored tomato fruit. Classification of low temperatureand room temperature-stored tomatoes was very highly significant (P < 0.0014).

PAGE 73

65 Figure 4.5. Comparison between ripe 'Solimar' tomatoes exposed to low temperature (5°C for 7 days) at early stages of ripening (breaker). Based on their ethylene treatment exposure time, tomatoes were divided into immature-green (IG) or mature-green (MG) at harvest. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of tomatoes stored at 5°C or 20°C. Classification of low temperature-stored tomatoes was significant (P< 0.0136).

PAGE 74

66 1 I if Internal-bruised Expected Int-bruised I I Non-bruised Expected Non-bruised I I -4 0 Canonical Score Distribution Figure 4.6. Comparison between internal-bruised and non-bruised tomatoes based on electronic aroma sensing. Histogram represents the frequency of canonical scores obtained after output from EN sensor array was analyzed using the first canonical function from the MVDA on EN sensor outputs. Line plots represent the expected scores for normal populations of immatureand mature-green tomato fruit. Classification of internal-bruised and non-bruised tomatoes was very highly significant (P < 0.00001).

PAGE 75

CHAPTER 5 AROMA VOLATILE PROFILES AND RIPE TOMATO FLAVOR ARE INFLUENCED BY PHYSIOLOGICAL MATURITY AT HARVEST Introduction Consumer dissatisfaction with fresh tomato flavor promoted extensive research during the late 1970's and early 1980's. Tomatoes harvested at early stages of maturity were perceived as less sweet, more sour, and as having less tomato-like flavor compared to those harvested more mature (Kader et al., 1977). Furthermore, vine-ripe fruits were considered sweeter by sensory panels without significant differences in soluble solids content or dry matter (Watada and Aulenbach, 1979). This observation highlighted the possible relevance of aroma compounds on tomato flavor perception. Physiological maturity of green-harvested tomatoes has been circumstantially related to ripe, fresh tomato flavor and quality by numerous researchers (Kader et al., 1977; Watada and Aulenbach, 1979; Al-Shaibani and Greig, 1979). Nonetheless, chemical composition data collected by these authors did not always concur with sensory panel results. Over 85% of Florida's fresh market tomatoes are harvested commercially at the green stage and are exposed to exogenous ethylene treatment to accelerate the onset of ripening processes. Difficulty in assessing harvest maturity leads to 25-40% of tomatoes being harvested at immature-green stage (Chomchalow, 1991), leading to compromised 67

PAGE 76

68 quality when ripe. Without an accurate and dependable non-destructive method of screening inferior quality immature fruits in green-harvest tomato operations, growers cannot provide the consistent quality that the markets demand. The development of locule tissue has been proposed as an accurate way to separate immature-green (M1+M2) from mature-green tomato fruits (M3+M4) (Kader and Morris, 1976). Data collected in our laboratory over the past two years showed that there is a strong direct relationship between immature-green fruit (M1-M2) and extended exposure to ethylene gas required to initiate ripening. The objectives for these experiments were: 1) to document changes in important volatile compounds from different fruit tissues as affected by fruit maturity at harvest; 2) to explore the potential use of electronic nose volatile sensing technology (EN) as a nondestructive tool to screen ripe tomato fruits harvested at immatureand mature-green stages; and 3) to identify and quantify tomato flavor and aroma differences utilizing difference discrimination tests and descriptive sensory panels. Materials and Methods In a series of experiments aimed at determining the effects of physiological maturity of green-harvested tomatoes on their quality at table-ripe stage, information was gathered from several commercial cultivars grown in different parts of the state of Florida. 'Agriset-761' (Agrisales Inc., Plant City, FL) and 'CPT-5' (Collier Farms, Naples, FL) tomatoes were harvested from experimental plots in Naples, FL; 'Solimar' (Asgrow Seed Co., Kalamazoo, MI) tomatoes harvested on two separate occasions from commercial plots in Palm Beach county, FL; and 'BHN-189' (BUN Research Inc.,

PAGE 77

69 Naples, FL) tomatoes harvested from a commercial farm in Quincy, FL. Following all harvests, tomatoes were transported to Gainesville, FL for postharvest treatments and subsequent ripening. Upon arrival, green tomatoes were sorted for defects, washed with chlorinated water, and towel dried. Prior to ethylene treatment, a random sample of fruits (n=50-100) were sliced equatorially to assess the physiological maturity distribution of the green tomato population. The physiological maturity stages were rated based on locule tissue development (M1-M4) following the internal maturity standards for tomatoes proposed by Kader and Morris (1976). Tomatoes were stored at 20°C and gassed in bulk with a humidified, 100 uL/L ethylene/air mixture in sealed chambers connected to a flow-through system. The flow rate calculation was based on maturegreen tomato respiratory rates for 20°C (mg C0 2 /kg fresh weight/hour) (Hardenburg et al, 1986) and total fruit mass (kg) inside each chamber. After 1, 3 and 5 days of ethylene treatment, tomatoes that attained breaker stage were removed from the gassing chamber and placed at 20°C and 95% RH for subsequent ripening. In order to compare tomatoes harvested at latter ripeness stages with those harvested at green stage using descriptive sensory panels, 'BHN-189' and 'Solimar' tomatoes were harvested at light-red stage (stage 5) from the same plots where green fruit were harvested approximately 8 days before. Tomatoes harvested at light-red stage were subsequently ripened along with the green-harvested tomatoes. Tomatoes were considered table-ripe upon attaining red stage (stage 6, >90% red coloration) and noticeable firmness loss (3-4 mm deformation threshold) measured by applying a constant 9.8 N force for 5 seconds on the fruit equator (Gull et al., 1980). Table-ripe tomatoes were removed from storage and utilized for compositional, aroma

PAGE 78

70 volatile, electronic nose and sensory analyses. In addition to the number of days required to attain table-ripe stage following ethylene treatment, physical traits characterizing tomatoes with different ethylene gassing exposure times were documented by recording fruit mass (g), fruit equatorial diameter (mm) and locule tissue content (% fresh weight) for 'Agriset-761' and 'CPT-5' tomatoes. During the first harvest of 'Solimar' tomatoes, the compositional differences between individual fruit tissues (locule, pericarp and whole fruit) were documented by excising the tissues and conducting compositional and GC analyses (Experiment 1). Samples from table-ripe tomatoes that required different ethylene treatments (1, 3, and 5 days) and tomatoes harvested at light-red stage ('BHN-189' and 'Solimar') were collected for compositional analyses. Six composite samples (3 fruit/sample) from each treatment and fruit tissue were homogenized and centrifuged at 18,000 X g„ and 5°C temperature. For vitamin C assays, homogenates (2 g/sample) were stabilized using 20mL of acid mixture (6% HP0 3 containing 2N acetic acid in water) prior to centrifugation. The supernatant was filtered using cheesecloth, stored inside scintillation vials and frozen at -20°C for later analysis. Titratable acidity, expressed as % citric acid, was determined by titrating 4 g of tomato supernatant to 8.2 pH with a 0.1 N NaOH solution using an automatic titrimeter (Fisher Scientific, Pittsburgh, PA). Soluble solids content, expressed as °Brix, was measured using a tabletop digital refractometer (Abbe Mark II, ReichartJung, Buffalo, NY) and pH measurements were conducted using a digital pH-meter (Corning model 140). Tomato fruit lycopene content was determined using a colorimetric method adapted from Umiel and Gabelman (1971). Four 10-g samples of tomato homogenate

PAGE 79

71 were mixed with 30 mL of acetone inside 100-mL glass vials covered with aluminum foil to minimize light-induced lycopene breakdown. After mixing for 60 sec using a Polytron blender, the samples were vacuum filtered (Whatman # 4 filter paper) into 500 mL sidearm Erlenmeyer flasks containing 45 mL of hexane. The hexane-lycopene phase was separated from the acetone through a series of deionized water washes using separatory funnels. Lycopene absorbance was read at 503 nm using a spectrophotometer (Beckman DU-20, Irvine, CA). Lycopene concentrations were determined using a standard curve derived from pure lycopene standards. Vitamin C and total soluble sugars were analyzed using spectrophotometric methods adapted from Terada et al. (1978) and Dubois et al. (1956), respectively, as described in Chapter 3. Individual sugar analysis (glucose and fructose) was performed for 'BHN-189' and 'Solimar' samples analyzed during the descriptive sensory panels using an adaptation of the high performance liquid chromatography (HPLC) method described by Baldwin et al (1991c). Approximately 20 g of tomato homogenate were extracted using 35 mL of 80% ethanol/deionized water solution. The homogenate/ethanol mixture was boiled for 15 min, then cooled prior to filtration (Whatman # 4 filter paper). The filtered solution was brought up to a 50-mL volume with 80% ethanol inside a volumetric flask. Ten mL of the filtered solution were then passed through a C-18 Sep-pak (Waters/Millipore, Milford, MA) then filtered through a 0.45 um Millipore filter. Sugars were analyzed using HPLC refractive index detector with a Waters Sugar Pak column and a 10^* M ethylenediaminetetraacetic acid disodium calcium salt (CaEDTA) mobile phase (0.5 mL min" 1 flow rate at 90°C). Glucose and fructose concentrations were converted to sucrose equivalents (Koehler and Kays, 1991),

PAGE 80

72 where values were multiplied by 0.74 and 1.73, respectively to better represent individual hexose sweetness perception potential. Tomato volatile compounds were identified and quantified using a gas chromatography (GC) headspace analysis technique (Baldwin et al., 1991b). Four samples of homogenate (40 mL) from each tomato treatment were combined with 10-mL of saturated CaCk solution, blended for 10 seconds, immediately frozen using liquid nitrogen and stored at -80°C. The saturated CaCl 2 solution was added to the tomato samples to help reduce enzymatic changes that might induce quantitative and qualitative changes in the tomato sample's volatile profile following tissue maceration (Buttery and Ling, 1993). For GC analysis, each tomato sample was thawed under nmning tap water and a 2-mL sample placed inside a 6-mL vial sealed with a crimp-top and Teflon/silicone septum. Vials were heated rapidly to 80°C and incubated for 15 minutes before injection to a Perkin Elmer HS-6 headspace sampler heating block. The analysis was carried out using a Perkin Elmer Model 8500 gas chromatograph equipped with a 0.53 mm X 30 m polar stabilwax capillary column (1.0-um film thickness, Restek Corp., Bellefonte, Pa.) and a flame ionization detector. Column oven temperature was held at 40°C for 6 min, then raised to 180°C at a rate of 6°C/min. The resulting GC peak heights for 16 important aroma volatile compounds (acetaldehyde, acetone, methanol, ethanol, l-penten-3-one, hexanal, cw-3-hexenal, 2+3-methylbutanol, /rarts-2-hexenal, /ra«s-2-heptenal, 6-methyl5-hepten-2-one, c/s-3-hexenol, l-nitro-2-phenylethane, geranylacetone, 2isobutylthiazole, and P-ionone) as reported in previous research (McDonald et al., 1996; Baldwin et al., 199 lab, Buttery et al, 1988; Petro-Turza, 1987) were quantified in uL/L

PAGE 81

73 utilizing standard curves determined by enrichment of bland tomato homogenate with authentic volatile compound standards (Baldwin et al., 1991abc). In addition to GC and compositional analyses, an electronic nose (EN) sensor array was utilized to discriminate treatments based on their volatile production. The EN consisted of a sampling head equipped with 12 polymer sensors, glass sampling vessel and purging valves (e-NOSE-4000, Neotronics Scientific Inc., Flowery Branch, GA). Each individual polymer sensor changed its electrical conductivity upon exposure to volatile compounds present inside the headspace of the sampling vessel. A computer recorded the sensor outputs over time. EN analysis consisted of placing individual 20 mL samples of fruit homogenate inside the sampling vessel (n=6 homogenate samples/treatment) and sealing it against the sensor head. Approximately 20 g from each frozen tomato sample was placed inside 113mL plastic cups, lidded and thawed in a 25 °C water bath. Immediately upon thawing, the lid was removed and the sample cup placed inside the glass vessel of an electronic nose (e-NOSE 4000, Neotronics Scientific, Flowery Branch, GA). EN analysis and sensor data acquisition was controlled by personal computer. EN sampling began with a 2-min purge of the glass vessel using compressed air. Next, the sampling head, containing twelve polymer sensors (manufacturer ID numbers: T301, T298, T297, T283, T278, T264, T263, T262, T261, T260, T259, and T258), was purged with compressed air for 4 min to eliminate any volatile compounds in contact with the polymer sensors, while volatile compounds from the tomato sample equilibrated inside the sealed sampling vessel. Both sampling vessel and sensor head were purged using compressed air at a 400 mL/min flow rate. Finally, the sensor head was lowered automatically into the sampling

PAGE 82

74 vessel to expose the polymer sensors to the volatile compounds produced by the sample in the headspace for an additional 4 min. EN analysis was carried out at room temperature (ca. 25°C) and relative humidity inside the sampling vessel was also recorded. The EN analysis lasted a total of 10 min per sample. For preliminary sensory analyses, a "difference from control" test was chosen because of its ability to identify overall differences between samples (when compared against a control), while allowing panelists to rate the extent of those differences by including a hidden control sample within the treatments. Samples from ripe 'Agriset-761' and 'CPT-5' tomatoes (Experiment 2) were presented to a group of 27 untrained panelists on two separate panel sessions, one session for each cultivar. Samples from tomatoes exposed to 1-day ethylene gassing to attain breaker stage were chosen as controls, and compared to tomatoes exposed to 3 and 5 days ethylene treatment. Panelists were asked to rate the degree of difference they perceived between the control sample and three other samples (1-day "hidden control", 3 and 5 days ethylene exposure time). A 12-point scale with verbal descriptors on either end was used to rate the extent of difference in the sensory test ballots (1 = no difference and 12 = extremely different). All samples were presented in plastic cups labeled with random numbers and presented in random order to panelists to avoid biased results. Following the sensory test sessions, the results from the ballots were computed, and in both cases, some degree of panelist screening was required when panelists rated the hidden control (1-day ethylene) extremely different than the control sample (same treatment). Based on the results from the "difference from control" sensory tests, a trained descriptive sensory panel was assembled. To organize and train a descriptive sensory

PAGE 83

75 panel, a group of 20 volunteers showing no dislikes for tomatoes were screened for proper sensory perception with the use of citric acid and sucrose solutions of varying concentrations. Over a period of three months, the group was reduced to 16 panelists, 10 males and 6 females comprising ages between 20 and 65 years old, and were trained to describe flavor and aroma attributes from fresh market tomatoes. Initially panelists were screened for proper sensory perception using sucrose, citric acid and sucrose/citric acid solutions of varying concentrations. During the initial training sessions, panelists were presented with a variety of tomato samples representing effects of ripeness stage, storage temperature and cultivar on characteristic tomato flavor. After having familiarized panelists with a wide range of tomato samples during the first five training sessions, the panel leader compiled a descriptor list from published literature on tomato flavor to aid panelists in verbalizing flavor and aroma characters perceived in the samples. The panel reached a consensus on five flavor attributes (typical tomato, sweetness, sourness, green/grassy and off-flavor) and two aroma attributes (ripe tomato and off-odor). Descriptor intensity was rated using a 150-mm, unstructured line scale with a low intensity on the zero (left) side and high intensity on the 150 (right) side as anchor terms. Approximately 20 minutes before sensory analysis, 'BHN-189' and 'Solimar' whole tomato samples (Experiment 3) were chopped into a coarse puree using 8 to 10 pulses of a food processor (M. Einstein, personal communication). Two tablespoons (ca 40-50 g) of tomato puree were placed in 113-mL plastic cups, sealed with lids and labeled with a two-digit random number. Evaluations were conducted in individual booths with dim lighting and samples were presented in random order. Panelists were

PAGE 84

76 instructed to open the lid from each tomato sample cup to rate the aroma descriptors, then to proceed with the flavor descriptors. Water and unsalted crackers were provided for panelists to rinse their palates between samples. In any given session, panelists were asked to rate the flavor attributes of 4 to 6 tomato samples. During sensory analysis, four 40mL samples of tomato homogenate for GC and EN analyses were combined with 10mL of a saturated CaCfe solution, blended at high speed for 10 seconds and flash-frozen with liquid nitrogen. CaCb was added to reduce enzymatic activity that could contribute to further volatile changes following tissue maceration and subsequent storage at -80°C (Buttery and Ling, 1993). Descriptive sensory panel scores for tomato flavor and aroma descriptors were analyzed as complete block design with panelists as blocks and maturity at harvest as treatments using GLM procedure of SAS (v6.12, SAS Institute, Cary, NC). All compositional, HPLC, and GC data were analyzed using multiple analysis of variance (MANOVA), Duncan's Multiple Range Test for means separation using SAS. Meanwhile, EN sensor outputs were analyzed using multivariate discriminant analysis (MVDA) with STATISTICA ( v4.5, Statsoft, Inc., Tulsa, OK). The differences between volatile profiles found in ripe tomato samples were identified and visualized through MVDA, which created two-dimensional canonical plots where descriptive linear functions (canonical functions) classified the different tomato samples based on the pattern of outputs from the polymer sensors In certain cases a forward stepwise procedure was utilized to optimize the number of variables considered for the descriptive linear functions, thus simplifying data collection and improving statistical significance. Relationships between instrumental and sensory parameters were explored through the

PAGE 85

77 use of correlation matrices (PROC CORR) and stepwise regressions (PROC STEPWISE) using SAS. Results and Discussion 'Solimar' (Experiment 1). The effects of maturity stage at harvest on tomato fruit tissue (locule and pericarp) aroma volatile profiles and chemical composition were documented in table-ripe 'Solimar' fruit. Even though no significant differences in pH were found for either locule or pericarp tissues, increasing pH values with increasing ethylene exposure time were documented for both tissues. Increasing pH values became significant in whole tomato samples exposed to 5 days ethylene when compared to those exposed to 1 day (4.24 and 4.20, respectively) (Table 5.1). No relationship between pH values and titratable acidity was evident, as pH values increased, titratable acidity either decreased (locule) or increased (whole) with increasing ethylene exposure time. Soluble solids were significantly higher in pericarp and whole tomato samples exposed to 5 days ethylene treatment (4.28° and 4.28°Brix, respectively) compared to those exposed to 1 and 3 days (whole fruit) ethylene treatment (3.95°, 4.10° and 4.12° Brix for pericarp and whole fruit samples respectively). No significant differences in vitamin C content were found between ethylene exposure times for any of the tomato tissues. However, reduced vitamin C content with increasing ethylene exposure time as observed in locule and whole fruit samples (Table 5.1).

PAGE 86

78 Table 5.1. Table-ripe chemical composition parameters for green-harvested 'Solimar' Compositional Parameters z Ethylene Gassing Exposure time 1-Day 3-Days 5-Days Whole Fruit PH 4.20 b 4.21 ab 4.24 a Soluble solids content (°Brix) 4.10 b 4.12 b 4.28 a Titratable acidity (% citric acid) 0.78 a 0.84 a 0.90 a Vitamin C (mg/lOOg fresh wt.) 13.8 a 13.6 a 12.1 a Locule Tissue PH 4.41 a 4.38 a 4.44 a Soluble solids content (°Brix) 3.90 a 3.72 a 3.98 a Titratable acidity (% Citric acid) 1.00 a 0.97 a 0.94 a Vitamin C (mg/lOOg fresh wt.) 24.6 a 18.6 a 16.8 a Pericarp Tissue PH 4.17a 4.18a 4.19a Soluble solids content (°Brix) 3.95 b 4.12 ab 4.28 a Titratable acidity (% Citric acid) 0.72 a 0.74 a 0.71 a Vitamin C (mg/lOOg fresh wt.) 12.7 a 18.6 a 16.8 a z Means for parameters with different level according to Duncan's Multiple letters within rows were significantly different at the 5% Range Test. (n=6 composite samples/trt). Aroma volatile compound concentrations were generally lower in table-ripe 'Solimar' tomatoes that required extended ethylene treatments to attain breaker stage. Five of 16 aroma volatile compounds showed significant differences between treatments. Table-ripe tomatoes that required 3 or 5 days ethylene treatment had significantly lower levels of l-penten-3-one, c/'s-3-hexenal, 6-methyl-5-hepten-2-one, 2-isobutythiazole and geranylacetone. In addition, tomatoes exposed to 5 days of ethylene treatment had significantly lower levels of c/s-3-hexenal and geranylacetone compared to those exposed to 3 days of treatment (Table 5.2). Differences in aroma volatile compound concentrations for pericarp tissue samples from 'Solimar' tomatoes followed a similar trend to those observed for whole fruit homogenate samples. Pericarp samples from tomatoes exposed to 5 days of ethylene treatment had significantly lower concentrations of l-penten-3-one, trans-2-

PAGE 87

79 heptenal, 6-methyl-5-hepten-2-one, 2-isobutylthiazole, geranylacetone and P-ionone compared to those exposed to 1 or 3 days of ethylene treatment (Table 5.3). Pericarp tissue samples from tomatoes exposed to 1 day of ethylene treatment had significantly higher concentrations of c/s-3-hexenal compared to pericarp samples from tomatoes exposed to 3 or 5 days of ethylene treatment. Concentrations of 2+3 methylbutanol were highest in pericarp tissue samples from tomatoes exposed to 3 days of ethylene treatment (Table 5.3). Table 5.2. Aroma volatile compound concentrations for whole table-ripe 'Solimar' t omatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage. Aroma Volatile Compounds z Ethylene gassing exposure time 1 day 3 days 5 days Acetaldehyde 12.64 a 13.54 a 12.04 a Acetone 0.45 a 0.48 a 0.41 a Methanol 310.82 a 304.13 a 307.41 a Ethanol 16.80 a 15.98 a 16.67 a l-Penten-3-one 0.27 a 0.19 b 0.17 b Hexanal 12.33 a 8.22 a 9.19a C/s-3-hexenal 9.04 a 6.18b 4.56 c 2+3-Methylbutanol 2.16 a 2.26 a 1.79 a 7ra/7s-2-hexenal 8.59 a 7.51 a 7.03 a 7>ans-2-heptenal 0.04 a 0.03 a 0.03 a 6-Methyl-5-hepten-2-one 0.83 a 0.55 b 0.42 b C/s-3-hexenol 0.05 a 0.04 a 0.05 a 2-Isobutylthiazole 0.09 a 0.06 b 0.05 b 1 -Nitro-2-phenylethane 0.06 a 0.06 a 0.05 a Geranylacetone 6.90 a 4.26 b 2.25 c P-ionone 0.12a 0.09 a 0.06 a TotaF 381.18 363.58 362.65 z Means for aroma volatile compounds (uL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile concentrations (uL/ L) based on the sum of the 1 6 compounds quantified. Aroma volatile compound concentrations in table-ripe 'Solimar' locule tissue samples showed significant differences in 5 of 16 compounds quantified. Locule tissue samples from tomatoes exposed to 1 day of ethylene treatment had significantly higher

PAGE 88

80 levels of l-penten-3-one, 2-isobutylthiazole, l-nitro-2-phenylethane, and p-ionone. Conversely, c/.y-3-hexenol levels were significantly higher for locule tissue samples from tomatoes exposed to 5 days of ethylene treatment (Table 5.4). Table 5.3. Aroma volatile compound concentrations for pericarp tissue (including columnella) from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene t reatment to attain breaker stage. Aroma Volatile Compounds 7 Ethylene gassing exposure time 1 day 3 days 5 days Acetaldehyde 1 H.J J a io.z / a 1 1 07 o 1 1.7/ a Acetone 0.45 a 0.47 a 0.38 a Methanol 340.43 a 379.90 a 375.78 a Ethanol 22.84 a 18.06 a 17.99 a l-Penten-3-one 0.21 a 0.18a 0.14b Hexanal 19.43 a 14.09 a 14.90 a C/s-3-hexenal 10.96 a 7.57 b 6.45 b 2+3-Methylbutanol 2.60 b 3.31 a 2.18b 7>a«.s-2-hexenal 9.78 a 8.66 a 6.98 a 7raws-2-heptanal 0.05 a 0.04 a 0.03 b C/s-3-hexenol 0.10 a 0.09 a 0.07 a 6-Methyl-5-hepten-2-one 0.95 a 0.76 a 0.50 b 2-Isobutylthiazole 0.11 a 0.09 a 0.05 b 1 -Nitro-2-phenylethane 0.06 a 0.06 a 0.06 a Geranylacetone 7.55 a 5.92 a 3.66 b P-ionone 0.13 a 0.11 a 0.05 b Total y 430.64 455.56 441.19 z Means for aroma volatile compounds (uL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. In general, table-ripe 'Solimar' tomatoes harvested at immature-green stage (exposed to >3 days of ethylene treatment) had significantly lower production of the aroma volatile compounds quantified. Pericarp tissue samples produced approximately a 20% higher concentration of aroma volatile compounds compared to whole fruit samples (443 and 369 uL/L, respectively). Buttery et al. (1988) found similar results and suggested that higher volatile production found in pericarp tissue could be due in part to

PAGE 89

81 tissue damage during separation. Tissue wounding would induce lipoxygenase activity, an important enzyme in the biosynthetic pathway of aldehydes and alcohols from membrane fatty acids (Hilderbrand, 1989). On the other hand, aroma volatile production in locule tissue samples was 43% lower than whole tomato samples (203.30 uL/L and 369 uL/L, respectively). In fresh market tomatoes, locule tissue constitutes 14.4% to 34% of the fresh weight (Stevens et al, 1977), thus showing a lesser contribution of locule tissue to whole fruit aroma volatile production compared to pericarp tissue. Table 5.4. Aroma volatile compound concentrations for locule tissue from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage. Aroma Volatile Compounds z Ethy ene gassing exposure time 1 day 3 days 5 days Acetaldehvde 19.95 a 17.43 a 16.45 a Acetone 0.49 a 0.48 a 0.47 a Methanol 146.63 a 144.74 a 136.33 a Ethan ol 18.84 a 19.33 a 18.78a 1 -Penten-3-one 0.18a 0.12b 0.14 b Hexanal 4.10a 3.02 a 4.78 a Cw-3-hexenal 6.53 a 4.89 a 5,59 a 2+3-Methylbutanol 2.38 a 2.66 a 2.95 a 7>aws-2-hexenal 6.83 a 5.99 a 7.55 a 7raHs-2-heptenal 0.03 a 0.03 a 0.03 a 6-Methyl-5-hepten-2-one 0.64 a 0.52 a 0.56 a C7s-3-hexenol 0.07 b 0.09 ab 0.14a 2-Isobutylthiazole 0.11 a 0.63 b 0.07 b 1 -Nitro-2-phenylethane 0.07 a 0.06 b 0.06 b Geranylacetone 2.98 a 2.69 a 2.72 a B-ionone 0.22 a 0.17 ab 0.08 b Total compound cone. y 210.05 202.28 196.68 z Means for aroma volatile compounds (uL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile concentrations (u,L/ L) based on the sum of the 16 compounds quantified. Significant reductions in the concentration of P-ionone, cw-3-hexenal, and 1penten-3-one found in table-ripe tomatoes exposed to extended ethylene treatment

PAGE 90

82 deserve special attention because of their low odor thresholds (Buttery et al., 1988). The odor unit values (OUV) represent a ratio between the concentration of an aroma compound present in a sample and the threshold concentration required for its sensory detection. A high OUV for a volatile compound would imply that small changes in its concentration would have greater impact during sensory evaluation, than similar magnitude changes in a volatile compound with low OUV. The levels of p-ionone, a volatile compound with one of the highest OUV among tomato compounds, dropped by as much as 70% in locule tissue samples which required extended ethylene treatment (5 days). In addition, geranylacetone levels dropped at least 50% in pericarp and whole tomato samples exposed to 1 day of ethylene treatment compared to those exposed to 5 days. Even though the levels of hexanal respectively dropped 33.4% and 27.5% in respective whole fruit and pericarp samples exposed to 3 days of ethylene gassing, those differences were not significant compared to 1 day treatment samples. It is important to note that methanol and hexanal concentrations were approximately 100% and 300% higher in the pericarp homogenate than in the locule tissue homogenate. Higher concentrations of hexanal present in the pericarp tissue could suggest its higher contribution to the perception of green/grassy flavor (Petro-Turza, 1987; Tandon, 1998) in whole tomato fruit. Cw-3-hexenol levels decreased significantly in pericarp samples that required extended ethylene treatment. Conversely, c/'s-3-hexenol concentrations increased with increasing ethylene exposure time in locule tissue samples. The possibility of opposite trends in the production of volatile compounds from different tomato tissues could explain lesser extent of volatile concentration difference found in whole tomato

PAGE 91

83 samples compared to pericarp samples. Three important aroma volatile compounds showed significant reductions in all three tomato tissues analyzed, l-penten-3-one, pionone, and 2-isobutylthiazole, the latter deserving special attention since it is a compound unique to tomato aroma (Petro-Turza, 1987). Reported aroma volatile concentrations from different tomato tissues (Buttery et al., 1988) do not entirely concur with our present results. Contrary to previous reports on the contribution of excised tissues to whole fruit volatile profiles, the concentrations of 6methyl-5-hepten-2-one and geranylacetone were lower in locule tissue samples from tomatoes exposed to 1 day of ethylene treatment when compared to pericarp tissue samples. However, locule tissue samples that required extended ethylene treatment produced slightly higher levels of both compounds than pericarp samples. The extent of these discrepancies might be related to cultivar variability as was shown by Baldwin et al. (1991b) and by maturity at harvest as suggested in this study. Nevertheless, the absolute levels of P-ionone, geranylacetone, 2+3 methylbutanol and 6-methyl-5-hepten-2-one were considerably higher in our results compared to those reported by Buttery et al. (1988). 'Agriset-76r and 'CPT-5' rExperiment 2\ Table-ripe 'Agriset-761' and 'CPT-5' tomatoes were utilized to document compositional and aroma volatile concentration changes, as related to ethylene exposure time (physiological maturity), and their possible effects on sensory quality. 'Agriset-761' tomatoes that required 1 day of ethylene treatment had significantly lower mass than those exposed to 3 or 5 days (195.5, 216.8 and 230.6 g, respectively). On the other hand, 'CPT-5' fruit exposed to 3 days ethylene

PAGE 92

84 gassing had significantly greater mass when compared to either 1 day or 5 days fruit (242.7, 195.2 and 204.9 g, respectively). Locule tissue content (% fresh weight) showed no significant differences between treatments for either cultivar, however, 'Agriset-761' tomatoes had slightly lower locule tissue contents when compared to 'CPT-5' fruits (19.8% and 21.1%, respectively) (data not shown). Both cultivars required approximately 10.5 days from breaker (following ethylene treatment) to table-ripe stage. During "difference from control" sensory tests, untrained panelists detected significant flavor differences between table-ripe tomato samples with different ethylene exposure time from both cultivars (Table 5.5). Panelists found lesser differences between samples exposed to 1 day of ethylene treatment (control and hidden controls), meanwhile the flavor differences between the control and 3 or 5 day ethylene treatment samples were significantly greater when compared to 1-day tomatoes. Table 5.5. "Difference from control" sensory tests conducted on table-ripe 'Agriset-761' and 'CPT-5' tomatoes. Flavor ratings were based on a 12-point scale with verbal descriptors ranging from 1 (no different) to 12 (extremely different) from control. Tomato cultivars Ethylene Gassing Exposure time Pr>F z 1-Day 3 -Days 5-Days 'Agriset761' 0.0008 3.85 a 6.65 b 6.80 b 'CPT-5' 0.0001 3.00 a 5.90 b 7.35 b z Means for tomato fruit sensory scores with different etters within rows were significantly different at the 1% level according to Duncan's Multiple Range Test. (n=27 untrained panelists). A close relationship has been reported between physiological maturity and ripening response under ethylene treatment, where green-harvested tomatoes that required 6 or more days of ethylene treatment to initiate ripening were likely immature at harvest (Kader at al., 1977; Kavanagh et al., 1986). In this study, sensory tests demonstrated that untrained panelists could detect overall flavor differences between

PAGE 93

85 table-ripe tomatoes as related to their ethylene exposure time. There was a trend of increasing degree of difference from the control with increasing ethylene exposure time to initiate ripening. Previous research (Kader et al., 1977; Watada and Aulenbach, 1979) has also alluded to sensory differences related to fruit maturity at harvest. Several panelists in this study noted "unpleasant", "metallic", "strange" or "lingering" off-flavors in tomato samples exposed to extended ethylene treatments. Hayase et al. (1984) reported tomatoes picked at green stage and ripened postharvest occasionally presented off-flavors, substantially weakening their characteristic tomato flavor. Significant differences in pH, titratable acidity, total sugars and vitamin C content were found for 'Agriset-761' fruits. In general, significantly lower pH values, found in 1 day ethylene samples, concurred with significantly higher titratable acidity content. Fruit exposed to 5 days ethylene had significantly higher pH values, while significantly lower total sugars, compared to those exposed to 1 and 3 days ethylene exposure time. Vitamin C content was significantly lower in 1 day fruit when compared to those exposed to 3 days or 5 days of ethylene treatment (9.96, 14.10 and 14.30 mg/lOOg fresh weight, respectively) (Table 5.6). In contrast, 'CPT-5' fruit showed significant differences only in titratable acidity, where fruit exposed to 3 days ethylene had significantly higher acidity than those exposed to 1 or 5 of ethylene treatment (0.86%, 1.04% and 0.94 % citric acid, respectively) (Table 5.6). The magnitude of pH and acidity variations required for sensory perception have been reported by Gould (1978), who concluded that changes in pH greater than 0.16 units and of 0.1% in titratable acidity were required for sensory detection. Therefore, flavor differences detected by untrained panelists may have been

PAGE 94

86 attributed in part to sugar and/or acid concentration changes related to harvest maturity for both cultivars. Table 5.6. Table-ripe tomato chemical composition for 'Agriset-761' and 'CPT-5' tomatoes exposed to 1, 3 or 5 days of ethylene treatment to attain breaker stage. Compositional Parameters z Ethylene Gassing Exposure time 1-Day 3 -Days 5-Days Agriset-761 PH 4.23 b 4.28 ab 4.34 a Soluble Solids Content (°Brix) 3.85 a 4.02 a 3.45 a Titratable Acidity (% Citric acid) 0.91 a 0.81 ab 0.73 b Total Sugars (% fresh weight) 1.93 ab 2.29 a 1.69 b Ascorbic Acid (mg/lOOg fresh weight) 9.96 b 14.10a 14.30 a CPT-5 PH 4.25 a 4.28 a 4.23 a Soluble Solids (°Brix) 3.90 a 3.90 a 4.35 a Titratable Acidity (% Citric acid) 1.04 a 0.86 b 0.94 ab Total Sugars (% fresh weight) 10.69 b 13.01 ab 14.33 a Ascorbic Acid (mg/lOOg fresh weight) 1.80 a 2.13 a 2.35 a z Means for parameters with different letters within rows were significantly different at the 5% level according to Duncan's Multiple Range Test. (n=6 composite samples/trt). Aroma volatile analysis by gas chromatography helped to identify additional factors contributing to the significant differences found during sensory panels. The concentrations of five aroma volatile compounds from table-ripe 'Agriset-761' tomatoes were significantly different between ethylene gassing exposure times. Significantly higher concentrations of acetone, hexanal and 2+3-methylbutanol were found in tomatoes exposed to extended ethylene treatment (5-days), whereas, 2-isobutylthiazole and 0ionone decreased significantly compared to other treatments (Table 5.7). Table-ripe 'CPT-5' tomatoes had significant differences in 9 of 15 aroma volatile compounds quantified. Tomatoes exposed to extended ethylene treatments showed increased production of hexanal, /ra/w-2-heptenal, 6-methyl-5-hepten-2-one, 2-isobutylthiazole, 1nitro-2-phenylethane and geranylacetone. Whereas, increased production of ethanol,

PAGE 95

87 2+3-methylbutanol and P-ionone was observed in tomatoes exposed to 3 days of ethylene treatment (Table 5.8). Table 5.7. Aroma volatile compound concentrations for table-ripe 'Agriset-761' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage (n=4 samples/treatment). A rnniQ \/r\l q ti 1 a /" , r\mi^r\iinrlc z rVIUIIla VUlalllC V^UIIipUUIlUS Ethylene Gassing Exposure time l day 3 days 5 days U.OZ D 0. 1 1 D l.oy a Mpthflnnl i V 1 \s L 1 Id 1 1 V ) 1 ^OJ.JJ a 9^8 10 a zoo.jz a Ethanol 52.07 a 55.77 a 84.65 a l-Penten-3-one 0.30 a 0.24 a 0.25 a Hexanal 13.24 ab 12.24 b 18.41 a C/s-3-hexenal 11.06 a 8.20 a 11.03 a 2+3-Methylbutanol 2.14 b 2.18 b 3.54 a 7>aMs-2-hexenal 8.32 a 6.56 a 9.23 a 7raws-2-heptenal 0.05 a 0.05 a 0.05 a 6-Methyl-5-hepten-2-one 0.85 a 0.68 a 0.83 a C/s-3-hexenol 0.05 a 0.05 a 0.05 a 2-Isobutylthiazole 0.14a 0.13 a 0.10b 1 -Nitro-2-pheny lethane 0.20 a 0.20 a 0.20 a Geranylacetone 6.54 a 8.14a 5.39 a p-ionone 0.17 ab 0.45 a 0.12 b Total y 392.96 381.01 403.86 z Means for aroma volatile compounds (uL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. For both cultivars, concentrations of 2+3-methylbutanol, geranylacetone and Pionone were significantly affected by physiological maturity at harvest. In general, volatile compound concentrations increased in fruit exposed to longer ethylene treatments to initiate ripening. This finding contradicts the results observed previously for 'Solimar' tomatoes with different ethylene exposure times. McDonald et al. (1996) reported reductions in six aroma volatile compounds from tomatoes exposed to 1 day of ethylene treatment, when compared to non-treated controls, changes in the concentrations of hexanal, 6-methyl-5-hepten-2-one, geranylacetone and 2+3-methylbutanol were also

PAGE 96

88 observed for 'Agriset-761' and 'CPT-5' samples. Significant reductions in 6-methyl-5hepten-2-one, 2-isobutylthiazole and geranylacetone found in table-ripe 'Solimar' tomatoes exposed to extended ethylene treatment further supports the relevant effects of harvest maturity on ripe tomato flavor/aroma. Table 5.8. Aroma volatile compound concentrations for table-ripe 'CPT-5' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to attain breaker stage (n=4 samples/treatment). r\l\Jill UlllUVJUllUo Ethylene Gassing Exposure time 1 day 3 days 5 days Acetone 1.01 a 1.72 a 1.23 a Methanol 268.50 a 301.65 a 282.44 a Ethanol 70.20 ab 100.48 a 38.53 b l-penten-3-one 0.28 a 0.28 a 0.33 a Hexanal 17.66 b 16.84 b 25.34 a C/.s-3-hexenal 8.80 a 11.30 a 9.51 a 2+3-methylbutanol 2.29 b 3.15a 2.80 b Trans-2-\\exenal 10.41 a 8.83 a 10.23 a 7raw.s-2-heptenal 0.06 ab 0.05 b 0.07 a 6-methyl-5-hepten-2-one 0.93 ab 0.75 b 1.11 a C/s-3-hexenol 0.06 a 0.06 a 0.05 a 2-isobutylthiazole 0.13 ab 0.11 b 0.14 a 1 -nitro-phenylethane 0.22 b 0.21 b 0.32 a Geranylacetone 6.64 b 7.83 ab 8.87 a (3-ionone 0.13 b 0.22 a 0.17 ab Total y 387.32 453.48 381.14 z Means for aroma volatile compounds (u,L/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. Preliminary "difference from control" sensory tests confirmed the perception of significant differences between table-ripe tomatoes with varying ethylene exposure times to reach breaker stage. It is probable that reported reductions in hexanal, 6-methyl-5heptene-2-one, geranylacetone, l-penten-3-one, methanol, 2+3 methylbutanol, and 1nitro-2-phenylethane by McDonald et al. (1996) were not entirely a direct effect of ethylene treatment but rather a consequence of tomato maturity at harvest.

PAGE 97

89 'BHN-189' and 'Solimar' (Experiment 3). Further information on the character/magnitude of flavor differences due to harvest maturity was obtained during descriptive sensory panels. Flavor attributes for table-ripe 'BHN-189' and 'Solimar' tomatoes picked at green stage and treated with ethylene were compared to those in tomatoes harvested at light-red stage. Table-ripe 'BHN-189' tomatoes that required 3 and 5 days of ethylene treatment were considered lower in ripe aroma, sweetness, tomato flavor, while higher in sourness (5 days) and green-grassy flavor compared to ripeharvested (light-red, stage 5) or 1 day ethylene treatment tomatoes. Ripe-harvested 'BHN-189' fruit were not considered significantly different than those exposed to 1 day ethylene treatment in any flavor descriptor except for off-flavor, where ripe-harvested fruit were rated slightly higher (Figure 5.1). Table-ripe 'BHN-189' tomatoes were stored at 20°C for an additional 14 days (table-ripe + 14 days) to re-assess flavor differences between harvest maturities in overripe tomatoes. At this time, ripe-harvested 'BHN-189' tomatoes were considered higher in ripe aroma (significantly), off-odors, sweetness, and tomato flavor compared to green-harvested tomatoes regardless of ethylene exposure time (Figure 5.2). Only slight sensory differences between ethylene treatments were documented in overripe tomatoes. For 'Solimar' tomatoes, slightly different sensory ratings between harvest maturities at table-ripe ripe stage were documented. Tomatoes that required 1 and 3 days of ethylene treatment were considered significantly higher in ripe aroma, while significantly lower in sourness, compared to 5-day ethylene treatment or ripe-harvested tomatoes (light-red, stage 5) (Figure 5.3). Similar to 'BHN-189', overripe 'Solimar' tomatoes (table-ripe +14 days) harvested at light-red stage were rated higher in ripe

PAGE 98

90 aroma, sweetness and tomato flavor when compared to the tomatoes picked at green stage, regardless of ethylene exposure time. In addition, tomatoes exposed to 1 and 3 days of ethylene treatment, although not significant, had considerably higher ratings for off-odors and off-flavors compared to those exposed to 5 days of ethylene treatment or harvested at light red stage (Figure 5.4). Compositional analyses from table-ripe 'BHN-189' tomatoes showed slight reductions in pH and increased soluble solids with increasing ethylene exposure time. Ripe-harvested tomatoes had higher pH than the 3and 5-day ethylene treatments, and lower or equal soluble solids than the 1or 3-day ethylene treatments (Table 5.9). Titratable acidity was highest in ripe-harvested tomatoes, whereas lowest in tomatoes exposed to 1 day of ethylene treatment. After the 14-day storage period, soluble solids and titratable acidity decreased considerably in all treatments. Conversely, pH values were lower in ripe-harvested and 1-day ethylene exposure time tomatoes, while higher in 3-day and 5-day ethylene treatments (Table 5.9). Table 5.9. Table-ripe and overripe chemical composition parameters for 'BHN-189' Chemical Composition Parameters z Ripe harvest Green stage ethylene exposure time 1-Day 3-Days 5-Days Table-ripe stage PH 4.32 4.34 4.31 4.27 Soluble solids content (°Brix) 3.8 3.7 3.8 4.0 Titratable acidity (% citric acid) 1.15 1.07 1.13 1.13 Sucrose equivalents 3.30 4.33 4.35 4.30 Lycopene content (pg/g fresh weight) 26 24 28 16 Overripe (table-ripe + 1 4 days) PH 4.30 4.25 4.38 4.35 Soluble solids content (°Brix) 3.6 3.3 3.6 3.7 Titratable acidity (% citric acid) 0.84 0.92 0.79 0.72 Sucrose equivalents 4.17 3.79 4.20 3.63 (n=8 fruit/sample).

PAGE 99

91 Sucrose equivalents were lower in ripe-harvested tomatoes when compared to green-harvested tomatoes regardless of ethylene exposure time. Nonetheless, sugar levels increased in ripe-harvested tomatoes while they decreased in green-harvested ones at overripe stage. Lycopene content was considerably lower in tomatoes exposed to extended ethylene treatment compared to remaining treatments. Table-ripe 'Solimar' tomatoes had increasing pH and soluble solids contents with increasing ethylene exposure time, while, ripe-harvested tomatoes had the lowest pH (4.2), lowest titratable acidity (0.92%) and next to lowest soluble solids content (Table 5.10). Table 5.10. Table-ripe and overripe stage chemical composition parameters for 'Solimar' Chemical Composition Parameters z Ripe harvest Green stage ethylene exposure time 1-Day 3-Days 5-Days Table-ri 5e stage PH 4.20 4.27 4.32 4.35 Soluble solids content (°Brix) 3.9 3.8 4.1 4.4 Titratable acidity (% citric acid) 0.92 1.19 1.19 1.08 Sucrose equivalents 4.39 5.20 5.55 4.17 Lycopene content (ug/g fresh weight) 19 14 15 12 Overripe (table-ripe + 14 days) PH 4.28 4.25 4.27 4.21 Soluble solids content (°Brix) 3.4 3.8 4.2 3.9 Titratable acidity (% citric acid) 0.67 0.72 0.71 0.74 Sucrose equivalents 5.60 5.90 5.75 3.60 z Means for compositional parameters represent the average of 2 composite samp es/ treatment (n=8 fruit/sample). After 14 days of storage in 'Solimar' tomatoes, pH values decreased in all treatments except for ripe-harvested tomatoes, whereas, titratable acidity was considerably lower in all treatments at the overripe stage. Soluble solids remained lowest for ripe-harvested tomatoes, while sucrose equivalents were consistently lowest in tomatoes exposed to extended ethylene treatments. At overripe stage, ripe-harvested

PAGE 100

92 tomatoes had comparable sugar levels to those in tomatoes exposed to 1 and 3 days of ethylene treatment. Lycopene content was considerably higher in tomatoes ripeharvested tomatoes, while tomatoes exposed to 5 days of ethylene treatment had the lowest contents. Table 5.11. Aroma volatile compound concentrations for table-ripe 'BHN-189' harvested at light-red or green stages. rvruiud VOlalllC conipouiiub rvipc harvest 11U1 V V^..5 1 Green stage ethylene exposure time 1 day 3 day 5 day Acetaldehvde 17.47 a 12.97 a 11.25 a 16.04 a Acetone 0.31 c 0.39 abc 0.37 ab 0.43 a Methanol 271.77 a 263.98 a 250.59 a 264.93 a Ethanol 20.85 a 21.47 a 21.62 a 23.00 a l-Penten-3-one 0.08 a 0.06 a 0.04 a 0.06 a Hexanal 1.78 a 2.66 a 1.14a 2.34 a C/.s-3-hexenal 4.35 a 2.10 b 1.85 b 2.51 b 2+3-Methylbutanol 0.81 c 2.04 b 2.20 b 2.79 a 7>aws-2-hexenal 3.29 a 2.83 a 2.71 a 2.99 a 7raw.s-2-heptenal 0.11 b 0.18a 0.18a 0.16a 6-Methyl-5-hepten-2-one 0.97 a 0.83 a 0.71 a 0.67 a C/5 , -3-hexenol 0.81 a 0.53 b 0.47 b 0.47 b 2-lsobutylthiazole 0.06 b 0.08 a 0.09 a 0.09 a 1 -Nitro-2-pheny lethane 0.08 a 0.06 a 0.06 a 0.06 a Geranylacetone 2.01 b 3.02 a 2.46 b 2.26 b P-ionone nd y nd nd 0.16a Total * 324.76 313.19 295.74 321.94 z Means for aroma volatile compounds (uL/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Volatile compound concentrations not detected. x Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. Aroma volatile analysis for table-ripe 'BHN-189' tomatoes sampled during descriptive sensory panels revealed that 7 of 16 volatile compounds had significant concentration differences due to harvest treatments. Notably, ripe-harvested tomatoes had the lowest concentrations of acetone, 2+3-methylbutanol, fram , -2-heptenal, 2isobutylthiazole and geranylacetone, and the highest concentrations of cw-3-hexenal and cw-3-hexenol (Table 5.11). Tomatoes picked at green stage, regardless of ethylene

PAGE 101

93 exposure time, were significantly lower in cw-3-hexenal, c/s-3-hexenol, while significantly higher in 2-isobutylthiazole compared to ripe-harvested ones. It is important to note that tomatoes exposed to extended ethylene treatment (5 days) were significantly higher in 2+3-methylbutanol and p-ionone concentrations compared to rest of treatments. Table 5.12. Aroma volatile compound concentrations for overripe 'BHN-189' tomatoes table-ripe + 14 days) harvested at light-rec or green stages. Aroma volatile compounds 1 Ripe harvest Green stage ethylene exposure time 1 day 3 days 5 davs Acetaldehyde 12.06 a 11.30a 14.35 a 12.67 a Acetone 0.36 a 0.34 a 0.45 a 0.36 a Methanol 250.06 a 248.92 a 287.26 a 253.56 a Ethanol 16.72 a 15.84 a 16.03 a 16.71 a l-Penten-3-one 0.11 a 0.06 b 0.06 b 0.07 b Hexanal 11.83 a 11.32a 8.26 ab 5.98 b C/s-3-hexenal 3.068 a 1.67 ab 1.63 ab 1.02 b 2+3-Methylbutanol 0.36 a 0.41 a 0.29 b 0.27 b 7>arts-2-hexenal 4.10a 3.18b 3.02 b 2.67 b 7ra«s-2-heptenal 0.03 a 0.03 a 0.02 a 0.02 a 6-Methyl-5-hepten-2-one 0.75 a 0.67 a 0.50 a 0.57 a C/'s-3-hexenol 0.58 a 0.38 b 0.30 b 0.33 b 2-Isobutylthiazole 0.07 a 0.07 a 0.05 b 0.04 b 1 -Nitro-2-phenylethane 0.05 a 0.06 a 0.055 a 0.06 a Geranylacetone 3.51 a 3.75 a 3.34 a 3.26 a P-ionone 0.13a 0.13 a 0.19a 0.30 a Total y 303.79 297.99 335.80 297.57 z Means for aroma volatile compounds (u.L/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. Following 14 days of storage, overripe 'BHN-189' tomatoes showed significant changes between treatments in 7 of 16 compounds quantified (Table 5.12). Four volatile compounds, cw-3-hexenal, 2+3-methylbutanol, cw-3-hexenol and 2-isobutylthiazole had significant differences between treatments at table-ripe and overripe stages. Ripeharvested tomatoes were highest in l-penten-3-one, cw-3-hexenal, /ra«s-2-hexenal, cis-3hexenol, and 2-isobutylthiazole. Conversely, concentrations of hexanal, c/.s-3-hexenal,

PAGE 102

94 2+3-methylbutanol, and 2-isobutylthiazole were lowest in those exposed to 5 days ethylene treatment. Green-harvested tomatoes, regardless of ethylene exposure time, had significantly lower concentrations of l-penten-3-one and cz's-3-hexenol compared ripeharvested ones (Table 5.12). For table-ripe 'BHN-189' tomatoes, some of the important aroma volatile compounds showed consistent trends, notably, c/.y-3-hexenal concentrations were approximately 2-fold higher in ripe-harvested tomatoes compared to green-harvested ones. Similarly, c/s-3-hexenol concentrations in ripe-harvested tomatoes were approximately 67% higher than those found in green-harvested ones. In contrast, concentrations of 2+3-methylbutanol were higher in green-harvested tomatoes, compared to ripe-harvested ones; furthermore, tomatoes exposed to extended ethylene treatment had the highest concentrations. Following 14 days of storage at 20°C, reductions in most volatile compounds were evident, probably due to volatile compound precursor consumption over time. Nonetheless, it is noteworthy to mention that hexanal concentrations increased at least two-fold in overripe tomatoes, contrasted by approximately 30% reductions in cw-3-hexenal and c/.s-3-hexenol. 7>an.s-2-hexenal, an isomerization product of c/.s-3-hexenal, experienced a 10% increase in overripe tomatoes. At table-ripe stage, 'Solimar' tomatoes showed significant differences in 10 of 16 aroma volatile compounds quantified. Tomatoes exposed to 5 days of ethylene treatment had the lowest concentrations of methanol, hexanal, c/s-3-hexenal, /raw5 , -2-hexenal, 6methyl-5-hepten-2-one, cz's-3-hexenol, 2-isobutylthiazole, and l-nitro-2-phenylethane, whereas concentrations of /raw.y-2-heptenal were highest compared to remaining treatments (Table 5.13). In contrast, ripe-harvested tomatoes had the highest

PAGE 103

95 concentrations of cw-3-hexenal, trans-2-hexenal, 6-methyl-5-hepten-2-one, cis-3hexenol, and 2-isobutylthiazole. Tomatoes harvested at green stage had significantly lower concentrations of 6-methyl-5-hepten-2-one regardless of ethylene exposure time. Table 5.13. Aroma volatile compound concentrations for table-ripe 'Solimar' harvested at light-red and green stages. Aroma volatile compounds z Ripe harvest Green stage ethylene exposure time 1 Hav 1 UaV J Udys 5 days Acetaldehyde 40.31 a 20.90 a 24.75 a 29.17 a Acetone 0.35 b 0.35 b 0.43 a 0.37 b Methanol 336.24 b 327.88 b 416.98 a 212.24 c Ethanol 49.99 a 29.73 a 43.22 a 25.69 a 1 -Penten-3-one 0.08 a 0.07 a 0.07 a 0.07 a Hexanal 4.18a 6.12a 5.40 a 2.35 a C/s-3-hexenal 3.94 a 3.24 a 3.69 a 1.29 b 2+3-Methylbutanol 2.09 a 1.03 a 1.61 a 0.75 a Trans-2-hexena\ 3.43 a 2.82 b 3.41 a 1.98 c 7>ans-2-heptenal 0.04 b 0.02 c 0.02 c 0.05 a 6-Methyl-5-hepten-2-one 1.55 a 0.53 b 0.62 b 0.44 b C/s-3-hexenol 0.93 a 0.56 b 0.89 a 0.12 c 2-lsobutylthiazole 0.11 a 0.07 b 0.08 ab 0.03 c 1 -Nitro-2-phenylethane 0.09 a 0.09 a 0.12 a 0.05 b Geranylacetone 2.43 a 2.82 a 2.85 a 0.53 a (3-ionone 0.16a 0.15 a 0.15 a nd y Total" 445.93 396.39 504.290 275.11 z Means for aroma volatile compounds (u.L/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Volatile compound concentrations not detected. " Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. After 14 days of storage, overripe 'Solimar' tomatoes showed a reduction in the production of aroma volatile compounds for all harvest treatments. The number of compounds showing significant differences increased to 14 of 16 compounds quantified. Similar to results at table-ripe stage, overripe tomatoes exposed to 5 days ethylene treatment had the lowest concentrations of all 14 volatile compounds showing significant differences. Conversely, ripe-harvested tomatoes had the highest concentrations in the same compounds except for methanol, 6-methyl-5-hepten-2-one and 2-isobutylthiazole

PAGE 104

96 (Table 5.14). Significant differences in acetone, methanol, c/s-3-hexenal, trans-2hexenal, /raws-2-heptenal, 6-methyl-5-hepten-2-one, cz's-3-hexenol, 2-isobutylthiazole, and l-nitro-2-phenylethane were present in tomatoes at table-ripe and overripe stages. Table 5.14. Aroma volatile compound concentrations for overripe 'Solimar' tomatoes (table-ripe + 14 davs) harves ted at light-red or green stages. Aroma volatile compounds 1 Rine harvpst Green stage ethylene exposure time 1 day 3 dsys 5 days Acetaldehyde 19 88 a 16.62 a 1 U.vO aH.s-2-heptenal 0.31 a 0.18b 0.18 b 0.05 c 6-Methyl-5-hepten-2-one 1.56 a 2.11 a 1.90 b 0.46 b C/.v-3-hexenol 0.47 a 0.47 a 0.40 a 0.09 b 2-Isobutylthiazole 0.05 b 0.06 a 0.04 c nd y 1 -Nitro-2-pheny lethane 0.09 a 0.07 b 0.07 b 0.04 c Geranylacetone 2.76 a 2.24 b 1.958 b 0.55 c p-ionone 0.11 a 0.08 a 0.113a 0.05 a Total" 375.49 371.65 420.94 256.86 z Means for aroma volatile compounds (u.L/ L) with different letters across rows are different at the 5% level according to Duncan's Multiple Range Test. y Volatile compound concentrations not detected. x Total aroma volatile concentrations (uL/ L) based on the sum of the 16 compounds quantified. 'Solimar' tomatoes seemed to be more susceptible to aroma volatile concentration changes in relation to harvest maturity than 'BHN-189' tomatoes harvested and stored under similar conditions. In fact, 'Solimar' tomatoes suffered a 12% average reduction in total production of volatile compounds at overripe stage compared to table-ripe stage, while overripe 'BHN-189' tomatoes suffered an average 1.6% of their volatile production. Nonetheless, it is safe to assume that such changes in concentration might not translate to the same magnitude in sensory changes once the OUV for the individual

PAGE 105

97 volatile compounds are considered. Some notable volatile compound differences related to harvest maturity in 'Solimar' tomatoes were higher levels of cw-3-hexenaL, 2+3methylbutanol, /raw.s-2-hexenal, and c/.s-3-hexenol found in ripe-harvested tomatoes compared to green-harvested ones. In addition, following 14 days of storage, hexanal levels increased approximately 2-fold while /ra/w-2-hexenal levels only 7%. It should be noted that levels of 6-methyl-5-hepten-2-one increased 3-fold in overripe tomatoes that required extended ethylene treatment while, levels remained unchanged in ripe-harvested or 1-day ethylene exposure time (mature-green) ones. Aroma volatile compounds have been proposed to have a significant effect on the human perception of tomato flavor (Petro-Turza, 1987; Kader et al., 1977) and the results from these studies further supports this contention. Greater volatile production has been related to greater flavor potential for tomato fruits (Baldwin et al., 1991b). In these present experiments, overall volatile production did not seem to be directly related to flavor quality. 'BHN-189' tomatoes exposed to 3 and 5 days ethylene treatment had total volatile concentrations comparable to other treatments. However, changes in the concentration of individual compounds, such as methanol, greatly affected total volatile production calculations without necessarily implying significant sensory changes. Volatile compound production based on biosynthetic pathways showed consistent trends related to harvest maturity. The production of carotenoid-derived compounds, notably geranylacetone, 6-methyl-5hepten-2-one and (3-ionone, was consistently lower in 'Agriset-761', 'Solimar' and 'BHN-189' tomatoes exposed to 5 days of ethylene treatment (immature-green). This group of volatile compounds have been described as providing fruity (Petro-Turza, 1987)

PAGE 106

98 or ripe tomato (Tandon, 1998) character during sensory studies. In fact, during descriptive sensory panels, tomatoes exposed to 5 days ethylene gassing were consistently rated as inferior in ripe aroma, sweetness and tomato flavor. Compositional studies support sensory results since both 'BUN189' and 'Solimar' samples exposed to 5 days ethylene gassing had considerably lower individual sugar (sucrose equivalents) and lycopene contents. Correlation matrices between compositional parameters and sensory descriptor ratings identified significant relationships between tomato flavor/aroma and instrumental measurements. Notably, ratings for "tomato flavor" were negatively correlated with acetaldehyde and ethanol concentrations, while positively correlated to lycopene concentrations (r = -0.60, -0.53, and 0.76, respectively)(Table 5.15). "Ripe aroma" ratings were positively correlated with fra/7s-2-heptenal concentrations (r = 0.49). "Offodor" ratings were positively correlated to 2+3-methylbutanol concentrations and negatively correlated to geranylacetone concentrations (r = 0.54 and -0.65, respectively). Meanwhile, "off-flavor" ratings were positively correlated to the concentrations of acetaldehyde, methanol, and ethanol (r = 0.71, 0.52, 0.65, respectively). At high concentrations, these volatile compounds have been related to incidence of off-flavors during controlled atmosphere and hypoxic conditions (Pesis et al., 1991), probably due to anaerobic metabolism. Surprisingly, numerous volatile compounds correlated with sensory ratings for "sweetness" and "sourness". Hexanal, geranylacetone and p-ionone concentrations were negatively correlated with sourness ratings (r = -0.62, -0.79, and -0.56, respectively). Furthermore concentrations of lycopene, m-3-hexenal, /ra/«-2-hexenal, and trans-2-

PAGE 107

99 heptenal were positively correlated with sweetness ratings (r = 0.70, 0.52, 0.51, and 0.58, respectively). Finally, "green/grassy" ratings were negatively correlated to the concentrations of /raws-2-heptenal r = 0.58 Sourness Soluble Solids r = 0.60 Hexanal r = -0.61 [3-ionone r = -0.56 Green/Grassy Trans-2-hexenal r = -0.51 Geranylacetone r = -0.60 P-ionone r = -0.58 Tomato Flavor Soluble Solids r = -0.61 Lycopene r = 0.76 Acetaldehyde r = -0.60 Ethanol r = -0.53 Off-flavor Acetaldehyde r = 0.71 Methanol r = 0.52 Ethanol r = 0.65 z Correlation coefficients (r) were significant at the 5% level according to F test statistics. In addition to the correlation matrix analysis, stepwise regressions helped to elucidate any interactions between compositional parameters and sensory descriptor ratings. According to the stepwise analysis, "ripe aroma" ratings were negatively influenced by acetaldehyde and acetone concentrations (model r 2 = 0.755). "Off-odor" ratings were best described by the concentrations of geranylacetone, methanol, 2+3methylbutanol and sucrose equivalents (r 2 = 0.88), whereas "off-flavor" ratings were

PAGE 108

100 influenced by /raw,s-2-heptenal, l-nitro-2-phenylethane and acetaldehyde concentrations (r 2 0.98). "Sweetness" ratings were influenced by soluble solids content, pH and trans2-hexenal concentrations (r 2 = 0.88) whereas, "sourness" ratings were affected by 1-nitro2phenylethane, methanol, and /raws-2-heptenal concentrations (r 2 = 0.68). Finally, "tomato flavor" ratings were influenced by the concentrations of acetaldehyde, 1-penten3one, lycopene, cw-3-hexenal, and l-nitro-2-phenylethane (r 2 = 0.86). Electronic Nose Analyses. The EN sensor array successfully segregated table-ripe 'Solimar' tomatoes according to the duration of ethylene treatment required to reach breaker stage. In general, significant increases in individual sensor responses resulted from the interaction with volatile compounds given off by the different tomato samples. To visualize the differences between the ethylene exposure treatments, multivariate discriminant statistical analysis (MVDA) was employed. Distinct clusters of points in a canonical plot illustrated resemblance among fruits from the same ethylene exposure times, and dissimilarity between fruits from different ethylene exposure treatments. The Mahalanobis distance (MD), the distance between the centroids of two clusters adjusted for probability (Srivastava and Carter, 1983), helped put in perspective the extent of differences between the tomato samples. The greater the MD between clusters, the greater the dissimilarity between the treatments. The forward stepwise MVDA procedure identified significant differences between the headspace volatiles present in whole tomato homogenate samples from 'Solimar' tomatoes (Figure 5.5). The MD between clusters corresponding to homogenate samples exposed to 1 and 3 days ethylene treatment was significantly greater than that

PAGE 109

101 between the 3 and 5 days ethylene clusters (24.14 and 6.52 units, respectively). To better understand the contribution of individual fruit tissues to overall differences in volatile profiles, EN analyses were conducted on excised tomato tissues. Forward stepwise MVDA from the sensor outputs successfully classified pericarp homogenate samples into three distinct clusters based on the length of ethylene treatment required to attain breaker stage (Figure 5.6). The MD between 1 and 3 day gassing clusters was ~175% greater than the distance between 3 and 5 day gassing clusters (116.31 and 65.04 units, respectively). Meanwhile, the forward stepwise MVDA analysis also segregated locule tissue samples into three distinct clusters based on the pattern of differences between treatments (Figure 5.7). The MD between 1 day and 3 days ethylene gassing clusters was similar to the distance between 3 and 5 days gassing clusters (15.74 and 13.28 units, respectively). The MD between clusters in all canonical plots for table-ripe 'Solimar' tomatoes showed significantly different volatile profiles as a function of the ethylene treatment exposure time. The differences in whole tomato homogenate seem to be influenced considerably by significant changes in pericarp (including columnella) tissue volatile profiles. The MD observed between 1 and 3 days ethylene treatments in pericarp homogenate samples was about 4.8 and 7.4 times greater than that observed in whole tomato and locular gel homogenate, respectively. In addition to the total concentration of aroma volatile compounds, the number of individual compounds showing significant changes could help explain the differences found during the EN analysis. The prolonged ethylene exposures required by immature-harvested fruits resulted in significant reductions in several important aroma compounds. The pericarp homogenate showed

PAGE 110

102 significant reductions in 8 out of 16 aroma compounds quantified, thus suggesting pericarp's greater susceptibility to harvest maturity. Even though the locule tissue aroma seemed to be less affected by the treatments (five compounds showed significant reductions), still less than one third of the aroma compounds quantified were significantly lower. EN analysis performed on the tomato homogenate samples used during the descriptive sensory panels served to further support results put forth by sensory and GC analyses. 'BHN-189' tomatoes were accurately classified into four distinct clusters based on their volatile compound analysis by the EN (Figure 5.8). The distances between clusters were all of statistical significance except for the distance between tomatoes harvested at light-red stage and those exposed to 1 day ethylene treatment. In addition, Mahalanobis distances were approximately 4-fold and 7.5-fold greater between ripeharvested and 3 day and 5 day ethylene treatments, compared to those with 1 day ethylene treatment samples. Based on their volatiles, 'BHN-189' tomatoes were classified by the MVDA procedure with accuracy ranging from 83-100%. Following a 14-day storage period, the degree of separation between EN classification clusters decreased considerably (Figure 5.9). The distances between the 1 day and 3 day treatment clusters were not significant. Mahalanobis distances were lowest between 1 day and 3 day ethylene treatments (6.75 units). Nonetheless, classification of tomato samples based on EN volatile analysis showed accuracy ranging between 99100%. EN analysis of table-ripe 'Solimar' tomatoes samples from the descriptive sensory panels classified harvest treatments into four distinct clusters. The distances between

PAGE 111

103 clusters were highly significant in all cases. The greatest MD identified in the MVDA analysis occurred between 1 day and 3 day ethylene treatments (11.58 units), and between 1 day and 5 day treatments (13.35 units). Ripe-harvested tomatoes clustered inbetween the 1 day and 3 day ethylene treatment clusters (Figure 5.10). The MVDA analysis was capable of classifying the 'Solimar' samples with accuracy ranging from 75.9 and 100%. According to the EN analysis, after 14 days of storage (overripe stage), the magnitude of the volatile profile differences between 'Solimar' harvest treatments increased considerably. Distances between all clusters were highly significant, meanwhile MD's were approximately between 2-fold and 10-fold greater than those observed prior to storage (Figure 5.1 1). By far, the greatest MD occurred between ripeharvested tomatoes and 1, 3 or 5 day ethylene treatments (133.04, 199.07 and 111.07 units, respectively), when compared to those distances in-between ethylene treatments. Due to the apparently large differences in volatile profiles between treatments, the MVDA procedure had no difficulties classifying treatments with 100% accuracy. The relationship between EN sensors and GC aroma volatile compounds was investigated through the use of correlation matrices where few of the volatile compounds quantified significantly influenced EN sensor outputs. Concentrations of c/s-3-hexenal were positively correlated with responses from EN sensors t278, t264, t263, t262, t261, t259, and t258 (r = 0.51 to 0.66), whereas negatively correlated to responses from sensors t301 and t263 (r = -0.49 and -0.55, respectively)(Table 5.16). Concentrations of trans-2heptenal were positively correlated with responses from EN sensors t301, t298, t297, t283 (r = 0.50 to 0.65). In addition, concentrations of 2-isobutylthiazole, a compound

PAGE 112

104 unique to tomato aroma, were negatively correlated to EN sensor responses from t301, t283, t264, and t263 (r = -0.49 to -0.63). Finally, concentrations of two carotenoidderived volatile compounds, geranylacetone and P-ionone, were negatively correlated with responses from EN sensors t283, t264, t262, and t261 (r =-0.41 to -0.56). Table 5.16. Correlation coefficients between electronic nose sensor outputs and compositional parameters or sensory panel descriptor ratings from 'BHN-189' and 'Solimar' tomatoes harvested at green or light-red stages. EN Sensor Sensory Descriptor Aroma Volatile Compound Type 301 C/s-3-hexenal r = -0.49 z rra«5-2-heptenal r = 0.50 2-Isobutylthiazole r = -0.52 Type 298 Ripe Aroma r = 0.47 7>a/is-2-heptenal r = 0.65 Type 297 7>ans-2-heptenal r = 0.64 Type 283 Green/Grassy r = 0.48 7>a/w-2-heptenal r = 0.59 2-Isobutylthiazole r = -0.49 Geranylacetone r =-0.49 Type 278 Green/Grassy r = 0.55 C/s-3-hexenal r = 0.63 Type 264 Green/Grassy r = 0.49 C/s-3-hexenal r = 0.64 2-Isobuylthiazole r =-0.54 Geranylacetone r = -0.4 1 Type 263 C/s-3-hexenal r = 0.60 C/s-3-hexenol r = -0.51 2-Isobutylthiazole r = -0.63 Type 262 Off-odor r = 0.50 Green/Grassy r = 0.55 C/s-3-hexenal r = 0.66 Geranylacetone r = -0.52 P-ionone r = -0.49 Type 261 Ripe Aroma r = 0.50 Green/Grassy r = 0.54 C/s-3-hexenal r = 0.62 P-ionone r = -0.56 Type 260 C/s-3-hexenal r = 0.58 Type 259 Ripe Aroma r = 0.51 C/s-3-hexenal r = 0.54 Type 258 C/s-3-hexenal r = 0.61 z Correlation coefficients (r) were significant at the 5% level according to F test statistics. Significant relationships between EN sensor outputs and sensory descriptor ratings for tomato samples were also identified. Ratings for ripe aroma were positively correlated to responses from EN sensors t298, t261, and t259 (r =0.47 to 0.51) (Table 5.16). Meanwhile, sensory ratings for off-odors were positively correlated with outputs

PAGE 113

105 from sensor t262 (r = 0.55). Finally, ratings for green/grassy flavor were correlated to responses from EN sensors t283, t278, t264, and t262 (r = 0.48 to 0.55). Conclusions Changes in chemical composition and aroma volatile compounds documented in four commercial tomato cultivars ('Agriset-761', 'CPT-5', 'BHN-189' and 'Solimar') as related to maturity at harvest induced significant differences in sensory perception. In addition, carotenoid-derived volatile compounds (geranylacetone, 6-methyl-5-hepten-2one and p-ionone) were consistently lower in tomatoes exposed to extended ethylene treatments (>3 days), which also had consistently lower concentrations of lycopene. Descriptive sensory panels revealed that fruit exposed to extended ethylene treatments (>3 days) had significantly lower ratings for ripe tomato aroma, flavor and sweetness, and higher ratings for sourness and green/grassy flavor when compared to tomatoes with brief ethylene exposure time (1-day) or harvested ripe. In addition to generally higher soluble solids contents and pH values found in tomatoes exposed to 5days ethylene, there were significant correlations between sensory panel ratings and aroma volatile concentrations. Concentrations of acetaldehyde, methanol and ethanol were negatively correlated to tomato flavor ratings; meanwhile, /r
PAGE 114

106 acetaldehyde, and 6-methyl-5-hepten-2-one were related to off-flavor and off-odor sensory ratings. The use of new technologies, such as the electronic nose (EN), may prove useful in correlating instrumental measurements to results from sensory panel evaluations. The EN sensor array discriminated between the different ripe tomato samples from the different ethylene treatments (i.e. harvest maturities) and these results were contrasted to GC aroma volatile profile changes. In this study, the GC headspace technique for aroma volatile quantification was useful to contrast the differences in volatile profiles observed during the EN analysis. Some lipid-derived volatile compounds (c/5-3-hexenal and trans2-heptenal) and carotenoid-derived compounds (geranylacetone and p-ionone) were significantly correlated to responses from numerous EN sensors. From these studies, it is concluded that maturity at harvest significantly influenced ripe tomato flavor and aroma. Tomatoes exposed to >3 days of ethylene treatment, most likely harvested immature, resulted in inferior flavor upon reaching tableripe stage. Conversely, tomatoes that required only 1 day of ethylene treatment, likely mature at harvest, were comparable or even superior in sensory qualities to ripe-harvested tomatoes. It was noted that green-harvested tomatoes, regardless of ethylene exposure time, had considerably lower sensory qualities at overripe stage when compared to those harvested at light-red stage.

PAGE 115

107 Figure 5.1. Sensory panel descriptor ratings for ripe 'BHN-189' tomatoes harvested at light-red or green stages. Significant differences (a=0.05) within ethylene treatment exposure times for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical range for mean separations at the 5% level.

PAGE 116

108 Figure 5.2. Sensory panel descriptor ratings for overripe 'BHN-189' tomatoes (table-ripe + 14 days) harvested at light-red or green stages. Significant differences (a=0.05) within ethylene treatment exposure times for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical range for mean separations at the 5% level.

PAGE 117

109 Sensory Panel Descripti Figure 5.3. Sensory panel descriptor ratings for table-ripe 'Solimar' tomatoes harvested at light-red or green stages. Significant differences (cc=0.05) within ethylene treatment exposure times for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical range for mean separations at the 5% level.

PAGE 118

110 Figure 5.4. Sensory panel descriptor ratings for overripe 'Solimar' tomatoes (table-ripe + 14 days) harvested at light-red or green stages. Significant differences (ct=0.05) within ethylene treatment exposure times for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical range for mean separations at the 5% level.

PAGE 119

Ill d o = g 1 § a 0 8 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 O Whole 1 day Whole 3 days O Whole 5 days -15 -10 0 5 10 Canonical Function 1 15 20 25 Figure 5.5. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in whole fruit homogenate samples from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to reach breaker stage. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P < 0.0066).

PAGE 120

112 12 10 8 (N 6 s I 4 o n. 2 1 o 1 -2 -4 -6 -8 -80 -60 -40 -20 0 20 Canonical Function 1 O O Pericarp 1 day Pericarp 3 days Pericarp 5 days 40 60 80 Figure 5.6. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in pericarp tissue homogenate samples from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to reach breaker stage. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P < 0.00001). The scale used for the canonical function 1 was increased compared to other fruit tissues to accommodate greater distances between pericarp tissue groupings.

PAGE 121

113 Figure 5.7. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in locule tissue homogenate samples from table-ripe 'Solimar' tomatoes exposed to 1, 3, or 5 days of ethylene treatment to reach breaker stage. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P < 0.00001).

PAGE 122

114 -5 -6 -7 -8 -30 o 1 day 3 days o 5 days A Light-red -20 -10 0 Canonical Function 1 10 20 30 Figure 5.8. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in whole fruit homogenate samples from table-ripe 'BHN-189' tomatoes harvested at light-red or green stages. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P< 0.0001).

PAGE 123

115 Figure 5.9. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in whole fruit homogenate samples from overripe 'BHN-189' tomatoes (table-ripe + 14 days) harvested at light-red or green stages. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P < 0.0007).

PAGE 124

116 Figure 5.10. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in whole fruit homogenate samples from table-ripe 'Solimar' tomatoes harvested at light-red or green stages. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P < 0.0001).

PAGE 125

117 -100 0 50 Canonical Function 1 100 150 Figure 5.11. MVDA canonical plot analysis for EN sensor outputs from headspace volatiles present in whole fruit homogenate samples from overripe 'Solimar' tomatoes (table-ripe +14 days) harvested at light-red or green stages. The ellipses around ethylene treatment groupings represent 95% confidence areas. Distances between groupings were very highly significant (P < 0.0001).

PAGE 126

CHAPTER 6 HARVEST MATURITY AND STORAGE TEMPERATURE AFFECT TOMATO VOLATILE PRODUCTION AND FLAVOR Introduction Evidence of the adverse effects of low temperature storage on tomato flavor has been published (Kader et al., 1978; Stern et al., 1994; McDonald et al., 1996). Tomatoes are subtropical plants native to Mexico and Central America and as such will incur significant chilling injury (CI) from postharvest exposure to temperatures below 13°C. CI symptoms in tomato include uneven or partial ripening, fruit softening, increased susceptibility to postharvest fungal pathogens, reduced flavor, and surface pitting (Hobson, 1987). The extent of injury depends on the storage temperature, the period of time fruits are exposed to that temperature, and the stage of fruit ripeness. Commercial storage temperature recommendations have been formulated as a result of chilling injury studies and are based on threshold temperatures which do not induce the development of visual CI symptoms or major compositional changes (Hobson, 1981). Greenand breaker-stage (<10% red coloration, USDA, 1976) fruits are considered more susceptible to chilling temperature storage than ripe fruits (Autio and Bramlage, 1986). In contrast, red-ripe fruit have been shown to withstand lower temperature storage for longer periods of time without development of CI symptoms (Cook et al., 1958). However, storage at 118

PAGE 127

119 recommended temperatures of 10° to 13°C might have a significant effect on tomato flavor even before any visual injury symptoms are apparent. The extent of CI symptoms has been ameliorated through the modification of temperature exposure regimes. In tomatoes, intermittent warming of fruit (Cabrera and Saltveit, 1990) and exposure to temperatures above 35°C prior to chilling (Lurie and Klein, 1991) have successfully reduced visual CI symptoms. As a consequence of short (2 or 3 days) exposures to temperatures over 35°C, a reversible inhibition of tomato ripening was documented (Lurie and Klein, 1990). Storage at elevated temperatures can interfere with ethylene and lycopene synthesis thus affecting fruit color (Masarirambi, 1997; Biggs et al., 1988; Ogura et al., 1975). During tomato fruit ripening, a series of quantitative and qualitative changes take place in tomato aroma volatile profiles. Organic acids, soluble sugars, amino acids, pigments, and over 400 aroma compounds contribute to characteristic tomato flavor (Petro-Turza, 1987). Due to the diversity of biosynthetic pathways that contribute to the formation of volatile compounds, tomato aroma could be a sensitive indicator of fruit injury as a result of improper postharvest handling. Nonetheless, the effects of storing tomatoes at currently recommended storage temperatures on flavor and aroma has not been thoroughly addressed. There is a need for nondestructive instrumental measurements to screen inferior flavored fruit and to identify postharvest handling scenarios that affect flavor quality. Electronic nose (EN) technology, combined with pattern recognition methods, has already been used successfully in nondestructive quality measurements of blueberries (Simon et al., 1996) and cantaloupe melons (Benady et al., 1995). In tomatoes, the EN

PAGE 128

120 was sensitive enough to differentiate ripe tomatoes which were harvested at either mature-green or immature-green stages, a task not easily achieved by any other nondestructive instrumental measurement (Maul et al., 1997b). These studies were carried out with the purpose of documenting the effects of low temperature storage on fresh market tomato quality. Experiments were designed to 1) document the effects storage of green-harvested tomatoes at temperatures below recommended thresholds on individual tomato tissues (locule and pericarp); 2) to quantify the flavor and aroma changes occurring in ripe-harvested tomatoes stored for up to 12 days at 5° to 20°C; and 3) to determine the feasibility of high temperature exposure (38°C for 2 days) prior to low temperature storage as a tool to ameliorate chilling-induced flavor quality changes. Materials and Methods For the first experiment 'Solimar' tomatoes (Asgrow Seed Co., Kalamazoo, MI) were harvested from a commercial field in Palm Beach County, Florida. After being transported to Gainesville, FL on the same day of harvest, green-harvested fruit were held at 20°C and treated in a flow-through system consisting of a humidified, 100 uL/L ethylene/air gas mixture to accelerate the onset of ripening. To reduce inherent variability due to the effects of fruit maturity at harvest, mature-green fruit were separated from immature-green fruit based on their ethylene requirement to attain breaker stage (stage 2, <10% red coloration, USDA, 1976). Fruits requiring up to 3 days of ethylene treatment to reach breaker stage were considered mature-green (MG) and, thus capable of ripening

PAGE 129

121 with proper flavor (Kader et al., 1978; Sargent et al., 1997). Meanwhile tomatoes requiring more than 3 days of ethylene treatment were considered to have been harvested immature-green (IG), and, therefore likely to have inferior flavor when fully ripe. Following ethylene treatment, tomatoes were held at 20°C and 95% RH until fully ripe (red, firm) then, transferred to a chilling temperature of 5°C for 7 days. There were four treatments: mature-green (MG) and immature-green (IG) harvest and storage at either 5°C or 20°C. Only tomatoes which were considered table-ripe were used for quality analyses (stage 6, >90% red coloration) and significantly losing firmness: 3-4 mm deformation threshold upon application of a 9.8N force for 5 seconds. Table-ripe 'Solimar' tomatoes were sliced equatorially and locule tissue was separated from pericarp tissue. Six 20-mL samples of whole fruit, locule, and pericarp tissues were individually homogenized for EN and compositional analyses. In a second experiment, 'Solimar' and 'BHN-189' tomatoes were harvested from commercial fields in Gainesville and Quincy, FL, respectively during fall 1997. Tomato fruit were harvested at light-red stage (Stage 5, <90% red coloration), washed, sorted for defects, and randomly divided into four groups (n = 50 fruits). Each group of 50 tomatoes was placed at one of four storage temperatures: 20°C (room temperature), 12.5°C (recommended temperature for green and breaker stage fruit), 10°C (recommended for ripe fruit), and 5°C (typical household refrigerator temperature). Samples of 9-10 fruits from each temperature treatment were removed after 2 days and 8 days ('BHN-189') or 4 days, 8 days and 12 days of storage ('Solimar'). At the time of sensory analysis, tomato fruit were at table-ripe stage as previously defined.

PAGE 130

122 In a third experiment, 'Trust' (De Ruiter Seeds Inc., Columbus, OH) tomatoes grown in greenhouses using hydroponic conditions were harvested at light-red stage (stage 5) and transported to Gainesville, FL for storage temperature treatments. Tomatoes were sorted for uniformity in color and size, then divided into two groups (n=40 fruit). One group was placed inside a humidified chamber at 38°C for 2 days, while the second group was placed at 20°C. Following the 2-day pre-treatment period, tomatoes stored at 38° or 20°C were transferred to a cold room at 5°C and 85-90% RH for 7 days. Five days into the low temperature treatment (7 days from initial harvest), a second group of 'Trust' tomatoes was harvested from the same greenhouses at light-red (stage 5) to red (stage 6) ripeness stages. Light-red tomatoes were stored at 20°C and red tomatoes were held at 38°C for 2 days. Finally, after the 2-day pre-treatment which coincided with a 7-day low temperature treatment period, all tomatoes were transferred to 20°C for 1 additional day prior to sensory, GC, EN and compositional analyses. Over a period of three months, a descriptive sensory panel (10 males and 6 females between 20 and 65 years old) was trained to describe flavor and aroma attributes from fresh market tomatoes. Training sessions and descriptive sensory analysis procedures were the same as those described in chapter 5. During sensory analysis, four 40-mL samples of tomato homogenate for GC and EN analyses were combined with 10mL of a saturated CaCl 2 solution, blended at high speed for 10 seconds and flash-frozen with liquid nitrogen. CaCl 2 was added to reduce enzymatic activity that could contribute to further volatile changes following tissue maceration and subsequent storage at -80°C (Buttery and Ling, 1993).

PAGE 131

123 Tomato homogenate samples from each of the temperature treatments were centrifuged at 18,000 X g n and 5°C. The supernatant was filtered through cheesecloth, and stored inside scintillation vials at -20°C for later analysis. Titratable acidity, expressed as % citric acid, was determined by titrating 1.5 g of tomato supernatant to 8.2 pH with a 0.1 N NaOH solution using an automatic titrimeter (Fisher Scientific, Pittsburgh, PA). Soluble solids, expressed as °Brix, was measured using a tabletop digital refractometer (Abbe Mark II, Reichart-Jung, Buffalo, NY) and pH measurements were conducted using a digital pH-meter (Corning model 140). Tomato fruit lycopene content was determined using a colorimetric method adapted from Umiel and Gabelman (1971) as described previously in Chapter 5. Individual sugar analysis (glucose and fructose) was performed using an adaptation of the high performance liquid chromatography (HPLC) method described in Chapter 5 (Baldwin et al., 1991c). Tomato volatile compounds were identified and quantified by GC using the same headspace analysis technique (Baldwin et al., 1991ab) detailed in Chapter 5. Previous studies (Baldwin et al., 1991ab; McDonald et al., 1996) evaluated 16 important tomato aroma volatile compounds (acetaldehyde, acetone, methanol, ethanol, l-penten-3-one, hexanal, cw-3-hexenal, 2+3-methylbutanoL /ra/w-2-hexenal, trans-2heptenal, 6-methyl-5-hepten-2-one, cw-3-hexenol, 1 -nitro-2-phenylethane, geranylacetone, 2-isobutylthiazole, and p-ionone) based on their OUV in aqueous solutions as determined by Buttery et al. (1988) and Tandon (1998). The GC peaks for the aroma volatile compounds of interest were quantified in uL/L using standard curves as determined by enrichment of bland tomato homogenate, obtained after roto-

PAGE 132

124 evaporating tomato volatiles for 4 hours at 50°C, confirm this with authentic volatile compound standards (Baldwin et al., 1991b). For EN analysis, approximately 20 g from each frozen tomato sample was placed inside 1 13-mL plastic cups, lidded and thawed in a 25°C water bath. Immediately upon thawing, the lid was removed and the sample cup placed inside the glass vessel of an electronic nose (e-NOSE 4000, Neotronics Scientific, Flowery Branch, GA), controlled by personal computer. EN sampling procedures were the same as described in Chapter 4. Descriptive sensory panel scores for the flavor descriptors were analyzed as complete block design with panelists as blocks and storage temperatures as treatments using GLM procedure of SAS v 6.12 (SAS Institute, Cary NC, 1996). GC aroma volatile compound concentrations, titratable acidity, pH, soluble solids, vitamin C, glucose and fructose concentrations and sucrose equivalents were analyzed by MANOVA and Duncan's Multiple Range Test (a =0.05) was used for mean separations with SAS. Electronic nose sensor outputs were analyzed using standard multivariate discriminant (MVDA) and canonical plot analysis with STATISTICA v 4.5 (Statsoft Corp., Tulsa, OK, 1994). Relationships between instrumental and sensory parameters were explored through the use of correlation matrices (PROC CORR) and stepwise regressions (PROC STEPWISE) using SAS. Results and Discussion Experiment 1. GC analysis for whole fruit homogenate samples of 'Solimar' tomatoes at table-ripe stage showed significant changes in the concentrations of five aroma volatile compounds (acetaldehyde, ethanol, 2+3-methylbutanol, l-nitro-2-

PAGE 133

125 phenylethane and geranylacetone) as a result of harvest maturity and storage temperature (Table 6.1). MG and IG tomatoes responded to low temperature storage inversely, in that the total volatile production (sum of the concentrations for the 16 aroma volatile compounds) increased 4.68% in MG fruit but decreased 15.18% in IG. Concentrations of 2+3-methylbutanol were significantly higher in tomatoes stored at 20°C (2.26 and 1 .79 uL/L for MG and IG, respectively) compared to those stored at 5°C (1.39 and 1.42 uL/L for MG and IG, respectively). Conversely, tomatoes stored at 5°C had significantly higher concentrations of acetaldehyde (MG), ethanol (MG), and l-nitro-2-phenylethane (IG) compared to tomatoes stored at 20°C (Table 6.1). Table 6.1. Aroma volatile concentrations from ripe 'Solimar' whole tomato homogenate samples harvested at mature-green (MG) and immature-green (IG) maturity stages, ripened and stored at 20°C or 5°C for 7 days. Aroma volatile Compounds z Postharvest storage treatments MG + 20°C MG + 5°C IG + 20°C IG + 5°C Acetaldehyde 13.54 b 19.95 a 12.04 b 12.24 b Acetone 0.48 a 0.38 a 0.41 a 0.34 a Methanol 304.13 a 314.41 a 307.41 a 260.60 a Ethanol 15.98 b 20.41 a 16.67 b 15.88 b l-Penten-3-one 0.19a 0.17a 0.17 a 0.178 a Hexanal 8.22 a 7.21 a 9.19 a 7.91 a C/s-3-hexenal 6.18a 6.21 a 4.56 a 6.26 a 2+3-Methylbutanol 2.26 a 1.39 b 1.79 ab 1.42 b 7>aws-2-hexenal 7.51 a 6.38 a 7.03 a 6.57 a rrans-2-heptenal 0.03 a 0.03 a 0.03 a 0.03 a 6-Methyl-5-hepten-2-one 0.55 a 0.52 a 0.42 a 0.55 a C/s-3-hexenol 0.05 a 0.05 a 0.05 a 0.05 a 2-Isobutylthiazole 0.06 a 0.06 a 0.05 a 0.06 a 1 -nitro-2-phenylethane 0.06 ab 0.05 b 0.05 b 0.06 a Geranylacetone 4.26 a 3.32 ab 2.25 c 2.84 be P-ionone 0.09 a 0.06 a 0.06 a 0.11 a Total y 363.58 380.58 362.16 307.18 z Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified.

PAGE 134

126 In general, IG homogenate samples had higher soluble solids than MG samples regardless of storage temperature. Low temperature storage did not induce significant changes in pH or titratable acidity. Vitamin C content, however, was significantly higher in tomatoes stored at low temperature, regardless of their maturity at harvest (Table 6.2) Table 6.2. Compositional parameters from table-ripe 'Solimar' whole tomato, locule and pericarp tissue samples harvested at mature-green (MG) and immature-green (IG) Compositional Parameters z Postharvest storage treatments MG + 20°C MG + 5°C IG + 20°C IG + 5°C Whole Fruit Soluble solids (°Brix) 4.12 a 3.98 a 4.16a 4.28 a pH 4.21 a 4.20 a 4.21 a 4.24 a Titratable acidity (% citric acid) 0.84 a 0.77 a 0.86 a 0.90 a Vitamin C (mg/lOOg fresh wt.) 13.58 be 18.11 a 12.06 c 16.86 ab Locule Tissue Soluble solids (°Brix) 3.72 b 3.86 a 3.98 a 3.88 a PH 4.38 b 4.26 c 4.44 a 4.35 b Titratable acidity (% citric acid) 0.97 a 1.04 a 0.94 a 1.07 a Vitamin C (mg/lOOg fresh wt.) 18.60 be 25.12a 16.77 c 22.93 ab Pericarp Tissue Soluble solids (°Brix) 4.12a 3.82 b 4.28 a 4.22 a PH 4.18a 4.21 a 4.20 a 4.20 a Titratable acidity (% citric acid) 0.74 ab 0.68 c 0.71 be 0.76 a Vitamin C (mg/lOOg fresh wt.) 14.53 ab 16.84 a 12.74 b 14.91 ab z Chemical composition parameter different at the 5% level according to means with different letters across rows are significantly Duncan's Multiple Range Test. Aroma volatile analysis from homogenate samples of 'Solimar' locule tissue stored at 5°C showed significant changes in concentration for of the 16 compounds quantified. There were significant reductions in acetone, cw-3-hexenal and 2+3methylbutanol. Locule tissue from IG tomatoes was found to contain the highest concentrations of l-penten-3-one, hexanal, Zraw.s-2-hexenal and cw-3-hexenol, nonetheless, their concentrations were dramatically reduced following storage at 5°C to the extent of having the lowest concentrations of all four treatments (except cis-3-

PAGE 135

127 hexenol). The total concentration of aroma volatile compounds following low temperature storage decreased 8.55% and 12.44% in MG and IG samples, respectively (Table 6.3). Locule tissue samples from MG tomatoes had the lowest soluble solids content. Locule tissue pH values were significantly lower in tomatoes stored at low temperature compared to those stored at 20°C. In addition, a significant interaction between maturity at harvest and storage temperature was evidenced in locule tissue pH, where IG tomatoes had significantly higher pH values regardless of storage temperature. Even though no significant differences in titratable acidity were shown, a trend of higher acidity was observed in tomatoes stored at 5°C when compared to those stored at 20°C. Vitamin C content was significantly higher in locule tissue from tomatoes stored at 5°C compared to 20°C samples (Table 6.2). GC analysis of pericarp tissue showed significant reductions in 11 of the 16 aroma volatile compounds quantified. Pericarp samples from MG tomatoes stored at 20°C were significantly higher in acetaldehyde, acetone, l-penten-3-one, 2+3methylbutanol, and 6-methyl-5-hepten-2-one, meanwhile IG tomatoes stored at 5°C were lowest. Pericarp samples from tomatoes stored at 20°C had significantly higher concentrations than those stored at 5°C in hexanal, 2+3-methylbutanol, and 6-methyl-5hepten-2-one. In addition, pericarp samples from MG-harvested tomatoes had significantly higher concentrations of 2-isobutylthiazole and geranylacetone compared to IG-harvested fruits. Total aroma volatile production in pericarp tissue from MG fruit decreased by 10.37% following low temperature storage, while IG samples underwent a 13.86% reduction (Table 6.4).

PAGE 136

128 Table 6.3. Aroma volatile concentrations from ripe 'Solimar' locule tissue homogenate samples harvested at mature-green (MG) and immature-green (IG) maturity stages, r ipened and stored at 20°C or 5°C for 7 days. Aroma volatile compounds z Postharvest Storage Treatments MG + 20°C MG + 5°C IG + 20°C IG + 5°C Acetaldehyde 17.43 a 20.51 a 16.45 a 17.15 a Acetone 0.48 a 0.40 b 0.47 a 0.31 c Methanol 144.74 a 132.49 b 136.33 b 123.37 be Ethanol 19.33 a 17.86 a 18.78 a 17.13 a 1 -Penten-3-one 0.12 ab 0.09 b 0.14a 0.07 b Hexanal 3.02 b 1.02 c 4.78 a 1.21 c C/s-3-hexenal 4.89 a 3.09 b 5.59 a 2.89 b 2+3-Methylbutanol 2.66 a 1.77 b 2.95 a 1.58 b 7>ans-2-hexenal 5.99 b 4.44 c 7.55 a 5.18 be 7>aAis-2-heptenal 0.03 a 0.03 a 0.03 a 0.02 a 6-Methyl-5-hepten-2-one 0.52 a 0.47 a 0.56 a 0.42 a C/s-3-hexenol 0.09 b 0.05 c 0.14a 0.13 a 2-Isobutylthiazole 0.06 a 0.07 a 0.07 a 0.07 a 1 -Nitro-2-pheny lethane 0.06 a 0.05 b 0.06 a 0.06 a Geranylacetone 2.69 a 2.51 a 2.72 a 2.45 a p-ionone 0.17a 0.14a 0.08 a 0.16a Total y 202.28 184.985 196.69 172.20 z Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. In general, pericarp tissue had considerably higher soluble solids while lower pH, titratable acidity and vitamin C content compared to the locule tissue from tomatoes of the same maturity and storage temperature treatments. Pericarp tissue samples from MG tomatoes stored at 5°C had significantly lower soluble solids compared to the rest of the treatments. A significant interaction between maturity at harvest and storage temperature was evidenced for pericarp titratable acidity, where MG pericarp samples showed significant reductions in acidity following low temperature storage (0.74 and 0.68% for 20° and 5°C, respectively), while IG samples showed significantly higher acidity in tomatoes stored at 5°C compared to those stored at 20°C (0.71 and 0.76%, respectively)

PAGE 137

129 (Table 6.2). Vitamin C content was generally higher in pericarp samples from tomatoes stored at 5°C compared to those stored at 20°C. Table 6.4. Aroma volatile concentrations from ripe 'Solimar' pericarp tissue homogenate samples harvested at mature-green (MG) and immature-green (IG) maturity stages, Aroma volatile comnounds 2 Postharvest Storage Treatments MG + 20 C MG + 5 C IG + 20°C IG + 5°C ArptfllHphvHf* r\L ^, Ltl lUtl 1 y 16.27 a 12.77 b 1 1.97 b 9.73 b 0.4/ a 0.33 be 0.38 b 0.29 c Methanol "J7Q on a j^to. j / a 17^ 78 0 j / j. lo a 3/O.U4 a Ethanol 18.06 a 18.70 a 17.99 a 18.57 a l-Penten-3-one 0.18a 0.14 b 0.14 b 0.12 b Hexanal 14.09 a 9.73 b 14.90 a 8.80 b C/s-3-hexenal 7.57 a 6.08 a 6.45 a 5.91 a 2+3-Methylbutanol 3.31 a 1.52 c 2.18 b 1.25 c 7>a«s-2-hexenal 8.66 a 6.82 b 6.98 b 5.92 b 7>a/j.s-2-heptenal 0.04 a 0.04 a 0.03 b 0.03 b 6-Methyl-5-hepten-2-one 0.76 a 0.60 c 0.50 b 0.45 c C/s-3-hexenol 0.09 a 0.08 ab 0.07 b 0.09 a 2-Isobutylthiazole 0.09 a 0.07 a 0.05 b 0.05 b 1 -Nitro-2-pheny lethane 0.06 a 0.06 a 0.06 a 0.06 a Geranylacetone 5.92 a 4.92 a 3.66 b 2.64 b (3-ionone 0.11 a 0.10a 0.05 a 0.09 a Total y 455.56 408.32 441.19 380.04 1 Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. In a previous study (Sargent et al., 1997), sensory evaluation of tomato fruit subjected to different postharvest storage conditions showed panelists were able to identify significant flavor differences due to low temperature storage and physiological maturity at harvest. After similar storage conditions (6 days at 5°C), tomato fruit harvested at green stage produced significantly lower levels of volatile compounds (Stern et al., 1994). Longer storage regimes (2°C for 14 days) significantly reduced the levels of hexanal, /raw5-2-hexenal, geranylacetone, l-penten-3-one, methanol, 2+3-methylbutanol

PAGE 138

130 and l-nitro-2-phenylethane (McDonald et al., 1996). Results from this study concur with such reductions in aroma volatile production following low temperature storage. In addition, the interaction between physiological maturity at harvest and susceptibility to low temperature storage was evidenced by greater reductions in aroma volatile production in IG fruit following low temperature storage. On average, pericarp tissue samples showed 123% higher total production of the sixteen volatile compounds than locule gel samples (421.28 uL/L and 189.04 uL/L, respectively). For whole tomato homogenate, the average volatile production was between that of the pericarp and locule tissue (353.38 uL/L), possibly due to the combination of low volatile-producing locule gel and high volatile-producing pericarp in the whole fruit homogenate. The higher aroma volatile production by the pericarp tissue, accompanied by a greater susceptibility to low temperature storage, are of great relevance when considering that pericarp tissue (including columnella) may constitute as much as 60% to 80% of the fruit mass. Nonetheless, the locule gel's contribution should not be downplayed because its fluidity would allow for faster access to sensory receptors than solid tissues during mastication. When volatile compounds were grouped based on their biosynthetic pathways, it was evidenced that MG tomatoes had considerably higher amino acid-derived (2+3methylbutanol, 2-isobutylthiazole and 1 -nitro-2-phenylethane) and carotenoid-derived (acetone, geranylacetone, 6-methyl-5-hepten-2-one, and p-ionone) volatile compounds compared to IG tomatoes. Regardless of harvest maturity or storage temperature, pericarp tissue samples produced an approximately 2-fold greater concentration of lipid-derived (hexanal, c/s-3-hexenal, /ra>w-2-hexenal, /ra/7s-2-heptenal, cw-3-hexenol, l-penten-3-

PAGE 139

131 one) and carotenoid-derived volatile compounds, and similar concentrations of amino acid-derived volatiles compared to locule tissue. The effects of low temperature storage on volatile compound production varied between tomato tissues and biosynthetic pathways. Locule tissue from tomatoes stored at 5°C suffered an average 43% reduction in lipid-derived, 38% reduction in amino acid-derived, and a 9% reduction in carotenoidderived volatiles, when compared to tomatoes stored at 20°C. Similarly, pericarp tissue from tomatoes stored at 5°C suffered a 26% reduction in lipid-derived volatiles, a 46% reduction in amino acid-derived volatiles, and a 21% reduction in carotenoid-derived volatiles, when compared to those stored at 20°C. EN analysis successfully classified the harvest and temperature treatments into four distinct clusters for 'Solimar' whole tomato homogenates (PO.003) (Figure 6.1). The Mahalanobis distance (MD) (the distance between the centroids of two clusters adjusted for probability) between homogenized whole fruit samples was greater between storage temperatures (6.92 and 6.30 units for MG and IG samples, respectively) than between harvest maturity (3.59 and 4.65 units for 20°C and 5°C treatments, respectively). Greater MD values represent greater degree of dissimilarity between tomato samples. MVDA classified whole tomato homogenate samples with accuracy ranging between 81.5 and 100%. MD's between locule tissue samples following 5°C or 20°C storage were greater in IG than in MG fruit (14.87 and 10.589 units, respectively) (Figure 6.2). These MD's were approximately 50% greater than those calculated for whole tomato homogenate. Probabilities for accurate EN classification of locule tissue samples based on their volatiles ranged from 99.90% to 100%. Such a high degree of classification accuracy

PAGE 140

132 indicates the possibility for greater contrasting features in headspace volatiles produced by the different locule tissue samples than by the whole fruit homogenate samples. Corroboration of MVDA results revealed that 75% of individual EN sensors (nine out of twelve sensors) had significantly higher outputs following exposure to locule gel samples from IG that were subsequently stored at low temperature (data not shown). According to the MVDA, the effects of low temperature on pericarp tissue were of slightly lower magnitude when compared to those observed in locule tissue samples (MD of 8.93 and 13.04 units for MG and IG samples, respectively) (Figure 6.3). However, during the individual EN sensor analysis, 92% of the sensors (eleven out of the twelve) had significantly different outputs. Comparable to results observed in whole fruit and locule tissue homogenate, pericarp samples from IG fruit that were subsequently stored at low temperature induced significantly higher sensor outputs. Differences in sensor outputs were found in the different 'Solimar' tomato samples analyzed (whole fruit, locule and pericarp tissue homogenates). The EN sensor outputs were, generally in agreement with the results from the GC volatile profile analysis. Greater MD between classification clusters from MVDA analysis of pericarp tissue and locule tissue samples coincided with greater number of GC aroma volatile compounds showing significant concentration differences, especially following low temperature storage. The differences in EN sensor responses suggest that storage at 5°C induced more dramatic changes in ripe tomato fruit profiles than physiological maturity at harvest. Reductions in the overall production of aroma volatile compounds following low temperature storage were evident in all tissues, however, pericarp homogenate samples seemed to be most affected. Consequently, pericarp tissue had lower total aroma

PAGE 141

133 volatile production than either locule tissue or whole tomato homogenate. It is important to note that, in general, volatile profiles from IG fruit were more disrupted by low temperature storage than MG fruit samples. Experiment 2. After only 2 days storage at different temperatures, 'BHN-189' tomato flavor was significantly affected as noted by sensory panels and corroborated by instrumental measurements. Panelists rated table-ripe 'BHN-189' tomatoes stored at 5°, 10° and 12.5°C significantly lower (PO.05) in ripe aroma, tomato flavor, and significantly higher in off-flavors (5°C stored fruit), when compared to fruit stored at 20°C (Figure 6.4). GC aroma volatile profiles showed that 2 of the 16 compounds quantified had significant concentration differences between temperature treatments. Notably, tomatoes stored at 5°C had the highest concentrations of cz's-3-hexenal and 2isobutylthiazole (Table 6.5). However, 12 of the 16 volatiles showed a trend toward reduced concentration with the lower storage temperatures. Fruit held for 2 days at 12.5° or 20°C were significantly lower in pH compared to those stored at 5° or 10°C (Table 6.6). There were no significant differences in soluble solids content or titratable acidity due to storage temperature. However, glucose levels were significantly higher in samples stored at 5°C when compared to those stored at 12.5° or 20°C. Fructose levels and sucrose equivalents were significantly higher in samples stored at 5° and 10°C when compared to those stored at higher temperatures. Sucrose equivalents were highest in fruit stored at 5° and 10°C. Meanwhile, lycopene content was significantly lower in tomato fruit stored at 20°C.

PAGE 142

134 Table 6.5. Aroma volatile compound concentrations for table-ripe 'BHN-189' tomatoes stored for 2 days at four different temperatures. Storage Temperatures Aroma Volatile Compound z 20°C 12.5°C 10°C 5°C Acetaldehyde 23.97 a 21.65 a 20.42 a 18.25 a Acetone 0.45 a 0.40 a 0.49 a 0.36 a Methanol 263.85 a 287.00 a 295.97 a 309.75 a Ethanol 27.05 a 25.83 a 25.82 a 23.88 a l-Penten-3-one 0.08 a 0.10 a 0.08 a 0.08 a Hexanal 4.17a 3.36 a 2.02 a 2.69 a C/s-3-hexenal 1.14 ab 1.48 a 0.80 b 1.51 a 2+3 -methy lbutanol 1.14a 1.03 a 0.91 a 1.07 a 7>a«.s-2-hexenal 3.00 a 2.51 a 2.65 a 2.70 a 7>a«s-2-heptenal 0.02 a 0.02 a 0.02 a 0.02 a 6-methyl-5-hepten-2-one 0.85 a 0.69 a 0.75 a 0.68 a Cz's-3-hexenol 1.03 a 1.06 a 0.86 a 0.92 a 2-Isobutylthiazole 0.06 be 0.05 c 0.10 b 0.16 a 1 -Nitro-2-phenylethane 0.06 a 0.06 a 0.06 a 0.06 a Geranylacetone 3.18a 2.57 a 3.47 a 2.81 a P-ionone 0.14 a 0.14 a 0.12 a 0.20 a Total y 330.18 347.93 354.52 365.14 z Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. 'BHN-189' tomatoes after 8 days of storage at 5°C were rated significantly lower in ripe aroma, sweetness and tomato flavor and were perceived more sour when compared to those stored at higher temperatures (Figure 6.5). GC analysis showed 13 of 16 compounds had significant concentration differences due to storage temperature treatment (Table 6.7). Tomatoes stored at 20°C had the highest concentrations of acetaldehyde, acetone, ethanol, hexanal, 2+3-methybutanol, /ra>w-2-hexenal, trans-2heptenal, 6-methyl-5-hepten-2-one, cw-3-hexenol, 2-isobutylthiazole, and geranylacetone. Conversely, the concentration of l-nitro-2-phenylethane was highest in tomatoes stored at 5°C. It is important to note that tomatoes stored below 10°C had

PAGE 143

135 significantly lower concentrations of /raws-2-hexenal, cz's-3-hexenol and geranylacetone compared to those stored at or above 12.5°C. 'BHN-189' tomatoes stored at 5°C for 8 days had the lowest pH whereas fruit stored at 20°C had the highest pH (Table 6.6). No significant differences in titratable acidity were evident between temperature treatments. Soluble solids were significantly higher in fruit stored at 10°C, while lowest in fruit stored at 12.5°C. Glucose, fructose and sucrose equivalents were highest in tomato samples stored at 12.5°C. Meanwhile, lycopene content was lowest in tomato fruit stored at 20°C. Table 6.6. Results from chemical composition analyses performed on 'BHN-189' after 2 and 8 days of storage at four different temperatures. Compositional Parameters z Storage Temperatures 20°C 12.5°C 10°C 5°C 2 days storage Titratable acidity (% citric acid) 1.19a 1.19a 1.17a 1.21 a Soluble solids (°Brix) 3.60 a 3.53 a 3.73 a 3.73 a PH 4.55 b 4.55 b 4.61 a 4.62 a Lycopene content (pg/g fresh wt.) 27.7 b 33.1 a 31.4a 37.1 a Glucose content (% fresh wt.) 1.14b 1.05 b 1.21 ab 1.24 a Fructose content (% fresh wt.) 1.27 b 1.18c 1.42 a 1.38 a Sucrose equivalents (% fresh wt.) 3.04 b 2.81 c 3.34 a 3.30 a 8 days storage Titratable acidity (% citric acid) 1.17a 1.03 a 1.11 a 1.21 a Soluble solids (°Brix) 3.73 b 3.37 d 3.87 a 3.60 c PH 4.76 a 4.66 c 4.70 b 4.57 d Lycopene content (pg/g fresh wt.) 19.9 c 27.6 a 28.1 a 23.9 b Glucose content (% fresh wt.) 0.99 c 1.13a 1.02 be 1.06 b Fructose content (% fresh wt.) 1.20 b 1.30 a 1.17b 1.24 ab Sucrose equivalents (% fresh wt.) 2.81 b 3.08 a 2.78 b 2.93 ab Chemical composition parameters between temperature treatments with different letters are significantly different at the 5% level according to Duncan's Multiple Range Test.

PAGE 144

136 Table 6.7. Aroma volatile compound concentrations for table-ripe 'BHN-189' tomatoes stored for 8 days of storage at four different temperatures. Storage Temperatures Aroma Volatile Compound z 20°C 12.5°C 10°C 5°C Acetaldehyde 19.01 a 13.01 b 11.35 be 10.03 c Acetone 0.42 a 0.39 b 0.33 b 0.34 b Methanol 241.42 a 252.76 a 213.27 a 249.86 a Ethanol 19.38 a 15.02 b 15.46 b 14.16b 1 -Penten-3-one 0.12 a 0.10a 0.06 a 0.06 a Hexanal 8.20 a 6.72 b 1.60 c 1.90 c Cw-3-hexenal 2.32 a 4.05 a 2.55 a 3.64 a 2+3-methylbutanol 0.76 a 0.67 a 0.51 b 0.64 ab 7>a/w-2-hexenal 4.52 a 4.26 a 2.68 b 2.86 b 7>ara:-2-heptenal 0.04 a 0.030 b 0.02 c 0.02 c 6-methyl-5-hepten-2-one 1.07 a 0.52 c 0.85 b 0.87 b C/s-3-hexenol 0.92 a 0.67 b 0.47 c 0.34 c 2-Isobutylthiazole 0.14 a 0.09 b 0.06 c 0.09 b 1 -Nitro-2-phenylethane 0.07 ab 0.07 b 0.07 b 0.08 a Geranylacetone 4.72 a 3.62 b 2.55 c 2.45 c (3-ionone 0.16a 0.17a 0.16 a 0.06 b Total* 303.25 302.14 251.97 287.48 z Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. In the other experiment, 'Solimar' tomatoes held at 5°C for 4 days were rated significantly lower in ripe aroma, sweetness, tomato flavor, and significantly higher in sourness compared to those stored at the higher temperatures (Figure 6.6). GC analyses showed that 4 of the 16 aroma volatile compounds (hexanal, 2+3-methylbutanol, trans-2heptenal and 2-isobutylthiazole) had significantly lower concentrations in samples stored at 5°C when compared to those stored at 20°C or 12.5°C (2+3-methylbutanol). All fruit stored below 20°C exhibited lower levels of 2-isobutylthiazole (Table 6.8).

PAGE 145

137 Table 6.8. Aroma volatile compound concentrations for table-ripe 'Solimar' tomatoes stored for 4 days of storage at four different temperatures. Storage Temperatures Aroma Volatile Compound z 20°C 12.5°C 10°C 5°C Acetaldehyde 28.84 a 27.82 a 24.52 a 32.14 a Acetone 0.41 a 0.38 a 0.51 a 0.37 a Methanol 245.55 a 327.82 a 309.26 a 312.50 a Ethanol 21.18a 20.46 a 25.79 a 23.16a 1 -Penten-3-one 0.10a 0.10a 0.09 a 0.10a Hexanal 4.19a 3.83 ab 3.67 ab 2.89 b C/s-3-hexenal 1.50 a 1.75 a 1.56 a 1.38a 2+3-methylbutanol 0.53 b 1.19a 0.79 b 0.48 b 7ra/w-2-hexenal 3.16a 3.25 a 3.37 a 3.34 a TVaws-2-heptenal 0.03 a 0.025 ab 0.02 b 0.02 b 6-methyl-5-hepten-2-one 0.88 a 0.70 a 0.69 a 0.64 a C/s-3-hexenol 1.03 a 1.02 a 0.90 a 0.99 a 2-Isobutylthiazole 0.36 a 0.20 b 0.19b 0.21 b 1 -Nitro-2-pheny lethane 0.07 a 0.08 a 0.08 a 0.07 a Geranylacetone 3.68 a 3.49 a 3.47 a 3.29 a (3-ionone 0.16 a 0.17a 0.15 a 0.12a Total* 311.67 392.30 374.76 381.71 * Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. There were no significant differences in pH or titratable acidity between temperature treatments of 'Solimar' tomatoes after 8 days of storage (Table 6.9). However, soluble solids were significantly higher in fruit stored at 5°C. Glucose content was highest in fruit stored at 5°C; in contrast, these same fruit had the lowest fructose levels. Sucrose equivalents were significantly higher in fruit stored at 5°C when compared to those stored at higher temperatures. In general, lycopene levels decreased in all 'Solimar' tomatoes after 8 days of storage; however, they attained the lowest level in 5°C stored fruit.

PAGE 146

138 Table 6.9. Results from chemical composition analyses for table-ripe 'Solimar' tomatoes stored for 4, 8, and 12 days of storage at four different temperatures. Compositional Parameters z Storage Temperatures 20°C 12.5°C 10°C 5°C 4 days storage Titratable acidity (% citric acid) 1.22 c 1.30 be 1.33 b 1.46 a Soluble solids (°Brix) 3.87 c 4.40 b 4.13 b 4.70 a pH 4.54 a 4.51 a 4.52 a 4.52 a Lycopene content (ug/g fresh wt.) 23.2 b 16.3 c 34.9 a 22.2 b Glucose content (% fresh wt.) 1.82 a 1.82 a 1.63 b 1.98 a Fructose content (% fresh wt.) 1.19c 1.36 a 1.27 b 1.39 a Sucrose equivalents (% fresh wt.) 3.41 b 3.71 a 3.41 b 3.86 a 8 days storage Titratable acidity (% citric acid) 1.36 a 1.33 a 1.33 a 1.40 a Soluble solids (°Brix) 4.30 b 4.30 b 4.10 b 4.63 a PH 4.52 a 4.50 a 4.51 a 4.49 a Lycopene content (ug/g fresh wt.) 21.0a 20.8 a 23.1 a 17.5 b Glucose content (% fresh wt.) 1.36 b 1.47 b 1.36 b 1.80 a Fructose content (% fresh wt.) 1.65 ab 1.72 a 1.69 a 1.57 b Sucrose equivalents (% fresh wt.) 3.86 b 4.06 b 3.93 b 4.78 a 1 2 days storage Titratable acidity (% citric acid) 1.15 be 1.11 b 1.42 a 1.28 ab Soluble solids (°Brix) 3.60 c 3.67 c 4.43 b 4.00 a pH 4.48 b 4.52 a 4.46 c 4.49 b Lycopene content (ug/g fresh wt.) 15.9a 20.4 a 20.3 a 16.0 a Glucose content (% fresh wt.) 0.83 c 1.04 b 1.07 b 1.37 a Fructose content (% fresh wt.) 1.24 c 1.26 c 1.35 b 1.67 a Sucrose equivalents (% fresh wt.) 2.76c 2.95 be 3.13 b 3.91 a z Chemical composition parameters between temperature treatments with different letters are significantly different at the 5% level according to Duncan's Multiple range Test. Similarly to 4 days storage, 'Solimar' tomatoes held at 5°C for 8 days were still rated significantly lower in ripe tomato aroma and tomato flavor, while significantly higher in sourness, when compared to higher temperatures (Figure 6.7). Conversely, fruit stored at 20°C were rated higher in sweetness while lower in sourness compared to chilled fruit. GC analysis showed eight aroma volatile compounds (methanol, l-penten-3one, hexanal, 2+3-methylbutanol, /r
PAGE 147

139 phenylethane, and geranylacetone) with significantly different concentrations between the temperature treatments (Table 6.10). Among these volatile compounds, fruit stored at 12.5° or 20°C produced the highest levels, except for l-penten-3-one, compared to other temperature treatments. Table 6.10. Aroma volatile compound concentrations for table-ripe 'Solimar' tomatoes stored for 8 days at four different temperatures. Storage Temperature Aroma Volatile Compound z 20°C 12.5°C 10°C 5°C Acetaldehyde 19.77 a 15.88 a 18.28 a 30.13 a Acetone 0.42 a 0.31 a 0.30 a 0.61 a Methanol 251.57 c 323.88 a 277.08 be 303.26 ab Ethanol 20.45 a 19.28 a 24.88 a 22.68 a l-Penten-3-one 0.11 ab 0.10 ab 0.09 b 0.13a Hexanal 7.81 a 5.39 ab 3.97 b 2.72 b C7s-3-hexenal 1.60 a 2.88 a 1.64 a 1.33 a 2+3-methylbutanol 0.57 be 1.21 a 0.85 b 0.42 c rra«s-2-hexenal 3.33 a 3.84 a 4.51 a 3.29 a 7>aws-2-heptenal 0.03 a 0.02 b 0.02 b 0.02 b 6-methyl-5-hepten-2-one 0.82 a 0.63 a 0.61 a 0.59 a Cw-3-hexenol 0.83 a 0.83 a 0.69 a 0.86 a 2-Isobutylthiazole 0.36 a 0.19 b 0.18b 0.19b 1 -Nitro-2-phenylethane 0.076 ab 0.081 a 0.078 ab 0.064 b Geranylacetone 4.06 a 3.07 b 2.12 c 2.10c (3-ionone 0.15 a 0.19 a nd 1 0.11 a Total y 311.82 377.59 335.29 368.51 1 Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. 1 Aroma compound concentrations not detected. After 12 days, 'Solimar' tomatoes stored at 5°C were still rated lowest in ripe aroma, sweetness, and tomato flavor (Figure 6.8). In addition to 5°C-stored fruit, those stored at 10° and 12.5°C became significantly lower in ripe aroma and sweetness ratings (10°C fruit), and all chilled fruit were rated significantly higher in green/grassy flavor and sourness compared to those held constantly at 20°C. GC analysis of aroma volatile

PAGE 148

140 compounds identified 13 out of the 16 compounds quantified (acetaldehyde, methanol, ethanol, hexanal, c/s-3-hexenal, 2+3-methylbutanol, trans-2-hexeml, /ra«5-2-heptenal, 6methyl-5-hepten-2-one, cw-3-hexenol, 2-isobutylthiazole, l-nitro-2-phenylethane and pionone) with significant concentration changes as a result of the prolonged exposure to the different temperature treatments (Table 6.11). As with shorter storage periods, the concentrations of hexanal, 1 -nitro-2-phenylethane and P-ionone were highest in fruit stored at 20°C. Tomatoes stored below 10°C had higher titratable acidity and those stored at 12.5°C had higher pH and soluble solids content compared to fruit from other treatments (Table 6.9). Meanwhile glucose, fructose, sucrose equivalents and soluble solids were highest in tomatoes stored at 5°C. Tomato flavor and aroma were significantly affected by low temperature (<12.5°C) for as short as two days duration for 'BHN-189' and four days for 'Solimar' tomatoes. In general, trained panelists found tomatoes stored at these temperatures to have lower tomato flavor and ripe aroma. Significantly higher off-flavor ratings were found in 'BHN-189' tomatoes after 2 days of storage at 5°C. In contrast 'Solimar', tomatoes stored at 5°C during all storage treatment times, and below 20°C after 12 days storage, were rated significantly lower in sweetness, while significantly higher in sourness. The faster ripening rates and metabolic activity associated with tomatoes stored at 20°C were not detrimental to sensory qualities. Fruit firmness was not significantly affected by the storage temperatures in 'BHN-189' fruit, and 'Solimar' tomatoes stored at 20°C only became significantly softer than those stored at lower temperatures after 12 days of storage (data not shown).

PAGE 149

141 Table 6.11. Aroma volatile compound concentrations for table-ripe 'Solimar' tomatoes Aroma Volatile Compound z Storage Temperature 20°C 12.5°C 10°C 5°C Acetaldehyde 15.91 a 22.76 ab 26.76 a 16.58 b Acetone 0.37 a 0.43 a 0.50 a 0.34 a Methanol 266.54 b 324.76 a 342.52 a 264.87 b Ethanol 17.96 b 26.36 ab 35.36 a 23.08 b 1 -Pent en-3 -one 0.13 a 0.11 a 0.11 a 0.08 a Hexanal 8.45 a 6.28 b 2.94 c 0.92 d C/s-3-hexenal 1.71 a 2.13 a 1.10b 0.81 b 2+3-methylbutanol 0.69 b 1.30 a 1.23 a 0.33 c 7>ans-2-hexenal 3.38 b 4.38 a 3.90 ab 2.29 c 7>aws-2-heptenal 0.03 a 0.03 a 0.02 b 0.02 b 6-methyl-5-hepten-2-one 0.87 b 1.05 a 0.84 b 0.58 b Cw-3-hexenol 0.83 b 1.22 a 1.27 a 0.80 b 2-Isobutylthiazole 0.24 b 0.30 a 0.24 b 0.13 c 1 -Nitro-2-phenylethane 0.07 b 0.08 a 0.08 a 0.08 a Geranylacetone 3.37 a 3.04 ab 2.54 b 2.63 b p-ionone 0.14 a 0.05 b 0.10 ab 0.08 ab Total y 320.67 394.28 419.51 313.62 across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. Chemical composition analyses revealed no consistent pattern that could explain sensory taste descriptor ratings (sweetness and sourness) given to tomatoes stored at the different temperatures. After 12 days storage, tomatoes held near room temperature (20°C) were still rated superior in tomato flavor, aroma and sweetness than those stored at lower temperatures, despite having higher metabolic activity and carbohydrate reserve consumption. Tomatoes stored at 20°C had significantly lower individual sugars (glucose and fructose) and titratable acidity contents compared to those stored at 5° or 10°C. 'BHN-189' tomatoes appeared to be more susceptible to low temperature storage than 'Solimar' tomatoes. After 8 days of storage, 'BHN-189' fruit had as many aroma volatile compounds showing significant differences as 'Solimar' fruit following 12 days

PAGE 150

142 of storage. McDonald et al. (1996; 1998) reported that, after 14 days of storage at 2°C, tomatoes showed significant changes in the concentrations of numerous important aroma volatile compounds without the presence of other visual CI symptoms. In this study, after only 8 days at 5°C, 'Solimar' and 'BHN-189' had significant changes in concentration for all volatile compounds reported by McDonald et al. (1996). It is important to note that hexanal, 2+3-methylbutanol, /ra/w-2-heptenal and 2isobutylthiazole concentrations were significantly lower in 'Solimar' and 'BHN-189' fruit stored below 20°C for all storage periods except for 'BHN-189' after 2 days. Three important aroma compounds were found to be positively correlated with ripe aroma sensory ratings: c/s-3-hexenal (r=0.75), 6-methyl-5-hepten-2-one (r=0.72) and p-ionone (r=0.67). In this study, the effects of low temperature storage (>10°C) on tomato aroma volatiles was consistently evidenced by considerable reductions in the concentrations of lipid-derived volatiles for both 'BHN-189' and 'Solimar' tomatoes at all storage times. The magnitude of changes in volatile concentrations due to low temperature storage, were also influenced by cultivar and time of exposure. In 'BHN-189', levels of lipidderived volatiles were reduced by -24% and -50% after 2 and 8 days of storage at or below 10°C, when compared to those stored at 20°C. For 'Solimar' tomatoes stored at or below 10°C, reductions in lipid-derived volatiles were -9%, 30% and 36% after 4, 8 and 12 days storage, respectively, when compared to those stored at 20°C. Storage temperature did not affect concentrations of amino acid-derived volatiles for 'BHN-189' tomatoes after 2 days storage and 'Solimar' tomatoes after 4 days storage. However,

PAGE 151

143 amino acid-derived volatile concentrations were slightly higher in 'Solimar' samples stored at or below 12.5°C for 8 and 12 days. Carotenoid-derived volatiles suffered slight reductions in concentration in samples stored at or below 10°C for 'BHN-189' after 2 days storage and for 'Solimar' after 4 days storage, when compared to those stored at 20°C (3.4% and 8.5% reductions, respectively). Concentrations of these volatiles were further reduced with longer storage times at or below 10°C; in 'BHN-189' tomatoes -42% reductions were documented after 8 days of storage, while in 'Solimar' tomatoes concentrations were -45% lower after 8 days storage when compared to those stored at 20°C. Fresh-market tomato acceptability is greatly affected by perceived sweetness and sourness (Jones and Scott, 1984). Another study indicated that panelists preferred tomatoes with intermediate sourness in combination with relatively high sweetness (Baldwin et al., 1998). The effects of sugars on sourness ratings are considerably less important than the effects of acids on sweetness ratings (Petro-Turza, 1987). Furthermore, Jones and Scott (1984) showed that, above a certain sugar threshold level, the impact of additional sugar concentrations on perceived sweetness may plateau. Such observations help understand how tomatoes stored at low temperature (5° or 10°C), with higher sugar levels and higher acidity, would actually be perceived by panelists to be more sour and less sweet. Also, in tomatoes with citric acid contents of approximately 0.80%, Malundo et al. (1995) found that increasing sugar concentrations could lead to improved flavor quality, although such benefits were not clearly achievable when acidity levels were higher. In the present experiments, acidity in 'BHN-189' and 'Solimar' fruit approached 0.80% when stored at 20°C for 8 to 12 days (0.99% and 0.87%, respectively).

PAGE 152

144 Therefore, any increased fructose concentrations during storage, such as those observed in 'Solimar' tomatoes stored continuously at 20°C, could indeed enhance sweetness perception in tomatoes. EN analysis separated the 'BHN-189' tomato samples into four distinct clusters after only 2 days of storage at different temperatures (Figure 6.9). The MD between clusters ranged between 5.73 and 10.50 units. MVDA classified tomato samples according to their temperature treatment with 96% to 100% accuracy based on volatile differences detected by EN after only two days of storage. After 8 days of storage, EN analysis separated the temperature treatments into distinct clusters, with MD's between clusters ranging between 8.8 and 18.58 units (Figure 6.10). MVDA classification of tomatoes from the different treatments had 100% accuracy after 8 days of storage, probably due to the greater magnitude of changes in volatile compounds, as previously noted by GC volatile profile differences. EN sensor output analysis using MVDA classified ripe, 'BHN-189' tomato samples into distinct clusters by storage temperature following two days of storage. In feet, increasing MD values from the EN analysis concurred with an increasing number of aroma volatile compounds that showed significant differences. This increased number of compounds with significant differences in concentration over time might reflect events occurring to the remaining 380+ volatile compounds present in the aroma/flavor bouquet characteristic of ripe tomatoes. There is considerable debate in relation to the contribution of any specific volatile compound to tomato aroma. Nonetheless, the ability of the EN sensor array to detect an increasing dissimilarity between these samples

PAGE 153

145 corroborates the contention that greater quantitative and qualitative changes occurred in the aroma volatile profile of tomatoes. Similar EN analysis performed on 'Solimar' tomato samples successfully separated treatments into four distinct clusters after 4 days of storage at different temperatures (Figure 6.11). MD values ranged between 3.74 and 12.23 units. Crossvalidation of the temperature treatments with EN sensor output classification had 83.7% to 100% accuracy. Following 8 days of storage, EN analysis classified the temperature treatments into clusters with a greater separation between treatments (Figure 6.12). MD values ranged from 4.79 units to 53.35 units, with classification accuracy ranging from 99.5% to 100%. EN analysis classified 'Solimar' tomatoes stored for 12 days with a greater degree of separation than those stored for either 4 or 8 days (Figure 6.13). MD values ranged between 52.47 units and 154.79 units, and crossvalidation accuracy based on EN sensor outputs was 100% in all cases. Experiment 3. Storage pre-treatments for 2 days at 38°C as a means to alleviate CI symptoms did not prove to be an effective method to suppress flavor changes during subsequent storage at 5°C for 7 days. Following this storage regime, 'Trust' tomatoes had significantly lower sensory ratings for ripe aroma, sourness and tomato flavor compared to those samples stored continuously at 20°C. Tomatoes stored continuously at 20°C or pre-treated at 38°C for 2 days without subsequent storage at 5°C were rated similar in ripe aroma, sweetness and tomato flavor, to those stored continuously at 20°C (Figure 6.14).

PAGE 154

146 'Trust' tomatoes stored at 20° or 38°C prior to 7 days at 5°C showed significant changes in 14 of 16 compounds quantified. Regardless of pre-treatment temperature (20° or 38°C), storage at 5°C resulted in reductions in hexanal, c/.s-3-hexenal, 2+3methylbutanol, /ra«s-2-hexenal and geranylacetone. However, concentrations of acetaldehyde, ethanol, hexanal, cw-3-hexenal, 2+3-methylbutanol, frww-2-hexenal, /ra/i.s-2-heptenal, 6-methyl-5-hepten-2-one, cw-3-hexenol, 2-isobutylthiazole and geranylacetone were significantly reduced by pre-treatment at 38°C when compared to those stored at 20°C. Furthermore, tomatoes pre-treated at 38°C and subsequently stored at 5°C were lower in the concentrations of all compounds showing significant differences, except for acetaldehyde and c/s-3-hexenol (Table 6.12). 'Trust' tomatoes pre-treated at 38°C for 2 days, stored at 5°C for 7 days, and transferred to 20°C for an additional 7 days had the lowest sensory panel ratings for ripe aroma, sweetness and tomato flavor, and the highest ratings for off-odors, sourness, green/grassy notes and off-flavors (Figure 6.15). Tomatoes pre-treated at 38°C without subsequent storage at 5°C and those pre-treated at 20°C then stored at 5°C, had ratings for ripe aroma, sweetness and tomato flavor similar to those samples stored continuously at 20°C. Significant differences in 12 of 16 compounds quantified still persisted after an additional 7 days of storage at 20°C, although, dramatic changes in the volatile profiles from the various treatments were evident (Table 6.13). Regardless of pre-treatment temperature, tomatoes stored at 5°C were lower in acetone, while significantly higher in cw-3-hexenal, 2+3-methylbutanol, /ra/w-2-hexenal, 6-methyl-5-hepten-2-one, cw-3hexenol, 2-isobutylthiazole, 1 -nitro-2-phenylethane and 0-ionone.

PAGE 155

147 Table 6.12. Aroma volatile compound concentrations for ripe 'Trust' tomatoes stored at 5°C for 7 days plus 1 day at 20°C with or without high temperature pre-treatment (2 days at 38°C). Aroma volatile compounds z Pre-treatment temperature 20°C 38°C Storage temperature 20°C 5°C 20°C •J V_/ Acetaldehyde 40.50 b 44.42 a 25 31 c 26 45 c Acetone 0.41 a 0.27 a 0 32 a 0 23 h Methanol 290.11 a 264 46 a 258 12 a 1 95 47 h Ethanol 46.52 a 50.56 a 23.92 b 30.52 b l-Penten-3-one 0.05 a 0.04 ab 0.04 a 0.02 b Hexanal 3.67 a 0.51 c 1.47 b nd x Cw-3-hexenal 4.21 a 2.53 b 1.95 b 0.85 c 2+3-Methylbutanol 2.39 a 0.55 b 0.34 c 0.22 c 7>a«s-2-hexenal 3.33 a 2.31 be 2.40 b 1.77 c 7>ans-2-heptenal 0.16a 0.04 b 0.04 b 0.02 b 6-Methyl-5-hepten-2-one 0.80 a 0.30 b 0.28 b 0.24 b C/s-3-hexenol 0.49 a 0.18a 0.13 b 0.15 a 2-Isobutylthiazole 0.06 nd nd nd 1 -nitro-2-phenylethane 0.07 a 0.04 a 0.03 a 0.03 a Geranylacetone 1.00 a 0.33 b 0.23 b nd (3-ionone 0.07 a 0.03 a 0.06 a 0.05 a Total y 393.46 366.58 314.99 256.28 1 Aroma volatile compound concentration means (uL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the concentrations of the sixteen compounds quantified. 'Aroma compound concentrations not detected. Tomatoes pre-treated with high temperature, then stored at chilling temperature, had the highest concentrations of 2+3-methylbutanol and 2-isobuthylthiazole (Table 6.13). After the additional 7 days of storage at 20°C the production of volatiles decreased by approximately 10% in stored continuously at 20°C, it increased by about 10% in tomatoes stored at 5°C for 7 days or pre-treated with 38°C for 2 days. Tomatoes that were pre-treated with 38°C temperature for 2 days followed by 7 days at 5°C had a 66.5% increase in total volatile production following the additional 7 days of storage at 20°C.

PAGE 156

148 Table 6.13. Aroma volatile compound concentrations for ripe 'Trust' tomatoes stored at 5°C for 7 days and 8 day at 20°C with or without high temperature pre-treatment (2 days at 38°C). Aroma volatile compounds z Pre-treatment temperature 20°C 38°C Storage temperature 20°C 5°C 7o°r cop Acetaldehyde 26.29 a 20 71 h ?n 1 7 h — — . 1 O D Acetone 0.37 a 0.30 b 0.34 a 0 11 h Methanol 256.62 a 292 56 a 286 65 a 115 16 a Ethanol 63.01 a 77.37 a 31.34 a 59 26 a l-Penten-3-one 0.06 b 0.07 a 0.04 b 0.05 b Hexanal 1.98 a 2.79 a 2.29 a 1.42 b Cw-3-hexenal 0.94 c 2.41 a 1.04 c 1.90 b 2+3-Methylbutanol 0.54 c 0.68 b 0.45 c 0.94 a 7raws-2-hexenal 1.96 b 2.65 a 2.09 b 2.42 a 7>a«s-2-heptenal 0.07 a 0.09 a 0.07 a 0.11 a 6-Methyl-5-hepten-2-one 0.60 b 1.25 a 0.69 b 1.08 ab C/s-3-hexenol 0.14 b 0.53 a 0.20 b 0.53 a 2-Isobutylthiazole nd 1 0.07 b nd 0.13 b 1 -Nitro-2-phenylethane 0.04 b 0.06 a 0.05 ab 0.06 a Geranylacetone 0.52 a 1.14a 0.73 a 0.80 a p-ionone 0.02 a 0.12 a 0.08 a 0.12a Total y 353.14 402.82 346.23 426.64 1 Aroma volatile compound concentration means (nL/L) with different letters across rows are significantly different at the 5% level according to Duncan's Multiple Range Test. y Total aroma volatile production represents the sum of the sixteen compounds quantified. "Aroma compound concentrations not detected. Due to the effects of high temperature and low temperature on tomato ripening, it is difficult to concisely separate the effects of storage temperatures on tomato flavor and aroma volatile profiles without addressing possible reversibility of CI symptoms. 'Trust' tomato samples subjected to 38°C pre-treatment and/or 5°C storage had drastic reductions in all aroma volatile compounds quantified. However, following an additional 7 days of storage at 20°C aroma volatile production increased to levels comparable (and sometimes higher) than those stored continuously at 20°C. This apparent recovery of volatile production was specially evident in tomatoes pre-treated at 38°C for 2 days then stored at 5°C for 7 days, in which total aroma volatile production increased from -256

PAGE 157

149 uL/L to -427 uL/L. Most surprising were the sensory ratings for tomatoes stored at 5°C without high temperature pre-treatment; after 7 days at 20°C they were rated significantly higher in ripe aroma and tomato flavor than those exposed to high temperature preconditioning. McDonald et al. (1998) found 10 of 15 aroma volatile compounds decreased following storage at chilling temperatures (2°C for 14 days), whereas heattreated tomatoes (38°C for 2 days) had volatile concentrations between those for chilled and non-chilled controls. Finally, EN analysis performed on 'Trust' tomato samples subjected to high temperature pre-treatment prior to storage at 5°C for 7 days found significant treatment differences in headspace volatiles and classified samples into 4 distinct clusters. There was greater proximity between clusters for samples stored at 5°C, regardless of pretreatment temperature. In fact, MD between samples stored at 5°C was -38% less than that between samples stored at 20°C (5.39 and 8.80 units, respectively) (Figure 6.16). Classification of 'Trust' tomato samples based on EN sensor outputs had better than 99.6% accuracy in all cases. Following an additional 7 days storage at 20°C to document any reversibility in the effects of either high and/or low temperature treatments, EN still classified samples into four distinct clusters. A slight overlap was evident between high temperature pretreated samples, regardless of storage temperature (Figure 6.17). Classification accuracy based on EN sensor outputs was greater than 96.4% for all treatments, except for samples pre-treated for 2 days at 38°C, in which one tomato sample was mis-classified. Throughout all EN analyses in this study, there was an apparent relationship between individual EN sensor outputs and the GC aroma volatile compounds quantified.

PAGE 158

150 Responses from all 12 EN sensors were negatively correlated to the concentrations of acetaldehyde, methanol and 2-isobutylthiazole (r = -0.46 to -0.50), while positively correlated to the concentrations of cw-3-hexenal (r = 0.49 to 0.50)(Table 6.14). Surprisingly, concentrations of l-nitro-2-phenylethane were highly correlated to the responses from all EN sensors (r = 0.97 to 0.98). In addition, EN sensor outputs also showed significant correlations with sensory ratings for sourness and green/grassy flavor, where outputs from all 12 sensors were negatively correlated with sensory ratings (r = 0.45 to -0.58). On the other hand, sensor outputs were highly correlated within themselves (r 2 from 0.80 to 0.99), thus supporting the contention of non-specificity among EN sensors. Previous studies have shown a relationship between volatile compounds and perceived sweetness or sourness (Watada and Aulenbach, 1979; Stevens et al., 1977). Baldwin et al. (1998) reported that c/s-3-hexenal concentrations correlated negatively with sweetness sensory ratings, while acetone negatively and hexanal positively correlated with tomato sourness ratings. In these tests, significant changes in the concentrations of important aroma volatile compounds in table-ripe tomatoes could partially explain the significant differences in descriptive sensory ratings for tomatoes stored at the four temperature treatments. The number of aroma volatile compounds showing significant differences increased dramatically with increasing storage time. Increased differences in volatile profiles with longer chilled storage periods below 20°C would be explained by a cumulative effect of CI similar to that reported for visual CI symptoms.

PAGE 159

151 Table 6.14. Correlation coefficients between electronic nose sensor outputs and compositional parameters or sensory panel descriptor ratings from 'BHN-189' and ' Solimar' tomatoes stored at four different temperatures for up to 12 days. EN Sensor Sensory Descriptor Aroma Volatile Compound Type 301 Sourness r = -0.48 z Methanol r = -0.46 C/s-3-hexenal r = 0.50 1 -Nitro-2-phenylethane r = 0.98 Type 298 Sourness r = -0.54 Green/Grassy r = -0.47 Acetaldehyde r = -0.46 Methanol r = -0.49 C/s-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.98 Type 297 Sourness r = -0.53 Green/Grassy r = -0.46 Methanol r = -0.48 C/5-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.97 Type 283 Sourness r = -0.55 Green/Grassy r = -0.47 Acetaldehyde r = -0.47 Methanol r = -0.49 C/s-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.98 Type 278 Sourness r = -0.5 1 Green/Grassy r = -0.45 Methanol r = -0.46 C/s-3-hexenal r = 0.50 l-Nitro-2-phenylethane r = 0.98 Type 264 Sourness r = -0.56 Green/Grassy r = -0.47 Acetaldehyde r = -0.47 Methanol r = -0.48 C/s-3-hexenal r = 0.48 l-Nitro-2-phenylethane r = 0.97 Type 263 Sourness r = -0.58 Green/Grassy r = -0.48 Methanol r = -0.49 C/s-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.98 Type 262 Sourness r = -0.52 Green/Grassy r = -0.46 Methanol r = -0.49 C/s-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.98 Type 261 Sourness r = -0.52 Green/Grassy r = -0.45 Methanol r = -0.48 C7s-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.97 Type 260 Sourness r = -0.50 Green/Grassy r = -0.44 Acetaldehyde r = -0.47 Methanol r = -0.49 Cw-3-hexenal r = 0.49 l-Nitro-2-phenylethane r = 0.98 Type 259 Sourness r = -0.54 Green/Grassy r -0.46 Methanol r = -0.46 Cw-3-hexenal r = 0.50 l-Nitro-2-phenylethane r = 0.98 Type 258 Sourness r -0.54 Green/Grassy r = -0.47 Methanol r = -0.48 C/s-3-hexenal r = 0.48 1 -Nitro-2-phenylethane r = 0.97 z Correlation coefficients were significant at the 5% level according to F test statistics.

PAGE 160

152 Individual aroma volatile compounds and compositional parameters also correlated significantly to sensory ratings. Ripe tomato pH was negatively correlated to off-odor, sourness and green/grassy sensory ratings (r = -0.54, -0.64 and -0.55, respectively), meanwhile, positively correlated to sweetness ratings (r = 0.53) (Table 6.15). Soluble solids content and sucrose equivalents were positively correlated to offodor (r = 0.48 and 0.49), sourness (r = 0.58 and 0.54), and off-flavor (r = 0.49 and 0.62) sensory ratings, respectively. Titratable acidity was negatively correlated to sweetness ratings (r = -0.48) while, positively correlated to sourness, off-odor and off-flavor sensory ratings (r = 0.66, 0.53 and 0.55, respectively). Concentrations of hexanal, trans-2heptenal, geranylacetone and P-ionone were positively correlated to ripe aroma ratings (r = 0.68, 0.51, 0.56, and 0.56, respectively). Meanwhile, concentrations of acetaldehyde, acetone and methanol were positively correlated to off-odor ratings (r = 0.53, 0.59 and 0.58, respectively). Sensory ratings for sweetness were positively correlated to the concentrations of hexanal, /raw.s-2-heptenal, 6-methyl-5-hepten-2-one, and geranylacetone (r = 0.59, 0.70, 0.49, and 0.66, respectively), while negatively correlated to methanol concentrations (r = -0.47) (Table 6.15). Conversely, sourness ratings were negatively correlated to the concentrations of hexanal, £ra/w-2-heptenal, geranylacetone, and P-ionone (r = -0.51, 0.57, -0.54, and -0.53, respectively), while positively correlated to methanol concentrations (r = 0.50). Hexanal and geranylacetone concentrations were positively correlated to tomato flavor ratings (r = 0.46 and 0.53), while l-nitro-2-phenylethane concentrations were positively correlated to green/grassy ratings (r = 0.47).

PAGE 161

153 Table 6.15. Correlation coefficients between descriptive sensory panel ratings and compositional parameters from 'BHN-189' and 'Solimar' tomatoes stored at four different temperatures for up to 12 days. Sensory Descriptors Compositional Parameter Aroma Volatile Compound Ripe Aroma Hexanal r = 0.68 z 7>aMs-2-heptenal r = 0.51 Geranylacetone r = 0.56 (3-ionone r = 0.46 Off-odor pH r = -0.54 Soluble Solids r = 0.48 Titratable Acidity r = 0.53 Sucrose Equivalents r = 0.49 Acetaldehyde r = 0.53 Acetone r = 0.59 Methanol r = 0.58 Sweetness pH r = 0.53 Titratable Acidity r = -0.48 Methanol r = -0.47 Hexanal r = 0.59 7>an,s-2-heptenal r = 0.70 6-Methyl-5-hepten-2-one r = 0.49 Geranylacetone r = 0.47 Sourness pH r = -0.64 Soluble Solids r = 0.58 Titratable Acidity r = 0.66 Sucrose Equivalents r = 0.54 Methanol r = 0.50 Hexanal r = -0.51 7>a/w-2-heptenal r = -0.57 Geranylacetone r = -0.54 (3-ionone r = -0.53 Green/Grassy pH r = -0.55 1 -nitro-2-phenylethane r = 0.47 (3-ionone r = -0.45 Tomato Flavor Hexanal r = 0.46 Geranylacetone r = 0.53 Off-flavor Soluble Solids r = 0.49 Titratable Acidity r = 0.55 Sucrose Equivalents r = 0.62 Methanol r = 0.46 z Correlation coefficients were significant at the 5% level according to F test statistics. Another approach to rate the importance of individual volatile compounds is to contrast the concentration of an individual volatile compound present in a fruit with the concentration threshold for sensory perception, referred to as the odor unit value (OUV) (Guadagni et al., 1963; Buttery et al., 1988; Tandon, 1998). Nonetheless, because of possible interactions with other compounds, odor unit values might not give a clear indication of an individual aroma compound's contribution when in a complex mixture (Baldwin et al., 1998). Cw-3-hexenal is considered of principal importance due to its high

PAGE 162

154 OUV. Odor threshold studies described c/s-3-hexenal in aqueous solutions as imparting a "fresh green" aroma (Kazeniac and Hall, 1970). More relevant to tomato sensory perception, cw-3-hexenal was described as having a "tomato/citrus" character when added to bland tomato homogenate (Tandon, 1998). Reductions in the concentration of cw-3-hexenal after 6 days of storage at 5°C were reported by Stern et al. (1994), and also were obtained in this study for 'Solimar' tomatoes stored at 5°C. Close relationships between fra/w-2-heptenal concentrations with ripe aroma and sweetness ratings deserve further investigation since there is no published information on this compound's sensory attributes. Low temperature storage consistently induced changes in 2-isobutylthiazole concentrations, a significant observation since this compound has only been isolated from tomato fruits (Petro-Turza, 1987). 2-Isobutylthiazole has been reported to impart "medicinal", "metallic" or "rancid" aroma notes in aqueous solutions with concentrations above 50 ppb (Kazeniac and Hall, 1970), and in recent odor threshold studies with bland tomato homogenate, it contributed a "pungent/bitter" character (Tandon, 1998). In this study, significantly higher 2-isobutylthiazole concentrations found in 'BHN-189' tomatoes after 2 days storage at 5°C suggest that higher off-flavor ratings might be influenced by this volatile compound. Six-methyl-5-hepten-2-one and p-ionone have been reported to impart a "fruity" aroma (Kazeniac and Hall, 1970) and "sweet/floral" aroma when added to bland tomato homogenate (Tandon, 1998). Both these ketone compounds are thought to result from the breakdown of lycopene and other carotenoid pigments (Buttery and Ling, 1993). The reduction in lycopene content for tomatoes

PAGE 163

155 observed during these tests could affect the production of such important ketone volatile compounds. Stepwise regression analyses determined that ripe aroma ratings were influenced by hexanal, p-ionone, methanol, 2+3-methylbutanol, cis-2-hexenal, and soluble solids content, while, off-odor ratings were influenced by acetone concentrations and pH values. Sweetness ratings were described by fram^-heptenal, l-penten-3-one and p-ionone concentrations while, sourness ratings were influenced by titratable acidity, p-ionone, hexanal, and l-nitro-2-phenylethane concentrations. Green/grassy ratings were described in part by the concentrations of methanol and p-ionone, while, tomato flavor ratings by geranylacetone concentrations. Off-flavor sensory ratings were influenced by sucrose equivalents and 1 -nitro-2-phenylethane concentrations. Conclusions These studies provide strong evidence that postharvest shipping, handling and storage of fresh-market tomatoes at currently recommended temperatures of 10° to 12.5°C could induce perceptible, negative alterations in flavor and aroma prior to the appearance or in the absence of visual CI symptoms. Trained descriptive sensory panels, GC aroma volatile profiles and EN sensor outputs clearly distinguished increasing negative effects of storage at these temperatures with increased exposure times. Furthermore, short-term storage of ripe tomatoes at typical household refrigerator temperatures (3° to 5°C) could be one of the most contributing factors to consumer complaints about inferior tomato flavor. Significant correlations between EN sensor

PAGE 164

156 outputs and several aroma volatile compounds, notably 1 -nitro-2-phenylethane and 2isobutylthiazole, point to their important contribution to EN discriminating ability. It is certain that the effects of 7 days of low temperature storage on tomato flavor and aroma volatile production were partially reversible following a similar period of time at 20°C. However, this does not downplay the dramatic effect that low temperature storage may play on consumer dissatisfaction with fresh tomato flavor; after all few consumers take their tomatoes out of the refrigerator several days prior to consumption.

PAGE 165

157 Figure 6.1. Canonical plots for mature-green (MG), immature-green (IG), mature-green and low temperature storage (MG + 5°C), and immature-green and low temperature storage (IG + 5°C) whole fruit homogenate samples based on forward stepwise MVDA from EN sensor outputs. Classification of intact tomato fruit using MVDA canonical plot analysis was very highly significant (PO.0003).

PAGE 166

158 Figure 6.2. Canonical plots for mature-green (MG), immature-green (IG), mature-green and low temperature storage (MG + 5°C), and immature-green and low temperature storage (IG + 5°C) locular tissue samples based on forward stepwise MVDA from EN sensor outputs. Classification of locule tissue using MVDA canonical plot analysis was very highly significant (PO.00001).

PAGE 167

159 Figure 6.3. Canonical plots for mature-green (MG), immature-green (IG), mature-green and low temperature storage (MG + 5°C), and immature-green and low temperature storage (IG + 5°C) pericarp tissue samples based on MVDA from EN sensor outputs. Classification of pericarp tissue samples using MVDA canonical plot analysis was very highly significant (PO.0001).

PAGE 168

160 Sensory Panel Descriptoi Figure 6.4. Sensory panel descriptor ratings for table-ripe 'BHN-189' tomatoes stored for 2 days at four different temperatures. Significant differences (a=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 169

161 u « u na U q c o in a u £ 4 Sensory Panel Descriptors Figure 6.5. Sensory panel descriptor ratings for ripe 'BHN-189' tomatoes stored for 8 days at four different temperatures. Significant differences (a=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 170

162 Figure 6.6. Sensory panel descriptor ratings for ripe 'Solimar' tomatoes stored for 4 days at four different temperatures. Significant differences (a=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 171

163 Sensory Panel Descriptors Figure 6.7. Sensory panel descriptor ratings for ripe 'Solimar' tomatoes stored for 8 days at four different temperatures. Significant differences (oc=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 172

164 Figure 6.8. Sensory panel descriptor ratings for ripe 'Solimar' tomatoes stored for 12 days at four different temperatures. Significant differences (a=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 173

165 Canonical Function 1 Figure 6.9. MVDA canonical plot analysis from EN sensor outputs for ripe 'BHN-189' tomatoes after 2 days of storage at different temperatures. Ellipses around clusters represent 95% confidence bands.

PAGE 174

166 Figure 6.10. MVDA canonical plot analysis from EN sensor outputs for ripe 'BHN-189' tomatoes after 8 days of storage at different temperatures. Ellipses around clusters represent 95% confidence bands.

PAGE 175

167 Figure 6.11. MVDA canonical plot analysis from EN sensor outputs for ripe 'Solimar' tomatoes after 4 days of storage at different temperatures. Ellipses around clusters represent 95% confidence bands.

PAGE 176

168 Figure 6.12. MVDA canonical plot analysis from EN sensor outputs for ripe 'Solimar' tomatoes after 8 days of storage of storage at different temperatures. Ellipses around clusters represent 95% confidence bands.

PAGE 177

169 s 5 3 s u. a 8 a o -80 -60 ^0 -20 0 20 Canonical Function 1 100 Figure 6.13. MVDA canonical plot analysis from EN sensor outputs for ripe 'Solimar' tomatoes after 12 days of storage at different temperatures. Increasing distance between clusters relates to greater dissimilarity between samples

PAGE 178

170 Sensory Panel Descripti Figure 6.14. Sensory panel descriptor ratings for ripe 'Trust' tomatoes after 7 days of storage at 5°C and 1 day at 20°C. Significant differences (a=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 179

171 Sensory Panel Descriptors Figure 6.15. Sensory panel descriptor ratings for ripe 'Trust' tomatoes after 7 days of storage at 5°C and 8 days at 20°C. Significant differences (a=0.05) within temperature treatments for each descriptor were determined using Duncan's Multiple Range Test. Deviation bars represent Duncan's critical ranges for mean separations at the 5% level.

PAGE 180

172 Figure 6.16. MVDA canonical plot analysis from EN sensor outputs for ripe 'Trust' tomatoes after 7 days of storage at 5°C plus 1 day at 20°C. Ellipses around clusters represent 95% confidence bands.

PAGE 181

173 Figure 6.17. MVDA canonical plot analysis from EN sensor outputs for ripe 'Trust' tomatoes after 7 days of storage at 5°C plus 8 days at 20°C. Ellipses around clusters represent 95% confidence bands.

PAGE 182

CHAPTER 7 CONCLUSIONS Data collected during nine harvests of green-stage tomatoes demonstrated a strong relationship between physiological maturity at harvest (M1-M4) and ethylene exposure (in days) to attain breaker stage (R 2 = 0.84). The regression equation developed from experimental data could be employed by tomato growers and packers to predict the proportion of green-harvested tomatoes from a lot that would require extended ethylene treatment and that would have potentially inferior sensory quality at table-ripe stage. In addition to the use of ethylene as a screening tool for immature-green harvested tomatoes, electronic nose sensor analysis non-destructively segregated whole, mature-green fruit from immature-green harvested fruit. Changes in chemical composition and aroma volatile compounds documented in four commercial tomato cultivars ('Agriset-761', 'CPT-5', 'BHN-189' and 'Solimar') were correlated to maturity at harvest and differences in sensory perception. Volatile compounds such as cw-3-hexenal, cz's-S -hexenol, 2+3-methylbutanol, geranylacetone, hexanal and p-ionone consistently showed significant changes in concentration in this study, thus, suggesting their relative susceptibility to the effects of harvest maturity. In addition, in tomatoes requiring extended ethylene treatments (>3 days) concentrations of carotenoid-derived volatile compounds (geranylacetone, 6-methyl-5-hepten-2-one and pV 174

PAGE 183

175 ionone) were consistently lower, as were amounts of lycopene, while soluble solids contents and pH values generally higher. Results from sensory panels provided strong evidence of the negative effects that immature-green harvest and low temperature storage had on ripe tomato flavor and aroma. Descriptive sensory panels revealed that tomatoes exposed to extended ethylene treatments (>3 days) were likely harvested at immature-green stage, since these tomatoes were rated significantly lower in ripe aroma, tomato flavor and sweetness, and significantly higher in sourness and green/grassy flavor at table-ripe stage. Conversely, tomatoes which attained breaker stage after 1 to 3 days ethylene exposure had sensory ratings comparable or superior to those harvested at light-red stage, thus suggesting that mature-green harvested tomatoes have the potential to ripen with very acceptable flavor and aroma. Green-harvested tomatoes at overripe stage (table-ripe +14 days at 20°C), had considerably lower sensory qualities when compared to those harvested at light-red stage, regardless of ethylene requirement. There were also significant correlations between sensory panel ratings and aroma volatile compounds. Concentrations of acetaldehyde, methanol and ethanol were negatively correlated to tomato flavor ratings, while £ram , -2-heptenal concentrations were positively correlated to ripe aroma sensory ratings. In general, carotenoid-derived volatile compounds were negatively correlated to green-grassy flavor, while lipid-derived compounds were positively correlated with sweetness ratings. Methanol, ethanol, acetaldehyde and acetone were correlated to offflavor and off-odor sensory ratings.

PAGE 184

176 The electronic nose (EN) sensor array successfully discriminated between ripe tomato homogenate samples from the ethylene treatments, and these results were contrasted to changes in aroma volatile profiles. Pericarp tissue showed higher concentrations of some of the important tomato aroma compounds, an observation that might downplay the importance of locule tissue fluidity in the impact of flavor perception during mastication. The gas chromatographic (GC) headspace technique for aroma volatile quantification was useful to contrast the differences in volatile profiles observed during the EN analysis. Some lipid-derived volatile compounds (c/s-3-hexenal and trans2-heptenal) and carotenoid-derived compounds (geranylacetone and (5-ionone) were significantly correlated to responses from numerous EN sensors. Significant correlations were identified between sensory ratings for ripe aroma, green-grassy flavor and offodors, and the outputs from numerous EN sensors. Results for 'Agriset-761' and 'CPT-5' tomatoes suggested that fruit color at tableripe stage was significantly affected by ripeness stage at harvest and harvest date. In general, lightness coefficients (L*) were lower, regardless of ripeness stage, for tomatoes harvested during December 1995 compared to subsequent harvests. Tomatoes harvested beyond the pink ripeness stage developed a darker red color compared to green or breaker-harvested fruit. Higher hue angles for ripe tomatoes that were harvested beyond breaker stage indicated fruit color approached an orange-red color compared to a lighter red coloration for breaker and green-harvested tomatoes. Time-temperature studies showed that postharvest storage of fresh-market tomatoes at currently recommended temperatures of 10° to 12.5°C could induce significant negative alterations in flavor and aroma in the absence of visual chilling injury

PAGE 185

177 symptoms. Trained descriptive sensory panels, GC aroma volatile profiles and EN sensor outputs distinguished increasing negative effects of storage temperatures below 12.5°C as exposure time increased. Sensory panel ratings for ripe tomatoes stored at or below 10°C were generally lower in tomato flavor, ripe aroma, sweetness, and higher in sourness when compared to fruit stored at or above 12.5°C. EN analysis demonstrated significant dissimilarities between ripe tomatoes stored at different temperatures, where increasing Mahalanobis distances between temperature groupings with longer storage times illustrated a cumulative chilling effect on ripe tomato volatiles. Furthermore, the number of aroma volatile compounds showing significant differences increased 7-fold between 2 and 14 days of storage at 5°C to 20°C. Correlation matrices showed significant relationships between compositional parameters and sensory ratings for tomatoes stored at different temperatures. Concentrations of hexanal, /ra/js-2-heptenal, geranylacetone and p-ionone were positively correlated with ripe aroma sensory ratings. While outputs from EN sensors were negatively correlated to concentrations of acetaldehyde, methanol and 2isobutylthiazole, and positively correlated with concentrations of cw-3-hexenal and 1nitro-2-phenylethane. Short-term storage of ripe tomatoes (2 days) at typical household refrigerator temperatures (3° to 5°C) could be one of the most contributing factors to consumer complaints about inferior tomato flavor. High temperature pre-treatment (38°C for 2 days) prior to storage at 5°C for 7 days did alleviate the negative changes in ripe tomato fruit sensory quality induced by storage at chilling temperatures.

PAGE 186

LIST OF REFERENCES Ahmed, E.M., R.A. Dennison, R.H. Dougherty, and P.E. Shaw. 1978. Flavour and odour thresholds in water of selected orange juice components. J. Agric. Food Chem. 26:187-191 Al-Shaibani, A.M.H. and J.K. Greig. 1979. Effects of maturity, storage, and cultivar on some quality attributes of tomatoes. J. Amer. Soc. Hort. Sci. 104(6):880-882. Anonymous. 1996. An introduction to electronic nose technology. Neotronics Scientific Inc., Flowery Branch, GA. Pages. 12-56. Artes, F., E. Sanchez and L.M.M. Tijskens. 1998. Quality and shelflife of tomatoes improved by intermittent warming. Food Sci. and Techno 1. L.W.T. 31(5):427431. Autio, W.R. and W.J. Bramlage. 1986. Chilling sensitivity of tomato fruit in relation to ripening and senescence. J. Amer. Soc. Hort. Sci. 1 1(2):201-204. Baldwin, E.A., J.W. Scott, M.A. Einstein, T.M. Malundo, B.T. Carr, R.L. Shewfelt, and K.S. Tandon. 1998. Relationship between sensory and instrumental analysis for tomato flavor. J. Amer. Soc. Hort. Sci. 123(5):906-915. Baldwin, E.A., M.O. Nisperos-Carriedo, and J.W. Scott. 1992. Tomato flavor volatile profiles: cultivar and harvest maturity effects. Proc. 1992 Tomato Quality Workshop and Tomato Breeders Roundtable, University of Florida, IF AS. Pages 18-29. Baldwin, E.A., M.O. Nisperos-Carriedo and J.W. Scott. 1991a. Levels of flavor volatiles in a normal cultivar, ripening inhibitor and their hybrid. Proc. Fla. State Hort. Soc. 104:86-89. Baldwin E.A., M.O. Nisperos-Carriedo, R. Baker, and J.W. Scott. 1991b. Quantitative analysis of flavor parameters in six Florida tomato cultivars. J. Agric. Food Chem. 39:1135-1140. Baldwin, E.A.; M.O. Nisperos-Carriedo and M.G. Moshonas. 1991c. Quantitative analysis of flavor and other volatiles and for certain constituents of two tomato cultivars during ripening. J. Amer. Soc. Hort. Sci. 116(2):265-269. 178

PAGE 187

179 Bartley, I.M., P.G. Stoker, A.D. martin, S.G. Hatfield, and M. Knee. 1985. Synthesis of aroma compounds by apples supplied with alcohols and methyl esters of fatty acids. J. Sci. Food Agric. 36:567-574. Benady, M, J.E. Simon, D.J. Charles and G.E. Miles. 1995. Fruit ripeness determination by electronic sensing of aromatic volatiles. Trans.ASAE 38(l):251-257. Biale, J.B. and R.E. Young. 1991. Respiration and ripening in fruits: retrospect and prospect. Academic Press, London, pp. 1-39. Biggs, M.S., W.R. Woodson and A.K. Handa. 1988. Biochemical basis of high temperature inhibition of ethylene biosynthesis in ripening tomato fruits. Physiol. Plantarum 72:572-578. Bisogni, C.A. and G. Armbruster. 1976. Quality comparisons of room ripened and field ripened tomato fruits. J. Food Sci. 41:333-338. Borjesson, T., T. Eklov, A. Jonsson, H. Sundgren and J. Schnurer. 1996. Electronic nose for odor classification of grains. Cereal Chemistry 73(4):457-461 . Brecht, J.K., R.L Shewfelt, J.C. Garner, and E.W. Tollner. 1991. Using X-ray-computed tomography to nondestructively determine maturity of green tomatoes. HortScience 26(l):45-47. Brecht, J.K. 1987. Locular gel formation in developing tomato fruit and the initiation of ethylene production. HortScience 22(3):476-479. Brecht, J.K. 1985. Location of development of ethylene biosynthesis during tomato fruit maturation. Proc. Fla. State Hort. Soc. 98:184-187. Brecht, P.E., L. Keng, C.A. Bisogni and H.M. Munger. 1976. Effect of fruit portion, stage of ripeness and growth habit on chemical composition of fresh tomatoes. J. Food Sci. 41:945-947. Brodnitz, M.H., C.L. Pollock and P.P. Vallon. 1969. Flavor components of onion oil. J. Agric. Food Chem. 17:760-763. Brown, E.D., RY. Yada and D.W. Stanley. 1989. Chilling injury and modified atmosphere storage of mature-green tomatoes. Can. Inst. Food Sci. Technol. 22(4):339-344. Burdach, K.J., J. Kroeze, and E.P. Koster. 1984. Nasal, retronasal, and gustatory perception: an experimental comparison. Perception Psychophys. 36:205-208.

PAGE 188

180 Buttery, R.G. 1993. Quantitative and sensory aspects of tomato and other vegetables and fruits. In: T.E. Acree and R. Teranishi (eds). Flavor science: sensible principles and techniques. Amer. Chem. Soc, Washington, DC. pp. 259-286. Buttery, R.G. and L.C. Ling. 1993. Volatile components of tomato fruit and plant parts: relationship and biogenesis. In: R. Teranishi, R.G. Buttery and H. Sugisawa (eds.) Bioactive volatile compounds from plants. Amer. Chem. Soc, Washington, DC. pp. 23-34. Buttery, R.G., R. Teranishi, L. Ling and J.G. Turnbaugh. 1990. Quantitative and sensory studies on tomato paste volatiles. J. Agric. Food Chem 38:336-340. Buttery, R.G., Teranishi, R., Ling, L., Flath, R.A., and Stern, D. 1988. Quantitative studies on origins of fresh tomato aroma volatiles. J. Agric. Food Chem. 36:12471250. Buttery, R.G., R. Teranishi, and L.C. Ling. 1987. Fresh tomato aroma volatiles: a quantitative study. J. Agric. Food Chem. 35: 540-544. Buttery, R.G., R.M. Seifert, D.G. Guadagni and L.C. Ling. 1969. Characterization of some volatile constituents of bell peppers. J. Agric. Food Chem 17:1322-1327. Cabrera, R.M. and M.E. Saltveit. 1990. Physiological response to chilling temperatures of intermittently warmed cucumber fruit. J. Amer. Soc. Hort. Sci. 1 15:256-261. Cain, W.S. and C.L. Murphy. 1980. Interaction between chemoreceptive modalities of odor and irritation. Nature 284:255-257. Cerra, F. 1975. Thick skinned tomato symbol of consumer discontent. New York Times 124 (April 15):33. Chomchalow, S. 1991. Storage conditions and timing of ethylene treatment affect uniformity and marketability of tomato fruit. Master's Thesis, university of Florida, Gainesville, pp. 35-58. Cook, C.C., C.S. Parsons, and L.P. McColloch. 1958. Methods to extend storage of fresh vegetables aboard ships of the U.S. Navy. Food Tech. 12:548. Crooks, J.R. and P.M. Ludford. 1984. Effect of chilling on volatiles of tomato fruit. Acta Horticulturae 157: 277-282. Davies, J.W. and E.C. Cocking. 1965. Changes in carbohydrates, proteins and nucleic acids during cellular development in tomato fruit locule tissue. Planta 67:242253.

PAGE 189

181 De Pooter, H.L. , M.R. Van Acker, and N.M. Schamp. 1987. Aldehyde metabolism and the aroma quality of stored Golden Delicious apples. Phytochem. 26:89-92. Dodds, G.H., P.N. Bartlett and J.W. Gardner. 1992. Odoursthe stimulus for an electronic nose, p 1-11 In: In: Gardner, J.W. and P.N. Bartlett (eds.) Sensor and sensory systems for an electronic nose. NATO ASI series E Applied Sciences, Kluwer Academic Publishers. Dodds, G.T., J.W. Brown and P.M. Ludford. 1991. Surface color changes of tomato and other solanaceous fruit during chilling. J. Amer. Soc. Hort. Sci. 1 16(3):482-490. Dubois, M., K.A. Gilles, J.K. Hamilton, P. A. Rebers, and F. Smith. 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28:350356. Fellman, J.K., D.S. Martinson, B.C. Bostick, and J.P. Mattheis. 1993. Ester biosynthesis in 'Rome' apples subjected to low-oxygen atmospheres. Postharvest Biol, and Technol. 3:201-214. Florida Tomato Committee. 1998. Annual Report 1997-98. Orlando, FL. pp. 45-49. Florida Tomato Committee. 1989. Annual Report 1988-89. Orlando, FL. pp. 19-32. Gardner, J.W. and E.L. Hines. 1997. Pattern analysis techniques. In: Kress-Rogers, E., and D. Phil (eds.) Handbook of Biosensors and Electronic Noses:Medicine, Food and the Environment. CRC Press Inc. p. 633-652. Gardner, H.W. 1995. Biological roles and biochemistry of the lipoxygenase pathway. HortScience 30:197-204. Gardner, W.H. 1966. Food acidulants. Allied Chemical Corp., New York, NY. pp. 76. Gerasopoulos, D. and D.G. Richardson. 1995. Effects of storage temperature on fatty acid composition of 'Anjou' pears. Acta Hortic. 379:321-327. Goodenough, P.W. 1990. Flavor of tomato, cucumber and gherkin. In: Morton, I.D.; Macleod, A.J. (eds.). 1990. Food Flavours: The flavour of fruits. Elsevier Publishers International, pp. 327-349. Gould, W.A. 1978. Quality evaluation of processed tomato juice. J. Agric. Food Chem. 26:1006-1011. Gray, J., S. Picton, J. Shabber, W. Schuch, and D. Grierson. 1992. Molecular biology of fruit ripening and its manipulation with antisense genes. Plant Mol. Biol. 19:6987.

PAGE 190

182 Grierson, D., A. Slater, M. Mounders, P. Crookes, G. Tucker, W. Schuch, and K. Edwards. 1985. Regulation of the expression of tomato fruit ripening genes: the involvement of ethylene. In: Roberts, J. and G.A. Tucker (eds). Ethylene and plant development. Butterworths, London, pp. 147-161. Guadagni, D.G., RG. Buttery, and S. Okano. 1963. Odor thresholds of some organic compounds associated with food flavors. J. Sci. Food Agr. 14:761-765. Gull, D., D. Cartagena and E.C. French. 1980. Analisis de calidad de tomate para lograr un mejor producto (in Spanish). IBTA, PRODES, UFLA, Cochabamba, Bolivia. Page 20. Hardenburg, R.E., A.E. Watada and C.Y. Wang. 1986. The commercial storage of fruits, vegetables, and florist and nursery stocks. USDA-ARS Agriculture Handbook 66. Pages 11-12. Harris, E.S. 1973. What's wrong with tomatoes? The Packer. 80(Feb. 10):24A. Hayase, F., T.Y. Chung, and H. Kato. 1984. Changes of volatile components of tomato fruits during ripening. Food Chem. 14:1 13-124. Hilderbrand, D.F. 1989. Lipoxygenases. Physiol. Plant. 76:249-253. Hobson, G.E. and D. Grierson. 1993. Tomato. In: Seymour, G.B., J.E. Taylor, and G.A. Tucker (eds). Biochemistry of fruit ripening. Chapman and Hall Co., London, pp. 406-442. Hobson, G.E. and L. Bedford. 1 989. The composition of cherry tomatoes and its relation to consumer acceptability. J. Hortic. Sci. 64:321-329. Hobson, G.E. 1987. Low temperature injury and the storage of ripening tomatoes. J. Hort. Sci. 62(l):55-62. Hobson, G.E. 1981. The short-term storage of tomato fruit. J. Hort. Sci. 56:363-368. Hobson, G.E. and J.N. Davies. 1971. The tomato. In: Hume, A.C. (ed). Biochemistry of fruits and their products. Academic Press, New York, NY. Vol. 2:437-482. Hodgins, D.and D. Simmonds. 1995. Sensory technology for flavor analysis. Cereal Foods World 40 (4): 187-1 91. Jones, R.A., and S.J. Scott. 1984. Genetic potential to improve tomato flavor in commercial Fl hybrids. J. Amer. Soc. Hort. Sci. 109:318-321.

PAGE 191

183 Kader, A. A., L.L. Morris, M.A Stevens, and M. Albright-Holton. 1978. Composition and flavor quality of fresh market tomatoes as influenced by some postharvest handling procedures. J. Amer. Soc. Hort. Sci. 103(1):6-13. Kader, A. A., M.A. Stevens, M. Albright-Holton, L.L. Morris, and M. Algazi. 1977. Effect of fruit ripeness when picked on flavor and composition in fresh market tomatoes. J. Amer. Soc. Hort. Sci. 102(6): 724-731. Kader, A. A. and L.L. Morris. 1976. Correlating subjective and objective measurements of maturation and ripeness of tomatoes. In: proc. 2 nd Tomato Quality Workshop Veg. Crops Ser. 178. Univ. of Calif, Davis. Kavanagh, E.E., W.B. McGlasson and R.L. McBride. 1986. Harvest maturity and acceptability of 'Flora-Dade' tomatoes. J. Amer. Soc. Hort. Sci. lll(l):78-82. Kazeniac, S.J. and R.M. Hall. 1970. Flavor chemistry of tomato volatiles J. Food Sci. 35:519-530. Klann, E.M., B. Hall, and A.B. Bennett. 1996. Antisense acid invertase (TIV1) gene alters soluble sugar composition and size in transgenic tomato. Plant Physiol. 112:1321-1330. Koehler, P.E. and S.A. Kays. 1991. Sweet potato flavor: quantitative and qualitative assessment of optimum sweetness. J. Food Qual. 14:241-249. Krammer, G.E., R.G. Buttery, and G.R. Takeoka. 1995. Studies on tomato glycosides. In: Rouseff, R.L. and M.M. Leahy (eds). Fruit flavors: biogenesis, characterization, and authentication. Amer. Chem Soc. Symp. Series 596:164-181. Kuo, J.M., A. Hwang, H.H. Hsu, and B.S. Pan. 1996. Preliminary identification of lipoxygenase in algae (Enteromorpha intestinalis) for aroma formation. J. Agric. Food Chem. 44:2073-2077. Laing, D.G., H. Panhuber, M.E. Willcox, and E.P. Pittman. 1984. Quality and intensity of binary odor mixtures. Physiol. Behav. 33:309-319. Land, D.G. 1994. Perspectives on the effects of interactions on flavor perception: an overview. Amer. Chem. Soc. Symposium Series 633:3-11. Lawless, H.T. and H. Heyman. 1998. Sensory evaluation of food: principles and practices. Chapman and Hall Co., Cambridge University Press, pp. 341-377. Luning, P. A. 1995. Characterization and occurrence of lipoxygenase in bell peppers at different ripeness stages in relation to formation of volatile compounds. J. Agric. Food Chem. 43:1493-1500.

PAGE 192

184 Lurie, S., A. Handros, E. Fallik and R. Shapira. 1996. Reversible inhibition of tomato fruit gene expression at high temperature. Plant Physiol. 1 10:1207-1214. Lurie, S. and J.D. Klein. 1991. Acquisition of low temperature tolerance in tomatoes by exposure to high temperature stress. J. Amer. Soc. Hort. Sci. 1 16:1007-1012. Lurie, S. and J.D. Klein. 1990. Heat treatment of ripening apples: differential effects on physiology and biochemistry. Physiol. Plant. 78:181-186. Malewski, W. and P. Markakis. 1971. Ascorbic acid content of the developing tomato fruit. J. Food Sci. 36:537. Malundo, T.M.M., Shewfelt, R.L., and Scott, J.W. 1995. Flavor quality of fresh tomato (Lycopersicon esculentum Mill.) as affected by sugar and acid levels. Postharvest Biol, and Tech. 6:103-110. Marsili, R. 1995. The electronic nose: a sensory evaluation tool. Food Prod. Design 6:53-66. Maruniak, J.A. 1988. The sense of smell. In: Pigott, J.R. (ed.). Sensory analysis of foods. Elsevier Science, UK. pp. 2562. Masarirambi, M.T. 1997. Physiological responses of tomato fruit to ethylene at high temperature. Doctoral Dissertation, University of Florida, Gainesville, FL. pp. 174. Massantini, R., E. Gobattoni, R. Botondi and F. Mencarelli. 1995. Role of alcohol dehydrogenase on ripening of ethanol treated tomato and kiwifruit. Acta Hort. 379:297-304. Mattheis, J.P. and R.G. Roberts. 1993. Fumigation of sweet cherry (Prunus avium 'Bing') fruit with low molecular weight aldehydes for postharvest decay control. Plant Disease 77(8):810-814. Maul, F., S.A. Sargent, M.O. Balaban, E.A. Baldwin, D.J. Huber, and C.A. Sims. 1998. Aroma volatile profiles from ripe tomatoes are influenced by physiological maturity at harvest: an application for electronic nose technology. J. Amer. Soc. Hort. Sci. 123(6): 1 106-1 110. Maul F., S.A Sargent, E.A. Baldwin, C.A. Sims and D.J. Huber. 1997a. Sensory evaluation correlates with flavor volatile profiles of ripe tomato fruits. HortScience 32(3):521. Maul, F., S.A. Sargent, D. J. Huber, M.O. Balaban, D.A. Luzuriaga and E.A. Baldwin. 1997b. Non-destructive quality screening of tomato fruit using electronic nose technology. Proc. Fla. State Hort. Soc. 1 10:188-194.

PAGE 193

185 McColloch, L.P. and J.T. Worthington. 1952. Low temperature as a factor in te susceptibility of mature-green tomatoes to Alternaria rot. Phytopatology 42:425427. McDonald, R.E., T.G. McCollum, and E.A. Baldwin. 1998. Heat treatment of maturegreen tomatoes: differential effects of ethylene and partial ripening. J. Amer. Soc. Hort. Sci. 123(3):457-462. McDonald, R.E., T.G. McCollum and E.A. Baldwin. 1996. Prestorage heat treatments influence free sterols and flavor volatiles of tomatoes stored at chilling temperature. J. Amer. Soc. Hort. Sci. 121(3):53 1-536. MacLeod, R.F., A. A. Kader, and L.L. Morris. 1976. Stimulation of ethylene and C0 2 production of mature-green tomatoes by impact bruising. HortScience 1 1(6): 604606. Mencarelli, F. and M.E. Saltveit. 1988. Ripening of mature-green tomato fruit slices. J. Amer. Soc. Hort. Sci. 113(5):742-745. Mitcham, E.J., K.C. gross, and T.J. Nig. 1989. Tomato fruit cell wall synthesis during development and senescence. Plant physio 1. 29:477:481. Mitchell, W.C. and G. Jellenkovic. 1995. Characterizing NAD-dependant and NADPdependant alcohol dehydrogenase enzymes of strawberries. J. Am Soc. Hort. Sci. 120(5):798-801. Moretti, C.L., S.A. Sargent, D.J. Huber, A.G. Calbo and R. Puschmann. 1998. Chemical composition and physical properties of pericarp, locule and placental tissues of tomato fruit with internal bruising. J. Amer. Soc. Hort. Sci. 123(4):656-660. Moretti, C.L., S.A. Sargent, C.A. Sims, and R. Puschmann. 1997. Flavor alteration in tomato fruit due to internal bruising. Proc. Fla. State Hort. Soc. 1 10:195-197. Morris, L.L. 1954. Field and transit chilling of fall-grown tomatoes. ProceedingsConference on transportation of perishables. UC Davis, pp. 5. Moskowitz, H. 1988. Applied sensory analysis of foods. CRC Press, Boca Raton, FL. Vol. 1, pp. 44-51. Nagodawithana, T. 1994. Flavor enhancers: their probable mode of action. Food Technol. 4:79-85. Nursten, H.E. 1970. Organic acids. In: Hulme A. (ed). The biochemistry of fruit and their products. Academic Press, London. Vol. 2, pp. 89-113.

PAGE 194

186 Ogura, N., H. Nakagawa, and H. Takenhana. 1975. Effect of high temperature short term storage of mature green tomato fruits on changes in their chemical composition after ripening at room temperature. J. Agric. Chem. Soc. 49:189-196. Pearce, T.C. 1997a. Computational parallels between the biological olfactory pathway and its analog the electronic nose. 1. Biological olfaction. Biosystems 41(1):4367. Pearce, T.C. 1997b. Computational parallels between the biological olfactory pathway and its analog the electronic nose. 2. Sensor-based machine olfaction. 41(1):6990. Perez, A.G., J.J. Rios, C. Sanz, and J.M. Olias. 1992. Aroma components and free amino acids in the strawberry variety Chandler during ripening. J. Agric. Food Chem. 40:2232-2235. Pesis, E. 1995. Induction of fruit aroma and quality by postharvest applications of natural metabolites or anaerobic conditions. In: Linskens, H.F. and J.F. Jackson (eds). Modern methods of plant analysis: fruit analysis. SpringerVerlag, Berlin, Germany. Vol. 18, pp. 19-35. Pesis, E. 1994. Enhancement of fruit aroma and quality by acetaldehyde or anaerobic treatments before storage. Acta Hortic. 368:365-373. Pesis, E., G. Zauberman, and I. Avissar. 1991. Induction of certain volatiles in feijoa fruit by postharvest application of acetaldehyde or anaerobic conditions. J. Sci. Food Agric. 54:329-337. Petro-Turza, M. 1987. Flavor of tomato and tomato products. Food Rev. Intl. 2(3):309351. Plattig, K.H. 1988. The sense of taste. In: Piggott, J.R. (ed). Sensory analysis of foods. Elsevier science, UK. pp 1-23. Resureccion, A.V.A. and R.L. Shewfelt. 1985. Relationships between sensory attributes and objective measurements of postharvest quality of tomatoes. J. Food Sci. 50:1242-1245. Rhodes, M.J. 1 980. The maturation and ripening of fruits. In: Senescence in plants. Academic Press, New York. pp. 157-205. Riley, J.C., C. Willemot, and J.E. Thompson. 1996. Lipoxygenase and hydroperoxide lyase activities in ripening tomato fruit. Postharvest biol. and Technol. 7:97-107. Rizzolo, A., M. Vanoli, and C. Visai. 1995. Effect of cold storage on volatile constituents of peaches and nectarines. Acta Hortic. 379:467-474.

PAGE 195

187 Rouseff, R.L. and M.M. Leahy. 1995. Advances in fruit flavor research: an overview. In: Rouseff, R.L. and M.M. Leahy (eds). Fruit Flavors. Amer. Chem Soc. Symposium Series 596:1-7. Sabehat, A., D. Weiss and S. Lurie. 1996. The correlation between heat-shock protein accumulation and persistence and chilling tolerance in tomato fruit. Plant Physiol. 110:531-537. Sargent, S.A., F. Maul, C.L. Moretti and C.A. Sims. 1997. Harvest maturity, storage temperature and internal bruising affect tomato flavor. Proc. 1997 Florida Tomato Institute, University of Florida, IF AS. Pages 22-24. Sargent, S.A., J.K. Brecht, and J.J Zoellner. 1992. Sensitivity of tomatoes at mature-green and breaker ripeness stages to internal bruising. J. Amer. Soc. Hort. Sci. 117 (1):1 19-123. Sargent, S.A., J.K. Brecht, and J.J. Zoellner. 1989. Assessment of mechanical damage in tomato packinglines. Amer. Soc. Agr. Eng. Paper no. 89-6060, St. Joseph, MI. SAS Institute Inc. 1996. SAS for Windows v 6.12. SAS Institute, Cary, NC. Schwartz, E. 1995. As premium tomatoes heat up, sales savvy becomes essential. Produce Business, 1 1 :48-52. Shewfelt, RL. and A.C. Purvis. 1995. Toward a comprehensive model for lipid peroxidation in plant tissue disorders. HortScience 30(2):213-218. Shurmer, H.V. 1990. An electronic nose: a sensitive and discriminating substitute for a mammalian olfactory system. IEE Proceedings 137(3): 197-204. Simon, J.E., A. Hetzroni B. Bordelon, G.E. Milesand D.J. Charles. 1996. Electronic sensing of aromatic volatiles for quality sorting of blueberries. J. Food Sci. 61(5):967-969. Speirs, J., E. Lee, K. Holt, K. YongDuk, N.S. Scott, B. Loveys, and W. Schuch. 1998. Genetic manipulation of alcohol dehydrogenase levels in ripening tomato fruit affects the balance of some flavor aldehydes and alcohols. Plant Phys. 117(3):1047-1058. Srivastava, M.S. and E.M. Carter. 1983. An introduction to applied multivariate statistics. Elsevier Science, New York, NY. Statsoft, Inc. 1994. STATISTICA for Windows v 4.5. Statsoft Inc., Tulsa, OK.

PAGE 196

188 Stern, D.J., R.G. Buttery, R. Teranishi, L. Ling, K. Scott, and M Cantwell. 1994. Effect of storage and ripening on fresh tomato quality, Part I. Food Chem. 49:225231. Stevens, M.A. 1985. Tomato flavor: effects of genotype, cultural practices, and maturity. In: Pattee, H.E. (ed.). Evaluation of Quality of Fruits and Vegetables. AVI Publishing, Westport. 41 Op. Stevens, M.A., A. A. Kader, M. Albright-Holton, and M. Algazi. 1977. Genotypic variation of flavor and composition in fresh market tomatoes. J. Amer. Soc. Hort. Sci. 102:680-689. Stone, H. and J.L. Sidel. 1985. Sensory evaluation practices. Academic Press, London, pp. 194-226. Stone, H., J.L. Sidel, S. Oliver, A. Woolsey, and R.C. Singleton. 1974. Sensory evaluation by quantitative descriptive analysis. Food Techno 1. 28:24:34. Strassburger, K.J. 1996. Electronic nose evaluation in the flavor industry. Food Testing and Analysis 2(5):22-27. Tandon, K.S. 1998. Odor thresholds and flavor quality of fresh tomatoes (Lycopersicon esculentum Mill). Master's Thesis, University of Georgia, Athens, GA. pp 20-31. Terada, M., Y. Watanabe, M. Kunitoma, and E. Hayashi. 1978. Differential rapid analysis of ascorbic acid and ascorbic acid 2-sulfate by dinitrophenilhydrazine method. Ann. Biochem. 84:604-608. Theologis, A., P.W. Oeller, L.M. Wong, W.H. Rottman, and D. M Gantz. 1993. Use of a tomato mutant constructed with reverse genetics to study fruit ripening, a complex developmental process. Dev. Genet. 14:282-295. Tomlinson, J.B., H.L. Ormrod and F.R. Sharpe. 1995. Electronic aroma detection in the brewery. J. Amer. Soc. of Brewing chem. 53(4):167-173. Tucker, G.A. 1993. The biochemistry of fruit ripening: an introduction. In: Seymour, G.B., G.E. Taylor, and G.A. Tucker (eds). Biochemistry of fruit ripening. Chapman and Hall Co., Cambridge University Press. Ulrich, R. 1970. Organic acids. In: Hulme, A. (ed). The biochemistry of fruit and their products. Academic Press, London. Vol. 1, pp. 89-1 18. Umiel, N. and W.H. Gabelman. 1971. Analytical procedures for detecting carotenoids of carrot (Daucus carota L.) roots and tomato (Lycopersicon esculentum Mill.) fruits. J. Amer. Soc. Hort. Sci. 96(6):702-704

PAGE 197

189 U.S. Department of Agriculture. 1976. United States standard grades of fresh market tomatoes. Agricultural Marketing Service, USDA, Washington, DC. Wang, C. C.K. Chin, C.T. Ho, C.F. Hwang, J.J. Polashock, and C.E. Martin. 1996. Changes of fatty acids and fatty acid-derived flavor compounds by expressing the yeast A-9 desaturase gene in tomato. J. Agric. Food Chem. 44:3399-3402. Warm, E.V., EX. Jourdain, R. Pressey, and B.G. Lyon. 1985. Effect of mutant genotypes hp, og and dg og on tomato fruit quality. J. Amer. Soc. Hort. Sci. 1 10:212-215. Watada A.E. and B.B. Aulenbach. 1979. Chemical and sensory qualities of fresh market tomatoes. J. Food Sci. 44:1013-1016. Watada, A.E., B.B. Aulenbach, and J.T. Worthington. 1976. Vitamins A and C in ripe tomatoes as affected by stage of ripeness at harvest and by supplementary ethylene. J. Food Sci. 41:856-858. Whitaker, B.D. 1994. A reassessment of heat treatments as a means of reducing chilling injury in tomato fruit. Postharvest Biol, and Techno 1. 4:75-83. Whiting, G.C. 1970. Sugars. In: Hulme, A. (ed). The biochemistry of fruits and their products. Academic Press, London. Vol. 1, pp. 1-31. Willaert, G.A., P.J. Dirinck, H.L. De Pooter, and N.M. Schamp. 1983. Objective measurement of aroma quality of 'Golden Delicious' apples as a function of controlled atmosphere storage time. J. Agric. Food Chem 31:809-813. Yahia, E.M., M. Rivera, and O. Hernandez. 1992. Responses of papaya to short-term insecticidal oxygen atmospheres. J. Amer. Soc. Hort. Sci. 1 17(l):96-99. Yang, S.F. 1985. Biosynthesis and action of ethylene. HortScience 20:41-45. Yang, S.F. and N.E. Hoffman. 1984. Ethylene biosynthesis and its regulation in higher plants. Annual Rev. Plant Physiol. 35:155-189. Zook, K.L. and J.H. Pearce. 1988. Quantitative descriptive analysis. In: Moskowitz, H. (ed). Applied sensory analysis of foods. CRC Press, Boca Raton, FL. Vol. 1, pp. 43-70.

PAGE 198

BIOGRAPHICAL SKETCH The author was born on October 12, 1973, in Guatemala City, Guatemala. He attended the Pan-American School for Agriculture "Zamorano" in Tegucigalpa, Honduras between 1992 and 1994. At Zamorano, he was awarded an agronomo degree on December 1994. He then transferred to the University of Florida to complete his undergraduate studies at the Horticultural Sciences Department, and was awarded a B.Sc. degree in December 1995. He was admitted into Graduate School to pursue a Doctoral degree majoring in postharvest physiology and technology of horticultural crops. Under the guidance of Dr. Steven A. Sargent, the author conducted research studies focusing on fresh tomato flavor quality, and was awarded a Ph. D. degree in May 1999. Upon receiving his degree, the author returned to Guatemala to pursue a career in industry and academia. 190

PAGE 199

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. \. Sargent, CHkir Steven A. Professor of Horticultural Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Pr Donald J. Huber Professor of Horticultural Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy Charles A. Sims Professor of Food Science and Human Nutrition I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Elizabeth A. Baldwin Courtesy Associate Professor of Horticultural Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. lufatTJTBalaban Professor of Food Science and Human Nutrition

PAGE 200

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May 1999 /Cj&C^-^ Dean, College of Agriculture Dean, Graduate School