CHEMICAL ANALYSIS OF CONCORD GRAPE ESSENCE AS RELATED TO CONCORD GRAPE FLAVOR Thesis for the Degree of Ph. D. IMICHTGAR STATE UNIVERSITY THEODORE WTLUAM‘ MUELLER ‘ 1971 mum“- "K9- LIBRARY 'VTE; ‘. Michigan State University This is to certify that the thesis entitled Chemical Analysis of Concord Grape Essence As Related to Concord Grape Flavor presented by Theodore William Moeller has been accepted towards fulfillment of the requirements for Ph-D degree in _Eo.od_S.cience Mfimrtmuun Date February 244 1971 0-7639 in am mum mc: ‘ LIBRARY amozns I " ABSTRACT CHEMICAL ANALYSIS OF CONCORD GRAPE ESSENCE AS RELATED TO CONCORD GRAPE FLAVOR BY Theodore William Moeller Twelve lSO-fold grape essences were obtained from three Michigan essence manufacturers for chemical analysis and flavor evaluation. Methyl anthranilate concentrations of volatile essence fractions ranged from 4 to 126 ppm, total esters ranged from 100 to 11,400 ppm (as ethyl acetate), total carbonyls ranged from 250 to 6600 ppm (as acetone), and chemical oxygen demands ranged from 20,000 to 200,000 ppm. Acetaldehyde and acetone levels were present from 18 to 2811 ppm and 0 to 440 ppm respectively. Similar varia- tions were evident in ultraviolet absorption spectra and gas-solid chromatographic analyses. Acetaldehyde was the single most abundant contri- butor to total carbonyl values. Ultraviolet absorption spectra indicated that unsaturated aldehydes were also present. Although compound concentrations in headspace vapors were not necessarily related to their respective solution concentrations, highly significant correlations between headspace and liquid acetaldehyde levels and Theodore William Moeller headspace ethyl acetate and total ester levels were present. Correlations between chemical and flavor panel data were generally poor. Flavor panel results showed that juices prepared from the essences were generally lower in flavor quality than commercial juices available to the consumer. Juices prepared from several essences, however, were acceptable to the panelists and were of the same overall quality as commercial juices. Using the methods described, no single component of Concord grape essence can be measured quantitatively for use as an absolute index of essence flavor quality. Thus, an intricate balance of components within the essence seems necessary for high flavor quality. This may best be measured by headspace vapor analysis via gas chroma- tography. CHEMICAL ANALYSIS OF CONCORD GRAPE ESSENCE AS RELATED TO CONCORD GRAPE FLAVOR BY Theodore William Moeller A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Food Science 1971 PLEASE NOTE: Some pages have small and indistinct type. Filmed as received. University Microfilms ACKNOWLEDGMENTS I am sincerely grateful to Dr. C. L. Bedford for his valuable guidance and suggestions throughout the course of this study and my degree program. I also thank members of my guidance committee, Drs. C. M. Stine, P. M. Markakis, W. M. Urbain, and D. R. Dilley, for their valuable contributions to my program. I extend my thanks to Mr. N. F. Roger of the A. F. Murch Company, Paw Paw, Michigan, for providing the essences and concentrate used in the study. To Dr. C. C. Sweeley and the Department of Bio- chemistry I extend thanks for use of the mass spectrometer unit. Also to Drs. R. Hammond and R. Laine who performed the analyses and data interpretation. I extend my gratitude to the Department of Food Science for financial aid during my program. The study was also supported in part by Public Health Service fellowship no. l-FOl-UI-42,748-01. Lastly, I am most grateful to my wife, Renee, for typing the initial manuscripts. To Renee and my daughter, Nicole, I give my deepest appreciation for their encourage- ment and sacrifices during these years. ii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . . . REVIEW OF LITERATURE . . . . . . . . . . . . . MATERIALS AND METHODS . . . . . . . . . . . . Materials . . . . . . . . . . . . . . . . Essences O O O O O O O O O O O O O O O concentrate O O O O O O O O I O 0 O O JUices O O O O O 0 0 O O O O O O O O 0 Methods of Analysis . . . . . . . . . . . Steam distillation . . . . . . . . . . Methyl anthranilate determination . . Total ester determination . . . . . . Chemical oxygen demand (COD) . . . . . Total carbonyl . . . . . . . . . . . . Thin layer chromatographic separation and quantitation of acetone and acetaldehyde 2,4-dinitrophenyl- hydrazones . . . . . . . . . . . . . Preparation of thin layer plates . . . Ultraviolet absorption . . . . . . . . Gas chromatography . . . . . . . . . . Peak identification . . . . . . . . . Peak quantitation . . . . . . . . . . Flavor evaluation . . ... . . . . . . Reference juice . . . . . Typicalness and acceptability ballot . Relative preference ballot . . . . . . Statistical analysis of data . . . . . iii Page vii 10 10 10 10 ll 12 13 14 14 15 17 17 19 20 20 21 23 24 24 Page RESULTS AND DISCUSSIONS . . . . . . . . . . . . . . . 26 Steam distillation . . . . . . . . . . . . . . 26 Methyl anthranilate . . . . . . . . . . . . . 28 Total esters . . . . . . . . . . . . . . . . . 30 Chemical oxygen demand . . . . . . . . . . . . 32 Total carbonyl . . . . . . . . . . . . . . . . 34 Acetone and acetaldehyde 2,4-dinitrophenylhydrazones . . . . . . . . 36 Ultraviolet absorbance . . . . . . . . . . . . 40 Gas-solid chromatography . . . . . . . . . . . 45 Flavor panels . . . . . . . . . . . . . . . . 57 Chemical test interrelationships . . . . . . . 67 Flavor panel vs chemical relationships . . . . 70 SUMMARY AND CONCLUSIONS 0 O O O O O O O O O O O O O O 76 LITERATURE CITED 0 O O I O O O O O C O O O O O O O O O 80 GENE RAL LITE RATU RE I O O O O O O O I O O O O O C O O O 8 3 APPENDIX 0 O O O O O O O O O O O O O O O O O O O O O O 86 iv ll. 12. 13. 14. 15. LIST OF TABLES Distillation recovery of methyl anthranilate . . . . . . . . . . . . . . Distillation recovery of ethyl acetate . . Distillation recovery of methyl anthranilate/ethyl acetate . . . . . . . Methyl anthranilate concentrations for the volatile fraction of essences . . . Total ester concentrations for the volatile fraction of essences . . . . . Chemical oxygen demand for the volatile fraction of essences . . . . . . . . . . Total carbonyl concentrations for the volatile fraction of essences . . . . . Acetaldehyde and acetone concentrations of the essences . . . . . . . . . . . . . . carbonYl data 0 O O O O O O O O O O O O O Ultraviolet absorbancies of essences . . . Gas-solid chromatographic peak identities Integrator count percentages for gas chromatography . . . . . . . . . . . . . Flavor panel data for reference juice determination . . . . . . . . . . . . . Flavor panel data for flavor difference and acceptability . . . . . . . . . . . Flavor panel data for preference by ranking Page 26 27 27 29 31 33 34 38 41 42 47 54 59 60 64 Table 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Chemical vs chemical correlations for the essences . . . . . . . . . . . . . Flavor vs chemical correlations for the essences . . . . . . . . . . . . . . . Methyl anthranilate standard curve . . . . . Methyl anthranilate data for the volatile fraction of the essences . . . . Total ester standard curve . . . . . . . . . Total ester data for the volatile fraction of the essences . . . . . . . . . Chemical oxygen demand standard curve . . . Chemical oxygen demand data for the volatile fraction of the essences . . . . . . . . . Total carbonyl standard curve . . . . . . . Total carbonyl data for the volatile fraction of the essences . . . . . . . . . Acetone 2,4-dinitrophenylhydrzaone standard curve . . . . . . . . . . . . . . Acetaldehyde 2,4-dinitrophenylhydrazone standard curve . . . . . . . . . . . . . . Thin layer chromatographic data for the essences . . . . . . . . . . . . . . . Ultraviolet absorbance data for the essences . . . . . . . . . . . . . . . Gas-solid chromatographic data for the essences . . . . . . . . . . . . . . . vi Page 68 71 86 86 87 87 88 88 89 90 91 91 92 93 94 LIST OF FIGURES Figure Page 1. Thin layer chromatographic separation of essence 2,4-dinitrophenylhydrazones . . . . . 37 2. Ultraviolet absorption spectra of 6 dilute essences . . . . . . . . . . . . . . 43 3. Ultraviolet absorption spectra of 6 dilute essences . . . . . . . . . . . . . . 44 4. Mass spectrum of gas-solid chromatographic peak B, acetaldehyde . . . . . . . . . . . . . 46 5. Mass spectrum of gas-solid chromatographic peak C, ethanol . . . . . . . . . . . . . . . 46 6. Gas-chromatographic separation of headspace volatiles over essence AJ68 . . . . . . . . . 48 7. Gas-solid chromatographic separation of headspace volatiles over essence AP68 . . . . 48 8. Gas-solid chromatographic separation of headspace volatiles over essence BJ68 . . . . 49 9. Gas-solid chromatographic separation of headspace volatiles over essence AJ69 . . . . 49 10. Gas-solid chromatographic separation of headspace volatiles over essence AP69 . . . . 50 11. Gas-solid chromatographic separation of headspace volatiles over essence AN69 . . . . 50 12. Gas-solid chromatographic separation of headspace volatiles over essence BJ69 . . . . 51 13. Gas-chromatographic separation of headspace volatiles over essence CJ69 . . . . 51 14. Gas-chromatographic separation of headspace volatiles over essence AJ6910/2 . . 52 vii Figure Page 15. Gas-chromatographic separation of headspace volatiles over essence AJ6910/7 . . . . . . . 52 16. Gas-chromatographic separation of headspace volatiles over essence AJ69lO/8 . . . . . . . 53 17. Gas-solid chromatographic separation of headspace volatiles over essence AJ69UK . . . 53 viii INTRODUCTION In many cases, the manufacturers of Concord grape juice and juice products have found it desirable to strip single-strength juice or puree of its essence and concen- trate the remaining portion to approximately 70° Brix. By utilizing this method of concentration, shipping weights and costs, container costs, refrigeration loads, man-hour‘ requirements, and the possibility of "chance" fermentation are all reduced (Murch and Ziemba, 1958). When these‘ essences are added back to the final product in proper amounts, manufacturers are able to produce full flavored products possessing all the natural Concord grape flavors and aromas (Roger, 1961). In the manufacture of such essences, the degree of concentration or essence strength is designated by the producer as being of a particular "fold". Thus, an essence of X fold theoretically has had each volatile component present in the original juice concentrated X times. This same figure is the value utilized by the end user of the essence to determine the amount of essence to be added to the final product. This value is actually derived by establishing a ratio of the rate of volumetric flow of input juice to the rate of volumetric flow of essence output of the concentration unit. It is assumed that uni- form concentration and 100 percent recovery of each chemical compound present in the original juice has been achieved. A flaw in utilizing this method, however, is the assumption that uniform concentration and 100 per cent recovery of each chemical compound present in the original juice has been achieved; obviously, this is not necessarily the case. The purpose of this study was to develop a simple, fast, and accurate test procedure which could be used to quantitate a single component and/or group of components present in various Concord grape essences, and to determine if this test may be used as a measurement of true essence strength. REVIEW OF LITERATURE Several investigators have attempted to charac- terize and elucidate the composition of not only Concord grape juice, but many other common juices. Methyl anthra- nilate was implicated in a study by Power and Chesnut (1921) as being a natural and apparently constant constit- uent of Concord grape juice. This compound possessed a decidedly grape-like odor in dilute aqueous solutions and was suspected to improve grape flavor when added to grape juices. It was found in other fruit juices by Power (1921). Methyl anthranilate was first discovered in 1895, in neroli oil. It has also been found in the oils of tuberose, ylang-ylang, Spanish orange blossoms, sweet orange rind, bergamot leaves, jasmin flowers, gardenia and certain varieties of apples. It was pointed out that even though methyl anthranilate has been found in all of the above, the compound is chiefly associated to the flavor in grape-type products (Scott, 1923). In a study involving varietal differences of grapes, it was found that pure-bred Zitig labrusca varieties (including Concord) contained relatively large amounts of methyl anthranilate. Varying amounts of the compound were found in hybrids of this species and tended to be present in higher concentrations when Vitis labrusca was the predominant genotype. On the other hand, Vitis vinifera (European-type grapes) and Vitis rotundifolia (Southern grapes) were found to be com- pletely devoid of methyl anthranilate. As a result, it was generalized the methyl anthranilate imparted a dis- tinctive Concord-type odor to those varieties in which it occured, although several exceptions were noted. For ex- ample, Catawba, a hybrid of Vitis labrusca-Vitis vinifera, was found to contain none of the compound, but had an odor characteristic of those grapes having relatively large amounts of methyl anthranilate (Power and Chesnut, 1923). A similar observation was made by Sale and Wilson (1926) on the variety Campbell, another Yitig’labrusca-Vitis vinifera cross. It was noted, however, that Campbell juices did contain exceptionally large quantities of volatile esters and that the total volatile ester content varied directly with the flavor and fragrance of juices of the Concord variety. No additional work was published on Concord grape flavor until volatile constituents present in aqueous solu- tions of Concord grape essences were examined by Holley et a1. (1955). Water-insoluble derivatives of the various compounds were prepared and examined chromatographically. Seven compounds were identified and quantitated. Ethanol and ethyl acetate were the most abundant compounds present, while the methyl anthranilate concentration was three orders of magnitude less than that of the former two compounds. A synthetic essence was prepared utilizing the concentra- tions of the seven compounds identified and quantitated. Comparing the ultraviolet absorption spectra of both essences, it was found that a weak maximum absorption at 280 nm in the natural essence was replaced by a minimum at 270 nm in the synthetic essence. The odor of the synthetic essence was lacking in components essential to natural Con- cord odor. However, when a chloroform extract, washed with acid to remove any methyl anthranilate present and concen- trated to obtain a small amount of oil, was added to the synthetic essence at the rate of 0.02 mg/ml, the resultant mixture had an odor which closely resembled that of the natural essence. The effect of this addition upon ultra- violet absorption was not discussed. In the more recent research, emphasis has been placed on essence composition. Most of the studies have been made using gas chromatography and mass spectrometry for compound separation and identification. Such an approach was used in a study of diethyl ether extracts of 100 "fold" Welch Concord essence. Sixteen volatiles were identified in that fraction of the original extract vapor- ized via bubbling nitrogen gas through the extract and subsequently collected in a dry ice-acetone trap at atmos- pheric pressure (Stevens et a1., 1965). Ethanol and ethyl acetate again were reported as being most abundant of all compounds identified while dimethyl vinyl carbinol, having a cresol-like odor, was the third most prevalent compound present. A small amount of material collected in a room temperature trap possessed a pungent odor, but was not examined. In a study of flavor variations of various Concord juices, methyl anthranilate was indicated as being essen- tial to characteristic Concord flavor (Clore et a1., 1965). The study showed the methyl anthranilate concentration in the grape itself increased steadily during the grape growing season until maturity. Total volatile esters as well as the total volatile ester-methyl anthranilate ratio fluctuated considerably during maturation and from season to season. Methyl anthranilate concentrations above a certain level did not increase the Concord flavor of the juices. Among compounds other than methyl anthranilate present in Concord essence, n-valeraldehyde was detected as an undesirable anomoly present in essences from atypi- cal Canadian Concord grapes (Neudoerffer et a1., 1965). A flavor described as "sweet and fruity" was evident in essences produced from grapes having ten times the normal amounts of this compound. Although n-valeraldehyde was the compound of primary interest, 31 additional compounds were identified when further analyses were conducted on ethyl chloride extracts of such essences. In the most recent report on Concord grape essence, 60 compounds were separated and identified in isopentane extracts of lSO-fold Seneca essence (Stern et a1., 1967). Relatively large amounts of ethyl acetate were found. How- ever, comparatively small concentrations of ethanol and dimethyl vinyl carbinol were found, in contrast to the results of Steven et a1. (1965). This was probably due to their relative insolubility in isopentane. These extracts contained unusually large amounts of crotonates, particu- larly the ethyl ester, and were reported to comprise a large percentage of the oil. Chemical compositions of the essences of grape varieties other than Concord have been reported. Seventy- seven compounds were identified as being present in isopentane extracts of Muscat of Alexandria essence (Stevens et a1., 1966). A relatively large percentage of this oil was made up of terpene alcohols with linalool and geraniol most abundant. Although several esters were present, they made up only a small percentage of the oil and no methyl anthranilate was found. Fifty-seven compounds were identi- fied by Stevens et al. (1967) in isopentane extracts of an essence from another Xitig vinifera variety, Grenache. Unlike the Muscat variety, a significantly larger amount of esters and a much smaller amount of terpene alcohols were present in this essence. The bulk of this essence was made up of alcohols. Trans-2-hexena1 and hexanal were two of six aldehydes identified and represented a large per- centage of the essence. Ketones were virtually absent and no methyl anthranilate was reported. This work was sub- stantiated by Stevens et a1. (1969) comparing composition of Grenache juices and Rosé wines. Analysis of trichloro- monofluoromethane (Freon ll) extracts, showed l-hexanol the most abundant of all the compounds present while large amounts of ethyl acetate and aldehydes were also present. In a study of White Riesling, another Yitis vinifera variety, diethyl ether extracts of the essence were shown to have alcohols as their major constituent. There was no methyl anthranilate found and ethyl and isoamyl acetates were the only esters reported (Van wyk et a1., 1967). As mentioned above, virtually all recent grape flavor research has been focused on the separation and isolation of the compounds present in essences. No serious attempt has been made to quantitatively correlate the presence, absence, or concentration of any one compound or group of compounds to the actual flavor potency of a given grape essence or its resultant product. Quantitation of various compounds and groups of compounds and their cor- relation to flavor has been attempted in this study. MATERIALS AND METHODS Materials Essences: Grape essences labelled 150-fold were obtained from three Michigan essence manufacturers and coded as follows: Manufacturer: A, B, or C. Type: P, Concord puree; J, Concord juice; N, Niagara juice. Year or date of production: 68, 1968, etc.; 10/8, October 8; UK, date unknown. An essence coded AP6910/6 for example, was manufactured by manufacturer A on October 6, 1969. Unless a specific date was given, it was assumed the essence was a representative composite of the entire year's production for that type. If the sample was stored at -23C until August 10, 1970, F was added to the code. Essences of 1968 grapes were obtained on August 12, 1969 and labelled AJ68, AP68, and BJ68. Each essence was transferred to one-pint glass bottles and stored at -23C until April 10, 1970. At that time, five bottles of each essence were removed from frozen storage and stored at 2C until analyzed. Two additional bottles of each essence 10 were removed from -23C storage on August 10, 1970, and stored at 2C as above. Essences of 1969 grapes were obtained on March 13, 1970, and labelled AJ69, AP69, AN69, AJ6910/2, AJ6910/7, AJ6910/8, AJ69UK, BJ69, CJ69. Each essence was transferred to glass bottles as above. Where essence quantities were sufficient, half the bottles of each essence were stored at -23C until August 10, 1970, when they were placed in 2C storage until analyzed. The remaining bottles were placed directly in 2C storage until analyzed. Concentrate: Two-gallons of Concord grape concen- trate, stripped of its essence, was obtained on March 13, 1970, and was of type AJ69; it was labelled by the manu- facturer as being 70° Brix but was measured by refactometry as 66.5° Brix. Juices: Three heat-processed and two frozen com- mercial Concord grape juices were purchased from a local supermarket for potential use as reference juices in flavor panel evaluations of the test essences. The following code was used to disguise the origin of each juice: Manufacturer: W, X, Y, or Z. Type: F, frozen; H, heat-processed. Methods of Analysis Steam distillation: Five ml of each cold essence (2C) was steam distilled in an all glass distillation ll apparatus of conventional design. A glass tube affixed to the end of the water cooled Graham condenser was immersed in ca. 30 m1 of distilled water in an iced 250 ml volu- metric flask. Approximately 200 ml of distillate were collected in about 10 minutes. The non-distillable residue of each essence was quantitatively transferred to a second 250 ml volumetric flask. Both flasks were made to volume with distilled water, labelled as volatile and non-volatile respectively, and stored at 2C until analyzed. Where essence quantities were sufficient, triplicate distillations were performed, otherwise, only one distillation was per- formed. Methyl anthranilate determination: A simple modi- fication of the method by White (1966) was used on both the volatile and non-volatile portions of each essence. Reagents: A. Hydrochloric acid (81 ml/100 m1) B. Sodium nitrite (3 g/200 ml H20) C. Hydrazine sulfate (5 g/200 ml H20) D. Potassium-l-naphthol-Z-sulfonate (5.6 g (90% practica1)/100 ml H20) E. Sodium carbonate (50 g/200 ml H20) Duplicate 10.0 ml aliquots of each essence fraction were transferred to 50 ml volumetric flasks and the above reagents were added in the following sequence: 1. Add 0.5 ml of A and 0.5 ml of B. Let stand exactly 2 minutes. 12 2. Add 1.5 m1 of C. Let stand exactly 1 minute. 3. Add 1.0 ml of D. Mix well. 4. Immediately add 1.5 ml of E. 5. Dilute to volume with distilled water and mix. After standing for 10 minutes at room temperature, the absorbance of each solution was read at 490 nm in 10 mm colorimeter tubes with a Bausch & Lomb Spectronic 70. 6. Results are expressed as ppm methyl anthranilate. The standard solution was prepared by dissolving 500 mg of methyl anthranilate in 50 ml of 95% ethanol. This solution was quantitatively trans- ferred to a 100 ml volumetric flask and diluted to volume with distilled water, yielding a 5 mg/ml solution. A 10 ug/ml solution was prepared from this solution by proper dilution; 1.0, 2.0, .... 7.0 ml aliquots of this solution were transferred to individual 50 ml volumetric flasks where the proper reagents were added as described above. The absorbance of each solution was measured and plotted as a function of methyl anthranilate con- centration. Total ester determination: The method of Thompson (1950), as described by Clore et a1. (1965), was applied to 2.0 ml duplicates of the volatile and non-volatile por- tions (or appropriate dilution thereof) of each essence. 13 After adding the usual reagents to the essences in 16 mm colorimeter tubes and thoroughly mixing, the absorbance of each was measured at 540 nm with a Bausch & Lomb Spectronic 70. Results are expressed as ppm ethyl acetate. The standard solution was prepared by dissolving 2.5 g of ethyl acetate in 50 ml of 95% ethanol. This solution was quantitatively transferred to a 100 ml volumetric flask and made to volume with distilled water, yielding a 25 mg/ml solution. From this solution, 25, 50, ..... 200 ug/ml solutions were prepared via proper dilutions; 2.0 ml of each of these solutions were then treated as described above. The absorbance of each solution was measured and plotted as a function of ppm ethyl acetate. Chemical oxygen demand (COD): The colorimetric method of McNary et a1. (1957) was applied to 50.0 ml duplicate samples of the volatile and non-volatile portions (or appropriate dilution thereof) of each essence. The absorbance of each sample was measured in 16 mm colori- meter tubes at 650 nm with a Bausch & Lomb Spectronic 70. Results are expressed as ppm COD. The standard solution was prepared by dissolving 2.0 g of reagent grade glucose, dried overnight at 105C, in distilled water and diluting to 100 ml in a volumetric flask. 10.0, 20.0, ...., 90.0 ml aliquots of this solution were each diluted to 200 ml in volumetric flasks, thus giving 106.7, 113.4, ...., 960.3 ppm COD respectively (1 g of glucose is equivalent to 14 1.067 g COD). Fifty ml of each solution were treated in the usual manner. The absorbance of each solution was measured and plotted as a function of ppm COD. Total carbonyl: The colorimetric method of Peleg and Mannheim (1970) was used to measure carbonyl levels of each essence. One change in the method was in carbonyl- free methanol preparation. Stock methanol was treated with 1 g of 2,4-dinitrophenylhydrazine and 4 g of trichloro- acetic acid per 500 ml, distilled through a ten-plate Oldershaw column, and stored in glass bottles. The authors reported the use of Girard P reagent to accomplish the same end. The test was applied to triplicate 1.0 ml aliquots of the volatile and non-volatile portions (or appropriate dilution thereof) of each essence. The results are ex- pressed as ppm acetone. The standard solution was prepared by weighting 1.25 g of acetone into 475 m1 of cold water in a 500 m1 volumetric flask. After diluting to volume with distilled water, aliquots of the solution were diluted to prepare 0, 5, 10, ...., 45 ppm solutions. One ml of each solution was treated in the usual manner. The absorbance for each solution was determined and plotted as a function of ppm acetone. Thin layer chromatographic separation and quantita- tion of acetone and acetaldehyde 2,4-dinitrophenylhydrazones: A modification of the method of Neuberg et a1. (1952) was used to prepare the DNPH derivatives. Triplicate 5.0 m1 15 aliquots of each essence were directly treated in 250 m1 separatory funnels with 2.0 to 4.0 m1 of the 2,4-dinitrophen- ylhydrazine reagent described by the authors. The amount of reagent used for each sample was determined by the results of the total carbonyl test. After standing at room temperature for 15 minutes, each solution was extracted with 5 X 15 m1 of carbonyl-free chloroform (distilled through a ten-plate Oldershaw column after treatment with l g of 2,4-dinitrophenylhydrazine and 4 g of trichloroacetic acid per 500 ml). The chloroform layers were combined in 250 ml standard taper Erlenmeyer flasks and evaporated to approximately 30 ml with a Bfichi rotary evaporator. Full aspirator vacuum was maintained in the evaporator while the bottom edge of the Erlenmeyer was placed in a 30C water bath. The concentrated extracts were then quantitatively transferred to 50 m1 volumetric flasks and diluted to volume with carbonyl-free chloroform. Preparation of thin layer plates: A silica gel GF-254 (Merck)/distilled water slurry (1:2) was blended at high speed in a Waring blendor for 1 minute and immediately spread on 10 cm X 20 cm glass plates; a Desaga/Brinkman spreader set at 0.5 mm was used. After setting overnight at room temperature, the plates were activated at 105C for 1 hour and stored in a desiccator at room temperature until used. Depending upon the derivative concentration, 5 to 20 ul of each extract was spotted on a single plate with l6 lambda pipettes. Triplicate plates were developed at room temperature to 10 cm in two consecutive solvent systems: System 1: Petroleum ether(75-92C)-diethyl ether- chloroform, 50:30:20 (v:v:v). System 2: Ethyl acetate-chloroform-hexane- methanol, 10:20:60:2.5 (v:v:v:v). Each developing tank was lined with filter paper to ensure tank saturation. The spots were visualized with ultraviolet light, removed with a Brinkman spot collector (cat. no. 0410139-8), and eluted into the collection flask with carbonyl-free chloroform. This solution was then quantitatively trans- ferred to a 5 m1 graduate cylinder and adjusted to 3 ml either by addition of carbonyl-free chloroform or by sol- vent evaporation with nitrogen. After transferring to 10 mm colorimeter tubes, each solution was measured for absorbance with a Bausch & Lomb Spectronic 70; the acet- aldehyde derivative was measured at 352 nm while the acetone derivative was measured at 358 nm. A blank value was obtained for each plate by removing absorbant from an unused portion of each plate at the same distance from the origin as the spot of interest. Identification was made by spotting known derivatives. Results are expressed as ppm acetaldehyde and acetone, respectively. The standard solu- tion was prepared by weighing 400 mg of each derivative into a 50 m1 beaker and dissolving in carbonyl-free 17 chloroform. After quantitatively transferring to a 100 ml volumetric flask and diluting to volume with carbonyl-free chloroform yielding a 0.4 ug/ul solution, 5, 10, ...., 40 pl aliquots of each were spotted on plates (6 repli- cates), developed, and treated in the usual manner. The absorbance of each resultant solution was determined and plotted as a function of ppm derivative. The conversion of ppm derivative to ppm acetaldehyde or acetone was made by multiplying ppm (derivative basis) by 0.1965 or 0.2438 respectively, the ratio of the molecular weight of each compound to that of its derivative. Ultraviolet absorption: Each essence was diluted, 1.0 ml to 25.0 ml, in a volumetric flask with distilled water. The absorbance of each solution was determined in 1 cm silica glass cuvettes with a Beckman DB spectrophoto- meter. Water was used as the reference. If the absorbance was above 1.00 for any essence solution, a 1.0 ml to 50.0 ml dilution of the essence was used. A Beckman strip chart recorder was used in the log mode to record absorb- ance directly. Results are expressed as absorbance at wavelengths where absorption peaks occurred. Gas chromatography: Gas-solid chromatography was utilized to separate various compounds present in the head- space over each essence. Chromatographic conditions were as follows: Instrument: Detectors: Columns: Carrier gas: Flow rates: Temperatures: 18 Hewlett Packard model 5750 research gas chromatograph equipped with a Mosley model 7127A recorder with Disc integrator. Dual flame ionization 1/8 in. X 10 ft. stainless steel packed with 80-100 mesh Chromosorb 101 (3.2 g/column). Columns were conditioned overnight at 250C with 10 cc/min helium flow. Helium Tank gauge pressure: 60 psig; flow rates adjusted via rotometers Column A: 35 cc/min Column B: Adjusted to provide proper base- line compensation during temperature programming. Hydrogen gauge pressure: 9 psig; flow rate: 28 cc/min Air gauge pressure: 25 psig; flow rate: 370 cc/min Injection port: 250C Detectors: 250C Collection vent: 285C Oven program: 80 to 200C Post injection interval: 2 min Linear program rate: 4C/min Upper limit interval: 10 min 19 Range: 102 Attenuation: Adjusted between 4 and 128 to obtain maxi- mum peak height less than 100 per cent recorder response. An effluent splitter was inserted between the end of column A and detector A, thus diverting 1/6 of the flow emerging from column A through the heated collection vent. This was necessary to maintain proper baseline compensa- tion during temperature programming. Headspace sampling and chromatographic analysis: Duplicate 10.0 ml samples of each essence (at 2C) were pipetted into each of two cold 30 m1 serum bottles. Single 10.0 ml samples were treated similarly for those essences where quantities were insufficient for duplicates. Each bottle was stoppered with a puncturable, resealable rubber stopper and heated in a vigorously stirred 50 i 0.2C water bath. After heating exactly 30 minutes, a 1.0 ml headspace vapor sample was withdrawn from the bottle headspace via a gas-tight syringe and injected into the chromatograph. After the oven program was begun, the baseline was adjusted as needed to provide zero recorder response. Peak identification: Where possible, peaks were tentatively identified by injecting known compounds and observing their retention times. When sufficient quanti- ties of given compounds were present, mass spectroscopic examination was used to provide positive identification. 20 This was accomplished by installing a single column in a LKB gas chromatograph coupled with a LKB 9000 magnetic deflection mass spectrometer. As compounds emerged from the column, they were ionized. The m/e was monitored via a Honeywell ultraviolet recorder. All data was recorded on magnetic tape and fed into a Digital 8/1 computer, which performed all calculations necessary to derive per cents abundance and total ionization. These results were plotted as bar graphs by a computer-controlled plotter. Peak quantitation: Identifiable peaks were quanti- tated using a Disc integrator; integrator counts were corrected for baseline drift with a drift corrector and adjusted for attenuation differences among peaks. Results are expressed as percentage total peak area. Flavor evaluation: Flavor panels were utilized to examine the typicalness of Concord grape flavor (relative to a reference juice), acceptability, and relative pre- ference of each essence; each essence was added to Concord grape concentrate diluted with cold tap water to yield single strength juice. Two sources of panelists were utilized. Initial flavor panels were each composed of twenty, randomly selected individuals from the Michigan State University Department of Food Science. A final "consumer panel" of fifty-two panelists was solicited from the author's apart- ment complex to examine consumer preference of juices prepared from two selected essences. 21 The soluble solids content, pH, and total titrat- able acidity were determined for each juice. Frozen commercial juices were diluted directly to 15.7-16.0° Brix with cold tap water. When necessary, citric acid was added to achieve a level of 0.70 per cent acid (as citric). In every case, each panelist was given approxi- mately 15 m1 of each juice in a small plastic cup. Each juice was kept iced until serving and was assigned a two- digit random number. Reference juice: Four of the juices purchased from a local supermarket were prepared as described above. Each panelist was given three pairs of juices, one pair at a time, at each of two sittings on successive days to determine juice acceptability and typicalness of Concord flavor. All possible pair combinations were presented to each panelist. The commercial juice possessing the most typical Concord flavor while still being acceptable was used as the reference juice in panels involving essences. The ballot accompanying each pair of juices is described as follows. 22 Concord Grape Juice Sensory Evaluation Before you are two samples of Concord grape juice. 1). Indicate which sample has the more typical grape flavor by placing an "X" in the appropriate box below. 2). Indicate whether each sample is acceptable or not acceptable by placing an "X" on the appropriate line below. Please rinse your mouth with water before tasting each sample. Sample Sample (Code) (Code) Acceptable Acceptable Not Acceptable Not Acceptable Essence-containing juices: Single strength, reconstituted juices were prepared using Concord grape con- centrate and each essence according to the following formula: 76.9 9 Concord grape concentrate. 2.0 m1 grape essence. Make to 319.8 g with cold tap water, yielding 300 m1 volume. In using this formulation, it was assumed that all essences were lSO-fold, as was indicated by the manufacturer. For each flavor panel, three reconstituted juices and the reference juice were judged together. At one panel sitting, panelists judged as to typicalness and 23 acceptability. At another sitting, panelists judged the relative preference of the juices by ranking. Ballots accompanying the juices in these panels are illustrated as follows: Typicalness and acceptability ballot: Flavor Difference Evaluation Instructions 1. Please make your evaluations based on your concept of typical grape flavor. 2. Determine the degree of flavor difference between each numbered sample and the reference sample R. a. If you do not detect any flavor difference, place a check opposite the words No Difference. b. If, in your judgment, any flavor difference does exist, place a check in one of the other four boxes opposite the term which best describes the degree of flavor difference. 3. After rating the flavor difference, place a check on one of the lines at the bottom of each column, indi- cating whether the flavor of the numbered sample is acceptable or not acceptable to you. Degree of Flavor Difference Sample Number Much More Typical Ill Slightly More Typical No Difference Slightly Less Typical II II Ill Much Less Typical Acceptable Not Acceptable |||l|JFlll ll lllllll 24 Relative preference ballot: Ranking Method of Flavor Evaluation Rank the samples in the order of how well you like them, giving the best sample or the one you like best a rank of l and rank the others below. You may use your own judg- ment whether to swallow or not to swallow the product, and the time to wait between samples. m l. 2. 3. 4O Statistical analysis of data: Analysis of vari- ance (Amerine et al., 1965a) was used to statistically analyze all chemical data. Since the number of replica- tions for distillations, extractions, and headspace vapor analysis were not equal for all essences, a randomized block design was used. Data showing significant sample differences were further subjected to Duncan's multiple range test (LeClerg, 1966). Where appropriate, coef- ficients of correlation were also determined (Amerine et a1., 1965b). Tukey's one-factor range test (Tukey, 1953) was used to test for differences of typicalness data obtained from flavor panels for the various juices. In each case, the reference juice was assigned a score of 3; The remainder of the scores were assigned as follows: much more 25 typical, 1; slightly more typical, 2; no difference, 3; slightly less typical, 4; much less typical, 5. The analysis was conducted on the basis of two samples, ie. the test sample vs the reference. The binomial test was used for acceptability scores (Amerine et al., 1965c). All rank tests were analyzed by using normal score transformation followed by analysis of variance (Li, 1957). Data showing significant sample dif- ferences were further subjected to Duncan's multiple range test (LeClerg, 1966). RESULTS AND DISCUSSIONS Steam distillation: The results of recovery studies of volatiles using the all-glass steam distillation apparatus are listed in Tables 1 through 3. Five ml each of a) 200 ug/ml methyl anthranilate solution, b) 25 mg/ml ethyl acetate solution, and c) 200 ug/ml methyl anthranilate- 25 mg/ml ethyl acetate solutions were distilled and quantitated. These solution concentrations were extra- polated (Holley et a1., 1955) as estimates of methyl anthranilate and total ester levels in lSO-fold essences. Table 1. Distillation recovery of methyl anthranilate. Solutions I II III Distillation pg MA/lO ml distillate 1 38.0 42.5 40.0 2 38.5 42.5 41.0 3 39.0 39.0 41.0 4 38.0 38.5 41.0 5 37.5 41.0 40.0 Mean 38.2 40.7 40.6 Reference 1 40.0 43.0 41.0 Reference 2 40.0 43.0 41.0 Mean 40.0 43.0 41.0 % Recovery 95.5 94.7 99.0 Mean % 95.7 26 27 Table 2. Distillation recovery of ethyl acetate. Solutions I II III Distillation ug EA/2 m1 distillate 1 910 850 900 2 910 900 950 3 920 920 920 4 960 920 980 5 920 900 940 Mean 924 898 938 Reference 1 980 950 980 Reference 2 960 950 1000 Mean 970 950 990 % Recovery 95.3 94.5 94.8 Mean % 94.8 Table 3. Distillation recovery of methyl anthranilate/ ethyl acetate. Solutions I II III Distillation (ug MA/lO m1 dist.)/(ug EA/2 ml dist.) 1 39/1000 41/950 41/980 2 40/1010 41/980 42/1010 3 40/1010 42/960 44/1020 4 42/1000 44/970 39/1000 5 42/1010 42/970 45/1020 Mean 40.6/1006 42.0/966 42.2/1006 Reference 1 41/1030 41/1000 42/1030 Reference 2 44/1040 45/1020 41/1040 Reference 3 41/1030 42/1030 41/1030 Mean 42.0/1033 42.7/1017 41.4/1033 % Recovery 96.7/97.4 98.5/95.0 102.0/97.4 Mean % 99.0/96.6 28 Recoveries were based on solution concentrations before distillation (labelled reference). Recoveries in excess of 95% could be consistently obtained if all glass joints were kept wet with water. No significant amount of methyl anthranilate or ethyl acetate was found in the residue of each distillation. Therefore, it was concluded that losses which did occur were not the result of incomplete or inef- ficient distillation, but from leakage through glass joints and/or failure to condense. Methyl anthranilate: Mean methyl anthranilate levels for each essence fraction are listed in Table 4. There were significant differences among the various essences at the 1% level. There were no significant dif- ferences among replicates. Complete data are listed in the Appendix in Table 19. Methyl anthranilate levels ranged from zero to 126 ppm in the volatile essence fractions. The levels in many essences were not significantly different from each other and there were no consistant trends as to manufacturer, type, or year of the essence. Daily production samples of essence AJ69 showed highly significant differences in methyl anthranilate levels. Frozen storage had no apparent effect upon the retention of methyl anthranilate in the essences. No methyl anthranilate was found in non-volatile fractions of the essences. 29 Table 4. Methyl anthranilate concentrations for the volatile fraction of essences. Statistical Statistical 1 Essence MA Significance Essence MA Significance ppm 5% 1% ppm 5% 1% AJ68F o3 a a AP69F 293 fg fg AJ68 42 b a CJ69 322 gh g BJ69 92 c c BJ68F 333 h gh AP68 102 c cd CJ69F 343 h h AP68F 113 cd cd AJ69UKF 353 h h BJ69F 143 d d BJ68 472 i i AJ69F 223 e e AJ6910/8 473 i i AJ69 222 e e AJ6910/7 543 j j AP69 272 f f AJ6910/2 783 k k AJ69UK 283 f f AN69 1262 1 1 lLike letters denote no significant difference among essences. 2Mean value of 6 determinations. 3Mean value of 2 determinations. Holley et a1. (1955) reported a methyl anthranilate concentration of 33 ppm for 83-fold Concord essence pre- pared in their laboratory. If this value is adjusted to a 150-fold level, a theoretical value of 60 ppm is obtained. This level, when coupled with the 0.80-1.49 ppm value for single strength juice reported by Scott (1923), is generally greater than methyl anthranilate concentrations in essences 30 examined in this study. This was possibly due to the relatively long storage and handling during commercial preparation. Essence AN69, having the highest methyl anthranilate content, was an essence stripped from the juice of Niagara grapes, a white labrusca variety, and was included in the study primarily for methyl anthranilate comparisons. Total esters: Mean total ester levels for each essence examined are given in Table 5. Total volatile ester levels of the essences ranged from 100 to 11,400 ppm and indicated very little similarity among samples. At the 5% level, only two pairs of essences (BJ68F, BJ68 and BJ69, AJ69F) were not significantly different. As with methyl anthranilate levels, there was no relationship between manufacturer, year, or freezing storage, and the total ester level. Total ester levels were significantly different (1% level) among daily production samples of essence AJ69. Essences prepared from puree, however, did contain generally greater total ester levels in their volatile fractions than those essences prepared from juices. Although several of the non-volatile fractions showed trace amounts of esters, the values obtained were within the experimental error of the procedure and may be considered insignificant. If, as above, the 83-fold essence examined by Holley et a1. (1955) is converted via calculations to a 31 Table 5. Total ester concentrations for the volatile fraction of essences. Statistical Statistical Essence EA Significance Essence EA Significance ppm 5% 1% ppm 5% 1% 3 3 . . BJ68F 100 a a AJ69UK 4600 j 1 BJ68 1002 a a AJ69 51082 k j AJ6910/2 2503 b b BJ69 53332 1 k AJ6910/7 7003 c c AJ69F 54003 1 k AJ68 9252 d a BJ69F 56753 m 1 AJ68F 10003 e d AP68 58002 n m AN69 16832 f e AP68F 62853 o n CJ69 23172 g f AP69F 88503 p o AJ69UKF 26003 h g AP69 91422 q p CJ69F 41003 i h AJ6910/8 11,400 r q lLike letters denote no significant difference among essences. 2Mean value of 6 determinations. 3Mean value of 2 determinations. 150-fold base, a theoretical ethyl acetate value of approxi- mately 6300 ppm is obtained. This essence therefore contained a relatively high total ester concentration when compared to those of essences examined here. Again, this difference is possibly due to essence age and method of preparation. 32 Chemical oxygen demand: Mean chemical oxygen demand (COD) levels for the essences examined ranged from 20,000 to 205,000 ppm and are listed in Table 6. No signi- ficant differences were present among replicates. Samples, however, were significantly different at the 1% level and suggested gross differences between essence folds. Puree essences, for example, exhibited significantly lower COD values than their juice counterparts. Thus, it was appar- ent that some component or series of components present in Concord puree and not present in juice was a cause of less efficient volatile recoveries for essences from puree than from juice. There were no apparent trends in COD as to the year or manufacturer of essences. However, as with methyl anthranilate and total volatile esters, COD values were significantly different (1% level) for daily produc- tion samples of essence AJ69. Complete data are given in the Appendix in Table 23. No oxidizable organic material remained in the non- volatile portion of any essence other than AJ68 and AJ68F. The levels were 2500 and 3000 ppm respectively and were probably due to a small amount of debris observed in these samples. Several authors have reported the use of chemical oxygen demand to indicate the concentration of organic volatiles present in fruit juices. Jensen (1961) reported a positive relationship between "oxidation number" (COD) 33 Table 6. Chemical oxygen demand for the volatile fraction of essences. Statistical Statistical Essence COD Significance Essence COD Significance ppm 5% 1% ppm 5% 1% AP68F 20,6503 a a AJ6910/7 74,0503 j AP68 22,1252 b a AJ68F 77,0503 k AP69F 26,5253 c b AJ68 77,1002 k AP69 27,3752 c b BJ69 85,9332 1 BJ68F 33,7503 d c BJ69F 89,1503 m BJ68 38,1252 e d CJ69 99,8002 n AN69 47,0502 f e AJ69UKF 105,0003 0 AJ69F 66,0003 g f AJ69UK 107,2003 p AJ69 68,1672 h CJ69F 109,6003 q AJ69lO/2 71,3503 i h AJ6910/8 204,8003 r q 1 Like letters denote no significant difference among essences. 2Mean value of 6 determinations. 3Mean value of 2 determinations. and the "fold" of volatiles in apple concentrates. This value, however, was not necessarily related to the flavor enhancing quality of essences and ethanol was suggested as a contributor to this discrepancy. Charley (1962) made a correction in oxidation number for ethanol and derived the term "aroma number", a value which more closely correlated with the true flavor contribution of the essence. He also suggested that carbonyls and esters be determined separately to help better understand the relationship between oxida- tion number, aroma number, and the flavor enhancing quality of essences. Total carbonyl: The mean total carbonyl level for each essence is listed in Table 7. Table 7. Total carbonyl levels Total carbonyl concentrations for the volatile fraction of essences. Statistical Statistical Essence Acetone Significance Essence Acetone Significance Ppm 5% 1% ppm 5% 1% BJ69F 2133 a a AP68F 6203 e fg BJ69 2572 ab ab AP68 7182 f BJ68F 2823 ab ab AJ69F 9583 h CJ69 2982 abc ab AJ69 11032 h l AJ6910/2 3083 abc ab AN69 11292 hi i BJ68 3382 bc abc AJ69UKF 12183 i 1 CJ69F 3983 cd bcd AJ69UK 15083 3 j AJ6910/7 4623 d cd AJ68 22642 k k AP69F 4923 a de AJ68F 23473 k k AP69 5962 e ef AJ6910/8 66133 1 1 1Like letters denote no significant difference among essences. 2 3Mean Mean value of 9 determinations. value of 3 determinations. 35 for volatile fractions ranged from approximately 200 to 6600 ppm and indicated definite trends as to the essence type and manufacturer. None of the essences from purees had high carbonyl levels while the essences from juices made by the same manufacturer (A) had relatively high levels of carbonyls. There were no significant differences among distillates or replicates. Complete data are listed in the Appendix in Table 25. Although several non-volatile fractions showed trace amounts of carbonyls, the values obtained were well within the experimental error of the procedure and may be considered insignificant. The presence of carbonyls in grape essence has been reported in the literature by several authors. Stern et a1. (1967) mentioned no fewer than six such compounds were present in isopentane extracts of Concord essence but made no effort to quantitate or correlate any of these compounds to their relative importance of flavor contri- bution. Thirteen carbonyls, eight of them aldehydes, were separated and identified from ethyl chloride extracts of Concord grape juices by Neudoreffer et a1. (1965). These compounds were present in relative amounts ranging from large to trace. n-Valeraldehyde was correlated with one group of essences having an undesirable flavor anomoly. 36 This study exemplified the importance of carbonyls as a contributor to Concord grape flavor. Acetone and acetaldehyde 2,4-dinitrophenylhydra- zones: Typical separations of acetone and acetaldehyde 2,4-dinitrophenylhydrazones via thin layer chromatography are illustrated in Figure 1. (To facilitate photography, plates were sprayed with 0.5% Rhodamine B in 95% ethanol.) The two derivatives were identified by spotting recrystal- lized knowns prepared from stock laboratory reagent. Upon spraying with 10% potassium hydroxide in 95% ehtanol, the acetone derivative turned dark brown while the acetaldehyde derivative showed a distinctive red-brown; both colors were short lived. When plates were developed exactly 10 cm in each solvent system, Rf values of 0.62 and 0.58 were obtained for acetone and acetaldehyde 2,4-dinitrophenyl- hydrazones respectively. Acetone and acetaldehyde concentrations obtained by this method of separation are listed in Table 8. The acetaldehyde values ranged from 18 to 2811 ppm. There were no statistically significant differences among plates. Complete data are listed in the Appendix in Table 28. It was found necessary to recrystallize the 2,4 DNPH before use since some lots gave positive acetalde- hyde values. For example, one lot gave values of 127 ppm acetaldehyde for each ml of reagent used in derivative pre- paration. No acetone derivative was obtained with the reagent. .monenmuumgamnwnmouuwcflcuv.N oocwmmo mo aofiumummmm canmmumoumfiouso nomad sane .H musmwm .occumuchnahconmouuficfiuae.N wohnowamuwom monoconse .econuomnamconmouuflcflolw.N ocouoom monocont .I. u v v w an . v v v v D D D d D P. d f D D P. D P. D N d D P. d D 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 6 6 6 6 6 8 8 8 6 6 6 6 6 6 6 6 6 8 8 8 0 a... J J J J J J n I T. I X X 0 0 0 d / / / 8 L Z noncommm 38 Table 8. Acetaldehyde and acetone concentrations of the essences. Statistical Statistical Essence Acetaldehyde Significance Acetone Significance ppm 5% 1% ppm 5% 1% BJ68 182 a a 2742 hi gh BJ69 352 ab 1362 d de BJ69F 432 ab 2112 fg f AP68 542 ab ab 4402 i CJ69F 552 ab ab 02 a a BJ68F 872 be ab 592 b abc AJ6910/2 1053 c b 473 ab ab AP69 2052 d c 3842 j i CJ69 2102 d c 842 bc bcd AP68F 2532 d c 2522 gh gh AP69F 2552 a c 1362 d de AJ6910/7 3093 e a 1113 cd cd AJ69F 3782 f e 1472 de de AJ69 4192 fg e 1842 ef ef AN69 4352 g e 3092 i h AJ69UKF 8073 h f 1923 ef ef AJ69UK 8153 h f 2143 fg fg AJ68 12882 i g 102 AJ68F 14882 j h o2 AJ6910/8 28113 k i 33 among essences. 2Mean value of 9 determinations. Mean value of 3 determinations. lLike letters denote no significant difference 39 The acetone values ranged from 0 to 440 ppm. Signi- ficant differences at the 5% level were found between the acetone extractions. There were no significant differences between the plates. Further analysis showed sample X extraction interaction significance at the 1% level. Extractions 1 and 3 were significantly different but not extractions l and 2 or 2 and 3. These differences may be due to variations in the time delay between derivative preparation and spotting and/or to the method of concen- tration. They were not considered extremely important as sample differences for acetone and acetaldehyde levels were quite large between essences. It is felt, however, a method of separation, eg. column chromatography, elimi- nating concentration after extraction, would be more desirable than the method used. Holley et a1. (1955) reported that their 83-fold Concord essence contained 300 and 30 ppm for acetone and acetaldehyde respectively. Placing these values on a 150-fold basis yields theoretical levels of 542 ppm acetone and 54 ppm acetaldehyde. The acetone levels for the essences examined in the current study were generally lower than those of this essence. The acetaldehyde levels, however were considerably greater. Relative essence age, methods of essence production, condition of grapes and/or juice prior to essence preparation, and methods of deriva- tive preparation could have contributed to these differences. 4O Carbonyl concentrations obtained by analysis of each essence were converted to u moles/m1 essence by dividing each respective value by the proper molecular weight. These conversions are listed in Table 9 and indi- cate the relative contributions of acetone, acetaldehyde, and other carbonyls to the total carbonyl level of each essence. Ultraviolet absorbance: Mean absorbancies of each essence at absorption wavelengths of the ultraviolet spec- trum from 200 to 300 nm are listed in Table 10. There were no statistically significant differences among repli- cates. Complete data are listed in the Appendix in Table 29. Figures 2 and 3 illustrate the typical ultraviolet absorption spectrum for each essence. Each essence absorbed very sharply in the region of 205-215 nm. At 243 nm, however, not all essences exhibited a distinct absorption peak. The ratio of absorbance at 210-213 nm to the absorbance at 243 nm are also listed in Table 10. Holley et a1. (1955) reported ultraviolet absorb- ance spectra for natural and synthetic essences. The spectra were identical except a maximum at 280 nm in the natural essence was replaced by a minimum at 270 nm in the synthetic essence. Although absorption peaks at these wavelengths were not observed for essences examined in 41 Table 9. Carbonyl data. Column 1 2 3 4 Total Acetaldehyde Acetone Total Carbonyls (2+3) Essence u moles u moles u moles u moles m m m m BJ69F 3.67 0.98 3.64 4.62 BJ69 4.43 0.80 2.35 3.15 BJ68F 4.86 1.98 1.02 3.00 CJ69 5.14 4.77 1.45 6.22 AJ6910/2 5.31 2.39 0.81 3.20 BJ68 5.83 0.41 4.72 5.13 CJ69F 6.86 1.25 0.00 1.25 AJ6910/7 7.97 7.02 1.92 8.94 AP69F 8.49 5.79 2.35 8.14 AP69 10.29 4.66 6.62 11.28 AP68F 10.69 5.75 4.35 10.10 AP68 12.38 1.23 7.59 8.82 AJ69F 16.51 8.59 2.54 11.13 AJ69 19.04 9.52 3.18 12.70 AN69 19.44 9.89 5.33 15.22 AJ69UKF 20.99 18.35 3.31 21.66 AJ69UK 25.99 18.51 3.69 22.20 AJ68 39.02 29.25 0.17 29.42 AJ68F 40.50 33.80 0.00 33.80 AJ6910/8 114.00 42 Table 10. Ultraviolet absorbancies of essences. Absorbance2 Ratio Statistical Statistical Essencé‘ at at Significance at Significance 210-213nm 205nm 213nm 5% 1% 243nm 5% 1% 243nm AJ68 0.435 ----- -- -- ----- -- -- ---- AJ68F 0.447 ----- -- -- ----- -- -- ---- CJ69 ----- 0.312 a 0.118 a ab 2.64 BJ69 ----- 0.315 ab 0.100 a a 3.15 BJ69F ----- 0.320 ab 0.110 a a 2.91 CJ69F ----- 0.322 ab ab 0.125 a abc 2.68 BJ68F ----- 0.430 ab abc 0.172 b cd 2.50 AP68 ----- 0.442 bc bcd 0.1554 -- -- 2.85 BJ68 ----- 0.457 cd 0.177 b d 2.58 AP68F ----- 0.462 de 0.1934 -- -- 2.39 AJ69 ----- 0.485 d ef 0.1704 -- -- 2.85 AJ69F ----- 0.489 de ef 0.1704 -- -- 2.88 AP69 ----- 0.504 de f 0.1604 -- -- 3.15 AP69F ----- 0.509 e f 0.1634 -- -- 3.12 AJ6910/7 ----- 0.557 f g 0.205 c de 2.77 AJ69UK ----- 0.579 g gh 0.165 b bcd 3.51 AJ69UKF ----- 0.602 h h 0.195 b d 5.04 AJ6910/2 ----- 0.699 i i 0.248 d e 2.79 AJ6910/8 ----- 0.876 j j 0.2874 -- -- 3.05 AN69 ----- 1.300 k k 0.463 e f 2.67 lEssences were diluted 1.0 ml to 25.0 ml with water. 2Mean values of 3 determinations. 3Like letters denote no significant difference among essences. 4No point of inflection or peak was present. This value was used only to compute the absorbance ratio and was not included in the statistical analysis. 43 .9. .9.F .91? It. 087* .80 .1. .77b .74- 0" 8 8 05‘” 5 8 e e 8 .e g 03" 02“ 01" 200 2i0 240 260 260 300 200 2i0‘ 240. 260 230 330 300 7230 :47 i 7: 1 Wavelength nn. Wavelength nu. «£1.82... :0 30 M68 1:25 dilution A968 1:25 dilution ”5. 1325611661.... OSWP .91. .STp '8‘. 081h 0.5? e71h 07‘P .1‘* 061- .“. 8 051 5.51b .e 3 .44 || 4... £0 .31 .34 .2-1 .24;- ‘1‘ 01‘? 200 22:0 241) 261) 2:80 3:00 0200%2120 2:40 2: : i o 4' 4 i A ‘77 60 280 300 20 ' Wavelength nu. Wavelength nn. 0 220 $321.32). :20 300 AJ69 1:25 dilution AP69 1:25 dilution 3.159 1,35 dilution Figure 2. Ultraviolet absorption spectra of 6 dilute essences. .11? o 4 A L l g :00 :20 310 260 :30 300 Wavelength nu. CJ69 1:25 dilution o J l l A 200 :20 250 260 230 360 Wavelength nu. AJ69 10/: 1.25 dilution 444 o 1 A 200 :50 210 260 230 330 Wavelength nu. AJ69 10/2 1:25 dilution o ., - 200 :30 220 2 Bo Wavelength nu. AJ69UK 1:25 dilution L 230 j. 300 l - l L l J 200 220 2:0 :60 :30 130 a, 1? Wavelength nu. AJ69 10/7 1:25 dilution 210 210 ‘u'o 23 380 Wavelength nu. 8W6, 1:50 dilution Figure 3. Ultraviolet absorption spectra of 6 dilute (BEBSHBIICNBEB. 45 this study, the slight absorption at 243 nm was reported by Holley et a1. Absorption in the 200-300 nm region is one of very strong absorbance by conjugated unsaturation and/or unsatu- rated aldehydes (Roberts and Caserio, 1965). Neudoreffer (1965) reported the presence of acetoin and crotonal in his study of Concord grape essence. Thus it is possible that peaks in the 200-300 nm region could be attributed to the presence of one or both of these compounds or to compounds of similar structure. Gas-solid chromatography: Compounds in the head- space over each essence which were separated and quanti- tated using gas-solid chromatography are listed in Table 11. In all, 7 peaks, ie. peaks A-D, F, H, and I, were posi- tively identified via mass spectrometry, retention times, and essence enrichment. Formaldehyde and n-propanol, peaks E and G respectively, were not present over the essences in quantities sufficient to permit proper m/e analysis and were identified using retention times and essence enrichment only. Peaks K and L were not identi- fied but were due to compounds of known masses. Peak K represented a compound of mass 70 while peak L repre- sented a mixture of compounds of masses 88 and 116. Further identification was not possible because of instru- ment limitations. Figures 4 and 5 illustrate typical mass spectra obtained. 46 100 .. 7 ~34 80 J _ 60 - _. 22 40._ _ % 20‘- __ l ‘ ‘I l 20 30 40 50 Figure 4. Mass spectrum of gas-solid chromatographic peak B, acetaldehyde. 1003' F34 BO-j _ 60.. _ z 40 " % j 20- r' 1 .I [ill 1 ll _ l l l I I 20 30 40 50 Figure 5. Mass spectrum of gas-solid chromatographic peak C, ethanol. 47 Table 11. GaS*solid chromatographic peak identities. Peak Compound Peak Compound A Methanol G n-Propanol B Acetaldehyde H Ethyl acetate C Ethanol I Iso-butanol D Acetone J * E Formaldehyde K * F Methyl acetate L * * Not identified. Figures 6 through 17 illustrate gas chromatographic separations for each essence headspace. Numbers at the tip of many peaks indicate the factor (attenuation) by which the area of that peak must be multiplied to normalize its area with that of the non-numbered peaks; the greatest instrument sensitivity used for any given essence analysis was range 102 and attenuation 4. Each peak is expressed as mean percentage total Disc integrator count in Table 12. No adjustments were made to account for sensitivity dif- ferences the flame detectors have for the various compounds quantitated. Headspace vapor samples are the simplest, and most precise method of sampling a food aroma for the chemical analysis. The aroma of a particular food product depends not only upon the qualitative nature of the compounds in DOwF 48 00+ 34 ll. 704» ' X32 ‘0”- 82 lsoii “004 i in») .1 T 10 0 U . t L .1 WM- If U V U r 7— K °c L1” 0‘: 196 1:0 :36 152 :1: :94 :90 ago nu. '0 3 3 :5 Y6 :5 2'4 50 5: in Figure 6. Gas-chromatographic separation of headspace volatiles over essence AJ68. ’0 I X32 1 ‘mb :3 in. 816 i 49.. 3 .. o ’0" 20" 104» ° ‘ 1.1 l>§liL— L- L4/\h~"‘=u-[j\-ullfi=h__. I C D P I .‘l ‘ I. °cveo on I, 134 :30 1‘35 1;: :90 up 290 :00 nun. o '0 'e 13 1'6 30 :3 2'. a; - 35 Figure 7. Gas-solid chromatographic separation 0f head- space volatiles over essence AP68. 49 ’01P @1th L1 c f1 I T. - so no : 1 1 a a o c _ A e! 42' 10 :gg 1;: :‘e 1!, 90 24 Win. a a e 1: :6 :o :4 a: a: 3: Figure 8. Gas-solid chromatographic separation of head- space volatiles over essence BJ68. ’01' .01? 701p KC ‘0 0 X13 0 header Wee u N if I c o ”a? e a a x L -c :0 no ll 12} 130 is: 152 1:: 1:4 zoo zoo via. 0' a i :2 is So 5: :e :2 E Figure 9. Gas-solid chromatographic separation of head- space volatiles over essence AJ69. 50 ’°.[ no.1 704L . T 1 3° 4L 10 m .haso so: mosuowamou .nsofiumswfiuouoc N no osmoa mus nosHM>N .moosunno csOEo oosouUMMfic usoowmwsmao 0s ouosoc uuouuod osta on on mH.o w m 00.0 m m mm.H mm s mo.mo so a mH.H U o -.v mummnc o no v~.o w m oo.o s « mh.m mo 0 mn.om a m mm.o n n wo.~ mmmwhm mm mm mv.c a m oo.o n n ao.~ m m oo.va n s om.H o m vm.o mmmmms no G om.o a a oo.o o o vm.~ o o ha.vm on o vv.v n s Hm.~ mmmwns mm mm w¢.o m m oo.o s : ah.m so a mm.va on on hh.H m m na.o mmmoms u m ~m.o m m oo.o p p ~m.~ mm s wv.hw s s mv.ma n n m~.~ mmmons woo on ~m.o w m oo.o s n mo.~ o w mm.mm w o mm.v o c w~.v mxsmohs s ms mq.o a a oo.o m m hm.~ c on am.>v o m ~m.m o o m¢.m mmxosamna a m no.0 o o va.o w n on.m so a mm.vh o o Nm.m m w oo.h mn\oamoh¢ m o oo.o n n mH.o H h as.m s m oo.sm o o mm.~ w u ms.h mwxosmmna no u mm.o o m oo.o m m mm.a m m mH.Nw n n mv.a to o vm.m mambc o no h~.o m w too.o s s oh.~ on o hm.om o m vm.o n n oo.~ Nmmhm m s oa.o m o oo.o x H om.m o c 5H.mv m m wo.> o o mp.m Namzs moo ow mv.o m m oo.o o o m~.~ no m ~m.mH n n o¢.H o m mm.o «moms moo mo He.o m m oo.o m m mw.~ o o vv.mm op mo mo.v o o 5H.m mamas no no mo.o o m oo.o o o om.~ mm s ha.mw n n vo.a m o Hm.o Nmmbm moo mo Hv.o m w oo.o n x om.m n n m~.HN a n mo.H m m ~m.o «moms n ma mv.o o m oo.o m m Hm.~ v as Ho.Hh s s om.~a n n mm.~ Nmmns ma am w wH mm a wH wm m ma mm a wa am a «A wm m Wosoowmasmwm assoc Wosmowmwsofim assoc Hoosmoflmwsmflm assoc Wcsmowmwsmufiw assoc Wosmowmwsmwm assoc oosoowmwsmfim assoc ousommm HMOMuuwumum Hosea Hsowummumuw Hmvom Hmcmmmwuoum Hmuom Hobaumwumum Hmumm Hobaummumum HouomsWHoowummumum1H0Mmm .m m D U m < xmom ‘ .asmmumoumfiouso mom you momsusoouom ussoo uoumumousH .NH wanna 55 .haso on: muoowdoou .usowsmswauouos oaoswu who mosao> n .aHso no: mwumowumou .nsOwusswEHouoo N no asses mum mosHm>N .uoosommo ososo ousoumwuwo assowmwsowm 0s ouoswo muouuoa oxwad m s 00.0 m cons no.0 as n o0.o m M 00.0 n U 0H.nN as as No.0 nhoooc o s oo.o o no No.0 o o m~.o a o 00.0 cu m oo.nv no on no.0 nhooom v 0 an.o w m 00.0 m u vm.0 w m 00.0 m M om.0o m M 00.0 nmoomd no on 00.0 m vcnm 00.0 m 0 on.0 m m 00.0 c m oh.mm no as No.0 nhooofl 0 w v0.0 m cons no.0 n 0 0H.H o a 00.0 M « nN.hh a m 00.0 nhoomd m m 00.0 on w hN.0 000 @ NN.0 m m 00.0 m n 00.NH no on no.0 nmooofl no no No.0 as $00 nH.o 0000 on mH.o a m 00.0 n 0 on.Nn as Dan no.0 nKDooofl o m 00.0 m cons $0.0 won on VH.0 m m 00.0 On 0 00.nn am 03 v0.0 no\oHooo¢ m m 00.0 can 00 mH.o m a 00.0 s m 00.0 o no mn.0H as can no.0 nh\0Hooot m s 00.0 Dam on mH.0 m m 00.0 m m 00.0 m m m0.m no can no.0 nN\oHooo¢ m a 00.0 no 000A NH.0 on n NH.0 m m 00.0 n 00 o0.vN AM on no.0 Nooou s m 00.0 m can v0.0 m0 6 vN.o M m 00.0 no u oo.nv no can no.0 Nooom c O 0a.0 c H 0n.0 w w Nv.0 m m 00.0 a mo o0.on as bad no.0 Nowzfl U 0 nn.0 m m 00.0 m m vm.0 s m 00.0 M w no.oh m m 00.0 No0m< n c 00.0 on ma 0N.0 m 0 on.0 m m 00.0 m0 0 hN.0m no as No.0 Noood u @ HH.~H w n mv.o m m 00.0 a n 0H.H m m hm.v c w on.o nooom 0 u nv.0 w m 00.0 a v nH.H w m 00.0 m a 0v.0h s m 00.0 N00m< m s 00.0 v m mm.o 000$ 00 oH.0 s m 00.0 m as ho.o n O 00.0 mmood «H mm 0 ad on 0 ad om v «H on a MA am w w# wm w ousoowwwsowm assoc oosoowwwsOMm assoc oossowuwsofim ussoc cosmowmwsoww assocfwosoowwwsowm assoc ousoowwwsowm assoc oosommm A Hmowumwuoum Hmuom H mommmoum smacks. H Hmowummumum smacks. H Honoummumum H38. Hmowumwuoum 138. H scumumwumum HEW A w h H m U xmom A.usocv .NH wansfi 56 the product, but also upon their concentration and relative ratios. There is a relationship between the concentration of a specific compound in the vapor phase at a given tem- perature and the vapor pressure of the compound, the type of medium in which it is distributed, its degree of solu- bility in the medium, its concentration in the medium, and its miscibility with other organic compounds in the mix- ture (Nawar, 1966). Kepner et a1. (1964) reported a method of quantitation using direct headspace sampling and stand- ard curves prepared from predetermined solvent systems closely resembling the product being examined. However, because of the complexity of the essence system examined in this study, integrator counts for each peak were ad- justed for instrument attenuation and expressed as a percentage of the sum total integrator counts for all peaks. The resultant percentages were used as a means of quantitation. Inspection of Table 12 and Figures 6 through 17 indicate each essence headspace contains virtually the same compounds. Quantities of each do vary considerably among essences although each was supposedly a lSO-fold product. The headspace over essence BJ68 (Figure 6) was rather unusual in respect to the number of major peaks it produced. Peak I was present in this sample while not present in any other. Peaks K and L (two compounds) were 57 present in other essence headspaces, but in levels several orders of magnitude less than that of BJ68. These two peaks represent unidentified compounds. The headspace over essence AJ6910/8 (Figure 14) contained something causing a broad "peak" to elute from the column at approximately 160C. This "peak" was not present in any other essence and has been neither identi- fied nor quantitated. To place the gas-solid chromatographic data in proper perspective, it should be noted that human olfac- tory sensitivities for many compounds are often much greater than that of a flame ionization detector. Thus, an organoleptically important compound may not even be detected with the flame ionization detector (Flath et a1. 1969). Olfactory thresholds are of primary concern in the determination of the relative contribution of any given compound to a products flavor (Guadagni et a1. 1963). Guadagni et a1. (1966) reported that some of the smaller peaks in gas chromatograms represent compounds giving the greatest odor intensity. Headspace vapor analysis does tend to overemphasize more volatile components present within the system, thus a low threshold compound respon- sible for the primary flavor of an essence may not be sampled at all (Forss et a1., 1967). Flavor panels: Flavor panel data for selection of a commercially available juice to be used as the reference 58 in flavor panel evaluation of juices prepared from essences are listed in Table 13. Emphasis for the juice selected as the reference was placed on acceptability as opposed to preference; the preference of any given juice is, by definition, dependent upon the juice with which it is paired in flavor panels. The acceptability test gives a more accurate indication of the intrinsic quality of a juice and is independent of other juices. Of the juices examined for reference juice use, two possessed highly acceptable flavors and were both pro- duced by manufacturer W. Juice WH, heat processed and bottled, was significantly acceptable on each of three presentations to semi-trained panelists. Similarly, juice WF, a frozen concentrate, was significantly acceptable on each of six presentations to the same panelists. When compared against each other in a paired comparison test, there was no significant difference between the two, thus indicating they were fully equivalent. Because of the ease of preparation of the heat-processed juice for flavor panels, it was chosen as the reference juice. Other juices were significantly inferior than these two in pre- ference and acceptability. Flavor panel data for flavor difference (relative to the reference juice) and acceptability of juices pre- pared from each essence are listed in Table 14. The method of scoring used for flavor difference panels was 59 Table 13. Flavor panel data for reference juice determination. Number of Number Number Number Juice Panelists Preferred Acceptable Not Acceptable Test 1 XH 18 14* 15* .- 3 YH 4 9 9 WF 17 13* 13* 4 YH 4 10 7 YH 18 7 9 9 ZF 11 ll 7 Test 2 ZF 19 10 17** 2 WF 9 l7** 2 WF 20 11 18** 2 XH 9 ll 9 ZF 20 ll 13 7 XH 9 7 9 Test 3 WH 20 15* 19** l XH 5 11 9 WF 19 11 14* 5 XH 8 12 7 XH 19 ll 10 9 ZF 8 9 10 Test 4 WF 20 ll 16* 4 WH 9 15* 4 ZF 20 6 13 7 WH 14 16* 3 WF 20 14 17** 3 2F 6 16** 2 *Indicates significance at the 5% level. **Indicates significance at the 1% level. 60 Table 14. Flavor panel data for flavor difference and acceptability. Flavor Number of Difference Number Number Essence Panelists Scorel Acceptable Not Acceptable Panel 1 20 BJ68 77* 14 6 AJ68 84* 8 12 AP68 87* 12 8 Panel 2 20 BJ69 64 15* 5 AJ69 77* 13 7 AP69 92* 6 14 Panel 3 20 None 72* 16* 4 CJ69 67* 17* 3 AN69 82* 7 13 Panel 4 20 AJ6910/2 77* 15* 5 AJ6910/8 84* 6 14 AJ69UK 79* 16* 4 Panel 5 l9 AJ69lO/7 75* 12 7 CJ69F 72* 14 5 AJ69UKF 82* 11 8 Panel 6 20 AJ68F 88* 6 14 AP68F 75* 14 6 AP69F 89* 7 13 Panel 7 17 BJ68 67* 10 7 BJ69 63 16** 1 CJ69 60 14* 3 Panel 8 19 BJ68 72* 14* 5 CJ69 69* 18** 1 BJ69 75* 15* 4 *Indicates significance at the 5% level. **Indicates significance at the 1% level. 1Relative to reference juice WH which was assigned a score of 60. 61 such that juices having Concord flavor less typical than the reference juice received greater numerical scores than did juices having more typical Concord flavor. For statis- tical purposes, the reference juice was assigned a flavor difference total of 60 (20 member panel). No attempt was made to determine if sample differences existed between flavor difference scores for juices prepared from essences; it is possible that two juices, one being much less typical than the other, could have maximum flavor difference scores of 100 (20 member panel). Flavor difference totals indicated that juices pre- pared from essences were generally less typical in Concord flavor than the reference juice. There were, however, two exceptions from this generalization; essences BJ69 and CJ69 were the only essences tested not significantly different from the reference juice on each occasion presented to the semi-trained panel. These juices were certainly two of the better prepared from essences in this panel series. The acceptability of these juices, measured in the same series of panels for typicalness of Concord flavor, indicated essences BJ69 and CJ69 acceptable on every occasion they were presented to the panel. These results substantiated the relatively good quality of juices pre- pared from essences BJ69 and CJ69. Other acceptable juices were made from essences AJ6910/2, AJ69UK, BJ68, and concen- trate diluted as usual but with no essence. 62 The acceptability of juice prepared using no essence whatsoever indicated the concentrate used had greater than minimum threshold concentrations of flavor compounds essential to acceptable Concord grape flavor. Analysis of the stripped concentrate diluted to 15° Brix indicated virtually no methyl anthranilate and approxi- mately 7 ppm total volatile esters (as ethyl acetate). The reference juice contained 2 ppm and 12 ppm methyl anthranilate and total volatile esters respectively. Methyl anthranilate (Scott, 1923, and Clore, 1965) and total esters (Sale and Wilson, 1926) have been implicated as making valuable contributions to Concord grape flavor. However, if these compounds did make valuable contribu- tions to the flavor of the reference juice in their respective concentrations, it is doubtful whether the acceptability of the juice made with stripped concentrate only could be attributed to such low concentrations of these two compounds. This statement can be made only if amounts of these compounds in the stripped concentrate were below threshold levels. Although juices made from essences were not signi- ficantly unacceptable to the flavor panels, neither were most significantly acceptable. This would indicate the presence of compounds in the various essences which, when the essence was added to the concentrate in the usual amounts, was related to a reduction of juice acceptability. 63 Results of ranking to determine juice preference are listed in Table 15. Juices prepared from essences BJ69 and CJ69 were not significantly different from the reference juice on each of four occasions while juices prepared from essences BJ68 and AJ6910/2 were not signifi- cantly different on two occasions. These latter essences made juices having relatively high preference, but were not placed so highly when considered for typicalness of Concord flavor and acceptability. Juices made from essence AJ69 were not significantly different in preference from the reference juice on only one of three occasions. Although no coefficient of correlation was deter- mined, it was noticed that a trend was developing between the total carbonyl levels of essences and flavor panel results. In general, juices made from essences having relatively high flavor difference totals (low typicalness), low acceptability, and high rank totals (low preference) contained relatively high total carbonyl levels. Thus, an essence containing these high total carbonyl levels was added to the concentrate in amounts such that the essence would contain approximately 300 ppm total carbonyls (as acetone) on a 150-fold basis. Juices were prepared from essence AJ69 in this manner for ranking. Duplicate panels indicated a definite flavor improvement of the juices pre- pared in this manner; these juices were preferred less in the second panel than in the first (see Table 15). 64 Table 15. Flavor panel data for preference by ranking. Data Statistical Juice or Number of Rank Transformation Significance Essence Panelists Total Mean Total Mean 5% 1% Panel 1 20 WH 36 1.80 9.40 0.470 a a BJ68 46 2.30 2.66 0.133 ab ab AJ68 55 2.75 -3.39 -0.l70 bc ab AP68 63 3.15 -8.97 -0.449 c b Panel 2 20 WH 36 1.80 9.57 0.479 a a BJ69 43 2.15 4.72 0.236 ab a AJ69 52 2.60 -l.20 -0.060 b ab AP69 69 3.45 -13.09 -0.655 c b Panel 3 20 WH 38 1.90 8.24 0.412 a a CJ69 42 2.10 5.45 0.273 a a AJ6910/2 51 2.55 -0.60 -0.030 a ab AN69 69 3.45 -13.09 -0.655 b b Panel 4 20 WH 31 1.55 12.96 0.648 a a AJ6910/7 47 2.35 2.06 0.103 b ab AJ69UK 53 2.65 -2.19 -0.110 b be AJ6910/8 69 3.45 -12.53 -O.627 c c Panel 5 20 WH 31 1.55 12.96 0.648 a a BJ69 41 2.05 5.79 0.290 b b AP68F 61 3.05 -7.98 -0.399 c c AJ68F 67 3.35 -ll.50 -0.575 c c Panel 6 20 WH 45 2.25 3.52 0.176 a a CJ69 47 2.35 1.86 0.093 a a BJ69 53 2.65 -l.93 -0.097 a a AJ69 55 2.75 -3.39 -0.l70 a a Panel 7 20 WH 44 2.20 3.99 0.200 a a No essence 45 2.25 3.26 0.163 a a BJ68 48 2.40 1.46 0.073 a a AJ69lO/2 63 3.15 -8.71 -O.436 a a lLike letters denote no significant difference among samples within panels. . 65 Table 15. (cont.) Data Statistical Juice or Number of Rank Transformation Significance Essence Panelists Total Mean TotaI’ Mean 5% 1% Panel 8 20 WH 38 1.90 7.34 0.367 a a BJ69 45 2.25 3.13 0.157 a a CJ69 50 2.50 0.00 0.000 a ab AJ69 67 3.35 -ll.50 -0.575 b b Panel 9 20 WH 41 2.05 6.05 0.303 a a BJ69 45 2.50 3.26 0.163 a a CJ69 48 2.40 1.46 0.073 a a AJ69 66 3.30 -12.83 -0.642 b b Panel 10 20 WH 39 1.95 7.38 0.369 a a 0.55 ml AJ69 51 2.55 -0.34 -0.017 a a 1.10 ml AJ69 54 2.70 -2.79 -0.l40 a a 0.00 ml AJ69 56 2.80 -4.25 -0.213 a a Panel 11 20 WH 28 1.40 15.02 0.751 a a 0.55 ml AJ69 52 2.60 -l.20 -0.060 b b 0.00 ml AJ69 53 2.65 -2.19 -0.110 b b 1.10 ml AJ69 67 3.35 -1l.63 -O.582 c b Panel 12 20 0.53 ml AN69 39 1.95 7.38 0.369 a a CJ69 43 2.15 4.72 0.236 ab a 0.23 ml AJ68 49 2.45 0.60 0.030 b a 1.00 ml AP69 69 3.45 -12.70 -0.635 c b Panel 13 52 (Consumer panel) CJ69 111 2.13 13.09 0.267 a a YH 132 2.54 -1.59 -0.039 a a 0.55 ml AJ69 135 2.65 -5.84 -0.112 a a AJ69 139 2.57 -5.66 -0.109 a a 1 among samples within panels. Like letters denote no significant difference 66 A similar experiment was conducted by preparing juices from essences AJ68, AN69, and AP69 as described above; juice prepared from essence CJ69 was used as the reference in a ranking panel. The juices of essences AJ68 and AN69 were not significantly different from the CJ69 juice while the juice of essence AP69 was different. Some factor other than total carbonyl level, adversely related to flavor preference, was evidently present in essence AP69. This essence did contain exceptionally high total ester and ethyl acetate levels, a possible explanation of its juice having relatively low preference. Fifty-two untrained and inexperienced panelists were solicited from the author's apartment complex for participation in a consumer flavor panel. They were given commercial juice YH and juices prepared from essences CJ69 and AJ69 and asked to indicate their preference by ranking. This combination of juices was chosen for several reasons. Juice YH, previously examined for reference use, was a commercial juice of relatively low acceptability, juice prepared from essence CJ69 was one of the better juices examined while juice prepared from essence AJ69 was not particularly favorable to semi-trained panelists. The latter essence was served to the untrained panelists as juices prepared using 2.0 ml and 0.55 ml essence per 300 ml final volume single strength juice respectively. The 67 results indicated that consumer preference for juices pre- pared from these essences is well within the range of products presently available to the consumer in super- markets. Chemical test interrelationships: Chemical test data were paired for each esSence and inspected for pos- sible interrelationships. Coefficients of correlation (Amerine, l965b) were calculated only for those pairings which appeared related. These coefficients are listed in Table 16. Although many coefficients of correlation were significantly different from r=0 (t=[l-r2]/[n-2], n-2 degrees of freedom, where n=the number of data pairs, [Amerine, l965b]), no practical significance should be placed on many, particularly those between r=-0.6 and r=0.6. Least squares regression lines were calculated for data with coefficients significantly different from r=0. The contribution of data to the linear portion of the line may be represented by r2 (Mendenhall, 1967b). Any r approaching zero from i 0.7, for example, would indicate less than 50 per cent of data points signifi- cantly contributing to the linear portion of the line. Thus, the standard error of estimate (Little, 1966) for the regression line was calculated and included in Table 16. Significant correlation coefficients do not necessarily indicate cause/effect relationships. 68 .ooafimu oso lumou u m.usmosum msflws Ho>oa wH osu um oosmcHMHson moumoflosH«« .ooawmu oso .umos u m.pso©sum msflms Ho>oa wm on» so oosmofimwsmflm moumoaosH« .msofluoasoamo oumuflafioom ou HE\moHoE_s ou oopso>soo mosam> mouoowosH H ma.amwmm.as+xmomoo.ous .4Hss.o ma mumumom assume omo+mumuoom Assam omo m> msoumo Hmuoe ms.omwmm.ss+xsomoo.ous .«mss.o ms 6060666 Hanum umo m> msoumo Hmuoe mmm.o ms maoumom omo m> osouooo one om.mwovv.o+xoso.ous .«Nmm.o as mvmnmosmumom omo m> oomsopaouooo one am.sflo.smuxm.mmns «.svm.o as wasnmcsmumom omo m> Hmsonsoo Hmuoe ssm.oums~.m+xmmm.ons «.omm.o as Honoumom oge+msmnmssmuoom one m> Hmsonuoo Hmuoe mms.on as Honeymom use m> Husossmo Hosea smm.anoom.suxmmm.ons ..som.o as Hmnmamssmumom use m> Ansonsmc Hopes m.osmus.smsuxmemsns .sss.o ms Asa msmuosmv mocmnHOmnm umsos>muuso m> Ansonsmo Hmuos soauosvm scammoumom soaumHossoc muflom sowumaouuoc mo usoflowmmooc s ..moosommo was How msOHuoHosHoo HMUHEoso m> HMOHEosc .ma manna 69 A significant coefficient (r=0.474) was calculated for total carbonyl vs ultraviolet absorbance at 210-213 nm. This indicated a possible contribution of unsaturated alde- hydes to total carbonyl levels (Roberts and Caserio, 1965). Total carbonyl levels were closely related to acetaldehyde levels as measured by both thin layer (r=0.904) and gas- solid (r=0.947) chromatography. Acetone was not related to total carbonyl levels (r=-0.l95) and thus made no major total carbonyl contributions. A good correlation existed between thin layer and gas-solid chromatographic determinations of acetaldehyde (r=0.862). This indicated that either method could be used to determine acetaldehyde levels of essences as those levels in essence headspaces were positively related to levels within the reSpective essences. This relationship did not exist, however, for acetone (r=0.353). Acetone volatility was evidently affected by a factor or factors within the essence causing inconsistencies between head- space and liquid essence concentrations. Liquid essence total ester levels were related to headspace ethyl acetate (r=0.773) and methyl acetate + ethyl acetate levels (r=0.77l). This was reasonable as ethyl acetate has been reported as the single most abundant ester in Concord grape essence (Holley et a1., 1955). No other significant correlations between the various chemical determinations were apparent. 70 Flavor panel vs chemical relationships: Flavor difference totals and per cent acceptability scores were inspected for possible significant correlations with chemi- cal test data. Where apparent relationships existed, coefficients of correlation were calculated and listed in Table 17. Coefficients were treated similar to those of chemical interrelationships. Rank total scores were not considered for correlation since scores are not independ- ent of other juices judged in the same panel. There was a general lack of significant correlation between flavor panel data and chemical data as only one coefficient (% acceptable vs ultraviolet absorbance at 243 nm) exceeded r=0.8. Even though no apparent relationship existed, the coefficient for flavor panel results vs methyl anthranilate was determined; considerable attention has been given to connecting this compound with Concord grape flavor. The concentration of this compound when paired with flavor difference totals and per cent acceptability scores yielded coefficients of r=-0.073 and r=-0.lO4 respectively. This, however, did not necessarily conflict with reports in the literature relating this compound to Concord grape flavor. Clore (1965), for example, pointed out that methyl anthra- nilate is a threshold factor in Concord flavor. Thus, threshold levels could be reached at very low methyl anthranilate concentrations; amounts exceeding the 71 msm.o as as: m¢m\ss msmnoamc mosMQHOQO poaow>msuas m> osoom oososoMMHo Ho>mam mos.ou as as: mv~\sc msmuoamc oosmnHOmnm uoH0H>oHuHo m> magnumooos w HvH.oH~om.0Ithwoo.onw «amas.o m AEs memo oosmnHOon uoHOH>msuao m> osoom oosonoMMHo Ho>mam voH.oHHma.0Ixhowoo.ouw **vmm.o| m AEs mvmv oosMQHOmnm uoHoH>muuHD m>_oanmumooo< w oom.onwm.olemmoo.onw 4mmv.o «a oostHOmnm 00H0H>osuao m> osoom oosouommwo so>mam mmv.oflmm>.o+xbm¢oo.0|nw «mmm.o| va Ass mamloamv mosMQHOmnm uoH0fl>msuHD m> magnumwood w mvm.o 5H mnoumo Houoe m> osoom oosoHoMMHo Ho>mHm mam.on ha msoumo Hmuoe m> magnumooos w mac.ou as moosscmurucm Hugumz m> osoom oosouommap Ho>mam sos.ou as momsscmunuam Hanna: m> magnumooos w soaumsom soflmmosmom soHumHosuoc mnflom soaumaossoc mo usofloflmmooc s .mocsmmmo on» How msoflumaousoo HMUMEoso m> so>mam .ha magma 72 .6mssmu oso .umou u m.usopsum osfims Ho>oa wa on» so oosmoflMHsmwm mopmoflosH%« .ooawm» mso .umou u m.usoosum msflms Ho>oH wm osu um mosMOfiMHsmwm moumoflosH« mam.onom.o+xsvso.ous swm.owmvs.o+xssooo.ouns on.mfiom.balxmmm.onw N0.mflmn.HH+anH.0InN hvm.m0HnHm.nHN+xonmalnw vortmoflhmh~NN+XmNhuN mam.o *«hmw.0I «omv.o «am00.0I omm.o non.OI nmv.o 0mm.ol *Hmv.0I «hvv.o 0H ma 0H 0H ma ma ma ma ma ma mb. m> m> m> m> ououoom Hmsuoa cmc whoom oososomwflo no>mam oumumom stuoe cmc m> magnumooos w moanmcsmuoom omw Houou ousmHoMMHU Ho>mHm monsooamumom cmc m> magnumooos w momgmosmumom one ouoom ocsosommwo Ho>mam monsooamuooo cue m> oanmumooos m Hmsonsmo Hmuoe osoom oosoHoMMHo Ho>mam Ansonumo Hmuoe m> oanmumooos m osoEoo soomxo HMOflEosc osoom oosoquMflo so>oam osmEop sommxo HMOHEosc m> magnumooos w 73 threshold concentration do not add to and possibly even degrade the flavor of these essences. Thresholds were not examined in this study. No significant correlations were found between flavor panel and total ester results. However, a trend indicated that greater total ester levels in essences tended to produce juices poorer in flavor quality. Absorbance in the 210-213 nm range of the ultra- violet spectrum correlated significantly with both acceptability (r=-0.535) and flavor difference data (r=0.485). Although the regression line attached to this data had a relatively large error term, the line indicated a general depletion of flavor quality with increased ab- sorption at this wavelength. Ultraviolet absorption at 243 nm correlated with flavor panel results better than other chemical data (acceptability, r=-0.854; flavor dif- ference, r=0.712). Again, an inverse relationship between absorption and overall flavor quality existed, indicating the possible importance of unsaturated aldehydes. The ratio of the absorptions of each essence at the above wavelengths was determined and when paired with acceptability and flavor difference data of flavor panels, no significant correlations existed (r=-0.168 and r=0.3l6 respectively). Significant correlation coefficients resulted when chemical oxygen demands were paired with flavor difference 74 totals (r=-0.451) and acceptability scores (r=0.447). Unlike the trend of other flavor/chemical correlations, chemical oxygen demand showed a general increase with essence flavor quality. This indicated a positive, but not necessarily absolute, relationship between total oxi- dizable organics and general essence flavor quality. This trend was in agreement with Charley (1962), who reported a very good relationship between "fold" and chemical oxygen demand. It was mentioned, however, that this relationship is not necessarily related on a flavor quality basis. Although an inverse trend did exist between total carbonyl levels and general flavor quality, no significant correlations were present (acceptability, r=-0.256; flavor difference, r=0.423). Those essences with relatively high total carbonyl levels were generally inferior in flavor quality to those having relatively low total carbonyl levels. Thin layer acetaldehyde data did not correlate significantly with flavor analysis (acceptability r=-0.393; flavor difference r=0.250). However, gas chromatographic acetaldehyde data did yield significant coefficients with flavor difference totals (r=0.436) and acceptability scores (r=-0.609). Both of these analyses showed the same trend as total carbonyl data. Correlation of gas chromatographic methyl acetate data with flavor panel results showed significant 75 coefficients (acceptability, r=-0.627; flavor difference, r=-0.513). Similar to other individual compound data, an inverse relationship existed between methyl acetate levels in essence headspace and overall flavor quality. SUMMARY AND CONCLUSIONS The range of compound quantities found in the essences of this study was truly remarkable if one con- siders that each essence was labelled as lSO-fold by its respective manufacturer. For example, methyl anthranilate was found in concentrations ranging from 4 to 126 ppm. Total esters were present in amounts ranging from 100 to 11,400 ppm (as ethyl acetate); total carbonyls ranged from 250 to 6600 ppm (as acetone). Any or all of these varia- tions could be possible and total organic carbon count, ie. chemical oxygen demand, could remain relatively con- stant. However, in these essences, even chemical oxygen demands ranged from 20,000 to in excess of 200,000 ppm. Volatile fractions of daily production samples taken within the same week for essence AJ69 yielded chemi- cal oxygen demands ranging from 70,000 to more than 200,000 ppm, total ester levels from 250 to 11,400 ppm, total carbonyl levels from 300 to 6600 ppm, and methyl anthranilate levels from 28 to 78 ppm. For these series of samples, total ester, total carbonyl, acetaldehyde, and chemical oxygen demand increased together, showing a posi- tive relationship among these factors. 76 77 Considering these variations, it is not difficult to understand essence add-back problems encountered by the end users of these products. Grapes themselves, the raw material of essence manufacture, undoubtedly contributed to these variations. But certainly the largest variations were introduced by non-standardized production practices. Chemical-chemical interrelationships indicated acetaldehyde as the single most abundant contributor to total carbonyl values. The generally high acetaldehyde levels found in the lower flavor quality essences could have been derived from products of spontaneous fermentation of grapes and/or juices prior to essence stripping. Neudoerffer et a1. (1965) reported relatively high levels of acetone and acetaldehyde in Concord essences having undesirable flavor anomolies. Many of the more prominent recorder responses in gas chromatograms of headspace vapors over the essences were from the presence of ethyl acetate. This compound was correlated (1% level of significance) with total vola- tile ester levels in the liquid portion of the essences and was reported by Holley et al. (1955) as being the- single most predominent ester present in Concord grape essence. Neudoerffer et a1. (1965) also reported rela- tively large amounts of ethyl acetate in the Concord essences they studied. 78 Correlations of chemical data with flavor panel results were generally not very good. Of all correlations, ultraviolet absorbance at 243 nm had the highest coeffi- cient (r=-0.854) with acceptability scores. Absorbance at this wavelength was only prevalent in those essences ranking relatively well in taste panels. Thus, absorbance at this wavelength would appear mandatory for an essence to be of high flavor quality. Flavor panel analysis of juices tested indicated stripped concentrate diluted to 15.5-16.0° Brix was quite acceptable to flavor panelists. This juice was not as typical in Concord flavor as the reference juice but ranked with it in preference. Therefore, in general, the addition of essences to the stripped concentrate caused a degrada- tion of the flavor quality of the diluted concentrate. This conclusion is quite reasonable, particularly if the essences possessing poorer flavor qualities were prepared from juices or purees which had undergone various degrees of spontaneous fermentation. Undesirable fermentation products could have been stripped from the juice and into the essence, thus making the essences poor and the concen- trate good in flavor quality. Several trends were evident in the chemical-flavor comparisons. These are as follows: 1. The level of methyl anthranilate in the essences apparently did not affect flavor typicalness, 79 acceptability, or preference of juices made from these essences. 2. As essence total volatile esters increased, general flavor quality in terms of flavor difference scores, acceptability, and preference for juices made from these essences decreased. 3. Chemical oxygen demand gave a general indica- tion of flavor quality, in a positive, but not absolute, manner. 4. As acetaldehyde and/or total carbonyl levels increased, general flavor quality decreased. The basic conclusion of this study was that no single component of Concord grape essence, using methods described, can be measured quantitatively and used to determine essence quality in terms of flavor enhancement capacity for Concord grape products. An intricate balance of components within the essence seemed necessary for high flavor quality and may best be measured by headspace vapor analysis via gas chromatography. Similar conclusions were reported for other food products by wolford et al. (1963), Heinz et a1. (1964) and Powers (1968). LIST OF REFERENCES LITERATURE CITED Amerine, M. A., Pangborn, Rose M., and Roessler, E. B., l965a, Principles of Sensornyvaluation of Food, Academic Press, New York, p. 451. Amerine, M. A., Pangborn, Rose M., and Roessler, E. B., l965b, Principles of Sensory Evaluation of Food, Academic Press, New York, p. 486. Amerine, M. A., Pangborn, Rose M., and Roessler, E. B., 1965c. Principles of Sensory Evaluation of Food, Academic Press, New York, p. 442. Buttery, R. G., Ling, Lousia C., and Guadagni, D. G. 1969, Volatilities of aldehydes, ketones, and esters in dilute water solution. J. Food Sci. 27,165. Charley, V. L. S. 1962, Volatile constituents of fruit juices. Proc. Scientific Tech. Commiss., Intern. Fed. of Fruit Juice Producers 4,365. Clore, W. J., Neubert, A. M., Carter, G. H., Ingalsbe, D. W., and Brummund, V. P. 1965. Composition of Washing- ton-produced Concord grapes and juices. Wash. Agr. Exp. Sta. Bull. No. 48. Dougherty, M. H. 1968. A method for measuring water- soluble volatile constituents of citrus juices and pro- ducts. Food Technol. 22,1455. Flath, R. A., Forrey, R. R., and Teranishi, R. 1969. High resolution vapor analysis for fruit variety and fruit product comparisons. J. Food Sci. 34,382. Forss, D. A., Jacobsen, Valerie M., and Ramshaw, E. H. 1967. Concentration of volatile compounds from dilute aqueous solutions. J. Agr. Food Chem. 15,1104. Guadagni, D. G., Buttery, R. G., and Okano, S. 1963. Odour thresholds for some organic compounds associated with food flavors. J. Sci. Food Agr. 14,761. 80 81 Guadagni, D. G., Okano, S., Buttery, R. G., and Burr, H. K. 1966. Correlation of sensory and gas-liquid chromato- graphic measurement of apple volatiles. Food Technol. 20,518. Heinz, D. E., Pangborn, R. M., and Jennings, W. G. 1964. Pear aroma: Relation of instrument and sensory techniques. J. Food Sci. 29,756. Holley, R. W. and Holley, Ann D. 1952. The identifica- tion of alcohols in dilute aqueous solution. Anal. Chem. 24,216. Holley, R. W., Stoyla, Brigitta, and Holley, Ann D. 1955. Some volatile constituents of Concord grape juice. Food Res. 20,326. Jensen, M. 1961. Determination of aroma in fruit juices and fruit juice concentrates with a quick method. Proc. Scientific Tech. Commiss., Intern. Fed. of Fruit Juice Producers. 3,63. Kepner, R. E., Maarse, H., and Strating, J. 1964. Gas chromatographic head space techniques for the quantitative determination of volatile components in multicomponent aqueous solutions. Anal. Chem. 36,77. LeClerg, E. L. 1966. "Experimental Methods for Extension Workers" Univ. of Calif. Agr. Ext. Service. p. 23. Li, J. C. R., 1957. Introduction to Statistical Inference, Edwards Brothers, Inc., Ann Arbor, Michigan, p. 459. Little, T. M. 1966. "Correlation and Regression; a sup- plement to Experimental Methods for Extension Workers" Univ. of Calif. Agr. Ext. Service, p. 17. McNary, R. R., Dougherty, M. H., and Wolford, R. W. 1957. Determination of chemical oxygen demand of citrus waste waters. Sewage and Ind. Wastes 29,894. Mendenhall, W. 1967a. Introduction to Probability and Statistics, Wadsworth Publishing Company, Inc., Belmont, California, p. 157. Mendenhall, W., 1967b. Introduction to Probabilit and Statistics, Wadsworth Publishing Company, Inc., Be1mont, California, p. 241. Murch, A. F. and Ziemba, J. V. 1958. Improves concentrat- ing three ways. Food Eng. 30:ll,81. 82 Nawar, W. W. 1966. Some considerations in interpretation of direct headspace gas chromatographic analysis of food volatiles. Food Technol. 20,115. Neuberg, C., Grauer, A., and Pisha, B. V. 1952. The pre- cipitation of carbonyl compounds with 2,4-dinitrophenyl- hydrazine. Anal. Chim. Acta 7,238. Neudoerffer, T. S., Sandler, S., Zubeckis, E., and Smith, M. Doreen. 1965. Detection of an undesirable anomoly in Concord grape by vapor phase chromatography. J. Agr. Food Chem. 13,584. Peleg, Y. and Mannheim, C. H. 1970. Determination of carbonyl concentration in aqueous citrus essences. J. Agr. Food Chem. 18,176. Power, F. B. 1921. The detection of methyl anthranilate in fruit juices. J. Am. Chem. Soc. 43,1741. Power, F. B. and Chesnut, V. K. 1921. The occurrence of methyl anthranilate in grape juice. J. Agr. Res. 23,43. Powers, J. J. 1968. Toward objective evaluation of food flavor. Food Technol. 22,383. Roberts, J. D. and Caserio, Marjorie C. 1965. Basic Principles of Organic Chemistry, W. A. Benjamin, Inc., New York, p. 249. Roger, N. F. 1961. Recovery of methyl anthranilate in Concord grape essence. Food Technol. 15,309. Sale, J. W. and Wilson, J. B. 1926. Distribution of volatile flavor in grapes and grape juices. J. Agr. Res. 33,301. Scott, R. D. 1923. Methyl anthranilate in grape beverages and flavors. Ind. Eng. Chem. 15,732. Stern, D. 1., Lee, A., McFadden, W. H., and Stevens, K. L. 1967. Volatiles from grapes: Identification of volatiles from Concord essence. J. Agr. Food Chem. 15,1100. Stevens, K. L., Bomben, J. L., Lee, A., and McFadden, W. H. 1966. Volatiles from grapes: Muscat of Alexandria. J. Agr. Food Chem. 14,249. Stevens, K. L., Bomben, J. L., and McFadden, W. H. 1967. Volatiles from grapes: Vitis vinifera (Linn.) cultivar Grenache. J. Agr. Food Chem. 15,378. 83 Stevens, K. L., Plath, R. A., Lee, A., and Stern, D. K. 1969. Volatiles from grapes: Comparison of Grenache juice and Grenache rose' wine. J. Agr. Food Chem. 17, 1102. Stevens, K. L., Lee, A., McFadden, W. H., and Teranishi, R. 1965. Volatiles from grapes: Some volatiles from Concord essence. J. Food Sci. 30,1006. Thompson, A. R. 1950. A colorimetric method for the determination of esters. Australian J. Sci. Research 3A,128. Tukey, J. W. 1953. Some selected quick and easy methods of statistical analysis. Trans. N. Y. Acad. Sci. Ser. II, 16:2,88. Van Wyk, C. J., Webb, A. D., and Kepner, R. 1967. Some volatile components of Vitis vinifera variety white Riesl- ing grape juice. J. Food Sci. 32,660. White, J. R. 1966. Methyl anthranilate content of citrus honey. J. Food Sci. 31,102. Wolford, R. W., Attaway, J. A., Alberding, G. E. and Atkins, C. D. 1963. Analysis of the flavor and aroma constituents of Florida orange juices bngas chromato- graphy. J. Food Sci. 28,320. GENERAL LITERATURE Bailey, S. D., Mitchell, D. G., Bazinet, M. L. and Weurman, C. 1962. Studies on the volatile components of different varieties of cocoa beans. J. Food Sci. 27,165. Bassett, R., Ozeris, Suheyla, and Whitnah, C. H. 1962. Gas chromatographic analysis of head space gas of dilute aqueous solutions. Anal. Chem. 34,1540. Byrne, G. A. 1965. The separation of 2,4-dinitropheny1- hydrazones by thin layer chromatography. J. Chromatog. 20,528. Denti, E. and Luboz, M. P. 1965. Separation of 2,4- dinitrophenylhydrazones of carbonyl compounds by thin layer chromatography. J. Chromatog. 18,325. 84 Flath, R. A., Black, D. R., Guadagni, D. G., McFadden, W. H., and Schultz, T. H. 1967. Identification and organic evaluation of compounds in delicious apple. J. Agr. Food Chem. 15,29. Hale, W. S. and Cole, E. W. 1963. A freezing technique for concentrating pre-ferments. Cereal Chem. 40,287. Hoffman, R. L., List, G. R., and Evans, C. D. 1966. Enrichment of volatile samples by syringe collection. J. Food Sci. 31,751. Jart, A. and Bigler, A. J. 1966. Improved techniques for thin layer chromatographic separation of 2,4-dinitrophenyl- hydrazones. J. Chromatog. 23,361. Kepner, R. E., van Straten, S., and Weurman, C. 1969. Freeze concentration of volatile components in dilute aqueous solutions. J. Agr. Food Chem. 17,1123. Mira, M. J. 1969. Gas chromatographic qualitative and semi-quantitative analysis of apple aroma by means of retention indexes. Anal. Chim. Acta 48,169. Murch, A. F. and Ziemba, J. V. 1957. Concentrating advances bring superior flavors. Food Eng. 29:12,90. Nelsen, P. E. and Hoff, J. E. 1968. Food volatiles: Gas chromatographic determination of partition coefficients in water-lipid systems. J. Food Sci. 33,479. Randerath, K. 1964. Thin Layer Chromatography, Academic Press, New York, p. 215. Rasmussen, H. 1967. Determination of keto acid 2,4- dinitrophenylhydrazones by a generally applicable quanti- tative thin layer chromatographic method based on photo- metric measurement of spot areas. J. Chromatog. 27,142. Ronkainen, P. 1967. Thin layer chromatographic resolution of mixtures of keto acid 2,4-dinitrophenylhydrazones. J. Agr. Res. 33,301. Schwartz, D. P. 1962. Separation of homologous series of 2,4-dinitr0phenylhydrazones by column partition chromato- graphy. J. Chromatog. 9,187. Schwartz, D. P., Shamey, Jennie, Brewington, C. Ro., and Parks, 0. W. 1968. Methods for the isolation and charac- terization of constituents of natural products: New and improved methods for the analysis of carbonyl 2,4-dinitro- phenylhydrazones. Microchem. J. 13,407. 85 Schwartz, D. P., Johnson, A. R., and Parks, O. W. 1962. Use of ion exchange resins in the micro analysis of 2,4- dinitrophenylhydrazones. Microchem. J. 6,37. Shapiro, J. 1961. Freezing out, a safe technique for concentration of dilute solutions. Science 133,2063. Shapiro, J. 1967. Concentration of volatile substances in aqueous solutions and production of water-free organics by freezing out. Anal. Chem. 39,280. Stanley, W. L., Ikeda, R. M., Vannier, S. H., and Rolle, L. A. 1961. Determination of the relative concentrations of the major aldehydes in lemon, orange, and grapefruit oils by gas chromatography. J. Food Sci. 26,43. Teranishi, G. and Buttery, R. G. 1961. Gas-liquid chrom- atography of the aroma of vegetables and fruits: Direct injection of aqueous vapors. Anal. Chem. 33,1439. Weurman, C. 1961. Gas-liquid chromatographic studies on the enzymatic formation of volatile compounds in rasp- berries. Food Technol. 15,531. Weurman, C. 1969. Isolation and concentration of vola- tiles in food odor research. J. Agr. Food Chem. 17,370. Wolford, R. W., Alberding, G. E., and Attaway, J. A. 1962. Analysis of recovered natural orange essence by gas chroma- tography. J. Agr. Food Chem. 10,297. APPENDIX 86 Table 18. Methyl anthranilate standard curve. Methyl Solution/Replicate Anthiggliite l/a 1/b 2/a 3/a 3/b ug Absorbance at 490 nm* 0 0.000 0.000 0.000 0.000 0.000 10 0.036 0.034 0.030 0.034 0.035 20 0.068 0.066 0.068 0.069 0.072 30 0.102 0.102 0.101 0.098 0.102 40 0.138 0.134 0.143 --- 0.141 50 0.174 0.157 0.163 0.172 0.169 60 0.201 0.206 0.204 0.196 0.208 70 0.246 0.237 0.238 0.246 0.246 r = 0.999 Y = 0.0035 X -0.0032 i 0.0066 *10 mm colorimeter tubes Table 19. Methyl anthranilate data for the volatile frac- tion of the essences. o o 1 I- Distillate 018tl Repli- Repli- late Essence cate l 72 3 Essence cate PPm MA PPm MA AJ68 a 4 5 5 AJ6910/8 a 46 b 4 4 4 b 47 AP68 a 11 ll 9 AJ69UK a 28 b 8 12 11 b 27 BJ68 a 42 50 53 AJ68F a 0 b 46 43 47 b 0 AJ69 a 20 20 23 AP68F a 11 b 24 23 20 b 11 AP69 a 28 25 28 BJ68F a 36 b 25 27 27 b 29 AN69 a 124 119 127 AJ69F a 21 b 128 130 128 b 23 BJ69 a 9 9 9 AP69F a 28 b 9 9 9 b 30 CJ69 a 31 32 31 BJ69F a 14 b 32 33 32 b 14 AJ6910/2 a 78 -- -- CJ69F a 31 b 78 -- -- b 36 AJ6910/7 a 52 -- -- AJ69UKF a 35 b 56 -- -- b 35 87 Table 20. Total ester standard curve. Ethyl Solution/Replicate Acetate pg/ml 1/a l/b 2/a 2/b 3/a 3/b diStll- Absorbance at 540 nm* ate 0 0.000 0.000 0.000 0.000 0.000 0.000 25 0.078 0.074 0.076 0.071 0.080 0.080 50 0.160 0.153 0.157 0.143 0.153 0.160 75 0.250 0.234 0.242 0.229 0.233 0.238 100 0.323 0.317 0.321 0.308 0.325 0.330 125 0.409 0.392 0.403 0.385 0.409 0.423 150 0.478 0.459 0.491 0.469 0.485 0.498 175 0.561 0.549 0.577 0.542 0.561 0.569 200 0.648 0.624 0.653 0.634 0.624 0.643 r = 0.999 Y = 0.0032 X -0.001:0.017 *16 mm colorimeter tubes Total ester data for the volatile fraction of the essences. Table 21. ——_—'-———————_r—'———_l—T—: Distillate Distll Repli- Repli- late Essence cate 1 2 3 Essence cate ppm EA ppm EA AJ68 a 900 950 900 AJ6910/8 a 11,400 b 900 950 900 b 11,400 AP68 a 5850 5850 5800 AJ69UK a 4550 b 5750 5750 5800 b 4650 BJ68 a 150 150 150 AJ68F a 1000 b 100 100 150 b 1000 AJ69 a 5100 5100 5050 AP68F a 6250 b 5100 5150 5150 b 6300 AP69 a 9100 9150 9300 BJ68F a 100 b 9150 9050 9100 b 100 AN69 a 1650 1650 1700 AJ69F a 5450 b 1700 1700 1700 b 5350 BJ69 a 5350 5350 5400 AP69F a 8900 b 5400 5300 5300 b 8800 CJ69 a 2400 2300 2300 BJ69F a 5650 b 2300 2300 2300 b 5700 AJ6910/2 a 250 -- -- CJ69F a 4050 b 250 -- -- b 4150 AJ6910/7 a 700 -- -- AJ69UKF a 2600 b 700 -- -- b 2600 88 Table 22. Chemical oxygen demand standard curve. a E COD Solution/Replicate PPm l/a l/b 2/a 2/b Absorbance at 650 nm* 0 0.000 0.000 0.000 0.000 107 0.036 0.036 0.040 0.040 214 0.076 0.076 0.078 0.077 320 0.112 0.118 0.115 0.114 427 0.153 0.152 0.155 0.152 533 0.192 0.191 0.192 0.189 640 0.231 0.231 0.227 0.231 746 0.266 0.270 0.264 0.264 853 0.308 0.305 0.303 0.301 960 0.344 0.342 0.337 0.337 r = 0.999 Y = 0.000355 X +0.002 : 0.0093 *16 mm colorimeter tubes Table 23. Chemical oxygen demand data for the volatile fraction of the essences. o o 1 . . - E Repli- Distillate Rep- Diztél ssence Essence 11-'——————— cate 1 2 3 1 cate PPm C0D ppm COD AJ68 a 76,700 77,500 76,700 AJ6910/8 a 208,000 b 77,500 76,700 77,500 b 201,600 AP68 a 22,500 22,250 21,750 AJ69UK a 107,200 b 22,150 21,600 22,500 b 107,200 BJ68 a 38,000 38,000 37,750 AJ68F a 77,600 b 38,750 37,750 38,500 b 76,500 AJ69 a 68,500 67,500 68,000 AP68F a 21,300 b 67,500 69,000 68,500 b 20,000 AP69 a 27,250 28,000 27,000 BJ69F a 33,800 b 27,500 27,250 27,250 b 33,700 AN69 a 47,200 47,500 47,500 AJ69F a 66,000 b 46,000 47,500 46,600 b 66,000 BJ69 a 85,000 86,600 85,800 AP69F a 26,400 b 85,800 85,800 86,600 b 26,650 CJ69 a 102,000 99,400 98,000 BJ69F a 89,800 b 99,000 101,000 99,400 b 88,500 AJ69lO/2 a 71,300 --- --- CJ69F a 110,200 b 71,300 --- --- b 109,000 AJ6910/7 a 72,600 —-- --- AJ69UKF a 105,000 b 75,500 --- --- b 105,000 89 Table 24. Total carbonyl standard curve. Acetone Solution/Replicate ppm 1/a 1/b 2/a 2/b 3/a 3/b Absorbance at 480 nm* 0 0.000 0.000 0.000 0.000 0.000 0.000 5 0.077 0.112 0.088 0.085 0.104 0.063 10 0.116 0.163 0.166 0.137 --- --- 15 0.202 0.240 0.235 0.216 0.192 0.206 20 0.305 0.305 0.314 0.297 0.290 0.301 25 0.310 0.367 --- 0.364 0.364 0.367 30 0.444 0.435 0.447 0.435 0.406 0.435 35 0.512 0.475 0.505 0.502 0.498 0.505 40 0.573 0.577 0.569 0.542 0.569 0.538 r = 0.996 Y = 0.0139 X +0.012 : 0.032 *16 mm colorimeter tubes Table 25. Total carbonyl data for the volatile fraction of the essences. . . Distil- Repli- Distillate Rep- late Essence cate l 2 3 Essence 11- ppm acetone cate ppm ace- tone AJ68 a 2150 2360 2240 AJ68F a 2370 b 2160 2480 2260 b 2300 c 2140 2150 2440 c 2370 AP68 a 725 750 750 AP68F a 590 b 725 625 625 b 575 c 760 750 750 c 695 BJ68 a 325 325 375 BJ68F a 260 b 300 350 375 b 260 c 340 340 375 c 325 AJ69 a 1175 1090 1165 AJ69F a 875 b 1015 1070 1140 b 955 c 1150 1070 1050 c 1045 AP69 a 575 640 555 AP69F a 460 b 560 610 620 b 535 c 570 645 590 c 480 AN69 a 1170 1150 1075 BJ69F a 225 b 1055 1130 1190 b 230 c 1165 1055 1175 c 185 BJ69 a 260 235 235 AJ6910/2 a 310 b 250 260 290 b 350 c 255 235 290 c 265 CJ69 a 240 285 340 AJ69lO/7 a 490 b 320 325 285 b 425 c 215 325 350 c 470 AJ6910/8 a 6800 -- -- CJ69F a 305 b 6240 -- -- b 610 c 6800 -- -- c 280 AJ69UK a 1425 -- -- AJ69UKF a 1190 b 1575 -- -- b 1235 c 1525 -- -- c 1230 91 Table 26. Acetone 2,4-dinitropheny1hydrazone standard curve. , , Solution/Replicate Derivative ug/3 m1 1/b 2/a 2/b 3/a 3/b Absorbance at 358 nm* 0 0.000 0.000 0.000 0.000 0.000 2 0.069 0.056 0.066 0.086 0.072 4 0.097 0.112 0.202 0.176 0.084 6 0.146 0.166 0.161 0.197 0.149 8 0.208 0.224 0.240 0.238 0.199 10 0.262 0.286 0.301 0.299 0.264 12 0.308 0.342 0.325 0.349 0.330 14 0.330 0.409 0.429 0.429 0.347 16 0.516 0.523 0.509 0.465 0.423 r = 0.979 Y 0.028 X +0.009 : 0.057 *10 mm colorimeter tubes Table 27. Acetaldehyde 2,4-dinitrophenylhydrazone standard curve. , , Solution/Replicate Derivative ug/3 ml l/a l/b 2/a 2/b Absorbance at 352 nm* 0 0.000 0.000 0.000 0.000 2 0.056 0.054 0.058 0.058 4 0.111 0.112 0.152 0.153 6 0.220 0.184 0.199 0.177 8 0.237 0.252 0.290 0.240 10 0.305 0.319 0.328 0.301 12 0.390 0.380 0.398 0.380 14 0.432 0.429 0.505 0.453 16 0.542 0.545 0.488 0.505 r = 0.994 Y = 0.033 X -0.007 : 0.035 *10 mm colorimeter tubes S92 Table 28. Thin layer chromatographic data for the eeaencee. Extraction Extraction Eaeence Plate 1 2 ‘ 2 pnggcetone ppm acetaldehyde AJ68 l 24 20 24 1271 1236 1106 2 0 0 24 1495 1240 1298 3 0 0 0 1361 1298 1291 A968 1 428 545 378 11 28 54 2 451 389 347 82 90 1 3 567 451 402 92 72 53 BJ68 l 338 219 219 19 5 l9 2 414 173 217 39 1 31 3 451 173 205 19 3 29 A369 1 170 175 166 461 467 394 2 183 197 166 442 391 422 3 244 183 170 369 414 414 AP69 1 451 424 334 178 257 206 2 401 377 341 178 225 192 3 341 334 451 170 237 206 BJ69 l 195 137 144 21 53 31 2 119 144 149 31 39 58 3 112 112 112 58 1 27 CJ69 1 49 71 37 220 243 225 2 95 117 85 170 182 ' 243 3 110 112 76 196 194 214 AN69 1 254 244 293 394 408 451 2 312 400 332 408 477 394 3 332 300 315 451 451 477 AJ6910/2 l 41 -- -- 119 -- -- 2 41 -- -- 121 -- -- 3 63 -- -- 175 -- —- AJ6910/7 1 117 -- -- 328 -- -- 2 95 -- -- 292 -- —- 3 112 -- -- 308 -- -- AJ6910/8 1 0 -- -- 2579 -- -- 2 0 -- -- 2854 -- -- 3 lo -- -- 2999 -- -- AJ68F 1 0 0 0 827 725 662 2 0 0 0 756 764 725 3 0 0 0 764 709 764 AP68F l 244 219 204 200 257 298 2 229 292 295 255 259 210 3 254 260 275 308 235 259 BJ68F 1 61 46 61 100 90 64 2 59 83 37 100 60 129 3 88 46 49 68 111 60 AJ69F l 141 144 124 337 420 373 2 129 180 158 336 394 355 3 158 129 158 394 355 434 AP69F 1 122 139 168 227 225 259 2 158 141 119 245 284 253 3 122 124 129 275 273 253 BJ69F 1 129 197 144 51 39 72 2 176 202 127 21 82 0 3 129 202 129 53 0 66 CJ69F 1 0 0 0 19 37 25 2 0 0 0 64 82 82 3 0 0 0 64 60 60 AJ69UK l 222 -- -- 844 -- ~- 2 176 -- -- 844 -- -- 3 244 -- -- 758 -- '- AJ69UKF 1 219 -- -- 797 -- ‘- 2 176 -- -- 828 -- ‘- 3 180 -- -- 797 -- -- 93 Table 29. Ultraviolet absorbance data for the essences. 1 . Wavelength Essence Replicate 205 nm 210_213 nm absorbance AJ68 a 0.420 --- 0.100; b 0.450 --- 0.1602 c 0.435 --- 0.1102 AP68 a --- 0.455 0.1852 b --- 0.425 0.1352 c --- 0.445 0.145 BJ68 a --- 0.460 0.185 b --- 0.465 0.190 c --- 0.445 0.1552 AJ69 a —-- 0.495 0.1902 b --- 0.485 0.1802 c --- 0.475 0.1402 AP69 a --- 0.510 0.1852 b --- 0.510 0.1702 c --- 0.490 0.125 BJ69 a --- 0.320 0.120 b --- 0.320 0.110 c --- 0.305 0.070 CJ69 a --- 0.315 0.130 b --- 0.310 0.125 c --- 0.310 0.100 AJ6910/2 a --- 0.695 0.245 b --- 0.710 0.275 c --- 0.690 0.225 AJ6910/7 a --- 0.555 0.205 b --- 0.545 0.180 c --- 0.570 0.2302 AJ6910/8 a --- 0.880 0.3102 b --— 0.880 0.2852 c --- 0.865 0.265 AJ69UK a --- 0.580 0.160 b --- 0.585 0.180 c --- 0.570 0.1552 AJ68F a 0.440 --- 0.1102 b 0.455 --- 0.1102 c 0.445 --- 0.1702 AP68F a --- 0.455 0.1552 b --- 0.470 0.2302 c --- 0.460 0.195 BJ68F a --- 0.420 0.165 b --- 0.435 0.190 c --- 0.435 0.1602 AJ69F a --- 0.480 0.1552 b --- 0.490 0.1652 c --- 0.495 0.1902 AP69F a --- 0.500 0.1452 b --- 0.505 0.1552 c --- 0.520 0.190 BJ69F a --- 0.310 0.090 b --- 0.330 0.140 c --- 0.320 0.100 CJ69F a --- 0.315 0.115 b --- 0.330 0.150 c --- 0.320 0.110 AJGQUKF a --- 0.585 0.180 b --- 0.600 0.215 c --- 0.620 0.190 AN69 a --- 1.280 0.430 b --- 1.330 0.520 c --- 1.290 0.440 1Essences were diluted 1.0 ml to 25.0 ml with water. 2No point of inflection or peak was present. 94 .00HucooH 0000 000 HH «Hana 000. 00.0 0H.0 0H.0 00.0 00.00 00.0 00.0 00.0 00.H 00.00 00.H H0.0 n H 00.0 0H.0 00.0 00.0 0H.00 00.0 0H.0 00.0 00.H 00.00 0H.H 00.0 o H 00000 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 00.H n H 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 a H 00000 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 0H.0H 00.H 00.0 n H H0.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 H0.0 00.0H 00.H 00.0 a H 00004 0H.0 0H.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 0H.0 00.0 n H 00.0 0H.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 H0.0 a H 00004 00.0 0H.0 0H.H 00.0 00.00 00.0 00.0 00.0 00.0 00.0H H0.H 0H.0 n H 00.0 0H.0 0H.0 00.0 00.00 00.0 00.0 00.0 H0.0 00.0H 00.H 0H.0 a H 00004 00.0 H0.0 00.0 00.0 00.HH 00.0 H0.0 00.0 0H.0 00.00 00.0H 00.0 n H 00.0 00.0 00.0 00.0 00.0H 00.0 00.0 00.0 00.0 00.00 00.0H 00.0 a H 00000 00.0 00.0 0H.0 00.0 00.00 00.0 00.0 00.0 00.H 0H.00 00.0 00.0 n H 00.0 00.0 0H.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 a H 000004 00.0 0H.0 0H.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 n H 00.0 HH.0 0H.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 a H 0\0H000< 00.0 H0.0 00.0 00.0 00.0H 00.0 00.0 0H.0 00.0 0H.00 00.0 H0.0 n 0 00.0 00.0 00.0 00.0 00.0H 00.0 00.0 0H.0 00.0 00.00 00.0 00.0 0 H 0\0H000< 00.0 H0.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 n H 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0H.0 00.0 00.H0 00.0 0H.0 a H 0\0H000< 00.0 0H.0 0H.0 00.0 0H.00 00.0 00.0 00.0 00.H 0H.00 0H.H 00.0 n 0 00.0 HH.0 HH.0 00.0 00.00 00.0 00.0 00.0 00.H 00.00 00.H H0.0 n 0 00.0 HH.0 HH.0 00.0 00.00 00.0 00.0 00.0 00.H 0H.00 00.H 00.0 n H 00.0 0H.0 0H.0 00.0 00.00 00.0 00.0 00.0 00.H 00.00 00.H 0H.0 a H 0000 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 a 0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 0H.00 00.0 00.H a 0 00.0 00.0 00.0 00.0 00.00 00.0 H0.0 00.0 00.0 00.00 00.0 00.H a H 00.0 00.0 00.0 00.0 0H.H0 00.0 00.0 00.0 H0.0 00.00 H0.0 0H.0 a H 0000 0H.0 00.0 00.0 00.0 0H.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 n 0 0H.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 H0.00 00.0 00.0 a 0 0H.0 00.0 00.0 00.0 00.H0 00.0 00.0 00.0 00.0 H0.00 00.0 00.0 n H 0H.0 00.0 00.0 00.0 00.00 00.0 H0.0 00.0 00.0 0H.00 00.0 00.0 0 H 00z< 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 0H.0 H0.0H 00.H 00.0 n 0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 0H.0 00.0H 00.H 00.0 0 0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.0H 00.H 00.0 a H 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 H0.0 H0.0H 00.H 00.0 0 H 0004 00.0 0H.0 00.0 00.0 H0.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 n 0 00.0 00.0 00.0 00.0 0H.00 00.0 00.0 00.0 00.0 00.00 00.0 00.0 a 0 00.0 0H.0 00.0 00.0 00.00 H0.0 00.0 00.0 00.0 00.00 00.0 00.0 n H 00.0 00.0 00.0 00.0 00.H0 H0.0 00.0 00.0 00.0 00.00 00.0 00.0 a H 0004 0H.0 00.0 00.0 00.H 00.0 00.0 00.0 00.0 00.H 0H.H0 00.0 00.0 n 0 0H.0H 00.0 00.0 00.H 00.0 00.0 00.0 00.0 00.0 00.00 00.H 00.0 a 0 00.0 00.0 00.0 00.0 H0.0 00.0 0H.0 00.0 00.0 00.H0 00.H 00.0 n H 00.0H 00.0 00.0 0H.H 00.0 00.0 00.0 00.0 00.0 00.00 00.H 00.0 a H 0000 00.0 00.0 0H.H 00.0 H0.00 00.0 H0.0 00.0 00.0 00.0H 00.H 00.0 n 0 00.0 00.0 0H.H 00.0 0H.00 00.0 H0.0 00.0 00.0 00.0H 00.H 00.0 0 0 00.0 00.0 00.H 00.0 00.00 00.0 00.0 00.0 H0.0 00.00 00.H 00.0 n H 00.0 00.0 00.H o0.0 00.00 00.0 H0.0 00.0 H0.0 00.00 H0.H 00.0 a H 0000 00.0 00.0 0H.0 00.0 00.0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 n 0 00.0 00.0 0H.0 00.0 00.0 00.0 00.0 00.0 00.0 00.H0 H0.0H 00.0 a 0 00.0 00.0 0H.0 00.0 00.0 00.0 00.0 00.0 0H.0 00.00 00.HH 00.0 n H 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.00 H0.0H H0.0 a H 000< ucooo uOuoumouCH Hauou umnucoouom a x o H a o m m a o m c 0000 le OHUUOm OOCOQMM exoflm Immu— .ooocomno on» new muse UH£QCHOOueEOqu vHHounooo .on oHnaa IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII élHIWILIHILIIZHMHIflilallljlfllwlllll MI W 2