IDENTIFICATION OF SOME CHEMICAL COMPONENTS IN CHICKEN FLAVOR Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY Lewis J. Minor I 1.964 345515 LIBRAR 1' Michigan Sta! UM, This is to certify that the thesis entitled IDENTIFICATION OF SOME CHEMICAL COMPONENTS IN CHICKEN FLAVOR presented by Lewis J. Minor has been accepted towards fulfillment of the requirements for Ph.D. Food Science degree in 7 M or professor Date May 29, 1964 0-169 ABSTRACT IDENTIFICATION OF SOME CHEMICAL COMPONENTS IN CHICKEN FLAVOR by Lewis J. Minor Some of the chemical components from cooked light and dark chicken tissue have been identified using gas, column, thin-layer and paper chro- matography, UV and I-R absorption spectra, a trap reaction technique utilizing functional group analysis, and a solubility classification method. Precursors and non-volatiles were studied by chemical methods. The protein, fat, moisture and ash contents were determined for the raw mus- cle and cooked-freeze-dried slurry from light and dark meat for three classes of birds. Concentrations of inosinic acid, creatine/creatinine, diacetyl/acetoin, sulfhydryl compounds, inorganic sulfides, cystine and methionine, and pH values were also determined for these samples. Yield studies gave average values of 52.7% for raw meat and 36.3% for cooked meat from roasters; 48.8 and 33.5%, from heavy hens and 47.7 and 35.5% from light hens, respectively. Organoleptic studies showed that after 50 hrs. cooking-distillation at 180°F, broth from heavy hens had better flavor than that from roasters or light hens. Broth from light muscle had typical chicken flavor, while broth from dark muscle tended to re- semble beef broth in flavor. After 50 hrs. cooking-distillation, the pH of the heavy hen breast meat slurry increased from 5.8 to 6.2 and that from the heavy hen leg meat slurry increased from 6.1 to 6.3 indicating that acidic and neutral constituents were distilled off. Some basic con- stituents (possibly sulfides and methylamine) were also distilled over Lewis J. Minor as evidenced by a change in the pH of the water trap (8.1-9.6) through which the volatiles were passed. The higher molecular weight acidic, neutral and basic compounds remained in the aqueous phase. Volatiles that were trapped in a liquid nitrogen medium (-l96°C) had pH values of 8.0 or higher. Gas chromatographic studies showed that 20 month old hens contained the same volatile components as 12 week old birds of identical origin that were fed the same ration, but some of the cooked volatile components from the older birds were present in higher concentrations. Chemical tests indicated that the intestinal contents from the heavy hens contained carbonyls, sulfides, disulfides and mercaptans, but in lesser amounts than for the meat from the same birds. Gas chromatography of the volatile fraction released upon cooking fryer breasts resulted in identification of ethane, propane and carbon dioxide. Hydrogen sulfide and carbon dioxide were identified in the fraction distilling off at -140°F. Ammonia was identified chemically in the cooked volatile fraction. Odor tests indicated the presence of car- bonyl sulfide, although further identification was not made. A steam distillate from heavy hen breast muscle contained two phos- phatidyl lipid components, which were tentatively identified as cardiolipin and either phosphatidic acid or phOSphatidyl inositol. Identifications were made using thin-layer chromatography. Volatiles derived by cooking-distillation of young birds in an oxy- gen containing atmOSphere were trapped in 2,4-dinitr0phenylhydrazine (2,4- DNP) and lead acetate solution, reSpectively. A total of nine carbonyls Lewis J. Minor were identified. Eight were monocarbonyls; namely, acetaldehyde, propanal, n-butanal, n-pentanal, n-hexanal, n-octanal, ethyl-methyl ketone and acetone. One was a polycarbonyl; namely, diacetyl. Results were in agreement with those of earlier workers. Sulfides, disulfides and mer- captans were identified by preparing their derivatives from the cooked volatiles and identifying the derivatives by acid decomposition followed by gas chromatography and chemical methods. Model systems demonstrated the importance of glutathione as a pre- cursor for the sulfur containing amino acids (specifically, cystine and cysteine) and of 2,3-butanedione as an active carbonyl in Strecker degra- dation. Copious amounts of hydrogen sulfide were evolved by this degra- dation which occurred in a warm aqueous solution. The product had a taste resembling chicken broth. Some of the chemical components of heavy hen leg muscle volatiles that were identified by the functional group trapping technique included a total of 29 compounds from 30 peaks separated from the total cooked volatile fraction. These compounds were methyl mercaptan acetone, methanol, dimethyl sulfide, methyl-ethyl sulfide, methylamine, diethyl sulfide, ethanol, acetaldehyde, methyl-iso-propyl sulfide, 2,3-butanedione, methyl disul- fide, acetoin, ethyl-n-prOpyl sulfide, ethyl disulfide, ethyl mercaptan, diprOpyl sulfide, n-prOpyl mercaptan, n-hexanal, 2,4-pentanedione, iso- amyl alcohol, nramyl alcohol, n-heptanal, ethanol-amine, n-hexanol, 2- heptanone, n-heptanol, ethane and propane. By the same method using an Apiezon-L column, a flame-ionization detector and temperature programming from loo-250°C with an F and M model Lewis J. Minor 500 gas chromatograph, a total of twenty-five peaks were obtained from heavy hen breast muscle. Compounds that were identified by means of the functional group analysis trapping technique were ethane, prOpane, ace- tone, methanol, dimethyl sulfide, methylamine, methyl formate, diethyl sulfide, ethanol, acetaldehyde, 2,3-butanedione, methyl disulfide, ace- toin, n-pentanal, ethyl-n-prOpyl sulfide, iso-butanol, ethyl disulfide, n-butanol, diprOpyl sulfide, n-pr0pyl mercaptan, n-hexanal, 2,4-pentane- dione, n-heptanal, 2-heptanone and n-heptanol. The concentrations of acetoin and diacetyl in light and dark muscle samples were determined by steam distillation followed by a colorimetric procedure. Results confirmed reports of earlier workers. Inosinic acid determinations were made using a convenient UV absorption method at a wavelength of 250 mu. Results were obtained for the raw light and dark tissue samples from roasters, heavy- and light-weight hens. Creatine/creatinine, cystine, methionine, sulfhydryl and inorganic sulfide determinations were made by colorimetric procedures on raw light and dark muscle samples from heavy hens, roasters and light hens, and also on the cooked-freeze-dried muscle slurries from these birds. In addition, cystine and methionine were determined by a microbiological method. Several important flavor characteristics of light and dark chicken muscle and of young and old chickens were revealed by these studies. Young and old chickens had the same components in their reSpective volatile fractions, but certain constituents were present in higher concentrations from older birds. Using a model system, glutathione was decomposed in Lewis J. Minor the presence of moist heat and 2,3-butanedione. Glycine, glutamic acid and cystine were formed and capious amounts of hydrogen sulfide, ammonia and carbon dioxide were liberated. A chicken-like flavor was obtained by this decomposition. Both glutathione and 2,3-butanedione are contained in chicken muscle and were shown to be precursors of hydrogen sulfide, organic sulfides, disulfides and mercaptans. Sulfur esters may be impor- tant flavor constituents also, but their existence in chicken muscle samples was not verified. Light meat contained more inosinic acid, creatine/creatinine, and acetoin/diacetyl than dark meat. Inosinic acid enhanced chicken flavor. Acetoin/diacetyl imparted a cooked, buttery and oily note to flavor, which confirmed earlier reports. Creatine/ creatinine were bitter but may have imparted desirable additive and/or synergistic flavor effects. Some steam-distillable phosphatidyl compon- ents, when freshly prepared, contained "chicken essence", but the desir- able aroma disappeared in a few hours. Results indicated that the phos- phatides may have acted as electrovalent bonding agents for the flavor volatiles. Sulfur compounds possessed a characteristic "meaty" aroma of greater importance to flavor than carbonyls. However, carbonyls may function at sub-threshold concentrations by exerting synergistic and additive flavor effects. Differences between flavor from light muscle versus dark muscle was not traceable to differences in the concentration of precursors. There may be a relationship between heme compounds and the characteristic flavors of the various red meats. Electropositive metallic bonds present in the iron of the heme moiety of myoglobin and hemoglobin could serve as binding media for electronegatively charged Lewis J. Minor flavor volatiles. It is certain that the removal of sulfur compounds by mercuric chloride and mercuric cyanide solutions resulted in an almost complete loss of "meaty" odor from the cooked volatiles of light or dark chicken meat. IDENTIFICATION OF SOME CHEMICAL COMPONENTS IN CHICKEN FLAVOR By Lewis J; Minor A THESIS Submitted to Michigan State university in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Food Science 1964 (9 Copyright by LEWIS JOSEPH MINOR 1967 ACKNOWLEDGEMENTS The author is sincerely grateful to A. M. Pearson, Professor of Food Science for his guidance in selecting the author's course of study and constructive criticism throughout the experimental work and various phases of the research project; to B. S. Schweigert, Chairman of Food Science for advice and encouragement; to C. M. Stine, Professor of Food Science for the use of laboratory facilities, advice, encouragement and forbearance; to J. R. Brunner, Professor of Food Science, for the use of equipment and suggestions; to L. E. Dawson, Professor of Food Science for securing chickens and other c00peration; and to L. J. Bratzler, Professor in Food Science for his advice and encouragement. The author wishes to thank Mrs. B. Eichelberger for her help and for typing the thesis. The author thanks his fellow graduate students and especially R, L. Bradley, J. Sargent, Mrs. G. McGinnis, C. Y. Peng, E. McCabe, G. Wells, R. Wilkinson, R. W. Porter, M. El-Gharbawi and others whose cooperation is deeply appreciated. To his wife, Ruth, children, eSpecially Ruth Ann, as well as manage- ment and personnel of the L. J. Minor Corp., the author gives Special thanks. ii TABLE OF CONTENTS INTROD mT ION . C O O O O C O C C O C O O O O O 0 EXPERIMENTAL OBJECTIVES . . . . . . . . . . . . REVIEW OF LITEMTURE O O O O O O O O O O O O O 0 Meat Flavor . . . . . . . . . . . . . . . . Beef Flavor . . . . . . . . . . . . . Pork and Mutton Flavor . . . . . . . . Fish Flavor . . . . . . . . . . . . . Undesirable Meat Flavor . . . . . . . Chicken Flavor . . . . . . . . . . . . Effects of Feed on Flavor . . . . . . How Cooking Method Affects Flavor . . Effects of Processing and Storage . . Other Factors . . . . . . . . . . . . Chilling . . . . . . . . . . . . Sex and age . . . . . . . . . . . Aging and Handling . . . . . . . MATERIALS AND METHODS . . . . . . . . . . . . . Chickens . . . . . . . . . . . . . . . . . Proxmate Analys is o o o o o o o o o o o 0 ‘Moisture . . . . . . . . . . . . . . . Ether Extract . . . . . . . . . . . . Protein . . . . . . . . . . . . . . . Ash . . . . . . . . . . . . . . . . . Creatine and Creatinine . . . . . . . Diacetyl and Acetoin . . . . . . . Inosinic Acid . . . . . . . . . . . . Hydrogen Ion Concentration pH . . Inorganic Sulfide Determination . Cystine and Methionine . . . . . . . . Sulfhydryl Content . . . . . . . . . . Gas Chromatography of Chickens Differing in Age Separation and Identification of C02, H28 and Carbonyl S ‘11 fide O O O O O O O O O I O O O O O O O O O O 0 Gas Chromatographic Analysis . iii Page 13 l4 19 23 33 36 37 39 39 40 40 42 42 45 45 45 45 46 46 50 53 53 55 55 57 64 66 68 Gas Chromatographic Comparison of Total Cooked Volatiles from the Carcass and Intestinal Contents of Old Hens . . Gas Chromatographic Analysis of the Low Boiling Fraction in Fryers of Unknown Origin . . . . . . . . . . . . . . . Cooked Meat Yields from Roasters, Heavy- and Light-weight Hens O O O O O O O O O O O . ‘Method of Distillation and Thin-Layer Chromatography of a PhOSpholipid Fraction Containing "Chicken Essence" . . . COOkingcoo-000.000.000.000... Thin-Layer Chromatographic Procedure . . . . . . . . Separation and Identification of Carbonyls and Sulfur Cooked Chicken . . Compounds in the Volatile Fraction of Cooking and Distillation . . . . Fractionation and Identification Column Chromatography . Preparing the Columns . ‘Melting Points . . . . . Paper Chromatography . . Preparation of Authentic Test for Sulfides . . . Test for Organic Disulfides of 2,4-DNPHS 2, 4- DNPHS Separation of Sulfur Containing Volatiles . . . . . Test for Thiols . . . . . . . . . . . . . . . . . . Separation and Identification of Carbonyls and Sulfur Compounds in the Volatile Fraction of Cooked Intestinal contents Of ChiCken O O O O O O O O O O I O O O O O I O 0 Cooking and Distillation . . . . . . . . . . . . . . Fractionation and Identification of 2,4-DNPHS . . . Paper Chromatography . . . . . . . . . . . . . . . . An Apparatus for Splitting and Trapping the Volatile Fraction of Cooked Chicken . . . . . . . . . . . . . . . Functional Group Analysis of Components from the Volatile Fraction of Cooked Chicken by Qualitative Color Tests . . . . . . Alcohols . . . . . . . . Carbonyls . . . . . . . Schiff's Reagent . . . . Esters . . . . . . . . . Alkyl Nitriles . . . . . Chemical Tests Page 69 70 71 72 72 74 77 77 79 79 8O 82 82 84 85 85 86 86 87 87 88 88 9O 99 100 100 101 101 101 101 ‘Mercaptans . . . . . . . . . . . Mercaptans . . . . . . . . . . . . . . . AlleSUlfideSooooooooooooo Alkyl Disulfides . . . . . . . . . . . Amines (primary, secondary and tertiary) Aromatic Nucleus and Aliphatic Unsaturation . . Tertiary Ainines O O O O O O O O O O O O O O O O O O 0 Nitroparaffins O O O O O O O O O O O O O O O O O O Phenols and Enols . . . . . . . . . . . . . . . . . . Amoniaooooooooooooooooooooooo Preparation of sulfide disulfide, Mercaptan and Carbonyl Derivatives. . . . . . . . . . . . . . . Solubility Classification of Compounds . . . . . . . . Functional Group Analysis by Gas Chromatography Chemical Trap Reaction Technique . . . . Gas Chromatography . . . . . . . Sampling for Gas Chromatography Control Samples . . . . . . . . Functional Group Analysis Samples Preparing the Glassware . . . . . Organoleptic Testing . . . . . . Hydrogen Ion Concentration Tests 'Method of Preparation, Sampling and Gas Analysis of Sulfur Derivatives . . . . . . . . . . . . . with the Chromatographic MOdElExperj-mentscocooooooooooooooo ‘Model Experiment with Sulfide and Lactate Fonmation . Model Experiment with Sulfide and Lactate Formation and Glutathione Decomposition . . . . . . . . . . . . EXPERIMENTAL RESULTS AND DISCUSSION . . . . . . . . . . . . Comparing Total Cooked Chicken Volatiles from Carcasses of Young and Old Female Chickens Hydrogen Sulfide . . . . . . . . Carbon Dioxide . . . . . . . . . Carbonyl Sulfide . . . . . . . . Gas Chromatographic Analysis . . the Whole Vapor Fractometer Tests on Cooked Breast Muscle A Phospholipid Fraction Containing "Chicken Essence" from Fryers Page 101 102 102 102 102 102 103 103 103 103 103 104 106 106 107 107 108 110 111 111 111 112 113 114 115 115 115 116 117 117 124 126 Comparison of Total Cooked Chicken Volatiles from the Whole Carcasses of Old Hens with the Volatiles from the Cooked Intestinal Contents of the Same Birds . . . . . . . . Gas Chromatography . . . . . . . . . . . . . . . Chemical Identifications . . . . . . . . . . . . Results of Carbonyl Tests . . . . . . . . . . . Results of Sulfide, Disulfide and Mercaptan Tests The Chemical Identification of Carbonyls and Sulfur Contain- ing Compounds Present in the Volatile Fraction of Cooked Ch 1Cken I I I I. O O O O O O O O C O O O O I O O O I O O O O Carbonyls . . . . . . . . . . . . . . . . . . . Sulfur Compounds in the Cooked Volatile Fraction Mercaptan Test . . . . . . . . . . . . . . . . . Sulfide Test . . . . . . . . . . . . . . . . . . Test for Disulfides . . . . . . . . . . . . . . Cooked Meat Yields from Roasters, Heavy- and Light-weight Hens O O O O O O O O I O O O O O O O O O O O I I O O smary O O O O O O O O O O O O O O C O O O O . Chemical Analyses . . . . . . . . . . . . . . . . . . Acetoin/diacetyl . . . . . . . . . . . . . . . . Inosinic Acid . . . . . . . . . . . . . . . . . Creatine/creatinine . . . . . . . . . . . . . . Cystine and Methionine . . . . . . . . . . . . . Sulfhydryl Content . . . . . . . . . . . . . . . Ammoniacal Nitrogen . . . . . . . . . . . . . . Organoleptic Evaluations of the Cooked Broth Samples Solubility Classification of the Volatile Fraction of Chi-Cken O O O O O O O O O O O I O I 0 O O O O O O O 0 Fraction of Cooked Chicken . . . . . . . . . . . Qualitative Chemical Tests . . . . . . . . . . . Sulfur Compounds . . . . . . . . . . . . . . . . Results of Sulfur Compound Identifications . . . Results for Sulfides . . . . . . . . . . . . . . Disulfides and Mercaptans . . . . . . . . . . . Gas Chromatography . . . . . . . . . . . . . . . Functional Group Analysis of Cooked Chicken Volatiles Light and Dark Muscle After 50 hrs. Cooking and Distillation Under Oxidation-inhibiting Conditions . . . . . . . . vi from Page 130 130 130 132 133 133 134 143 144 144 144 146 156 157 157 161 161 162 163 163 164 166 166 170 172 172 172 173 173 176 Page Organoleptic Evaluation of the Residual Effluents Contained in the Liquid Nitrogen Traps Following Gas Chromatography 183 Madel systems 0 O O O O O O O O O O O O O O O O O O O O O O 186 Chemical Spot Tests . . . . . . . . . . . . . . . . . 189 SIMMANDCONCLUSIONS00000000000000.0000.192 BIBLIOGRAHIYOOOOOOOOOOOOOOODOOOOOOO.O.196 APPENDIX 0 O O C O O 0 O O O O O O O O O O O O O O O O O O O O O 214 vii Table 10 11 11 12 13 LIST OF TABLES Page Volatile components separated from old hens and younger female chickens . . . . . . . . . . . . . . . . . . . . . 121 Vapor fractometer separations of the low boiling fraction in fryer breast muscle distillate . . . . . . . . . . . . 125 Thin-layer chromatographic results for "chicken essence" from heavy hen breast muscle . . . . . . . . . . . . . . 127 Physical prOperties of 2,4-DNPHS of carbonyl compounds from the cooked intestinal contents of old hens . . . . . 132 Properties of 2,4-DNPHS of carbonyl compounds isolated from cooked chicken volatiles . . . . . . . . . . . . . . . . 136 Paper chromatography and UV absorption maxima of Spots of 2,4-DNPH derivatives . . . . . . . . . . . . . . . . . . 138 Summary of the average uncooked yields obtained from 35 chickens of each class by weight and percentage . . . . . 147 Summary of the average cooked yields obtained from 35 chickens of each class by weight and percentage . . . . . 148 Summary of the waste, solids, separation loss and liquids obtained from 35 chickens of each class by weight and percentage . . . . . . . . . . . . . . . . . . . . . . . .150 Average yield of cooked lean meat and total cooked edible meat based on uncooked eviscerated weight (EW) . . . . . 152 Components of raw-frozen breast and leg muscle and cooked- freeze-dried breast and leg muscle meat-broth slurry from roasters, heavy- and light-weight hens . . . . . . . . . 158 Components of raw-frozen breast and leg muscle and cooked- freeze-dried breast and leg muscle meat-broth slurry from roasters, heavy- and light-weight hens (continued) . . . 159 Organoleptic evaluation of broth samples after 50 hrs. cooking at 180°F under "oxidation-inhibiting conditions", and pH values for original muscle/H20 slurry and cooked broth . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Organoleptic evaluation of volatile fractions components by the solubility classification method (Cheronis and Entrikin, 1961) . . . . . . O O O O O O O O O O I O O O O 167 viii Table 14 15 16 17 18 Organoleptic evaluation of volatile fraction effluent aroma from various reagent traps . . . . . . . . . . . . 168 Results of qualitative chemical tests on the volatile fractions of cooked chicken based on the solubility classification method of Cheronis and Entrikin (1961) and for sulfur components by the method of Folkard and Joyce (1963) . . . . . . . . . . . . . . . . . . . . . . . . . 171 Gas chromatographic analysis of volatiles obtained by acid decomposition of a mixture of the mercuric cyanide and mercuric chloride derivatives of cooked chicken volatiles 175 Organoleptic characteristics and pH values of liquid nitrogen trap contents after passage of the total cooked fractions of light and dark muscle through the reagent trap and chromatographic analysis . . . . . . . . . . . . 184 Gas chromatographic analysis of a model system in which glutathione was decomposed in an aqueous media containing other components of chicken muscle . . . . . . . . . . . 188 ix LIST OF FIGURES Figure Page 1 Zn Chloride creatinine standard curve . . . . . . . . . . 49 2 Diacetyl standard curve . . . . . . . . . . . . . . . . . 52 3 Inosinic acid standard curve . . . . . . . . . . . . . . 54 4 Inorganic sulfide standard curve . . . . . . . . . . . . 56 5 L-cystine standard curve . . . . . . . . . . . . . . . . 58 6 L-methionine standard curve . . . . . . . . . . . . . . . 59 7 Cystine standard curve . . . . . . . . . . . . . . . . . 60 8 Methionine standard curve . . . . . . . . . . . . . . . . 61 9 Sulfhydryl standard curve . . . . . . . . . . . . . . . . 63 10 Cooking and distillation apparatus . . . . . . . . . . . 65 11 Liquid-nitrogen-dipped trap . . . . . . . . . . . . . . . 67 12 Radialdmanifold stream-splitter...Large absorption trap with ball and socket plugs clamped on. A 12 cm. rule. Small absorption trap with surgical seal and plug in place to retain volatiles for gas chromatographic analysis . . 92 13 PerSpective view of cooking-distillation apparatus showing nitrogen gas tube-chicken slurry in flask water-trap and cluster of reagent traps surrounding the radial-manifold Stream-Splitter o o o o o o o o o o o o o o o o o o o o o 95 14 Closeup view of water trap and a cluster of 8 reagent traps surrounding the radial-manifold stream-splitter . . . . . 97 15 Closeup view of cooking-distillation set-up for trapping cooked volatiles with dry ice-ethanol and liquid nitrogen traps connected in series to a small trap containing a functional group reagent connected to the radial-manifold stream-splitter . . . . . . . . . . . . . . . . . . . . . 98 16 Gas chromatograms of 10 ml. volatiles from cooked whole carcass old hens and young birds--trap warmed to (23°C) 1/4" x 13' (DEHS) column-70°C and 21 p.s.i. helium . . . 119 Figure 17 18 19 20 21 22 23 24 25 26 27 Page Gas chromatograms of 10 m1. volatiles from cooked whole carcass hens and young birds-trap warmed to 75°C-1/4" x 13' (DEHS) column-70°C and 21 p.s.i. helium . . . . . . . 120 Gas chromatogram of 5 m1. volatiles from cooked intestinal contents of old hens. Trap warmed to 23°C - 1/4" x 13' (DEHS) column-~70°C and 20 p.s.i. helium . . . . . . . . 131 Infra-red Spectra for the authentic and unknown samples of acetone-2, 4-DNm o o o o o o o o o o o o o o o o o o o o 139 Infra-red Spectra for the authentic and unknown samples of diacetYI-Z’A'DNPI'I O o o o o o o o o o o o o o o o o o o o 140 Infra-red spectra for the authentic and unknown samples of ethyl-methyl-ketone-Z,4-DNPH . . . . . . . . . . . . . . 141 Infra-red Spectra for the authentic and unknown samples of n-hexanal'2,4'DNmooooooooooooooooooo142 Percentage of cooked meat as: A-breast, B-leg, C-back, D-wing, E-neck and F-total compared with the percentage of bone for these same components, in heavy hens, light hens and roasters . . . . . . . . . . . . . . . . . . . . . . 151 Percentage of Armeat, B-skin, C-fat, D-total cooked edible portion in heavy hens, light hens and roasters . . . . . 153 Chart for calculating the cost/lb. of cooked lean meat for three classes of chickens . . . . . . . . . . . . . . . . 155 Nonlinear temperature programmed separation of 3 m1. of volatiles from acid decomposition of mercuric cyanide and mercuric chloride derivatives of cooked chicken volatiles 174 Nonlinear temperature programmed separation of 2 m1. of total cooked volatiles from heavy hen leg muscle . . . . 177 Composite of gas chromatographic analysis of cooked vola- tiles from heavy hen leg muscle by a functional group mathod..........o...o..........178 Nonlinear temperature programmed separation of 2 m1. of total cooked volatiles from heavy hen breast muscle . . . 180 Composite of gas chromatographic analysis of cooked vola- tiles from heavy hen breast muscle by a functional group method 0 O O O O O O O O O O O O O O O O O O O O O I O O 181 xi Figure Page 28 Nonlinear temperature programmed separation of75 ml. of volatiles from aqueous decomposition of glutathione and other compounds in a model system . . . . . . . . . . . . 187 xii Table LIST OF APPENDIX TABLES Page Ration fed to White Leghorn laying hens (20 months old) and younger female chickens (12 weeks old) raised from the same flock . . . . . . . . . . . . . . . . . . . . . 214 Determinations of L-cystine and L-methionine content of raw-frozen and cooked-freeze-dried meat slurry from roasters, heavy- and light-weight hens . . . . . . . . . 215 Henderson and Snell's (1948) single median for determining 14 different amino acids with the appropriate amino acid Mitted O O I I O O I O O O O O O O O C O O O O O O O O 216 Colorimetric determination of percent transmission of cystine in diluted hydrolysate samples of raw-frozen and cooked- freeze-dried meat slurry from roasters, heavy- and light- weight hens . . . . . . . . . . . . . . . . . . . . . . 217 Colorimetric determination of percent transmission of methionine in diluted hydrolysate samples of raw-frozen and cooked-freeze-dried meat slurry from roasters, heavy- and light-weight hens . . . . . . . . . . . . . . . . . 218 xiii INTRODUCTION For nearly two centuries scientists have studied the chemical com- position of meat extracts searching for the key to meat flavor. Hydrolyzed vegetable proteins, casein digests and yeast autolysate extracts are preparations having characteristic flavors of their own, yet are compatible with certain meats or meat and vegetable mixtures. ‘Monosodium glutamate, disodium guanylate and disodium inosinate have also proven useful in hmproving the flavor scores of some meat and vegetable mixtures (Gaul and Raymond, 1963; Kurtzman and Sjgstrom, 1963; Toi 25 $1., 1963), but none of these products or any known combination of them can replace meat fla- vor. Their primary functions are to provide additive or synergistic flavor improvement and to extend meat flavor. Of all the condiments, meat extract provides the truest resemblance to the mouth satisfaction indigenous to freshly prepared meat stock. Relatively little is known of the many chemical reactions that occur on heating meat. The complexity of meat and its associated flavors are responsible for the difficulties encountered in elucidating the chemical basis of meat flavor. DesPite the efforts that many workers have expended for more than one hundred and fifty years, the available information concerning meat flavor is at best fragmentary. In addition to the continuing research on meat extract, elucidation of the volatile and non-volatile flavor components associated with beef, pork, mutton, chicken and fish has been undertaken. Scientists in many countries have contributed to the present status of our knowledge of meat flavor. The natural resources of individual nations have played an im- portant role in determining whether support and research emphasis should be given to investigations on fish, fowl, animal flavor or to the flavor of meat extract. Fat flavor has been of special interest due to the large amounts of fat present in meat and the tendency of fat to undergo chemical changes, even at sub-freezing temperatures. Lipid oxidation in lean tissues has been reviewed by Watts (1961). Some components present in pork fat and associated with sex odor have been reported by Craig and Pearson (1959). Hofstrand and Jacobson (1960) stated that the presence or absence of fat did not affect the taste of mutton broth, but that water soluble com- pounds in depot fats volatilize and influence the aroma. Off-flavors in meat caused by irradiation have been the subject for numerous investigations, as any value that this method for preserving meat may have is limited by palatability scores. Deve10pment of new techniques and scientific instruments has made flavor research more objective. Recent developments in instrumentation provide both accuracy and sensitivity that supplements chemical and organoleptic testing procedures, which have been established and stan- dardized. Physical methods have now become indisPensable to the chemist or food scientist. Every modern laboratory has its Spectrographic and chromatographic equipment. To these methods gas chromatography has been added, which is in a state of flux due to constant improvement in the sensitivity of instruments and new identification techniques. Enthusiasm is generated from the fact that gas chromatographic analysis, as it is practiced today, not only achieves separation, but also identification and quantitative determination. By means of gas chromatography joined with other modern processes an automatic analysis is almoSt within reach. It may now be possible to achieve an automatic analysis of volatile con- stituents from cooked meat by passing the volatiles through a gas chroma- tograph to separate them, an infrared and/or ultraviolet Spectrograph to identify molecular configuration, a time-in-flight mass spectrometer to determine molecular weight and nuclear magnetic resonance equipment to confirm the molecular structure of each compound. Molecular separations of non-volatiles can now be made by dialysis through.membranes of controlled porosity size, by elution through columns known as melecular sieves, and ion-exchange or neutral columns. Ion exchange paper is also available for rapid separation of nucleotides as reported by Smillie (1959). Thin layer chromatographic media and tech- niques are likewise undergoing rapid development. X-ray diffraction can be used to identify different compounds present in a Single impure crystal. Complete protein analysis can be made in less than 24 hours. Histochemical techniques and enzyme research deve10pments will undoubtedly assume important roles in developing the future knowledge concerning meat flavor and its precursors. EXPERIMENTAL OBJECTIVES The overall experimental objective of this study was to elucidate the components responsible for chicken flavor; more Specifically, the objectives were as follows: 1. To develop methods for separating and collecting the volatile compounds from cooked chicken in order to identify them by chemical and gas chromatographic analysis. 2. To determine some chemical compounds in raw-frozen and cooked- freeze-dried light and dark muscle from three classes of chicken that would indicate changes in precursors and differences in flavor volatiles. 3. To place special emphasis on the identification of sulfur com- pounds in cooked chicken volatiles and to determine their precursors. 4. An added objective was to determine meat yields from three classes of birds, which were used as a major source of material for the flavor studies. REVIEW OF LITERATURE Numerous definitions of flavor can be given but the one chosen is that of Kazeniac (1961), in which he used four designations as proposed by Sjgstrom (1955). The first was taste, consisting of saltiness, sweet- ness, sourness and bitterness. The second was aroma, which was used to describe perceptible odor sensations. The third was body, which re- ferred to the texture or sensation caused by chewing, but having nothing to do with taste or aroma. The fourth was mouth satisfaction, which implied increased salivary stimulation, blending and pleasantness, but had little to do with taste or aroma. Thus described, flavor is an interaction of these four basic sensations. Meat Flavor A classical report on the constitution of meat extract and its mode of preparation from lean muscle was published in 1847 by the immortal chemist, Justus von Liebig (1803-73). Liebig related how Berzelius, Gmelin, Braconnot and Chevreul some 40 years earlier had tried without success to find a meat substitute. Liebig extracted not only ox, but chicken, fish, pork, deer and fOX‘mUSCIG as well. His procedure was to mix 10 lbs. of muscle with 5 lbs. of water by hand. The solution was pressed out and collected. He repeated this procedure three times. The combined extracts were then filtered through linen cloth and placed in a large glass flask which was immersed and heated in a water bath. Fat was removed constantly as it rose to the top of the solution. When all of the fat was separated, Liebig filtered the solution and simmered it -5- in a water bath. After prolonged heating the solution darkened, and was concentrated under a vacuum to complete the process. Liebig (1847) detenmined the creatine and creatinine contents of the muscle extracts, and concluded that creatinine is an important con- stituent of muscle. He extracted the remaining concentrate with alcohol, after first removing the creatine and creatinine and isolated a new acid, which he called inosinic acid after the Greek word meaning muscle. When inosinic acid had been extracted, lactic acid remained behind. He then determined the amount of lactic acid and inorganic constituents in the extract. Liebig also isolated sarcosine (methyl aminoacetic acid), which he named after the Greek word meaning meat, and tyrosine, which he named after the Greek word meaning cheese. After examining these ex- tracts, he concluded that the odor and taste coming from the lean muscle resembled roast meat. After they were cooked and concentrated, there was no fundamental difference between ox, doe, fox, pork or chicken mus- cle. The amount of albumin released by heating depended upon the age of the animal, with lesser amounts coming from older animals. Liebig's ex- tract is recognized in commerce as a derivative of ox or beef muscle. Many scientific studies have been devoted to the elucidation of its chemical nature and value. Papers on meat extract have been published by Micko (1908, 1909, 1914a,b), Wolff (1913), Krimberg and Israilskii (1914), Salkowski (1914), Smith (1916), Waser (1917), Geret (1918), Kluender (1920), Korchow (1927), Kappeler-Adler and Stern (1931), Hordh (1936), Boon (1937), Kokovikhina (1942), Tempus (1956), Wood (1956, 1957, 1961), Lissi ggual. (1961), Bender and Ballance (1961), Pocchiari 5531. (1962) and Wood 392.1.- (1962). In a review on meat flavor, Doty 25 a1, (1961) indicated that modern objective techniques have replaced the purely subjective approaches that were used in earlier flavor research. He stated that two approaches have been used for meat flavor research. The first involves the isola- tion and identification of both volatile and non-volatile flavor com- ponents from cooked meat. A second method depends upon the isolation and characterization of flavor precursors from raw meat. He cited some obvious advantages and disadvantages inherent in each of these approaches. He pointed out that one may isolate many compounds from cooked meat, yet have no means of determining the relative importance of each of the isolated constituents to the overall meat flavor. This would entail quantitative determinations of the volatile and non-volatile constitu- ents, and then recombining them at the correct concentration to again achieve the original cooked meat flavor. Doty pointed out that if one isolates and identifies the precursors of meat flavor from raw meat, it is still necessary to elaborate the chemical changes which are re- sponsible for the development of typical meat flavor. Beef Flavor waser (1920) extracted dried beef broth with alcohol and isolated an aromatic fraction from the alcohol insoluble portion, which he then fractionally dialyzed through parchment paper. Twelve percent of the dialysate was assuned to contain inosine, inosinic acid, carnine and similar substances. Analysis of the remaining 88% of the dialysate indicated that it was composed of 47% inorganic and 53% organic consti- tuents. The organic fraction consisted of the following compounds in percentage: taurine or cystine 1.6, ammonia 4.4, creatine 5.4, creati- nine 2.7, hypoxanthine 1.4, carnosine 16.6, methylguanidine 1.3, glutamic acid 7.0, formic acid 1.4, acetic acid 23.9, lactic acid 12.9 and organic phosphorous 2.4. Of the total ash, 10% was KCl and the balance was composed of various phosphates. Feulgen and Voit (1925) histologically demonstrated the existence of plasmal, which was derived from the lipid, plasmologen. They noted a relationship between plasmal and the odor of cooked beef. In early work, Crocker (1948) stated that typical meaty flavors were contained in the fibers of cooked meat rather than in the SXpress- ible fluid. Barylko-Pikielna (1957) reported taste panel evaluation indicated that the typical flavor of roast beef was present in the water- insoluble residue, whereas, the water-soluble fraction had a very intense but atypical beef flavor. More recently, Kramlich and Pearson (1958) found that fluids expressed from raw meat deve10ped a concentrated flavor upon cooking, and that the cooked beef fibers bound the flavor components more tenaciously than the raw fibers. In further studies on beef flavor, Kramlich and Pearson (1960) and Hornstein gt El. (1960) removed the flavor precursors from raw beef with cold water, and identi- fied some of the volatile compounds emanating from cooked beef. The classical report on ox muscle extract by wood and Bender (1957) exemplifies the painstaking investigations on meat flavor. These workers identified more than 30 volatile and non-volatile compounds from ox muscle extract. Melanoidins, which form as a result of interactions between free amino acids and reducing sugars, were noted in meat broth by' Lobanov and WOlfson (1958). By means of two-dimensional paper chromatography, they separated and identified glucose and glucose-6-ph03phate, cystine, histidine, aSpartic acid, glutamic acid, serine, glycine, lysine, arginine, threonine, alanine, proline, tyrosine, methionine, valine, leucine, phenylalanine, carnosine, and anserine. In a model system, they produced a broth-like odor by refluxing a solution containing 1% of a mixture of certain amino acids and 0.2% of a mixture of glucose and glucose-6-phOSphate. The glucose and glucose-6-ph03phate had disappeared upon evaporation to dryness, thus showing evidence of the browning reaction. Browning of the acetone-soluble constituents of ox muscle extract was studied by WOod (1961). Nucleic acid decomposition was an important factor in browning and meaty flavor. He isolated inosinic acid and ribose-S-phOSPhate as the active ingredients. He described the taste of inosinic acid as meaty. By using model systems, the relative effica- cies of ribose, ribose-S-phosphate and glucose with amino acids were studied. Ribose-S-phOSphate was more active in browning than the other reducing sugar moieties. Major emphasis at the American Meat Institute Foundation has been placed on the study of flavor precursors from raw beef. This work was summarized in a review by Doty gt 31, (1961). This approach was selected due to the fortuitous observation that the fat fraction from the third acetone extract of raw ground beef gave a typical broiled steak odor q-10- when heated. Subsequent research on this fraction revealed that the flavor was not due to the fatty material pg; fig, but was caused by sub- stances that diffused through a semi-permeable membrane upon dialysis with water. On separation of this fraction, these workers found a white granular material. This fraction was unstable and became a brown, tarry, mass when Stored under vacuum. Ammonia and/or amines were released as browning progressed, and the deteriorated material assuned a character- istic stale meat odor. Concurrently, WOod (1961) also reported the same or a similar acetone-soluble material from ox muscle extract that decom- posed in an identical manner. Batzer gtflgl, (1960) undertook the formidable task of trying to characterize and identify the components in the water-soluble fraction that yielded the typical meat odor on heating. Upon further fraction- ation of this diffusate by dialysis through a sausage casing and separ- ation of the dialysate on a Sephadex G-25 column, two fractions were obtained. One was a protein fraction, and the other was characterized as the basic meat flavor precursor in beef. By extracting chicken or pork loin tissue and treating the extracts in the same manner as that used for ground beef, they were successful in isolating similar protein fractions. These fractions had almost the Same basic odor as that iso- lated from beef muscle. In a classical series of experiments, the same investigators showed that the basic meat flavor precursor was a glycoprotein, which gave a strongly positive carbohydrate reaction prior to hydrolysis with perchloric acid and a strongly positive phOSphate test after-hydrolysis. Ultra-centrifugation at 60,000 rpm resulted in -11.. no distinct peak. Thus, the compound was established as being of rela- tively low molecular weight. Ninhydrin tests were positive after 8 Spots had been separated by paper chromatography of the acid hydrolysate. Two of the ninhydrin positive Spots could not be positively identified, but the others revealed the presence of leucine, proline, isoleucine, alpha alanine, valine, serine and beta alanine, with trace amounts of glycine and glutamic acid. When the diffusate fraction from the secondary dialy- sate was separated on Dowex-50 ion-exchange resin and eluted with acid, a spectra of nucleotide peaks matching published results for hypoxanthine,‘ inosinic acid and inosine were obtained. Some further studies on the identification of beef flavor precursors by Batzer 25 51, (1962) resulted in the identification of inosinic acid, inosine and a glycoprotein with a glucose moiety. It was concluded that these are simple, water-soluble components of beef muscle tissue. Batzer and Landmann (1963) stated that the consistent isolation of the glchprotein from various runs was a function of the efficacy of the sausage casing which was used for dialysis. 01d casings were ineffective and fractions obtained failed to exhibit the characteristic broiled steak odor when the standard test was applied. Furthermore, they were unsuccessful in identifying the eighth amino acid of the peptide group in the glycoprotein moiety. When a mixture of the glyc0protein, inosine or inosinic acid, inorganic phOSphate and glucose was heated in fat, an odor similar to the original broiled Steak odor was obtained. It is interesting that the phOSphorous-containing nucleic acid fraction isolated in these studies was previously reported as present in ox muscle extract -12- by Wbod (1956). The glycoprotein fraction was not reported present in beef extract and Doty §£_§l, (1961) concluded that this was understand- able, since the glchprotein undoubtedly decomposed upon heating. 'With reference to precursors, first Crocker (1948) and then Kram- lich and Pearson (1958) and Hornstein gghal, (1960) demonstrated that flavor precursors can be leached from raw beef with cold water, although none of these workers attempted to establish their identity. In a recent report on the flavor of beef and whale meat, Hornstein ggugl. (1963) made a chemical comparison between freeze-dried extracts of lean whale and beef tissue and showed the volatile fractions were identical, there- by confirming their earlier conclusions. In working with beef and pork (1960a, b) and lamb (1963), the same authors found an identical meaty aroma associated with the lean portion of all red meats, whereas, flavor differences resided in the fat. A paper by Harries ggmgl. (1963) described the evolution of their methods for assessing beef quality which have become standard since 1955. Organoleptic techniques were perfected by rigidly standardizing cooking and judging procedures. Judges were chosen for their flavor evaluating acuity. All samples were coded at random, and no more than four samples were judged at any one session. Only one session was held per day. Bread and water were used as palate cleansers and tasters were encouraged to take their time and retaste as often as necessary. Hot and cold tasting results were compared. -13- Pork and Mutton Flavor Only a few studies have been reported on pork and mutton flavor volatiles. Hornstein and Crowe compared the volatiles from lean pork, beef and lamb in three separate studies (1960a, b, 1963). Common com- ponents of the volatile fraction from these three different meats were carbon dioxide, formaldehyde, acetaldehyde, acetone, ammonia and hydro- gen sulfide. The same investigators also theorized that fundamental flavor differences between meat reside in lipid-soluble foreign com- pounds, and the capability of a Specific fat to produce different flavor components in different ratios. This was in contrast to the report of Cracker (1948), who suggested that bones contribute little to meat fla- vor while marrow and tissue fats supply aroma, but add nothing to flavor. Observations on the objectionable reaction of some individuals to mutton flavor were made by Ziegler (1958) and Kean (1959), and gave impetus to research on the chemical constituents in mutton flavor. IMutton flavor studies by McInnes g£_§l, (1956) indicated that a group of steam volatile fatty acids occurred in mutton fat, which may be important in mutton flavor. They found a series of normal acids from formic to capric (01 to Clo) together with isobutyric, isovaleric, and alpha methyl butyric acid. Hofstrand ££L§£9 (1960) proved that lamb broths lost their identifying aromas and the fat odor was volatilized by the application of heat. Jacobson gtflgl, (1962) characterized mutton flavor as "fragrant," "oily," "sweefl'and somewhat "musty." By comparing the volatiles from cooked lamb with chemical odors, they found Similarities to ethyl oleate, -14- diacetyl and a number of sulfur compounds. An extensive study on lamb flavor was reported by Jacobson‘ggugl, (1963). They found glucose, fructose, inositol and 19 amino constituents in water extracts from raw and cooked lamb. Aroma was strongly affected by the numerous carbonyls present, including several alkanals, alkanones and possibly 2-methyl- cyclOpentanone. ‘Williams (1962) obtained a patent on a process for improving the tenderness and flavor of mutton. Freshly slaughtered old crop (yearling) lamb and mutton were pumped by the stitch method with a solution con- taining 0.4-0.6 oz. monosodium glutamate (MSG/lb) at a dosage level of 3% by weight of the carcass. The carcass temperatures were raised to 108-115°F by microwave heating or in a steam room. He claimed that after chilling the carcass, the meat had the flavor and tenderness of Spring lamb. Fish Flavor A voluminous literature on the chemistry of fish has been compiled by Borgstrom (1961). However, no attempt was made to correlate the chemical data with flavor. Jacquot's (1961) contribution on the organic constituents of fish and other aquatic animal foods provides an insight into some similarities and differences between fish and mammalian muscle. He reported that seafoods contain a series of Specific basic proteins. These protamines are indigenous to a certain Species and differ from one another due to the nature, content and position of the amino acids. Salmine from salmon, iridine from rainbow trout, fontinine from brook -15- trout and clupeine from carp all have polypeptide chains ending in pro- line. Sturine from sturgeon has alanine and glutamic acid at the ter- minal end of the protamine molecule. Fat composition varies between ocean and fresh water fish according to Lovern (1950). In fresh water fish, 20% of the unsaturated acids were C10, 40% 618, 13% C20 and 2.5% 022, whereas, in ocean fishes 10% of the unsaturated acids are Clo, 25% C18, 25% 020 and 15% C22. 'Wide variations were found in the composition of lipids due to Species, diet, temperature and salinity of the environment, selective mobilization and selective distribution. Lovern stated that fish lipids contained little:0r no linoleic acid. Palmitic acid was the predominant saturated fatty acid in fish oil and constituted 10-18%. Myristic and stearic acids occurred in lesser quantities, with the latter rarely exceeding 2%. Lovern also concluded that the three alcohol sites on the glyceride mole- cule could be esterified, resulting in the same or different fatty acids. Lovern (1950) also reported that the most common unsaturated fatty acids encountered in fish oils could be classified as monoenoic acids and polyenoic acids. The principal monoenoic acids reported were pal- mitoleic with C16, gadoleic with C20, cetoleic with 622 and selacoleic acid with C24. The latter acid is abundant in Shark oil but has not been encountered so far among the telosteans. The principal polyenoic acids found were clupanodonic with 022, which is the most abundant and contains five double bonds, arachidonic with 020 and four double bonds, hiragonic with 016 and three double bonds, and minisic and thynnic acids with 024, whose compositions are so far not well understood despite numerous studies. -l6- According to Tarr (1950) and Jones (1958) glucose and ribose are the principal free sugars in fish. Original evidence for sugar phos- phates was obtained by Tarr (1950), who used a non-specific barium frac- tionation procedure. Recently, Jones and Burt (1960) used an ion-exchange column and separated fructose-l-phOSphate, fructose-6-ph03phate, ribose- l-phOSphate and ribose-S-phOSphate. These workers characterized the flavor as being "sweetiSh-Salty" at pH 6.8. Amino acids and peptides differed between fatty and non-fatty fish according to Lukton and Olcott (1958). Fatty Species contained consider- able histidine in the flesh, whereas, non-fatty species contained anser- ine (beta alanyl-1-methylhistidine). Jones (1961) stated that anserine and taurine have pronounced pleasant effectson the tongue, which were described as "mouthfulness" by the panel. ASpartic and glutamic acids were detected at low concentrations and described as "acid." The other amino acids were characterized as "sweetish," with proline being rated as the sweetest. He also found that a combined amino acid solution simulating non-fatty fish muscle had a "pleasant, sweet-sour, meaty, yeasty flavor." Tarr (1958) reviewed the biochemistry of fishes. This article in- cluded sections on proteins, non-protein nitrogen, nucleic acids and related compounds, phOSpholipids, enzymes, sterols and the nutritional value of marine invertebrates. A total of 265 references are included in this comprehensive review. Extensive investigations have been made on nucleotides and nucelo- tide changes in fish muscle by Tarr and his associates (1957, 1958, 1959, -17- 1960 1962), Fujimaki and Kojo (1953), Shewan and Jones (1957), Jones and Murray (1957) Saito and Arai (1958), Tomlinson and Creelman (1960), Tomlinson et a1. (1961) and Olley (1961). Jones (1961) reported a high inosine-5'dmonophosphate content in fresh fish and showed that it was one of the major flavoring components. It was des- cribed as possessing a strong salty-acid flavor, with overtones described " "almonds." as "meat extract, " "yeasty, An electrodialysis method was devised by Yoshida and Kogeyama (1956) for separating inosinic acid from dried fish muscle. A contemporary patent granted to Motozaki 35 El- (1963) described the preparation of inosine from mutants of Bacillus subtilis, which require adenine and an amino acid for growth in an aerobic media containing carbon and nitrogen sources. Such nutrients can be yeasts, or yeast derivatives, meat ex- tract or corn steep liquor. The pH must be maintained between 4 and 9 preferably between 5.0 and 6.7 during fermentation. It is interesting to note that the first commercial production of inosinic acid was from fish muscle by the Japanese. Saito and Arai (1957) reported that noteworthy changes occurred in carp muscle when it was held at low temperatures. At 16°C, ATP and ADP broke down to give inosine monOphOSphate (IMP) as the major product, while in liquid air no change took place. According to Jones (1961), inosinic acid broke down to form.hypoxanthine after l--12 days storage of fresh fish on ice. Hypoxanthine was produced and gave bitter flavor, which was indicative of breakdown. -13- Several important chemical factors are responsible for undesirable flavors in fish. *Tanikawa(l959) traced sourness in stored fish to an increase in free fatty acids. The sour taste was noticed after 90 hours at 10-13°C in dry Storage or after 21 days on ice. Spoiled flavors in fish (Hughes 1960) resulted when a reducing atmosphere prevailed, thus causing formaldehyde, dimethylamine and trimethylamine to be formed and volatilized. Hughes (1961) also reported that the combined creatine- creatinine fraction plus histidine were reduced during heating, and than large amounts of ammonia were liberated. Tanikawa (1958) measured the amount of increase in some objectionable amines in fatty fish. He found that as cold Storage time increased so did the content of hista- mine, cadaverine, putrescine, agmatine and iso-amylamine. Reay and Shewan (1949) reported that hydrogen sulfide increased in direct prOpor- tion to the time the fish were chill-stored. Methyl mercaptan was found in aged samples. In working with fish oils, Lundberg (1957) attributed the oxidized flavor and odor accompanying deterioration, to an increase in the un- saturated carbonyl and dicarbonyl compounds. Privett g; 31, (1958) suggested that malonaldehyde is the active carbonyl compound in the thiobarbituric acid test (TBA test) for oxidative rancidity. Mangan (l959a,b) demonstrated the presence of acetaldehyde, methanol, ethanol and probable dicarbonyl and alpha hydroxy-compounds in frozen haddock, and suggested that they may be important flavor and odor com- ponents. The production of carbonyls in heat-processed herring has been reported by Hughes (1961, 1963). He correlated the degree of freshness -19- of iced herring prior to canning with total carbonyl production and the classes of carbonyls that were produced. He isolated five major car- bonyls, namely; acetaldehyde, propionaldehyde acetone, iso-butyraldehyde and 2~methylbutyra1dehyde from the cooked flesh of herring. These were the main components of the neutral volatile fraction with smaller quanti- ties of longer-chain carbonyl compounds. Jones (1961) characterized the most acceptable raw and cooked fish odor as "fresh seaweedy." The most undesirable odors were described as "sulfide," "ammoniacal" (due to amines) and "indole." Sweet flavors were found in fresh fish and flavors ranging from "bitter," "rubber- like" and "sulfide-like" to ”nauseating and putrid" occurred as spoilage ensued. He described some intermediate odors as "sour" and "condensed milk" and some intermediate flavors as "chewing cotton wool" or "sourness with no bitterness." Undesirable Meat Flavor The influence of browning in foods may be desirable if controlled, but is undesirable if excessive. Maillard (1912) first explained the fundamental browning mechanism as an interaction between the carbonyl group of reducing sugars and the amino group of amino acids, peptides, polypeptides or proteins. This is known as nonenzymatic browning. Lea and Hannan (1950a,b) reported on the reaction between proteins and reduc- ing sugars in the dry state, and on the biochemical and nutritional significance of browning. In 1952, Lea and Schwartz reported further Studies on the mechanism of browning in the dry State. Speck (1952) -20- showed that primary and secondary amines catalyze dealdolization and conversion of Six carbon sugars to pyruvaldehyde and diacetyl. Lewin (1957) investigated the effect of initial pH on the behavior of glu- cose and amino acid-glucose interactions. Subsequently, he studied the reactions of amino and imino compounds with reducing sugars, including the reaction of histidine with glucose. Later Lewin.and McCall (1957) reported on the interaction of glycylglycine with glucose. A classic study on the Strecker degradation of amino acids was made by Schgnberg‘ggngl, (1952). Previously these workers (1948) showed that degradation of alpha amino acids to aldehydes and ketones resulted from interactions between these amino acids and certain carbonyl compounds. Alpha amino acids are degraded to the correSponding aldehydes or ketones containing one less carbon atom by the action of Specific carbonyls. Degradations were carried out in water or water-glycerol media by hold- ing the mixture at the boiling point for 15 minutes in a 002 atmosphere. Active carbonyls were the C0:C0 type such as glyoxal, methyl glyoxal, pyruvic acid, diacetyl (acetoin), ketoglutardialdehyde, alloxan, phenyl pyruvic acid and others. Various mechanisms were postulated, but eXper- imental evidence favored Strecker degradation with the accompanying loss of 002 and NH3 in an oxygen atmOSphere. Schiff base intermediates were formed between the amino acid and the active carbonyl. Degradation products of the amino acids were identified by their 2,4-dinitrophenyl- hydrazones (2,4-DNPHS). Akazawa and Conn (1958) obtained little experimental evidence supporting the pyridoxal cleavage theory of amino acid degradation. -21- This mechanism requires the oxidation of reduced pyridine nucleotide by peroxidase in the presence of catalytic amounts of'Mn++ and certain phenols. The active phenols were found to be monohydric phenols or resorcinol. The reaction was inhibited by catalase, certain phenols, cysteine, cupric ions and inhibitors of heavy metal enzymes. It was shown by'Mazelis (1962) in his studies on the potential precursors of volatile sulfur compounds that highly purified horseradish peroxidase will not oxidatively decarboxylate methionine unless both pyridoxal phosphate and Mn++ are present. Phenols were found to be catalysts of this reaction. In a separate Study (Mazelis 3511: 1962) methylthiOprOpionamide was identified as a product of decarboxylation of‘methionine. Glucosone is regarded as an important intermediate of browning. Kato (1960, 1962, 1963) has presented data on its formation from N- glucoside and proposed a mechanism for the reaction. D-glucosone was isolated from the browning degradation mixture of N-D-glucosyl-n-butyl- amine in methanol solution neutralized with acetic acid. It was iden- tified by its 2,4-dinitr0phenylhydrazone derivative. The author also found 3-deoxy-D-g1ucosone and Showed it to be an intermediate substance in the browning reaction. The yield of N-D-glucosone was far less than that of 3-deoxy-D-glucosone, but increased considerable when air was bubbled through the reaction mixture. This indicated that its formation is dependent on the amount of dissolved oxygen. Burton 25 21, (1963) showed that non-enzymatic browning of phenolic compounds in the presence of nitrogen occurs more rapidly than non- -22- nitrogenous browning, regardless of whether in the solid state or in, weakly acid solution. Sulfite was effective in retarding the onset and development of such browning. They stated that this is clearly an example of the browning associated with carbonyl and amino groupings, which sulfites will prevent in some foodstuffs. Both carbonyl groups and reducing agents also effectively retarded this particular browning reaction. Lobanov and WOlfson (1958) have studied browning of meat broth while Wbod (1961) studied browning of meat extract. According to Wbod, reducing substances in fresh muscle include the reduced form of diphos- phOpyridine nucleotide (DPNH), sugar phosPhateS and free reducing sugars. He stated that the Maillard reaction involves Splitting of ortho-phOSphate in aged muscle, by which free sugars, principally ribose, are released to react and impart the brown color and meat flavor characteristic of ox muscle extract. WOod used model systems, in which various amino acids were reacted with glucose and glucose-G-phosphate, or ribose and ribose-S-phOSphate to produce effects similar to those observed in foods. Recently, Ballance (1961) showed that Strecker degradation of methionine by ninhydrin resulted in formation of methional, methyl mer- captan, dimethyl sulfide, isobutyraldehyde and acrolein. Schgnberg 35 ‘31. (1948, 1952) demonstrated that many compounds apart from ninhydrin will decompose amino acids on heating. This included several that are indigenous to living tissues. ‘Off-flavors in meat caused by irradiation have been studied by numerous workers. Papers on irradiated meat flavor have been published -23- by Proctor £3 31, (1955), Batzer and Doty (1955), Groninger g£_§l, (1956), Marbach and Doty (1956), Witting and Batzer (1957), Burke £3 31, (1957), Cavallini Egg. (1959), Pearson £13.21: (1959), Merritt £31. (1959), and Hedin gt_gl. (1960). According to Pearson g£_gi, (1959), irradiation flavor was associated with high levels of hydrogen sulfide, methyl mer- captan and carbonyls. Their studies were made on irradiated pork, beef and veal by both chemical and organoleptic evaluation procedures. Oxidative rancidity of adipose and lean tissues of meat was reviewed by watts (1961). She reported that lean muscle tissues of land animals or fin fish tended to oxidize very rapidly following heat denaturation of heme proteins. The hematin compound was found to be the active cata- lyst in this oxidation. In heat-processed meats, the oxidation occurred only to a limited extent due to the development of antioxidants during the severe heat treatment. Watts stated that a promising approach to the elimination of lipid oxidation is through the use of water SOldble antioxidants which can be incorporated in raw meat or applied to the sur- faces of the cooked meat. Chicken Flavor According to Lineweaver and Pippen (1961), the ten-fold increase in per capita consumption of broilers since 1947 constituted strong evidence that broilers as presently produced, processed and marketed have highly acceptable flavor. They also cited the fact that the low cost of poul- try compared with other meat has been an important factor in acceptance. -24- However, there is still a lack of scientific knowledge on the effects of modern production upon chicken flavor. Lineweaver mentioned the laxity of the poultry industry in failing to rate the comparative effi- cacy of modern broilers and old-fashioned roasting chicken. He stated that on the basis of current experimental evidence it is difficult to determine any flavor difference among various types of birds, and then concluded that differences are ordinarily negligible. Studies on the chemical nature of chicken flavor began when Liebig extracted various meat, including hen meat in 1847. He quantitatively determined creatine, creatinine, lactic acid, inorganic salts and other constituents of lean chicken muscle. He then isolated and identified inosinic acid in extracts from chicken. Osborne and Heyl (1908) hydrolyzed the edible parts of mature hens, which had been bled and freed from fat and connective tissue. After hydrolysis the amino acids were determined to be present in the follow- ing percentages: glycine 0.68, alanine 2.28, leucine 11.29, proline 4.74, phenylalanine 3.63, aspartic acid 3.21, glutamic acid 16.48, tyro- sine 2.16, arginine 6.50, histidine 2.47, lysine 7.24 and ammonia 1.67, thus comprising a total of 62.75%. Trypt0phane was present. Cystine and hydroxyproline were not determined. Valine and serine were not identified in the hydrolysate. It was concluded that except for the higher lysine content, the protein of hen muscle is similar to the protein of legunin- ous seeds. Houghton (1911) studied the effects of low temperature storage on ground meat. After five months at 35°F, ammonia increased slightly in -25- light meat. There was a large increase in the water-soluble nitrogen content of light meat and a slight increase in dark meat. There was an increase in free amino acids and proteoses, and a decrease in peptones. Insoluble phOSphorous increased during the first 90 days storage, eSpec- ially in dark meat. There was also an increase in the iodine number of the fat. Five different lipolytic enzymes were detected in the stored fat and muscle fractions. Sadikov 25 a1. (1934) ascribed an important role to glutathione as a flavor precursor in chicken. These workers found that hydrogen sulfide_ was formed on cooking chicken muscle. The hydrogen sulfide appeared to be due to the total decomposition of glutathione and the partial decom- position of cystine and methionine. They concluded that part of the hydrogen sulfide was reabsorbed by the protein, but that larger amounts were lost by formation of free sulfur. Modern studies on the chemical nature of chicken flavor in the United States probably began with a classic distillation of tissues from chicken, pork and beef by Crocker (1948). Some of the same simple com- pounds, namely; hydrogen sulfide, ammonia and acetaldehyde were found in each of the three distillates. Crocker concluded that all meats may possess identical fundamental flavor factors, and that individual Species differences may be due to low concentrations of Specific compounds char- acteristic of the particular Species. Bouthilet (1951) replaced Crocker's simple distillation procedure by a high vacuum distillation technique, and used it in fractionating chicken broth. Hydrogen sulfide and ammonia were found in the distillates. Upon completion of numerous tests using -26- chicken broth. Bouthilet (1949, 1950, l951a,b) postulated that gluta- thione is the major muscle precursor of chicken flavor. He further concluded that fat affects the aroma of the broth. Distillation studies were made on the aqueous phase of chicken broth by Pippen and Eyring (1957). These workers also found hydrogen sulfide and ammonia in the distillate and demonstrated that removal of the ammonia from the distillate resulted in enhancement of chicken fla- vor. Thus Bouthilet's (1949) conclusion that a progressive lowering of pH raised the chicken flavor level in the broth distillate was confinmed.. Pippen and Eyring also confinned Bouthilet's work showing that upon Standing, desulfuration of the broth continued as long as true chicken flavor existed. Results led both groups of workers to conclude that sul- fides are important in chicken flavor. From these experiments, Pippen and Eyring also concluded that chicken flavor is associated with the neutral or acidic constituents. A concentrated chicken flavor precursor was prepared by Peterson (1957) using a water extract of ly0philized lipid-free muscle. He con- cluded that fat does not contribute to the taste of chicken broth but contributes to its aroma. A flavorful broth was made by heating a water solution of the chicken flavor precursor. These results confinmed the earlier findings of Pippen ggugl. (1954). A classical study on the carbonyl compounds in the volatile frac- tion of cooked chicken was published in two parts by Pippen gtflgl, (1958, 1960). Many carbonyl compounds were isolated, separated and identified as their 2,4-dinitr0phenylhydrazone derivatives (2,4-DNPHS). Elegant -27- separations were made by column and paper chromatography, and identifi- cations of the carbonyls were substantiated by melting point comparisons with known compounds, as well as by infrared Spectroscopy. When suffi- ciently large fractions of derivative were available to permit carbon, hydrogen and nitrogen analyses, empirical formulas for the 2,4-DNPHS were determined. These same investigators showed that the cooking proce- dure and isolation method had a pronounced influence on the yield of carbonyls. Normal cooking consisted of heating an equal weight of water and raw cut-up chicken at 100°C for 4 hours in a vessel arranged for distillation. Cooking in an oxidative atmOSphere consisted of refluxing the chickendwater mixture for 16 hours. The carbonyls were carried past the low-efficiency condenser with an air stream passing through the trap containing 2,4-dinitr0phenylhydrazine solution (2,4-DNP) at a rate of two bubbles per second. Carbonyl compounds were condensed along with the water from the cooking mixture. Cooking in an inert atmOSphere con- sisted of passing nitrogen through the system instead of air. Pippen £5 a1. (1958, 1960) found that the yield of volatile carbonyls was influenced markedly by the cooking method employed. Large yields resulted from cooking the chicken in an oxidative atmosphere. Carbonyl production was 30 times greater when an oxidative atmOSphere was used than when a nitrogen atmosphere was utilized. Normal cooking gave an intermediate yield of carbonyl derivatives. By using the oxidative at- ‘mosphere for cooking, the yields of 2,4-DNPHS was so large that the recognition of minor carbonyl compounds was facilitated. Diacetyl, acetaldehyde, n-hexanal, n-hept-2-ena1 and n-deca-2,4-diena1 were present -23- in the largest amounts. Many other carbonyls were identified including n-butanal, n-pentanal, n-heptanal, n-octanal, n-nonanal, acetone, pro- panal, methyl-ethyl-ketone, 2-pentanone, 2-heptanone, n-pent-Z-enal, n-butenal, n-hepta-2,4-diena1, n-hex-Z-enal, n-dec-Z-enal, n-oct-Z-enal, and n-non-Z-enal. A total of 7.8 g. of 2,4-DNPH precipitate was obtained from 31.4 kg. of chicken. Long chain aldehydes from C16 to 013 that were found in the normal cooking did not appear in the volatile fraction after air entrainment. N-hept-Z-enal, which was one of the most abundant carbonyl compounds present in the air entrainment volatile fraction, was found only as a minor component after the normal cooking procedure. 'Lineweaver and Pippen (1961) stated that specific chemical knowledge concerning the precursors of carbonyls is incomplete or non-existent. Pippen g£_§1, (1960) determined that the concentration of acetoin exceeded that of diacetyl by a factor of 18 in cooked aqueous extracts of chicken meat. It was concluded that acetoin is the immediate precursor of some if not all of the diacetyl. By studying the effects of heating time and temperature on the concentration of acetoin in aqueous extracts of raw chicken meat, they found that heat labile precursors of acetoin were preSent, These authors concluded that the remaining carbonyls may ori- ginate from either the triglyceride or phOSpholipid fractions. Rapport (1959) has demonstrated that the plasmologens or acetal-phospholipids can yield long chain aldehydes. Lineweaver and Pippen (1961) concluded that this phospholipid fraction may be the precursor of 016 to C18 alde- hydes in chicken meat, but is not the precursor of decadienal. -29- Experimental evidence for the importance of carbonyls to chicken flavor is not yet conclusive. According to Lineweaver and Pippen (1961), carbonyls represent some of the end products of chemical reactions that occur during cooking, and probably have a definite role in the flavor of cooked chicken. Pippen and Nonaka (1960) estimated that the average carbonyl concentration in chicken broth samples was approximately 14 x 10"5 moles/liter. Lea and Swoboda (1958) determined that n-decanal could be tasted at a level of 5 x 10"8 moles/liter. According to Patton Iggugl, (1959) decadienal was detected at concentrations as low as 0.5 part per billion in water. Lineweaver and Pippen (1961) found that decadienal had a deep-fat-fried odor, that may contribute to the flavor of foods where fat mingles with moisture at elevated temperatures. Thus appears to be a staling or rancidity factor. Limited tests were made by Pippen SE a1. (1960) to ascertain whether diacetyl and acetoin contribute to the flavor of chicken broth. They concluded that normal concentrations of acetoin and diacetyl in chicken broth cannot be detected. However, if substantial amounts of acetoin was oxidized to diacetyl, its presence could easily be detected. They postulated that diacetyl contributes a transient buttery-oily type aroma in freshly cooked chicken. The conclusions reached by Lineweaver and Pippen (1961) after re- viewing the available knowledge on chicken flavor indicated that flavor is fairly independent of the type of bird. Meat was the most important flavor component in the carcass. The chemical composition of the vola- tile fraction from cooked poultry included sulfide and carbonyl compounds, -30- which are probably of importance to flavor. .Although diacetyl/acetoin may contribute to flavor, further work is needed to confirm the impor- tance of carbonyls to chicken flavor. Kazeniac (1961) reported on a project aimed at isolating and identi- fying the components and precursors of chicken flavor, establishing the Optimum conditions necessary for flavor develOpment, developing process- ing procedures that will maintain high levels of flavor, discovering suitable ways for preventing flavor deterioration during storage and raising chickens high in flavorful components and precursors. He postu- lated that fat may be a trapping agent for flavor volatiles, Since broth with some fat gave a more desirable flavor than broth with the fat removed. Diacetyl values were found to be higher for cooked skins than for raw skins or dark meat, and for light meat than dark.meat. Light meat broth had stronger taste and mouth Satisfaction, which was attributed to a higher inosinic acid content, while dark meat had more body and a stronger aroma. Taste in chicken broth was attributed to the various classes of compounds, including a mixture of amino acids, peptides, carbohydrates, inorganic salts, sulfides, carbonyls and nonramino nitrogen compounds, such as ammonium sulfide, creatine/creatinine, carnitine, hypoxanthine, inosine and inosinic acid. Addition of either glutathione or homocy- steinethiolactone improved broth flavor. Other sulfur compounds such as cysteine, cystine, and homocysteine increased the broth sulfide content after heat-processing, but also gave objectionable off-flavors. Kazeniac concluded that sulfide released must be considered as a direct contri- butor, as well as a possible indirect indicator of flavor. He found that -31- the sharp aldehydic flavor of light meat distillates became similar to the flavor of dark meat distillates by the addition of ammonia, and stated that the sulfide/ammonia relationship has apparently been over- looked in flavor problems. Ammoniun sulfide contributed a sweet taste to broth and improved the aroma. Kazeniac (1961) stated that acetoin gave a desirable buttery, oily- type flavor with improved body, but when levels were too high, the characteristic sour notes of diacetyl made the flavor undesirable. Acetaldehyde gave a scorched flavor note, which is common to chicken broth flavor profiles. Carbonyls derived from chicken Skins contributed a very bitter off-flavor, whereas, those higher aldehydes derived from muscle added to the aroma and body, as characterized by oily-fatty notes in the broth. Creatine/creatinine content was higher in light than in dark meat, and was related by Kazeniac to the bitter after-taste more often found in light meat than in dark meat broths. Carnitine enriched broths developed strong fishy aromas and showed intense browning. Hypo- xanthine and inosine imparted a bitter taste, whereas, inosinic acid made a major contribution to mouth satisfaction and intensified the effects of other compounds. Collagen and lipids gave more body to the flavor of chicken broth. Kazeniac showed that when certain amino acids including lysine, arginine, alpha alanine, glutamic or aSpartic acid 'were added to chicken broth the overall flavor was improved. Glutamic acid gave greatest mouth satisfaction at levels of 0.02-0.04%. Lysine ‘with glutamic acid gave the highest intensity of mouth satisfaction when levels of glutamic acid were 0.02-0.05% and lysine between 0.05 and 0.08%. -32- Alpha alanine imparted a sweet taste to broth together with some mouth satisfaction. Lactic acid contributed to the sour, astringent taste in broths and improved mouth satisfaction slightly at levels of 0.02-0.04% lactic acid and 0.06-0.08% lysine or arginine. Kazeniac (1961) further reported that glucose, fructose and ribose were the principal sugars present in chicken broths, and that inositol was suSpected. Dialysates of the raw meat extracts showed appreciable amounts of ribose, but upon boiling the extracts turned brown with noticeable losses of ribose. Results were similar to the effect found by Wbod (1961) in ox muscle extract and Doty 25 El. (1961) using an ace- tone extract from raw beef. In Kazeniac's work he found that sulfhydryl compounds such as glutathione decreased browning and improved broth flavor. He reported that the salty taste in broth was derived from inorganic salts and salts of the amino acids. He also found that the alkaline ash of chicken meat improved broth flavor Slightly, whereas, Pippen and Klose (1955) found that the neutral ash of water extracts of chicken produced a Similar effect. Kazeniac (1961) attributed the improvement in flavor to the added inorganic phOSphates and to the alkalinity of the ash, which raised the pH and released ammonium sulfide. Ammonia in the presence of hydrogen sulfide gave the broth a sweet aroma and taste, provided the con- centration of ammonia was relatively high. Kazeniac concluded that chicken flavor is a complex blend of different compounds, and that precursors 'hold more promise for improvement in chicken flavor than the volatile flavor fraction. He planned in future research to use radioactive sulfur compounds to study the reactions of sulfur containing compounds. -33- Glucose and fructose were Shown to be the principal free sugars in chicken muscle by Lilyblade and Peterson (1962). Red muscle contained more than twice as much free inositol as white. Inositol, fructose and ribose increased during storage in both kinds of muscle in both older and younger birds. Carnosine levels were found by Davey (1957) to be 1.04 g/100 g. in the breast and 0.18 g./100 g. in the leg muscle of the hen, whereas, anserine levels were 0.58 g./100 g. and 0.28 g./100 g., respectively. The free amino acid level in the cock was reported to be approximately 0.22 g./100 g. by Florkin (1957). All of these values were obtained on lean somatic muscle. PhOSphatide components in fowl were reported by Rates and James (1961). These workers found phosphatidyl ethanolamine in the Sphingo- myelin fraction together with lysophosphatidyl ethanolamine and small amounts of phOSphatidyl inositol, phOSphatidyl serine, lysolecithin and phOSphatidic acid. Each had a different fatty acid composition. The phosphatidyl ethanolamine fraction had a higher prOportion of polyunsat- urated fatty acids, and stearic acid had a lower proportion of palmitic acid than lecithin. Both phOSphatides were also present in plasmologen as the palmitaldehyde and stearaldehyde moieties. Effects of Feed on Flavor Early ration studies by Maw (1935) suggested that fat replacement of moisture in the bird is due to ration, and that fat carries the fla- vor. Thus, a higher fat content in tissues would mean more flavor. -34- Wheat-fed birds had dry tissue, whereas, the tissue of corn-fed birds was moist, soft in texture, best in flavor and had the highest nutri- tional value. In single grain rations after corn, barley ranked second, oats third and wheat last. Broilers fed single grain rations, such as wheat and oats or wheat and barley, were found by Odland gt 51. (1955) to be equal or superior to those raised on more complex rations of barley, wheat, bran Shorts, wheat bran and/or corn. Fish solubles pro- duced no off-flavors, but did not improve palatability scores. When levels of 8% animal fat were fed to broilers for 10 weeks by Essary (1961), the tissues contained more fat than birds raised on a standard commercial diet. In separate studies reported by Marble gtflgl. (1938), Garrick and Range (1926) and Asmundson gtugl. (1938) it was found that fish meals or fish oils imparted a fishy flavor to chicken meat. A study on the effect of milk products on broilers by weisberg (1956) indicated that addition of 1.5-5.0% of dry milk solids to the rations improved the flavor of the meat. Brant ggflal, (1958) found the ration to have no effect when compar- ing the flavor of modern-type broilers with that of those raised on older type rations. In a separate study, Hanson gtflal. (1959) compared the 1930 (old-type) diet with the 1956 (new type). No Significant flavor difference was found to be associated with diet with either equal age or equal weight birds. By feeding a low-fat purified diet and a standard diet as a control from the chick stage to an age of 8 to 10 weeks, Lewis gtflgl. (1956) demonstrated that birds raised on a standard diet had more intense flavor in the broth and in both light and dark.meat. -35- An attempt was made by Newman £5 31. (1958) to season poultry meat by feeding flavor ingredients to eight week old White Rock broilers for two weeks prior to slaughter. Garlic, celery seed, allspice, cloves or monosodium glutamate were added at a level of 3 to 5 ozs./100 lb. feed. Taste panels rated the garlic-flavored birds the least acceptable, while the monosodium glutamate fed birds were Slightly more acceptable than the controls. Spice-fed birds were rated about the same as the controls. No flavor differences occurred when birds were fed either semi- synthetic or standard commercial rations according to Kahlenberg gtflgl. (1960). Oxidized oils, such as those reclaimed after potato-chip manufacture, caused an edemous condition in chickens when included in standard rations according to Brew gt a1. (1950). A high chlorine content occurred in these birds, especially in the pectoral muscle, but was liberated during cooking. Both the broth and meat from the birds fed oxidized oil had an off-flavor. The effect was attributed to a compound contained in the unsaponifiable fat fraction of the toxic fat. A concentrate was obtained by column chromatography that was 3200 times as toxic as that contained in the original fat. The molecular weight was determined by a mass Spectrograph, and the UV Spectrum confirmed the compound as a benzene substituted material or a product of cholesterol. Patrik (1962) reported better growth resulted when 2% of phOSpha- tides from vegetable oils were added to a chick ration. The meat from phosphatide-fed birds graded higher, and a higher yield of edible meat with lower muscle water content was obtained. -36- How Cooking;Method Affects Flavor Simering or pressure-cooking was recommended by Hanson 313-1.];- (1950) to increase tenderness in cooked poultry. They reported that roasting had no advantage over Simmering or pressure-cooking in produc- ing typical "roast turkey flavor." Furthermore, roasting had the disad- vantage of accelerating rancidity development. A study of the relation- ship of cooking method, grades and frozen storage to quality of cooked ‘mature Leghorn hens was reported by Swickard ££.§i3 (1954). They found that the meat of steamed hens rated higher than that obtained by pressuree cooking. Conventionally-cooked lO-week old fryers were compared with the similar electronically-cooked birds by Schano and Davidson (1958). Electronically-cooked chicken was inferior to roasted or rotisserie- cooked birds, although the yields by the former method were higher. Old fowl was cooked by boiling, Simmering and pressure-cooking in experiments by Kahlenberg and Funk (1960). Simmering gave higher yields, and pressure- cooking more tender meat, but with a lowered fat content. A comparison of the effects of various cooking methods on old fowl with and without the incorporation of Salt in water was made by these workers. Cooking losses, degree of tenderness in the breast meat and amount of fat in the thigh meat was determined. Cooking in salt solutions had no advantage over cooking in water. Significantly lower non-fat cooking losses were obtained by Simmering than by boiling. Pressure-cooking resulted in no significant change in non-fat cooking losses as compared to Simmering. Pressure-cooking increased tenderness of breast meat, but the fat content -37- of the thigh meat was reduced by pressure-cooking as compared to boiling or simmering. A statistical treatment of palatability scores showed that cooked dark meat was significantly more flavorful, juicier and more tender than light meat from the same bird. Effects gf_Processing and Storage In a study made on cold storage of chickens (1907), Pennington reported that after 10 months storage the birds degenerated as evidenced by microscOpic appearance and taste. subsequently Pennington and Hep- burn (1913) reported the presence of catalases, oxidases and lipases in chicken fat, and showed that lipolysis occurred even in fat that was frozen solid. Storage time ranged from 3 days at 40°F to 151 days at -9.4°F. Stewart gt 31. (1945) showed that quick-frozen broilers lost flavor during 51 days storage at -10°F. In a recent review, Brant (1963) expressed concern over the possibility that the broiler and fryer indus- tries may decide to market their products frozen rather than fresh. He pointed out that 85% of all turkeys sold in the United States are ‘marketed frozen, whereas, an almost equal percentage of broilers and fryers are sold fresh. Consumers prefer cooked meat of fresh fryers over the frozen according to Mountney gtflgl. (1960). CellOphanewwrapped fryers were evaluated by a panel of 1500 visitors at the 1957 Texas State Fair. Each was given 3 cooked samples to taste; namely, (a) fresh, (b) frozen-stored 3 months, and (c) frozen-stored 9 months. Half of the panel preferred the fresh fryers over those frozen and stored 3 months. About 60% preferred the fresh over frozen-stored 9 months. They concluded -33- that there is enough flavor difference to create a Slight product resis- tance toward frozen chicken stored for 3 to 9 months. Biochemical changes occurring in chickens stored at above freezing temperature were reported by Van den Berg 25,31, (1963). Poultry meat was stored at 0°C. under an inert atmosphere of nitrogen. Changes in odor, flavor, tenderness and juiciness of the meat were observed after cooking. Odor and flavor scores differed markedly from that of control samples held at -40°C. After 5 weeks storage at 0°C. the breast meat was less tender. Tenderness and juiciness increased in leg meat during the first week of storage, indicating tenderization during that period. The amount of extractable protein in leg meat also increased appreciably during this period, whereas, no change was noted for breast meat during storage. Proteolysis was appreciable in both breast and leg meat, re- sulting in free amino acids and other breakdown products. Ion-binding prOperties as measured by loss of weight and minerals during cooking changed markedly during storage. The water-binding capacity of breast meat decreased considerably during the first week of storage, whereas, no appreciable change occurred in leg meat. In a sequel to this study, Khan and Van den Berg (l96i9 reported on the quantitative changes in myofibrillar, sarcoplasmic, stroma and non-protein-nitrogen fractions of breast and leg muscles from chickens 10 weeks, 4 months and 8 months of age stored under aseptic conditions at 0, 2 and 5°C for 7 weeks. Changes were small in the stroma-protein fraction, actomyosin fraction, and the myosin-adenosine triphosphatase activity of the actomyosin fraction. The myosin fraction increased during storage except in breast -39- muscle of 10 week old birds. The sarc0plasmic-protein fraction decreased in the leg muscle of 10 week old birds and the breast of 4 and 8 month old abirds but not in the breast of 10 week old birds. Non-protein- nitrogen and protein breakdown products increased in both muscles regard- less of bird age. Proteolysis increased with storage time and temperature. Results obtained in this study were compared with those for storage at below-freezing temperatures. Zabik and Dawson (1963) compared poultry coated with acetylated mono- glyceride with controls wrapped in polyvinylidene film. When stored in the absence of other foods, flavor scores for the coated breast meat and the control were about equal. In the presence of other foods, the coated meat scored significantly lower than the control. Other Factors Chilling: Broth from half carcasses immersed in ice water for as little as 5 hours was found by Pippen gt 31. (1954) to have less flavor than broth from halves cooled in air. A subsequent study of the effect of chilling by immersion in ice water on chicken flavor was made by Pippen and Klose (1955). They demonstrated that loss of flavor resulted from ice water chilling, and that part of the decrease in flavor was traceable to the loss of neutralized ash by leaching. A comparison of broilers chilled for 21 hours by immersion or with "Spin-chill," first in water then in water plus ice by Kahlenberg ggdgl. (1960) resulted in a 22.7% gain by "spin" and a 13% gain by immersion. No Significant -40- difference was noted in the flavor of thigh meat. Greater losses of flavor occurred in liquid-chilled poultry as compared to air-chilled according to Hurley gtflgl. (1958). After water leaching by par-boiling, no detectable differences in roasted meat flavor were noted between either chilling procedure. Sex and Age: .A notable difference in flavor between sexes was demonstrated by Gilpin.g£.§i, (1960). ‘Males yielded more meat and were taStier, but females yielded a higher percentage of breast meat and fat. The age factor was found to have an important effect on flavor in poul- try by Peterson gELgl, (1959). These workers found the dark muscles were more flavorful from old hens than that from 3 month old pulletS. Also with older birds, the dark muscles were more flavorful than the breast muscle. A comparison of freeze-dehydrated breast muscle from 28 month-old, 19 week-old and 9 week-old White Leghorn females by Wells 23 .21. (1962) indicated that although color was best in the 19 week-old birds, tenderness and flavor were best in the 28 month-old birds. Aging_and Handling: Scalding caused flavor loss and toughening in poultry according to Wise and StadeLman (1961), but could be reduced by aging the birds for 24 hours. Greater cooking losses resulted in lowered flavor scores, and were shown by Ziolecki (1963) to increase from 17.8 to 22.8%, when chickens were aged 24 hours as compared to 7 days at 2'406. -ql- De Fremery and Pool (1960) demonstrated that mechanical handling resulted in a rapid onset of rigor and excessive losses of ATP. They also showed that muscles became tough and glycogen was lost more quickly by any treatment which caused rapid lowering of pH. WOrking with red meats, Briskey and'Wismer-Pedersen (1961) reported that rapid chilling Slowed the rate of glycolysis and retarded pH change. MATERIALS AND METHODS Chickens For the first phase of this study, old and young laying hens of known origin, which had been raised on identical rations (Appendix Table 1) were obtained. The birds were slaughtered, dry-picked and eviscerated. The entire carcass was ground in a meat grinder. Four of the laying hens were used; whereas, seven of the young hens were required to give appro- ximately an equal weight of meat after grinding (8.25 lbs.). In order to preclude bacterial spoilage, the meat was frozen.immediately after grind-I ing. The young hens, which were used in the preliminary chemical identi- fication of cooked volatiles, were prepared as described above. After 20 old laying hens had been killed, the intestinal contents were collected (1.5 lbs.) and saved for subsequent analysis. The old hens were then processed in the same manner as described earlier. The intestinal con- tents were used for subsequent studies, in which attempts were made to relate the volatiles from the cooked intestinal contents with those ob- tained from eviscerated ground whole carcasses. Fresh fryer breasts of unknown origin were purchased from a retail store, skinned and boned. They were subjected to gas chromatography and used for the determination of some of the gaseous components, including oxygen, nitrogen and carbon dioxide together with some lower alkane hy- drocarbons. Sixty heavy weight hens (Cornish cross hens), 65 light weight hens (White Leghorn hens), and 65 roasters (Cornish cross males) were selected -42- -43- from known sources. All of the birds were raised and maintained on Stan- dard commercial rations to an age of 16 months for the hens and 16 weeks for the roasters. Evisceration was carried out in the usual manner, after which the birds were divided into two separate lots. One of these lots was used for yield studies and the other for identification of chem- ical components. Thirty-five heavy weight hens, 35 light weight hens and 35 roasters were selected for the yield studies. The birds were packed in ice and held 14 hrs. at 30°F for further processing. The remainder of this group of birds consisting of 25 heavy hens, 30 light hens, and 30 roasters were used for the identification of chemi- cal components. The birds were packed in ice and held at 30°F overnight before preparing them for frozen storage. The birds were then de-iced, packaged as individual whole carcasses, and sealed individually under vacuum in Cryovac bags. Each bird was tagged for identification purposes. The birds were then separated according to class and placed on the freezer racks to be individually quick-frozen at -30°F. The birds were next crated according to class in marked crates and Stored at -10°F for fur- ther processing. Prior to analysis, the frozen birds were unpacked, sawed into halves at 60°F., and then the frozen halves were sawed in a manner that gave a crude separation of light from dark meat. Skin was removed after partial thawing. Kidney fat was removed. Then the bones, tendons and veins were cut away to give a careful separation of light from dark.musc1e. After grinding, the two kinds of muscle from each of the three classes -44- of birds were weighed as individual 100 g. portions and packaged in 4 x 6 inch Cryovac bags. The bags were sealed to preclude air, tagged indi- vidually for identification, quick-frozen at -30°F, and stored at -10°F for subsequent analysis. This group of birds was used for the following purposes: (1) to compare gas chromatograms of the cooked volatiles from light and dark muscle; (2) to identify sulfur compounds and functional groups present in the cooked volatile fractions of light and dark muscle; (3) to tenta- tively identify a phosphatidyl component in a "chicken essence" distillate; (4) to determine the diacetyl/acetoin and inosinic acid contents of raw light and dark muscle from the three classes of birds; and (5) to also determine pH, and possible precursors or precursor indicators including creatine, creatinine, cystine, methionine, sulfhydryl (glutathione equi- valent) and inorganic sulfide in both light and dark raw muscle and the cooked-freeze-dried meat-broth slurry for all three classes of birds. After cooking and distilling 2 kg. of muscle and 3 l. of deionized distilled H20 Slurry in a 12 l. flask for 50 hrs. at 180°F (Variac at 60), the flask contents were emptied into an 18 qt. stainless steel pail and mixed by hand. When the mixture containing broth, collagenous and fibrous residues became homogeneous, it was poured into two 12 x 12 x 2 inch stainless steel trays to a depth of 1 inch. The trays were covered with aluminum foil, cooled to room temperature and placed in a freezer overnight at -10°F. Then the trays were put on separate Shelves in a Stokes Model 2003 F-2 freeze-dryer and dried 36 hrs. at a tray tempera- ture of 42°C. Vacuum in the freeze-dryer compartments was maintained -45- between 75-100 u Hg. In this manner the cooked-freeze-dried meat-broth slurry aliquots of light and dark muscle from roasters, heavy- and light- weight hens were obtained for the study of changes in precursors in raw meat which occurred after prolonged cooking. The freeze-dried muscle- broth aliquots were friable and broke up when packaged in Cryovac bags and were held for analysis at -10°F. Proximate Analysis 'Moisture: ZMoisture was determined by placing 10 g. tissue in a loosely covered aluminum dish (75 mm. dia. x 15 mm.), and drying to a constant weight for 24 hrs. at 95°C. Moisture was calculated using the following formula: % H20 I original sample wt. - sample Wt. after drying x 100 original sample wt. Ether Extract: The dried sample obtained after moisture analysis was weighed into a Soxhlet extraction thimble of a porosity that permitted rapid extraction of the sample with anhydrous ether. Complete extraction of the fat from the tissue required 16 hours in a Soxhlet Extractor at a solvent condensation rate of 4 draps per second. The extracted tissue was dried 30 min. at 100°C, cooled in a desiccator and weighed (A.O.A.C. 1960). Ether extract was calculated using the following formula: % ether extract a sample wt. after drying - sample wt. after ext'n.x 100 original sample wt. Protein: The Kjeldahl method followed the procedure outlined by Benne $531. (1956). A sample of tissue weighing 1.5-2.0 g. was weighed on 2 x 2 inch vegetable parchment paper. The paper was then rolled -46- around the meat and dropped into a Kjeldahl flask. For digestion of the meat, the HgO or Hg used in the A.O.A.C. (1960) method was replaced with l g. CuSO4. Five grams of anhydrous Na2804 and 40 ml. of H2804 (N-free -93-98%) were added. The flask was heated gently until frothing ceased, and was boiled briskly until the solution cleared and then for an addi- tional hour. Before distilling, 200 ml. of H20 were added to the digestion mixture, and the mixture was cooled by placing the flask in the cooler for 2 hrs. Then 4 drOpS of mineral oil and a few granules of mossy Zn were added. The flask was tilted and 100 m1. of a pre-cooled 50% NSOH solution was added to form a layer without inter-mixing. The flask was immediately connected to the distilling bulb on the condenser. The tip of the condenser was immersed beneath the surface of the standard acid in the 250 ml. Erlenmeyer receiver, which contained 50 m1. of 0.2N H2804 and 2 drops of methyl red indicator. The excess standard acid was back titrated with Standard 0.2N Na0H. Protein was determined by multiplying the number of g. of nitrogen by 6.25 and expressing as percentage protein in the meat. :AEEF Two grams of tissue were weighed into a porcelain crucible and placed in a muffle furnace that had been preheated to 600°C. It was held at this temperature 2 hr. The crucible was transferred directly to a desiccator, cooled and weighed. Ash was calculated as follows: % Ash . sample wt. after ashing x 100 original sample wt. Creatine and Creatinine: Several colorimetric methods for determin- ing creatine and creatinine were compiled by Snell and Snell (1954). The -47- method chosen for this study was that of Peters (1942). However, the tissue samples were prepared for colorimetric analysis by the water ex- traction method described by A.O.A.C. (1960). A 10 g. sample of tissue was weighed into a 150 ml. beaker, 10 ml. of cold (15°C) deionized-distilled H20 were added and stirred to a homo- geneous paste. Then 50 m1. cold H20 were added, and stirred at 3 min. intervals for 15 min. After standing 2-3 min., the liquid was decanted through a quantitative filter, and the filtrate was collected in a 500 ‘ml. volumetric flask. The beaker was drained by pressing the liquid out of the meat fibers with a glass rod. Fifty m1. of cold H20 were added to the fiber residue. It was stirred for 5 min., allowed to Stand for 3 min., and decanted as before. The fibers were removed from the filter and returned to the beaker with a glass rod. Extractions of the fiber residue were repeated using two 50 ml. and four 25 m1. portions of cold H20. After the final extraction, the entire fiber residue was transferred to the filter and washed with three 10 m1. portions of H20, allowing time to drain between washings. The filtrate was diluted to the 500 ml. mark and mixed thoroughly. .A 150 m1. aliquot of the extract was measured into a 250 ml. beaker and evaporated to 40 ml. on a steam bath with occasional stirring. The solution was neutralized using phenolphthalein to spot test small aliquots. Chromogens were removed by adding 1 m1. of 0.1N acetic acid and boiling gently for 5 min. The coagulum was filtered on quantitative filter paper, and the beaker was washed 4 times with hot H20. The coagulum on the -43- filter was washed 3 times and the residue was discarded. For duplicate measurements of both creatine and creatinine, 5 ml. of the diluted extract were placed in a flask, and 40 m1. N/12 H2804 and 5 ml. sodiun tungstate solution were added. The mixture was shaken thoroughly, and the protein was removed by filtering. Eight ml. of the protein-free tungstic acid filtrate were measured into each of 4 colori- metric tubes, while 8 ml. of H20 were measured into a fifth tube as a blank. The mouths of two of the tubes were covered with aluminum foil, ‘ and the tubes were autoclaved for 45 min. at 15 lbs. pressure. The tubes were removed from the autoclave and cooled to room temperature. .A fresh solution of alkaline picrate was prepared (important to prepare just 5 min. prior to use) by adding 1 volune of 10% NaOH to 5 volunes of the picric acid solution. Four ml. of the alkaline picrate solution were added to each of the 5 tubes (the 2 which had been autoclaved, the 2 containing unautoclaved filtrate and the water blank). The tubes were set aside for 30 min. to permit complete deve10pment of color before colorimetric readings were taken. After color development, the readings were taken in a Model 6A Coleman spectr0photometer at 490 mu. The blank tube was read first with the galvanometer set to give a transmittance reading of 100. The colors of the other tubes were then read in the usual manner. The galvanometer readings were converted to creatinine by interpola- tion on a curve (Fig. l) constructed by plotting readings for standard solutions of creatinine zinc chloride (Hawk and Bergeim 1937). The standard semi-logarithmic curve was obtained by plotting percent trans- -49- {5 °/° TRANSMI T TANCE 3 FIG.I 2°“ Zn CHLORIDE CREATININE .510- m (WAVE LENGTH 490 mu) o I 2 3 4 5 s 7 CONC'N m, PERCENT OF CREATININE -50- mittance against concentration of creatinine zinc chloride in mg. percent of creatinine. The difference between readings of autoclaved and unauto- claved samples represented creatinine in terms of eaeatine. This was converted to creatine by multiplying by the factor 1.16. Dilutions of the tissue extract were made so that readings could be taken from the curve in the concentration 1 to 5 mg. percent. In this range, the accur- acy was i 4%. At concentrations higher than 5 mg. percent, the curve gave larger deviations. Diacetyl and Acetoin: Spectrophotometric determinations of the diacetyl and acetoin contents of raw-frozen light and dark muscle samples were made by the method of Prill and Hammer (1938) as modified by Stotz and Raborg (1943) and Pippen'gtugl. (1960). For diacetyl, a 50 3. sample of tissue was weighed into a 250 ml. beaker, and 50‘m1. of deionized-distilled H20 were added while stirring with a glass rod to make a uniform slurry. The slurry was transferred to a 500 ml. distillation flask and the beaker was rinsed with 10 ml. H20, which was likewise added back to the Slurry. Several granite chips were added to prevent bumping. The flask was connected to the distillation apparatus and an inert atmOSphere of carbon dioxide was maintained in the flask during a 3 hr. refluxing period. Steam was introduced slowly below the surface of the sample, and 5.0 to 5.2 ml. of distillate were collected during 25-30 min. Collection of the distillate in 1 m1. of hydroxylamine acetate solution facilitated the partial conversion of diacetyl to dimethyl- glyoxime. The stOppers of the tubes of distillate were loosened and the tubes were heated in a water bath at 85°C for 1 hr. to complete the for- -51- mation of the dioxime. The tubes were then removed from the water bath and while still warm, 1 ml. of acetone-disodiun phOSphate was added to each and allowed to react for 30 min. This removed any excess of hydroxy- lamine. Then 0.3 ml. of ammoniun hydroxide and 2.2 ml. of saturated tartrate solution were added to each of the tubes, which were then Stappered and inverted repeatedly to facilitate mixing. Then 0.2 m1. of ferrous sulfate solution was added to each tube and mixed by inverting the tubes. Thirty minutes were required to develop the color. A reagent blank was prepared by adding 5 ml. deionized-distilled H20 to an equiva- lent amount of reagents. The blank was treated in the same manner as the diacetyl tubes. After color development, readings were taken in a Model 6A Coleman SpectrOphotometer at 530 mu.(Krishnaswamy and Babel 1951). The blank tube was used to adjust the galvanometer to give a transmittance of 100. The colors of the other tubes were then read against the blank. The concentration of diacetyl was measured as ammono-ferous dimethylglyoxime by referring to a standard semi-logarithmic curve prepared for the Spectro- photometer (Fig. 2). Acetoin plus diacetyl was determined using a 50 g. sample of tissue, which was weighed into a 250 m1. beaker. Then 60 m1. of a 40% ferric chloride solution was added gradually with stirring to facilitate thorough mixing. The mixture was carefully transferred to a 500 ml. distilling flask and refluxed for 3 hrs. to oxidize acetoin to diacetyl. Due to the high concentration of diacetyl, 10.0-10.5 ml. of distillate were collected in 2 m1. of hydroxylamine acetate solution by careful Steam distillation for 25-30 min. The development and measurement of the color were the same -52- no. 2 2 ,. DIACETYL STD. CURVE (WAVE LENGTH 530 m) o l 2 ' 3 4 5 6 7 ‘ co~c~ nus/ml. or DIACETYL -53- as in the diacetyl measurements, except that the amount of reagents were doubled throughout. Inosinic Acid: Inosinic acid was determined for the raw, light and dark muscle samples using a semi-quantitative Spectr0photometric method suggested by Kazeniac (1963). After cooking and distillation of the raw ‘muscle slurry, the collagenous material coagulated and floated to the top of the solution. The fibrous material collected at the bottom of the flask and a comparatively clear broth formed an intermediate layer. A 50'm1. aliquot of the clear broth was pipetted off, and the slight haze which remained therein was removed by adding chloroform (v/v) and centri- fuging. After one-half hour of centrifugation, the haze concentrated at the interface between the aqueous and chloroform layers. The clear super- natant was then pipetted into a cuvette and ultraviolet absorption was measured at 250 mu with a Beckman DU Model 2400 spectrOphotometer using a hydrogen lamp. According to Kazeniac (1963), 90% of the absorbancy in chicken broth samples at 250 mu was due to inosinic acid. Absorbancies were compared with those of a standard curve for inosinic acid (Fig. 3). Values reported were based on the assunption that 90% of the nucleotide absorption at 250 mu was due to inosinic acid, and were corrected by cal- culation to a 100% inosinic acid equivalent basis. Hydrgggg Ion Concentration (pH): The pH.measurementS in this study were made with a Beckman.Mbdel G battery operated pH.meter equipped with small glass and calomel electrodes, or with a Beckman Zeromatic line- 0perated pH meter using a standard calomel half-cell and glass electrode. The instruments were Standardized with reference buffer solutions to read ABSORBANCE -54- 0.50 ' 0.45 r 0.40 .. 0.35 " 0.30 » 0.25 . (120 » FIG. 3 OJ: INOSINIC ACID SID. 9.0.8315. 0J0 0.05 L 1 L 1 l A) 00 2 3 5 0 7 ‘ CONC N. a. ”In /Ioo ml INOSINIC ACID -55- accurately in the range of pH 6.8-7.0 with the temperature adjustment knob set at that of the test solution. The results were expressed to the near- est one-tenth pH unit. Inorganic Sulfide Determination: Sulfide determinations were made on the raw-frozen and cooked-freeze-dried light and dark muscle samples, by the spectr0photometric method of Sands g£_§l, (1949). A 10 g. sample of tissue was extracted with water by the procedure used in the creatine/ creatinine determinations. A 25 m1. aliquot of the water extract was added to 25 ml. of zinc acetate solution. Then 5 ml. of the diamine reagent were added and stirred. One ml. of ferric chloride solution was stirred in next. The resulting blue color was allowed to develop for 15 min. be- fore reading the absorbance at 745 mu. A standard curve was develOped using standard sulfide solutions (Fig. 4). Readings were made on a Model 6A Coleman SpectrOphotometer at 745 mu by comparing the transmittance readings with a reagent blank set at 100% transmittance. Cystine and Methionine: Cystine/methionine contents of light and dark muscle from raw, frozen roasters, heavy- and light-weight fowl and from the cooked-freeze-dried slurries of these Same birds were determined by the microbiological assay method of Henderson and Snell (1948). The method utilizes specific strains of lactic acid producing bacteria with a single medium, which is made deficient in any single amino acid. The method can be used to determine any of 14 different amino acids titrimetrically. The colorimetric method of Friedmann and Graeser (1933) as modified by Barker and Summerson (1941) was also used for determining cystine and methionine values. ‘00 °/. TRANSMI 1' TA NC E 904 80-! 70 «I A O 30< -30- FIG. 4 INORGANIC SULFIDE STD CURVE (Sands £21., 1949) A n l A l J I 00MI 0.0002 0.0003 0.0004 0.0005 0.0000 0-0007 GRAINS OF SULFUR ISOII- OF SOLUTION (1 gram - 15.43 grains) -57- Assay media were obtained from Digestive Ferments 00., Detroit, Michigan. The media were Specific, with one being deficient in cystine and the other in methionine. A stock culture of Lgyggngsjgg_magenternides P-60, ATCC 8042 was obtained from the American Type Culture Collection, 2029 M. St. N.W. Washington, D. C. Stab cultures and inocula were main- tained and grown by standard procedures (A.O.A.C. 1960). Duplicate tubes containing 0, 0.2, 0.4, 0.6 and 1.0 ml. of Standard solutions of L-cystine and L-methionine were used to construct titrimetric and colorimetric standard curves. One or 10 g. raw, frozen or cooked-freeze-dried meat slurry were hydrolyzed by refluxing with 2 N HCl overnight for cystine determinations or with 3 N HCl for 24 hrs. for methionine determinations. The hydrolysates were neutralized carefully to pH 6.8 and diluted as necessary to conform with the makeup of the standard curves. Then from 0.2 to 2.0 ml. were added to 5 or 10 incubation tubes, reSpectively. Ti- trimetric determinations were made using a Beckman Automatic Titrator and checking the end point at pH 6.8 with bromothymol blue indicator. ‘Milli- grams of cystine or methionine per gram of original samples were calculated from m1. 0.0987 N KOH used by referring to the standard curves for cystine and methionine (Figs. 5 and 6). Colorimetric determinations were made using a Spectronic 20 calorimeter set at a wavelength of 560 mu. Percentage transmission values for the various samples were read off from the standard curves for cystine and methionine (Figs. 7 and 8). The milligrams of cystine or methionine per gram of original sample were then calculated. Sulfhydryl Content: The sulfhydryl content of the raw-frozen light and dark muscle of roasters, heavy- and light-weight hens and the cooked- -58- 13 \R\ sou-I m2.._.m>U I J 2 UZOU 5N 0N .N o. n. N. a d 4 d u a w>¢30 .OPm U£_._.m>UII_ m .67.. -59.. \£\x:\ 2.20.152 .. 4 2.028 0. V. N. O. 0 4 d d d - w>¢DU Ohm w2_ZO_I._.u§II_ 0 .07"— .IN 1 O 1 1‘ 718 l 9 1 O HON Lm'O fix -60- FIG. 7 CYSTINE srn CURVE (WAVELENGTH 560 um) l l l l I 2 3 4 r CYSTINE/n‘. -61- FIGS METHIONINE STANDARD CURVE (WAVELENGTH 5603).“) ‘7. p- I' 1 l 2 4 0 B IO Y METHIONINE/ ml. -62- freeze-dried meat-broth slurries from these same birds was determined SpectrOphotometrically by the method of Grunert and Phillips (1951) as modified by Batzer and Doty (1955). A Standard curve was constructed in the range of 10 to 200 ug of glutathione per 10 ml. of solution. The color intensity was directly proportional to the concentration of gluta- thione. Absorbance was plotted against glutathione concentration. In order to decrease the possibility of errors due to non-homogeneity of the Sample, a relatively large sample was used. Either 10 g. of meat or of dried slurry was mixed in a waring Blendor with 60 m1. cold 3% meta- phOSphoric acid and 20 m1. cold distilled H20 for l min. About 30 g. NaCl was added for saturation and the mixture was mixed for an additional 1 min. The solution was centrifuged for 10 min. and the supernatant was filtered into a 100 m1. volumetric flask and made to volume with a satur- ated NaCl solution. Eight ml. of this solution were added to each cuvette, and allowed to stand for 10 min. to equilibrate. One ml. of 2% nitro- prusside solution and 1 m1. of sodium carbonate-0.067M sodium cyanide solution were added. Each tube was read immediately in a Coleman Model 6A spectr0photometer at 520 mu. Two percent metaphosphoric acid solution saturated with NaCl was used as a blank. The standard curve (Fig. 9) was prepared from the appropriate amounts of reduced glutathione in a 2%wmeta- phosphoric acid-saturated sodium chloride solution. Absorbance values were proportional to the amount of glutathione in the range of 0 to 200 gammas of glutathione in 10 ml. of final solution. Determinations were made in duplicate. -53- .zuo» .1o_\.an 32:52.30 .z.uzou com oo. oo- o! 03 as oo oo as on o n J u u d d d u d Ass-own 5023933 m>m30 dew J>¢0>Im43m 0 .0: 00.0 0..0 I 0N.0 I 0nd I 0‘0 I 00.0 I 00.0 I 05.0 I 00.0 I 00.0 L 00.. JONVSUOSBV -64- Gas Chromatography_of Chickens Differing in Age Within 1-2 days after freezing, 8.25 lbs. of the ground chicken mixture was placed in a stainless steel bucket and immersed in a hot- water bath. After adding 4 l. deionized-distilled H20, a smooth slurry was prepared by kneading the cold meat and warm water together by hand. This slurry was filtered into a 12 l. flask, and the flask was attached to the apparatus shown in Fig. 10. The slurry was cooked under influx at 102°C for 4 hrs. with the normal cooking conditions described by Pippen eta]: (1958). After cooking 4 hrs., the cooling water in Condenser B was shut off and distillation began. The distillate was collected in a 1 l. flask (Fig. 10 - Trap D) and cooled in ice-water. The volatiles, which passed through the ice-water trap were collected in a liquid nitrogen trap (Fig. 10 - Trap F). When 2/3 1. of watery distillate had accumulated, the ice- water cooling trap (Fig. 10 - Trap F) was disconnected at K. A UV absorption test for carbonyls was made on the watery distillate by placing an aliquot of distillate in a cuvette and reading the O.D. maxima at 228 mu for monoenals and 282 mu for dienals (Kazeniac, 1961). Part of the absorption in the 228 mu range was likely due to sulfide, which according to Kazeniac (1961) also gives a maximum absorption in this region. No attempt was made to correct for sulfide as the main object was to verify that volatiles had been entrained by the watery distillate. The watery distillate was then subjected to vacuun and heated to 45-50°C in order to expel the volatiles into Trap F. After 25-30 min. under vacuum, distillation was Stapped, and an aliquot was pipetted from the l 1. flask into a cuvette for another UV absorption reading. This time the test was 3.38-3 3:.“ a notice road ... .52.: 3va n or 5.50 028 .5223 2:33.. .- 9323 3:2: 2.5;.- .c 55% .354: 2:22- .a ewe-338v. .cujocpzou use; .a .54.: 533 .o 2.343 26.. 5x8» 92 .3:— Eu. 6 75-83 29 :9.» .558 .5526 33 a}. :9. u 2:- 443 oh. wow-.- zoazut-u .5 a .z .54.: 5.83 a}. co.- . .5550 j:- Qu- 22 2.2. gum: 0.3.: 883.35. 1.3:. .a .54.: 5.83 a3- 5243 .84.:- .x x»: o..- uu- .u o/ Hus-o» 2:- ..«E. C35.- .: .o .3: EN. mum-.- zo-mzupxu .Eoo :- Sunooo zupaaczunzuozou .o .3: 3a- 0» 3308 52323538 .n \ .Saaueotvun cut-Ba axe-.3.- .x .65.. 2.34.533 u. .4 \ . ”mamaacuflmmmq \\ mnb oz< wzm: ado mm 3.20. no m2wumnmneOO mun swan nonwan we coxowuo mo pcwx menu you upon on» umnu mouooavsw A+v swam osHo .ooaouommap unmowmanwam on n .a.m.zn .nouu nowmuoaaa nowouuwo vwsvaa u m away Inw om A+vn¢ M\a mH m mm A+vo¢ «\H ma m A.n.m.zv an NH m\n ca 0 A.n.m.zv an ma ma o. Mm sq A+vam a m an A+V~o «\H on m 1 . mm A+vem «\m o a A+Vme SN m a ma A+vom «\H m m me A+vnm ~\H o m am am N\H m N «N nA+vo¢ ~\n m N 33.5 a m in m H £5.95 n a e H 1 1 «\H ~1~\H H new 1 u mu¢\H N new maoxoano was: A.OHEM. .oz maoxowso moo: A.GHEV .oz wan-ow Bo med. ammo wS-ow 3o 25. 33m Nunwaon xoom.x saga noausOquM, Newman; goon x mafia nowunouOMM moenao> sowuaouon o>aumnomeoo mossao> nowuaouou o>Huoumneoo comm um omenuu soum ooxmu moHHumHo> .oomm um um noun eoum noxmu moafiumao> .mnoxOH50 Odessa Howonom use moon pHO scum noumuonom mucoaonaoo oaauoao> .H wanna -122- in Table l which were appreciably higher for older hens than for younger birds. The volatiles obtained from younger birds gave only one peak out of seven at 23°C, Peak 4, that was higher than the corresponding peak for the old hens. Two of the seven peaks, Peaks 1 and 6, obtained from each type of bird were small peaks and these Showed no significant difference in retention volunes. Sampling at 75°C. gave four peaks out of seven (Peaks 3, 4, 5 and 7) in Table l which were appreciably higher for older hens than for young birds. The three remaining peaks at 75°C showed no significant differences in height for either type of chicken. The results of gas chromatographic analysis indicate that the vola- tiles from young and Old chickens were identical in chemical composition. This was demonstrated by the good agreement between retention times. The retention volumes for the older birds were consistently higher. This experiment indicated that the cooked volatiles from whole car- casses of young and old chickens were essentially alike, but that larger quantities of certain volatile components were evolved from Older birds upon prolonged heating. The origin and chemical character of these \rolatiles were not elucidated. Studies on cooked chicken volatiles by Bouthilet (1949, 1950, l951a,b), Pippen and Eyring (1957), Pippen 25.31. (1958), Pippen 33.31. (1960), Kazeniac (1961) and Pippen and Nonaka (1963) Show the importance that these workers attached to the relation of the composition of the volatile fraction to chicken flavor. The results ob- tained in the present study substantiate Lineweaver and Pippen's (1961) observation that the flavor of chicken pgrigg is fairly independent of type of bird. Although the flavor components of Old hens and young female -123- chickens were found to be the same, the older birds appeared to have more flavor per unit of cooked volatiles than the younger birds. Whether the particular portions of the volatile fraction of Old hens that were pre- sent in higher concentrations indicated a better flavor rather than simply ‘more flavor cannot be decided from the results alone. In fact, our pre- sent knowledge Of chicken flavor includes scarcely more than speculative information as to the origin of the so-called flavor volatiles. It is entirely possible that cooking serves to eliminate or reduce the amounts of certain undesirable flavor components in addition to pro- ducing the desirable "chickeny" flavor that develOps only after heating the meat. For example, Pippen and Eyring (1957) showed that the flavor of chicken broth was improved as a result of removal of ammonia. Kaze- niac (1961) Stated broth with a high hydrogen sulfide content had an undesirable "eggy" taste; whereas, the addition of a high concentration of ammonia to the solution reduced the egginess and gave the solution a sweet flavor and an improved aroma. Kazeniac (1963) suggested that ammoniun sulfide and ammoniun polysulfide may make a positive contribution to chicken broth flavor. Kazeniac (1961) also showed the need for ammonia flavor by ammonia and sulfide balance experiments. Carbonyls contribute to overall chicken flavor, both positively and negatively, according to Pippen (1958) and Kazeniac (1961). Accordingly, further evidence would be needed in order to conclude that old hens have a more desirable flavor than young birds simply because a higher concentration of certain volatile components was Obtained by cooking and distilling meat from old hens as compared to younger females. ~124- This study of heavy hen and younger female chicken volatiles showed that the volatile components from the two kinds of birds were essentially alike. It was also shown that the concentration of volatile constituents from old hens was higher than from young female chickens. This concen- tration difference was verified by gas chromatography and chemical tests. Hydrogen sulfide and carbon dioxide were identified in the low boiling portion of the volatile fraction which was distilled from the total cooked volatile fraction at -l40°C. Sulfide and carbon dioxide concentrations were markedly higher in the low boiling fraction from Old hens than in that from young female chickens. Vappr Fractometer Tests on Cooked Breast Muscle from Fryers The first peak from cooked chicken volatiles as determined with a hydrogen flame detector and a Model 500 F and M instrunent would not re- act with any of the functional group reagents, including concentrated sulfuric acid (Hoff and Feit, 1963). Thus, the presence of low molecular weight paraffin(s) was suspected. To determine whether this was true, vapor fractometer tests were made using pre-packaged fryer breasts that were obtained from a local market. The manner of preparation of the muscle has been previously described. The presence of oxygen, nitrogen and carbon dioxide in the volatile fraction was verified using the method of Brenner_cfl 21. (1959). The sam- ple passed through the Linde 5A molecular sieve within 5 min. and 3 peaks were obtained. These peaks were tentatively identified as oxygen, nitro- gen and carbon dioxide, in order of emergence from the column by using a -125- standard chromatogram for comparison (Table 2). The presence of nitro- gen and oxygen may have been due to a leak in the Buchler apparatus, which would allow liquid air condensed from the laboratory atmosphere to accumulate in the liquid nitrogen trap. The presence of carbon dioxide in the concentration indicated by the retention volume of that peak could not be attributed to the small amount of carbon dioxide that was present in the laboratory atmosphere. A test sample from the laboratory atmos- phere verified the low concentration of carbon dioxide and indicated that the carbon dioxide must have originated from the cooked chicken. Table 2. Vapor fractometer separations of the low boiling fraction in fryer breast muscle distillate Retention-time (min.) Tentative identity columnsa column Linde 5A HMPA on Linde 5A Peak molecular sieve Chromasorb molecular HMPA on No. Unknown Known Unknown Known sieve Chromasorb 1 2 2 1/2 1/2 oxygen nitrogen 2 4 4 l l nitrogen ethane 3 6 6 2 2 carbon dioxide 4 2 l/2 2 1/2 prOpane 5 5 3/4 none unknown available aColumns were connected in parallel. By using the tentative method for natural gas analysis of the Nation- al Gasoline Association of America (421 Kennedy Building, Tulsa 3, Okla- homa), which utilizes gas chromatography, the presence of ethane and propane was tentatively verified. Upon passing through the HMPA column -126- a total of 5 peaks were recorded on the chromatogram. Four of these peaks were identified as nitrogen, ethane, carbon dioxide and propane. The fifth peak was not identified. All of the peaks were eluted within 6 minutes. When authentic samples of ethane and prOpane were injected into the frac- tometer, the retention times matched those obtained on the chromatogram of the cooked breast meat condensate taken at -196°C. as shown in Table 2. A Phospholipid Fraction Containing,"Chicken Essence" When an experiment aimed at using smaller amounts of muscle for gas chromatographic and chemical analysis of cooked chicken volatiles failed, a modified distillation procedure resulted in the isolation of a phOSpho- lipid fraction. By cooking under "oxidation-favoring conditions" (Pippen ‘gg‘gl., 1958) and refluxing under a partial vacuum some yellow, green, amber and brown colored material collected in the small water trap (Fig. 12). This material had been steam-distilled over into the first trap and amounted to about 4 m1. of solution. ‘Members of the laboratory staff agreed that the freshly prepared distillate had a several-fold concentra- tion of true chicken odor. After standing overnight the color changed from dark amber to clear yellow. The pH was determined on a Model G pH.Meter and was 8.6. After 48 hrs. the "chickeny" aroma was nearly gone, which was probably due to the escape of sulfur compounds (Bouthilet, 1951; Pippen and Eyring, 1957). The fugitive nature of sulfur components contained in onion has been re- ported by Challenger and Greenwood (1949). -127- The solubility of the fraction was tested in water, ethanol, ammonia water, chloroform, and carbon disulfide. The material was readily soluble in carbon disulfide, water and ammonia water; moderately soluble in ethan- ol, and slightly soluble in chlorofonn. The fraction was subjected to thin layer chromatography. Two unknown spots moved faster than the lecithin contrOl spot. The two unknown spots were close together. By changing the phase from "runtral" to "alkaline" the spots moved more slowly. However, they still moved faster than the lecithin standard during the 75 min. development time at room temperature (Table 3). The identity of the spots was tested using ninhydrin Spray for amines, fuchsin-sulfurone reagent for aldehydes, hydroxylamine-ferric chloride reagent for esterified fatty acids, ammoniacal silver nitrate for glycerol and inositol, ferric chloride-sulfosalicylic acid reagent for phosphate groups, molybdic acid for phOSphatides and Dragendorf's reagent for choline compounds (Skiemore, 1962). The lecithin Spot devel- Oped with the Dragendorf reagent Spray. The slower moving unknown spot reacted positively to the ammoniacal silver nitrate. Both of the unknown spots reacted to give a positive test with the molybdic acid spray. Table 3. Thin layer chromatographic results for "chicken essence" from heavy hen breast muscle Rf values in cm. Sample 1a 2b 3b Lecithin ll 10 8 Unknown-l 19 1/2 14 12 Unknown-2 20 16 14 aNeutral" chromatOplates. Chloroform-methanol-Hzo (195:75:12) b"Basic" chromatOplates. Chloroform-methanol-g1acia1 acetic acid-H20 (62:25:8:4) c"Basic" chromatOplates. Chloroform-methanol-glacial acetic acid-H20 (50:25:8:4) -128- Authentic compounds were not available to test the Rf values of the two unknown compounds. From the literature values of Skipski grflgl. (1962), it appeared that these Spots were due to phOSphatidyl-ethanolamine and lyso-phosphatidyl-ethanolamine which Kates and James (1961) had pre- ‘viously isolated from old hens. However, the fact that a negative ninhy- drin test was obtained ruled these compounds out. It seems likely that these compounds may be cardiolipin and phosphatidyl inositol which Gray and Macfarlane (1961) have isolated from pigeon breast muscle. Bouthilet (1951) extracted an aqueous solution of redistilled and _ strongly flavored chicken broth distillate with isopropane. He obtained three fractions, and found that desulfuration continued upon Standing so long as true flavor was evidenced by hydrogen sulfide formation. He con- cluded that chicken flavor consists of at least two portions, a sulfur containing compound and a fatty acid-like material. Bouthilet (1951) post- ulated that the sulfur compound that is the precursor for the transitory hydrogen sulfide contributing to chicken flavor is glutathione. Peterson (1957) also found that the principal aroma component in chicken is sulfide, which he stated possibly arises from the heat treatment of the protein. Kazeniac (1961) found that steam distillates obtained from the fat fraction of chicken broth showed faint, chicken-like aroma, which indicated that fat may serve as a trapping agent for some of the volatiles. The volatile sulfur component(s) present in the phospholipid fraction isolated in the present study were not due to the presence of glutathione as the ninhydrin spray test (Skidmore, 1962) was negative. This indicates that the phosphatide itself may serve as a trapping agent for small amounts -129- of sulfur compounds, since only small amounts are needed to produce or- ganoleptic response (Guadagning£.§1., 1963). Furthermore, Schgberl (1933, 1936) observed that butylmercaptan decomposes in the presence of oxygen in a mild alkaline media to liberate dipropyl disulfide, hydrogen sulfide and propionaldehyde. Thus, results suggest that the phosphatide may serve as an electrostatic binding media for both the acidic mercaptans and the sulfides which have two unshared pairs of electrons on the sulfur atom or the disulfides which have one unshared pair of electrons on each sulfur atom. (Challenger, 1959). Due to the pH of the fraction (8.6), which was probably associated with the hydrogen ion binding prOperties of the aqueous sulfide, it appears that the only sulfur compounds present would either be sulfides or disulfides or both. .Mercaptans would be con- verted to disulfides and/or aldehydes (Schgberl, 1933, 1936) under the basic condition of the media. Aldehydes and alcohols, due to their acidic natures (Fieser and Fieser, 1956), would also be bound by the media. In conclusion, evidence indicates that a phosphatidyl lipid fraction or fractions in chicken may be responsible for the observation of Kazeniac (1961) that chicken fat may serve as a trapping agent for some of the volatiles. This lipid fraction could also be the fatty acid-like material that Bouthilet (1951) found was responsible for cooked chicken flavor, together with a highly labile sulfur compound contained in the chicken flesh, probably glutathione. ~130- Comparison of Total Cooked Chicken Volatiles from the Whole Carcasses of Old Hens with the Volatiles from the Cooked Intestinal Contents of the Same Birds The purpose of this study was to determine whether any chemical Simi- larities could be demonstrated between the intestinal contents of older hens and the cooked carcasses from the same birds. This was an explora- tory type of experiment in which comparative parameters were not closely controlled. The intestinal contents collected from 20 old laying hens (1.5 lbs.) were mixed with 500 ml. deionized-distilled H20 and cooked by refluxing under "normal cooking conditions" for three hours and then by cooking and distilling under "oxidation-inhibiting conditions" for 17 hours. Gas Chromatography: The results of gas chromatographic analysis are Shown in Fig. 18. None of the peaks was identified. A hydrogen flame detector was used in this study so that a positive correlation of reten- tion volumes with those obtained for the whole carcasses of old birds as shown in Figs. 16 and 17 could not be made as the parameters were differ- ent. However, Fig. 18 represents a chromatogram for the total cooked volatiles from the intestinal contents of old hens. Chemical Identifications: Prior to gas chromatography, the low boil- ing fraction was condensed at -196°C. and boiled off at -140°C. The results obtained matched those for the Same fraction of volatiles obtained from cooking the whole carcasses of heavy hens. Carbon dioxide and hydrogen sulfide were positively identified and carbonyl sulfide was tentatively identified. -131- .223: 3..- 8 a: .uoOsncEzou 3.1on -n_ x .-¢\_ .uonm or pop-53 goth .23: 20 5 8:02.80 .2532. 3.300 ES... 3:2; .3 n .0 68022520 30 .m .coE...oaxm 0. 23¢ .23 I wsE. zo_._.zw._.mm o N v o o o.- N._ a o._ m.- cm 1 ‘ 1 3 S NOdS 3 8 83080338 -l32- Results of Carbonyl Tests: A small quantity of DNPH was isolated (200 mg.). By means of column chromatography, UV and I-R.spectra and paper chromatography, acetaldehyde, acetone and n-hexanal were tentatively identified from the monocarbonyl fraction. Polar microscOpy and I-R Spectra were used to identify diacetyl, which was separated from the poly- carbonyl fraction by fractional crystallization. (Rice ££H£l-: 1951; Ellis EEHél-: 1958; and Pippen ggugl., 1958). Results for the carbonyl identification are shown in Table 4. Table 4. Physical properties of 2,4-DNPHS of carbonyl compounds from the cooked intestinal contents of old hens UV absorption Column Rf Rf ‘maxima Additional band 2,4-DNPH unknown known Unknown Known ‘methods of No. derivative cm. cm. ‘mu mu characterization fO rerun unknown - - - - - l n-hexanal 0.82 0.80 360 361 1 2 acetone 0.88 0.86 363 363 1 3 acetaldehyde 0.81 0.79 354 355 l 4 unknown - - - - - - diacetyl 0.00 0.00 - - l, 2 l. The infrared spectrum was identical to that of the authentic sample. 2. Microscopic examination Showed that the crystals had identical refrac- tive indices and phase transformations. A polar micrOSCOpe was used. Three monocarbonyls, namely, acetone-, acetaldehyde- and n-hexanal - 2,4-DNPHS, and one polycarbonyl--diacetyl--were tentatively identified in the cooked volatile fraction from the intestinal contents of old hens. The small amount of derivative recovered from the solvent cuts taken from -133- the center of each band (center cuts), which had sharp absorption maxima, was too small to make melting point determinations practical. Tests on the forerun solvent fraction (which contained no clearly defined bands) and on the chloroform extract of bands remaining on the column adsorbent, showed that in addition to those compounds which were tentatively identi- fied, other monocarbonyls and polycarbonyls were present. A violet color with alcoholic KOH gave evidence of the presence of at least one or more additional polycarbonyls, other than diacetyl. By reading the UV absorp- tion maxima in chloroform solvent, evidence was found that 2-en-l-al with a UV absorption maximum of 373 mu and 2,4-dien-l-a1 with a UV maxima of I 339 mu may be present. Results of Sulfide, Disulfide and Mercaptan Tests: The quantity of lead acetate derivative obtained from the cooked volatiles from the intes- tinal contents was extremely small as compared to the amounts obtained in the whole carcass experiments. Sulfides, disulfides and mercaptans were present but in low concentrations. The exploratory experiment for evaluating the characteristics of cooked intestinal contents indicates that the same classes of compounds probably exist in this media as in the whole carcass of chicken but in varying prOportions and in lower concentrations. The Chemical Identification of Carbonyls and Sulfur Containing Compounds Present in the Volatile Fraction of Cooked Chicken The purpose of this study was to attempt to confirm the classical work of Pippen §£.§l. (1958) on carbonyls, and also if possible, to deter- mine some of the classes of sulfur containing compounds that were present -134- in the cooked volatile fraction of chicken. Although the cooking was done under the atypical "oxidation-favoring conditions" used by Pippen .2£.El- (1958) for carbonyls, the distillation was carried out under "oxi- dation-inhibiting" conditions by using nitrogen gas to purge the whole carcass chicken Slurry and to provide a carrier gas for ebullition of the cooked volatiles through the 2,4-DNP and saturated lead acetate reagent absorption traps. The apparatus used is shown in Fig. 10, but was modi- fied by adding an absorption train as previously described. Carbonyls: The fractionation of the hydrazone precipitate was carried out according to the scheme described by Pippen ggugl. (1958). A total of 1.8 g. of crude precipitate of 2,4-DNPHS was obtained from 3 kg. of chicken, which was a considerably higher yield than that reported by Pippen et a1. (1958). It may have been caused by a more efficient air entrainment method associated with modifications in the cooking and dis- -tilling apparatus of Pippen $5.31. (1958). The separation and isolation of diacetyl-2,4-DNPH was made by a direct crystallization from a nitrobenzene solution of the polycarbonyl derivatives in the manner described by Pippen 2E 31. (1958). A yield of 0.15 g. of polycarbonyl-derivatives was Obtained, which was taken up in hot nitro-benzene and filtered. After several days, the crystals that separated were washed with ethanol and dried. Comparison of the melting point and I-R.spectra of this unknown material with those of authentic diacetyl-2,4-DNPHS showed that they were identical. During chromatography of monocarbonyl 2,4-DNPHS, other polycarbonyl 2,4-DNPHS were found as small bands. These were eluted from the colunn following acetaldehyde- -135- 2,4-DNPH. Treatment with alcoholic KOH gave a blue-violet color which is characteristic for 2,4-DNPHS having carbonyl groups on adjacent carbon atoms (Newberg and Strauss, 1945). The amounts of 2,4-DNPH derivatives present in the bands were small, and this fact as well as the lack of authentic derivatives precluded any further study. The unknown bands were easily collected, except for the butanal and ethyl-methyl ketone bands, which ran together (Pippen g; 31., 1958). To separate these 2,4-DNPHS, the eluted fraction was dried, taken up with chloroform and separated by chromatography according to the method of Lynn ggngl. (1956). An average distance of 5 cm. separated the centers of the two spots, which were outlined on the paper by pencil as the Spots were examined in a dark room under UV light. Pure ethyl-methyl ketone-- and butana1--2,4-DNPHS were recovered by elution from the paper with chloroform, drying and recrystallizing from hot ethanol. A single passage of equal amounts of the monocarbonyl derivatives through 11-75 cm. beds of adsorbent gave 9 preliminary bands by gradient elution. Two Slower bands were cut out of the column adsorbent media and these were eluted from the adsorbent with hot chlorofonm. Solvent cuts, that were taken from the centers of the 9 preliminary bands, were tested for UV absorption maxima in chloroform solvent in the range of 300-425 mu using a Beckman DU Spectrophotometer. Additional UV maxima readings were taken on the dried spots following paper chromatography according to the method of Nonaka E£.§l- (1959). The results are shown in Table 5. ‘Melting and mixed melting point results (Table 5) indicate that the center cuts taken from bands 9, 8, 7, 4, 3 and 1-gave sharp melting .mnowumeuommnmuu amuse mam moowvow O>Huomumou on noonmou suaa oxwam mums mamumzuo man umnu poSOSm omoomouoae wsauwumaon o Suez mamumhuo mo nowuosaamxo OaaoomouOHZM .oaeamm cannonuso man mo noun ou aoowusona mp3 sonuOOQm nouumumsw ones .moaeaom.oausonuno sues noxmono mp3 mewuwoumeouso.nonun no noauouwwzn .o>wuo>auop fiend-aoamnaam mo oaumwnouoouono soauauomnm uoaow>ouuan msoamo .O>Huo>wuom oaosmxamnm N mo monumauouomuoso coauauomnm uOHOH>oHuH= msonmn .o>«uo>anom annexamun no mo mogumwnouomuono soauQHOmnw uOHOa>mHuHS maonmo m.o com o>onm .p can o>ono .p Hanooowv com o>onm .m ahnonuoo u n u nsonxnn n umaoe u n u caosxs: monem OH 0 .m wo2-e02 me2-mo2 mexnmenmnmom mon-¢o2 m m .e .n a~2-¢~2 5N2-m~2 muonnom a~2-e~2 m n .m wm2-em2 ann-mn2 Hmnmnonn-n mm2-¢n2 a - m an an #HH1HHH maanaaa ocouox ahsumuahnuma mHHIHHH o w m J «2.22 «2.22 2.2-San-.- m2-~2 m o no oHHIwOH oaanmoa Honmuooman oHHuon a m .e .m Son-m02 o22-mo2 nmnmnne-n moH-no2 m o u OHAoHHm>mon Houaucoapueaw emanoma N m .m aOn-¢O2 mo~-~ca Hnnmnoo-n no2-eo2 2 u u u ssosxsn mmanw annouom nowumufiuouooumno 0o ..m.z_ 0o 44m.z. powmaunopa Uo ..m.z .02 mo mposuoa magnum nausea pnuoneoo Hmaonumo coauooum pawn Hmooauamp< Owuaonunm oausonun< mouwaamummuoom canaoo sues poxaz moHHuoHo> sexuaso poxooo scum nouoHOmH mpsooaaoo ahoonnoo mo mmmznnqam mo mowuuomoum .n manna -137- points. The center of band 10 was cut out of the adsorbent, eluted with hot cthroform (UV reading taken), dried and recrystallized from hot ethanol. Band 11 was small and poorly defined so that no tests were made on it. The center cuts from bands 3 and 4 and 5 and 9 gave melting points that were so nearly alike that paper chromatography together with UV and I-R spectra readings were required to establish their reSpective identi- ties. The UV readings are shown in Table 6 together with the Rf and Rh values for 8 of 11 bands. Results of I-R spectra studies on acetone, diacetyl, ethyl-methyl ketone and n-hexanal are shown in Figs. 19, 20, 21 and 22. No authentic samples were available for confirming the identities of fractions 2, 10 or 11. On the basis of melting point and mixed melting point results (Table 5) together with paper chromatography and UV absorption results (Table 6) fraction 1 was identified as n-octana1-2,4-DNPH. USing melting points (Table 5) and UV absorption data (Table 6), band 2 was tentatively identi- fied as 2,4-dien-l-a1-2,4-DNPH (Pippen SE El-: 1958). A total of 8 monocarbonyls were identified by 3 or more criteria in- cluding melting point, chromatographic behavior, crystallographic proper- ties, ultraviolet and infrared absorption spectra. Two ketone derivatives, namely; methyl ethyl ketone and acetone were isolated and identified together with 6 aldehydes; namely, the n-alkanals containing 2, 3, 4, 5, 6 and 8 carbon atoms. One polycarbonyl, namely, diacetyl was separated by the solubility and crystallization method of Pippen ggflgl. (1958). Identity of diacetyl was also confirmed by melting point, I-R spectra and microscopic examination of its crystallographic prOperties. - - - - - emamnnnnenns 22 - - - - - ennunnnmeana an mom 2m.o mm.o o.m m.o2 meanoeamumon a mum n¢.o o¢.o m.ma ~.ma onouoom m can 2e.o ne.o o.~2 2.mn Hanmnonn-n a .a «an mm.o «a.o m.- m.a~ sconnx Hannm-2mnnne s 3 A. can qn.o mm.o m.m2 m.e2 Hanannn-n m can 05.0 mn.o o.- c.2N Hanannnn-n a cam om.o em.o o.o~ H.mN Hmnmxmn-n m mam - - - - emauannnenns N onm «H.H o~.H mm mm HonouOO1a a 1 1 1 1 1 nowmwuaomaan anuouom :8 .80 .80 .80 .80 o>wum>wumm .oz annexes nsonx manage: sauna nausea: mmza-¢.~ enmn coauQHOmno >2 am mm mm mm nasaoo mo>auo>wuom mmzn1¢«N mo muoam mo «Beams soauauomnm >3 pom anonymouoeouno Hoamm .c manoH -l39- .mmzm1qam1osouooo mo meanemm nsosxo: mam Ofiuaonuoo man How uuuoonw moa1onmsH .mH .wam Iazo v-N uZOhuu< mmroh<2 hmJme co: 2. mm4.h<1_0> luv-0.10 30.: 2320 '1N nzocoi z. xhozu4u>4I -l40- .mmzn1qau1ahuoooan mo moanadm naonxon mam owunosuoo onu pom auuoonm non1mumaH. .oN .mam Into v1N 4>hw0(.o mw20h(8 bwaqwa CO! 2. mu4.h<1_0> tutu-20 30¢... 11.20 '-N a. a ntOCga 8. Shotuau>¢fi 2 O. o I h 0 a q d T q 4 an .1 q huAJwa max 2. Into ¢1N J>hu045 thwthat on! ..e On... Obr 8 t 00.1 u d d - + D a q q — .- p P b r p p r b F 00. 8°. 00: 8a. 009 8! 80. 008 .30 lid)! 8 it 8 i! 8 SI 9 Si 8 -14l- .emzn-a.~ 1onouox1ahnuoa1amuuo mo mOHQEOm neonxn: use owusonunm aha you caveman wos1mumnH .HN .me 112° v1N uZOhwx Arrhu814>2hu mwxuhdl huJJua co: 2. 1061:5340) 29.0.20 30¢... 2120 C1N clan-=8 8. Shelua u>(’ c. o. v. n. u. : O. o o s o o o n a O m a d A 4 d 4 d q 1—I 4 d d 4 o o. T 1 o. o~ fir 1 On On 1.1 1 3 oc _T 1 O. on .1 1 00 oo .1 1 8 as I 1 on 0.1 1o. oo .1 1 on 00- 1 00. p p h b p p P b h - p p P b bunnwd max 2. Iazo ¢1~ uZOhux 421.521.;th e_hzuzhad 0 H1 4 a I a q _ 1 A 2 e a a a 1 O o. .F 1 O. 8 T 1 On an _r 1 On 0. 1 1 O. on 1 .1 Go 0.1 .1 1O. 1 Op 1 1 On one 1.8 on T 1 00 00. p P P F p F p L L b p p p b 06. one 02 8. can SO. 8: ocu- 89 oo! oom- ooou 33 89. 3983 .180 3883.3); -142- .mmzo1aam1fionuxon1a mo magmawm neonxon moo oeuaonuno How ouuoomm-mou1oumnH .NN .wam znzo Tu 34:45:12 3.8:: 5.3”...- zox z. mun-=40) 29.2.8 :2: zero Ta 22......- z- 523:!- o- 9 v. n. u. : o. a a h o a v n n O I J. A a J d d d 4 — a d d d 1 O o- .1 1 o. 8 1 1 8 on I 1 0» co 1. 1 00 on r 1 o. 8 1 8 2 1 1 2 o. 1 1 8 2 1 1 8 00. F p F p p p F p P _ b p p 00. 5.3”:- cmx 2_ 11.3 Tm 44288-112 onu-taa o .1 a _ _ _ 2 _ a _ a - - a T 1 o o- 1 1 o- Ou .1 1 ca 04 1 2 9 1 1 8 on 1 1 8 o. .1 1 o. 2 1 12 8 1 1111 1 8 o. 1 1 o. 8. 8- b p p k p P P p p p p p h h S. 82 8. 8. 8o- 8: 8a- 8». 8c- 82 88 88 8» o8 88 7.8 5252!:- -l43- The.e1egantinvestigation of carbonyls in the cooked volatile fraction of chicken by Pippen ggdgl. (1958) was confirmed in a limited manner by the results obtained in the present study for 8 monocarbonyls and l polycar- bonyl. Pippen gtflgl. (1958) used larger chickens that probably had a higher fat content than those used in the present study. This may have been the reason that a larger number of carbonyl fractions were obtained than in the present study. Nevertheless, a total of 9 carbonyl derivatives were obtained, which had the same physical properties as 9 of the 18 car- bonyls that were identified by earlier workers. This showed that the cooked volatiles from both kinds of birds were similar, at least in re- spect to the 9 carbonyl components. Lineweaver and Pippen (1961) have pointed out that the importance of carbonyls to chicken flavor is not well understood. In a recent study using gas chromatography, Pippen and Nonaka (1963) found that n-hexanal and n-2,4-decadienal were two of the principal components in the cooked volatiles of rancid chicken, whereas, n-hexanal was prominent in the cooked volatiles from either fresh chicken or turkey, as well as in rancid chicken volatiles. Larger amounts of these two carbonyls were found in chicken skin and the fat attached to the skin than in the lean leg or breast muscle of chicken. The present study seams to confirm Pippen and Nonaka's (1963) findings concerning carbonyls in fresh chicken, although the identity of the n-2,4-decadienal derivative was merely tentative. Sulfur Compounds in the Cooked Volatile Fraction: Sulfur derivatives formed in the reagent traps in prolific amounts. Their formation started -144- within one-half hour after the cooking-distillation temperature of 180°F was reached and continued without abatement until the end of the cooking- distillation period. This same phenomena was previously reported by Yueh and Strong (1960) in their studies on cooked beef volatiles. Mercaptan Test: A positive test for mercaptans was obtained in the form.of a light yellow precipitate of 2,4-dinitrophenyl thioether(s). This test was made using the method of Folkard and Joyce (1963) in the manner previously described. Sulfide Test: A positive test for sulfide(s) was obtained as evi- _ denced by the formation of methylene blue using the method of‘Marbach and Doty (1956). Test for Disulfides: A positive test for disulfides was obtained using the methods of Stahl and Siggia (1957) and Rittner gt il° (1962). Bouthilet (1951), Pippen and Eyring (1957) and Kazeniac (1961, 1963) have recognized and stressed the importance of sulfides and other sulfur compounds in chicken flavor. In the present study, sulfur derivatives were visible in the traps within one-half hour after heating of the chicken- water slurry and ebullition of the volatiles through the reagent traps began. In contrast, the formation of carbonyl derivatives took place slowly, and a visible precipitate of 2,4-DNPH was not visible until about two hours after heating began. Thus, it appears that organic sulfur com- pounds are probably the first odorous compounds obtained in quantity in the volatile fraction, apart from hydrogen sulfide and ammonia both of which have low boiling points. This suggests that at least part of the typical cooked chicken flavor may be related to sulfur compounds; princi- -145- pally, hydrogen sulfide, organic sulfide(s), organic disulfide(s) and to a lesser extent to mercaptan(s). Speculatively, carbonyls may have a split role. The lower carbonyls, principally acetaldehyde may play a major part in the "browning reaction". Acetaldehyde has also been described as adding a scorched flavor note in 'flavor profiles of chicken broth (Kazeniac, 1961). Low flavor threshold values have been found by Lea and Swoboda (1958) for many carbonyls; for example, n-decanal was detectable in water at a concentration of 5 x 10"8 moles per liter. Patton ggngl. (1959) reported that n-2,4-decadienal, which Pippen and Nonaka (1963) had shown to be a major component in cooked chicken volatiles, has a "deep-fat fried aroma." However, Pippen and Nonaka (1963) have noted that it can also have a desirable cooked fat or fried chicken odor. 0n the other hand, these workers, found that on exposure to air at room temperature, n-2,4-decadiena1 first developed stale and then rancid odors. Furthermore, n-hexanal was found to be the most prominent carbonyl compound in rancid chicken by Pippen and Nonaka (1963). The concentration of n-hexanal in rancid chicken was found to be much higher than in fresh chicken (Pippen and Nonaka, 1963). Accor- dingly, it can be concluded that the higher carbonyls, such as n-hexanal and n-2,4-decadienal, can contribute to desirable aroma but may also be immediate precursors of stale or rancid odors. This leads to the specu- lation that carbonyls may be responsible for over-cooked and off-flavors as well as to some desirable flavors in cooked chicken volatiles. Per- haps, some pleasing flavors in cooked chicken may be due to the presence of low or even sub-threshold concentrations of Specific carbonyls. The ~146— importance of acetoin and n-2,4-decadienal, which may impart transient buttery-oily or fried chicken odors, respectively, has been shown by Pippen and Nonaka (1960, 1963). Cooked Meat Yields from Roastgllfij Heavy- and Light-Weight Hens The purpose of this study was to determine the cooked meat yield from three classes of chickens. This was accomplished by dividing the carcasses into wings, necks, backs, legs, breasts, giblets, and skin, which constitute the so-called solid portion of the chicken, and into the broth and fat, which constitute the liquid portion. The waste due ' to cutting and handling was also determined for each class of bird. Table 7 shows that the uncooked White Rock roasters were two times and Cornish Cross hens two and one-half times as heavy as the White Leghorn hens. On a percentage basis the breast meat content of the light hens was higher than that of either the roasters or heavy hens, but on a weight basis was considerably less. Breasts from the heavy hens weighed considerably more than those from roasters. On a percentage basis, the difference was significant at the 5% level. Legs from roasters and heavy hens were about equal on a weight basis, but on a percentage basis the roaster legs showed a greater yield with the light hens second and heavy hens last. Heavy hens yielded the heaviest backs. From Table 8, it is evident on a percentage basis that cooked roasters yielded more leg meat than light hens, which in turn yielded more than heavy hens. On a weight basis, the light hens yielded less than one-half as much leg meat as heavy hens or roasters. The breast meat yield from .HH>HH NH Hm HamuHmHame= .Hm>ma NH Hm HaHoHHHamHm. .mnOHuoH>ov photomumv .Ho>oa NH um unoOHMHanmHH .Hm>oa Hm Hm HnHoHHHamHmH .waonoo Hound vo>oawu mwaHs was «mxoon «axoun scum ame musomoumou Gmeu .waonoo opomon ame mug was unmoun «waonoo Hound voowHos was vo>oaou ohms ame good was Hcme mGHs Hame xoomn .maonoo ou HOHHQ wo>oaou moa £UH£3 unmoun was mmoa scum onm mo Hmuou coanaoo can muaomoumou amem «.o H o.ooH mad H HHOmou 0.0 H o.ooH mom H stmou q.o H o.ooH 00H H HHNon Houoa m.o H a.o N H om m.o H m.o w H mm m.o H m.o q H Ha mommog . H.H H ¢.m «N H oh N.H H m.m mm H HHmoH ¢.~ H :m.N mm H mm muoHAHu m .. .. .wHHS o.~ H Tm 8 H m: .. .. - .933 3H HHESE N.H H m.N OH H «Hung H.~ H .m.w «o H HHNHN ¢.H H 5.0 ma H HHNN monm o.H H ¢.HN #0 H ssnmq m.a H ¢.m~ mm H HHomn m.H H :o.¢~ mm H *HHoN umoe unmoum o.N H N.H~ mm H HHqu a.~ H :H.m~ moa H Hsmoo o.~ H o.H~ mm H HHmNN xomm m.H H =o.m~ on H own m.H H =H.- am H mom m.H H =0.HN RN H HHHHN smug m.o H :m.¢ Hm H HH¢¢ 5.0 H :m.~ om H HHNN H.H H :N.m HH H «seq xooz m.H H =m.~H an H mom m.H H =H.OH mm H Hum H.o H =o.HH «H H HHHNH conga Hmuou mo N maouw Hmuou mo N mawuw Hmuou mo N mawuw uoononaoo mnoumeM moon hbmom mam: ustA owmunoouoa mam uanoa mo mmmao sumo mo mamMOHno mm scum vmsHmuoo melo voxoooqs owouo>m can no ..N canoe Summary of the average cooked yields obtained from 35 chickens of each class by weight and percentage Table 8. Roasters % of total Grams % of total Heavy hens Grams % of total t hens Li Grams Component Wings 3.9 i 1.0 86 i 19 3.2" i 0.7 81 i 17 3.9 i 0.6 2.1 9 42** i Meat 0.5 +| 2.3 11 +l 49 0.3 +l +l 55 3.5" i 0.6 1.9 7 38** i Bone 0.9 +l 52** i 21 2.0" i 0.9 70** i 21 3.3 i 0.6 6 21** i Skin Neck i 0.4 1.2" i 0.2 1.8 +l 37 1.4" i 0.4 4 1.8 i 0.4 35 i 11 1.8" i 0.4 20** i Meat 0.2 +l 5 5 1.1 25** i 23* 0.8" i 0.3 1.0 i 0.4 6 22* 25** i 11 1.7" i 0.4 3 4 15** i 14** 1 Bone Skinb Legs i 33 13.1' i 1.1 274 i 36 10.7" i 0.9 284 125** i 17 11.5' i 2.1 Meat -148- 0.8 +| 3.6 75** i 15 0.3 +1 7 4.7" i 0.9 92** i 9 3.5 51** i Bone Back 5.7" i 1.1 4.0 i 24 120* 83** i 13 129* i 21 5.1' t 0.7 107** t 21 4.0 137* 154* 4.6' i 0.9 ‘5.5" i 1.3 50** i 11 Meat 0.7 +l i 0.8 60** i 13 Bone 1.5 +l 6.2 i 33 1.9 +l 5.8 i 55 i 1.9 58** i 11 5.4 Skin Breast 12.4' i 0.9 259** i 36 40 i 1.2 13.1 147** i 16 13.5' i 1.7 348** i 53 Meat 2.2" i 0.5 10 +l 1.6" i 0.2 6 +l 2.7" i 0.6 43 7 29** i Bone i 0.7 1.8 38** i 11 i 0.6 i 15 1.9 52 i 11 4.4" i 1.7 48 Giblets See footnotes Table 7. ~149- 1ight hens (percentage basis) was greater than that from heavy hens, and from heavy hens was greater than from roasters. On a weight basis, the heavy hens yielded more than twice as much breast meat as the light hens and one and one-third times that from roasters. Back meat ranged from 50 grams, or4u€1percent from light hens, to 120 grams or 5.7 percent from roasters, and 137 grams or 5.1 percent from heavy hens. Neck meat varied from 20 grams or 1.8 percent from light hens, to 35 grams and 1.8 percent from roasters and 35 grams or 1.4 percent in heavy hens. Wing meat amounted to 42 grams or 3.9 percent from light hens to 81 grams or 3.9 percent from roasters and 86 grams or 3.2 percent from heavy hens. waste, in Table 9, refers to veins and cartilaginous material plus drip. Results are graphically summarized in Fig. A. In Table 10, the average cooked lean meat content with reference to the uncooked eviscerated carcass weight is given. Light hens yielded 35.5 percent of lean meat, heavy hens 33.5 percent and roasters 36.3 percent. Table 10 also shows the average yield of cooked total edible meat (lean, skin and fat) based on the uncooked eviscerated carcass weight. Light hens yielded 47.7 percent of total cooked edible meat whereas heavy hens yielded 48.8 percent and roasters yielded 52.7 percent. A graphical summary of these results is shown in Fig. B. The significance of these data may be limited by several factors including boning method, cooking method, meat recovery method, eviscerated weight and eviscerated price per lb. as follows: -150- .2 magma mOHOGHOOM 60m w.m H m.om QNH H HHmmN H.m H :H.m¢ mma H HquHH H.@ H N.Nm moH H HHomm Houoa o.o H :¢.n om H «HNHH 0.0 H :m.o cm H HHNNH m.o H :m.m HH H «Hem umm m.m H m.om moa H Hano N.N H :N.om Nqa H HHmNm m.N H «.mu mm H «Homm nooum HeHsuHH N.m H .N.mo NNH H HHHmmH H.m H :¢.om mmH H HHmamH N.N H .m.no Nq H HHHmN Hmuou moHHom mm. Hm NN. an an. on mmoa cOHumummmm o.~ H 0.0H mq H HNN m.a H :m.m mq H mmu m.~ H o.m Hm H HHom oame m.H H :m.mH «m H HHwNN 5.0 H :o.NH mm H *Hmam o.~ H :o.wH Hm H HHmmH moon N.m H .m.om «Ha H HmeN H.~ H :m.mm ooH H HHoaw o.¢ H .m.mm. «m H HH¢wm umoz mcHHom m.o H H.H HH H mm m.o H w.o m.w H Hm H.o H m.o H H *Hm mamas Houou mo N msouw Houou no N madam Hwoou mo N mawuw ucoaomaoo mumummom mam: h>oom mam: uSMHA . mo maoonno mm scum voaHmuno .owouaoouon paw uanmsan mmoao 30mm 1 mvHovHH poo mmoH aOHuonnom «mvHHom «cummz moo mo humasom .m manna . 12.111,- 35.01 30.0. 25.0 _ .3) 10.01 Heavy Hons ‘ .i-___.] 1) :1" W1 1) :1. - 1| F’! H a . A o , a C ,’ ”at A 9 o E 1 EL BE z’r. 7 _ f 1 1%“. Light Ohm Roasters GROUPS Fig.1.. Percentage of coalud mad as”? A—breast, B-Ieg,C-bock D-winq, E—neck and F- totat compared wflt the percentage 0!. bone for hens Dona than some carnponenta,1n heavy hem, light roosters. - E111... .Awao axooB oHv muoummou zoom muHszv .Avao mnuaoa oav moms macho anauoo ouHszo .Aeao nausea oHv mama aHoame HHHazn .GOHuoH>ov wuaunouma L. 5 .4 H.~m HaoH “Ha H omom H.m H n.8m HHH H wmh Baa H omom HHHHHHHQM w.ms HmNH mom H mmom H.N H m.mm sea H cam mom H mmom swam: H>Hmm H.2H on ooH H HmoH o.¢ H m.mm Hm H «mm sea H “woa ammo; HamHH Hauou mo o\o mEMHU 9:me Hmuou mo N mEmHU $9:me CHAD mo mmMHU HHHHH‘.>< 3m eoxoooc: HHHHH .>< 3m Hoxooopa umma.oHnHwo voxooo Houoa uooa_amoH voxoou A3mV uanoa wouanoomH>o woxoooo: no woman umoa oHAHwo woxooo Hmuou wow umoa Gama voxooo mo melo omnum>< .oH oHan PERCENT (Va) 50. A O 0‘ O 20 60'» -153— U D A A A B Heavy Hens Light Hens Roasters G R O U P Fig. B. I Meat Percent age of E2 Skin "ll A— meat, 8- skin, C- fat. D—totai edible portion in heavy hens, light hens and roaster: . Fat cooked D Total -154- (l) A preliminary study revealed that when the carcass was boned prior to cooking,the cooked meat yield obtained was lower than when the carcass was cut up before cooking, but not boned until after cooking. (2) Conventionally cooked chicken was compared with electronically cooked 10 week old fryers by Schano and Davidson (1958). They showed that flavor quality of electronically cooked chicken was lower but that the yield was higher than with roast or rotisserie cooked birds. (3) 01d fowl was cooked by boiling, simmering and pressure-cooking as reported by Kahlenberg and Funk (1961). Simmering gave higher yields and pressure-cooking more tender meat but with a lowered fat content. (4) In the present study it was noted that the same number of man-hours were required to separate the lean meat from the birds whether roasters, heavy or light hens were used. Since the yield of lean meat from the 35 light hens by weight was about one-half of that ob- tained from the other classes, the cost of hand separation was doubled. However, cooked lean meat from light hens could be separated commer- cially by the Harris flotation method as described by waskiewicz (1962), which would make the cooked lean meat from light hens less expensive than that from heavy hens or roasters. (5) It was also noted that heavy hens which gave the highest yield of cooked lean meat on a weight basis required less man-hours of labor per unit weight of lean meat separated. In other words, heavier birds resulted in a reduction in the cost of hand separation. Fig. C provides a chart for each of the three classes of chickens. A horizontal line from the price per 1b. eviscerated weight of the class 455: LIGHT HENS HEAVY HENS ROASTERS (2.5 lb.) (6.0 lb.) (5.0 lb.) Price/ lb Cost/lb Price/ lb Cost/lb Price/lb Cost/lb Eviscerated Cooked Meat Eviscerated Cooked Meat Eviscerated Cooked Meat (Uncooked) (Lean) (Uncooked) (Lean) (Uncooked) (Lean) [.874 L2 P30 .836 .35 |.04 .425 l.l7 .29 .8l8 .34 LC! .4l0 l.l2 .28 .789 .33 ..98 .393 l.08 .27 .76! .32 .95 .380 l.04 .26 .733 .3! .92 .385 LCD .25 .705 .30 .89 .350 ‘ .90 .24 .677 .29 .86 .335 .9l .23 .648 .23 .83 .320 .87 .22 .620 .27 .80 .305 .83 .2 l .592 .29 .77 .290 .79 .20 .564 .25 .74 .275 .75 .I9 .536 .24 .7I .260 .7l ,l8 .507 .23 .68 .245 .67 .l7 .479 .2 2 .65 . 230 .63 ,l6 .45l ,2i .62 ,2 IS .59 .l5 .423 , .20 .59 i .200 .55 .395 i .56 .5I Fig. 6. Chart for calculating the cost/ lb of cooked lean meat for three classes of chickens. -156- of interest extended through the cost per lb. on the cooked lean meat axis will indicate the cost/1b. of cooked lean meat. This cost does not include labor or overhead and may vary up or down with lighter or heavier birds and in accord with the other parameters which have already been discussed. Summary: A total of 105 birds (35 roasters, 35 heavy fowl, and 35 light-weight hens) were weighed and cut up, and individual parts were then weighed to ascertain the uncooked weights and yields of these various parts. The parts of each bird were then sealed in coded containers and pressure cooked. The meat was then removed from the bones so that the. meat and bone contents of each part could be determined. Skin from legs and breasts was removed and weighed prior to cooking. Neck, back, and wing skin was separated and weighed after cooking. Broth and fat of each bird were separated and weighed. A composite value for the total cooked weight was ascertained for each bird. Analysis of variance was applied to the data to test for differences among classes, and standard deviations were calculated. Roasters yielded the highest percentage of cooked lean meat (36.6%), light hens next, and heavy hens the least. Yield of total edible product (lean, fat, and skin) was likewise highest for roasters (52.7%), followed by heavy hens and light hens. Light hens had the high- est percentage of bone, whereas heavy hens yielded the highest percentage of total liquid (broth and fat). Light hens yielded the highest percent- age of cooked breast meat (calculated from dressed weight); whereas, roasters had the highest percentage of cooked lean meat from the legs. The percentage of breast meat was higher from heavy hens than from roasters, -157- but lower than from light hens. From the data, a chart was develOped for sight-calculating the cost/lb. of cooked lean meat for these three class- es of chickens. Chemical Analyses The purpose of the chemical studies was to correlate the chemical composition of a given raw or cooked muscle sample with that of the cooked volatile fraction. Besides determining the proximate analysis for each raw muscle sample and each sample of cooked-freeze—dried slurry several other chemical analyses were made on these samples for: Creatine/creati- nine, cystine, methionine, sulfhydryl compounds, inorganic sulfide and pH. Determinations were made on the raw muscle samples only for acetoin/diacetyl and inosinic acid concentrations. These results are shown in Table 11. Representative paired samples taken from a 100 g. aliquot of either the raw muscle or the cooked-freeze-dried slurry of each kind of muscle from each class of bird was analyzed. Table 11 shows the protein, fat, ash, and moisture content of the various samples. Attempts to relate differences in composition to the organoleptic characteristics which were noted for the same samples gave negative results. The cooked breast muscle slurry from heavy hens had better flavor than other samples. AcetoinZdiacetyl: Table 11 shows a higher acetoin and diacetyl con- tent and a lower pH value for heavy hen breast muscle. These results agree with those obtained by Pippen ggugl, (1960) and Kazeniac (1961). Acetoin is 2,3 butanedione, which can be oxidized to diacetyl (acetyl- methyl carbinol) by heating it in an oxidizing atmOSphere. Acetoin in .AHHHHHH :HOHnsuooa waHanouaaHmnmaxooo "av Aoaumsa_aau0Hm13oH "Hv .woa nan uanauo was HumooHn can ustaum “waa HaumoOan mumoaHn HaumoOHum ”on can h>oonum mumaaHn no: h>wan1a Havoc camewm .H o.m mm.o no.o m.H o.~ m.¢ 0.0m m.¢a o o m.m a~.o wo.H N.H o.H m.~ m.Hm N.¢H o m m.m HN.o Nw.o o.~ m.~ H.w o.mm N.ma o q m.a am.o mo.H w.H H.H ¢.m m.mw ~.HH o m as is 8.0 3 .2 as new 92 o H m.“ 3 was 2; 3 a; 2 2w 0.3. u H .4 m.~ o~.o Hm.o H.nm o.H ~.m o.o~ N.m H o m.N Hm.o om.o m.qn o.H o.m m.H~ m.m H m ¢.~ Hmwc mm.o o.mn m.H ¢.m o.o~ ~.m H q N.N um.o mm.o ¢.NN o.H N.¢ m.HN n.m H m m.m mm.o om.o o.mm a.H N.m n.NH w.~ H N 3 $5 a}. 2.2 3 9H 0.2 2 H H I434? H. H H .\. .\. H H .02 maHuth aaHaHumoHo oaHuaoHo cum nm< umm nHoHOHm amwOHqu oHnEMm moo: uan031u£wHH can 1h>oan HmHoumoOH EOHH mHHon nHOHnnuan voHHvuouoon 1coxooo one aaomsa woe one umooHn ooHHwnauoonuvoxooo mam noNOHmusmH mo muaoaomaoo .HH manoH -159- .AHHHsHm HHOHnnuwoa 0oHH010NaoHH10oxooo Hov Aoaomsa.aoN0Hm13mH HHV .woa no: HnMHH10 0am HumomHa so: uanH10 “woa Houmo0H1¢ mumoaHn HoumoOHnm mmoH nos m>oon1N mummaHn no: h>oo£1H Havoc manaom .H 0.0 mmH 0.0N o 0 m.0 00m N.NN o m 0.0 0mg c.0m o q N.0 mwm ¢.¢N o m 0.0 00m 0.0N o N N.0 MMN 0.NN u H H.N N.0 0.H 0.0N NN 0.N H 0 0.N 0.0 0.0 0.NNH mm 0.0 H m 0.N 0.0 H.N w.Hm 0H H.m H 0 m.~ m.m H.m 0.HmH .aHH H HHHH mm a.0 H m 0.N H.0 0.N 0.00 mmaH Ho mm N.m H N m.m w.m w.m ~.H0H HHHHHHH Hon Hm n.0 H H .ldme: 0Hom mat .HB_00H\.wd .Ha 00H\.w: .enm. .Bam. .w\.w8 .oz oHaHmoaH MahuoooHn mHHouao< NovaHaw Hthhsmadm oHHGOHSHoz oaaawm AvasoHuaoov moon uanosnuanH 0am 1h>mms HmHaumoOH eon NHHoam SHOHnnumoa 0oHHvuouoaHm rvaxooo 0am oHomsa on 0am unmoHp onHvuaNoonnmoxooo 0am HoNOHmasmH mo mucoaooaoo .HH manna -l60- sufficient concentration imparts a transient buttery-oily aroma to cooked chicken or to chicken broth, whereas, high diacetyl levels impart sour notes (Kazeniac, 1961). Bouthilet (1951) showed that lower pH values enhance chicken flavor. Thus, these two effects, that is a higher ace- toin content and lower pH value for heavy hen breast muscle than the other samples, may contribute to the superior flavor characteristics of breast muscle from the heavy hens. Using model systems, Self gt El. (1963) recovered hydrogen sulfide from the decomposition of cystine or cysteine in air or nitrogen atmos- pheres. Furthenmore, methanethiol, dimethyl sulfide and dimethyl disul- fide formed when methionine was decomposed in air. Only methanethiol formed in a nitrogen atmoSphere. Cystine and methionine as well as glutathione (glutamyl-cysteinyl- glycine) are indigenous to chicken muscle. Sulfhydryl concentration is used as an index to glutathione content. From the concentrations of methionine, cystine and sulfhydryl in Table 11, it is evident that these compounds were present in larger amounts in the samples from light weight hens than in the heavy hen or roaster muscle samples. Studies by Kazeniac (1961) showed that sulfide and nitrogen balance was an important factor in good chicken broth flavor. The principal source of ammonia and carbon dioxide is from Strecker degradation of amino acids (Self ggflgl., 1963). The importance of an active poly—carbonyl compound in such amino acid decompositions was demonstrated by Schgnberg ggugl. (1948, 1952). Acetoin (2,3-butanedione) which is indigenous to chicken muscle is one of these active carbonyls. Therefore, it is con- ceivable that the amount of amino acid decomposition encountered in -16l- heating muscle samples is a function of the active polycarbonyl concen- tration present in the sample. On this assunption, the flavor acceptabi- lity would improve if conditions prevailed favoring ammonia and sulfide balance. This could explain the richer and more "chickeny" flavor of the heavy hen breast and leg muscle samples. It would also clarify the incon- sistency of organoleptic and gas chromatographic results when compared to the concentrations of sulfur compound precursors shown in Table 11. Inosinic Acid: Inosinic acid is an important non-volatile flavor constituent. The concentrations of inosinic acid ranged from 2.3 to 3.3 um/g. in the raw breast muscle samples, and from 2.1 to 2.9 um/g. in the raw leg muscle samples. Heavy hen leg and breast muscle had the highest concentrations of inosinic acid with the roaster muscle samples next highest and light-weight hen muscle samples contained the lowest concen- trations. The higher inosinic acid content in heavy hen breast and leg muscle samples than in the other samples probably contributed to their superior taste. Creatine/creatinine: According to Kazeniac (1961), creatine/creati- nine make up a major part of the non-amino nitrogen compounds in chicken muscle. The creatine/creatinine contents of raw and/or cooked muscle in the present study were higher in light than in dark meat (Table 11). Heavy hen muscle samples had a higher creatine/creatinine content than the other samples (Table 11). Both compounds have a bitter taste and in high concentrations impart a bitter after-taste to the muscle slurry. In his work, Kazeniac (1961) found that after 0.2 g. of creatineJlOO g. of ground light chicken meat and 100 m1. of water were made into a slurry, -l62- canned and processed 30 min. at 120°C, the creatine additive had little effect on the volatiles. The role of creatine/creatinine in flavor is probably additive and synergistic. Due to the high content of creatine/ creatinine in chicken, it is possible that they may serve a useful pur- pose as precursor tag compounds. It is interesting to note that the creatinine concentrations found in the raw muscle samples were almost equal to those of creatine. These results were in contrast to Kazeniac's (1961) observation that creatinine contents were very low if present at all in raw meat extracts. Creatine is converted to creatinine on heating. However, the creatine/creatinine concentrations in the cooked-freeze-dried muscle-slurry were about equal. Furthermore, results (Table 11) did not indicate that any appreciable conversion had occurred. However, some results (Table 11) indicated that losses of both creatine and creatinine occurred after 50 hrs. cooking-distillation. Cystine and Methionine: Methionine concentrations were considerably higher for both raw and cooked samples than cystine concentrations. How- ever, the ratio of concentrations between raw and cooked samples was about the same for methionine as for cystine. The pathway to cystine from glu- tathione via cleavage and oxidation is well known. That of methionine in cooked muscle has not been clearly defined. The role of glutathione and methionine in the production of hydrogen sulfide and methyl mercaptan during irradiation of meat has been reported by Martin 3511;. (1962). Recently Self gtugl. (1963) in their work on potato volatiles have de- fined the roles of cysteine, cystine and methionine in model systems. Again, the production of hydrogen sulfide from cystine and cysteine was -l63- observed as a result of Strecker degradation of these alpha amino acids. Methionine degradation resulted in the formation of methyl mercaptan and dimethyl disulfide. The cystine concentrations of all the samples were approximately equal, but were somewhat higher for the light than for the dark muscle samples (Table 11). Uncooked heavy hen and roaster breast muscle samples had considerably higher methionine concentrations than the corresponding leg muscle samples. The light-weight hen raw leg muscle sample had a ‘markedly higher concentration of methionine than the breast muscle sample. There was no clearly defined trend in the cooked-freeze-dried slurry samples (Table 11). Sulfhydryl Content: Both raw and cooked heavy hen breast muscle samples had a considerably lower sulfhydryl concentration than either the roaster or light-weight hen samples. Yet the sulfhydryl concentra- tion was notably higher in raw heavy hen leg muscle than in either the roaster or light-weight hen.samples. Sulfhydryl is used as an index of glutathione concentration. However, the cystine concentrations do not reflect the results one would eXpect by comparing sulfhydryl to cystine concentrations. Ammoniacal Nitrogen: Ammoniacal nitrogen values were obtained for the raw muscle samples only and varied between 0.05 and 0.06%. The significance of the chemical analyses (Table 11) probably extends beyond the limited data obtained. For example, the absence of sulfide indicates that a loss of ammonium sulfide and other sulfides must have occurred as a result of freeze-drying. It likewise confirms the fact that sulfide formation is the result of degradations and possibly con- -l64- densations and rearrangements which occur when the raw muscle is cooked. Unfortunately, the sulfide determinations were not made on the cooked muscle slurry prior to freeze-drying. Such determinations would have given more significant results for sulfide. Organoleptic Evaluations of the Cooked Broth Samples The purpose of this study was to evaluate the flavor of the cooked broth samples from heavy hen, roaster and light-weight hen breast and leg muscle samples. Table 12 shows that broth from heavy hen breast muscle had the most acceptable, strongest and truest chicken flavor of any of the samples following 50 hrs. of cooking-distillation at 180°F. Broth from heavy hen leg muscle had a good strong flavor, but it was not "chickeny". The odor of the hot broth resembled that of a kettle of hot pork fat while being rendered. With the atypical "oxidation-inhibiting conditions" (Pippen ggflgl., 1958) used in these experiments, the broth from roaster breast muscle was not pleasant smelling or tasting and that from roaster leg muscle was less pleasing. Chemical volatiles that were tentatively identified as chlorine and ammonia and possibly nitrite were liberated from the reagent traps. This phenomena occurred only in the experiments on roaster muscle. The least acceptable of any of the broth samples were those from light-weight hen breast and leg muscle. Objectionable flavors in the breast muscle broth sample described as "oniony" was noted. This may have been due to the presence of higher concentrations of disulfide, sulfide and/or mercaptan than normally occur in chicken broth. -165- Table 12. Organoleptic evaluation of broth samples after 50 hrs. cooking at 180°F under "oxidation-inhibiting conditions", and pH values for original muscle/H20 slurry and cooked broth. Broth ApH sample Slurry Broth Odor and flavor characteristics Heavy hen breast 5.8 6.2 excellent, rich, "chickeny", pleasant muscle after taste that lingered rich, "chickeny" odor taste, strong glutamate-inosinate effect Heavy hen leg 6.1 6.3 bland, sweet, odor like hot pork fat, muscle not "chickeny". - taste, strong gluta- mate-inosinate effect. Roaster breast1 5.8 6.2 bland, no odor of chicken, strong muscle taste but not "chickeny", metallic, soapy, astringent Roaster leg1 6.6 6.8 bone-stock overtone, astringent muscle metallic after taste, taste like wet chicken feathers, nauseating unpleas- ant odor, wet chicken feather odor, grassy or seaweedy, soapy or tallowy, fecal taint. Light-weight hen 6.0 6.3 weak odor of chicken, strong sulfury breast muscle odor, wet chicken feather odor and taste, disagreeable taste (oniony) Light-weight hen 6.2 6.4 sweet, rubbery and burnt odor - astring- leg muscle ent, wet chicken feather taste, disagree- able, gags. fecal taint 1. Capious amounts of chlorine and ammonia were liberated during these tWO runs 0 All of the laboratory glassware surrounding the apparatus and the apparatus itself was covered with a dense coating of ammonium chloride. -166- In making up the raw muscle slurries from the frozen dark muscle samples, a pronounced "beefy" odor was noted on warming the meat-water mixture in a hot-water bath. Similarities in flavor between the red meats, such as beef, pork, lamb and whale, have been reported by Hornstein and Crowe (l960a,b, 1963, 1964). Chicken leg muscle also belongs in this category according to its organoleptic characteristics and the presence of heme compounds. The presence of heme compounds was proven by extract- ing roaster leg muscle with water at 30°F overnight and then taking UV readings on a Beckman dU spectr0photometer on aliquots of the clear solu- tion. Absorption maxima readings of 425 mu.for heme compounds and 675 ‘mu.for metmyoglobin were obtained. Oxymyoglobin was lost as a result of prolonged frozen storage of the muscle samples prior to analysis. The role of heme compounds in meat flavor chemistry is not yet known. Solubility Classification of the Volatile Fraction of_Cooked Chicken: The purpose of this study was to separate the volatile fraction according to solubility characteristics (Cheronis and Entrikin, 1961) by passing the volatile fraction through a cluster of large traps (Figs. 12, 14) containing the reagents previously described. As the volatiles were swept through the radial manifold stream-Splitting system, the odor characteristics of the effluent streams emanating from each of these traps were evaluated. After a run was completed, the con- tents of each trap were subjected to specific qualitative tests. Separate cooking and distillation experiments, lasting for 50 hrs. at 180°F were made on light and dark muscle from heavy hens, roasters and light-weight hens. The organoleptic evaluations are summarized in Table 13. AH00H «0HxHH00m 000 0H0000nov 000H0H>H0 00H000HHH00000100HHH00000 0H0 z 000 H01 N0 - 1 N00 0H0H 00000 0000 .H00o 0ww0 00000H H000 00Hm H000 0H00000m 10H00 «0H0000H «0H00 M000 0000H. H0 «00H00 «0000 0000 00000. 00 H000 00000. H0 «00Hm 003 «0wm0 00000H 000000 M0H 0000M «H0000 «H0000H0 H0000 «0wm0 000000 «H000 H0000 «000H00 1000 «H000H0er0.H0 «0H0an0 «0H00000 000 .03.0H hmmwn .Hm vLar—Hmdw 0H0000H «0H00 «00HHH00 H000 00HwH=0 m00H00 00000H00.H0 «00030 003 «0ww0 00000H 00000E 000000 00000 «H0000 «H0000H0 .00000 «00H00 H000 H0000 000 00H00 «H000H0oea0.H0 «0H0oea0 «0H00000 000 .03.0H 0H00 1HH0V H000 000Hw «+ 1000 00H00 «mw0 000000 .+ 0000 H0000 000N000 «0H00E30 0000 H0000 0000000 H000 0000 0000H «0H0OEE0 «0H0000H 0000000 0000 .H0000 + 0000 H0000 0000 0000 .00000 0000 «+ 0000 H0000 0000 00HHH00 A0H00000000 0H H0000H0 «H000 00030 1000 000H «H000 00m H000 0000H00 00Hm 1000 000H «H000 00w 00H0 0Nm 0Hom00 0HHo00 0H0008 F «0HH00 000 00HH0H00 10H00 «0ww0 00000H 1H00 000 «0000 00000 10H00 H000H0oea0.H0 1000B «0HHHH00>«0003wv w0H H0000QM 6 1 . .0 000H0 00HH0H00 H000 00HH0H00 H000 00HH0H00 0HHE mo 0000 0000000, 0H0000H m00H00 «H000 H000 «H000 0000 0000H 10H0 «00030 «H000 wm0 00000H «0H 0H0008 00030 «H0000 «0000000 0000H 00HHH:0 0HHE 00HHH00 0000:000H0 H000 0000H 00HHH00 100850 «0H000o0 00H00 0000Hn 0000000 H000 «000 00mmH mHsmHom .0H0oeam 0H00 0H00=E H0000 «00000H0 «00030 H000 00000 00HmHom 00HHH00 0000:000H0 .0000 H000 00HHH00 1000 «0000000 0000000 m0H 000 0300m A.000 00HmV .000 £00H0.H0 «0HDMH00 «0H .H000 H0000 «0000000 «000H00 «00:00m H000 0000H00 1000 00on. «0000M «0H0H00 0H0 1&000 0000 0000H w00H00 «0H00 00000 0HDHH00 1000 00H00000 «00030 0H0008 0000Hn 100H «0H0000H «00030 00H3 H000 0000H00 «0H0000H «00030 «0000 00on00 «0000000 000000H 000 0>00m z .>H0 H0 .>Ha 000m NH .>H0 000m 0 .>Ho 000 .>Ho 0H0:em 0omNm 0000H00000o Zm.NHHmoom0z.mm.H Zm.NHHmoom0z Zm.N Hum zN.H 00H0 H0003 0HH00H0> 0H00H000v 000005 00H000HMH000H0 00HHH0=H00 000 00 0000000800 00H000Hm 0HH00H0> AHQHH .HHHHHHHH was mo 00H00=H0>0 0H000Ho00wH0 .MH 0000a -168- Table 13 indicates that a more characteristic chicken odor was ob- tained from heavy hen breast muscle volatiles than from any of the samples tested. There was a prevalence of sulfide and roast beef odor in the effluent streams emanating from the traps which contained alkaline solu- tions as well as from the HCl trap. A stronger roast beef-like odor came from the more alkaline (Div. A2) trap than from the 1.2N HCl (Div. B) trap. This suggests that certain components of the volatile fraction having an additive effect in producing a roast beef-like aroma were ab- sorbed in the alkaline trap but were not absorbed in the HCl trap. Thus, alkaline conditions seemed to favor the maximum development of a roast beef-like odor. According to the solubility classification tables of Cheronis and Entrikin (1961) the alkaline trap (Div. A2) removes sulfonic and sulfinic acids. The HCl trap (Div. B) removes amino thi0phenols and amino sulfonamides. The alkaline trap (Div. A2) also removes mercaptans and thiophenols, aminosulfonamides, amino sulfonic acids, amino thiophenols, sulfonamides and thioamides. Another possible explanation for the stronger beef-like odor from the alkaline trap is that it may remove compounds which mask or nullify the roast beef odor. Panel judges of chicken broth as reported by Peterson (1957) classi- fied the odor as meat broth, sulfide, bready, burnt and ammonia-like. This substantially agrees with the results found in the present study. In addition to the reagents shown in Table 13, other reagents were used including nitrochromic acid, mercuric chloride, mercuric cyanide, bismuth nitrate 2,4-DNP, l-Cl-2,4-DNB, and ceric ammonium nitrate (Table 14). One of the most surprising results was that the volatile effluent Table 14. -168a- from various reagent traps Reagent trap nitrochromic acid Some meaty and nonameaty odor characteristics of effluent stream sweet, toasted marshmallow, beefy, water solution of effluent had a pro- nounced beef taste from breast muscle of heavy hens (heavy dark green ppt) Organoleptic evaluation of volatile fraction effluent aroma pH of effluent trapped in H20 after 40 hrsa 7.0 mercuric chloride sweet, chemical odor from reagent only 6.6 mercuric cyanide chemical odor from reagent, acidic odor 6.9 (blackgpptl bismuth nitrate strong saurkraut odor, acidic, sulfide 4.3 reagent odor, sl. beefy (brown-black ppt.) 2,4-DNP strong sulfide and pronounced roast - beef odor (orange, yellow-brown ppt.) l-cl-2,4-DNB sulfide, beefy (yellow ppt.) - ceric ammonium sweet sulfury chicken odor with roast - nitrate beef background aroma, lead acetate paper gave + test. (yellow-amber ppt.) Div. A1 sweet, fatty, oily (gray ppt.) 6.9 Div. A2 sweet, rubbery, strong chicken aroma 8.9 from heavy hen breast muscle (black ppt.) Div. B glue odor 7.0 Concentrated H2804 water trap sweet, ester, alcohol, aldehydes, acro- lein (black ppt.) roasted peanuts, burnt pOpcorn, pungent strong sulfide and rubber smell 8After 40 hrs. of cooking and distilling, the samples were taken. -169- from the concentrated sulfuric acid trap had a pH value of 7.0 after being trapped in water. In addition, carbonyl removal by 2,4-DNP did not detract from the roast beef aroma emanating from this trap, but rather seemed to enhance it. The sauerkraut odor from the bismuth nitrate rea- gent and the toasted marshmallow aroma from the nitrochromic acid trap were other surprising phenomena that were observed. The effluent stream from nitrochromic acid also contained a high concentration of roast beef- like aroma. The roast beef aroma was obtained from either chicken breast or leg muscle by passing the cooked volatile fraction through an acid oxidizing media, an alkaline media, an acidic solution of DNP, an acidic solution of l-Cl-2,4-DNB or through 1.2N HCl. Results indicated that chicken flavor probably contains the same volatile components as beef plus some additional sulfur containing compounds, phOSpholipid fractions, enols and other reduced forms of compounds, which are responsible for the "chick- eny flavor". The importance of sulfur compounds in meat flavor was also made evi- dent from the fact that mercuric chloride, mercuric cyanide, bismuth nitrate, and concentrated sulfuric acid removed the meaty aroma from the cooked volatile fraction. These findings confirmed the observation by Pippen‘gtngl. (1958) that chicken flavor is associated with the neutral or acidic constituents. Later, Kazeniac (1961) reported that "chickeny" sulfide flavor could be removed from broth by dialysis. Furthermore, separate observations by Bouthilet (1951) and Pippen and Eyring (1957) to the effect that sulfides are important in chicken flavor were also con- firmed. The importance of carbonyls to normal meat flavor is questionable. Under the "oxidation-inhibiting conditions" of cooking-distillation, the carbonyl yield was negligible and the amounts of DNPH formed from the -170- volatiles of several runs were trivial, and amounted to much less than the DNPH obtained from a single run under "normal cooking conditions". Large amounts of sulfur derivatives were obtained under either type of cooking atmosphere, indicating the importance of sulfur compounds to fla- vor. Qualitative Chemical Tests: In order to determine the efficacy of the solubility classification system of Cheronis and Entrikin (1961) as a means of separating various components which exist in the volatile fraction of cooked chicken muscle, a series of qualitative chemical tests were made on the contents of the reagent absorption traps according to the methods outlined by Cheronis and Entrikin (1961), Marbach and Doty (1956), Folkard and Joyce (1963), Siggia and Stahl (1957), and Rittner .2£._$- (1962). The qualitative chemical tests are summarized in Table 15. Table 15 shows the qualitative chemical results obtained by using the solubility classification system augmented by other absorption rea- gent traps. By this method, a total of 20 major classes of compounds and a few Specific compounds were found to be present in cooked chicken volatiles. The compounds identified included nitro-sulfonic acid, ethanol, ethanal, methyl ketones, aldehydes, mono-, di- and polybasic acids, enols, phenols, sulfinic acid(s), ammonia, amines, mercaptans, alcohols, esters, hydrogen sulfide, organic sulfides and disulfides . Chemical tests require concentrations of parts per million or more in order to get a reaction to occur. Furthermore, Guadagni gt .31. (1963) recently demonstrated that aromas may consist of several 33006 0.003 + 8030 8.80. 28 008090 3030 .0000 $5000 03 $300.30 0.0.03 + £33 30000 08 H080 0000000H0E A.000v + 00000 000 0H0xH00 AonoH «0000 000 0000H0zv 00H80H01000H00000 00>H00>HH00 000HMH00 000 00HMH00 00on000 Am0H0v + 1m «000w00H 000H0H0 00080Hm H0mH0m 0H0000 a00H100H0V + 00H00H000H000 00HH0H00 0HHH0m 0H0000H0 a00H0v + 0H00 0H00H000H0H0 0H0000H0 A030H0V + 0H00 0H0oH z mGMUQQOHmflH + mznl+u «NIHOIH 0380 0830 + 0830 0303-333 N... 00H00 0HHH00H00 A00HV + 0000 00HH0H00 0HHH00 1.. 033 $35 + 0330 «380-333 H0 ”0 _ 000He0 A00H01000Hmv + 00H0000HHH0M m 0H0oaa0 A00H0v + 0000 H00000 00HH0H000000 00H000 0H00 0H0HMH00 A00HV + 0000 00HH0H00 0HHH0m 0Ho0000 000 0Ho00 a00H0 00 000V +. .000 owm 0H 00H00H0 A0H0000Ho0 000 0H000H0v 00H00 A00H0v + 00H000 00H00H100000H 0000000H0 a00H130HH00V + 000H 0HH000 0000000 H0000a «H000000 «H000000 A.000 30HH00V + 00H00Ho000 00H000 0H00 0H000H0010H0H0 A.000 030H0100Hv + 0000 00onH000 000HH0m um 00000000 ~00v 000H0 H0002. 0H0000 000008 Ho 000w00m 000000 00000000 Ho 00H0H>H0 .8003 00000. 000 0H00Ho0 mo 000000 000 N0 0000000 1800 H0MH00 How 000 AHomHv 0HxHH00m 000 0H00H00o mo 000000.00H000H0H000H0 00HHH00H00 000 00 00000 0000H00 003000 mo 00H000Hm 0HH00Ho> 000 00 00000 H00H8000 0>H000HH000 mo 00H000m .mH 0HA0H -172- chemical entities, which exist in the fraction at subliminal concentra- tions but are organoleptically recognizable as a result of an additive effect. This means that odors may be perceptible in mixtures when each component is present in subthreshold concentrations. Sulfur Compounds: Attempts to purify the mercuric cyanide and mer- curic chloride derivatives obtained from cooked chicken volatiles by recrystallization from hot water, ethanol and ethyl acetate failed to provide material with a sharp melting point. This is supported by the earlier work of Challenger (1959), who found that mercury forms coordin- ation complexes with sulfur compounds and makes the separation of mixtures of sulfur compounds as mercury derivatives extremely difficult. Results of Sulfur Compound Identifications: The small trap method of testing for thiols, sulfides and disulfides was used as described on p. 85 in the experimental section. A 50 mg. aliquot of mercuric chloride or mercuric cyanide derivative was used for each of the tests. Thiol tests were made on the mercuric cyanide derivative; whereas, sulfide and disulfide tests were made on the mercuric chloride derivative as previ- ously described. Results for Sulfides: Hydrogen sulfide gave a black precipitate when moist lead acetate paper was exposed to the effluent stream emanating from the mercuric chloride trap when acid decomposition of the precipitate started. Sulfide tests were also made on a spot test plate by the method of Marbach and Doty (1956). A few drops of solution contained in the water traps that were used for the disulfide and mercaptan test were pipetted -173- on a spot test plate. Upon the addition of a dr0p or two of N,N-dimethyl- p-phenylenediamine reagent and Reissner solution a deep blue color formed. This gave positive evidence that sulfides were present. Disulfides and Mercaptans: Positive tests were obtained for both disulfides and mercaptans in the form of light yellow precipitates of their reSpective 2,4-dinitr0phenyl thioethers as previously described. Gas Chromatography: In addition to fonming secondary derivatives of mercaptans sulfides and disulfides as their 2,4-dinitr0pheny1thio- ethers, which were identified qualitatively as colored precipitates, gas chromatographic evidence of their identities was also obtained. The amounts of 2,4-dinitrophenylthioethers recovered were not sufficient to permit recrystallizations for melting point determinations. Accordingly, to provide additional evidence of the presence of mercaptans, sulfides and disulfides, 25 mg. aliquots of mercuric chloride and mercuric cyanide derivatives were decomposed together in a small trap as previously des- cribed. This procedure was ad0pted instead of reacting 50 mg. aliquots separately. The washed volatiles from the mercury derivative decomposi- tion trap were collected at -196°C and were then tested by gas chromato- graphy. These results are shown in Fig. 23. Fig. 23 shows that a total of nine peaks were obtained when the mer- curic chloride and mercuric cyanide derivative mixture was decomposed with 8N HCl and the washed volatiles were subjected to gas chromatography. The retention times for these peaks were compared with those obtained using.authentic known samples of the various sulfides, disulfides and mercaptans as shown in Table 16. .00000000> 0000000 000000 00 00>000>0000 00000000 00000008.000 0000000 00000008 00 0000000080000 0000 8000 00000000> 00 .H8 m 00 0000000000 008800w000 00000000800 000000002 0052.: -0 z .0 O. m 0 N m — 0 q u _ com 00008.2 000 0 0 00000 n m 00 0000000000 .0.0.0 omu00000000 800003 .ooomm o0 ooHu0w000 00000000509 3 0000080000 Nmm 00 A 0000000 Nmauw000000 000000 000000 .00 m x .0.0 .00 m\Hu080Hoo 0000000000 08000u00000000 no . no VA -174- .z_2\o on. XB-B .v . xs-L XZE-S In: n ‘ X8-9 .mm .000 LHGBH )Wad ~175- Table 16. Gas chromatographic analysis of volatiles obtained by acid decomposition of a mixture of the mercuric cyanide and mer- curic chloride derivatives of cooked chicken volatiles. Retention times Peak Unknown Known Tentative identity No. min. min. ofgpeaks l l. 2 1.4 dimethyl sulfide 2 1.7 1.8 ethyl mercaptan 3 2.0 2.5 methylethyl sulfide 4 2.5 2.8 n-Pr0py1 mercaptan 5 3.0 3.2 diethyl sulfide 6 3.7 3.8 methyl disulfide 7 3.9 4.1 ethyl-n-propyl sulfide 8 5.0 5.2 Di-n-pr0py1 sulfide 9 5.8 6.0 n-hexyl mercaptan Table 16 and Fig. 23 show that the principal compounds present in the mixture of sulfur derivatives were dimethyl sulfide, ethyl mercaptan, methyl-ethyl sulfide, n-pr0pyl mercaptan, diethyl sulfide, methyl disul- fide, ethyl-n-pr0pyl sulfide, di-n-propyl sulfide and n-hexyl mercaptans. These sulfides, disulfides and mercaptans probably constitute the princi- pal volatile flavor fraction in cooked chicken. Due to the low threshold levels, sulfides, disulfides and mercaptans are easily perceptible organ- oleptically but are not easily obtainable in sufficient amounts for chemical identification. Furthermore, there is prior evidence that sul- fides are important constituents of chicken flavor (Pippen and Eyring, 1957; Bouthilet, 1951; Kazeniac, 1961). The absence of methyl mercaptan may be explained by the fact that its high vapor pressure might have -176- caused it to be lost instantaneously before the reaction trap could be closed following addition of the acid. Possibly a more precise technique would result in the recovery of methyl mercaptan. A small preliminary peak may have been due to methyl mercaptan since it has a retention time of about one-half minute at 100°C on the Apiezon L column. Functional Group Analysis of Cooked Chicken Volatiles from Light and Dark Muscle After 50 hrs. Cookingiand Distillation Under "Oxidation- inhibiting7Conditions" The purpose of this study was to tentatively identify the compounds present in the cooked volatile fractions from light and dark chicken muscle. As previously described, the volatiles were passed through Specific reagents according to the methods of Walsh and Merritt (1960), Bassette 23 El: (1962) and Hoff and Feit (1963). A trap reaction tech- nique was devised and used as previously described (Fig. 15). A gas chromatogram of the total cooked chicken volatiles from heavy hen leg muscle is shown in Fig. 24. Chromatograms obtained after pass- ing the total cooked chicken volatiles from heavy hen leg muscle through traps containing specific functional group reagents (Fig. 15) were com- pared with control chromatograms. Results of these comparisons are summarized in Fig. 25. The reagents used in obtaining the analytical results were as follows: (1) acidic hydroxyl-amine, (2) basic hydroxy- lamine, (3) a saturated solution of potassium penmanganate, (4) mercuric chloride (3% w/v), (5) mercuric cyanide (4% w/v), (6) acetic anhydride, (7) sodium nitrite, (8) sodium borohydride, (9) 1.2N HCl, (10) concen- trated H2804, and (11) deionized distilled H20 (control). Each of the .000008 w00 000 00000 8000 00000000> 000000 00000 00 .08 N 00 0000000000 0008000m000 00000000800 000000002 .0N .w00 . mmhaziuuzz. o. m 0 h m m t n N . o w 0 0 0 0 0 0 0 0 0 O .258 .0. . con 00002 0 05 0 N 0 00000 I w 00 0000000000 .0.0.0 on .. 00000000 0500000 Uoomw 00 OO0 u 0w000 00000000800. I. A n v .7 n d 00 00 . 3 3 0000080000 0.3 00 .0 0000000 030000000 C. V 000000 000000 .00 m x .0.0 .00 02:08:30 I. WM X .h 0000000000 0800000000000 6 u: _ . 9 I. V H N a x c B . x I. m . 70 X I O I I 9 1 n l ... 0 ... x 9 I | x m ”an a 3 X . x w. 0... 1 0 n... x 0.. me. XX 0.3 - 0 70X X m 0 0 X ~178- QQIKWXx 058,0 0Y>\QRU>SK Y 0% .0 VDMDV< 0.00 >QI \SvQI $90K 900000.003 QNXOOQ k0 3.000??? QIQVQDQKYEOQID (206 0‘0 LEG/900360 hm .910 402 QNXQQQ KG MSWXNXSV‘ bkaTQbQKVEQQID M66 KO M§WQQV.:._.2uoH m>..—.<.P2u._. macaw £29523... 6%.: 4 INS 3.2 0242 «qu ... ..U 3!: ON 6% \NMVWQNQ kbwkkw WKEMQYMQ Q3006 “EUR b>3k -182- A comparison was made of retention volumes shown in Figs. 24 and 25 and Figs. 26 and 27 for the total cooked volatile fractions from leg and breast muscle, respectively. The cooked volatiles from heavy hen leg muscle contained a larger proportion of sulfur compounds than car- bonyl compounds. Heavy hen breast muscle contained a larger proportion of carbonyls than sulfur compounds as judged by gas chromatography. How- ever, this was not confirmed by total derivative yields. Furthermore, Guadagni gt El- (1963) found that the subliminal concentrations of sulfur compounds are much lower than those of carbonyl compounds. Accordingly, the sulfur compounds would appear to dominate the overall flavor of the total cooked volatile fraction from both breast muscle and leg muscle, but to a somewhat lesser extent for the former. The presence of hydrogen sulfide and ammonium sulfide in typical volatiles from cooked chicken together with a higher carbonyl content resulting from oxidation under "normal cooking conditions" tended to increase the sulfide effect on overall flavor. No attempts were made to study the flavor effect of add- ing known compounds corresponding to those tentatively identified, in the prOportions indicated by the comparative retention volunes. It is likely that the nunber of peaks found for total cooked volatiles of breast and leg muscle in the present study represent only a fraction of the true number of compounds present. Further studies using capillary columns and electron capture procedures are needed to probe deeper into the problem. Neither were the acidic constituents detectable with the gas chroma- tographic procedures used in this study. Evidence for their presence was revealed, however, by the chemical identification system. -183- The complexity of the problem of chicken flavor has been well stated by Kazeniac (1961), and the present study confirms those observations. Quite recently Hornstein and Crowe (1964) stated in a review on meat flavor that the high boiling fraction contains the compound(s) responsi- ble for the meaty aroma. This was also confirmed in the present study. Organoleptic Evaluation of the Residual Effluents Contained in the Liquid Nitrogen Traps Following Gas Chromatography: The purpose of this study was to determine the organoleptic charac- teristics and pH of the residual contents of the sample traps after removal of Specific functional groups and gas chromatographic tests on the effluent.‘ The dry ice ethanol trap (-80°C) gave only three or four peaks for total volatiles in the low boiling range 1oo-175°c (0 to 5 min.) but peaks typical of those from the liquid nitrogen trap were eltmed from the Apiezon L colunn in the ZOO-250°C range (7 to 10 min.). A pronounced sweet and roast beef odor was noted in the -80°C trap and a white preci- pitate was present on the trap inlet tube at that temperature. Upon warming the precipitate, it disappeared and may have been solid carbon dioxide. However, it was not identified. When the contents of the trap were treated with 2-3 ml. deionized-distilled H20, the pH was 8.0 or slightly over. The water solution of the volatiles residue tasted "meaty". Table 17 summarizes the pH and odor characteristics of the liquid nitrogen traps {-196°C). Results for the effluent stream with no treat- ment (control) and treatment by several typical functional group reagents Table 17. -184- Organoleptic characteristics and pH values of liquid nitrogen trap, contents after passage of the total cooked volatile fractions of light or dark muscle through the reagent trap and chromatographic analysis Trap treatment ApH Odor and taste characteristics none 9.0 buttery, sulfury, petroleum odor, meaty taste basic hydroxyl-amine 10.0 ammonia odor, bitter, disagreeable taste acidic hydroxyl-amine 8.4 sweet-strong roast beef odor, beefy tas te acetic anhydride 5.0 sweet, beefy odor and taste 1.2N HCl roast beef odor, sl. bitter, but meaty flavor mercuric perchlorate 8.0 carbonyl, ester odor-not meaty, bitter taste 2,4-DNP - strong-sweet roast beef odor, aromatic undertone, meaty taste bismuth nitrate - sweet, sulfide odor, sauerkraut taste, acidic 2.5N NaHCO3::2.5N NaOH - sulfide, esters, aromatic odor, bitter, astringent, rubbery taste 1.5N NaHCO3::2.5N NaOH - sweet, fatty, ester odor, grassy, bitter astringent taste permanganate - oxidized fat, acetone, sulfide odor, burnt bitter taste sodium borohydride - strong mercaptan odor, disagreeable putrid flavor mercuric cyanide 9.0 sweet and carbonyl odor, bitter, astrin- gent, sweet, oily, bitter taste mercuric chloride 9.0 sweet and carbonyl odor, bitter, astrin- sodium nitrite gent, bitter taste nitrite odor, bitter, astringent chemi- cal flavor -185- after the effluent passed from the reagent trap through the -80°C trap and into the -196°C trap are included. All tests were made on the -196°C trap only as random chromatogrqahic tests on the contents of the -80°C trap showed only a few peaks as previously described. Table 17 shows the effects of various functional group reagents on the odor and flavor of total cooked chicken volatiles. Carbonyl removal or reduction enhanced the meat flavor. An increase in mercaptans by the reduction of disulfides had an adverse effect on odor and taste. Oxi- dation of carbonyls and sulfides caused a burnt flavor and/or removed the meaty flavor component. Acidic reagents enhanced the meaty or roast beef-like flavor character. Basic reagents imparted a disagreeable bitter or rubbery flavor. Sulfide elimination caused a loss of the meaty taste and left a sweet, oily, bitter, burnt taste, which was probably due to esters, acids and carbonyls. Perhaps some sulfur esters exist in the volatile fraction, which have not been identified. These would probably be found in the medium or higher boiling fractions. In a private communication, Kazeniac (1963) advised that Sasin (1962) prepared a series of sulfur esters. As a result of a private communication, Sasin (1964) advised that these thio- esters had been sent out to another laboratory for Spectrographic analy- sis and were not available. Although the mode of preparation of these thiol esters of fatty acids is quite simple, this line of investigation was not pursued further. Since the beef flavor of the volatile fraction was enhanced by the acetic anhydride treatment, this enhancement may have been due either to possible thiol ester formation, acetylation of amines or removal of basic sulfides. However, these effects were not studied. -186- Model Systems The purpose of the model system studies was to attempt to relate chicken flavor to glutathione and to some simple compounds which are known to exist in chicken broth. These precursors and compounds included car- bamyl phOSphate, lactic acid, ammonium sulfide, glutathione, cystine and Inethionine. Organoleptic and gas chromatographic evaluations of three separate preparations were made. In the first system, hydrogen sulfide was evolved by adding lactic acid to sodium sulfide. When carbamyl phosphate was added to this mix- ture and the mixture was warmed by a water bath to 180°F, an odor similar to chicken volatiles resulted. In the second system sodium sulfide, lactic acid, carbamyl phosphate, and 2, 3—butanedione were made up in solution and ammoniun hydroxide was added to give a pH reading of 6.9. The mixture when warmed to 60°C gave only 3 small chromatographic peaks as compared to 7 small peaks that were obtained from the first system. Again a chicken-like flavor resulted. The importance of ammoniun lactate as a component of meat flavor was re- cently reported by Hornstein (1964). In the third system, which represented the concluding part of the model study, glutathione was added to a mixture of lactic acid, carbamyl phosphate and 2,3-butanedione. After adjusting the pH to 7.9, 3 ml. of the mixture were put into a small reagent trap (Fig. 12) and heated to 180°F in a water bath. The volatiles were trapped in a small liquid nitrogen trap (Fig. 12) and were subjected to gas chromatography. The chromatographic results are shown in Fig. 28 and Table 18. .Eoumum Hence a mo muaoa0d§o umnuo paw onoflsumuoaw @0968 no ooaugmoaaooon msoosvm 80am $5.339, mo .1: m mo nowumuuag pofifimumoun ousumuomfimu Moonwaaoz .mm .wfim mmhazilmzrr o. m o h w 0 ¢ n N . o u q a a _ . . a J a . O ... com Homo: 2 new m 1 N H owoum .. w um Hougaouu< ..m.m.e on .. ousmmoud E130: ooonu ou ooH u mwaau manuaHoQEmH Swan: 2 £3383 N8 so a conga... aflbfixoum 9 ..n 4. H9300 poaqoo .um m x 5.0 .a.“ whugaoo . 3 oowuafiaoa madam u Houoouon m C. W 07.. ... 1 H 1.. x c u 9 m. n n .. o .. h n c -188- Table 18. Gas chromatographic analysis of a model system in which gluta- thione was decomposed in an aqueous media containing other components of chicken muscle. Retention time Peak unknown Known Tentative identity No. min. min. of peaks 1 0.7 0.7 methyl mercaptan 2 1.2 1.4 dimethylsulfide 3 1.6 . 1.6 methylamine 4 1.8 1.8 acetaldehyde 5 2.2 2.4 methyl disulfide 6 2.8 2.8 2,3-butanedione acetylmethyl carbinol 7 ethyl n-prOpyl sulfide 8 4.0 3.8 ethyl disulfide Fig. 28 shows that a total of 8 well-defined peaks are obtained. These were tentatively identified as 1 mercaptan, 2 disulfides and 3 sul- fides plus 2,3-butanedione acetaldehyde and acetylmethyl-carbinol (Table 18). These results indicate that glutathione may be the source of methyl mercaptan, dimethyl sulfide, methylamine, acetaldehyde, methyl disulfide, ethyl-n-prOpyl sulfide and ethyl disulfide, which were previously shown to be indigenous to the total cooked volatile fraction of chicken leg muscle. With the exception of methyl mercaptan, all these compounds were likewise indigenous to the total cooked volatile fraction of breast muscle. The final pH of the reaction products mixture was 4.7 as compared to the initial pH value of 7.9. The predominant taste was sulfide, and the flavor and odor resembled chicken broth. A mild mercaptan odor was -189- evident. However, the final product did not have a complete chicken broth flavor. A total of 50 ug of inosinic acid in a 1 m1. aliquot was added to 2 m1. of the reaction product solution. The pH was then ad- justed to 5.6 with 0.01N NaOH and l/4 mg. of creatine was added. These additives, in turn, improved the flavor but the product as before did not constitute a true chicken broth flavor. Chemical Spot Tests: The sodiun hypoiodite test (Cheronis and En- trikin, 1961) gave a positive reaction for ethanal (acetaldehyde) in the form of a yellow precipitate. A few draps of the unknown solution on a Spot plate reacted with ferricyanide reagent to give a green-blue color, which indicated that an amine was present. The addition of sodiun boro- hydride to a few drOps of the solution gave a mercaptan odor which indi- cated that disulfide(s) was/were present. A positive methylene blue test for sulfides was obtained with p-phenylenediomine reagent (Marbach and Doty, 1956). A positive test for mercaptans was obtained by adding 1 ml. of ethanolic NaOH and a few drops of l-Cl-2,4-DNB to the solution which remained in the tube. A light yellow precipitate of the thioether derivative(s) separated upon cooling the tube in ice. Tests on the effluent stream from the liquid nitrogen trap gave a positive hydrogen sulfide test (black precipitate) with lead acetate paper. The number and complexity of the reactions which would be required to produce these seven products from the original preparation envisages many reactions that are inexplicable by the Strecker degradation of amino acids alone. Studies on glutathione's decomposition in aqueous media have previously been reported by Kendall 2; 31. (1930) and Mason -190- (1931). Cleavage of glutathione to cysteinyl-glycine and pyrollidone carboxylic acid occurred by incubating glutathione at 37-62°C. Judging from the reaction products obtained, further cleavage to cysteine, gly- cine and glutamic acid probably occurred at the higher temperature (180°F). The formation of acetylmethylcarbinol and methanethiol indicated that reduction reactions must have occurred in addition to Strecker degradation of the amino acids. There were also probably numerous rearrangements. There is no evidence to support the finding of methanethiol as a product of cysteine and cystine. In their model systems, Self ggbgl, (1963) obtained methanethiol from.methionine, but not from either cystine or cysteine. Furthermore, their investigations showed that no sulfides or disulfides were formed from cystine or cysteine in either a nitrogen or oxygen atmosphere. They found that methionine yielded dimethyl sul- fide in an oxygen mediun, but not when nitrogen was used. Perhaps the presence of glycine glutamic acid, glutathione and the other compounds that were present with cystine were responsible for the mercaptans, sul- fides and disulfides found in the present study. These compounds may also possibly result from interactions with hydrogen sulfide, cysteine and cystine in the vapor state. It is also conceivable that 2,3-butanedione may cause a different reaction pattern with cysteine and cystine than that observed by Self ggugl. (1963) with dehydroascorbic acid. Further work is needed to verify the tentative identifications that were made in the present study of glutathione in a model system. Gluta- thione's role in hydrogen sulfide production upon heating chicken flesh as first reported by Sadikov 33 a1. (1934) was substantiated by this study. -191- Glutathione also fits Bouthilet's (1951) description of a labile sulfur compound which together with a fatty acid-like material is reSponsible for cooked chicken flavor. A postulation was made by Bouthilet (1951) concerning a steam distillable, weakly acidic constituent of cooked chicken meat that had its origin in the glutathione of chicken muscle. He found that this constituent imparted a strong flavor to chicken broth in the pH range of l to 5. In the present study this characteristic was demon- strated by glutathione. Upon decomposition of glutathione, the pH of the original media dropped from 7.9 to 4.7. The association of desulfuration and sulfide losses with a reduction and final complete loss of "chickeny" flavor in broth as demonstrated by Pippen and Eyring (1957) was also ob- served in this study of glutathione. Addition of glutathione to a stan- dard broth solution prior to heating resulted in a more "chickeny" flavor according to Kazeniac (1961) and his finding was further substantiated by the results obtained in the present study. Furthermore, the association of chicken flavor with neutral or acidic constituents by Pippen and Eyring (1957) was likewise confirmed. SUMMARY AND CONCLUSIONS Chemical components from cooked light and dark chicken muscle were identified using gas, column, thin-layer and paper chromatography, UV and I-R absorption Spectra, functional group analysis by a trap-reaction technique and a solubility classification method. A total of 30 peaks were obtained by gas chromatographic separation of the total cooked volatile fraction from leg muscle and 25 peaks from breast muscle. Identifications were made by passing the volatiles through traps containing Specific functional group reagents or by comparing chromatograms of the total cooked volatile fractions with those obtained using authentic compounds as reference standards. Constituents identified in the total cooked volatile fraction from leg muscle were ethane, propane, methyl mercaptan, acetone, methanol, dimethyl sulfide, methyl-ethyl sulfide, methyl-amine, diethyl sulfide, ethanol, acetaldehyde, methyl-isopropyl sulfide, 2,3-butanedione, methyl disulfide, acetoin, ethyl-n-propyl sulfide, ethyldisulfide, ethyl mercap- tan (symmetrical trithiane), diprOpyl sulfide, n-prOpyl mercaptan, n- hexanal, 2,4-pentanedione, iso-amyl alcohol, n-amyl alcohol, n-heptanal, ethanol-amine, n-hexanal, 2-heptanone and n-heptanol for a total of 29 organic compounds. Constituents isolated from the total cooked volatile fraction of breast muscle were ethane, propane, acetone, methanol, dimethyl sulfide, methylamine, methyl formate, dimethyl sulfide, ethanol, acetaldehyde, 2,3-butanedione, methyl disulfide, acetoin, n-pentanal, ethyl-n-prOpyl sulfide, iso-butanol, ethyl disulfide, n-butanol, dipropyl sulfide, -192- ~193- n-prOpyl mercaptan, n-hexanal, 2,4-pentanedione, n-heptanal, 2-heptanone and n-heptanol giving a total of 25 organic compounds. Inorganic compounds identified in the cooked volatile fractions by cryogenic trapping of volatiles at -196°C followed by cryogenic distilla- tion at -l40°C and gas chromatography were hydrogen sulfide, and carbon dioxide. Ethane and propane were also identified. Ammonia and ammonium sulfide(s), hydrogen sulfide and carbon dioxide were identified using chemical and physical methods. The presence of carbonyl sulfide was indicated by odor tests, but further identification was not made. Gas chromatographic studies showed that whole carcasses from 20 ‘month old hens contained the same volatile components as 12 week old birds of identical origin and raised on the same ration. Intestinal contents from 20 month old hens had the same volatile organic constituents as meat from the same birds; namely, carbonyls, sulfides, disulfides and mercaptans, but in lesser amounts. A steam distillable phosphatidyl lipid fraction was isolated from chicken breast meat. By subjecting this fraction to thin-layer chroma- tography, two compounds were identified; namely, cardiolipin and either phosphatidyl inositol or phOSphatidic acid. The freshly prepared dis- tillate had a several-fold concentration of true chicken odor and was called "chicken essence". A total of 8 mono- and 1 poly-carbonyl were isolated from cooked chicken volatiles by preparing 2,4-DNPHS and identifying the derivatives by column and paper chromatography, UV and I-R absorption spectra, melting points, and mixed melting points using authentic samples, and microscopic examination. Acetaldehyde, acetone, n-propanal, methyl- ethyl ketone, n-butanal, n-pentanal, n-hexanal, n-octanal and diacetyl -194- were identified. A 2,4-dien-l-a1 was also present, but no authentic sample was available for comparison. Gas chromatographic analysis of volatiles obtained by the acid decomposition of organomercuric derivatives of cooked chicken volatiles resulted in the identification of dimethyl sulfide, ethyl mercaptan, methylethyl sulfide, n-pr0pyl mercaptan, diethyl sulfide, methyl disul- fide, ethyl-n-prOpyl sulfide,monnm mummoun no; h>mosua "mnoo oaqamm .H 0 mm ow N 0 0H ¢.nH 0.0 00.0 n.- 0.0a on ~.~N Nu.~ 0.MH «.ma m.m m0.0 w.m~ 0.0a om 0.0m 00.m N.mH n.0H m.m mm.0 m.a~ 0.NH od . ¢.¢~ «0.N ~.~H m.mH m.0 no.0 w.m~ 0.0a on ”D 0.¢~ 0¢.~ m.mH 0.m~ 0.0 00.0 H.0N n.0H ow nfi w.n~ w~.~ m.mH 0.mH m.w mw.0 m.H~ «.ma 0H 0.5 0n.0 0.0a m.mH m.~ mN.0 m.m~ 0.NH an 0.5 0n.0 0.0a «.ma N.N hm.0 H.5N 0.0a an 0.0 00.0 «.ma 0.0a m.~ mN.0 m.n~ 0.NH H0 m.m Hm.0 ¢.HH 0.NH 0.N 0~.0 0.0m 0.0a Hm H.m Hm.0 ~.0H N.HH 0.N 0N.0 N.¢N 0.NH MN n.0 no.0 0.ma 0.0H n.~ m~.0 m.m~ 0.0a Ha oHeEMm madman murmhaouphn mummhwoumhs camera camera ouwmhaoumhs mummhaounhn H.oz Hmawwwuo Hmdfiwwuo pmuaawp nonnawn HmGHmHuo Hmaflwwuo nonnawp nousawp magamm a“ .w\.wa aw N .HE\.w= .HE\mQM Ga .w\.wa aw N .HE\ws .HE\mQM 25000. .H8 25000. .H8 aoaumawEQouon oofisognuoeua coaumoaeHoump megaphonq nosuoa HmoaonOMnoaoae Awamav m.HHmum 0cm commando: he moon unwwo3uunwaa 0am uh>mon «muoummou eoum huuaam numoa 0o¢n0:ouomum:0oxooo new nououmuamu mo uaouaoo ocHooHSumanq 0am oawumzouq mo mcoaumaaauouon .N manna -216- Table 3. Composition of Henderson and Snell's (1948) single medium for determining 14 different amino acids with the apprOpriate amino acid omitted. Composition of Medium Amount/ Amount/ Component 100 m1. Component 100 m1. Glucose 2 g. Ca Pantothenate 100 ug. Sodium Citrate (U.S.P. X111) 2 g. Niacin 100 ug. Sodium acetate (anhydrous) 0.1 g. p-aminobenzoic acid 20 ug. Ammonium Chloride 0.3 g. Biotin 1 ug. KZHP04 0.5 g. Folic Acid 1 ug. Salts C 2 ml. D-L-Alanine 100 mg. Adenine-SO4 1 mg. D-L-Aspartic Acid 100 mg. Guanine-HCl 1 mg. L-Glutamic Acid 100 mg. Uracil 1 mg. L-Arginine-HCl 20 mg. Xanthine 1 mg. L-Lycine-HC1:H20 20 mg. Thiamine 100 ug. Other Amino Acids L-forms 10 mg. ea. Riboflavin 100 ug. or DL-forms 20 mg. ea. Pyridoxal 20 ug. Notes: 1) Magnesium and manganese ions from salts C overcome citrate toxicity. 2) Use a 2% sodium citrate buffer for Leuconostoc mesenteroides. 3) Hydrolyze the protein containing material for 5 hrs. at 15 lbs. pressure with 3N HC1. -217- Table 4. Colorimetric determination of percent transmission of cystine in diluted hydrolysate samples of raw-frozen muscle and cooked- freeze-dried meat slurry from roasters, heavy- and light- weight hens. Standard Curve A standard solution containing 5 ug of cystine was used. m1. standard % Transmission of solution duplicate tubes 0.0 12 12 0.2 55 59 0.4 78 78 0.6 89 93 0.8 102 95 1.0 102 65 The % transmission values were plotted against ug of cystine on the standard curve (Fig. 5a). Each hydrolysate sample was diluted 1-4000 and various amounts added to incubation tubes. The % transmission obtained were as follows: m1. of diluted sample 9:; 9_,_4 _o_._6 SL8. _1_._g Sample No.1 1 33 50 61 72 77 2 34 52 62 71 78 3 26 43 54 64 71 4 27 44 55 63 68 5 35 53 66 73 79 6 34 55 61 71 81 7 33 50 64 68 80 8 31 49 63 69 78 9 34 55 66 74 84 1o 38 55 68 72 76 11 37 53 65 75 82 12 38 54 66 74 82 Using the standard curve (Fig. 5a), ‘Milligrams of cystine per gram of sample were calculated from % transmission obtained. 1. Sample code: l-heavy hen breast; 2-heavy hen leg; 3-roaster breast; 4-roaster leg; 5-1ight hen breast; 6-1ight hen leg. (r: raw-frozen muscle) (c: cooked-freeze-dried meat- broth slurry). -218- Table 5. Colorimetric determination of percent transmission of methio- nine in diluted hydrolysate samples from raw-frozen muscle and cooked-freeze-dried meat-slurry from roasters, heavy- and light-weight hens. Standard Curve A standard solution containing lOug of methionine was used. m1. standard solution % transmission of duplicate tubes 0.0 15 18 0.2 42 41 0.4 68 67 0.6 88 91 0.8 100 103 1.0 107 107 The % transmission values were plotted againstug of methionine on the standard curve (Fig. 6a). Each hydrolysate sample was diluted 1-4500 and various amounts added to the incubation tubes. The % transmissions obtained were as follows: m1. of diluted sample g,_2_ o_.4_ $2 _q_._§ r,_o_ Sample No.1 1r 31 49. 64 77 89 Zr 31 50 62 78 86 3r 29 44 58 71 83 4r 30 43 61 70 80 Sr 31 46 64 107 85 6r 30 48 63 91 89 1c 30 43 59 70 82 2c 29 43 58 72 81 3c 36 52 68 84 93 4c 35 53 67 82 92 5c 31 50 63 79 90 6c 32 50 64 78 86 Using the standard curve (Fig. 6a), milligrams of methionine per gram of sample were calculated from % transmission obtained. 1. Sample code: 1-heavy hen breast; 2-heavy hen leg; 3-roaster breast; 4-roaster leg; 5-light hen breast, 6-1ight hen leg. (r: raw-frozen muscle) (c: cooked-freeze-dried meat- broth slurry). 316 at: 0'0 '1“ 5k 'ch ‘Jk‘CAMDENLNEWJERSEY‘Jk 'k 910 380 516 510 * Novanber l, 1963 Mr. L. J. Minor Meat Laboratory Food Science Department Michigan State University East Lansing, Michigan Dear Mr. Minor: It was a pleasure to talk to you and loam that you are doing research work on chicken flavor. I hope that the following consents may be helpful to you on this problem. The major non-protein nitrogen compounds of chicken my be grouped as follows: 1. Creatine/creatinine 2. Amino acids/peptides . Purine derivatives . Carnitine . Ammonium salts (probably phosphates and lactate) - 6. Possibly N-bases of phospholipids (e.g., choline, ethanolamine) These are several procedures for the detensination of creatine/ creatinine, of which, the following would seem to be the best: 1. The alkaline picrate method (Methods of Anal sis, Association of the Official Agricultural Chemists, p.6389. 1955)). Also see T. Wood and A. E. Bender, Biochem. J. _‘L, 36 (195?). 2. The alpha-naphthol/diacetyl color reaction (D. R. Anderson .11. Williams, 8. x. Krise and n. n. Dowben, 81661:... J. £1, 258 (1957)). 3. The Sullivan-Irreverre test with greater specificity (14.x Sullivan and F. Irreverre, J. Biol. Chem. 232, 530, (1953)). Amine acids/peptides can be analysed colorimetrically with ninhydrin by may or the various chromatographic methods or directly. There are several good procedures for the determination of purine derivatives, using ion eaehenge resins. Thee compounds also can be estimated readily by thin-layer chromatography. The spots can be .leeated with ultraviolet light, scraped from the glass plate, dissolved and measured by their ultraviolet absorption. Oar early work indicated -220- Campbell Soup Company Page Two Novanber l, 1963 that total purines with no differentiation as to type could be approxinnted by direct measurement of the absorption of the extract or the broth. Dr. N. R. Jones at the Torry Research Station in Aberdeen, Scotland, has developed a method for determining purines enzymically with xanthine oxidase. The procedure is designed to measure the freshness of fish, since it is known that the nucleotides are at their higiest level in fresh fish decreasing as the fish ages. Carnitine can be determined either as the periodide (J. 3. Wall, D. C. Christianson, R. J. Dimler and F. R. Senti, Anal. Chem. 1;, 870, (1960)) or with bismuth triiodide (A. E. Bender, '1‘. Wood and J. A. Palgrave, J. Sci. Food Agr. 2, 812, (1958)). Incidmtly, the latter paper contains many useful procedures for the analysis of meat extracts. During our telephone conversation, you expressed interest in thiolesters. Dr. Sasin at the Drexel Institute in Philadelphia, Penna... has prepared thiolesters of long chain fatty acids and possibly might have samples available. Some of the references to his work are: 1. G. S. Sasin, J. Am. Oil Chemists' Soc. %, 14.88 (1962). 2. G. s. Sasin, et al., J. Org. Chem. g, 22 (1959). 3. G. S. Sasin, et al., J. Org. Chem. g_1_, 852 (1956). Details of the sulfide procedure we use are attached. If you have any other questions, I shall be glad to try to answer them. Good luck to you on chicken flavor. Best regards to Dr. Pearson. Sincerely yours, .) ";0,V'I\(r ' / S. J. Kazmiac Inorganic Sulfide Determination A. E. Sands, H. A. Grafins, H. w. Wainwrigm and M. H. Wilson, Report of Investigations, 3.1. [1.51.], United States Department of the Interior - Bureau of Mines, Septanber 1914.9). Reaggt s : l. Zinc acetate solution - dissolve 20 g. zinc (083000) -2320 in 500 ml. water. Add a few drops of glacial acetic ac d to prevent precipitation of insoluble zinc salt s. Precipitation may occur, if ammonia is present at high levels in solution being analyzed, but this does not seem to affect the remlts. 2. l,N-Dimethyl-p-phenylmediamine reagmt - dissolve 150 m. of l,N-dimethyl-p-phenylenediamim dihydrochloride in 150 m1. sulfuric acid (two volumes of concentrated sulfuric acid added to sone volume of water). 3. Ferric chloride reagent - dissolve 2.7 g. Feel 4320 in 100 ml. hydrochloric acid (one volume of concentrated gydrochloric acid in one volume of water). - Pro cegre :2 Add 25 ml. of sample containing from 1-50 ppm sulfide to 25 ml. zinc acetate solution. Next add 5 ml. of the diamine reagent and stir. Then add 1 ml. of the ferric chloride solution, stir and allow blue color to develop for 15 minutes before reading absorbance at 715 mu. The curves attached are taken from the paper by Sands Gt ‘1 lbOVOe We developed our own curve using sodium sulfide standard solutims. It checked closely with the reported values. ~222- Method for preparation of derivatives: (2,4 DNPH) I. Preparing 2,4 DNPH solution: -to 0.4 g of 2,4-Dinitr0phenylhydrazine in a 25 m1 Erlenmeyer flask is added 2 m1. of concentrated H280 -3 m1. of H 0 is added dropwise with swirling or stirring until solution is comp ete -10 m1 of 95% ethanol is added to this solution II. Preparing Carbonyl solution in ethanol: dissolve 0.5 g of the compound in 20 m1 of 95% ethanol III. Makinggderivative: the freshly prepared 2,4-DNPH in solution is added, and the result- ing mixture is allowed to stand at room temperature, crystallization of the hydrazone usually occurs within 5-10 min. If no precipitate forms - mixture is allowed to stand overnight.