EFFEFT 0F COGKE'JG PRC» CEDURE ON THE

“-1

RAVDR ’“Z‘MPOE éfiils’S 0! BE?

I“: CHEESA ' ESTATE UN? EcEz.‘ :aS: iEEiE

Ame- :m‘edersen
7.96.6

 

LIBR A E. 1."

Michigan State
University

 

 

 

ABSTRACT
EFFECT OF COOKING PROCEDURE ON THE FLAVOR.COMPONENTS 0F BEEF

by Anne Sanderson

The chemical components volatilized when beef was cooked in water
were compared and contrasted with those obtained on cooking beef in fat.
Equal quantities of fat-free beef were cooked by the two procedures for
eight hours in an inert atmOSphere. The flavor volatiles produced were
swept through a series of reagent traps with nitrogen gas. Special em-
phasis was placed on the carbonyl and sulfur compounds in the flavor
volatiles produced by both methods of cooking.

The carbonyl compounds in the flavor volatiles were precipitated
as their 2,4-dinitr0pheny1hydrazones (2,4-DNPS). The yield of 2,4-DNPS
from beef cooked in fat was three times as large as that from beef cooked
in water, with mean values of 0.17 g. and 0.05 g., respectively. An
equivalent amount of beef fat cooked alone yielded less than 0.01 g.

The aldehydes and ketones were released from their 2,4-DNPS by ex-
change with levulinic acid. The carbonyl compounds from beef cooked in
both fat and water had a sweet and faintly caramel aroma, which was indi-
cative of their role in producing a cooked beef aroma. When chromato-
graphed, the carbonyl compounds from beef cooked in water and fat
separated into 14 peaks. Butanal, 2-butanone and the 5-carbon carbonyls
were identified as the major components of the carbonyl compounds volati-
lized by both cooking procedures. Smaller amounts of formaldehyde,
acetaldehyde, prepanal, acetone, 2-methy1 propanal and 3-methy1 butanal

were shown to be present. Five smaller peaks, too small to identify,

Anne Sanderson

were also obtained. In addition there was one small peak which was found
to be a breakdown product of levulinic acid.

Chemical tests for polycarbonyl compounds, such as 2,3-butanedione,
with 2% alcoholic potassium hydroxide showed that polycarbonyls were not
present in the 2,4-DNPS derivatives from the flavor volatiles produced
by both methods of cooking. Other chemical tests for confirmatory iden-
tification of the aldehydes and ketones proved unsatisfactory.

Although beef cooked in fat yielded the same number and type of
aldehydes and ketones as beef cooked in water, the quantities of carbonyl
compounds volatilized varied with cooking procedure. Beef cooked in
water consistently yielded less 2-butanone than butanal, whereas, beef
cooked in fat produced more 2-butanone than butanal. Since 2-butanone
was the carbonyl compound found in greatest proportion on heating fat
alone, it appears that the increase in 2-butanone in the flavor volatiles
of beef cooked in fat was mainly due to the fat. However, smaller quan-
tities of 2-butanone were found in the flavor volatiles obtained on
cooking beef in water, together with larger amounts of butanal, so it
appears that these two carbonyls play an important role in producing
typical beef aroma and flavor.

Hydrogen sulfide and the mercaptans in the flavor volatiles were
precipitated in a reagent trap containing 4% mercuric cyanide, while any
volatile sulfides passed through this trap and were precipitated in 3%
mercuric chloride. The mercaptans were released from their corresponding

mercuric mercaptides with acid and then were oxidized to disulfides with

Anne Sanderson

iodine in ether for easier separation and identification. Only one mer-
captan was identified in the flavor volatiles of beef cooked in both fat
and water; namely methyl mercaptan. The black hydrogen sulfide precipi-
tate remained in the reaction flask as it was not attacked by acid.
Identification of the precipitated sulfides by most normal tech-
niques was not possible due to their extremely small amounts. 0n regen-

eration of the sulfides with concentrated sodium hydroxide, the sulfide

in greatest proportion was determined by its characteristic odor. Dimethyl

sulfide was tentatively identified in this manner.

Although no differences were found in the number and type of sulfur
compounds in the flavor volatiles of beef cooked in both water and fat,
there is no doubt that slight variations in the concentration of methyl
mercaptan, hydrogen sulfide and/or dimethyl sulfide could account for
some of the flavor differences. This is especially true since the aroma
and flavor of any sulfur compound is very dependent on concentration.
When instrumentation and techniques have been deve10ped to measure the
minute amounts of methyl mercaptan and dimethyl sulfide and their inter-
action with the carbonyl compounds, it should be possible to identify
the exact compound or compounds responsible for the difference in flavor

between beef cooked in fat and water.

EFFECT OF COOKING PROCEDURE ON THE FLAVOR.COMPONENTS OF BEEF

By

Anne Sanderson

A THESIS

Submitted to
iMichigan State University
in partial fulfillment of the requirements
for the degree of

MASTER.OF SCIENCE

Department of Food Science

1966

ACKNOWLEDGEMENTS

The author is sincerely grateful to Dr. A. M. Pearson, Professor
of Food Science, for his continuous guidance and encouragement during
the course of this investigation and preparation of this thesis and to
Dr. B. S. Schweigert, Chairman of Food Science, for advice and encour-
agement.

Sincere thanks are also extended to the author's fellow graduate
students and especially to C. E. Bodwell and F. H. Pepper for their

suggestions and friendship.

ii

TABLE OF CONTENTS
Page

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

LITERATURE REVIEW . . . . . .~. . . . . . . . . . . . . . . . . 3
‘Meat Flavor . . . . . . . . . . . . . . . . . . . . . . . . 3
Flavor Precursors . . . . . . . . . . . . . . . . . . . . . 8

.Method of Cooking . . . . . . . . . . . . . . . . . . . . . 9

EXPERIMENTAL PROCEDURE . . . . . . . . . . . . . . . . . . . . . 13
Sample Preparation . . . . . . . . . . . . . . . . . . . . 13
Cooking Procedure . . . . . . . . . . . . . . . . . . . . . 13

Method of Cooking in water . . . . . . . . . . . . . . 13
IMethod of Cooking in Fat . . . . . . . . . . . . . . . 15

Separation of Carbonyl Compounds . . . . . . . . . . . . . 15
Separation of Sulfur Compounds . . . . . . . . . . . . . . 16
Regeneration of Carbonyls from 2,4-DNPS . . . . . . . . . . 17

Regeneration of'Mercaptans from Mercuric Salts Precipitated
by Mercuric Cyanide . . . . . . . . . . . . . . . . . 18

Regeneration of Sulfur Compounds from.Mercuric Salts
Precipitated by Mercuric Chloride . . . . . . . . . . 22

Gas Chromatography of Carbonyl Compounds . . . . . . . . . 22

Gas Chromatography of Sulfur Compounds . . . . . . . . . . 23
RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . . . . . 26
Carbonyl Compounds . . . . . . . . . . . . . 26
Sulfur Compounds . . . . . . . . . . . . . . . . . . . . 37
Hydrogen Sulfide and Mercaptans . . . . . . . . . . . Z;

Sulfides . . . . . . . . . . . . . . . . . . . . . . .

iii

Page

SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 45

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . 48

iv

Table No.

I

II

III

IV

LIST OF TABLES
Page

Values of prepared mercaptan derivatives . . . . . . . 21

Yield of 2,4-DNPS by two different methods of cooking 27

Retention times of carbonyl compounds regenerated
from 2,4-DNPS by exchange with levulinic acid . . . . 31

Organic disulfide standards . . . . . . . . . . . . . 39

Figure No.

I

II

III

IV

LIST OF FIGURES

Page
Schematic diagram of the cooking apparatus and
trapping system used for the collection of the
volatile components from beef cooked in water . . . . 14

Gas chromatogram of the volatile carbonyl compounds
isolated from beef cooked in water . . . . . . . . . . 29

Gas chromatogram of the volatile carbonyl compounds
isolated from beef cooked in fat . . . . . . . . . . . 32

Gas chromatogram of the disulfide obtained on oxidation
of the mercaptan volatilized by beef cooked in either
water or fat . . . . . . . . . . . . . . . . . . . . . 41

vi

INTRODUCTION

With the expanding population of the modern world, problems of an
insufficient food supply have already arisen. It is becoming increasingly
apparent that the use of animals as a major source of food is inefficient
and costly, and that in future man may have to rely more heavily on plant
material. However, it is likely that man will be reluctant to relinquish
his taste for meat, and for beef in particular. Consequently, it is
necessary to elucidate the exact nature of the flavor constituents that
make meat palatable. By the turn of the century, much of our "meat" will
probably consist of protein derived from plant sources with chemicals
being added to produce the desired flavors. There have been several
attempts to produce "synthetic meat”. Although the texture is similar,
the flavor leaves much to be desired.

The characteristic aroma and flavor of meat have long interested the
researcher. Since the field of flavor chemistry began over a century ago,
the question of origin of the volatile flavor components has been a topic
of wide speculation. Raw meat has little flavor with the characteristic
"meaty" aroma developing on heating. Flavor precursors can be extracted
from raw meat with water and the characteristic odor of meat can be pro-
duced by heating the isolated precursors in fat, (Batzer gt 31., 1962).
Hornstein and Crowe (1964) suggested that the aroma derived on heating
the water-soluble precursors from meat is the same regardless of the type
of meat, while the characteristic species flavor differences are due to

the contribution of the volatiles derived from the fatty tissues.

-2-

The composition and quantity of flavor volatiles varies with the
method of cooking. Both temperature and time of cooking influence the
yield of the volatile components. The quantity and composition of the
flavor volatiles is also affected by the collection conditions. This was
demonstrated by Lineweaver and Pippen (1961) who showed that the amount
and type of volatile carbonyls could be varied by cooking in an inert
atmosphere of nitrogen, an oxidative atmosphere of oxygen or by distilla-
tion. The influence of fat on the yield of flavor volatiles is of
special interest since there are large amounts of fat present in meat,
which readily undergo oxidation on heating.

The present study was undertaken to compare and contrast the flavor
volatiles derived from beef cooked in water and in fat. Particular empha-
sis was placed on the identification of the carbonyl and the sulfur

compounds.

L ITERATURE REVIEW

Meat Flavor

The definition of flavor is as complex as the sensation itself. Of
the many definitions, that of Kazeniac (1961) appears to be the most
appropriate. He described flavor as a combination of taste, aroma, body
or texture and mouth satisfaction.

In 1847 Justus von Leibig published the first analysis on meat ex-
tract. He found no difference between the flavor of meat from ox, pork,
chicken, doe or fox after cooking and concentrating the extracts. He
concluded that the odor and taste of cooked muscle extract was similar
to that from roast meat. More recently, Wood and Bender (1957), Bender
_§£_§l. (1958) and Bender and Ballance (1961) have studied meat extracts.
Wood and Bender (1957) reported the isolation of more than 30 volatile
and non-volatile compounds from commercial ox-muscle extract.

Crocker (1948) studied the flavor volatiles of beef, pork and
chicken. He found ammonia, hydrogen sulfide and acetaldehyde present in
the flavor volatiles derived from all three sources of meat. He came to
the conclusion that the flavor components originated from the fibers of
the meat instead of from the expressible fluid. Further studies by other
researchers (Kramlich and Pearson, 1958), showed that the expressible
fluid deve10ped more flavor on cooking than the fibers, although the
flavor components were more strongly bound to the cooked than to the raw
fibers. This was further substantiated by Hornstein gtflgl. (1960), who

found that water-extracted hamburger was essentially tasteless and odorless.

However, the water extract developed a definite "beefy" aroma on heating.
Kramlich and Pearson (1960) identified acetaldehyde, acetone, carbon
dioxide, methyl mercaptan and tentatively methyl sulfide from a cooked
beef slurry.

In a series of articles, Hornstein and Crowe (1960, 1963, 1964) and
Hornstein ggnal. (1960, 1963) studied the flavor of beef, pork, whale and
lamb. They found that lean meat regardless of source, gave a basic ”meaty"
aroma on cooking. They postulated that the species differences in flavor
are due to the heated fat. The characteristic odor of lamb was found to
be due to the volatiles from the fat, a major portion of which was con-
tributed by the carbonyl compounds. They reported finding ammonia,
acetaldehyde, acetone, formaldehyde, hydrogen sulfide and carbon dioxide
in the volatiles obtained on cooking both beef and pork. In addition they
isolated formic, acetic, pr0pionic, butyric and isobutyric acid from beef
broth.

Pippen §£_§l, (1954) found that more flavor was produced on cooking
a cold water extract of chicken meat than the fibers. They further indi-
cated that the fat contributed to the overall chicken flavor. More
recently, Pippen 25 31. (1958), Pippen and Nonaka (1960, 1963) and Line-
weaver and Pippen (1961) stressed the importance of the volatile carbonyl
compounds in chicken flavor. They isolated and identified many carbonyls
present in the volatile fraction. They observed an increase in the yield
of carbonyls on cooking chicken in an oxidative atmOSphere as compared to

a non-oxidative atmOSphere. Bouthilet (1950, 1951a), working with chicken,

reported that fat played an important role in producing chicken flavor.

He also found that hydrogen sulfide was a major component in the volatiles
from cooked chicken. Fractionation of chicken broth gave a sulfur-contain-
ing fraction with a definite "meaty" odor, and a fraction having a typical
"chickeny" flavor. He concluded that the "meaty" aroma of chicken was

due to a compound associated with the meat fibers, which could be ex-
tracted with water, and stressed the importance of the sulfur compounds

in producing the ”meaty" odor on heating.

Pippen and Eyring (1957) identified the sulfur-containing volatiles
from cooked chicken as consisting mainly of hydrogen sulfide. They
suggested that the hydrogen sulfide released on cooking played an impor-
tant role in chicken flavor as well as in the flavor from other meats.
Mecchi gt 31. (1964) quantitatively analyzed the hydrogen sulfide produced
by heating chicken muscle. They indicated glutathione as the precursor
of hydrogen sulfide. Since the sulfur in glutathione is present as cys-
tine and/or cysteine, they suggested that the rate of hydrogen sulfide
evolution from cooked chicken could be approximately predicted from the
cystine content.

Pippen and Eyring (1957) failed to detect mercaptans in the flavor
volatiles of cooked chicken, but recent work on the flavor of chicken by
Minor gt al. (1965a, b, c) has indicated the presence of not only mercap-
tans but also of disulfides. In a comparative study of the flavor
volatiles from old and young hens, Minor gtflgl. (1965b) found few, if any,

qualitative differences. Minor gg_gl. (1965c) identified carbonyl compounds,

 

hydrogen sulfide and ammonia in the volatile fraction. Pippen and Eyring
(1957) identified the volatile nitrogen as consisting almost completely
of ammonia, and showed by organoleptic tests that ammonia was not impor-
tant to chicken flavor. In fact, the volatile ammonia in chicken flavor
may exert a negative influence according to a report by Lineweaver and
Pippen (1961).

Pippen and Nonaka (1963) reported a comparison of the volatile flavor
components of chicken and turkey, although they did not identify any of
the compounds isolated. They found that carbonyl compounds played an
important role in the production of rancidity in chicken. The importance
of carbonyl compounds was stressed by Jacobson and Koehler (1963), who
suggested that the differences in flavor between breeds of sheep could
be due to differences in the quality and composition of the carbonyls.

Until recently, very little information was available on the flavor
components of cured meats. Ockerman st 31. (1964) presented a gas chroma-
tographic analysis of the volatiles derived from dry-cured hams. They
found that the Spectrum of compounds was very similar to that reported by
other investigators for the volatiles of uncured meat. Cross and Ziegler
(1965) found that the main differences between cured and uncured ham were
due to the different pr0portions of the carbonyl compounds. uncured ham
yielded appreciable quantities of n-valeraldehyde and hexanal on cooking,
but these compounds were barely detectable in the volatiles from cured
ham. Acetone was found to be a major constituent in the volatiles from

both cured and uncured ham. Although the authors found methyl mercaptan

and hydrogen sulfide present in the volatiles from both cured and uncured
ham, they did not attempt to compare the relative amounts of the sulfur
compounds from the two types of meat.

Apart from the identification of hydrogen sulfide in the flavor
volatiles of beef, less emphasis has been placed on the sulfur compounds.
However, many workers agree that sulfur plays an extremely important role
in production of the typical "meaty” aroma. Yueh and Strong (1960), in
addition to hydrogen sulfide, reported finding small quantities of dimethyl
sulfide. Increased amounts of hydrogen sulfide were evolved on prolonged
heating. 0n quantitative analysis, Hornstein gtugl. (1960) determined
that 0.1 mg. of hydrogen sulfide was recovered per gram.of dried beef
powder after vacuum distillation at 100°C. Neither groups of researchers
found any trace of mercaptans or disulfides, even though Yueh and Strong
(1960) suggested the possibility of oxidation of the mercaptans to disul-
fides during cooking. After cooking beef in an inert atmosphere by
bubbling nitrogen gas through the cooking slurry, Kramlich and Pearson
(1960) identified methyl mercaptan in the volatile flavor constituents.
They also tentatively identified dimethyl sulfide, which is in agreement
with Yueh and Strong (1960).

Sulfur compounds in high concentrations are also known to be respon-
sible for off-odors. Batzer and Doty (1955) found methyl mercaptan and
hydrogen sulfide present in the off-odors from gamma irradiated beef.
Although low concentrations of sulfur compounds may greatly enhance the

flavor of a food, they are probably responsible for many unpleasant odors.

Extremely low concentrations of sulfur compounds can be detected by the
human nose, 1.6., one part ethyl mercaptan per fifty billion parts of

air can be detected by the olefactory nerve cells of the human nose
(Noller, 1965). The majority of the flavor volatiles from onions (Carson
and Wong, 1961) and garlic (Oaks 35 31., 1964) are sulfur compounds, such
as mercaptans, sulfides and disulfides, which are normally considered to

have a displeasing aroma.

Flavor Precursors

 

In the water extract of meat, Justus von Liebig (1847) isolated
inosine, creatine, creatinine, lactic acid and amino acids. Over a century
later, inosinic acid, in combination with a ribose-S-phOSphate complex,
was shown to be important in the develOpment of browning and "meaty"
flavor by Wood (1961). Bouthilet (1951b) indicated glutathione was a
major precursor for the "meaty" flavor of chicken, and stressed the im-
portance of the sulfur-containing fraction to meat flavor. Mecchi 35:31.
(1964) found that glutathione was the precursor of hydrogen sulfide in
cooked chicken.

After an elegant separation procedure, Batzer 35 21. (1960, 1962)
isolated a low molecular weight glycoprotein, which Upon heating with
glucose, inosine and inorganic phosPhate in fat, produced an odor similar
to that of broiled steak. In a review of meat flavor and flavor precur-
sors, Doty gt El- (1961) indicated that the glyc0protein fraction could
be isolated from beef, pork or chicken and resulted in the same basic

"meaty" aroma regardless of the source. Further studies using the

fractionation procedure of Batzer et al. (1960) have been carried out by
Wasserman and Gray (1965), who have questioned the role of sugars as one
of the "meaty" flavor precursors. In addition to the low molecular
weight fraction, which gave a broiled steak odor, they isolated a high
molecular weight fraction having a brothy odor on heating. The fraction
responsible for the steak aroma was separated on ion-exchange resins and
a complete amino acid analysis was carried out on all fractions. Horn-
stein and Crowe (1960, 1963) found a similar water-soluble, low molecular
weight compound in the lean meat portions of pork, beef and lamb, which

they indicated to be a precursor of the typical "meaty" odor.

Method.2£_Cooking

In one of the earliest reports on meat flavor, Crocker (1948) boiled
various types of meat in water in an open beaker. Since then numerous
refinements have been made to simulate "actual cooking conditions" and to
trap the volatile flavor components, either by fractionation in a series
of cold traps or by trapping in various reagents. In their studies on
beef, pork, lamb and whale, Hornstein gt El- (1960, 1963) and Hornstein
and Crowe (1960, 1963) removed the volatiles from.meat by vacuum distilla-
tion and then captured them in a series of traps held at the temperature
of either liquid nitrogen or of a dry-ice isoprOpyl alcohol mixture.
They chose 100°C as the most satisfactory operating temperature for vacuum
distillation. A similar temperature fractionation procedure was used by
Kramlich and Pearson (1960) to separate the flavor volatiles produced by

cooking beef in a non-oxidative system. A second method for identification

-10-

was also employed by bubbling the volatiles through a solution of 2,4-
dinitr0pheny1hydrazine (2,4-DNPH) and a solution of lead acetate in order
to identify the carbonyl and sulfur compounds, respectively. Bouthilet
(1949, 1950, 1951a,b) found that the flavor components could be released
from the fibrous material of chicken muscle by simmering in tap water
but were further released from the resulting chicken broth using either
steam or vacuum distillation. He stressed the importance of the inter-
relationship between the fat and lean meat and the production of flavor,
and suggested that the flavor components may be soluble in the fat. This
was later explained by Kazeniac (1961) in a review on chicken flavor.

He suggested that the fat might serve as a trapping agent for some of
the volatiles. Steam distillation of chicken broth with some fat gave a
more desirable flavor than broth from which the fat was removed. In
earlier work, Pippen gt 31. (1954) also found that fat contributed to the
aroma of chicken broth. They found that cooking method and isolation
procedure strongly influenced the yield of carbonyl compounds. When
chicken was cooked under non-oxidative conditions by passing a stream of
nitrogen through the system, the lowest yield of carbonyls was obtained.
Under oxidative conditions, utilizing air to entrain and sweep the vola-
tile components into reagent traps containing 2,4-DNPH, they obtained the
greatest yield of carbonyl compounds. Under "normal cooking conditions"
or distillation, they found an intermediate yield of 2,4-dinitropheny1-
hydrazones (2,4-DNPS). In a later paper, Pippen ggngl. (1958) identified

the 2,4-DNPS released by air-entrainment of cooking chicken. The volatile

-]_]_-

carbonyls produced under these conditions included those which arose
through oxidative processes not present under normal cooking conditions,
and may contribute to both typical and off-flavor.

Steam distillation produced aldehydes with longer carbon chains than
air-entrainment (Pippen and Nonaka, 1960). Chicken boiled in the pre-
sence of air yielded not only a larger volatile fraction, but a more
complex fraction (Pippen and Nonaka, 1963). Temperature was found to be
important in the production of flavor and aroma, as shown by the fact
that raw chicken had little flavor as compared to chicken cooked at 100°C.
They found that n-2,4-decadiena1 was produced on cooking fresh chicken
in the presence of oxygen, and also showed that this compound was a major
volatile component of cooked rancid chicken. This suggests that n-2,4-
decadienal may be an immediate precursor of stale or rancid chicken.
Nonaka and Pippen (1966) found that steam distillation gave a greater
yield of volatiles than vacuum distillation and suggested that the off-
flavor of fried chicken might be due to volatile products produced by the
oxidation of lipid material. They also concluded that the relative amounts
of n-2,4-decadienal could be used as an indication of freshness.

Utilizing the "oxidation-inhibiting conditions" of Pippen £3.2l:
(1958), Minor g£_§l, (1965a, b, c) separated and identified a variety of
flavor volatiles using both a temperature fractionation procedure and a
reagent trapping system. They indicated that the sulfur compounds were
reaponsible for the "meaty" odor of chicken, while the typical "chickeny"

flavor was due to the carbonyl compounds.

-12-

The difficulty of transferring the isolated volatiles from any system
for concentration, identification and analysis has always been a problem
to the flavor chemist, because of the minute quantities of compounds
volatilized, and their often unstable nature. Precipitation in reagent
traps is only suitable for certain classes of volatile compounds, and
temperature fractionation can lead to large losses of volatiles by eva-
poration. Libbey 25 El: (1962) deve10ped a refrigerated trap for the
collection of food volatiles. In this system the volatiles could be
directly transferred to a gas chromatograph by gently heating. Hornstein
and Crowe (1962) used a similarly designed helical trap, but added a
chromatographic column packing of 25% Castorwax on 30/60 mesh Chromosorb
W. This gave the added advantage of a preliminary separation before
transferring the sample to the gas chromatograph. A greatly improved
yield of flavor volatiles was obtained for analysis, although the authors
did not mention the problem of samples containing considerable quantities
of water, which is usually the case with most of the volatiles produced

on cooking meat.

EXPERIMENTAL PROCEDURE

Sample Preparation
‘Most of the meat used in this study was removed from the lumbar

region of the longissimus dorsi muscle of U. S. Choice or Prime grade

 

cattle of approximately 20 months of age. After removal of all the sub-
cutaneous and intermuscular fat, the meat was finely ground and stored
at —20°C prior to cooking. The fat used for cooking was removed from
the same carcasses as the meat samples. The fat was chilled, finely
ground and stored at -20°C until used. Before cooking, the samples were

allowed to thaw slightly at room temperature.

Cooking Procedure

Method of Cooking ELI}. M- A total of 500 g. of ground beef was cooked
with 1000 m1. of distilled water in the cooking apparatus shown in Fig. I.
A "resin reaction" flask was used as a cooking vessel as it offered
special advantages for cleaning after cooking. The cooking flask was
heated with a heating mantle, which could be held at a constant tempera-
ture using a rheostat. The meat slurry was simmered gently for eight
hours after it reached the boiling point. All volatile constituents were
removed from the cooking vessel by constantly sweeping the surface of

the cooking slurry with nitrogen gas, and then trapping the volatiles in
a series of reagent traps. At the end of eight hours of cooking, off-
odors began emanating from the meat and were considered to be atypical

of cooked beef, so cooking was discontinued.

-13-

 

Noun? a.“ woo-000 moon. Eoum muse-.6950 33-39, mafi- mo noun-00:00
usu- uom mom: Emu-who moi-mafi- woo emu-summon wad-000 9-”- mo Baum-3v ”Judas-Em .H mun-swam

 

w
Fan-Sm umoa IILI J

xmmam ..ao.-uomou III.
canon: .H N

8338 Naowm an - e a m 333.

coausaom NAzovwm Nd a a d m macaw

codunHom mmznnaam .uam a N a H wonky

 

 

 

 

-14-

no..- 0805 can III. A

v.
j j) A) 414 #J Samoan-fl.-

mono-once
.' t._ . in , r AH, E...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-15-

Method gf_Cooking in Fat. The apparatus used for cooking beef in fat was
similar to that used for cooking beef in water except that no condensor
was utilized. Instead, any water that was produced was collected in a
500 ml. trap immersed in crushed ice before the remaining volatiles were
swept through the reagent traps.) A total of 500 g. of ground beef was
cooked with 500 g. of fat. The fat was melted in the cooking flask prior
to the addition of the meat in order to prevent sticking or burning.

The meat was cooked in fat for eight hours so that results would be com-
parable to those for the sample cooked in water. The temperature at

which the volatiles emerged was recorded.

Separation of Carbonyl Compounds

The volatiles from the cooking meat were bubbled through a capillary
tube into a saturated solution of 2,4-DNPH in 2N HCl. A second tube of
2,4-DNPH was used to prevent any loss of carbonyls, although it was not
necessary to change either the first or the second 2,4-DNPH trap in this
study. The water collected on cooking the meat in fat was added to a
solution of saturated 2,4-DNPH and the precipitated 2,4-DNPS were added
to those in the first 2,4~DNPH trap. After the eight hour cooking period,
the traps were disconnected and the yellow-orange precipitate of 2,4-
DNPS was centrifuged, washed with 2N HCl and distilled water and then

dried in a desiccator. The yield of 2,4-DNPS from six samples cooked in

fat and six cooked in water was recorded.

-16-

Separation 2f_Sulfur Compounds

The volatile sulfur compounds were passed through a series of traps
containing 4% mercuric cyanide and then 3% mercuric chloride according
to the method of Challenger (1959). This method is based on the fact
that hydrogen sulfide and mercaptans form an insoluble precipitate with
mercuric cyanide, whereas, alkyl sulfides do not form a precipitate.
Alkyl sulfides form an insoluble white precipitate with mercuric chloride.
Thus, any alkyl sulfides produced on cooking would pass through the
mercuric cyanide traps and be precipitated in the mercuric chloride traps.
This method was useful in that it separated the mercaptans from the sul-
fides and disulfides.

After the eight hour cooking period, the traps were disconnected.
The small yield of sulfur compounds from cooking in either fat or water
made it necessary to combine the precipitates from two runs, giving a
total of three identifications from beef cooked in fat and three from
beef cooked in water. A greater yield of sulfur compounds resulted when
the mercuric cyanide and mercuric chloride traps were placed before the
2,4-DNPH traps. After the first trial, this procedure was considered to
be necessary. Mercuric cyanide was added to the water collected from E
the beef cooked in fat to give a 4% solution, and the resultant precipi-
tate was added to that in the other mercuric cyanide trap. The black
precipitate from the mercuric cyanide traps was centrifuged and washed
with distilled water prior to regeneration.

Precipitation of the sulfur compounds as their mercuric salts was

chosen to preference to their lead salts because of the ease of regenera-

-17-

tion. In comparison to the mercuric salts, the lead salts of sulfur
compounds tend to be amorphous and are, therefore, very difficult to

crystallize or obtain pure (Challenger, 1959).

Regeneration of Carbonyls from 2,4-DNPS

 

A stock mixture of one volume of water to nine volumes of levulinic
acid was made up according to Keeney (1957). For the regeneration of
the 2,4-DNPS, 5 ml. of the stock solution and 1 ml. of distilled water
were added to the dried hydrazones in a small distillation flask. The
temperature of the flask was raised very slowly to 100°C in an oil bath
over a period of 30 minutes. The levulinic acid exchanged with the 2,4-
DNPS liberating the free carbonyl compounds, which were distilled together
with a little water into a small tube chilled in ice-water. If regener-
ation were carried out too rapidly, levulinic acid breakdown occurred.

The oil bath was held at 100°C for a further period of five minutes, in
order to distill over the majority of the carbonyl compounds.

Ockerman et a1, (1964) used a-ketoglutaric acid to free the carbonyl
compounds from their 2,4-DNPS. With this method it is difficult to attain
reproducible results, as all the 2,4-DNPS do not always exchange in the
dry state. Cross and Ziegler (1965) have reported good results using 20%
H2304 to regenerate carbonyls from their 2,4-DNPS, however, levulinic
acid was preferred in this study, because of the strong oxidizing prOper-
ties of sulfuric acid.

Since the gas chromatograph used was not suitable for samples con-

taining water, the carbonyl were extracted with an organic solvent. Most

-13-

of the more commonly used solvents, such as diethyl ether, were not con-
sidered to be satisfactory, as they masked the earlier emerging carbonyl
peaks on the gas chromatograph. Methyl phenyl ether was finally chosen
as a suitable solvent, as it emerged on the chromatograph at least 25
minutes after the last of the carbonyl peaks from the regenerated sample.
After extraction, approximately 0.1 g. of powdered anhydrous sodium
sulfate was added to remove any traces of water.

The procedure using levulinic acid described above was chosen since
the extremely small amounts of 2,4-DNPS released on cooking could be
successfully regenerated and separated by gas chromatography. Retention
times of known reagent-grade aldehydes and ketones were employed to
characterize the carbonyls from beef cooked in both water and fat on two

separate columns under differing conditions.

Regeneration of Mercaptans from Mercuric Salts Precipitated bz_Mercuric

 

Cyanide

Challenger (1959) reported the occasional formation of co-ordination
compounds with mercuric cyanide and mercaptides of the type (RS)2Hg.Hg(CN)2
as well as mercaptides such as (RS)2Hg. The co-ordination compounds can
be removed by boiling or shaking the solution of the mercaptides, but
this was found to be unnecessary in the present study. The mercuric
'mercaptides were regenerated with acid, whereas the mercuric sulfide
formed from hydrogen sulfide was not reduced (Challenger, 1959). This
chemical test was utilized to distinguish the difference between mercap-

tides ( (RS)2Hg) and mercuric sulfide (HgSHg). The test was found to be

-19-

extremely useful, since the large quantities of black mercuric sulfide
normally would mask the small quantities of white mercuric mercaptides.
The free mercaptans were liberated from the mercaptides with 1N H2804,
which did not appear to cause much oxidization of the mercaptans to their
corresponding acids.

Trapping the liberated mercaptans presented a particularly difficult
problem due to their extremely volatile nature, their unpleasant odor
and the small amounts present. Collecting the mercaptans in a gas-tight
bottle and subsequent injection into the gas chromatograph with a gas-
tight syringe also resulted in problems. At the high sensitivities
required, contamination of the large surface area of the syringe and the
septum in the injection port of the gas chromatograph could lead to ex-
traneous peaks on subsequent analyses. Trapping the regenerated
mercaptans in a coil held at liquid nitrogen temperatures for direct
release into the chromatograph could be used to eliminate losses in the
gas transfer system. However, the small amounts of water produced during
the regeneration procedure made this method unsatisfactory, since the
gas chromatograph was not suitable for samples containing water.

Adams et a1. (1960) reported a technique for trapping malodorous,
sulfur-containing gases from waste processes of the paper industry. By
passing the regenerated mercaptans through a tube of Drierite at 55°C,
it was possible to trap all traces of water. The mercaptans, still in
the vapor state, passed over to a "U" tube of silica gel at ~78°C. The

‘mercaptans were absorbed on the silica gel and could easily be transferred

-20-

to the gas chromatograph for analysis by gentle heating with a heating
tape. Although this method is ideal for large amounts of known samples,
it was not practical for the extremely minute amounts of regenerated
mercaptans obtained from the mercuric salts of the volatile mercaptans
in the present investigation.

Because of the difficulties encountered in transferring small amounts
of liberated mercaptans into the gas chromatograph, the possibility of
converting them into a more stable form was investigated. Sporek and
Danyi (1963) reported good yields on converting mercaptans to a series
of disulfides by a simple and fast oxidation with iodine in ether. The
disulfides were easily soluble in ether, and owing to their relatively
high boiling points and chemical stability were more convenient to handle
than the original mercaptans. As mixed disulfides are formed from two
or more mercaptans, and give rise to a series of peaks on the gas chroma-
tograph, this offers an additional check on the existence of a particular
thiol.

The black precipitate was transferred to an amber flask fitted with
a water condensor. A total of 10 m1. of diethyl ether and 25 m1. of 1N
H2804 was added, and the mixture stirred vigorously with a magnetic stirrer
for ten minutes at room temperature. After the flask and contents were
cooled in iceawater for five minutes, 10 ml. of diethyl ether containing
an excess of iodine (approximately 0.5 g.) were added to the flask through
the condensor. The mixture was stirred for a further five minutes and
then decanted into a separatory funnel. The resultant liquid contained

considerable unchanged black mercuric sulfide and was brown in color due

:‘3 ’

-21-

to excess iodine. The aqueous phase was discarded and the organic layer
washed with 25 ml. of 0.1M sodium thiosulfate to remove the excess iodine.
Again the aqueous layer was discarded and the organic layer washed with

a solution of saturated sodium chloride. After washing three times with
distilled water followed by saturated sodiun chloride, the organic layer a
was concentrated under vacuum. The concentrated disulfides were dried
with anhydrous sodium sulfate powder and injected into the gas chromato-

g raph.

 

Mercuric salts of known mercaptans were prepared by adding the thiol 5i
to a 4% mercuric cyanide solution. The white mercaptides precipitated
and were recrystallized from ethyl alcohol. The meJJing points of these

mercaptides are recorded in Table l and are compared with those given by

Wild (1960).

Table 1. Values of prepared mercaptan derivatives.

 

Mercury salt

 

 

Boiling point Melting point of Melting point
Mercaptan (°C) prepared mercaptide reported by Wild
(° C) (°C)a
Methyl 6 168 175
Ethyl 36 74 76
n-prOpyl 68 71 72
ISO-propyl 59 63 63
n-butyl 98 85 86
Iso-butyl 88 -- 95
n-amyl 127 66 75

 

aValues reported by F. Wild (1960) in "Characterisation of Organic
Compounds", Cambridge University Press, Cambridge, England p. 104.

-22-

Retention times of the disulfides regenerated from the unknown mer-
curic mercaptides were compared with the retention times of the disulfides

synthesized from the known, crystalline mercuric mercaptides.

Regeneration 9f Sulfur Cpmpounds from Mercuric Salts Precipitated by_
Mercuric Chloride

Even after several runs, the amount of white precipitate in the first
mercuric chloride trap was almost negligible, making identification ex-
tremely difficult. 0n regeneration of the sulfides with cold, concentrated
sodium hydroxide according to Challenger (1959), the amount of material
available was so small that identifications were not possible except by
comparing the odor with that of known sulfides. By comparison with known

dialkyl sulfides, the sulfide in greatest pr0portion could be detected1)y

its characteristic odor.

Gas Chromatography pf Carbonyl Comppunds

The instrument used was a Barber-Colman, Model 20 gas chromatograph
equipped with a radium ionization detector. A six foot, 1/4 inch O.D.
c0pper column packed with 10% diisodecylphthalate on acid washed Diaport
W (60/80 mesh) was shown to satisfactorily separate mixtures of known
aldehydes and ketones. Ockerman E; El! (1964) reported good results
using a similar column packing to separate regenerated carbonyl compounds
from the volatiles of cured ham. The column was Operated at 100°C, with
the flash heater and detector cell at a temperature of 155°C. Argon was

utilized as the carrier gas at 8 p.s.i., giving a calculated flow rate

-23-

of 43.5 ml./minute through the column. The current across the detector
cell was set at 1250 volts.

A second six foot, 1/4 inch O.D. c0pper tubing column, packed with
10% Carbowax 20M on acid washed Diaport W (60/80 mesh) was prepared for
confirmatory identification of the volatile aldehydes and ketones of
cooked beef. It was operated at 75°C with the detector cell and flash
heater settings at 155°C. The cell voltage setting was 1250, and argon
was used as the carrier gas at 8 p.s.i., giving a flow rate of 43.5 ml./
minute through the column.

Retention times of known carbonyl compounds were employed to charac-
terize the unknown aldehydes and ketones regenerated from their 2,4-DNPS

using both columns.

Gas Chromatography pf Sulfur Compounds

 

The disulfides regenerated from the mercuric mercaptides were analyzed
with an Aerograph Model 204 gas chromatograph equipped with both an elec-
tron capture and a hydrogen flame detector. The combination detector was
eSpecially suitable since the electron capture detector is extremely
sensitive to sulfur-containing compounds. Oaks g£_gl, (1964) used a
similar dual channel system in the analysis of sulfur compounds in garlic
with excellent results. The flame detector responds almost equally to
all organic compounds, while the electron capture detector is sensitive
only to certain organic compounds such as halogenated compounds, conju-
gated carbonyls, nitro compounds and sulfur-containing compounds. The

sensitivity between the two detectors varies with the type of compound

-24-

and for this reason the ratio of the electron capture response to the
flame response becomes an important means of component identification.
Oaks 35 31. (1964) designated this value as 9. The value O is an empir-
ical function of the two detectors and is based on the same attentuation
for both detector electrometers and assumes a 1:1 split in the column
effluent going into each detector. Thus, 6 coupled with retention time
gives a definite and almost complete proof of identification when compared
to known standards.

A six foot, 1/4 inch O.D., copper column was packed with 30% (w/w)
Triton X-305 on acid washed Chromosorb W (30/60 mesh). This column was
found to satisfactorily separate known mixtures of sulfur-containing
compounds without the extensive bleeding obtained with some other columns.
Adams and Koppe (1959) investigated several Triton series from Rohm and
Haas as stationary phase solvents for the separation of mercaptans, sul-
fides and disulfides. They obtained the best separation with low
retention times using the Triton X-305 solvent on Chromosorb W.

The column was Operated at 155°C with the flash heater and both
detectors set at 175°C. The flame detector was Operated with a hydrogen
generator set at a hydrogen flow rate of 25 ml./minute. Both detectors
were operated at a fixed potential of 90 volts. Nitrogen was used as a
carrier gas with a flow rate of 50 m1./minute, giving a flow of 25 m1./
minute through each detector, with a 1:1 split after the column. The
responses from both detectors were recorded on a two-pen Westronics

recorder, Model DllA, which allowed identical retention times for each

-25-

channel with a single compound. Both channels were Operated on a range
of 1, usually with an attenuation of 16 times.

Retention times of the disulfides regenerated from the unknown
mercuric mercaptides were compared with the retention times and the E
values of the disulfides synthesized from the prepared mercuric mercap-

tides.

RESULTS AND DISCUSSION

CarbonylgCompounds

The yield of 2,4-DNPS from beef cooked in fat was three times as
large as that Obtained from beef cooked in water. The mean values were
0.17 and 0.05 g., respectively. The data from each cooking replication
are presented in Table II. Although there was considerable variation
between yields of 2,4-DNPS in each replication by both methods of cooking,
the yields were consistently higher on cooking beef in fat. The greater
yield on cooking in fat could have been due to the fact that the conden-
sor used for cooking beef in water may have caused the retention of
carbonyl compounds with the refluxing water, and may have thereby reduced
the yield on cooking in water. Other aldehydes and ketones may have
resulted from the reaction of the meat with the heated fat on cooking
beef in fat. It is also quite possible that the fat itself contributed
volatile aldehydes and ketones, especially since Hornstein and Crowe
(1960, 1963), Cross and Ziegler (1965), and other workers have reported
carbonyl compounds to originate from fat. Nevertheless, cooking the
meat in fat appears to have been responsible for the greater yield of
2,4-DNPS. The fact that cooking 500 g. of fat alone yielded only 0.0098
g. of 2,4-DNPS after six hours of cooking suggested that the greater
Yield on cooking the meat in fat, as compared to water, was actually
re8ponsible for the difference in yields between the two methods of

cooking.

-26-

-27-

Table II. Yield of 2,4-DNPS by two different methods of cooking.

 

Grams of 2,4-DNPS from 500 g. beef

 

 

 

Replication Cooked in water Cooked in fat
1 0.1255 0.1610
2 0.0492 0.1895
3 0.0330 0.1070
4 0.0092 0.0403
5 0.0540 0.2564
6 0.0481 0.2874
Mean: 0.05 0.17

 

The 2,4-DNPS obtained on cooking fat alone were difficult to
regenerate with levulinic acid, and on chromatographic analysis gave one
peak which corresponded to 2-butanone. Other very small peaks obtained
were too minute to identify. The low yield of 2,4-DNPS on cooking in
fat alone would suggest that the differences in yield between the two
methods of cooking meat were largely the result of variation in the
cooking procedures.

Immediately after regeneration of the 2,4-DNPS with levulinic acid,
the carbonyl compounds released from both the beef cooked in water and
that cooked in fat had a sweet, faint caramel aroma. This characteristic
aroma of the aldehydes and ketones is indicative of their role and rela-

tive importance in the flavor volatiles from cooking meat. A typical

_ —‘_-'P—r"g
.' - "w.
__.. '7- ‘1‘.-

;,...--n--::rz=:s:s‘f'.-‘F’-‘i E - '-
4 u I; o L -‘ a.

édfl

-28-

gas chromatogram of the carbonyl compounds from beef cooked in water is
shown in Fig. II. The aldehydes and ketones from beef cooked in water
included formaldehyde, acetaldehyde, propanal, acetone, 2amethyl prOpanal,
butanal, 2-butanone and 3-methyl butanal. Another peak, correSponding

to the 5-carbon aldehydes and ketones, which may have contained either

at?
. :AI’

pentanal, 2-pentanone or 3-pentanone, or any combination of these three

_ __,._F.-aq:~m-

compounds was observed. However, it was not possible to separate and

.xx-g ts- -.

. .—

identify known reagent-grade standards of pentanal, 2-pentanone and

3-pentanone by altering the conditions of the gas chromatograph, since

‘7'

the range in boiling point of the three carbonyls is within 1°C.

Five small peaks were also obtained from the regenerated carbonyl
compounds volatilized by cooking beef in water, but due to the small
sample size identification was not possible. None of the peaks corres-
ponded to known samples of carbonyls such as hexanal, 2- or 3-hexanone,
2-methyl butanal, 3-methyl 2-butanone, or similar short chain aldehydes
and ketones. Another small peak was identified as a breakdown product
of levulinic acid. Butanal, 2-butanone and the 5-carbon carbonyls were
found to be the major carbonyl compounds produced by cooking beef in
water, with smaller amounts of 3emethy1 butanal, 2-methy1 prOpanal,
acetaldehyde, formaldehyde, acetone and prOpanal. The low concentrations
of formaldehyde and acetaldehyde, and especially of the former, were
undoubtedly due to losses by volatilization during the regeneration
procedure.

For identification of the regenerated carbonyls from the flavor

volatiles obtained on cooking beef, the retention times of these aldehydes

:C QUE—.Eu—Jfiufiudxiuuqu—hful r en. \tfw-

-29-

 

.kuwB

ea pmxooo moon Scam woumHomw mvnnooEOo Hhconuwo ofiwumfio> Ono mo Emnmouoeonfio new .HH muomam

Ammusowav mEHH dowuamumm
ea «a N- m- m m w N o

u I. u
I
I

  

      

 

m m. m m.
a
I.-
u o m. c. 1 c. m
m mm a m. m .m
9 m M: 39
l T. a 7ud an.
m. m mm m.” an.
H
u n .A .6 a .d .L a
TL .1. - 3.99 P .d
e I. Q. u.o"u a a
T. u a. n [A 9.: u. a
P I. n 30. 1.11.5 V...
e o 1 a n o n.
U; 9 U1» du 9 U;
rA e u o e .4 a a
P 0 B uu o I.
a T... I 38d 8
S D. 21% U.
3.
B O. 8
u l T.
D. a
e
N. N.
e n.
1.
o
u
a
S

 

o- - aun>auamsom
muonHE\Ha m.m¢ u some 30am
.fi.m.m m n whommowo oomw<
UoooH : munuuHOaEOH

Agave ow\oov 3 nnoamao ;
so muaaoaueaaaooeomaae eoa - easaoo

 

 

 

 

 

-30-

and ketones were compared with those of known reagent-grade standard
carbonyl compounds. The retention times are shown in Table III. The
agreement between the retention times of known standards and unknown
carbonyls on both chromatographic columns was excellent.

Fig. III shows a typical gas chromatogram of the regenerated carbonyl
compounds obtained on cooking beef in fat. Although the beef cooked in
fat gave a greater yield of volatile carbonyl compounds than beef cooked

in water, both cooking procedures produced the same number and type of

.___—-—-_ 1'.:.- ‘l.’"5.. - I:..' ’

aldehydes and ketones. However, the relative proportion of the various
carbonyl compounds differed according to the method of cooking. A com-
parison of the gas chromatograms shown in Fig. 11 and III will show that
beef cooked in water yielded more butanal than 2-butanone, whereas, the
beef cooked in fat usually had more 2-butanone than butanal. The relative
amounts of carbonyls volatilized from beef cooked in fat were less con-
sistent than those from beef cooked in water. Hornstein and Crowe (1960)
have also noted similar variations in volatile aldehydes and ketones
released from heated lamb fat.

As the.major carbonyl compound produced by heating beef fat alone
was 2-butanone, it appears highly probable that the increase in this
compound on cooking beef in fat was due to the contribution of the fat
itself. At first thought this may suggest that 2-butanone does not con-
tribute to the aroma of beef itself. However, this is not necessarily
true since the fat present in the meat itself may result in production
of 2-butanone, and thus, it could still be responsible for the character-
istic difference in aroma between beef cooked in fat and that cooked in

water.

\Pfi—

«Juan ZQIQV A

fflc (Ir in

I‘ll‘Jlll

 

n.mH 0.0H N.m H.w moHuNoH

ma%noanmo nonwmoum

 

 

 

w.oH o.HH N.o m.o mm Hosanna Hmnuoanm
m.m m.m m.¢ o.¢ om moooousnsm
m.m ¢.w m.a N.¢ an Hmamusn
_ 0.0 m.o q.m a.m so Hoseaowa Hmauoauw
m“ m.m ¢.m w.m m.m om Odouoom
H.m N.m n.m o.N ma HmomQOHQ
q.m ¢.m o.N w.H Hm ethnowamuwow
H.N o.N m.H m.H Hm: OwkfimvaEnow
1IweoaM esoaxo: osonx osooxopll noov unwoo poammeoO Hemmanmu
ll o.ma um .11. Iluooo- um menace wuaaaom
cesaoo Sow xmsonumo mumamzunaamompOmwwm

 

mounaaa aw OEDHOO no main coauuoumm

.pwom OHGHH5>OH auw3 owamnoxo

an mmza-e

aN eowm woumwmcowon monsooeoo ahnonwmo mo moEHu coauaouom .HHH manna

.umm DH
poxooo womb Eowm wmumHOmH mwGDanoo Hhaonhmo mawumao> msu mo EOHwODNEOHJO mow

 

    

  

 

 

 

 

 

 

 

.HHH whswwm
Ammuscflev mafia soauamuom
a s a a w .s .N .N o
n I.
u c. I. o
a a I m m
m m m ... m. a m
m m. a a w a d a e
w. u m m. m m. mm.
m m. n m. 4. ”mm m...
D. u o.na u.a"u a
. a. a. m as was a
M“ ,A 1 e s "u 0. mm
. D. s e "u u .d a e s
a «a u o p. 1 x.
S I. B u T o
e D. TL 3 .w u.
u a. u m.
D. l B on
a I u
w a a
m. P
w m
S
OH I %uw>wuwmflmm
Ouoafie\aa m.mq u mush 30am
.H.m.o m u ounmmoua coww<
UoooH - ousuwwmoeoa
Emma 838 3 anon-3m L
so OumaonunmahquOmHflw NOH u oasaoo

 

-33-

Butanal, 2-butanone and the 5-carbon carbonyls appeared to be the
carbonyl compounds in greatest proportion in the flavor volatiles obtained
from beef cooked in fat. Smaller amounts of formaldehyde, acetaldehyde,
acetone, propanal, 2wmethyl prOpanal and 3-methyl butanal were also iden-
tified. In addition, there were five small peaks which were too small
to be identified, and one small peak which emerged before the 5-carbon
carbonyl compounds. The latter peak was due to the breakdown of levulinic
acid.

Known samples of poly- and dicarbonyls, such as 2,3-butanedione
(diacetyl), either gave very broad or reverse peaks and could not be used
for identification. As the peaks from the regenerated carbonyl compounds
from beef cooked in water and fat were sharp, it appears that di- and
polycarbonyls were not present in appreciable quantities. As another
test for polycarbonyls, a 2% solution of alcoholic potassium hydroxide
was added to the 2,4-DNPS dissolved in ethyl acetate. If any polycarbonyl
compounds were present, the 2,4-DNPS solution would have turned brownish-
purple. No color change was observed in the solution of 2,4-DNPS from
beef cooked in either fat or water, indicating that polycarbonyls were
absent in the flavor volatiles produced by both methods of cooking.

Any dehydroxy aldehydes and ketones present in the flavor volatiles
would have formed osazones with 2,4-DNPH. For example, 2,3-butanedione
would have been formed on regeneration of the osazone of 3-hydroxy buta-
none (acetoin). The absence of polycarbonyls in the regenerated carbonyl

sample from the flavor volatiles eliminated the possibility of the exis-

tence of any dehydroxy aldehydes and ketones as well as of polycarbonyls.

-34-

A soluble ring compound is formed when 2,4-DNPH in 2N HCl is added
to 2,4-pentanedione, and the hydrazone does not precipitate. If 2,4-
pentanedione were present in the flavor volatiles, it is unlikely that
it would precipitate as a 2,4-dinitrOphenylhydrazone. It is possible
that 2,4-pentanedione could be present in either or both beef cooked in
fat and water, but, if present, it would not be identifiable as a hydra-
zone. Although 2,4-pentanedione has been identified as a constituent of
the flavor volatiles from the breast and leg muscle of chicken by Minor
‘§£.§1. (1965a), there have been no reports of the isolation of 2,4-pentane-
dione from the volatiles obtained on cooking beef.

The possibility of oxidation of the regenerated aldehydes to their
corresponding acids with Tollen's reagent was investigated for confirma-
tory identification. If ammoniacal silver nitrate is added to a basic
solution of aldehydes, it is reduced to a silver mirror, while the alde-
hydes are oxidized to their corresponding acids (Noller, 1965). Tollen's
reagent will oxidize a variety of organic compounds, including aldehydes,
although it will not affect ketones. Although some of the aldehyde peaks
from the regenerated sample were decreased on the gas chromatograph,
consistent results were unattainable, so the method could not be used for
confirmatory identification of the aldehydes.

According to Wild (1960), aldehydes fonm crystalline derivatives
with an excess of dimedone (5,5-dimethy1 cyclohexa-l,3-dione). As the
reaction is specific for aldehydes, and no precipitate is fonmed with

ketones, the possibility of precipitating the aldehydes of the flavor

-35-

volatiles was investigated for confirmatory identification. Unfortunately,
the dimedone derivatives are formed rather slowly, sometimes taking days
to crystallize, thereby making the method unsatisfactory for precipitation
of the aldehydes of the volatiles obtained by cooking beef.

In a review on meat flavor, HOrnstein and Crowe (1964) listed the
aldehydes and ketones found in the volatiles obtained on cooking beef,
which has been reported by other investigators. These included formalde—
hyde, acetaldehyde, propanal, 2-methyl prOpanal, 3~methyl butanal, 2-
butanone and acetone. These are in agreement with the findings in the
present study. In addition, butanal and the 5-carbon carbonyls were
shown to be present in the flavor volatiles obtained in the present in-
vestigation.

Yueh and Strong (1960) isolated 2,3-butanedione from cooked beef
broth. When reagent-grade 2,3-butanedione was chromatographed, either a
very broad or a reverse peak was obtained. If present, 2,3~butanedione
could have been masked by the large peaks of butanal and 2-butanone on
the gas chromatogram. However, it is unlikely 2,3-butanedione was present
in the flavor volatiles since chemical tests with 2% alcoholic potassium
hydroxide on the 2,4-DNPS were negative.

Hornstein and Crowe (1960) isolated hexanal and small quantities of
nonanal from.steam distilled beef fat. They also reported finding un-
saturated aldehydes and dialdehydes. If they heated beef fat in air
without steam.distillation, dialdehydes and unsaturated aldehydes were
not produced. The three carbonyl compounds they obtained by heating beef

fat in air at 100°C were acetaldehyde, prOpanal and acetone. In the present

-35-

investigation, the beef was cooked in an inert atmOSphere, in the absence
of air or oxygen, by both methods. The temperature of the emerging vola-
tiles from both beef cooked in fat and water did not rise above 100°C,
although when beef was cooked in fat the temperature of the fat itself
was considerably higher than that Of the emerging volatiles. It is pro-
bable that sweeping the surface of the cooking meat with air or oxygen
would have produced a greater yield of carbonyl compounds, especially
from the beef cooked in fat.

It seems likely that the difference between the flavor volatiles
produced by beef cooked in water and fat may arise in part from the
quantity and composition of the aldehydes and ketones released on heating.
However, it is not possible at present to identify the exact carbonyl or
carbonyls responsible for the difference. The greater concentration of
Z-butanone in the flavor volatiles Obtained on cooking beef in fat, as
compared to that cooked in water, was undoubtedly due to the contribution
of the fat itself. This was evident when 2-butanone was shown to be the
carbonyl obtained in the greatest proportion when fat was cooked alone.
It is possible that the relative proportions of 2-butanone and butanal
may account for part of the difference in aroma and/or flavor obtained
on cooking in fat and water, but as yet this has not been verified. When
it is possible to accomplish quantitative analysis of the carbonyls in
the volatiles rather than just determining the relative proportions of
the aldehydes and ketones, it will be possible to elucidate the exact
differences between the carbonyl compounds obtained on cooking beef in

fat and water.

-37..

Sulfur Compounds

Hydrpgpn Sulfide and Mercaptans It was not possible to obtain the yield
of mercuric mercaptides because of the large quantities of mercuric sul-
fide present in the same mercuric cyanide trap. The white mercaptides
were not visible in the first mercuric cyanide trap because the black
mercuric sulfide precipitate masked their appearance. As hydrogen sul-
fide is the only sulfur compound which forms a black precipitate with
mercuric cyanide, this is a positive identification for the presence of
hydrogen sulfide in the flavor volatiles of both beef cooked in both fat
and water. 'Mercurous oxide is another black mercury compound, but it
would not be formed by bubbling the flavor volatiles of cooking meat
through a 4% solution of mercuric cyanide.

Although white mercuric mercaptides were not visible in the mercuric
cyanide trap, the mercaptans were released on regeneration with acid.
Black mercuric sulfide was not decomposed with the IN H2804. This showed
that mercaptans as well as hydrogen sulfide were present in the flavor
volatiles of beef cooked in both water and fat. After regeneration with
1N H2304, the mercaptans were converted to disulfides with excess iodine
in ether according to the procedure described by Sporek and Danyi (1963).
The resulting diethyl ether solution was pale yellow in color due to the
disulfides and impurities. 0n submitting the mixture to gas chromato-
graphy, the impurities completely masked all peaks produced by the electron
capture detector. Thus, it was necessary to wash the mercuric mercaptide

precipitate with diethyl ether prior to regeneration in order to remove

-33-

the impurities. Washing the mercaptides prior to regeneration and oxi-
dation removed the majority of the masking impurities, which were produced
on injection of the disulfide sample.

In order to identify the mercaptans, the retention time and the value
of O (the ratio of the response of the electron capture detector to the
response of the flame detector) for disulfides regenerated from known
‘mercuric mercaptides were recorded. These values are reported in Table
IV. Oaks E£;§l: (1964) reported D values for mercaptans from 0.1 to 0.5,
for monosulfides from 0.01 to 0.09 and for disulfides from 0.5 to 10.0.
The highest O values were obtained for trisulfides, with an increase of
200- to 300-fold over the D value of the disulfides. High 0 values were
observed in the present investigation, if oxidation of the mercaptans to
disulfides were carried out too vigorously, presumably because of the
formation of trisulfides.

Oaks g£_§l, (1964) noted that the values for U may vary as much as
10% from day to day and that the reproducibility for a single compound
within a given set of chromatographic conditions was within a deviation
of 4%. Variation of the E value could be partially compensated for by
normalizing with known standards. In the present study, the value of O
varied considerably according to the amount of standard disulfide injected
into the gas chromatograph. Injection of a consistent volume of known and
unknown disulfides produced a more consistent fl value. All the values of
¢ obtained for disulfides were within the range of 0.5 to 10.0 as found

by Oaks SE 31. (1964).

-39-

Table IV. Organic disulfide standards

Separated on'Triton X-305 volumn at 105°C

 

Retention time

 

Disulfide in minutes fl value
Dimethyl 2.4 1
Methyl ethyl 3.2 2
Methyl iso prOpyl 3.7 3
Diethyl 4.2 2
Methyl prOpyl 4.5 3
DiisoprOpyl 5.2 4
Ethyl propyl 5.9 3
Methyl butyl 6.5 l
Dipropyl 8.3 3
Dibutyl 12.2 3

 

The inconsistency of H is mainly due to the response of the electron
capture detector. This response is dependent on the standing current,
which is partially determined by the column bleed and the contamination
of the tritium foil. In order to attain consistent results, the tritium
foil was cleaned at least every two months according to the Atomic Energy
Commission regulations. Oaks gt El. (1964) reported 9 values for stan-
dard sulfur compounds to two significant places and one significant place
for the unknown sulfur compounds extracted from garlic. Variations in
E up to 1.6 between the standard and unknown disulfides made this accur-
acy appear unnecessary. In the present investigation, the value of D

was reported to the nearest whole number.

-40-

The only disulfide produced from the unknown mercuric mercaptides
from beef cooked in both water and fat was identified as dimethylmdisul-
fide. A typical chromatograph of dimethyl disulfide from either beef
cooked in water or fat is shown in Fig. IV. The first peak is diethyl
ether, which was used as the solvent. The second peak is dimethyl
disulfide obtained upon oxidation of the mercaptan in the flavor vola-
tiles. It has a retention time. of 2.4 minutes and a E value of l,
which is in agreement with the values obtained for the standard sample
of dimethyl disulfide and the values reported by Oaks g£_§l, (1964).
This indicates the presence of methyl mercaptan in the volatiles from
beef cooked in both water and fat.

If more than one mercaptan had been present in the flavor volatiles
of cooking meat, this would have given rise to a series of peaks on
oxidation to disulfides. For example, if ethyl mercaptan had been pre-
sent as well as methyl mercaptan, diethyl and dimethyl and methyl ethyl
disulfide would all have been produced on oxidation. Oxidation of three
mercaptans would have produced six disulfides, four mercaptans would
have produced ten disulfides and five mercaptans would have given 15
permutations. Thus, if more than one mercaptan had been present in the
flavor volatiles of beef cooked in either fat or water, a series of
disulfides would have resulted on oxidation to disulfides. Within the
sensitivity of the gas chromatograph, there is a definite indication that
cooking beef by either method produced only one mercaptan; namely methyl

mercaptan.

-41-

 

 

 

 

 

 

 

 

 

 

 

1
Column - 30% Triton X-305 on
2 Chromosorb W(30/60 mesh)

Temperature - 150°C
Flow rate - ml/minute
Attenuation - 16-x
Flame Ionization

o \\\c

U)

c

8.

a: lo I l l

a? 2 4 6 minutes

'3 I I l I Time

u

0 li_

a

32 4

J Electron Capture

l
i ‘
Figure IV. Gas chromatogram of the disulfide obtained on oxidation 0f

the mercaptan volatilized by beef cooked in either water or
fat.

-42-

Kramlich and Pearson (1960) found methyl mercaptan and hydrogen
sulfide were released on heating a slurry of beef. Besides hydrogen
sulfide and methyl mercaptan, Hornstein and Crowe (1964) noted that ethyl
mercaptan had been found in the flavor volatiles of cooking beef. Ethyl
mercaptan was not found in the flavor volatiles from beef cooked in

either fat or water in the present investigation.

Sulfides The volatile sulfides obtained from cooking meat were passed
through the mercuric cyanide trap and precipitated in the mercuric chlor-
ide trap. According to Challenger (1959), organic sulfides form co-
ordination compounds with mercuric chloride Of the type RZS.ng012. The
co-ordination complex from dimethyl sulfide with mercuric chloride has
the composition 2(CH3)ZS.3HgC12, and that from diethyl sulfide with
mercuric chloride has the fonmula (C2H5)23.2Hg012. There is no bond
between the mercury and carbon as the sulfur donates electrons for the
formation of the co—ordination compound with mercuric chloride. Because
of this fact, the co-ordination compounds are readily decomposed by heat

or hot water. When decomposed with cold sodium hydroxide, the character-

istic odor of the sulfide is apparent.

The combined precipitate from several cookings was filtered and

washed with distilled water. Cold concentrated sodium hydroxide was

added to the precipitate and the odor produced was compared with dimethyl,

diethyl, methyl ethyl and dipropyl sulfide. A brilliant yellow precipi-

tate of mercuric oxide was produced as the dialkyl sulfide or sulfides

was/were released. According to Challenger (1959) the mercuric oxide

-43-

produced does not oxidize the dialkyl sulfides either in hot or cold

suspension.

A strong odor, characteristic of dimethyl sulfide, was produced
when sodium hydroxide was added to the white precipitate from the mercuric
chloride trap of beef cooked in both fat and water. This indicates that

dimethyl sulfide is the sulfide in greatest proportion in the flavor

volatiles of beef cooked in both water and fat. This is in agreement

with Kramlich and Pearson (1960) who tentatively identified dimethyl

sulfide in the flavor volatiles of a cooking beef slurry. Yueh and

Strong (1960) have also identified dimethyl sulfide in the flavor vola-

tiles of cooking beef.

Only hydrogen sulfide and methyl mercaptan were positively identified

from the volatile sulfur derivatives obtained on cooking beef. However,

dimethyl sulfide was tentatively identified by its characteristic odor.
These same three compounds appeared to be present in both beef cooked

in fat and that cooked in water. It is possible that the amounts of

these compounds may have varied quantitatively, but the methods used were
only qualitative, and the differences could not be determined.
The fact, that no differences were found in the type of sulfur com-

pounds produced by beef cooked in either fat or water, does not rule out

the possibility of their role in producing flavor differences by the two

methods of cooking. The odor of hydrogen sulfide, mercaptans and sul-

fides varies considerably with concentration. Hornstein 35 31. (1960)

have already determined the quantitative amounts of hydrogen sulfide

-44-

produced when beef is cooked in water and work on the hydrogen sulfide

produced from heated beef fat is currently under investigation. Although

quantitative analysis of the minute amounts of methyl mercaptan and di-
‘methyl sulfide presents a more formidable task, there is no doubt that

the improved instrumentation and techniques of the future will determine

the exact role and importance of each flavor constituent.

SUMMARY AND CONCLUSIONS

Chemical components volatilized on cooking beef in water were com-

pared with those obtained on cooking an equal amount of beef in fat. The

flavor volatiles were separated in reagent traps, regenerated and identi-

fied by gas chromatography.

The yield of carbonyl compounds precipitated as their 2,4-DNPS from
beef cooked in fat was three times that from beef cooked in water, with

‘mean values of 0.17 and 0.05 g., respectively. Although the fat itself

contributed some carbonyl compounds, these carbonyls did not completely

account for the increase in 2,4-DNPS on cooking beef in fat. The greater

proportion of 2-butanone volatilized by beef cooked in fat was probably
due to the contribution of the fat, although it could still be a major
contributor to the aroma and/or flavor of beef cooked in fat.

On regeneration with levulinic acid, the carbonyl compounds from

the volatiles obtained on cooking beef in both fat and water had a sweet,

faint caramel aroma. A total of 14 peaks was obtained from the aldehydes

and ketones regenerated from their 2,4-DNPS from beef cooked in both fat

and water. Butanal, 2-butanone and the 5-carbon carbonyl compounds con-

stituted the major aldehydes and ketones produced by both methods of

cooking. Smaller amounts of formaldehyde, acetaldehyde, acetone, prepanal,

2~methy1 propanal and 3-methyl butanal were also identified. Five other

peaks were Obtained but identification was not possible due to the small

sample size. A small peak caused by the breakdown of levulinic acid was

observed on the gas chromatograms of the carbonyl compounds from beef

cooked in both water and fat.

-45-

-46-

Polycarbonyls, such as 2,3-butanedione, were shown to be absent in
the regenerated carbonyls from the flavor volatiles Obtained on cooking
beef in fat and water. Various chemical tests for the confirmatory
identification of the aldehydes and ketones were found to be unsatis-
factory.

The carbonyl compounds in the flavor volatiles obtained on cooking
beef in water had a greater concentration of butanal than 2-butanone,
whereas, beef cooked in fat yielded more 2-butanone than butanal. Since
2—butanone was found to be the carbonyl compound released in greatest
pr0portion from heated beef fat, it is possible that the larger propor-
tion of 2-butanone found in the volatiles obtained on cooking beef in
fat may have been partially due to the fat itself.

Free mercaptans were released on regeneration of the precipitated
mercuric mercaptides with acid. The freed mercaptans were oxidized to
their corresponding disulfides for easier identification and analyzed by
gas chromatography. Methyl mercaptan was the only mercaptan identified
in the flavor volatiles obtained on cooking beef in either fat or water.

Large quantities of hydrogen sulfide were found to be present in the
flavor volatiles obtained from beef cooked in both water and fat. In
addition, dimethyl sulfide was tentatively identified as the sulfide in
greatest proportion in the volatiles of beef cooked in water and fat.

The importance of the sulfur compounds in flavor is well known.
Although no differences were found in the type of sulfur compounds pro-

duced by the two methods of cooking, it is evident that further investi-

-47-

gations are necessary to elucidate the exact role and importance of each

component. Studies to increase the knowledge of interactions between the

carbonyl and sulfur compounds in the flavor volatiles of meat is also

necessary.

BIBLIOGRAPHY

Adams, D. F., and R. K, KOppe. 1959. Gas chromatographic analysis of

hydrogen sulfide, sulfur dioxide, mercaptans, and alkyl sulfides and
disulfides. TAPPI 42:601.

Adams, D. F., R. K. Koppe and IL PL Jungroth. 1960. Adsorption sampling

and gas chromatographic analysis of sulfur compounds in waste process
gases. TAPPI 43:602.

Batzer, O. F., and D. M. Doty. 1955. Nature of undesirable odors formed
by gamma irradiation of beef. J. Ag. Food Chem. 3:64.

Batzer, O. F., A. T. Santoro, W. A. Landmann and B. S. Schweigert. 1960.
Precursors of beef flavor. J. Ag. Food Chem. 8:498.

-. -flufiififia .' 3.“???

tr

Batzer, O. F., A. T. Santoro and W. A. Landmann. 1962. Identification
of beef flavor precursors. J. Ag. Food Chem. 10:104.

Bender, A. E., and P. E. Ballance. 1961. A preliminary examination of
the flavor of meat extract. J. Sci. Food Ag. 12:683.

Bender, A. E., T. Wood and J. A. Palgrave. 1958.

Analysis of tissue
constituents. Extract of fresh OX‘mUSCIE.

J. Sci. Food Ag. 9:812.

Bouthilet, R. J. 1949. A note on the nature of a flavor constituent
from poultry meat. Food Technol. 3:118.

Bouthilet, R. J. 1950. Chicken flavor:

The source of the meat flavor
component. Food Res. 15:322.

Bouthilet, R. J. 1951a.

The fractionation of volatile constituents.
Food Res. 16:118.

Bouthilet, R. J. 1951b. Chicken flavor: Separation and concentration
of its volatile components from broth. Food Res. 16:201.

Challenger, F. 1959. Aspects pf the Organic Chemistry pf Sulfur.
Butterworth's. London, England.

Crocker, B. C. 1948. Flavor of meat. Food Res. 13:179.

Cross, C. K., and P. Ziegler. 1965. A comparison of the volatile frac-
tions from cured and uncured meat. J. Food Sci. 30:610.

Carson, J. F., and F. F. Wong. 1961.

The volatile flavor components of
onions. J. Ag. Food Chem. 9:140.

-48-

-49-

Doty, D. M., O. F. Batzer, W. A.'Landmann and A. T. Santoro. 1961.
Meat flavor. Proceedings Flavor Chemistry Symposium. Campbell Soup
Co., Camden, N. J. p. 7.

Hornstein, I., P. F. Crowe and W. L. Sulzbacher. 1960. Constituents of
‘meat flavor: Beef. J. Ag. Food Chem. 8:65.

Hornstein, I., and P. F. Crowe. 1960. 'Meat flavor chemistry: Studies
on beef and pork. J. Ag. Food Chem. 8:494.

Hornstein, 1., and P. F. Crowe. 1962. Gas chromatography of food vola-
tiles - An improved collection system. Anal. Chem. 34:1354.

Hornstein, I., and P. F. Crowe. 1963. Meat flavor: Lamb. J. Ag. Food
Chem. 11:147.

'Hornstein, I., P. F. Crowe and W. L. Sulzbacher. 1963. Flavor of beef
and whale meat. Nature 199:1252.

Hornstein, I., and P. F. Crowe. 1964. Meat flavor - A review. J. Gas
Chromatography 2:128.

Kazeniac, S. J. 1961. Chicken flavor. Proceedingp Flavor Chemistry
Symposium. Campbell Soup Co., Camden, N. J. p. 37.

Keeney, M. 1957. Regeneration of carbonyls from 2,4-dinitrophenylhydra-
zones with levulinic acid. Anal. Chem. 29:1489.

Kramlich, W. E., and A. M. Pearson. 1958. Some preliminary studies on
‘meat flavor. Food Res. 23:567.

Kramlich, W. E., and A. M. Pearson. 1960. Separation and identification
of cooked beef flavor components. Food Res. 25:712.

Libbey, L. M., D. D. Bills and E. A. Day. 1963. A technique for the
study of lipid-soluble food flavor volatiles. J. Food Sci. 28:329.

Liebig, Justus von. 1847. Uber die Bestandtheile der Flgssigkeiten
des Fleisches. Annalen der Chemie und Pharmacie 62:257.

Lineweaver, H., and E. L. Pippen. 1961. Chicken flavor. Proceedings
Flavor Chemistry Symposium. Campbell Soup Co., Camden, N. J. p. 21.

 

Mecchi, E. P., E. L. Pippen and H. Lineweaver. 1964. Origin of hydrogen
sulfide in heated chicken muscle. J. Food Sci. 29:393.

-50-

Minor, L. J., A. M. Pearson, L. E. Dawson and B. S. Schweigert. 1965a.
Chicken flavor: The identification of some chemical components and
the importance of sulfur compounds in the cooked volatile fraction.
J. Food Sci. 30:686.

‘Minor, L. J., A. M. Pearson, L. E. Dawson and B. S. Schweigert. 1965b.
Gas chromatographic analysis of volatile constituents from cooked
carcasses of old and young chickens. Poultry Sci. 44:535.

‘Minor, L. J., A. M. Pearson, L. E. Dawson and B. S. Schweigert. 1965c.
Separation and identification of carbonyl and sulfur compounds in
the volatile fraction of cooked chicken. J. Ag. Food Chem. 13:298.

Noller, C. R. 1965. Chemistry pf Organic Compounds. W. B. Saunders Co.,
Philadelphia.

Nonaka, M., and E. L. Pippen. 1966. Volatiles and oxidative flavor
deterioration in fried chicken. J. Ag. Food Chem. 14:2.

Oaks, D.‘M., H. Hartman and K. P. Dimick. 1964. Analysis of sulfur
compounds with electron capture/hydrogen flame dual channel gas
chromatography. Anal. Chem. 36:1560.

Ockerman, H. W., T. N. Blumer and H. B. Craig. 1964. Volatile chemical
compounds in dry-cured hams. J. Food Sci. 29:123.

Pippen, E. L., A. A. Campbell and J. V. Streeter. 1954. Flavor studies.
Origin of chicken flavor. Poultry Sci. 34:1139.

Pippen, E. L., and E. J. Eyring. 1957. Characterization of volatile
nitrogen and sulfur fractions of cooked chicken and their relation
to flavor. Food Technol. 11:53.

Pippen, E. L., A. A. Campbell, F. T. Jones and F. Stitt. 1958. Volatile
carbonyl compounds of cooked chicken. 1. Compounds obtained by air
entrainment. Food Res. 23:103.

Pippen, E. L., and M. Nonaka. 1960. Volatile carbonyl compounds of
chicken. II. Compounds volatilized with steam during cooking. Food
Res. 25:764.

Pippen, E. L., and M. Nonaka. 1963. Gas chromatography of chicken and
turkey volatiles; the effect of temperature, oxygen, and type of
tissues on composition of the volatile fraction. J. Food Sci.
28:334.

Sporek, K. F., and M. D. Danyi. 1963. Detection and identification of
mercaptans by gas liquid chromatography. Anal. Chem. 35:956.

-51-

Wasserman, A. E., and N. Gray. 1965. IMeat flavor.

I. Fractionation
of water soluble flavor precursors of beef.

J. Food Sci. 30:801.

Wild, F. 1960. Characterisation g£_0rganic Compounds. Cambridge Univer-
sity Press, Cambridge, England.

Wood, T. 1961. Browning of OX'mUSCIB extracts. J. Sci. Food Ag. 12:61.

Wood, T., and A. E. Bender. 1957.

Analysis of tissue constituents.
Commercial ox muscle extract.

Biochem. J. 67:366.

Yueh,‘M. H., and F. M. Strong. 1960. Some volatile constituents of
cooked beef. J. Ag. Food Chem. 8:491.

”'Cl’it'fim‘juElm“filliflilj’lil’llulffiflm“