EFFECT OF COLCR AND FLAVOR
MODIFICATION ON THE PALATABILITY

AND GEL STRUCTURE OF
STANDARD BAKED CUSTARD

Thesis far Hm Degree of M. S.
MICKEGRN STATE UNIVERSITY

Mary Eve‘iyn
1964

 

THESIS

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

IIIIIIIII III IIII III III III IIIIIIIIIIIIIIL

3 1293 00073 9643

LIBRARY

Michigan {State
Univermty

 

”a... .—

 

s I

PLACE IN RETURN BOX to remove this checkout from your mood.
To AVOID FINES return on or before data due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative ActioNEqual Opportunlty lmfltutlon

 

 

 

 

‘I‘ III Iva -

ABSTRACT

EFFECT OF COLOR AND FLAVOR MODIFICATIONS
ON THE PALATABILITY AND GEL STRUCTURE
OF STANDARD BAKED CUSTARD

By Mary Evelyn Garlick

The purpose of this investigation was to study the ef-
fect of the addition of the nine possible combinations of
three different peach colors and three levels of peach fla-
vor on the palatability and gel structure of standard baked
custard.

The basic formula contained a constant proportion of
ingredients using reconstituted dried whole milk, fresh eggs
and sucrose. Each of the nine treatment variables contained
one of three peach colors - designated light, medium, and
dark color - and one of three levels of peach flavor - desig-
nated low, medium, and high flavor level. A control custard
having neither added color nor added flavor was used as a
reference for the objective tests. Six replications of each
of the nine treatment variables were prepared and evaluated
at room temperature by subjective and objective methods.

A taste panel (4 women and 3 men) evaluated nine char-
acteristics on a 7-point rating scale: crust color, crust
tenderness, crust flavor, inside color, aroma, inside flavor,
consistency, texture, and syneresis. Objective measurements
included pH of the custard before and after baking, gel

strength as indicated by the penetrometer (crust on, crust

Mary Evelyn Garlick

off) and by per cent sag, and syneresis as indicated by per
cent drainage.

Analysis of variance on the subjective scores for the
nine treatment variables indicated no significant differences
in six characteristics: crust tenderness, aroma, inside fla-
vor, consistency, texture, and syneresis. The custards were
significantly different in crust color (1% level of probabil-
ity), inside color (1% level of probability), and crust fla-
vor (5% level of.probability). For crust color and inside
color the medium color custards, as a group, scored highest.
The custard with medium color and high flavor level scored
highest for crust color and second highest for inside color.
For crust flavor, the custard with medium color and high fla-
Ivor level scored the highest followed by two custards con-
taining the low flavor level and dark and medium colors re-
spectively. There was an indication of the "halo effect" of
ncolor impression on the judges' scores for flavor. The
judges' scores and comments indicated the color and the fla-
vor were concentrated in the crust. In considering all three
significantly different subjective characteristics, it ap-
peared that the custard containing the medium color and high
flavor level was scored highest by this taste panel.

No significant differences existed in any of the objec-
tive measurements. High standard deviations in the measure-
ments for both gel strength and per cent drainage cast doubt
on the reliability and suitability of these two objective

measurements for baked custards. The addition of color and

Mary Evelyn Garlick

flavor did not alter the pH of either the mix or the baked
custards as compared to the control custard.

No significant correlations were found between objec-
tive and subjective measurements for gel strength or for
syneresis. The highly significant positive correlations be-
tween crust color and crust flavor and between inside color
and inside flavor indicate further the "halo effect" of color
impression on flavor judgment. Highly significant correla-
tions existed between crust color and inside color, between
crust flavor and inside flavor, and between texture and syn-
eresis.‘ Positive correlations (5% level of probability)
existed between inside flavor and aroma and between con-

sistency and texture.

EFFECT OF COLOR AND FLAVOR MODIFICATION
ON THE PALATABILITY AND GEL STRUCTURE
OF STANDARD BAKED CUSTARD

BY

Mary Evelyn Garlick

A THESIS

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

MASTER OF SCIENCE

Department of Foods and Nutrition

1964

<32fl045
7/9 #04)

ACKNOWLEDGMENTS

The writer wishes to express her gratitude and appreci-
ation to Dr. Theodore F. Irmiter for his generous guidance
and encouragement throughout this study.

Grateful acknowledgment is also due to Dr. L. E. Dawson,
Professor of Food Science, Michigan State University, and to
Dr. T. I. Hedrick, Professor of Food Science, Michigan State
University, and their staffs. Without their invaluable assist-
ance, a controlled study would not have been possible. The
assistance of Donald Kiel, Michigan State UniVersity Computer
Center, and Programmer, Agricultural Experiment Station, in
the statistical analysis of the data is also greatly appreci-
ated.

The writer expresses her sincere appreciation to
Givaudan Flavors Inc.‘, New York, and Food Materials Corpora-
tion‘, Chicago, Illinois, for supplying flavor samples and
information used in this study.

I Grateful appreciation is expressed to Dr. Grace Miller,
Miss Mary Morr, Mrs. Norma Gilmore, Mrs. Bette Smith, Spiros
Constantidines, H._Crane Day, and Albert Hefner, who served
on the taste panel.

The writer wishes to acknowledge her special indebted-
ness to General Foods Corporation, White Plains, New fork,
for granting her a fellowship to study for this degree and

for additional funds which provided support for this research.

 

1WMention of manufacturers or their products in this thesis
does not constitute endorsement by Michigan State University.

To my Mother and Father

TABLE OF CONTENTS

ACKNOWLEDGMENTS.......,...........................,...
LIST OF TABLES........................................
LIST OF FIGURES..................................,....
LIST OF EXHIBITS......................................
INTRODUCTION.......................................,,.
REVIEW OF LITERATURE..................................
Color in Foods...,...............................
Importance of color...,.....................

Classes of colors...........................

Addition of colors to foods.................

U. S. government regulations.,.........

Certified food colors.....,.,..........

FD&C Red No. 4....................

FD&C Yellow No. 5......,..........

Color blends......................

Evaluation and measurement of color.........
Subjective evaluation.........,........

Objective measurement................,.

Flavor in Foods..................................
Psychological aspects....,..................

Factors influencing flavor..,,...,,.........

Addition of flavors to food products..,.....

Fruit flavors.........a.....................

iv

Page
ii
viii

>4
\OODCD-xlflmmmmk-D'J-‘HH

F‘ r4 14 h‘ P4 re
Ox er \» P‘ +4 (D

TABLE or CONTENTS (CONT.)
Page
Classifications........................ 16
True Fruit Flavoring.............. 17

True Fruit with Other
Natural Flavoreoeevooeoeo900000909 l7

Imitation Fruit Flavoring......... l7

Imitation Fruit Flavoring
With True Fruitooqoooooqoooooooooo l7

PeaCh flavorOO’O..'.....,..0,0.00...... 17

Custards as a Medium for Experimental
work with Color and Flavor....................... 19

Composition................................. 19
Heat coagulation of proteins..........,..... 19

Effect of acid, alkali, salts,
and organic solvents,.....,.oooq........so.. 20

pH of custard mix and cooked custard....,... 2l
Gelation temperature and rate of cooking.... 21
Method of baking............,............... _ 22

Use of dried whole milk in baked custards... 24
Evaluation of Foods.............................. 24
Subjective methods................,......... 25
Types of tests and judgments........... 25
Considerations for taste panels........ 26

Objective methods........................... 27
EXPERIMENTAL PROCEDURE...............,................ 29
Preliminary Investigations....................,.. 29

V

TABLE OF CONTENTS (CONT.)

Page

Design of Experiment........................,.... 30
Color-flavor combinations..............,.... 30
Basic Custard Formula............................ 32
Ingredient Procurement...................,....... 32
Basic formula ingredients................... 32
Colors........................,.....,....... 33
Flavors..................................... 33
Preparation of Color Solutions................... 34
Preparation of Custard Mix....................... 34
Baking Procedure........,........................ 35
Subjective Tests................................. 37
Objective Tests......................,.,......... 38
Penetration................................. 39

Per cent sag................................ 39
Syneresis................................... 41
pH....................,..................... 41
Statistical Methods.....,........................ 43
RESULTS AND DISCUSSION................................ 44
Subjective Tests................................. 45
Crust color................................. 45
Crust tsnderness,..,......,................. 48
Crust flavor................................ 48
Inside color..............,................. 53

vi

TABLE OF CONTENTS (CONT.)
Page
Aroma,....................,...,....,........ 56
Inside flavor....................,..,....... 58
Consistency...........,..................... 6O
Texture.,................................... 6O
Syneresis.....,....,........................ 63
Objective Tests..................,............... 63
Penetration.........,....................... 63
Per cent sag...........,,................... 68
Syneresis................................... 70
pH readings.........................,,...... 72
Correlation between Selected Measurements......,. 72
SUMMARY AND CONCLUSIONS.................,............. 77
LITERATURE CITED..,................................... 82
APPENDlx......................................,,...... 87

vii

LIST OF TABLES
Table Page

1. Mean Scores and Standard Deviations
for CPUBL COloroooeeooeoooooo'yooeooeocoeooooooo 46

2. Analysis of Variance for Crust Color............ 47

3. Mean Scores and Standard Deviations
for Crust Tenderness.....,...................... SO

4. Mean Scores and Standard Deviations
for crust FlavorQOQQOOOOQQOOQOOOOIOOCOQOOQ'0.000 51

5. Analysis of Variance for Crust Flavor........... 52

6. Mean Scores and Standard Deviations
for Inglde Caloroooocoooeoooooyoo99099900190900. 54

7. Analysis of Variance for Inside Color..,........ 55

8. Mean Scores and Standard Deviations
for aromaoooeooooooooeooooeeoooo0000009000000... 57

9. Mean Scores and Standard Deviations
for Inside FlavorOOOOOOOOOQOOOOOOOQQOOOOOO0.0909 59

10. Mean Scores and Standard Deviations
for conalstencyOOOO’QOOOOO’OOOQO'OOOOOOOO090090O 61

11. Mean Scores and Standard Deviations
for Texturepoooesooco9oecoooooeooooooooeeeoooooo 62

12. Mean Scores and Standard Deviations
for SynereSIS.........,...o.oo,o.p.o9.g......... 64

13. Mean Penetrometer Values and Standard
Deviations on Baked Custards with the
CPUBL onooeeoeoeeooooeoeqoeoooeoccooo'QQCQOoqooo 65

14. Mean Penetrometer Values and Standard
Deviations on Baked Custards with the
crust OffOOOOOOQOOOOOOOOOQOQOOOQOQQOOQOOOOQ0,... 66

15. Mean Penetrometer Difference Values
and Standard Deviations......................... 67

16. Per Cent Sag and Standard Deviations
of Baked Custards............................... 69

viii

LIST OF TABLES (cour.)
Table Page

17. Per Cent Drainage and Standard
Deviations of Baked Custards.,,..,.,.,.......... 71

18, Range of pH Readings over Six Replications...... 73

19. Correlation Coefficients of
Selected Measurements........................... 74

20. Colors and Flavors Used in Baked
Custards for Objective Tests..,..........,...... 88

21. Colors and Flavors Used in Baked
Custards for Subjective Tssts................... 89

22. Description of Flavors Used..,.............,,... 90

23. Composition of Color Solutions........,.,....... 91

ix

LIST OF FIGURES

Figure

1.

2.

Water bath pan equipped with stopcock
and perforated stainless steel rack,-
showing placement of thermocouples..............

Measuring penetration after release
of 35-gram cone attachment......................

Apparatus for height measurement of
free-standing inverted baked custard............

Equipment used for drainage weight
determinations from freenstanding
inverted baked custard..........o.........ouoo.

Three different peach colors in
baked custaardla(cru8t on)009000000000001000000030

Three different peach colors in
inverted baked custards..................,......

Page

36

42
49

49

LIST OF EXHIBITS
Exhibit ‘ Page

1. Procedure for Preparation of
Colored and Flavored Custard Mixes............. 92

2. General Instructions for Peach
Custard Taste Panel Members.................... 94

3. Score Sheet for Peach Custard.................. 96

4. Sample Calculation of Studentized
Multiple Range Test...........,................ 97

xi

INTRODUCTION

For many decades home economists have been concerned
with the importance of color and flavor of foods in the
skillful planning of normal and modified diets. These food
. attributes have been recognized as important determinants of
consumer acceptance and satisfaction. More recently, the
research of psychologists and physiologists has reflected
increased interest in and recognition of the effect food
colors and flavors exert on the human appetite (2). Food
technologists continue to focus their attention on the basic
factors involved in the preservation of color and flavor in
foods in the development of new processing techniques, new
products, and new forms of established products.

Some dietitians have recognized the psychological im-
pact of color as'a key to successful meal planning and ser-
vice (60). The appetite inhibiting effect of monotony in
the diet is an ever present concern of all nutritionists.
Although careful selection of food and food combinations is
of prime importance, numerous workers believe that variation
of color and flavor can be of considerable assistance in re-
lieving monotony in all types of diets (2, 58).

The problem of monotony is, perhaps, more acute in the
formulation of restricted diets than in the planning of nor—
mal ones, especially in the area of permissible dessert
items. Baked custard, basically a soft, bland, high protein

food consisting of milk, eggs and sugar, is used extensively
l

for many types of modified diets. For the most part, when
baked custards are served they are yellow in color and vanil-
la flavored. A search of the literature failed to reveal re-
ports of investigations concerned with the use: of fruit
colors and flavors in the preparation of baked custards. If
selected fruit colors and flavors could be added successfully
to a basic custard formula, it is conceivable that variations
of this standard dessert item would be useful to both the
homemaker and the professional dietitian.

Preliminary experimental work indicated it is possible
to produce a peach colored and peach flavored baked custard
without noticeably altering the gel structure of the product.
On the basis of these limited observations a controlled in-
vestigation was planned to study the effect of peach color
and flavor modification on the palatability and gel structure
of standard baked custard: The objectives of this study were:

1. To identify the types and proportions of food colors
which can be combined with the natural carotene pig-
ments present in a standard baked custard mix to
produce a desirable peach color.

‘2. To determine the type and levels of peach flavors
which, when added to the standard baked custard
formula, produce custards of acceptable peach fla-
vor compatible with the peach colors selected for
study.

3. To study the effect of the addition of three

different peach colors and three levels of peach
flavor on the palatability and gel structure of

standard baked custard.

REVIEW OF LITERATURE
Color in Foods

m .ce Co r

Color is constantly a part of food, a visual element to
which human eyes, minds, emotions and palates are very sensi-
tive. Through the ages, man has come to build strong and in-
tuitive associations between what he sees and what he eats.
People demand the right shade in foodstuffs and will accept
or reject a product on its appearance - quite apart from nu-
tritional considerations (2). i .

_ Color, the first quality attribute a consumer perceives,
plays a major part in his willingness to accept a food. It
is often regarded as an index of the general quality of the
product, and may influence the consumer's Judgment of flavor
(28, 42, 57). >

What is perhaps basic to color and appetite are certain
direct associations and known responses to the stimulation of
color. According to Birren (2), bright and "warm" colors
(red, orange, and yellow) tend to stimulate the autonomic
nervous system of man, including digestion, whereas soft and
"cool" colors tend to retard it. Of these warm colors, red
and orange have a more stimulating effect than yellow.

People learn to associate colors with various kinds of
experiences with food such as their taste, odor, or the total
complex of stimuli associated with eating, and ultimately

4

5

with the resulting degree of satisfaction. However, to
interpret preference for a food on the basis of color alone
without regard to its role as a member of a group of stimuli

could result in invalid conclusions (57).

C ass a of lore

The two general classifications of colors in food are
natural and synthetic. The natural color classification in-
cludes the pigments and appearance factors actually found in
food, such as the homes, carotenoids, chlorophylls, antho-
cyanins, and flavonoids. Carotene is present in egg and
milk and lutein is the yellow coloring matter of egg yolk
(42). Natural colors also include those derived from vege-
table, animal, and mineral sources. Natural vegetable colors
most commonly added to foods are caramel, from burnt sugar,
and annatto, an extract of annatto seed (60).

Synthetic colors are compounds of known structure pro~
duced by chemical synthesis and conforming to standards es-
tablished by the U.’§. Food and Drug Administration. Syn-
thetic colors are most widely used in the food industries
because of the variety of available shades, their brilliance

and uniformity, solubility and high tinctorial strength (60).

Addition 9; gglgrs to Foods

Foods fall naturally into two groups when examined on
the basis of color:

1. Foods which have an acceptable natural color,
or in which an acceptable color is deve10ped
through cooking.

2. Rxds which usually require added color in
processing (margarine, cheese, desserts such
as ice cream and sherbets, gelatin desserts,
puddings, candy, cakes, cookies and pastries
as well as many beverages) (60).

According to Jablonski (24), colors have long been used
to improve the appearance of products. The number of color-
ing materials has increased in the past fifty years from a
few simple ones to a vast range of tints and shades, mostly
of synthetic'production, and their use» has become an inte-
gral part of our civilization.

U, §, government regulations

Added colors in foods are regulated by the Color Addi-
tives Amendment of 1960 to the Federal Food, Drug,end Cos-
metic Act and are deemed "color additives" in the Code of
Federal Regulations(lo). In order to protect the health of
the public, the U. 5. Government permits food manufacturers
to use only batches of coal-tar dyes which have been tested
by the U. 8. Food and Drug Administration and certified by
this agency to be harmless and suitable for use (24).

Cert fied o co ors

Jablonski (24) reports when the Federal Food and Drug
Act was first enacted (1907) the number of permitted coal-
tar colors certifiable for food purposes was limited to
seven. By 1950 the number had increased to nineteen dyes -
four oil-soluble dyes and fifteen water-soluble dyes. Since

1950 the Food and Drug Administration has removed the four

oil-soluble colors and four of the waterwsoluble colors from

the list of permitted colors because these delisted colors
did not meet the requirements of the Color Additives Amend-
ment of 1960 (10). At the present time there are eleven
certified food colors, all of which are water-soluble (10).
Most of the primary certified food colors have two names, the
official or FDacC1 designations and the common trade name.

FD&C gag “9;-4,'.2932°au §§

FD&C Red No. 4, commonly called Ponceau 8X, is a red

 

powder easily soluble in water, giving an orange-red solu-
tion. In acid solution (pH 2.9) the color becomes very
slightly redder. In alkali solution (pH 8.4) the color is
not changed (24). Ponceau Ex is the disodium salt of 2~(5-
sulfo-2,4-xy1yl-azo)ul-naphthol-4-sulfonic acid (10). Its

structural formula (24) is:

- CH:
H3C .NzN

Na038
N8033

 

 

F a Yel w N _ - a tra ne
FDdC Xellow No. 5, commonly call Tartrazine, is an
orange-yellow powder, easily soluble in water, giving a
goldeneyellow solution, There is no visible color change in

either weak acid or weak alkaline solution (24). Tartrasine

 

I Ffiaa signifies'Foods, Drugs and Cosmetics and means that
any certified color with this designation may be used in
foods, in drugs and in cosmetics,

8

is the trisodium salt of 5-carboxy-S-hydroxywl-pesulfophenyl-

4-p-sulfophenyl-azo-pyrazole (10). Its structural formula

N8035.N=N-fiwfi“COONa

- N
HO Q\\/

(24) is:

  

‘03Na
.01 r en
Because the list of permitted colors represents a se-
lection of shades far too limited to cover food industry
needs, certified blends containing two or more permitted
colors to produce numerous intermediate shades are avail-

able (60).

Evalugttgg gag fleastggment 9; 90122

According to Brice (4), the measurement and specifica-
tion of color is "color science" combining segments of phys-
ics, chemistry, physiology, and psychology for its complete
understanding. With modern photoelectric instruments, and
psychophysical data adopted by international agreement, the
measurement and specification of color can be quite straight-
forward.

Although color evaluations may be made either by subjec-
tive or objective methods, there is considerable controversy
as to the relative merits of each method“in evaluating color

in foods. Color evaluation in food products has become

increasingly important as food technologists place more em-
phasis on product quality, yet color is one of the hardest
quality factors to evaluate. To satisfy the desire for obs
jectivity in recording colors and to overcome the short-
comings of the human eye in color perception, various instru-
ments and devices have been constructed. Some instruments
answered a need; others exchanged one difficulty for another.
As color recording became mechanized, the interpretation of
the data became more difficult (18, 56).

’ b so, v ev uation

wilson et a1. (62) state it is desirable to establish
objective criteria which could be related to organoleptic
evaluations. However, no specific criteria were given. De-
velopment of such quality indices would greatly alleviate
the difficulties encountered in maintaining a trained panel
of judges and eliminate certain variations due to fatigue in
routine subjective testing of a large number of samples. The
subjective methods of color measurement and standardization
based on visual comparisons are subject to shortcomings of
human observers, such as the variability in the reactions of
different observers, and in those of one observer at differ-
ent times and under different viewing conditions, and the un-
reliability of color memory. The difficulties associated
with sensory testing have encouraged the development of ob-
Jective methods for the measurement of color which eliminate

the human retina in favor of the photoelectric cell (28).

10

b 0 ve eas rem nt

An objective evaluation of color may be made in a num-
ber of different ways, including the use of charts, discs,
and instruments. Maerz and Paul's Dictionary of Color and
the Munsell Book of Color provide charts for the visual com-
parison of colors. The Maxwell spinning disc principle has
been used for color measurement for over 100 years. Avail-
able types of instruments for measuring color include:
visual colorimeters, comparators, spectrophotometers, and
tristimulus photoelectric colorimeters (52).

Robinson et al. (56) report the Maxwell spinning disc
to be less objective than the photoelectric instrument, and
its use is more tedious and time consuming.

The Hunter Color Difference Meter and the Gardner Color
and Color-Difference Meter, which measure color directly by
reflectance, have been used for color measurements in many
research projects (5. 22, 34, 62, 63). Since the Hunter
Color Difference Meter had been used successfully to measure—
color differences in tomato juice, strawberries, and other
fruit and vegetable products, Huggart and Wenzel (22) under-
took to determine if this instrument would be satisfactory
for measuring the color of citrus juices and concentrates.
They concluded that although color differences can be objec-
tively measured with the Hunter Color Difference Meter,

neither this nor any other instrument available at that time

was satisfactory for determining the actual visual color of

11

a citrus juice as seen by an observer.

Longree (36) found the Hunter Color Difference Meter
unsatisfactory for evaluation of colors in baked custards.
The subjective method of evaluating color was ultimately
used in her study. According to Kefford (28), no one instru-
ment to date has been devised by which it is possible to
measure all the characteristics of colored bodies discern-

ible to the human eye.

Flavor in Foods

Ps c o ical A ects

According to Meyer (48), flavor is the subtle and com~
plex sensation that is the source of much of the delight man
finds in food. To both connoisseur and layman, flavor is of
utmost importance in determining preferences.. One's appreci-
ation of flavor and one's judgment are influenced by many
psychological factors. Some of the psychological factors
that cannot be controlled are the experience background of
the individual and his personal problems and reactions (15,
48) . I

Pettit (54) states the psychological aspects of flavor
include attitudes, experiences, memory, expectations, sug-
gestions, motivations, and other factors of learned behavior.
The setting, the questions, the personality and attitude of
the investigator may all recall past experiences and influ-

ence the consumer's interpretation of his immediate flavor

12

experience.

.Flavor memories are very keen and enduring and must be
taken into account whenever food flavors are-being originaw
ted or changed. If the first meeting an individual has with
a flavor is in medicine, there is a possibility that person
will always dislike that flavor. Sometimes "innocent" fla-
vors or textures such as those of oatmeal, cereal, cornstarch
pudding, or jelly recall with great distaste some episode of
sickness (12). I

A person's reactions toward flavors are exceedingly
characteristic and natural w a conservative person tends to
be conservative in the flavors he likes. However, a person
may train himself to recognize new flavors and to like them
(12).

Sternberg in 1914 as reported by Crocker (14), states
one naturally eats and swallows pleasantwtasting food rapid-
ly, rather than lets them linger in the mouth. If one holds
food long in the mouth, it is an indication the taste is not
entirely a pleasant one.

ApprOpriateness is of great importance in our enjoyment
of flavors in foods. An onion flavor may be delectable in a
stew or soup but objectionable in a custard. We become con-
ditioned to expect certain sensations from certain foods and
while a slight variation is titillating, a completely unex-
pected taste is unacceptable (48).

Caul (7) has analyzed the pattern of good flavor as the

13

following sensations: "(1) an early impact of appropriate
flavor; (2) rapid development of an impression of highly
blended and usually full-bodied flavor; (3) pleasant mouth
sensations; (4) absence of isolated unpleasant notes; and

(5) anticipation of the next mouthful."

actors nfluenc n l vor

Taste depends on a number of factors, the most important
of which is chemical composition. Saltiness is a prOperty of
electrolytes, the halides in particular. Sweetness is found
in a number of organic compounds, however, alcoholic hydrox-
yls are most effective in endowing a compound with sweetness.
Bitterness is a property of some organic and inorganic com-
pounds; some of the alkaloids such as quinine and brucine are
exceedingly bitter. Sourness is a property of the hydrogen
ion and its concentration is of primary importance in deter-
mining whether the sensation of sour is detected (48).

Taste is also influenced by temperature, texture, and
the presence of other compounds. According to Cracker (l2),
temperature has a less noticable influence on the sweet and
bitter taste components than it does on salty and sour com-
ponents.

Texture preperly is part of flavor; an unaccustomed
texture places the senses of taste and smell under a handi-
cap. Texture partially controls the quantity of sapid matter
that can reach the taste buds at a given time; as a substance

becomes thicker, the flavor becomes weaker. This weakening

14

of taste is probably mechanical because the viscosity of the
fluid interferes with the diffusion of the soluble substances
to the sensory receptors (12, 13).

Mackey and Valassi (41) measured the taste threshold
for four primary tastes (sweet, salty, sour and bitter) in
water as well as in tomato juice and in egg-milk custard,
prepared as liquid, gel, and foam. They found the primary
tastes were hardest to detect in the gel state, easiest to
detect in the liquid state, and intermediate in the foam.

The presence of other compounds has been shown to in-
fluence flavor in different ways. According to Pangborn (53)
the interactions among taste qualities in foods have been a
subject of opinion and speculation with relatively little
experimental support. The early literature on the subject
of taste interrelationships in aqueous solutions of pure
compounds is controversial since conflicting results were
obtained from similar experiments.

Results of studies on apricot, peach, and pear nectars
carried out by Valdes et a1. (59) show that sucrose enhanced
fruit flavor up to an optimum sweetness level, beyond which
it masked flavor. This relationship between sweetness and
fruit flavor was found to be greatly influenced by the acidi-

ty of the product.

Addition of Flavors to Food grgducts

Flavoring extracts are widely used in the food industry

and are well standardized in the United States. The standard

15

for a flavoring extract, as established by the Secretary of
Agriculture in 1919 and accepted by the flavor industry, is
as follows: "a solution in ethyl alcohol of preper strength
of the sapid and odorous principles derived from an aromatic
plant, or parts of the plant, with or without its coloring '
matter, conforming in name to the plant used in its prepara-
tion". If artificial coloring or synthetic flavoring com-
pounds are added, these must be declared on the label (23,
48).

Flavoring agents are chemical substances which are able
to impress themselves on the senses of taste, smell, and
feeling by way of food and drink. These chemicals are of
many types and origins; some occur naturally in foods; some
develop within food while it is being cooked or otherwise
processed; and others are added substances of natural or
artificial origin. Nearly all of the flavoring chemicals
that have been identified in foods have been duplicated syn-
thetically and, in many instances, these synthetics have
been used in flavors (l2, l4).

Imitation food flavors are composed of synthetic organ-
ic aromatic chemicals sometimes used alone or in combination
with natural products. These chemicals can be used as food
flavors for controlled accentuation of a particular note or
notes, where a highly concentrated flavor is required and
for economy (25).

From birth, modern generations have been reared and

16

accustomed to accept taste trends, and flavors are a daily
and accepted necessity in life. The desire for a variety of
palatable and appetizing taste sensations that makes the con-
sumption of food a pleasure, not a necessary daily task, is
a basic consumer characteristic of our times (25, 46).
Literature with regard to food flavors has been very
meager during the last fifty years, according to Merory (46).
During this same period of time numerous compounds have been
synthesized which are useful in formulating synthetic or imi-
tation flavors. Such compounds are known as flavormatics
and there are over 200 such flavormatics available for use
in imitation flavor compositions (25, 46). In 1963 the Fla-
voring Extract Manufacturers' Association issued a new list
of flavoring ingredients suitable for use in the preparation

of imitation flavors (16).

Fruit Flavors

The skin and the peels of fruit carry most of the fla-
vor. There is a correlation between sugar content, color,
acidity, and flavor as fruit ripens. The flavor usually de-
ve10ps to its fullest when the sugar content of the fruit is
at its maximum and the color of the skin or peel acquires

the richest shade of brilliance (46).

Classifications

There are several types of flavoring materials available
for use in food products. Legal definitions of these flavor-

ing materials are lacking and the issues of "Imitation" and

17

"Natural" flavors are being discussed at the present time by
governmental agencies and the flavor industry (16). The
flavor industry and unofficially the F.D.A. accept the fol-
lowing definitions:

:rue Fruit Flavoring - extracts or concentrates
derived directly from a particular aromatic
plant or fruit, of which a minimum of 25# of
fruit are used to yield 1 gal. of juice (16).

True- rui w th 0 he Nat ral av rs - fruit
flavors which are fortified with other natural
flavors. At least 51% of the flavor strength

must be derived from the true fruit named and
not more than 49% from other natural flavors

(23)

Imitation Fruit Flavoring - flavorings synthesized
entirely from chemica 8. There are no standards
set for these flavorings, except that their con-
stituents are to conform with the Food Additives

Amendment of 1958 (16).

Imitation Fruit Flavoring with True Fruit - a com-
bination of synt etic ngredients an true fruit
extractives in which not less than 5% of the
total flavor is derived from true fruit extrac-

tives (16).

Peach Flavor

The peach was one of the first fruits to be studied in
order to determine the chemicals responsible for the flavor
of the fruit. Power and Chestnut (55) state peach oil or
peach essences consist for the most part of purely empirical
mixtures of esters and essential oils with other more spe-
cific aromatic substances. The general character of these
empirical mixtures indicates some of their components may
not actually be found in the pulp of the fruit whose flavor

they are supposed to represent. For example, benzaldehyde

18

is frequently regarded as one of the proper constituents of
peach aroma, although this is contained in the kernels of
the peach and not in the pulp of the fruit.

Toward the end of the 19th century, the flavoring prop-
erties of gamma-n-heptyl butylrolactone were discovered and
the chemical was synthesized. This became known as Aldehyde
C-l4 which is a misnomer, since it is neither an aldehyde
nor does it contain 14 carbon atoms (27, 47). The formula

for Aldehyde 0-14 is: ‘ 7315

so

According to Merwin (47),cmce the existence of gamma
undecalactone became known, it became relatively simple to
prepare a flavor resembling peach. It is difficult for any
flavorist toconceive of a peach flavor today which will be
acceptable to the American public without Aldehyde 0-14.
Besides gamma undecalactone, those ingredients which would
be considered essential to peach flavor are benzaldehyde,
some of the linalyl esters and several valeric acid esters
(47).

I Many foods can be flavored or reinforced with a peach
flavor by using the preper level of gamma undecalactone and
adjusting the color and acidity of the food product involved.
Great care must be exercised because of the ease of over-

flavoring and obtaining an objectionable flavor (47).

l9

Custards as a Medium for Experimental Work with Color and
Flavor
Baked custard represents a type of cooked product in
whicthhe thickening property of the.egg protein is used,
since only 0.75% of the heat‘coagulable protein is provided
by the milk (38). Color, flavor, and odor are also detected
easily in custards as they are made of relatively few essen-

tial ingredients (44).

Composition

Custard is made from milk, eggs, and sugar and some-'
times salt and vanilla flavor. The same proportions of in-
gredients can be used for both stirred and baked custard
(20). The baked custard is firmer than the stirred custard
and appears to be in one piece or clot. All custards are
highly sensitive to slight modifications in the egg-sugar—

milk mixture.

Heat Coogulogion of Erotoins

The formation.of a gel in custards depends on the coag-
ulation of protein which then holds within its structure the
solution from which it was precipitated (51). Although pro-
tein may be coagulated by many means, heat has been one of
the most important methods. Heat coagulation of the egg and
milk proteins in custards may be influenced by acid, alkali,

salts, and organic solvents, as well as other factors (3, 38,

51).

2O

Effeot of Acid, Alkali, Salts ond Qrganio §olvonts

Weiser (61) indicates that certain protein sols (albu-
mins, globulins, and globins) coagulate at some definite
temperature which is fairly constant for each protein. With
the lowering of the pH by the addition of acid, the coagula-
tion temperature is raised. Chick and Martin (8) point out
denaturation, the first part of the heat coagulation process,
is not hastened by acid solution, but the clotting or coagu-
lation process is accelerated. The second part of the coag-
ulation process, the clotting or coagulation, does not occur
in alkaline solution. If, after heating, the alkali is neu-
tralized with acid, the coagulation process occurs (9).

Although the salt content of milk is low, it is very
important in the behavior of the proteins in food prepara-
tion (38). A good example is custard, which fails to thicken
when heated unless sufficient salts are present. Lows (38)
reports the concentration of the salt and the valence of the
ion have an effect on the coagulation of custards. Chick and
Martin (8) state an increase in the concentration of salts
raises the coagulation temperature. Many anions such as
iodide,bromide, and chloride are denaturing agents (48);
however according to Fruton and Simmonds (19), not all anions
are denaturing agents - caprylate and aromatic carboxylate
ions actually protect egg albumin from heat denaturation.

Denaturation may be caused by organic solvents such as

ethanol, methanol, and acetone. These have been used for

21

the precipitation of proteins from aqueous solutions (19).
There is‘a decrease in the protein stability of milk toward
alcohol as a result of homogenization, according to Carr and

Trout (6).

ate d i x d C e One
The pH of the eggs used in custard has an effect on the
pH of the custard mix - fresh eggs are less alkaline than
aged ones. Cooked custards are usually more alkaline than
the custard mix (3, 35, 38). Bittner (3) reports the pH
readings on the custard mix and baked custard made with fresh

eggs and dried whole milk as 7.03 and 7.09 respectively.

Gelation Tomoerature ond Rooe of Cookiog

Many factors affect the coagulation of protein and the
gelation temperature of custards. These factors include the
proportion of ingredients, the pH of the mixture, and the
rate of cooking. As a result, ru>one internal end tempera-
ture can be recommended as the optimum temperature to which
a custard should be baked. Lows (38) states with ordinary
rates of cooking, and using unhomogenized milk, a serving
consistency for custard is usually obtained between 82° and
84°C. Carr and Trout (6) state the rate of heat penetration
is slower in custards made with homogenized milk than in
those made with unhomogenized milk. The firmness of the cus-
tards indicates the custards prepared with homogenized milk

could be baked to a higher internal end temperature without

22

seriously affecting the stability of the gel. Miller et al.
(49) observed the optimum gelation temperature of custards
prepared from fresh shell eggs and homogenized milk to be
86° to 88°C.. Cook and Husseman (11) state the temperature
required to reach gels of similar consistency is 1°C higher
when dried whole milk is used instead of fresh whole milk.
Lows (38) reports the rate of coagulation increases as
the oven temperature is raised; in addition, a firmer custard
is obtained as the temperature is increased until at a spe-
cific internal temperature, dependent on the rate of cooking,
optimum gelation occurs. Further heating to a higher inter-
nal temperature causes porosity and finally syneresis.
Experimentation shows the rate at which custard is
cooked affects the gelation temperature - with a slow rate
of cooking, gelation of custard occurs at a lower tempera-
ture, while gelation takes place at a higher temperature
with a fast rate of cooking. A slower rate of cooking is
considered most desirable because optimum gelation is more

easily perceptible (38).

Meghod of Baking

Two factors to consider in baking custards are the oven
temperature and the temperature of the water bath. Griswold
(20) states even when baked at relatively low oven tempera-
tures, custards are improved by being placed in a pan of hot

water for protection from oven heat. Bittner (3) and Carr

23

and Trout (6) used an oven temperature of 325°F in their
research work with custards. McBride (45) and Mastic (44)
regulated the oven at 350°F. Jordan et al. (26) report an
oven temperature of 350°F is preferred to 325°F or 400°F
because the baking period is unduly long at 325°F and so
short at 400°F that the custard can easily be overcooked.
In addition, custards baked at 400°F are reported less de-
sirable in appearance than those baked at the other two temw
peratures. To insure the relatively long heating which is
necessary to make the food product bacteriologically safe,
it is recommended that an oven temperature of 350°F or
below be used for baking custards (37).

_For many years the recommended initial temperature for
the water bath for baked custards has been boiling or hot
water. Various initial temperatures have been reported for
the water baths used in research work on custards in the
past few years: Bittner (3) used 35°C, Carr and Trout (6)
used 30°C, and Mastic (44) used 250-2700. Nagler (50) re-
ports the texture of the custards was more delicate and ten»
der if the initial water for the baking pan was 35° to 40°C
rather than 97° to 100°C. The results of a special problem
undertaken by the writer at Michigan State University in
1963 on the hot water bath versus the cold water bath for
baked custards indicated that custards started in a 20°C
water bath were superior to those started in a 40°C water

bath for the following characteristics: inside appearance,

24

crust tenderness, texture, flavor, consistency, syneresis,

and general acceptability.

Use of Qgied Whole Milk in Bokeo Cusgards

Bittner (3) reports the results of using dried whole
milk (DWM) in place of other milks in baked custards, all of
which were baked to 86°C. Baked custards prepared with DWM
were similar in appearance to custards made with homogenized
milk or with pasteurized milk, however the crusts of the DWM
custards were slightly more wrinkled and less tender than
those made from homogenized milk. .The inside color for the
DWM custards scored slightly higher than when homogenized or
pasteurized milk was used. The DWM custards had a highly
significantly different flavor than pasteurized milk cus-
tards — a decided cooked flavor was indicated and the DWM
custards were significantly sweeter than those made with
pasteurized milk. The DWM custards were firmer than cus-
tards made from homogenized milk but less firm than those

made from pasteurized milk.
Evaluation of Foods

Food products can be evaluated by either of two methods:
the subjective method in which a particular characteristic
of a product is scored or otherwise rated by an individual
with a score being decided upon by judgment, or the objec-
tive method where the outcome is largely independent of

human judgment (43).

25

Although the limitations of a taste panel for judging
the palatability or eating quality of food are realized by
investigators, there are still organoleptic factors which
cannot be expressed by objective measurements. Factors such
as appearance, color, and flavor are better judged by a
scoring panel. Results of objective tests may often be cor-
related with results obtained by subjective means (3, 44).

Lowe and Stewart (39) state it is desirable to employ
both subjective and objective tests in connection with re.
search on the functional and organoleptic preperties of food
products. During the development of an objective test it is
particularly appropriate to run parallel organoleptic tests

in order to test the degree of correlation between them.

Subjecgiye Metoooo

Subjective tests measure the qualities of food as they
make their impression individually and collectively on the
sensory organs. They are subjective because the individual
is required to go through a mental process in giving his
opinion as to the qualitative and quantitative value of the
characteristic or characteristics under study.

Typos of goats and judgoents

According to Lowe and Stewart (39), subjective tests
may be conveniently classified into 2 categories: a) prefer-
ence or acceptance tests and b) psychometric or difference

tests. In preference testing the degree of acceptance is

26

obtained and, when conducted with a large enough number of
people within a population, permits the determination of
consumer acceptance of a product.

The psychometric tests are for determining quantitative
differences, such as rating or scoring food quality factors.
They can be used to evaluate quality differences among breeds,
varieties, or formulas. This makes these tests valuable re-
search and development tools (39),

Mason and Koch (43) state that two reasons for choosing
a scoring system in preference to a ranking system where the
treatments are ranked according to a preference for a given
product are:

l. a scoring system permits ties, whereas a
ranking system forces judgment even though
the taster can detect no definite difference.

2. a scoring system permits the spread of treat-

.- ments to be influenced by the magnitude of the
differences found. If there are very distinct
differences between two samples, one may be
scored 1 and one scored 10, but in rank ng
they must be 1 and 2.

Often a numerical scale is used so that the scores of
the individual taste panel members can be added readily to
give a composite score (48).

e de t ns 0 t ste e 8

Owing to practical considerations, most panels are com-
posed of from four to twelve members and if only three or four
acceptable judges are available, results comparable to those

of a larger panel can be obtained if each sample is scored

two or three times during the investigation (20, 30, 39).

27

According to Dawson et al. (15) a panel of three to ten
is usually adequate for testing flavor differences but panel
size depends on the variability associated with the experi-
ment and on the magnitude of the difference between samples
to be detected. A

Experimental work by Langwill (33) indicates that women
between the ages of seventeen and thirty years of age have a
greater sensitivity of taste in distinguishing between sweet,
salty, sour, and bitter foods than do men in the same age
bracket. Laird and Breen (32) report from their study that,
in comparison with men, women at all ages have more prefer-
ences for tart tastes, and less for sweet tastes,

The age of the judge is also a factor. Young people
can distinguish differences between different strengths of
sweet, sour, and salty food when middle-aged peOple cannot
(l). Sensory ability decreases with age and preferences
change also. Because of this it is felt that taste panel
members should be between the ages of twenty and fifty (31).

Objective Methods

Types of objective evaluations are many and varied, in-
cluding chemical, histological, and physical tests, Griswold
(20) states objective methods are reproducible and less sub-
ject to error than sensory methods.

There are many examples in the literature of objective
tests for measuring such varied acceptance characteristics

of foods as texture, color, tenderness, consistency, and

28

juiciness (39). There is no objective method for measuring
flavor pggwgg. However, some flavor components may be meas-
ured by chemical means, i.e., sugar, salt, and acid.
According to Lowe and Stewart (39) objective tests pos-
sess obvious advantages with respect to reproducibility and
in the ease of applicability to the needs of the research
laboratory. However, the ultimate test of whether an objec-
tive test is measuring a quality is its agreement with re-

sults of subjective testing (20).

EXPERIMENTAL PROCEDURE
Preliminary Investigations

Preliminary experimental work with the addition of food
colors to baked custards indicated the possibility of devel-
oping in baked custard several pastel colors typical of dif-
ferent varieties of fruit. Strawberry, raspberry and peach
colors were shown to be feasible, and of these, peach showed
the most promise..

The kinds and amounts of food colors required to pro-
duce a peach color in baked custards were determined in this
preliminary study. Various combinations of spec Red No. 4-
and Yellow No. 5 were found to yield the most satisfactory
range of tints and shades of peach color in baked custards.
An arbitrary decision was made as_to the color range selec—
ted for study. Two randomly selected preference panels of
twenty persons each indicated which of these colors would be
most acceptable in baked custards.

After the colors were established, various types and
amounts of peach flavors individually and in combination
were tried in the baked custards. 0n the basis of the trial
tests, an arbitrary decision was made as to the flavor char-
acter and strengths that would be included in the study. A
randomly selected preference panel of twenty-two persons
evaluated custards containing three different combinations
of peach flavors and indicated that all three combinations

29

30
were acceptable.
Design of ExPeriment

Baked custards prepared from a standard formula with
the addition of all possible combinations of three different
peach colors and three levels of the same combination of
three peach flavors were compared subjectively and objec—
tively, A control custard having neither added color nor

added flavor was used as a reference for the objective tests.

golgg-Elavgr Qombinatiogs

The nine treatment variables (A through I) represent all
possible combinations of three different peach colors and

three levels of the same peach flavor.

_ C Elavgr Lgvgé
Variable Eight Me u 2355, .£23 he gm
x

A x
B x ' x

c x x
D x

E x

F x
G x x

a x

I x

Quantities of the colors and flavors for the nine treat-

ment variables are presented in Tables 20 and 21, pages 88

31

and 89 in the Appendix.

Six replications of each of the nine color-flavor com-
‘hinations were prepared and evaluated. The baked custards
were cooled to room temperature by a standard procedure be—
fore evaluation. Objective tests were conducted the first
three weeks of the eXperiment and the subjective tests were
conducted the following six weeks. Certain of the objective
tests were performed daily on custards made from the same
mix used for panel evaluation. This procedure was followed
as a check against the objective tests previously conducted.

Four treatment variables1 were prepared each day for the
objective tests and baked in two batches. 'Sixteen custards,
eight each of two treatment variables, werebaked at one
time. Of these eight custards, seven were used for objec~
tive tests and one to obtain a timentemperature record during
baking.

For the subjective tests, three treatment variables
were prepared each day in two bakes. Sixteen custards were
baked at a time. In the first bake four custards of each of
the three treatment variables were used for the taste panel,
one custard of each of the three treatment variables was re-
served for color determination. and one custard used to ob-
tain a time-temperature record during baking. .

The experiment had to be modified with respect tO‘the

objective determination of color. Due to mechanical

 

I 'Infthis'manuscript, treatment variable refers to a single
color-flavor combination.

32

difficulties it was impossible to collect reliable data,
therefore no further comment will be made on this point.

A predetermined randomized arrangement was used for the
preparation of the treatment variables throughout the experi-
ment and for the order of presentation of the samples to the
judges each day. The weakest flavored custard in a group
was always,presented first to the taste panel.

Room temperature and humidity in the experimental foods
laboratory were recorded on each day that custards were

baked and evaluated‘.
Basic Custard Formula

The formula selected for study. based on that of Lows
(38) consisted of 3815 grams of reconstituted whole milk,
748 grams whole egg and 391 grams sucrose. Percentages of
ingredients, based on total weight of the mixture. were:
whole milk, 77.0; whole egg, 15.1; and sucrose. 7.9. The
percent ratio of dried whole milk to water was l2.6 to 87.4.

Ingredient Procurement

la ents
A common lot of dried whole milk was prepared to speci-
fication by the Michigan State University Dairy Plant. The
reconstituted milk contained 3.2% butterfat. The dried
whole milk was packaged in small sealed polyethylene bags,

 

' hygro-Thermograph, Model 594, Price Instrument Division,
Bendix Aviation Corporation, Baltimore. Maryland.

33

each containing 504 grams, and refrigerated at 38-400F until
needed. The dried whole milk was removed from the refriger-
ator and brought to room temperature before reconstitution.

Fresh eggs were obtained from a designated flock of
White Leghorn hens at the Michigan-State University Poultry
Farm. The eggs used were fresh daily. ‘Egg yolk color de-
terminations were made at the beginning and in the middle of
the experiment. The Cargill-Nutrena Yolk Color Meter read-
ings were 5-6. The Heiman-Carver Yolk Color Roter determina-
tions were 11-12 on all samples tested (21).

A common lot of sucrose. obtained from the Michigan
State University Food Stores, was packaged in closed poly-
ethylene bags, 391 grams per package. and stored in a covered

metal container at room temperature.

The FD&C certified food colorswere obtained from Food
Materials Corporation in dry, powdered form and stored at

room temperature in their original glass containers.

Eiélfififi

In the preliminary work no single peach flavoring materv
ial imparted a satisfactory peach flavor to the baked cus-
tard. It was necessary to use combinations of peach flavor-
ing materials. A combination of three different peach fla-

voring materials was selected. A description of the flavors

and the quantity used are found in Tables 20, 21, and 22.

34

pages 88, 89, and 90 in the Appendix. The flavoring materi-
als were used full strength as received from the manufac-

turer.
Preparation of Color Solutions

For addition to the custard mix, three color solutions
(light, medium. dark) were prepared by dissolving varying
proportions of spec Red No. 4 and spec Yellow No. 5 in dis.
tilled water. The weights of the two colors used in each of
the color solutions are found in Table 23 in the Appendix.

Preparation of Custard Mix

Each day 4500 c.c. of the basic custard mix was pre-
pared. The detailed procedure for the preparation of the
custard mix is given in Exhibit 1. page 92 in the Appendix.
Fresh, shell eggs were broken out. blended, and weighed.

One package of prevweighed sucrose was added to the eggs

and the mixture blended. The reconstituted milk was added
to the egg-sugar mixture and the entire mixture was blended.
The custard mix was strained through a mediumwfine, wire-
mesh household strainer to remove unmixed material prior to
the addition of color and flavor.

The basic custard mix was divided into smaller portions
for addition of color-flavor variables. For the objective
tests the basic mix (4500 c.c.) was divided into four por-

tions of 1100 c.c. each. For the subjective tests it was

35

divided into three portions of 1500 c.c. each.

All mixings to incorporate colors and flavors were done
on Hobart K—SA mixers equipped with quuart stainless steel
bowls and a paddle attachment. Portions of each final mix~
ture reserved for the second bake were left at room tempera-
ture in the mixing bowls and covered with polyethylene

coated freezer paper secured with a rubber band.
Baking Procedure

The custards were baked in Seounce pyrex custard cups.
Each cup was filled with custard mix to a measured depth of
1 3/4 inches (approximately 130 c.c. custard mix). The cups
were placed in a specially designed galvanized water bath
pan equipped with a bronze stepcock for drainage and a per-
forated stainless steel sixteen-hole rack to support the
cups (Figure l). The cups were placed in a predetermined
randomized order for the subjective and objective tests.

Water at 18—20°C was added to the pan until it came up
to the level of the custard mix. The baking pan was placed
on an oven rack 3 7/8 inches from the bottom of an apartment
size General Electric Oven preheated to 350°F. Three lead
wires from a Brown Eleotronik Potentiometer High Speed Multi—
ple Point Recorder were clamped in place.‘ One thermocouple
was centered in a custard located in a central position of
the baking pan, a second was positioned in the water bath

just to the right of the custard containing the first

37

thermocouple. The third thermocouple recorded the tempera—
ture at the center of the oven. All custards were baked to
an internal temperature of 86°C. The oven was turned off
and the water partially drained out of the water bath pan
through the stopcock. The rack containing all the custard
cups was then removed from the water bath pan. The oven was
again preheated to 350°F and the second bake was carried out
following the same procedure as the first bake. The cusq
tards were placed on wire racks and allowed to cool at room
temperature for approximately three hours before objective

or subjective tests were conducted.
_Subjective Tests

The samples for the subjective tests were selected so
that, in the course of six replications of each variable.
each judge had an equal number of samples from inside bake
and outside bake positions (see diagram below) as well as

front and rear oven bake positions.

 

. Outside bake
0 Inside bake

 

 

 

 

The panel was composed of 7 judges (4 women, 3 men)
from several related departments of the university. A one-

hour training session was held for all taste panel members

38

at which they were given instructions as to how the samples
would be presented. how to score the factors and the share
acterietics of a good baked custard. A sample of the in-‘
struction sheet is shown on pages 94 and 95 in the Appendix.
To acquaint the taste panel members with peach flavor, sever-
al brands of peach ice cream were tasted.

The taste panel evaluated the room temperature custards
for nine characteristics; crust color, crust tenderness.I
crust flavor, inside color, aroma. inside flavor. consisten-.
cy, texture, and syneresis. Three different custard varia-
tions were scored each day. The custards were presented
one at a time and the panel was requested to evaluate each
sample individually without comparison with the other sam-
ples.

A 7epoint rating scale was used in the evaluation with
7 as the highest score. The panel was instructed to score
each custard on all nine characteristics and to check the
most appropriate descriptive terms for those factors they
scored 4 or below. A sample of the score sheet is shown on

page 96 in the Appendix.
Objective Tests

On the day of preparation, objective tests were made on
the room temperature custards to determine the gel strength,
measured both as the ability to resist penetration and as

thegability to hold a rigid shape; syneresis; and pH.

39

Continuous recordings of room temperature and relative hu-

midity were made during each test period.

e tr 1 .
Gel strength was determined by a Micrometer Adjustment
Penetrometer1 equipped with a 35-gram grease cone attach-

2. Each sample was placed on a level platform directly

ment
beneath the cone attachment, with the point of the cone just
touching the surface of the custard. The cone was released
for a period of 5 seconds and the depth of penetration.re-
corded to the nearest 0.1 millimeter. Measurements were’.

. made both with the crust on (crust loosened from the edges
of the cup) and the crust off (crust entirely removed). Two
different samples were used for each of these measurements:
one from an inside bake position and one from an outside
bake position. Readings were recorded and averaged. Figure

2 shows the measurement of penetration after the release of

the cone..

Pe gent Sag

1h order to compare the firmness of the different cus-
tards, measurements of the height of the inverted sample be-
fore and after a specified time interval were made. The pro-'
cedure used for this objective test was that described by

Zhbik (64). The apparatus shown in Figure 3 was used to

 

; 715thur H. Themes Co.

Precision Scientific Co. - Cat. No. 73525

41

minimize damage to the gel structure of the sample during
testing and to facilitate the collection ofsuch data.' The
per cent sag for each sample was calculated from the differ-
ence between readings divided bythe initial height of the

inverted custard.

§yngggsi§

Th. method of Miller et al. (49) was used to determine
the weight of the drainage from the custards. The equipment
for this test is shown in Figure 4.

The baked custard with the crust intact was loosened
carefully and inverted on fine wire screening (15 wires per
inch) which had been placed over a weighed Petri dish. The
assembly was immediately covered with a glass bowl to pre-
vent evaporation and allowed to stand one hour. it the end
of this period. the wire screen with the custard was removed
and the weight of the drainage recorded to the nearest 0.1
gram. Per cent drainage was calculated from the ratio of
the weight of the drainage to the weight of the sample be-

fore inversion.

£2

The pH of 50 ml of the fluid custard mix was recorded.
The pH of the baked custard was determined after thoroughly
blending one custard (approximately 130 c.c.) in a Waring

Blender for one minute. All samples for the pH determina-

tions were at room temperature (220-2506) and the

(\)J. J! J: «4 [N 9.4 i‘Jl W Isl ll \ ‘1 1|: 0

 

43

measurements were made on a Beckman‘Zeromatic pH meter with

a glass electrode and a saturated calomel electrode.
Statistical Methods

The data obtained from the subjective and objective
tests were evaluated by the use of two computer programs on
the CDC 3600 computer at the Michigan State University Coma
puter Center. The FAOREP Routine 2 Option 1 (one-way face
toriel with replications) was used to calculate analyses of
variance and the CORE Routine was used to calculate simple
correlations. Significant differences among variables were
evaluated through use of the Studentized range tables (17).
An illustration of the Studentized range calculations for
one of the quality factors is shown in the Appendix, pages
97, and 98. 1 ‘

Correlation coefficients were calculated on the follow-
impairs of items: penetrometer readings (crust off) versus
per cent sag, penetrometer readings (crust off) versus con-
sistency, per cent sag versus consistency, per cent drainage
versus syneresis score, texture versus syneresis. crust
color versus crust flavor, inside color versus inside flavor,
crust color versus inside color, crust flavor versus inside
flavor, inside flavor versus aroma, and consistency versus

texture.

RESULTS AND DISCUSSION

The baked custards prepared from the nine possible com-
binations of three different peach colors and three levels
of peach flavor were evaluated bysubjective and objective
tests. The identifying code for the combinations of colors
and flavor levels used as treatment variables is as follows:

Variable legre Elgvgg Qombinatign

Light color. low flavor level
Light color, medium flavor level
Light color. high flavor level

Medium color. low flavor level
Medium color. medium flavor level
_Medium color. high flavor level

i Dark color. low flavor level
Dark color. medium flavor level
Dark color, high flavor level

CONTROL No added color,-no added flavor

Hm$ wmu am»

The scores obtained from the objective and subjective
tests were analyzed by a oneeway analysis of variance and
by the Studentized multiple range testril7) when significant
differences were found in the analysis of variance.

The mean taste panel scores and objective test readings ‘
for each replication of the baked custard variables and the
standard deviations of the means are given in the tables
accompanying the discussion of the results for each of thei.‘

quality factors.

44

45

Subjective‘Tests

Crust Color

The mean scores and standard deviations for crust color
are listed in Table 1. Analysis of variance of these data
(Table 2) revealed significant differences in crust color
attributable to the treatments. ComparisOn of treatment
means indicated custard F scored significantly higher than
custards A. I, H,.B, and C at the 1 per cent level of prob-
ability and, in addition, significantly higher than custard
G at the 5 per cent level of probability. Custards F. D.
and E were not significantly different at the 5 per cent
level of probability, indicating for crust color the medium
peach color was equally acceptable regardless of the flavor
level used, Custards A, D, and G (low flavor level) were
not significantly different at the l per cent level of prob—
ability, indicating for crust eclor the low level of flavor
was acceptable, regardless of the peach color used. However,
the effect of flavor on color does not appear to be con-
sistento

The Judges indicated the crusts of the light colored
custards (A, B. and C) were too yellow. too pale. too light.
and unappetizing. This objectionably light crust color ac-
counted for the low scores assigned custards B and C butp
does not explain the higher score given to custard A. The

higher score for custard A (light color, low flavor level)

may have been due to first samplelbiaa since this sample was

46

 

 

4.71

Table 1. Mean Scores and Standard Deviations for Crust
Color.

Variw e c ns Grand
ssia._n..l;--w.-2”mq---l. WW,,... all, an fi.lfl sass..ss2i
A1 5.142 4.83 5.00. 5.29 ~ 4.45 4.85 4.92:0.30
B 5.57 5.43 3.57 4.00 4.00 5.57 5.69:0.25
c 3.71 3.43 3.50 3.86 _ 3.86 x 3.71 3.68:0.18
D 5.00 5.14 5.00 5.57 5.29; 5.71 5.29:0.30
s 4.71 5.83 5.50! 5.29 4.86 5.14 5.22:0.41,
F 5.57 5.71 6.00 5.71 .5.29 5.55 5.6oto,27'
e ‘ 4.86’ 4.86 4.57 5.29" 5.86 ' 5.14 5.10:0.45
H '4.29 4,43 4.83 4.57‘ 4.86  4,67 4.61:0.22
1 5.00 5.00 '5.14 4.45 4.57» ‘ 4.81:0.28

 

rw-Yr'vv'vr' rry—wrv

WV“ rrvwav

.v—v Y’W‘ .

vmw—‘Y'y-f vvr 1‘—

IF-“ firv

.1-wfiv-1Vw—Y'r‘v

r—wwv

W vwv-Y ‘VV'va' '-

1 Refer to code on page 44 for variable identification.

2 Mean score of 7 Judges' evaluations of 1 replication of

each variable.

1W

47

Table 2. Analysis of Variance for Crust Color.

 

 

 

 

 

333333"3? Degrees—SF‘ “ “1635‘"“ “T” "F'""
I££$£££2e;. “.m, aa’232.‘Zv‘..§$.1.S?..‘¥iw :w.ll:§£22£sw.w.-sn--R&£12.g.
Treatment means 8 . 2.7569 29.41aa_1
Error 45 0.0937 L

Total 53

** Significant at l per cent level of probability.

 

a- ,lmmll_.l_,mw-, a, . w . ,.
Duncan's Studentized Multiple Range Test]

 

1% Level:

13§81362 4H 61 4181 4. 92 §,-;Q 5:22, figfifi. §;_Q

1*er w-v—v—wv ‘1‘v— 'm ,_., Y'

 

 

 

 

.l T “1...? a _.,.. w“ ., W Y
5% Level°
. .
- —vv 7 -V W y— V7 vw
_ l l
uvw T fiwar —r—r . ‘r —rv v r—‘r—‘VVV‘ w- —v—- v- w—nt t r v y —-r v v—v 17 “V wv—— thww '7' ~, W .V'V‘ “‘V ‘

 

‘ Means underscored by the same line are not significantly

different (17).

48

always evaluated first whenever it was served.

The dark colored custards (G, H, and I) scored lower
than the medium colored custards, although treatment H was
the only one of the three that scored significantly lower
than any of the medium colored custards (D. E, and F). The
Judges indicated the crusts of the dark colored custards
were too red, too dark, and more orange colored than peach.

It was observed that the crust color was consistently
darker than the inside color for all the variables and the
baking procedure did not produce browning of the crust for
any of the samples regardless of the peach color used.
Figure 5, page 49, illustrates'crust colors for the three

different peach colors used in the study.

r t T _ se-
The mean scores and standard deviations for crust ten—
derness are listed in Table 3. 'Analysis of variance of
these data revealed no significant difference in the crust

tenderness of the nine treatment variables.

qust Elavog

Table 4 lists the mean scores and the standard devia-
tions for crust flavor. The individual custards were sig-
nificantly different for crust flavor only at the 5 per cent
level of probability as shown by the analysis of mariance,
Table 5. Comparison of the variables by the Studentizcd

multiple range method indicated custards F, G. and D scored

49

 

 

 

_’_——

Figure 5. Three different peach colors1
in baked custards (crust on) .

 

 

W“ ————_— __———~_~—_—

‘Figureug:j_Three different peach colo s
in inverted baked custards .

 

 

I 'Zér£"1¢ night 3'light color, medium color, dark color.

Table

7.
JO

Tenderness.

50

Mean Scores and Standard Deviations for Crust

 

 
 

 

' w‘v—vwv

Grand

 

: - , - Massage 5:94.
A‘ 4.149 4.55 5.17 6.00 5.45 4.67 4.96:0.71
B 5.57 4.57 5.29 5.43 5.29 5.43 5.26:0.36
c 5.00 5.00 5.67 5.45 ' 5.57 6.00 5.45:0.59
D ,.5,71 5.71 4.50 5.7i 5.86 6.14 5.61:0.57
s 4.00 » 5.67 5.50 6.00 6.14 5.86‘ 5.53t0.78
F _ ,»5,00 5.29 5.57 5.57 5.29 5.50 5.37:0.22
0 4.86 5.14 5.29 5.29 5.71 5.86 5.36t0.37
H _ 5.14. 4.86 5.67 5.14 - 5.45 5.50 5.29:0.50

‘1 4.43 ' 5.35 5.55 5.71 7 5.14. 5.57 5.25:0.45

'Y ,. v WW7, yrs-"fir wrr'. Y V‘w—w-w yw.~'w~—vw. Y'“

‘ Refer to code on page 44 for variable identification.

2 Mean score of 7 Judges' evaluation of l replication of
each variable.

51

 

 
    

 

 

 

Table 4. Mean Scores and Standard Deviations for Crust
Flavor. -
,m.-1w 1 - a 111m-“ m. 1--.;anwn
Variw Grand
. . 1 . .. .1! 1 11.....- 1.1MM8 .‘ S D
4' 4.292; 4.67 5.17 5.71 5.45 4.67 ~ 4.99:0.54
B 4.45 5.00 4.45 4.57 5.00 4.86 4.72:0.27
0 4.57 4.86 5.35 5.14 4.57 4.86 _ 4.89:0.50 '
D 4.57 5.14 5.55 5.71 ‘ 5.57 5.45 5.2940.40
s 5.00 5.17 5.00 5.29 5.57 5.43 5.24:0.23
F 5.14 5.45 5.45 5.57 5.45 5.50 5.42:0.15
G 4.45 4.71 v 5.45 5.57 6.14 5.71 5.35:0.64
H 5.14 5.00 5.00, 5.00 5.00 5.85 v5.16t0.33
I 5.14 4.50 5.50 4.86 4.14 4.8840.49

4
«i
J
1

r- rvvww—rvuv—un

 

V‘W‘vuV—rva w—y Irv-'nw‘ v w'v' ‘rv '- u—"' w

1 Refer to code on page 44 for variable identification.

2 Mean score of 7 Judges' evaluation of l replicationof

each variable.

52

Table 5. Analysis of Variance for Crust Flavor

 

senses of ” "'"W Degreee‘sr 5*" " “ " ‘ wean" WW” ”3‘"

 

 

 

V. 3"? 11- am .,. 1-11.11. W -11 5.4.212.
Treatment means 8 . 0.3514 2,17e
Error 45 . ‘0.1619

Total 53 ‘

 

W 77‘" yv . v w -: v T 1 VV- 1 wyw . v‘xv v. “Yr—y . v.‘ —u (xv—f— V‘V‘l wrr—l rwfi w—-- v— ,fiv‘ W
1

* Significant at 5 per cent level of probability.

 

r—V—fi w 7* —--y—-7 77' —r 7 rwr .fiwvv-w v—w-ur .—.,,Ylvl rvvr—r-w v v77r T . -— WW v—wr ’T' '—V . 1 WV

Duncan's Studentized Multiple Range TestI

 

V—‘V'Tl—wa‘ —-w w r Y ry r—w W. 7 vvv—w V7 w ‘fi-v if" W q—q w—w ‘7 ‘I' w—w—— T r w .- Yr—w 7—71— 'w‘ 1 quu— v e‘V—Wr—

5%Level;

 

 

s 1 . 0 1 a E 0' 0 F
4.72 $45—55 4.89 44995.16 5.;4 5,32 5,55 5.42

 

' Means underscored by the same line are not significantly
different (17).

53

significantly higher than custard B at the 5 per cent level
of probability.

Panel members' comments indicated the crust flavor for
custard B (light celor, medium flavor level) was too weak,
artificial, bitter, and slightly offvflavor. -These results
indicate the possibility of a "halo effect" of crust color
impression on crust flavor Judgments since custards D and G
were low flavor level, mediumand dark color respectively,
and scored higher than custard B which was medium flavor
level but light color. It was also noted the flavor was
more concentrated in the crust than in the inside portion

of the custard.

Insi C r
The mean scores and the standard deviations for inside

color are listed in Table 6. Analysis of variance of these
data (Table 7) revealed significant differences for inside
color at the l per cent level of probability. Comparison of
treatment means indicated custard G scored significantly
higher than custards A, B, and 0 (light color) at the l per
cent level of probability and, in addition, significantly
higher than custard E at the 5 per cent level of probability.

. Custards D, E, f, G, H, and I were not significantly
different at the l per cent level of probability, indicating
for inside color, the medium and the dark peach colors were
equally'acceptable regardless of the flavor level used. Cus-

tards A, B, and C contained the light peach color and scored

Table 6 0

Color.

54

Mean Scores and Standard Deviations for Inside

 

Hmmunriuow

 

 

4.292 4.83
4.29 5.86
4.29 5.86
4.86 5.14
4.71 5.35
5.00 5.00
5.71 5.29
4.86 4.86
5.14 5.17

 
         

 

4.83

3.86
4.17
5.17
5.17
5.57

‘5.14

5.00
4.83

fr—vyrwu-l-

4.71)

4.00
3.57
5.57

5.00 N

5.43
5.43
5.29
5.29

 

4.85
4.00
5.86
5.29
5.29
5.50
5.29
5.55
5.14

4.58:0.35
5.95:0.24
3.8930.30
5.22:0.23
5.04:0.28
5.27c0.26
5.45:0.28
5.1l20.22
5.12:0.15

 

y.-

‘1—1

v v w

vw—r vww'wrfi‘rn'v " Ir

Trv‘ v

w ‘v-‘wv w

“.7 ww—fi.‘ "VFW ‘VT'W'F' 'fi‘y

w—r—w-r—v I'r'r rv' .-q var—vv—w—Tv

1 Refer to code on page 44 for variable identification.

2 Mean score of 7 Judges' evaluations of_l replication of

each variable.

55

Table 7. Analysis of Variance for Inside Color.

 

p. with . .1. .4 r—ww m1... w .7..- -wl—r W.T . “.4 m. ”m
Eource of fiegrees of Mean F

 

 

 

Treatment means 8 2.0232' 29,55as
Error 45 0.0685
Total 53

*.. Significant at l per cent level of probability.

 

7— r... y y vwwvwvv—wrr-w v y 7] why '77 vr.‘ . v ‘— w—v'vvwfi .. W m ‘ we

Duncan's Studentized Multiple Range Test‘

 

1% Level:
G B A E H I D F G

5% Level:

 

 

rv—‘V‘VT 1 r wv vr' r‘ Yvwr'W-m'r , “VT" I'Vr-w vv-r #7 w . .7- r 11". I-vyryrv- ‘r by. ’v-y r. VW' v w w—y— Few .. 1 v vr—vn ‘-
4

' Means underscored by the same line are not significantly

different (17).

56

significantly lower for inside color than all the other cus-
tards. The Judges indicated the inside colors of the cus-
tards with the light color (A, B. and C) were too pale, too
light, and too yellow for a peach color in baked custards.

It was observed the inside color was consistently light»
or than the crust color for all variables. Custards B and C
were scored significantly lower than all other custards for
both crust color and inside color at the l per cent level of
probability.“ Figure 6, page 49, illustrates inside colors
for the three different peach colors used in the study.

' Issac

Table 8 shows the mean scores and the standard devia-
tions for aroma. The analysis of variance showed no sign
nificant difference for aroma among the nine treatment vari-
ables at the 5 per centlevel of probability. I

Aroma in custards is a factor often difficult to evalun
ate. In this study the aroma seemed to be more difficult to
evaluate if the crust had been removed from the custard for
several minutes. The descriptive terms indicated frequently
on the Judges' score sheets fer all of the treatment vari-
ables were: no peach aroma, too weak, and "eggy". It ap-
pears the addition of the three levels of peach flavor used
in this experiment did not contribute to the peach aroma of

any of the baked custards.

Table 8.

57

Mean Scores and Standard

Deviations for Aroma.

 

 
     

    

F1

0:3

H 32.43

4.142

4.14
4.43
3.71
4.14
4.14

" 4029

3.57
3986

 

   

.- 42%2_.§42+

 

'4.00
4.86
4.29
4.29
4.29
4.29
4.14
4.71
4.57

   

Y

4943
4029

yw—‘q

4.14’

4.14
5.86
4.29

4.86.

4.29

4.29.

4.29
4.29
4.29
4.45

‘4.55 ’
4.29'

5.17
4.71

3.67

4.04:0.25
4.31:0.39
4.2020.26
4.10:0.26
4.5440.55
4.2940.16
4.19:0.46
4.52:0.56
4.4340.32-

 

I

W 'x y—Y I urwfi 1.....—

w

Iwrfi' 1"

"T

v... .

M v

rr— '7 Y'r'

Refer to code on page 44 for variable identification.

2 Mean score of 7 Judges’ evaluations of l replication of

each variable.

58

inside Flavor

The mean scores and standard deviations for inside flag
vor are given in Table 9. No significant difference at the
5 per cent level of probability was indicated by analysis of
variance of the inside flavor scores. The high standard de-
viation for custard A may have been caused by the low score
for the first replication, the first custard the taste panel
evaluated in this experiment.

Custard I received the highest mean score for inside
flavor, however it was not significantly higher than the
other custards. Custard 0 (light color, high flavor level).
received the lowest mean score for inside flavor, indicating
the Judgments for inside flavor were influenced by impres-
sions of the color. 0n the average, the group of custards
with the dark color (G, H, and I) received the highest score
for inside flavor. Treated as a group, it seems that the
dark color may have exerted some “halo effect” on the inside
flavor.

When comparing the inside flavor scores with the crust
flavor scores (which were significantly different at the 5
per cent level of probability). it should be pointed out
some of the panel members indicated the flavor appeared to
be concentrated in the crust of the custards.

The taste panel members indicated the inside flavor of
all the custards was too weak, "eggy", and artificial. All

of the custards contained the same amount of sugar but

59

Mean Scores and Standard Deviations for Inside
Flavor.

Table 9.

 

 
 

 

     

 

Vari~
‘ , . A ... -1111 1,.“1111 ,. -,-.-1;M393h—£Ldlt
A1 3.142 4 17 4.33 ’ 4.29 4.57 4.33 4.14:0.51
B 4.29 4.14 4.29 4.00 4.43 4.14 4.22:0.15
c 4.14 4.14 4.00 4.29 3.86 4.14 4.10:0.15
D 3.86 3.86 4.33 '4.29 . 4.71 A 4.86 4.3210.42
E 3.71 4.50 5.00 4.57 4.71 4.57 4.51:0.43
F‘ 3.86 4.86 4.43 4.86 4.57 4.67 4.54:0.37
c 4.29 3.86 4.71 4.86 4.71 4.71 ' 4.52:0.38
H 4.71 4.29 4.71 4.57 4.14 5.17 4.51:0.39
I 4.57 4.33 5.00 4.71 5.14 4.43 4.70:0.32

—, WW Yv‘v‘rvv-v- 'W rvwv—yrv PWW‘W—T ww
-

1 Refer to code on page 44-for variable identification.

T—vv

W ‘—‘YV'

—- Trvrvv PVT" wvwwv—T V—r v—v'r' 1 w—Wv-gvr-qr w ‘~—‘—‘— 1'

W

2 Mean score of 7 Judges' evaluation of l replication of

each variable.

——...,—.

6O

Judgments on the sweetness of the interior custard varied
considerably among Judges. For all nine variable treatments
one panel member indicated the inside flavor was too sweet
while another member indicated it was not sweet enough. The
panel member who indicated the custards were too sweet had a
low threshold for sucrose (0.015M) and the panel member who
indicated the custards were not sweet enough had a high.
threshold for sucrose (0.035M) as determined by taste threSh-
.old studies conducted at the beginning of this experiment.

W

,Table 10 shows the mean scores and standard deviations
for consistency. Analysis of variance of the scores for
consistency indicated there was no.eignificant difference
at the 5 per cent level of probability. Therefore, con-

sistency showed no effect due to treatment.

Texture

The mean scores and standard deviations for texture are
listed in Table 11. There was no significant difference at
the 5 per cent level of probability as indicated by analysis
of variance.

The high standard deviations for custards C, D, and G
may have been caused by the low scores in replications 2, 3,
and 4 where the panel members indicated the interior texture
was not smooth and the custards contained several holes.

These custards were baked to the same end temperature as all

61

Table 10. nean Scores and Standard Deviations for Consist?

ency.

 

yu-r-‘rv “VT—fl. -—v—1wv firr—wv Tue—'— ruv-«p‘r

 
   
      

 

 

Grand
1. .11.-iw11.1- ..- - 4......Esae__§a21
41 5.712 5.83 5.33 5.57 5.43 5.33 5.53:0.21
B 5.29 4,86 5.57 5.14 5.57 5.71 5.36t0.32
c 5.86 5.57 4.33 5.29 5.71 5.43 5.37:0.55
D 5.00 5.29 4.83 5.29 4.86 5.86 5.1910.39
E 5.00 5.50 5.67 5.14 5.43 5.57 5.39:0.26
F 5.57 5.86 6.00 5.29 5.29 5.50 5.59:0.29
e 5.43 5.14 5.57 5.43 5.43 5.43 5.41:0.14
H 4.71 5.29 5.00 5.14 5.29 5.33 5.13:0.24
1 5.14 5.67 5.83 5.71* 5.86 5.86 5.6810.28

1
2

Refer to code on page 44 for variable identification.

Mean score of 7 Judges’ evaluations of l replication of
each variable. ’

62

 

 

 
 

 

 

    

 

Table 11. Mean Scores and Standard Deviations for Texture.
Variw Grand
82}? -1’ 7 -11. ,1-11111. -... -1. Mass.s§s21
4' 4.002 5.50 4.50 5.00 4.57 3.67 4.54:0.66
s 5.43 4.86 6.00 4.00 4.57 5.14- 5.17:0.77
0 6.14 3.86 3.33 4.00 5.86 4.71 4.66:1.14
D 6.00 3.57 3.83 4.14 4.43 5.57 4.59:0.98
E 4.43 4.67 4.67 4.00 5.86 4.29 4.6530.64
F 5.57 6.00 5.43 4.86 5.43 4.00 5.05:0.75
a 5.29 6.14 5,71 3.57 5.86 5.43 5.33:0.92
H 5.43 4.14 3.67 4.71 4,57 4.00 '4.42:o.62
I 4.14 4.17 4.33 4.29 4.71 5.14 4.46:0.39

 

rrvr—y— 1v

r.~—vvvvw~,*1erT'—1r

‘ r w—W—W—rl

w—rwq -vvv

Y .— 1 ,

, v.—

fi—vv

'f-rvv

wv—r

7.“ II1Y'WT

1 Refer to code on page 44 for variable identification.

2 Mean score of 7 Judges' evaluations of 1 replication of
each variable.

63

other custards and the time required to reach 86°C was with-
in the range of times for all other custards. Examination
of the baking records does not reveal the reason for these

high standard deviations.

esi*

Table l2 lists the means scores and standard deviations
for the syneresis scores. Analysis of variance indicated no
significant difference at the 5 per cent level of probabili-
ty. The high standard deviation for custard C may have been
caused by the low score for replication 3 and the relatively
high scores for replications l and 5. The texture score for
these replications of custard G showed similar variation.
on scoring texture, the taste panel indicated this custard
contained some holes and therefore a greater amount of syn~

eresis would be expected.

Objective Tests

genetratigg

The depth to which the penetrometer cone penetrated the
baked custards was used as a measurement of the firmness of
the custards. The mean penetrometer values with the crust
on and with the crust off and the differences in the mean
penetrometer values (crust off minus crust on) are included
in Tables 1), l4, and 15 respectively.

Analyses of variance of the penetrometer values (from

which the values of the control were excluded) indicated

64

Table 12. Mean Scores and Standard Deviations for Syneresis.

 

 

 

 

 

 

A‘ 5.002 6.00 4.83 5.86 4.71 5.00 5.23:0.55
B 5.86 6.14 5.86 4.71 5.14 5.43 5.5230.53
C 5.86 4.57 4.00 4.71 5.29 5-57 5.1710.88
D 5.71 5.00 4.83 5.14 5.14 6.14 5.3310950
a 4.86 5.67 -‘5.33 4.29 6.14 5.29 5.26:0.64
F 5.29 6.00 5.71 5.71 5.00 5.17 5.4810.38
c 5.71 6.14 5157 4.57 6.29. ‘5.43 5.62:0.61
H 5.43 4.71 4.50 5.71 4.71 5.00 5.0120.47
I 4.57 5.33 5.17 5.43 5.29 5.86 5.2830.42
k-

 

 

1
2

v 7*?—

fir

IVY u-vv 7" WWW

wwwj

nw—r‘

vr y—v-vvvj

Refer to code on page 44 for variable identification.

Mean score of 7 Judges' evaluations of l replication of
each variable. _

65

 

   
 

 

Table 13. Mean Penetrometer Values and Standard Deviations
on Baked Custards with the crust Uni (Values in
millimeters). '

Variw -' Grand W-th
A' 28.02 27.6 29.2 28.0 28.7 28.6 28.420.59
8 27.6 27.3 26.7 27.2 27.7 28.1 27.430.48
c 27.9 28.1 28.9 26.7 ‘27.} 26.5 27,620.91
D 27.1 29.4 28.5 29.6 28.6 27.0 28.421.11
a 28.0 29.0 28.6 27.2 26.9 26.5 27.720.99
F 26.9 '27.8 27.4 28.0 26.9 27.7 27.520.47
c 27.6 27.1 27.3 27.0 26.4 28.1 27.340.58
H '26.8 28.4 27.0 27.1 28.5 28.4 272710.81
I 26.9 28.3 28.4 27.8 27.9 27.7 27.820.54

CON- 26.3 26.8 ' 27.5 27.1 27.330.70

TROL

28.0

'28.1

 

MW , V

‘5 v-v1rry

upwv-V-rrrv

V'vv‘ ,vv‘w‘

w.

‘v 'Yvr' 1 ""1"

1 Refer to code on page 44 for variable identification.

2 Mean score of 2 evaluations of l replication of each

variable.

66

Mean Penetrometer Values and Standard Deviations
on Baked Custards with the Crust off. (Values in
millimeters).

Table 14.

 

'W‘TT'“ V Y . fir ~u-V v'vv y—vv—wyw —v———v v wr—vvvw—rr .-

Grand

  

 

, _ w. - .. -2..1. efissaw.§424

A1 29.82 29.1 29.9 30.1 31.0 30.9 3o.1t0.72
s 29.2 30.6 29.8 30.3 30.9 30.2 * 30.220.60
C 31.5 ‘31.2 33.5 30.2 30.4 29.9 31.111.32
D 29.1 32.3 31.1 31.1 .31.6 30.1 30.911.13
E 30.1 ~31.0 32.3 30.5 30.1 30.5 30.8:0.83
F 30.3 30.7 31.1 31.1 29.8 29.7 30.5i0.62
c 31.7, 28.8 31.2 29.9 29.1 30.4 30.2ii.14
H 31.0 31.1 30.0 30.0 29.6 29.6 30.220.67
I 31.3 30.2 31.8 28.2 30 3 30.4 30.5:1.25
gfigi '30.0 30.4 30.7 29.7 32,4 30.6 30.620.94

'xru-r

I'v'ri'

vwrw —‘-—~ .yv-vw—uwalw
.

:v—v v
I

- v y wwj

‘vwvrv'T-Y

1 Refer to code on page 44 for variable identification.

2 Mean score of 2 evaluations of l replication of each

variable.

‘V

67

Table 15. Mean Penetrometer Difference Values and Standard
Deviations. (Values in millimeters).

 

Grand

. , . . _ . .1 M- . -Mesaa.§+2l
41 1.82 1.5 0.7 2.1 2.3 2.3 1.78:0.61
B 1.6 3.3 3.1 3.1 3.2 2.1 2.73to.71

C 306 30]- 496 305 301 304 3955:0955

 
 

 

 

0 2.0 2.9 2.6 1.5 3.0 3.1 2.52:0.64
2 2.1 2.0 3.7 3.3 3.2 4.0 3.05:0.83
r 3.4 2.9 3.7 3.1 2.9 2.0 3.0010.58
c 4.1 1.7 3.9. 2.9 2.7 2.3 2.93:0.92
a 4.2 2.7 3.0 2.9 1.1 1.2 2.52:1.18

I 494 199 304 004 299 207 2c6a1936
CONH
TROL 3.7 3.6 3.2 1.7 5.3 2.5 L 3.33:1.22

Y - WV—WV'Y“ yww—qwmvrww as "vvv‘r wwwvw .w .m —. firr—rwvv—v n w— v V

1 Refer to code on page 44 for variable identification.

2 Mean score of 2 evaluations of l replication of each
variable.

68

there were no significant differences in firmness with the
crust on. the crust off, or for the difference in the mean
penetrometer values. There is no real explanation for the
high standard deviations for several of the treatment vari-
ables since each value is an average of two readings for one
replication (one inside bake and one outside bake). There
was also the same number of front and back bakes for each
treatment. When comparing the mean penetrometer values
(crust on and crust off) for the nine treatment variables
with those of the control custard, it does not appear the
addition of the small amounts of food colors and peach fla-
vors altered the firmness of the gel structure to any extent.
Differences between mean penetrometer values for sam-
ples with and without crust were recorded because it was
thought this comparison might provide a possible measurement
of resistance due to the crust for each of the treatment
variables. No consistent trend appears with respect to the

three different colors or the three levels of flavor.

W

The mean per cent sag and standard deviations are listed
in Table 16. Analysis of variance, excluding the control
custard. showed no significant difference among the nine
treatment variables.

The very high standard deviation for the control cus-

tard appears to be caused by the extremely high value for
the sixth replication. It is felt that difficulty in

Table 16.

69

Per Cent Sag and Standard Deviations of Baked
Custards.

 

I
GONw
TROL

 

 

 

4.58
6.99
7.52
9.52
4.92
4.80
7.34
4.92
7.34
7.16

7.91
7.52

10.52

4.92
4.80
7.16
7.71

,w r‘v‘

4. 92 15. 42

7.34 7.71
5.04 7.16
7.52 5.17
7.71 5.04
7.71 7.52
4.92 7.34
4.92 5.69
7.52 7.52
7.16 6.99

4.80
9.30
6.99
14.68
9.09
7.34
6-99
10.00
”4.80
20.00

7.27:4.15
7.48:0.97
7.22:1.26
9.05:3.19
7.23:1.86
7.4Eil.85
6.51:1.24
6.39t2.02
6.9221.05
8.97:5.50

 

,wer r"

r-r

.r—v~—r7 v-Vvyv—‘wvvy—

V-W" 17W Y'T'T j—rerV— v vrv

fir...

wwv .

V‘v‘r‘ y'v w—vr

1 Refer to code on page 44 for variable identification.

7O

removing the custard from the cup caused this high value;
this was evident with the sixth replication of custard D and
the fifth replication of custard A.

The high standard deviations for all treatment vari-
ables and the difficulties encountered in removing the cus-
tard from the cup may indicate the method of performing the
test for per cent sag used in this experiment is not suit-
able for use on conventional baked custards. The per cent
sag test as conducted in this experiment has been used with
satisfactory results in other types of gels (64). The per
cent sag. as measured by a different technique, has been
used successfully on baked custards by Mastic (44) and

Miller at al. (49).

§ yne;:esis

The mean per cent drainage values and the standard de-
viations are listed in Table 17. Analysis of variance of
per cent drainage, excluding the control custard, revealed
no significant difference.

The high standard deviations for custards D, E, and F
were caused by the very high values in the second and third
replications. These high values appeared to be connected
with the difficulty involved in removing the custard from
the cup. The custard would crack causing a greater amount
of drainage. It must be noted that all but two of the stand-

ard deviations are higher than the corresponding means.

These data do not represent true measurements of syneresis

7l

Table 17. Per Cent Drainage and Standard Deviations of
Baked Custards.

 

 
 

0.00

WWI W‘rv‘y‘v-f 7’!“

‘V

 

rrv, (fir—7w w - VT—
1

Grand

. liegn Sch

 

0.90

0.00
0.84
3.85
2.30
3-83
0.09
0.00
0.09
0.00

1.20
0.44
1.64
2.74
2.26
1.51
0.17
0.17
0.00
1.46

, hr.

0934

1.60
0.00
0.00
0.00
0.41
0.00
1.02
0.00
0.56

iv fiv

0.57:0.52
0.43:0.61
0.71:0.85
1.52:1.55
1.14:1.04
1.04:1.48
0.2020.37
0.2610.39
0.1410.30
0.35:0.58

v —v Y

1 Refer to code on page 44 for variable identification.

72
and no definite conclusions should be based on them.

H a n s

The range of pH readings over six replications for the
custard mix and for the baked custards are shown in Table 18.
The ranges were so close no analysis of variance was carried
out on the pH readings. In all cases the baked custards
were more alkaline than the custard mix. This result is in
accord with the findings reported by Logue (35). Lows (38),
and Miller et al. (49).

By comparison with the control custard, it appears that
the addition of the small amounts of colors and flavors did
not alter the acidity or alkalinity of the custard mixes or

the baked custards under the conditions of this experiment.
Correlation between Selected Measurements

Correlation coefficients were calculated between selec-
ted objective measurements, between eelected objective and
subjective measurements and between selected subjective
measurements. The results are shown in Table 19.

The penetrometer values (crust off) and the per cent
sag values were compared to determine the relationship be-
tween these two obJective measurements for firmness. There
was no significant correlation between these two objective
tests indicating that one or both of these tests may be un-
reliable as an objective method of determining firmness of

baked custards.

73

 

Table 18. Range of pH Readings over Six Replications.
Vari~ Range of pH for Range of pH for
able _ i CREEEPd,ME§. w. ‘Baked Custard
A 6.90 n 7.00 7.01 - 7.10

B 6.92 - 7.00 7.01 — 7.10

C 6.90 - 6.99 7.05 - 7.10

D 6.90 - 6.99 7.05 - 7.10

E 6.90 - 7.00 7.02 . 7.10

F 6.90 - 7.00 7.02 - 7.10

G 6.90 - 7.00 7.01 - 7.10

H 6.90 e 7.00 7.01 - 7.10

i 6.90 ~ 7.00 7.01 - 7.10 .
CONTROL 6.90 - 7.00 7.02 - 7.10

 

 

1

W'Y

v f

w—r

Refer to code on page AA-for variable identification.

74

Table 19.

Correlation Coefficients of Selected Measurements.

 

 

*v-v v VY—v—r w , yvv— ‘1 7 www ‘7 .. wV—p-

I. Between Objective Measurements

Penetrometer Values (Crust Off) vs.Per Cent

Sag (Inverted, Crust Cn) +0.0768

II. Between Objective and Subjective Measurements

Penetrometer Values (Crust Off) vs.
Consistency Scores

Per Cent Sag (Inverted, Crust 0n) vs.
Consistency Scores

Per Cent Drainage (Inverted, Crust On) vs.
Syneresis Scores

III. Between Subjective Measurements

Crust Color vs. Crust Flavor

Inside Color vs. Inside Flavor

Crust Color vs. Inside Color

Crust Flavor vs. Inside Flavor

Inside Flavor vs. Aroma

Texture vs. Syneresis_

Consistency vs. Texture

fiv’v‘ v—T W W 7v rrw—r-fi juvvv-r #— r v *V‘wv'wv— f fi" r I—vv-

** Significant at 1% level of probability.
* Significant at 5% level of probability.

~0.1795

+0.0857

"O Q 2210

+O.SOO4**
+0.4334**
+0.8082**
+0.5649**
+0.3592*

+0.8559**
+0.3070*

wwv—vv— 'V‘w

75

Correlation coefficients were calculated on the taste
panel scores for consistency with both the penetrometer
values (crust off) and the per cent sag values to determine
the reliability of these latter two tests. The results in-
dicate there were no significant correlations between either
of these two objective measurements and the subjective evalu-
ation for firmness. Panel scores for syneresis were not cor-
related with per cent drainage values, indicating this ob-
jective measurement was not reliable. The results for firm—
ness obtained with the penetrometer in this experiment are
in accord with those of Bittner (3) and MacDougall (40),
both of whom indicated that the curd tension meter is a
better objective measure of gel strength than the pens.
trometer.

Table 19, part III shows comparisons between several of
the factors scored by the taste panel. There were highly
significant positive correlations between the following
pairs of judges' scores: crust color and crust flavor, in-
side color and inside flavor, crust color and inside color,
crust flavor and inside flavor. A "halo effect" of color
impression on flavor judgment may be operating part of the
time. The inside flavor and aroma scores were positively
correlated at the 5% level of probability. The highly sig-
nificant positive correlation existing between texture and
syneresis scores indicates the texture scores increased as

the syneresis scores increased, inferring that with a better

76

texture (fewer holes) there was a smaller amount of drainage
in the baked custards. The judges' scores for consistency

and texture gave a positive correlation significant at the

5% level of probability.

SUMMARY AND CONCLUSIONS

The purpose of this investigation was to study the ef-
fect of the addition of the nine possible combinations of
three different peach colors and three levels of peach fla-
vor on the palatability and gel structure of standard baked
custard.

The basic eXperimental formula consisted of constant
proportions of reconstituted dried whole milk, fresh whole
eSB. and sucrose. Each of the nine treatment variables con-
.tained one of three peach colors a designated light, medium.
and dark color - and one of three levels of peach flavor -
designated low, medium, and high flavor level. A control
custard having neither added color nor added flavor was used
as'a reference for the objective tests.

Six replications of each of the nine treatment variables
were prepared and evaluated by subjective and objective meth-
ods. Preparation of the custard mix, additionof the celors
and flavors, and baking procedure were standardized as much
as possible. All samples were cooled to room temperature
before objective and subjective evaluations.

Room temperature and relative humidity of the labora-
tory during preparation and testing were recorded. Time-
temperature relationships during baking were recorded on a
Brown Electronik Potentiometer High Speed Multiple Point

Recorder.
77

78

The palatability of the baked custards was evaluated
subjectively by a taste panel of 7 persons (4 women and 3
men) on nine characteristics: crust color, crust tenderness,
crust flavor, inside color, aroma, inside flavor, consist-
ency, texture, and syneresis. In rating the samples, a 7.
point scale was used in which 7 was the highest possible
score. Objective measurements included pH of the custard
before and after baking, gel strength as indicated by the
penetrometer (crust on, crust off) and per cent sag, and
syneresis as indicated by per cent drainage.

Analysis of variance on the subjective scores for the
nine treatment variables indicated there were no significant
differences in six of the characteristics: crust tenderness,
aroma, inside flavor, consistency, texture, and syneresis.
Therefore it may be concluded that for these characteristics
any of the nine variable treatments would be equally accept-
able to the taste panel used in this experiment.

Significant differences among the nine treatment vari—
ables were found in the subjective scores for crust color
(1% level of probability), inside color (1% level of probes
bility), and crust flavor (5% level of probability). For
crust color and inside color the light color custards as a
group scored the lowest, and the medium color custards as a
group scored the highest with the exception of custard G
which received the highest mean score for inside color.

This high inside color score for custard G indicates some

79

of the taste panel members did not object to the dark inside
color of the custard but they did object to the dark crust
color. This reinforces the observation that the added color
concentrates in the crust, making the crust darker in color
than the inside portion of the custard.

For crust flavor, custard F (medium color, high flavor)
scored the highest, followed by custards G and D, both con-
taining the low flavor level and dark and medium colors, rev
spectively. This may indicate influence of the "halo effect"
of crust color impression on the judges' scores for crust
flavor. Since the flavor seems to concentrate in the crust,
it would not be unreasonable to find the low level of flavor
:‘rating high. In considering the three subjective character-
istics which were significantly different, it appears that
custard F (medium color, high flavor) was scored highest by
this taste panel.

For inside flavor, custard I (dark color, high flavor)
scored the highest, although the variable treatments were
not significantly different. The medium color and high fla~
vor level custard (custard F) scored second highest and cus-
tard G (dark color, low flavor) scored third. It should be
noted the custard with high flavor level but low color (cus-
tard C) scored the lowest. As a group, the custards con-
taining the dark color received the highest score for inside
flavor. The evidence again seems to indicate the ”halo ef-

fect'of color impression on flavor judgment.

80

it is evident the inside flavor scores were consistently
lower than the crust flavor scores, reinforcing the observa~
tion that the flavor, as well as color, concentrated in the
crust of the custard.

The results of the color and flavor scores indicate
there were more differences in the color characteristics
than in the flavor characteristics of the nine treatment
variables.

Analyses of variance of the objective measurements ins
dicated no significant difference in the nine variable treat-
ments for gel strength and per cent drainage. High standard
deviations for all of the objective measurements raise the
question of the reliability and suitability of these objec-
tive evaluation methods for baked custards.

There was no significant correlation between penetrome-
ter values (crust off) and per cent sag values, and no sig-
(nificant correlations between subjective and objective mea-
surements for gel strength or for syneresis. These results
may indicate the unreliability of the objective methods used.

Highly significant positive correlations were found be-
tween several selected subjective measurements. The highly
significant positive correlations between crust color and
crust flavor and inside color and inside flavor indicate the
"halo effect" of color impression on flavor judgment.

Highly significant positive correlations were also found

81

for crust color versus inside color, crust flavor versus in,
side flavor, and texture versus syneresis. Positive correla-
tlons at the 5% level of probability were found for inside

flavor versus aroma and consistency versus texture.

10.

ll.

12.

13.

LITERATURE CITED

Baten, W.D. Reaction of age groups to organoleptic
tests. Food Tech. fig277, 1950.

Birren, F. Color and human appetite. Food Tech. 11:
553. 1903-

Bittner, B.A. The effect of various milks on the timer
temperature relationships in baking custard, and on
the quality of the product. Unpublished M. S. thesis.
Michigan State College. 1950. ‘

Brice, B.A. Color in Foods. Symposium. Nat'l. Acad.
Sci.sNat'l Res. Council. 1954.

Buck, R.E. and Sparks, R.A. Relation of ketchup color
to tomato color as determined by the Hunter instru-
ment. Food Tech.Ԥ:122, 1952.

Carr, R.E, and Trout, G.M. Some cooking qualities of
homogenized milk. I. Baked and soft custard. Food
Research,1:360, 1942.

Gaul, J.F. The profile method of flavor analysis.
Advances in Food Research. Vol. 7. Edited Mrak,
E.M. and Stewart, G.F., New York, Academic Press

Inc., 1957,

Chick. H. and Martin, C.J. On the "heat coagulation"
of proteins. Part II. The action of hot water upon
egg albumen and the influence of acids and salts upon
.reaction velocity. J. Physiol. flzcl, 1911.

Chick, H. and Martin, C.J. On the "heat coagulation”
of proteins. Part III. The influence of alkali upon
the reaction velocity. J. Physiol. 3:;61, 1912.

Code of Federal Regulations. Title 21 - Foods and Drugs.
'Parts 8 and 9. Washington, D.C. Office of the Federal
Register, National Archives and Records Service, U.S.

Gov't. Printing Office, 1965.

Cook, H.L. and Husseman. D.L. Consumer acceptance of
dry milks in quantity cookery. Wisc. Agr. Expt. Sta.
Research Bulletin 164, 1950.

Cracker, E.C. Flavor. New York, McGraw-Hill Book
Company, Inc., 1945.

Crocker, E.C. Measuring food flavors. Food Research
g;273. 1937.

82

14.

15.

16.

17.

180

19.

20.

21.

22.

23.

24.

27.

28.

(D

,3,

Cracker, E.C. and Platt, w. Food flavors - a critical
review of recent literature. Food Research,23183,
1937. .

Dawson, E.H., Brogdon, J.L., and McManus, S. Sensory
testing of differences in taste.. II. Selection of
panel members. Food Tech. 11:1251. 1963.

Di Geneva, J. Information on flavors (Private communi-
cation), New York, Givaudan Flavors, Inc.. 1964.

Duncan, D.B. Multiple range and multiple F tests.
Biometrics 11:1, 1955.

Esau, P. Procedures for conversion of color data from
one system into another. Food Tech.,l23167, 1958.

Fruton, J.S. and Simmonds, 8. General biochemistry.
2nd ed. New York, John Wiley and Sons, Inc., 1958.

Griswold, R. The experimental study of foods. Boston,
Houghton Mifflin Company, 1962.,

Holman, V. and Carver. J.S. The yolk color index.
U.S. Egg and Pltry. Mag.,fil(no. 8):40, 1935.

Huggart, H.L. and Wenzel, F.W. Color differences of
citrus juices and concentrates using the Hunter Color
Difference Meter.. Food Tech. 2c27. 1955.

Ingle, J.D. Information on flavors (Private communi-

cations). Chicago, Illinois, Food Materials Corporaw
tion. 1963, 196 0

Jablonski, C.F. Coloring matters in foods. The Chem-
istry and Technology of Food and Food Products. Vol.
I. 2nd ed. Edited Jacobs, M.B., New York, Inter-
scisnce Publishers, Inc., 1951,

Janovsky, H.L. Flavomatics in food. Food Tech.,2:

Jordan, R., Wagner, E.S., and Hollander, H.A. Non~
homogenized vs. homogenized milk in baked custards.
J. Am. Dietet. A. 2931126, 1954.

Katz, A. Highlights in newly developed flavoring
aromatics. Food Tech.,2:636, 1955.

Kefford, J.F. Objective measurement of colour. C.S.
I.R.O. Food Pres. Quarterly 22(no. l):8, 1963.

29.

30.

31.

32.

33.

34.

37.

38.

39.

40.

41.

42.

32+

Kessler. H.P. Information on flavors (Private communi-
cation). New York, Givaudan Flavors Inc.. 1963.

Kramer, A., Murphy, E.F., Briant, A.M., Wang, M., and
Kirkpatrick, M.E. Studies in taste panel methodology.
J. Agr. Food Chem. 2;224, 1961.

Krum. J.K. Truest evaluations in sensory panel testing.
Food Eng.,g1:74, 1955.

Laird, D.A. and Breen, W.J. Sex and age alterations in
taste preferences. J. Am. Dietet. A. 1§;549, 1939.

Langwill, K. Taste perceptions and taste preferences
of the consumer. Food Tech. fijl36, 1949.

Leonard, S.J., Luh, E.S., Chiohester, 0.0., and Simone,
M. Relationship of fresh clingstone peach color to
coéor and grade after canning. Food Tech.,1fi:492.
l9 1.

Logue, L.E. Some qualities of eggs affecting the gel
strength of custards. Unpublished M.S. thesis. Iowa
State College. 1940.

Longree, K. Color determinations for baked custards
(Private communication). Institutional Management
Dept. Cornell University. 1963.

Longree, K., Jooste, M., and White, J. Time—temperature
relationships of custards made with whole egg solids.
III. Baked in large batches. J. Am. Dietet. A. 28;
147. 1961.

Lowe, B. Experimental cookery. 4th ed. New York,
John Wiley and Sons, Inc., 1955.

Lowe, F. and Stewart, C.F. Subjective and objective
tests as food research tools, Food Tech.,1:30, 1947.

MacDougall, M.J, Cooking qualities of several concen-
trations of various types of nonfat dried milk solids.
Unpublished M.S. thesis. Michigan State College.

1953.

Mackey, A.C. and Valassi. K. The discernment of primary
tastes in the presence of different food textures.

Food Tech» 19t238, 1956.

MacKinney, G. and Little, A. Color of foods. West-
port, Conn., AVI Publishing Co., Inc., 1962.

430

44.

46.

47.

48.

49,

50.

51.

52.

53.

54.

55.

56.

85

Mason, D.D. and Koch, E.J. Some problems in the design
and statistical analysis of taste tests. Biometrics

Mastic, M.E. The effect of homogenization on the gela-
tion and palatability of baked custards prepared with
dried whole egg solids. Unpublished M.S. thesis.
Michigan State University. 1959.

McBride, B.M. Practicality of using dried whole a g in
custards for institutional food service. Unpubl shed
M.S. thesis. Iowa State College. 1947. ‘

VMerory, J. Food flavorings - composition, manufacture,

in% use. Westport, Conn., AVI Publishing Co.. Inc.,
9 O.

Merwin, E.J. The peach v another stone fruit. The
Givaudan Flavorist. No. 4, 1965.

Meyer, L. Food chemistry. New York, Reinhold Pub-
lishing Corp., 1960. . '-

Miller, G.A., Jones, B.M., and Aldrich, P.J. A compari»
son of the gelation preperties and palatability of
shell eggs, frozen whole eggs, and whole egg solids in
standard baked custard. Food Research gg;584, 1959.

Negler, L. Acceptabllity of baked custards made with
increased amounts of dry milk solids. Special problem
FM 405. Foods and Nutrition Dept. Michigan State
University. 1957.

Mason, E.H. Introduction to experimental cookery. New
York, McGraw-Hill Book Company, Inc., 1939.

Nickerson, D. Color measurement and its application to
the grading of agricultural products. USDA Misc.
Publ. 580. March, 1946.

Pangborn, R.M. Taste interrelationships. Food Research
g§;245, 1960.

Pettit, L.A. Informational bias in flavor preference
testing. Food Tech” 12:12, 1958.

Power, F.B. and Chestnut, V.K. The odorous constitu-
ents of peaches. J. Am. Chem. Soc. 42;l725, 1921.

Robinson, W.B., Wishnetsky, T., Ransford, J.R., Clark,
W.L. and Hand, D.B, A study of methods for the
measurement of tomato Juice color. Food Tech. g:

64.

86

Schutz, H.G. Color in Foods. Symposium. Nat'l.
Acad. Sci.-Nat'1. Res. Council. 1954.

Smith, B.H. Modern trends in flavors. Food.Research

Valdes, R.M., S mone, M. J. and Hinreiner, E. H. Effect
of sucrose an organic acids on apparent flavor inv
tensity. II. Fruit nectars. Food Tech.,1gz387, 1956.

Warner-Jenkinson Manufacturing Co. Certified Food
Colors. St. Louis, Mo. 1961.

Weiser, H. B. Colloid chemistry. 2nd ed. New York,
John Wiley and Sons, Inc., 1949,

Wilson, D.E., Meyer, J. C., Robinson, W.B., and Hand,
D.B. Objective evaluation of color and consistency
in peach puree, Food Tech. ,_1:479. 1957.

Yeatmen, JoNo. Sidwell, AOPQ and Norris, KeHo Dfirlvaw
tion of a new formula for computing raw tomato juice
color from objective color measurement. Food Tech.
14:16,1960,

Zabik, M.E. The effect of sucrose and calcium cyolamate
upon the gel strength of selected gelling agents in
the preparation of jellied custard, Unpublished M. S.
thesis, Michigan State University. 1961.

APPENDIX

88

Table 20. Colors and Flavors Used in Baked Custards for
Objective Tests.

 

"' V mfir'rva-v'w w v— v fi-‘v W‘l‘r'r‘v—yi v Wyr—rfiy rv w—y -w~y7~*vw ‘- rw ,— W T—v _, ,-...—_..—7_.~» v

(Quantity for 1100 c. c. custard mix)

Color Color Color
Solution Solution Solution Flavgr Flavor Flavgr
2 ~ 3

 

 

Variable 2
W "T“m ml m TW“"'"'“5m ' m ' m
A‘ 1.00 _- - 0.55 0.28 0.11
B 1.00 - - 1.10 0.55 0.22
C. 1.00 — - 1.65 0.83 0.33
D - 1.00 . - 0.55 0.28 0.11
E - 1.00 - 1.10 0.55 0.22
F - 1.00 - 1.65 0.83 0.33
0 ~ - 1,00 0.55 0.28 0.11
H - ‘. - 1.00 1.10 0.55 0.22
I - . - 1.00 1.65 0.83 0.33

 

*7 '7 rv r‘v— *- 7 WV w—w; r r “VT 7, v~ F—vw .v—V v V'V'i v .1

1 See page 91 for composition of color solutions.

2 Imitation Peach Extract No. 7442, Food Materials Corpora-.
tion, Chicago 18, Illinois.

3 'Imitation Peach Flavor with True Fruit F-4710, Givaudan
Flavors Inc., New York 56, New York.

4 Flex-Sol Imitation Peach Concentrate F-4711, leaudan
Flavors Inc., New York 36, New York.

Table 21.

89

Colors and Flavors Used in Baked Custards for
Subjective Tests.

 

*‘r‘ rw w.‘ vv-v-uvw'r va .V—Vfi—v- 7' 7r Yr v .—v v. . r—w .17 .1 1*...— TV Frw ,71 r, .—

(Quantity for 1500 c. c. custard mix)

Color Color Color

Vari b1

_c: c: rm >-

#4 :2 Q a: DJ

Solution Solution Solution
'2

m

1.36
1.36
1.36

m

p

1036
1.36

1.56

Flaygr Flavgr Flavor

 

“w

m

1.36
1.56
1.36

WV

m
0.75
1.50
2.24
0.75
1.50
2.24
0.75
1.50
2.24

m,
0.38
0.75
1.13
0.38
0.75
1.13
0.38
0.75
1.13

m
0.15
0.30
0.45
0.15
0.30
0.45
0.15
0.30
0.45

 

V'Vv . v—vv

fl.

'vv WWW "rvv—vwvv

vfi—v v fl

r— 7"-

1 See page 91 for composition of color solutions.

my —

2 Imitation Peach Extract N0. 7442, Food Materials Corpora-
tion, Chicago 18, Illinois.

3 Imitation Peach Flavor with True Fruit F-4710, Givaudan

Flavors Inc., New York 56, New York.

Flex-Sol Imitation Peach Concentrate F-47ll, Givaudan
Flavors Inc., New York 56, New York.

Table 22. Description of Flavors Used,

w—rwvw

W —v www,' V7 V7. 7 WW

90

WT- v

_x

‘rv w m "a v'v‘v—‘Y v—v‘w 7r v—y v—
z

 

Composition Solvent Solubility pH

Imitation Combination of 95% - Completely 6.0a
Peach Extract natural oils alcohol soluble in 7.5
No. 7442 such as orange (23) alcohol; some (23)
(Food oil and syn. of the mater-
Materials thetic flavorq ials not
Corporation) ing materials soluble in

‘ (23). water (23).
Imitation Blend of alcohol Absolute 4.2
Peach Flavor synthetic and alcohol (16)
with True ingredients water and water
Fruit F-4710 with concen- ' '
(Givaudan trated peach
Flavors Inc.) extract (29).
Flex-Sol An emulsion water Not completely 4.1
Imitation of oil soluw soluble in ale (16)
Peach ble chemicals cohol due to
Concentrate and gum um arabic
(Givaudan arabic (29). 29). Soluble
Flavors Inc.) in water.

rw—r— jT—Ww—u-rww—V‘yl- .r

91

Table 23. Composition of Color Solutions.

 

 

Color FD&C ,* FD&C
Solution‘ Red No. 4 Xeiiot no. 5
5 w" '5" 3”." 5’ 3"" " '3 ’ fifg""’ " "' “53' ‘ ‘" W mgfw "
#1 (Light) . 330 220
$2 (Medium) 490 220
#3 (Dark) 730 660

Y‘rfiw—fi'lfi‘

1 Powdered colors made up to 100 ml. solution with dis-
tilled water. pH of all three color solutions was 7.3.

92

Exhibit 1. Procedure for Preparation of Colored and Flavored
Custard Mixes.

P e ar t f u*d m 1k:

1. Remove one package of dried whole milk (504 grams)
from refrigerator and allow to come to room tempera«
ture.

2. Weigh out 3496 grams tap water into aluminum sauce-
pan and heat to 46°C.

3. Pour heated water into a 12-quart bowl of a Hobart
Mixer1 and add DWM.

4. Mix DWM and water on speed #1 for 30 seconds using
the paddle attachment.

5. With a rubber spatula mix the milk by hand to break
down any lumps.

6. Mix the milk an additional 30 seconds on speed #1.

7. Place the mixing bowl in a cold water bath to cool
the milk rapidly to room temperature.

Preparation39§ basic custard gig:

1. Break 15 fresh, sheil eggs into a 5- -quart bowl of
a Kitchen Aid Mixer

2. Beat eggs for 3 minutes on speed #2 using a whip
attachment and then for 1 minute on speed #4,

3. Weigh out 748 grams of blended egg into a 5-quart
Kitchen Aid mixing bowl.

4. Add 391 grams of sugar to egg and blend egg-sugar
. mixture for 1 minute on speed #4 using the paddle
attachment

5. Pour egg-sugar mixture into a 12-quart Hobart
mixing bowl.

6. Weigh out 3815 grams of the reconstituted whole
milk and add to the egg-sugar mixture.

 

v’v‘ (W‘v vr-v '—-v-~——r—1

1 Hobart model A-QOO mixer, The Hobart Mfg. 00., Troy, Ohio.
2 Kitchen Aid K5-A mixer, The Hobart Mfg. 00., Troy, Ohio.

93

7. Mix the egg-sugar-milk mixture on speed #1 for 5
minutes using the paddle attachment.

8. Strain custard mix through a medium-fine wire-mesh
household strainer to remove any undissolved parti-
else.

9. Divide custard mix into smaller portions in Sequart
stainless steel bowls. Four portions of 100 c. c.
each for objective tests; three portions of 1500
c.c. each for subjective tests.

Addition of 2910; and flavors:

1. Add color solution to custard mix using a 1 m1.
pipette.

2. Add flavors, one at a time, to custard mix using 1
m1. pipettes.

3. Mix the added color and flavors into the custard
mix using a Kitchen Aid K5~A mixer with the paddle
attachment on speed #1 for 2 minutes.

94

Exhibit 2. General Instructions for Peach Custard Taste
Panel Members.

1. Please do not eat, smoke, or chew gum for 1/2 hour prior
to the time of tasting.

2. Do not give any reactions, such as grimace, smile, or
vocal expression as you evaluate the sample.

3. You will receive three custards, one at a time, to judge
each day. All samples will be at room temperature. You
are to evaluate one custard at a time and complete your
scoring on that custard before another one is presented
to you. There will be a separate score sheet for each
custard. Since this is a color and flavor problem you
will be receiving custards of different color and differ-
.ent levels of flavor. Remember to score each sample
independently of the others.

4. Judge the nine factors in the order in which they are
listed on the_score sheet. Place a mark (X), using a
red pencil, in the bIOck which most nearly fits your
evaluation rating of each sample. Be sure to score each
of the factors listed on the score sheet and to mark (X)
the appropriate descriptive term(s) for each factor you

rate 4 92 below.

5. Between each sample eat some unsalted cracker and take
a drink of the water provided. '

6. This sheet will be given to you each day you score to
remind you of the definitions and instructions below.

*************

PALATABILITY FACTORS: DEFINITIONS AND INSTRUCTIONS

Crust Color. Evaluate the crust color of the custard as it
appears n the custard cup. The crust color will always be
more intense than the inside color.

Crust Tenderness. With your spoon, break through the crust
in t e center 0 the custard and evaluate the tenderness of
a piece of the center crust.

Crust Flayg . Evaluate the flavor of a piece of crust taken
'rom he center to the outside of the custard cup.

Inside Color. Spoon some of the custard onto the plate pro-
videdland'evaluate the inside color.

 

95

Aroma. Evaluate aroma on the basis of a sniff obtained from
the custard remaining in the custard cup.

Inside Flavor. Taste some of the custard in the cup (without
crust) and evaluate the flavor.

Consistency. Spoon some of the custard from the cup onto
the p ate, cut through it with the edge of the spoon, and

make an evaluation. The custard should hold its shape when
spooned out but not make a brittle break when out.

exture. Look at the custard in the cup, at the bottom and
around the sides. There should be no holes present. 'Taste
the custard to determine smoothness.

Synegesis. Look at the custard remaining in the cup and on
your p ate. There should be little or no separation of the

liquid from the gel structure.

CHECK THE SCORE SHEET TO MAKE SURE N0 FACTORS
HAVE BEEN OMITTED AND THE APPROPRIATE DESCRIPTIVE
TERM(S) FOR EACH FACTOR SCORED 4 OR BELOW HAVE BEEN MARKED (x)

 

 

 

96

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Exhibit 3. SCORE SHEET FOR PEACH CUSTARD
Judge Code no.
Date
Score Key: 7 - Excellent Instructions: In the appropriate columns,
6 - Very good place a mark (X) for the score which best
5 - Good expresses your evaluation of that factor.
4 — Medium For those factors scored 4 or below, mark
3 - Fair (X) the descriptive terms which best describe
2 - Poor the sample.
1 - Very poor
PALATABILITY 5 SCORE VALUES
FACTORS 7 6 5 4 3 2. 1 DESCRIPTIVE TERMS
_Unappetizing _Too red
CRUST COLOR _Too dark _Too yellow
_Too pale _Other
CRUST
T ENDERN ESS _T ough _Rubbe ry
_Too strong _Bitter
CRUST FLAVOR _Too weak _Artificial
_Other
_Too dark _Too red
INSIDE COLOR _Too pale _Too yellow
_Other
_Too strong _Perfumy
AROMA _Too weak _Other
_Too sweet _Bitter
INSIDE FLAVOR _Not sweet enough—Artificial
_Too strong _Eggy
_Too weak _Other
___Too firm _Rubbery
CONSISTENCY _Not firm enough _Brittle
_Other
_Not smooth _Many holes
TEXTURE _Several holes
_Slight syneresis
SYNERESIS _Pronounced syneresis

 

 

 

 

 

 

 

 

 

97

Exhibit 4. Sample Calculation of Studentized Multiple Range
Test.

Based on data for Crust Color

General formula for standard deviation of the mean:

SE : s3 s2 = variance (error mean square)
n n = number of replications

Sample (see Table 2, page 47):

31 :JC.‘ 927): :: 0.125

Studentized Multiple Range Values (1% level of probability)‘:
LLLLLLLL

3.82 3.99 4.10 4.17 4.24 4.30 4.34 4.37

Shortest Significant RangesQ:

L21 Elli). ii). 1.6.111). 1.3). 1.2).

.48 .50 (.51 .52 .53 .54 .54 .55

Crust Color Mean Scores:

Treatment
variable: C B H I A G E D F

Mean Score: $.68 3.69 4.61 4.81 4.92 5.10 5.22 5.22_§,6O

 

 

-v-V'

Any two means not underscored by the same line are sig-
nificantly different. Any two means underscored by the same
line are not significantly different (17).

 

Y—v

Duncan (1?).

2 Shortest Significant Ranges = Studentized Multiple Range
Values x 8x. If the difference between any two scores
exceeds the shortest significant range value, those
scores are significantly different.

d

98

Conclusions:

Treatment variable F scored significantly higher than
treatment variables A, I, H, B, and C.

Treatment variables D and E scored significantly higher
than treatment variables H, B, and C.

Treatment variables G, A, I, and H scored significantly
.higher than treatment variables B and C.