DEVELOPMENT AND APPLICATION O F A

TEXTURE MEASUREMENT PROCEDURE FOR

TEXTURED VEGETABLE PROTEIN*

W I L L I A M M. B R E E N E and T H O M A S G. B A R K E R * *
University of Minnesota, Dept. of Food Science and Nutrition,
St. Paul, Minnesota 55108, U.S.A.

(Received 21 July 1975; in final form 6 October, 1975)

Abstract. A procedure was developed for separating different rehydrated commercial soy ‘protein
products’ and cooked ground beef: soy mixtures (75 :25) into textural classes. For comparison,
cooked ground beef was included in the study. Samples were ground in a meat grinder (3/8” holes)
and compressed at 5 cm min-’ in an Ottawa Texture Measuring System 10 cm2 cell, with eight wires
grid, mounted in an Instron tester. Maximum force, average maximum force, hardness, cohesiveness
and chewiness were determined from the force-distance curves. Parameters which detected the
greatest differences between textural classes were maximum force and the average maximum force,
followed - and in this order - by chewiness, cohesiveness and hardness.

1. Introduction
1.1, IMPORTANCE OF TEXTURED VEGETABLE PROTEIN
Textured vegetable protein products have assumed some degree of acceptance and
popularity in the past few years (Cumming et al., 1972) because of the steadily in-
creasing costs of meat, poultry, dairy and fish products (Wolf, 1972). It might be
pointed out that to-date, the terms ‘vegetable’ and ‘soy’ as used in the trade to de-
scribe textured plant proteins (TPP) have been practically synonymous. The term
‘textured vegetable protein’ has been sweepingly applied to soy-derived products
containing about 50% protein and varying considerably in textural properties. The
abbreviation TVP, which has heretofore been used quite loosely, is a registered
trademark; therefore, TPP will be used as a generic name in this paper. Along with
the increased uses and acceptance of the vegetable proteins, particularly soy protein,
has come the approval by the U.S. Department of Agriculture in February 1971 of
combining textured vegetable proteins with ground beef in foods used in the school
lunch program (USDA, 1971). Several large grocery chains in the U.S.have begun
marketing soy-ground beef mixtures as substitutes for regular ground beef.
Textured soy proteins are not used as fillers. Rakosky (1972) explained that soy
protein is a highly nutritious food (it contains the eight essential amino acids required
for human nutrition) which just happens to have been processed so that it looks and
tastes like something else. Furthermore, textured soy protein can be used as a distinct
product (e.g., along with ground beef) as well as a simulated product (e.g., a meat
* Paper No. 8816, Scientific Journal Series Article, Minnesota Agricultural Experiment Station,
St. Paul. Presented at the AACC Annual Meeting, Montreal, Canada, Oct. 20-24, 1974.
** Present Address: Central Soya Co., 1825 N. Laramie Ave., Chicago, Illinois 60639.

Journal of Texture Studies 6 (1975) 459-472. All Rights Reserved

Copyr&ht 0 1975 by D. Reidel Publishing Company, Dordrecht-Holland
460 WILLIAM MBREENE AND THOMAS G-BARKER

analog) because, as a major source of protein in the human diet, it is both nutritious
and acceptable (Koury and Hodges, 1968).
Textured soy protein products can be made by continuous extrusion under heat
and pressure. The procedure is based on preconditioning and passing through an
extruder under specific conditions of moisture, temperature, time and pressure, a
premix of soy flour and added components (e.g., fat, flavorings, carbohydrates,
coloring, etc.). When the product coming out of the extruder reaches atmospheric
pressure, it will usually expand. Different particle shapes and sizes are obtained by
varying the shape of the die and the speed of the revolving knife at the end of the
extruder (Smith and Circle, 1972).
Textured plant proteins are produced by three methods : (I) extrusion-expansion,
(2) compaction and (3) spinning. The second method results in a product lacking
‘puff or expanded characteristics, but may have a plasticized nature (Smith and
Circle, 1972). This third process involves pumping an alkalinized ‘dope’ through
spinnerets into an acid bath. The latter products were not included in this research
because of their more limited use due to higher cost.
1.2. SELECTION
OF SPECIFICTEXTUREDVEGETABLE PROTEIN

When textured soy proteins are used as substitutes for other products, their textural
as well as nutritional properties should be similar to those of the product being re-
placed. Specific products used in the school lunch program, based on texture charac-
teristics, have been chosen empirically (Anon., 1972); other users have probably
chosen products in a similar way. Since sensory texture analysis is time consuming,
laborious and not easily replicated, a suitable objective procedure must be sought.
1.3. TEXTURE
OF TEXTURED VEGETABLE PROTEIN

Textured (soy) protein products have been defined as:

fabricated palatable food ingredients professed from edible protein sources including among others
soy grits, soy protein isolates and soy protein concentrates with or without suitable optional in-
gredients added for nutritional or technological purposes. They are made up as fibers,sheds,chunks,
bits, granules, slices or other forms. When prepared for consumption by hydration, cooking, re-
torting or other procedures, they retain their structural integrity and characteristic ‘chewy’ texture
(Anon., 1972).

The U.S.Department of Agriculture (1971) has defined textured vegetable protein

products for use in the school lunch program as:
food products made from edible protein sources and characterized by having a structural integrity
’
and identifiable structure such that each unit will withstand hydration and cooking, and other
procedures used in preparing the food for consumption.

The Food Protein Council of the National Soybean Processors Association has
defined textured vegetable protein as being:
derived from edible vegetable protein sources, and mixtures thereof, by physical and/or chemical
processes. These materials are converted into a structural form having definable textural properties
similar to those of the food system for which its use is intended. Food ingredients andlor additives
 TEXTUREMEASUREMENTPROCEDUREFORrrXTUREDVEGETABLEPRO~IN 46 1

may be incorporated to enhance its nutritive value and organoleptic properties. Textured vegetable
protein shall contain not less than 35% protein (Nx6.25) on a moisture free basis. The protein
efficiency ratio (P.E.R.) shall be not less than 1.8 on the basis of P.E.R. of 2.5 for casein.

A very important part of this definition is “definable textural properties”. It was

the purpose of this research to develop an objective method for determining differences
in textural characteristics of textured vegetable (soy) protein products, other than
spun fiber types, as a first step toward defining what is meant by the descriptive term
“definable textural properties” as it relates to them. Since ground beef is being ex-
tended with textured soy products, its textural characteristics were also studied, as
were those of ground beef: soy mixtures.

2. Materials and Methods

2.1. SAMPLEDESCRIPTION
Sixteen different TPP products (approximately 50% protein) from nine commercial
suppliers were used in developing and applying objective procedures. These de-
hydrated products were stored at room temperature in closed plastic bags during
the study. Products were selected to include those likely to exhibit textural character-
istics across a fairly wide range. Thirteen products, listed in Table I, were subjected
to an in-depth analysis using the developed objective method.

TABLE 1
Textured protein products evaluated by the Minnesota Texture Test

Product Code Supplier

TVP 240 U Archer Daniels Midland

Mira-Tex 210 b A. E.Staley
Bontrae Crumbles c General Mills
Vita Pro d Lauhoff
Supro 50 Regular e Ralston Purina
Toasted Nutrisoy Flakes f Archer Daniels Midland
Textrasoy 412 l? H.B. Taylor
Soyabits 25T h Central Soya
Supro 50 Minced i Ralston Purina
Textrasoy 24 i H.B. Taylor
Maxten k Worthington Foods
Meatone Grits I Archer Daniels Midland
Mira-Tex 240 m A. E. Staley

Two lots of ground beef containing 20% and 28% fat, respectively, were obtained
from two supermarkets in 1-lb packages. The fat content was determined on duplicate
1 oz samples by a modified Babcock test using concentrated H2S04 digestion. An-
other lot was obtained from a third supermarket. The packages were kept refrigerated
during the testing period, which usually did not exceed three days.
462 WILLIAM M. BREENE A N D THOMAS 0. BARKER

2.2. SAMPLE
PREPARATION

Dry TPP products were sized with Cenco-Meinzer sieves and shaker. That portion
of each product which fell through a 4-mesh sieve and was caught on a 20-mesh
sieve was used. Rehydration was accomplished by mixing 100 g of dry product with
200 g of 1 "C water, stirring, covering, and allowing to stand for 2 h at room tem-
perature. Before instrumental testing, the rehydrated products were ground in an
electric meat grinder using a grid with 3/8 inch diameter holes.
In preparing ground beef-soy mixtures, a 25 g portion of a rehydrated soy product
was mixed by hand with 75 g of raw ground beef. It should be noted that supermarkets
usually mix a rehydrated soy product with boned beef and grind the mixture.

Fig. 1. OTMS 10 cm2cell and the eight wires grid.

 TEXTURE MEASUREMENT PROCEDURE FOR TEXTURED VEGETABLE PROTEIN 463

Patties were formed by pressing 80g of ground beef, or of beef/TPP mixtures,

into a 4 inch diameter circular mold. Patties were cooked 2 min on each side in a
teflon-coated aluminum fry pan over a gas range with burner control set at 350"F,
and ground as described above.

2.3. TEXTURE
TESTING

Samples were compressed-extruded in a I0 cm2 Ottawa Texture Measuring System

cell(Voisey, 1971a)with the eight wires grid (Figure 1) affixed to the Instron Universal
Testing Machine (Figure 2). The cell was filled by hand and sufficient pressure was
applied to provide uniform compactness without exceeding the elastic limit of the

Fig. 2. Instron set-up with the OTMS cell and plunger.

464 WILLIAM M.BREENE AND THOMAS 0.BARKER

material. The cell was slightly overfilled and a knife was drawn across the top to
remove the excess material. Each sample was tested in five replicates.
The Instron 100 kg CTM compression load cell was calibrated using a Morehouse
Model 5C proving ring force gauge. The crosshead speed was 5 cm min-' and the
chart speed was 10 cm min-'. The gauge length was set to stop the piston 0.5 cm
from the wire grid.
Textural parameters were determined from the force-distance curves essentially as
suggested by Voisey (1971b) and Voisey et al. (1972). Seven parameters were deter-
mined (Figure 3), five of which were deemed useful in this study. The definitions of
these seven parameters (most of which are based on those published by Szczesniak,
1963) are as follows:
(a) Hardness is the force necessary to attain a given deformation. According to
Voisey (1971b) the average slope of the initial (approximately linear) portion of the
curve in kg cm-' can be interpreted to indicate hardness.
(b) Cohesiveness relates to the strength of the internal bonds making up the body
of the product. It is quantified as the force in kg at which shearing and extrusion are
initiated.
(c) Extrudubility is a function of hardness and cohesiveness. It is taken as the aver-
age slope of the curve in kg cm-' after the onset of shearing and extrusion.

Fig. 3. A typical force-distance curve for compression of a soy protein product on the lnstron
UTM-M using the OTMS wire extrusion cell: A =hardness, B=cohesiveness, C=extrudabiIity,
D=chewiness. E=maximum force, F=average maximum force, G=packability.
 TEXTURE MEASUREMENT PROCEDURE FOR TEXTURED VEGETABLE PROTEIN 465

(d) Chewiness is the energy required to masticate a solid food product to a state
ready for swallowing. It is interpreted as the area under the curve. The area represents
the energy used to compress, shear and extrude the sample. Area measurements were
made with a planimeter calibrated to read in square inches (multiply by 6.452 to
convert to cm’).
(e) Maximum force is the maximum force in kg attained during extrusion.
(f) Average maximum force is the average force in kg required to maintain a
constant extruding force. This parameter is less prone to variability than the maxi-
mum force, since extreme values due to individual hard particles or voids are not
included.
(g) Puckability is the distance (cm) traveled by the plunger before an average
linear slope is reached. Packability indicates the relative compactness of the sample
in the test cell.
Extrudability was not included in this work since all products produced similar
relatively flat curves. Values for packability were also similar for the various products.
Therefore, the parameters recorded were hardness, cohesiveness, chewiness, maximum
force and average maximum force.

3. Results and Discussion

3.1. DEVELOPMENT
OF A STANDARD TEST

Certain factors relating to sample preparation were found to influence parameter

values. The test procedure detailed in Materials and Methods section was selected
after studying the effects of the most important factors. Criteria for the test included
simplicity, rapidity, repeatability and the ability to detect differences in texture.
The size of individual particles and the range in particle size both vary considerably
among different dry commercial soy protein products. Preliminary testing revealed
that size effects would be minimized by eliminating particles at the extremes of the
particle size range, i.e., testing the material which passes through a 4-mesh sieve and
catches on a 20-mesh sieve. Particles which fall through the 20-mesh sieve, when
rehydrated, probably will be slightly smaller than the space between the wires of the
employed test cell. This will not allow each small particle to be sheared in the cell.
Figure 4 illustrates this size effect on parameter values for Vita Pro (produced by
Lauhoff) ground in a blender, sized and rehydrated.
Increasing the rehydration ratio (water:dry product, weight basis) from 2: 1 to
3 : 1 to ‘full rehydration’ (excess water followed by draining) generally decreased
parameter values, everything else being equal, We recommend that products con-
taining approximately 50% protein be rehydrated using the 2: 1 rehydration ratio.
Time of rehydration and the temperature of the water are somewhat interrelated.
Less time is required for rehydration in boiling than in ice water. Commercial users
of TPP usually rehydrate in ice water to reduce the possibility of microbial growth.
Textural parameter values were generally slightly higher for products rehydrated 2 : 1
in ice water (1 “C) vs those rehydrated in boiling water (100°C). Ice water was selected
 -
466 WILLIAM M. BREENE A N D THOMAS G . BARKER

CAUGRT 3N A #b MESH SCREEN

CAUGHT ON A #10 MESH SCREEN

0 CAUGHT ON A #20 MESH SCREEN

FELL THRWGH A #20 MESH SCREDi

A B C D E
Kgbm Kg Kg Kg Sq in.
TEXTURE PAR A M E T E R S
Fig. 4. Hardness (A), cohesiveness (B), average maximum force (C), maximum force (D) and
chewiness ( E ) of Vita Pro as influenced by particle size. Product was ground in a blender, screened
and rehydrated.

for the developed method and a rehydration time of 2 h allowed all products to
rehydrate uniformly.
Grinding after rehydration improved homogeneity, especially of products having
a relatively wide range of particle size and shape, and facilitated uniform packing in
the cell. Although textural parameter values were generally decreased by grinding,
relative values among products were essentially unaffected. Grids with holes smaller
than 3/8 inch (3/16or 118 inch) produced too much lowering of textural parameter
values.
Bulk density differences among dry products were lessened by sizing, rehydration
and grinding. Consequently, parameter values were found to be about the same whether
the cell was filled by volume (full) or by weight (84 g).
The developed standard testing procedure will be referred to hereafter as the
Minnesota Texture Method.
 TEXTURE MEASUREMENT PROCEDURE FOR TEXTURED W E T A B L E PROTEIN 467

3.2. OBJECTIVE
TEXTURE CHARACTERIZATION OF SOY PRODUCTS

Mean values and 95% confidence intervals for maximum force, average maximum
force, cohesiveness, hardness and chewiness for products listed in Table I are given
in Figure 5. The values are plotted in the order of decreasing magnitude. Mean
values whose 95% confidence intervals do not overlap can be assumed to be signi-
ficantly different. Products a-e are generally considered in the trade to be ‘textured’.
Products i, k and m are extruded, expanded, but have a relatively small particle size.
Maximum force values separated the products roughly into three textural groups
or classes. Product a (TVP 240) had the highest mean maximum force value; in fact,
it was considerably higher and significantly different from all the other products.
Products 6 , c and d (Mira-Tex 210, Bontrae Crumbles and Vita Pro) formed a group
having relatively high mean maximum force values of about equal magnitude.
Product e (Supro 50 Regular) overlapped this group and a second group of products,
f, g, h, and i (Toasted Nutrisoy Flakes, Textrasoy 412, Soyabits 25T and Supro 50
Minced). A third group included Textrasoy 24, Maxten, ADM Meatone Grits, and
Mira-Tex 240 which showed the lowest mean maximum force. Supro 50 Regular, an
extruded-expanded product which overlapped the first and second groups, showed a
high degree of variability. The lowest value for the highest maximum force group
(a, b, c, d ) was approximately 25 kg, 6 kg greater than the highest value for the other
products, excluding Supro 50 Regular.
In spite of eliminating high and low peaks which should decrease the variability
of data, the average maximum force was quite variable for Supro 50 Regular. Mean
values showed a trend similar to that of maximum force, i.e., means for products
a, b, c and d were higher than those for the other products. Again, TVP 240 ex-
hibited a higher mean value and was significantly different from all the other products.
Products b-d had similar mean values; Supro 50 Regular overlapped high (a, 6 , c, d )
and intermediate (f,g, h, i ) groups. Products j-m comprised a third group, with
Mira-Tex 240 (m)being the lowest at 3.0 kg. The lowest mean value for group a-d,
again excluding Supro 50 Regular, was 22 kg. With the highest mean value in the
second group (f-i) being 15 kg, it appears that the average maximum force can be
used to separate soy protein products into textural classes.
Mean cohesiveness values showed trends similar to maximum force and average
maximum force values. Products a-d were higher in mean cohesiveness than the other
products. However, unlike maximum and average maximum force values, cohesive-
ness of product a was not significantly different from that of products b, c and d ; this
was due in part to greater variability as evidenced by higher 95% confidence intervals.
Supro 50 Regular (e)mean cohesiveness did not overlap the first and second groups; it
was in the second group (e-i). Products in the third group (j-m) were about equal in
cohesiveness. The lowest value for group a-d was about 15 kg. The highest value for
the other products was approximately 12 kg. Although absolute differences in co-
hesiveness among classes were less than in the case of maximum force or average
maximum force, the range of mean values was also proportionately less. It may, thus,
468 WILLIAM M.BREENE A N D THOMAS 0.BARKER

be concluded that cohesiveness appears to be a useful parameter for textural classi-

fication of soy products.
Highest mean chewiness was exhibited by TVP 240 (a) followed by Mira-Tex
210 (b), Vita Pro (d) and Bontrae Crumbles (c). Supro 50 Regular (e) fell in the second
group (e-i). Productsf-i showed essentially the same trends as in the cases of the other
textural parameters discussed above. The final four products ranged from 4 in' to
2in' with Mira-Tex 240 (m) showing the lowest mean value. Chewiness clearly
separated the products into three distinct textural classes and appears to be useful
as a texture-defining parameter.

Fig. 5. Mean values and 95 % confidence intervalsfor textural parameters of 13 soy protein products
rehydrated and tested as described in Materials and Methods.
 TEXTURE MEASUREMENT PROCEDURE FOR TEXTURED VEOETABLE PROTEIN 469

Mean hardness values were the highest for TVP 240 (a) and Vita Pro (d). Two
other extruded, expanded products, Mira-Tex 240 (b) and Bontrae Crumbles (c),
were similar to each other and next highest in mean hardness. Toasted Nutrisoy
Flakes (f ) had the highest mean values for maximum force, average maximum force
and cohesiveness of the products in the intermediate texture class. However, its mean
hardness was among the lowest of the 13 products. Hardness does not, at first ob-
servation, appear to distinguish as well among the poorer textured products as the
other parameters. Products i, k and m were intermediate in hardness, whereas
productsf, g and h were low in hardness. However, products i, k and m are extruded-
expanded products having a relatively small particle size. Products .f,g, h, j and i are
not extruded-expanded.It is assumed that when a considerable proportion of particles
are smaller than the spaces between the wires in the test cell, each particle will not
undergo shearing during testing; lower values can, therefore, be expected for param-
eters whose values are affected by shearing. Since hardness is determined from the
slope of the compression curve before any shearing occurs, it is probably much less
influenced by particle size. Hence, hardness might be very useful in defining texture
of some of the more finely divided products,
The Minnesota Texture Method separated soy protein products into distinct
textural classes. Parameters which detected the greatest differences between classes
were average maximum force and maximum force, followed, and in this order, by
chewiness, cohesiveness and hardness. More data will be required before definitive
correlation of the instrumental texture parameters with organoleptic texture sensa-
tion is proven and specifications established.
3.3. TEXTURAL
COMPARISON OF GROUND BEEF,SOY PRODUCTS AND
THEIR MIXTURES
Another potential application of the Minnesota Texture Method is in comparing the
texture of TPP with the texture of ground beef, and in demonstrating the effect on
cooked patties of partial substitution of ground beef with soy protein.
Figure 6 shows mean textural parameters of 20 and 28% fat cooked ground beef as
well as of four soy protein products. The soy products included two extruded-ex-
panded products (TVP 240 and Bontrae Crumbles) and two compacted products
(Textrasoy 412 and Meatone Grits). Hardness, cohesiveness, maximum force and
average maximum force values were lower in the higher fat ground beef as compared
to lower fat samples, whereas chewiness was slightly higher. The tested textural
parameters of TVP 240 were very nearly the same as those of 28% fat ground beef.
Although hardness and cohesiveness were similar for 28% fat ground beef and
Bontrae Crumbles, maximum force and average maximum force values indicated
that Bontrae Crumbles was not as highly textured as either ground beef or TVP 240.
Textrasoy 412 and Meatone Grits were lower than ground beef and extruded-ex-
panded soy products in all the textural parameters. These data suggest that cooked
ground beef: soy patties (75: 25) should be texturally similar to pure cooked ground
beef if the soy ingredient is texturally similar to cooked ground beef. A 25% content
470 WILLIAM M. BREENE AND THOMAS G . BARKER

2@ FAT GROUND BEEF BONTRAE CRUMBLES

28% FAT GROUND BFXF TEXTRASOY #412

50 n ADM MEATONE GRITS

r
TW &LO

E
Kg/cm Kg Kg Kg Sq in.
TEXTURE PARAM E T E R S
Fig. 6. Hardness (A), cohesiveness (B), maximum force (C), average maximum force (D), and
chewiness (E) of two cooked ground beef samples and four rehydrated soy protein products prepared
and tested as described in Materials and ‘Methods.

El CONTROL GROUND BEEF AD14 MEATONE GRITS t GROUND BEEF

rn BONTRAE CRUMBLES + GROUND BEEF
a 90 PDI soy FLAKES + GROUND BEEF
MIRA-TEX 240 + GROUND BEEF

Kg icrn Kg Ka Sq in.
TEXTURE PA R A M E TE R S
Fig. 7. Hardness (A), cohesiveness (B), maximum force (C), average maximum force (D), and
chewiness ( E ) of a cooked ground beef sample and four cooked ground beef-soy (3:l) mixtures
prepared and tested as described in Materials and Methods.
 TEXTURE MEASUREMENT PROCEDURE FOR TEXTURED VEGETABLE PROTEIN 47 1

of any poorly textured soy product would probably alter the textural characteristics
of the patties.
Figure 7 shows how the texture of a ground beef control compared with that of its
75: 25 mixtures with a rehydrated product from the highest texture class (Bontrae
Crumbles), with a rehydrated product from the lowest texture class (either Mira-
Tex 240 or Meatone Grits) or with the unextruded raw material (90 PDI soy flakes).
Control ground beef was a single lot purchased from one supermarket. The fat
content was constant throughout and was not determined. Hardness appeared to be
the least discriminatory of the five parameters. Cohesiveness, maximum force,
average maximum force and chewiness data indicated that the textural quality was
maintained reasonably well only in the mixture containing the Bontrae Crumbles.
In a similar comparison, hardness values of 75 :25 mixtures of ground beef with
Supro 50 Regular or Supro 50 Minced were about equal (ca. 12-13 kg cm-I). Values
for the other four parameters were lower than the control in the Supro 50 Regular
mixtures and lower still in the Supro 50 Minced mixtures.
The Minnesota Texture Method appears to be adequate to quantify objectively the
textural characteristics of soy protein products. Our next step in this research will
be to determine how well these objective measurements correlate with the sensory
evaluation.
4. Conclusions

An objective texture profile procedure has been developed which is capable of

separating soy protein products into different texture classes. Of the five parameters
in the profile, average maximum force, maximum force, chewiness and cohesiveness
appear to be very useful. Although hardness appears to be less useful overall, it may be
applicable to the textural characterization of the more finely divided products. The
method can be used to test rehydrated soy products, ground beef, or ground beef: soy
mixtures. Texture of ground beef varies due to differences in fat content, among other
factors. If textural parameter values of a given soy product are lower than those of a
given ground beef product, they should also be lower for cooked mixtures of the two.

Acknowledgement
We thank the Food Protein Council of the National Soybean Processors Association
for their counsel and financial support. We are particularly grateful to Mr Richard
Kellor of Cargill, Inc. We also appreciate the help and suggestions given to us by
Mr Peter Voisey and Mr Martin KIoek of the Engineering Research Branch, Canada
Department of Agriculture.
References
Anonymous: 1972, ‘Notebook on Soy: Textured Vegetable Protein’, School Food Service J. 26, 51.
Cumming, D. B., Stanley, D. W. and deMan, J. M.: 1972, ‘Texture-Structure Relationships in
Textured Soy Protein. II: Textural Properties and Ultrastructure of an Extruded Soybean Product’,
Can. Inst. Food Sci. Technol. J. 5, 124.
472 WILLIAM M.BREENE AND THOMAS G.BAKKER

Fischer, R. W. : 1972,Soy Protein Products Marketed by American Companies, Am. Soybean Assoc.,
Hudson, Iowa 50643.
Koury, S. D., and Hodges, R. E.: 1968,‘Soybean Proteins for Human Diets?’, J. Am. Dietet. Assoc.
52, 480.
Rakosky, J. Jr.: 1972,‘Soy Protein -Nutritious Answer to High Food Costs’, Cunner Packer 141,8.
Smith, A. K.,and Circle, S.J.: 1972,‘Protein Products as Food Ingredients’, Soybeans: Chemistry
and Technology, ed. by A. K. Smith and S.J. Circle, Vol. I, AVI Publishing Co., Westport, Conn.
Szczesniak, A. S.: 1963,‘Classification of Textural Characteristics’, J. Food Sci. 28, 385.
USDA, ARS: 1971,‘Textured Vegetable Protein Products (B-I)’, FNS Notice 219.
Voisey, P. W.: 1971a, ‘The Ottawa Texture Measuring System’, Can. Inst. Food Sci. Technol. J. 4,
91.
Voisey, P.W.:1971b, ‘Use of the Ottawa Texture Measuring System for Testing Fish Products’,
Internal Report 7022,Eng. Res. Service, Canada.
Voisey, P. W.,MacDonald, D. C.,Kloek, M., and Foster, W.: 1972,‘The Ottawa Texture Measuring
System - an Operational Manual’, Engineering specification 7024, Eng. Res. Service, Canada.
Wolf, W. J.: 1972,‘Purification and Properties of the Proteins’, Soybeum: Chemistry and Technology,
ed. by A. K. Smith and S. J. Circle, Vol. I, AVI Publishing Co., Westport, Conn.