1. 1:. a 5...... a... .m \w. .. am .5154» .. M325: _ . .. 3 . i : Wm. . n... <1— ugh? a“. . axismmfi 9 fl. .1 ibnmu . ”mm .. an? .. , ufiwgwnwé in... H- UJ This is to certify that the thesis entitled Rheology and Sensory Analysis of Hot Cereals presented by Julie Jean DeJongh has been accepted towards fulfillment of the requirements for MS degree in Food Science JEWM M Major professor Date '5/973/551 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution ll LIBRARY Michigan State University PLACE IN RETURN Box to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 cJCIRC/DateDuepes-m 5 RHEOLOGY AND SENSORY ANALYSIS OF HOT CEREALS By Julie Jean DeJongh A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 2002 ABSTRACT RHEOLOGY AND SENSORY ANALYSIS OF HOT CEREALS By Julie Jean DeJongh The vane method was used to determine the rheological behavior of hot cereals. Five cereals were evaluated: Nabisco 2 1/2 Minute Cream of Wheat (COW), Nabisco Instant Original Cream of Wheat (ICOW), Malt-o-Meal Co. Malt-o-Meal (MOM), Quaker Oats Co. Quick One Minute Quaker Oats (lMIN), and Quaker Oats Co. Instant Regular Oatmeal (IOAT). The apparent yield stress of the prepared cereals was calculated using the vane method from raw data collected with a low speed rotational viscometer. The effects of time (3 to 20 minutes) and cooling on the apparent yield stress were observed. As time (hydration time) increased and temperature decreased, the apparent yield stress of each cereal increased. A descriptive analysis panel was used to generate scores relating to four sensory parameters: thickness (stirrability), stickiness to spoon, stickiness to self, and viscosity. The sensory scores were compared to the apparent yield stress values. Through regression analysis, it was determined that there is no correlation between any of the sensory parameters and the instrumentally obtained apparent yield stress values for any of the cereals, at any of the hold times studied. The apparent yield stress method is, however, an effective technique for evaluating and comparing the rheology of hot cereals. ACKNOWLEDGEMENTS I would like to thank Dr. James Steffe for his unfaltering support and encouragement during the time I have spent at Michigan State University. I hope to follow his example throughout my personal and professional lives. I would also like to acknowledge Dr. Kirk Dolan and Dr. Perry Ng for serving on my committee. Additionally, Dr. Janice Harte and Ms. Christine Ebeling were indispensable to my understanding and practice of sensory science. Ms. Maryn Zengerle and all the individuals who spent the time and effort to serve on my descriptive analysis panel deserve recognition, as well as Mr. Richard Wolthuis, who provided me with excellent technical support. I must also acknowledge and thank Brookfield Engineering for providing me with the Yield Rheometer, which made my research possible, as well as the excellent technical support from Mr. Greg Krysko. Lastly, I wish to thank my parents, Mr. and Mrs. Jay DeJongh for encouraging me, and my boyfriend, Mr. Eugene Chio for keeping me grounded and focused. iii Table of Contents Page List of Tables vi List of Figures ix 1. Introduction 1 2. Literature Review 4 2.1. Breakfast Cereals 4 2.1.1. Oat-Based Cereal 4 2.1.2. Wheat-Based Cereal 6 2.2. Starch 9 2.2.1. Starch Molecules 9 2.2.2. Gelatinization 12 2.2.3. Starch Characteristics From Wheat and Oat Species 13 2.3. Yield Stress 15 2.3.1. Static versus Dynamic Yield Stress 16 2.3.2. Methods to Measure Yield Stress 17 2.3.3. Vane Method 19 2.4. Sensory 23 2.4.1. Evolution of Sensory Science 23 2.4.2. Descriptive Analysis 24 2.4.2.1. Texture Profile Method 25 2.4.2.2. Quantitative Descriptive Analysis® 26 2.4.2.3. SpectrumTM Method 26 2.4.2.4. Free Choice Profiling 26 2.4.2.5. General Descriptive Analysis 27 2.5. Correlative Studies 29 3. Materials and Methods 33 3.1. Rheological Measurements 33 3.2. Descriptive Analysis Measurements 39 3.2.1. Screening Panel 39 3.2.2. Trained Descriptive Analysis Panel 40 4. Results and Discussion 47 4.1. Rheological Behavior 47 4.2. Descriptive Analysis 54 4.3. Correlation of Rheological and Sensory Parameters 63 iv 5. Future Research 6. Summary 7. Appendix 7.1. Apparent Yield Stress Measurements for Wheat-Based Cereals 7.2. Apparent Yield Stress Measurements for Oat-Based Cereals 7.3. Sensory Scores Based on 15cm Line Scale for Wheat-Based Cereals 7.4. Sensory Scores Based on 15cm Line Scale for Oat-Based Cereals 7.5. ANOVA Results for Descriptive Analysis for Wheat-Based Cereals 7.6. ANOVA Results for Descriptive Analysis for Oat-Based Cereals References 71 72 74 74 76 77 81 85 89 91 Table 2.1. 2.2. 2.3. 2.4. 3.1. 3.2. 3.3. 3.4. 3.5. 4.1. 4.2. 4.3. 4.4. 4.5. 4.6. 4.7. LIST OF TABLES US sales by calendar year of all hot cereals. Amylose content of starches. Properties of amylose and amylopectin. Some properties of whole granular starches. Proportions of water and cereal for sample preparation. Spindle size used for measurement of each sample. Comprehensive list of descriptors for Nabisco Instant Original Cream of Wheat and Quaker Oats Instant Regular Oatmeal. Descriptors with definitions and testing methods from Days 1 and 2 of Session 1. Reference samples, sample code numbers, and sample identifications for Repl, Rep2, and Rep3. Average apparent yield stress values. Results for descriptive analysis panel scores for wheat cereals and oat cereals for the parameter “thickness (stirrability).” Results of descriptive analysis panel scores for wheat cereals and oat cereals for the parameter “stickiness to spoon.” Results of descriptive analysis panel scores for wheat cereals and oat cereals for the parameter “stickiness to self.” Results of descriptive analysis panel scores for wheat cereals and oat cereals for the parameter “viscosity.” Equations and cereal ages used to find apparent yield stress values evaluated by sensory panel. Correlation coefficient for regression analysis for wheat- and oat-based cereals for four sensory parameters. vi Page 11 14 33 35 41 41 46 49 55 56 57 58 64 64 7.1. 7.2. 7.3. 7.4. 7.5. 7.6. 7.7. 7.8. 7.9. 7.10. 7.11. 7.12. 7.13. 7.14. 7.15. 7.16. 7.17. 7.18. Measurements for apparent yield stress for each hold time for 2 '/2 Minute Cream of Wheat. Measurements for apparent yield stress for each hold time for Instant Regular Cream of Wheat. Measurements for apparent yield stress for each hold time for Malt- 0- Meal. Measurements for apparent yield stress for each hold time for One Minute Regular Oatmeal. Measurements for apparent yield stress for each hold time for Instant Regular Oatmeal. Sensory scores for thickness (stirrability) for wheat-based cereals. Sensory scores for stickiness to spoon for wheat-based cereals. Sensory scores for stickiness to self for wheat-based cereals. Sensory scores for viscosity for wheat-based cereals. Sensory scores for thickness (stirrability) for oat-based cereals. Sensory scores for stickiness to spoon for oat-based cereals. Sensory scores for stickiness to self for oat-based cereals. Sensory scores for viscosity for oat-based cereals. AN OVA table for wheat cereals for the sensory parameter “thickness (stirrability).” Effect of subject variation between wheat cereals for the sensory parameter “thickness (stirrability).” Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “thickness (stirrability).” AN OVA table for wheat cereals for the sensory parameter “stickiness to spoon.” Effect of subject variation between wheat cereals for the sensory parameter “stickiness to spoon.” vii 74 74 75 76 76 77 78 79 80 81 82 83 84 85 85 85 86 86 7.19. 7.20. 7.21. 7.22. 7.23. 7.24. 7.25. 7.26. 7.27. 7.28. 7.29. 7.30. Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “stickiness to spoon.” ANOVA table for wheat cereals for the sensory parameter “stickiness to self.” Effect of subject variation between wheat cereals for the sensory parameter “stickiness to self.” Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “stickiness to self.” ANOVA table for wheat cereals for the sensory parameter “viscosity.” Effect of subject variation between wheat cereals for the sensory parameter “viscosity.” Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “viscosity.” ANOVA table for oat cereals for the sensory parameter “thickness (stirrability).” ANOVA table for oat cereals for the sensory parameter “stickiness to spoon.” AN OVA table for oat cereals for the sensory parameter “stickiness to self.” AN OVA table for oat cereals for the sensory parameter “viscosity.” Effect of subject variation between oat cereals for the sensory parameter “viscosity.” viii 86 87 87 87 88 88 88 89 89 89 90 90 Figure 2.1. 2.2. 2.3. 2.4. 3.1. 3.2. 3.3. 3.4. 4.1. 4.2. 4.3. 4.4. 4.5. 4.6. 4.7. 4.8. LIST OF FIGURES Chemical structures of amylose and amylopectin. Static and dynamic yield stresses. Cup and vane dimensions for the vane method. Torque-time curve at constant angular velocity. Dimensions of vanes used with Brookfield Yield Rheometer. Dimensions of the sample containers used for measurements. Ballot presented to panelists in Session H. Questionnaire for test sessions. Temperature versus hold time for wheat-based cereals. Temperature versus hold time for oat-based cereals. Apparent yield stress versus hold time for wheat-based cereals Apparent yield stress versus hold time for oat-based cereals Scalar illustration of results of descriptive analysis scores for wheat cereals and oat cereals for the sensory parameter “thickness (stirrability)” Scalar illustration of results of descriptive analysis scores for wheat cereals and oat cereals for the sensory parameter “stickiness to spoon” Scalar illustration of results of descriptive analysis scores for wheat cereals and oat cereals for the sensory parameter “stickiness to self” Scalar illustration of results of descriptive analysis scores for wheat cereals and oat cereals for the sensory parameter “viscosity” ix Page 10 17 19 20 36 36 43 44 48 48 50 50 55 56 57 58 4.9. 4.10. 4.11. 4.12. 4.13. 4.14. 4.15. 4.16. Sensory scores for thickness of wheat-based cereals versus apparent yield stress measurements Sensory scores for stickiness to spoon of wheat-based cereals versus apparent yield stress measurements Sensory scores for stickiness to self of wheat-based cereals versus apparent yield stress measurements Sensory scores for viscosity of wheat-based cereals versus apparent yield stress measurements Sensory scores for thickness of oat-based cereals versus apparent yield stress measurements Sensory scores for stickiness to spoon of oat-based cereals versus apparent yield stress measurements Sensory scores for stickiness to self of oat-based cereals versus apparent yield stress measurements Sensory scores for viscosity of oat-based cereals versus apparent yield stress measurements 65 65 66 66 67 67 68 68 1. Introduction Rheological properties of food materials have significant impact on consumer acceptance of the particular product. Consumers make judgments about products based on sensory perceptions. “Texture and mouthfeel are major determinants of consumer acceptance and preference for foods and beverages” (Guinard and Mazzucchelli 1996). Consumers may not like chewy steak, mushy apples, or pasty oatmeal. All of these problems involve the texture of the food. To monitor and evaluate issues such as these, it is first necessary to define the problem. The definition of texture varies depending on the source. Szczesniak (1963) published an article discussing this issue in which eight different definitions for the term “texture” were enumerated. Two key elements were distilled from the plethora of ideas: physical structure and mouthfeel (Szczesniak 1963). From this she created a method of classification for texture where characteristics are divided into three categories: 1) mechanical characteristics, 2) geometrical characteristics, and 3) other characteristics (relating to moisture and fat content). Mechanical characteristics refer to the reaction of food to stress, such as the stress of mastication. These characteristics can be described by the terms “hardness, cohesiveness, viscosity, elasticity, adhesiveness, brittleness, chewiness, and gumminess.” Geometrical characteristics are most often observed visually in the arrangement of the constituents of the food. They can be divided into two groups: qualities related to particle size and shape, and qualities related to shape and orientation. “Other” characteristics involve mouthfeel and other factors that do not fit into the first two classes, such as oiliness and greasiness. This expanded definition of texture lends itself to both instrumental and sensory evaluation of food. The next step after defining the parameter is to measure texture parameters of food. This can be done quantitatively through instrumental methods. Several instrumental methods exist to characterize food texture. The method used depends upon the parameters that one desires to explore. The parameter of interest to this thesis is yield stress (0'0). Yield stress is defined as the maximum shear stress required to initiate flow (Steffe 1996). One simple way to measure yield stress is by the vane method. Using a vane and sample cup of correct proportions, the vane is lowered into the sample and rotated at very low speed. The torque on the spindle is measured, and a yield stress is reached at the maximum torque recorded before initiation of flow. Texture perceptions in food are too complex to evaluate simply by instrumental means. To give meaning to the objective measurements, they must be related to sensory evaluation. Sensory data can be collected using a descriptive analysis panel, however; sensory panels are often time consuming and expensive. Hence, industry would prefer to use instruments that are simple and generate reproducible results. Understanding the relationship between texture of food and sensory perception allows growth in several areas of food science including product development, process engineering, and quality control. One segment of the food industry that could benefit from characterization of texture by instrumental and descriptive analysis is the hot cereals area. Currently, companies use only subjective techniques to evaluate the flow behavior of this product. A simple quantitative method that correlates to sensory parameters is needed. In response to this need, three objectives were formulated for the current research: 1) Develop experimental protocol to characterize flow behavior of hot cereals using the vane method to measure a yield stress parameter; 2) Evaluate the sensory characteristics of hot cereals with a trained panel using descriptive analysis; 3) Evaluate the relationship between instrumental measures of yield stress and sensory evaluation data. 2. Literature Review 2.]. Breakfast Cereals Hot cereals are often forgotten at the breakfast table while the more convenient cold, or ready-to-eat (RTE) cereals are preferred. However, as an alternative to breakfast foods other than RTE cereal, hot cereals are an inexpensive option that are undeniably a force in the market. Table 2.1 shows the annual trends in consumption of hot cereal between 1995-1998. Traditional grains used for hot cereals are wheat, rice, corn, oats, and barley. The hot cereals of interest to the current work are oatmeal and wheat-based, farina products. Table 2.1*: US Sales by Calendar Year of All Hot Cereals (Excluding Corn Grits)“ Product 1995 1996 1997 1998 Total hot cereals 347 (100.0%) 358 373 357 (100.0%) Total hot oatmeal 271 (78.1%) 286 303 293 (82.1%) Total hot otherc 76 (21.8%) 72 70 65 (18.2%) Total standard 234 (67.4%) 232 239 225 (63.0%) (non-instant) hot cereals Total standard oatmeal 170 (49.0%) 180 190 180 (50.4%) Total old-fashioned 57 (16.4%) 64 69 66 (18.5%) rolled oats Total quick oatmeal 105 (30.2%) 114 116 109 (30.5%) Total instant hot cereal 113 (32.6%) 126 134 132 (37.0%) Total instant oatmeal 101 (29.1%) 106 113 112 (31.4%) Total instant non— 12 (3.5%) 20 21 20 (5.6%) oatmeal *Taken from Caldwell et a1. 2000. ‘A.C. Neilsen data courtesy of The Quaker Oats Co. bData (other than %) are in millions of pounds (1.0 million lb = 453,600 kg). “Primarily wheat-based, farina products. 2.1.1. Oat-Based Cereals Ninety-five percent of oat crops are used for animal feed, but the remainder is consumed by humans (Caldwell 1973). Before the oats reach the breakfast table, they are processed to shorten the amount of cooking time required for preparation. The treatment of oats involves chemical methods that modify the oat starch for rapid gelatinization, and/or physical methods that involve grinding, heat treatment, or addition of gums to improve hydration and dispersability of fine particles (Daniels 1974). The basic process for preparing oatmeal is described in the following section. Whole oats are received, then separated and cleaned to remove trash, weeds, chaff, and dust. Oats are then hulled, typically by an impact huller, which uses centrifugal force to cause oats to strike the wall of the huller. The impact causes the groat to separate from the hull. Groat is composed of starchy endosperm, and makes up 50-55% of the whole oat. Afier hulling, the groats may be pearled or scoured, then dried and conditioned with heat to inactivate enzymes that may cause rancidity. The groats are sized next: the largest are used to make old-fashioned rolled oats. Remaining groats are cut into three to five pieces with a process called steel-cutting. Steel-cutting creates smaller pieces that hydrate more rapidly, without creating a lot of fine flour that would lead to a pasty hydrated product (Caldwell 2000). Three types of products arise from oat processing: old-fashioned rolled oats, quick oatmeal, and instant oatmeal. In making old-fashioned rolled oats, the whole (uncut) groats are steamed to a moisture content of 10-12%. They are then flaked on flaking rolls, and dried and cooled for the final product. Quick oatmeal undergoes the same treatment, however the starting material is steel-cut groats, which are then rolled thinner than the old-fashioned rolled oats. For instant oatmeal, steel-cut groats are used again and the flakes are thinner yet. The groats are also exposed to additional heat treatment that pregelatinizes the starch. A hydrocolloid gum is usually added to improve hydration of the instant cereal (Caldwell 2000). Several methods were patented in the past forty years for the optimized process of making oatmeal for hot cereal. In 1961, Quaker Oats Co. patented the use of edible polysaccharide gums, which are dry mixed into the rolled oats. Upon addition of boiling water, the gums hydrate rapidly and form gelatinous films on the oat flakes (Huffman and Moore 1961). In 1970, Tressler developed a treatment that reduced the cooking time of oats from ten to twelve minutes down to thirty to forty seconds. After the oats are rolled, they are exposed to two treatments that crack the flakes to form internal capillaries that rapidly absorb boiling water (Tressler 1970). The two treatments are: l) a dry heat treatment, ideally 300°F for ten to eighteen minutes, to remove the “raw” flavor and to make the flakes highly absorptive, and 2) placing the flakes under high pressure to produce thin, small flakes that also absorb water more rapidly. The pressure should be sufficient to flatten the flakes to half their original thickness (Tressler 1970). Also in 1970, National Oats Co., Inc. patented the use of slow-coating starches such as potato, tapioca, wheat, corn, and waxy-maize to permit boiling water to complete the gelatinization of the flakes while providing a coating which gives uniform texture and flavor to cooked oatmeal (Hanser and Martin 1970). In 1972, Nabisco, Inc. obtained three patents: addition of cereal hydrolysate, addition of oat fiactions, and addition of pregelatinized starch. These patents decrease hydration time of the oatmeal (Ronai and Spanier 1972a, 1972b, 1972c). 2.1.2. Wheat-Based Cereals Wheat, or farina-based hot cereals are second in popularity to oatmeal. Farina is defined in the United States Federal Code of Federal Regulations (1999) as: “(a) Farina is the food prepared by grinding and bolting cleaned wheat, other than durum wheat and red durum wheat, to such fineness that, when tested by the method prescribed in paragraph (b) (2) of this section, it passes through a No. 20 sieve [0.850mm openings], but not more than 3 percent passes through a No. 100 sieve [0.150mm openings]. It is freed from bran coat, or bran coat and germ, to such extent that the percent of ash therein, calculated to a moisture—flee basis, is not more than 0.6 percent. Its moisture content is not more than 15 percen .” In short, farina is wheat endosperm in granular form (Caldwell 2000). Farina comes from hard wheat, typically hard red spring or winter wheats. The endosperm of soft wheats disintegrates in hot water and would lead to a pasty and unacceptable cereal. During the milling of wheat, the first stream off the mill is the hard chunk of endosperm. This is referred to as the middling stream, and constitutes farina, which hydrates in boiling water. Instant farina products mirror the process for instant oatmeal. The whole wheat middling is saturated with water, pressure-cooked, flaked on flaking rolls, and dried. The process is constantly being optimized to create an instant farina with the same texture as the whole farina (Caldwell 2000). In the 1930’s and 1940’s, Cream of Wheat Corporation found that addition of disodium phosphate and pepsin reduced the cooking time of whole farina (Caldwell 2000). Cantor et al., in 1959, obtained a patent to incorporate a gum or a thickening agent into the cereal. Gums such as arabic, karaya, guar, and tragacanth decrease the amount of water needed by one third, and cut hydration time down from three minutes to thirty seconds. The added agents help suspend the farina particles in the hot water to increase the surface area available for hydration and gelatinization. In 1970, The Quaker Oats Co. patented a heat treatment process to denature wheat proteins in farina (Hyldon 1970). A year later, Ralston Purina patented a process for instant wheat cereal in which milled wheat was tempered at 85-104.4°C to a moisture content of 15-16% before being flaked in rolls to 0007-0008 inch thickness. Once the flakes are dried to 8-9% moisture, they rehydrate instantly in boiling water (Spring, Jr. 1971). Nabisco Brands, Inc. patented technology for the process of preparing instant, flaked wheat farina. The process includes mixing farina and guar gum, adding water and then tempering with agitation. The mix is then cooked at 110-120°C to gelatinize the starch before it is dried, tempered, and flaked (Karwowski 1985, 1986, 1987) 2.2. Starch The primary component of hot cereals is starch. Starch is the major stored form of carbohydrates in plants. In cereal grains in particular, the endosperm of the kemel contains most of the starch (Shannon and Garwood 1984), which is stored as granules. These granules are 2-150 pm in diameter and are partially crystalline in nature (Zobel 1984). Although primarily carbohydrate, starch granules also contain small amounts of lipids, proteins, and ash (French 1984). Starch granules develop in organelles called amyloplasts. When there is one granule per amyloplast, it is called a simple granule; the presence of two or more granules in an amyloplast is called a compound granule. In addition to varying in granule structure, starch also varies in relation to species, cultivars, growth environment, and genetic mutations (Shannon and Garwood 1984). 2.2.1 Starch Molecules The molecules that compose starch are amylose and amylopectin. Generally, amylose is present in the amount of 15-30%, depending on the plant source (Table 2.2). On the basis of amylose content, starch can be divided into three categories: waxy starch having 0-8% amylose, normal starch having 20-30% amylose, and high-amylose starch having greater or equal to 50% amylose (Jane 2000). Table 2.2. Amylose Content of Starches“ Starch Amylose (%) Barley 22 Corn 28 Oat 27 Rice 1 8.5 Wheat 26 * Taken from Young 1984. Amylose is essentially a linear polysaccharide composed of ( l+4)—linked a—D- glucopyranosyl units (Fig. 2.13), with a total average molecular weight of 250,000 daltons (Zobel 1984). Some studies have found that amylose may also have some slight branching (Shannon and Garwood 1984). Amylopectin is also (1 +4)-linked a-D- glucopyranosyl, but it has branches at (1 +6)«linkages over 5% of the structure (Fig. 2.1b), and the molecular weight ranges from 50 million to 100 million Daltons (Zobel 1984). Amylopectin is the main contributor to the crystallinity of starch granules (Swanson 2000). CHZOH CH20H CH20H “0 o o 0 OH OH OH OH OH OH OH CHZOH 0 OH \ CHgOH OH C H; CHZOH (b) o o 0 OH OH OH OH OH OH OH Figure 2.1. Chemical structures of (a) amylose and (b) arnylopectin (Jane 2000). 10 When exposed to heat and water, amylose is unstable and quickly precipitates to initiate gelation. Rigidity develops as the starch gel cools and ages. Amylose gels are firm and require temperatures of 1 15-120° to reverse (Zobel 1984). The branched structure of amylopectin gives it greater stability in water (Young 1984). Gelation of amylopectin occurs at a much slower rate and requires higher concentrations and temperature. The gels are soft and reversible at 50—85° (Zobel 1984). The properties of amylose and amlepectin are summarized in Table 2.3. Table 2.3. Properties of amylose and amylopectin“ Amylose Amylopectin Molecular shape Essentially linear (with few Branched branches) Molecular weight ~10° daltons ~108 daltons Retrograde Rapidly Slowly Film property Strong Weak and brittle *Taken fiom Jane 2000. Gel structure not only depends on individual starch fractions, but also on the interactions and synergy of the two together. For example, waxy starch, which is 0-8% amylose, produces a clear paste that disperses easily because amylose is not present to intertwine with amylopectin. Normal starch, which is so called because it is within the mid-range of amylose content with 20-30% amylose, produces a stronger gel (Jane 2000). Jane and Chen (1992) explored the role of amylose present in different concentrations and with different lengths, as well as amylopectin with different branch chain lengths on gel strength. They found that higher concentrations of amylose led to stronger gels. Amylopectin with greater branch chain lengths also tended to produce stronger gels. This was explained by the branch chains interacting with amylose: longer length branches interact to a greater extent. 11 2.2.2 Gelatinization Without the addition of heat and water, hot cereal is just powder. This section will describe what happens to the starch in the hot cereal when the consumer prepares the product for consumption. When dry starch is exposed to water at 0-40°C, the granules absorb a small amount of the water and undergo limited reversible swelling (French 1984). When heat is added in the presence of excess water, above a characteristic temperature the starch granules irreversibly swell and the amylose is solubilized. When this solution cools, the amylose associates to form a matrix in which the swollen granules are embedded (Ellis and Ring 1985). The umbrella term for the phenomenon is “gelatinization.” One definition of gelatinization comes from Daniel and Weaver (2000): “Starch gelatinization is the collapse (disruption) of molecular order within the starch granule manifested in irreversible changes in properties such as granular swelling, native crystallite melting, loss of birefi'ingence and starch solubilization. The point of initial gelatinization and the range over which it occurs is governed by starch concentration, method of observation, granule type and heterogeneity within the granule population under observation.” When starch is heated with excess water, several processes occur. French (1984) gives a comprehensive overview: As water is imbibed, “swelling begins in the least organized, amorphous, intercrystallite regions of the granule. As this phase swells, it exerts tension on neighboring crystallites and distorts them. Heating leads to uncoiling, or disassociation, of double helical regions and break-up of amylopectin crystallite structure. The liberated side chains of amylopectin become hydrated and swell laterally, further disrupting crystallite structure. The starch molecules are unable to stretch longitudinally, and may even contract to approach random coil formation. Increased molecular mobility with further hydration permits a redistribution of molecules; and the smaller, linear arnylose molecules may diffuse out. Further heating and hydration weaken the granule to the point where it can no longer resist mechanical or thermal shearing, and a sol results.” 12 This process takes place over a temperature range of 10-15°C (French 1984). The leaching out of the amylose takes place at a temperature of 57-100°C, while the granule is still intact (Young 1984). Gelatinization and subsequent cooling is accompanied by an increase in viscosity. Traditionally, this thickening was believed to be the result of swollen granules coming into contact with each other and inhibiting flow. Other research indicates that it is the exudate, amylose, forming a network outside of the granule that causes the system to become viscous (Miller et al 1973). 2.2.3 Starch Characteristics from Wheat and Oat Species The type of gel formed when a gelatinized starch solution cools depends on the molecular components of the starch, which varies with plant species. Wheat and oat starches are distinct from each other. In wheat, the starchy endosperm composes 82% of the kernel. The endosperm also contains some protein, pantothenic acid, riboflavin, and minerals (Swanson 2000). Doublier (1987) described wheat as 0.003% protein, 0.008% lipid, and 28% amylose. The percent of amylose in wheat ranges from 17-29%, depending on the cultivar (Shannon and Garwood 1984). Wheat starch is composed of simple granules, which come in two types depending on the growth stage of the plant. The first granules produced in the endosperm cells develop into large lenticular, disc- shaped granules. After two weeks, additional granules are produced which are small (<10um) and spherical. The large granules make up only 12.5% of the total starch by number, but 93.0% by weight (Shannon and Garwood 1984). Upon hydration plus heat, the lenticular granules transform into a saddle shape, but do not thicken (French 1984). The small granules swell as spheres. l3 Oat endosperm, which constitutes 50-55% of the grain (Caldwell 1973), has more protein and oils than other cereals (Swanson 2000). Oat granules are compound. They start out as round granules, but become angular as they are packed together in the amyloplast. For oats, swelling occurs in three dimensions, with the swollen granules appearing as larger versions of the unswollen granules (Williams and Bowler 1982). Also, in oat starch, contrary to other cereals, amylose and amylopectin are leached simultaneously. As a result, oat gels have a stronger internal network at equivalent concentrations of other cereal starches, such as wheat (Doublier et al 1987). Table 2.4 summarizes the properties of granules from different cereals. Table 2.4. Some properties of whole granular starches“. Source Gelatinization Granule shape Granule size (nm) Amylose content temperature range (%) (°C) Barley 51-60 Round or 20-25 22 elliptical 2-6 Wheat 58-64 Lenticular or 20-35 23-27 round 2-10 Oat 53-59 Polyhedral 3-10 23-24 Corn 62-72 Round or 15 28 polyhedral Rice 68-78 Polygonal 3-8 17-19 *Taken from Swanson 2000. 14 2.3. Yield Stress The texture of fluid foods is often used by the food industry as a benchmark for quality in product development and consumer acceptance of new foods, and in grading and quality control of traditional products (Timbers and Voisey 1987). Production of commercial foodstuffs must be reproducible to satisfy consumers. An effective way to determine reproducibility is to measure the rheological properties, which can be correlated to performance of the materials in storage stability, ease of pumping, and sensory perception, to name a few (Walton 2000). Marr and Pederson (1999), in their research on low-fat and full-fat mayonnaise, found that rheological properties are related to product quality. Similarly, in a paper concerning objective and subjective methods of characterizing the rheological properties of buffalo cream butter, Kulkami (1986) explained that 40% of the grading in butter comes from rheological attributes. Industries appreciate simple low-cost tests that produce a single value that can easily be translated into practical significance relevant to product usage (Barnes 2001). One such rheological parameter that fits these criteria is yield stress (0'0). As a point of reference, Barnes (1999) gives us a list of approximate yield stresses of some common fluid foods: so (1).) Product 15 Ketchup 25 Spaghetti sauce 60 Mustard and apple sauce 90 Mayonnaise 125 Tomato paste During storage and preparation, many foods form an internal network structure due to chemical changes over time. These changes can result in weak gels that may explain the phenomenon of yield stress in terms of structural deformation and breakage 15 of network bonds (Rao and Steffe 1997). Yield stress has traditionally been defined as the stress below which no flow takes place. This definition has been challenged by modern rheometry, which can take measurements at very low shear rates. The measurements Show that low levels of flow do occur at stresses below the “yield stress.” This information prompted Barnes (1999) to conclude that in reality, a yield stress does not exist because, given the appropriate time range, everything flows. To illustrate his point, he refers to a Biblical quote of the prophetess Deborah: “ ‘the mountains flow before the Lord,’” meaning that even rocks would flow if the time scale of observation is extended to geological scale. For engineering quality control purposes, yield stress is accepted as a reality. Usefirl definitions of yield stress include: 1) the minimum shear stress required to initiate flow at the shear rate used (Briggs 1996, Walton 2000). 2) The maximum stress that can be applied before the structure breaks down (Daubert 1998). 3) The stress that must be exerted to just move one fluid layer past another (Missaire 1990). 2.3.1. Static versus Dynamic Yield Stress There are two commonly recognized types of yield stress values: dynamic and static. Static yield stress is taken on a material at rest while dynamic measurements are taken on a material where the internal network system has been destroyed. Since the static measurements are taken on an undisturbed sample, the yield stress values are usually larger in magnitude than dynamic measurements. Figure 2.2 illustrates the concept of static and dynamic yield stresses (Steffe 1996). Yield stress is an important consideration in the food industry because it directly anticipates how a fluid will flow over a range of shear stress values over a given time 16 Static 0'0 / Equilibrium Flow Curve Shear Stress Dynamic Go Shear Rate Figure 2.2. Static and dynamic yield stresses. scale (Rao and Steffe 1997). Several studies have found that yield stress is important in coating solid surfaces and in keeping small particles in suspension (Y 00 et al. 1995) as well as for predicting pumping requirements for a desired flow rate (Lang and Rha 1981). Yield stress can also be used to calculate the thickness of layers of fluid food products leftover on the wall of vessels in a food processing system (Barnes 1999). Lang and Rha (1981) studied yield stresses of hydrocolloid dispersions and found them to be associated with the use of hydrocolloids as binders through retaining their shape. Yield stress has also been directly related to spreadability of products such as cream cheese (Breidinger and Steffe 2001) and butter (Mortensen and Danmark 1982) and plays a role in sensory evaluation and consumer acceptance of these products. 2.3.2. Methods to Measure Yield Stress There are numerous ways to measure yield stress. Steffe (1996) details several of the methods, but cautions that yield stress values are defined by the techniques used for measurement and the values from one method will not necessarily match those from 17 another. Therefore, it is essential to specify the conditions of the test when reporting the results to allow for reproducibility. Traditional methods to determine yield stress involve extrapolation of shear rate ( 7' ) and shear stress (a) data according to flow models. Popular equations for this purpose are Binghamza = 00 + r/py' (1) Casson: J5: J3:+ 1MP)" (2) Herschel-Bulkley: a = 00 + K(}/' )” (3) where the intercept of each equation is the yield stress (00). Taking measurements for these equations involves destructive shear on the product, which can affect the yield stress values; therefore, a direct determination may be more desirable. Established methods for direct determination include cup and bob, plate coating, and the centrifugal slump test. Wendin et a1 (1997) used a cup and bob to measure the yield stress of mayonnaise. In another study, Wendin and Hall (2001) employed a cone and plate to measure the yield stress of salad dressing. Marr and Pedersen (1999) measured the yield stress of mayonnaise with parallel plates. Omura and Steffe (2001) used a centrifugal viscometer to predict yield stress of fluid foods by measuring the slump in the materical induced from centrifugal acceleration. Mortenson and Danmark (1982) used three methods to determine yield stress of butter: direct measurements using a disc penetrometer and a sectilometer, and indirectly using measurements from a cone penetrometer. All three methods correlated closely with spreadability of butter. Several of these tests have limitations. For example, a cup and bob apparatus may experience wall slip, and plate coating is limited by the adhesive properties of the plate surface (Lang and Rha 1981). 18 2.3.3. Vane Method The vane method was developed by Dzuy and Boger (1985) to address some issues from indirect and direct determination methods. The technique measures the stress to initiate flow from a vane immersed in the test material (Steffe 1996). The vane and sample cup dimensions must meet certain specifications to achieve essentially an infinite cup. Steffe summarized dimensional requirements (Figure 2.3): 1.5 g h/d 5 4.0; Z z/d 3 0.5; Z, = 0.0 or Z1/d 3 1.0 if the vane is completely immersed in the sample; D/d 2 2.0 where D is the diameter of the container if circular, or the minimum crossectional dimension if some other shape is used. I4— ”—N MM + Figure 2.3. Cup and vane dimensions for the vane method. +s+|+— =—> 1‘1" +| The vane circumscribes a cylinder in the sample with the boundaries defined by the edges of the vane blades. Material within the cylinder acts as a solid body and the material outside experiences shear (Barnes 1999). The test material yields along a 19 cylindrical surface as the torque on the spindle per unit time is recorded. The total torque (Mo) to overcome yield stress may be described as (Steffe 1996) ”d3 h 1" M0=T 2+3 0'0 (4) 2M, h 1“ therefore, 00: 7rd3 2+3 (5) A typical torque-time curve resembles graph show in Figure 2.4. Torque Time (seconds) Figure 2.4. Torque-time curve at constant angular velocity. The vane geometry confers some distinct advantages. The presence of blades rather than a smooth cylinder eliminates the effects of slip. Submerging the vane in the sample causes negligible effect on the property being measured because the actual yielding surface is located at the outer edge of the blade, near the surface (Breidinger and Steffe 2001). Another advantage of the vane method is that oftentimes, the original food container can serve as the sample cup, providing it meets the vane-vessel dimensions. 20 There are two ways to take yield stress measurements using the vane method: by controlled-shear rate (C 7' ) or controlled-shear stress (Co). For C 7' , the vane is rotated in the sample at a very low (<1.0rpm) constant speed. In the Co, the torque is increased step-wise until the point at which strain increases rapidly. Yoo et al (1995) found that yield stresses taken by the C 7' method were more reproducible than those taken by the Co method. C 7' measurements also proved to be more sensitive for both dynamic and static yield stresses in determining the extent of structure development. The researchers concluded by saying, “The C 7' method appears to be superior to the Co method, Since it is simple, unequivocal, and sensitive to determine the yield stresses and the dimensionless yield numbers of food dispersions.” Several studies have validated the vane method as a reliable technique to measure yield stress. Missaire (1990) used a six-blade vane to measure the yield stress of unstructured and structured food suspensions. This study found that in unstructured suspensions the vane method and Casson’s equation produced yield stresses of comparable magnitude. In structured suspensions, the vane-determined yield stresses were much higher. Rao and Steffe (1997) concluded that the vane method at a controlled shear rate gives more reliable values for yield stress than extrapolation fi'om the Casson equation. A six-bladed vane was used by Cantu-Lozano et al (2000) to compare yield stresses of apple pulp suspensions to produce stresses calculated by the Casson model. Values from each method were nearly identical. The vane method has now become an accepted technique to measure yield stresses of foods. Several recent studies have taken advantage of the technique to characterize food. The vane method was used by Qiu and Rao (1988) to correlate yield 21 stress with pulp content and particle size of apple sauce. Wilson et al (1993) suggested yield stress as a rapid and simple method for quality assurance purposes with molten chocolates. Briggs et al (1996) used the vane method to determine the “scoopability” of ice cream as a direct result of yield stress. Daubert et a1 (1998) measured yield stresses of thirteen spreadable food products. Higher yield stresses were found in products which were more difficult to spread. Kovalenko and Briggs (2002) determined that the vane method is a rapid and inexpensive way for effectively detecting textural differences of soy-based yogurts. 22 2.4. Sensory Instrumental measurements alone are insufficient for the characterization of food products due to the complexity of food systems. Instruments are incapable of detecting the interactions of flavor, texture, and other components of food, which create the effect perceived by humans. Sensory evaluation tests are designed to use humans as instruments capable of yielding analytical data. 2.4.1. Evolution of Sensory Science Sensory science for the food industry began to develop early in the 20th century. In 1936, a researcher by the name of Moir demonstrated the role of color in food acceptability (Moskowitz 1983). He served a dinner for the food group of the Society of Chemistry and Industry where foods were prepared in the conventional manner. Some of the dishes, however, were unusual colors. Several of the scientists indicated feelings of nausea after eating the oddly colored food, even though there was nothing wrong with it. With the coming of World War H more focus was placed on sensory science as the feeding of the military became important. Food was shipped and stored for months at a time, resulting in a decrease in food quality. The morale of soldiers was affected by the quality of the food, so it became necessary to understand methods to assess and improve product acceptability (Moskowitz 1983). Sensory evaluation evolved further as a science through the eventual amalgamation of concepts taken from psychology, such as scaling techniques and psychophysics, with trade practices, such as grading. In the late 1940’s and early 1950’s, techniques began to emerge which allowed food products to be analyzed by sealing methods that were determined by sensory perceptions (Moskowitz 1983). 23 2.4.2. Descriptive Analysis Descriptive analysis is one method of sensory evaluation that has become widely used. Einstein (1991) gives the following definition of descriptive analysis: “Descriptive analysis is the sensory method by which the attributes of a food material or product are identified, described, and quantified using human subjects who have been specifically trained for this purpose.” Moskowitz (1983) highlights one advantage of descriptive analysis over consumer panels by saying that it encourages a broader range of terms to be generated to describe a product; otherwise, consumer panels may limit themselves to buzz words used in advertising. O’Mahony (1991) describes descriptive analysis as “a system where particular sensory characteristics of a food are identified and defined for trained panelists using physical standard stimuli.” These characteristics are measured using one of several scaling techniques. In general, trained judges are used to describe intensity of product attributes, while untrained judges are more useful in hedonic applications (Roberts and Vickers 1994). Training panelists standardizes the concepts used for judgment. Untrained subjects do not have a common language for communication; therefore, the same perception could be given several different definitions by different judges (O’Mahoney 1991, Munoz and Civille 1998). Lawless (1991) says “a term should be used in the same way by an individual upon repeated occasions (intraindividual consistency), be used in the same way by different individuals on a panel (interindividual consistency), and be used by the panel consistently upon subsequent evaluation sessions or experimental replications (consistency of panel mean scores across replicates)” Training a descriptive analysis panel is an effective way to create the desired consistency of terminology. 24 There are several accepted methods of descriptive sensory analysis. Although these methods were deve10ped for flavor and textural attributes, textural studies will be emphasized in this discussion because they relate directly to the objective of the current work. In the 1940’s, Arthur D. Little, Inc. developed the Flavor Profile Method (F PM), which allowed generation of sensory data concerning the use of seasonings in cooked foods (Einstein 1991). The information was applied to product development projects. Over the years, several other methods were born from the FPM, including the Texture Profile Method, Quantitative Descriptive Analysis® (QDA), and Spectrum”. 2.4.2.1. Texture Profile Method The Texture Profile Method (TPM) was developed in the 1960’s under the direction of Dr. Alina Sczcesniak at the General Foods Technical Center (Szczesniak 1963, Szczesniak et al 1963). The method is based on rheological principles and applies descriptive analysis to the sensory evaluation of food texture (Lawless 1991, Munoz et al 1992, Brandt et al 1963). Properties of food texture are placed into one of three categories: mechanical attributes, geometrical attributes, and attributes related to moisture and fat (Munoz et al 1992). Szczesniak et al (1963) developed scales with points anchored by foods that represent a specific intensity of an attribute. Screening procedures select a minimum of ten panelists who are trained for approxirnatelyl 30 hours over a six to seven month period (Murray et a1 2001). Panelists use one of several scaling methods to generate data. The original TPM standard scales involved only oral judgments to assess the entire texture of the product from first bite through mastication, but in later modifications of the TPM, non-oral assessments were also considered. Munoz (1986) explains that 25 manual characteristics are important as an evaluation technique because consumers use such methods to judge certain quality aspects of food products. 2.4.2.2. Quantitative Descriptive Analysis® Quantitative Descriptive Analysis® (QDA) was developed in the 1970’s as a modified version of the F PM that allowed statistical analysis of profile data. Participants in this type of panel are users and likers of the products in question, and reference standards are used only as necessary (Murray et al 2001). The panel leader trains a panel over a period of ten to fifteen hours, and then data are collected by rating attribute intensities on an unstructured line scale. 2.4.2.3. Spectrum"M Method Also in the 1970’s, Gail Vance Civille designed the SpectrumTM Method. This method involves extensive training in which the full “spectrum” of product attributes are considered rather than focusing on just flavor or texture (Murray et al 2001). The SpectrumTM Method also makes use of universal scales made up of points of food reference samples. Non-oral judgments can also be made with the Spectrum Method as well as with TPM. In 1999, Civille participated in a study with Drake and Gerard, which determined that either hand evaluation or mouth evaluation can be used to discriminate cheese texture (Drake et al 1999). 2.4.2.4. Free Choice Profiling Free Choice Profiling (FCP) is a technique that was developed on the 1980’s by Williams and Arnold at the Agricultural and Food Council in UK (Meilgaard et a1 1991). Each panelist creates his or her own list of descriptors and does not use other terms. This method takes less training time than other systems. A study done on a model 26 system using menthol solution found that similar results were found between descriptive analysis and F CP for nasal and oral sensory parameters of the solution (Gwartney and Heymann 1996). 2.4.2.5 Generic Descriptive Analysis Generic descriptive analysis, as the name implies, is a general method of descriptive analysis that takes bits and pieces from the established techniques to best suit the project at hand. A descriptive analysis consists of the following universal factors: Screening to select panelists: The screening may include a discrimination sensory evaluation and a personal interview. It is important that the panelist be able to differentiate between the products or attributes of products, which is determined by a discrimination test. Even more importantly, the subject must have the motivation to participate in the descriptive panel. Piggott (1991) recommends using a panel of about twelve members, Moskowitz (1983) suggests greater than or equal to ten participants, and the QDA® typically uses ten to twelve panelists (Einstein 1991). Meilgaard et al (1991) suggests that the best panelists should be chosen from participants with the best potential rather than those with the highest performance in a selection test. Some tests that can be used for selection are detection/discrimination tests, description tests, and ranking/rating tests. Training: Training sessions are used to generate terms and develop procedures as well as to create standard definitions of descriptors. The most successful vocabularies are the lists chosen by assessors (Piggott 27 1991). Moskowitz (1983) discusses four ways to create a list of descriptors: l)Unprompted Description — the panelist sits alone with the product and creates a personal list of terms that describe the product. 2) Focus Group Interviews - discussions follow a pre-specified guide that covers specific points. 3) Prompted Description in the form of a Kelly Repertory Grid — in this method, the panelist is presented with three items, in pairs, and creates a list of similarities and differences between the combinations. 4) Prompted Description in the form of a checklist — panelists are asked to narrow down a list of terms by categorizing them as critical or non-critical. Assessment: A variety of scales can be used to measure product attributes in descriptive sensory analysis. Meilgaard et al (1991) discusses three commonly used scales: 1) Category scales are limited sets of words or numbers with equal intervals between categories. 2) Line scales use a six- inch or fifteen-centimeter long scale. Panelists make a mark on the line to indicate the intensity of a product attribute. 3) Magnitude estimation (ME) scales are based on free assignment of the first number, and all following numbers are proportional to the first. QDA® uses a six inch line scale with word anchors at each end. SpectrumTM uses line scales with ends labeled “none” and “extreme.” TPA uses any of the three scales described above. 28 2.5. Correlative Studies Descriptive analysis can be very time consuming and expensive to perform, so many industries use correlation studies (between sensory and physical properties) with the hopes of reducing the number of descriptive analysis panels needed. Kokini (1985) explains the usefulness of such methods by explaining “if textural attributes could be related to a single physical parameter or a combination, then these parameters could be used to monitor quality during processing and storage. Measurements of physical properties are quicker to perform and vary much less than sensory data.” Quality control is just one reason that correlations are desirable. Other goals are to predict consumer response, to understand what is being perceived in sensory texture assessment, and to develop improved/optimized instrumental test methods that will lead to a texture testing apparatus that duplicates sensory evaluation (Szczesniak 1987). It is necessary to be careful when making correlations to determine the relationships that actually exist. Any variables can be plugged into a computer program and correlations can be generated, because a computer cannot determine what relationships are logical and valid. Szczesniak (1987) enumerates several factors that affect correlations: 0 Test conditions: Better correlations are achieved when the instrumental test conditions are very similar to the conditions of the sensory test. The type of test depends on the product and the characteristic being measured. 0 Test Material: Homogenization of samples can improve correlations by elimination of variation within the product due to location of sampling. 29 o Sensory Terms: Complex terms, which may describe several parameters with one word will decrease the likelihood of finding meaningfiil correlations. o Sensory Scales: The type of scale used can influence the correlation depending on the range of intensities of the textural parameters. Correlation studies allow food scientists to determine how humans perceive physical and chemical factors of foods. They are mostly used in the areas of quality control and product development. Moskowitz (1983) reminds us that correlations do not say why the relationships exist. He also adds that if correlations are being used as predictors of consruner attitudes, the predictability will be poor because consumers do not expose themselves to stimulus in a controlled environment. Functional relations are determined by regression analysis that yields the expected sensory ratings for a specific measurement (Moskowitz 1983). Correlation studies have been carried out for both fluid and nonfluid food systems and model systems: a Szczesniak et al (1963) presented the TPM standard rating scales for hardness, brittleness, chewiness, gurnminess, adhesiveness, and viscosity. The scales were all correlated with objective measurements from a General Foods Texturometer or a Brookfield Viscometer. All of the scales displayed good correlation with instrumental measurements of texture. 0 Richardson et a1 (1989) performed a study on the perceived texture of thickened systems with dynamic viscosity measurements. The results found that small deformation measurements of dynamic viscosity under oscillatory shear at a 30 single frequency correlated directly with panel scores for perceived thickness of both true solution (fluid materials that do not exhibit yield stress) and weak gels. Also, panel scores for slirniness and stickiness were also directly correlated with dynamic viscosity. Morris et a1 (1984) found that perceived texture, using the descriptors “thickness” and “stickiness,” of random coil polysaccharide solutions correlated well with instrumental measures of maximum viscosity at low shear rates and zero shear viscosity. A wide range of foods was used by Muellenet et al (1998) to develop a model of the relationship between sensory and instrumental texture profile attributes. High correlations were found between the attributes of hardness and springiness; however, no significant correlations were found for the attributes cohesiveness and chewiness. Tang et al (1999) correlated instrumentally measured texture parameters of cooked wheat noodles with data from a trained descriptive analysis panel using a generic descriptive analysis technique. Multivariate analysis approaches indicated that overall results from instrumental measurements and sensory analysis of texture were in good agreement. All of the preceding examples indicate that the concept of correlation of sensory data with instrumental data is a valid method that can yield useful results. Some caution is warranted. Peleg (1983) notes that “the existence of a correlation between sensory and instrumental parameters does not necessarily permit meaningful extrapolation or even interpolation, especially to untested foods of a different textural character.” Sherman 31 (1988) concurs, stating that identical instrumental tests cannot be used for all foods. Adjustments must be made to test conditions based on physiological adjustments made by panelists in sensory assessment. The current work is particularly concerned with correlations of the instrumentally measured parameter of yield stress with sensory characteristics. Several studies have involved yield stress, correlating it with various sensory parameters: Kulkarni and Murthy (1986) correlated yield stress measurements with sensory parameters of butter. In sofi butter, yield stress correlated significantly with hardness and stickiness scores. Yield stress is directly related to the scoopability of frozen ice cream (Briggs et a1 1996) Wendin et al (1997) found that mayonnaise yield stress correlated with thickness and fattiness. Yield stress has been found to be related to spreadability in several food products from processed cheese spread to whipped topping (Daubert 1998), and cream cheese (Breidinger and Steffe 2001). Mortensen and Danmark (1982) also showed a high correlation of yield stress with spreadability of butter. Wendin and Hall (2001) correlated several sensory parameters with yield stress in salad dressing, including thickness and fattiness. 32 3. Materials and Methods 3.1. Rheological Measurements Yield stress measurements were taken using a Brookfield Yield Rheometer called the YR-l (Brookfield Engineering Laboratories, Middleboro, MA). This rotational viscometer was designed specifically to use the vane method to determine the yield stress of test materials. The system includes software that allows the user to custom design the test program. A temperature probe, as well as three vanes, are also included. Along with temperature, the instrument also measures time interval, % torque, delta torque (%), stress, and strain. For the current work, five hot breakfast cereals were measured: Nabisco 2 l/2 Minute Cream of Wheat (COW), Nabisco Instant Original Cream of Wheat (ICOW), Malt-o-Meal Co. Malt-o-Meal (MOM), Quaker Oats Co. Quick One Minute Regular Oats (lMlN), and Quaker Oats Co. Instant Regular Oatmeal (IOAT). Each cereal was prepared according to package instructions using Nanopure® (Barnstead/Thermolyne Corp., Dubuque, IA) filtered water. Table 3.1 indicates the proportions of cereal and water to prepare a batch of four servings. Table 3.1. Proportions of water and cereal for sample preparation. Cereal Water (mL) Dry cereal (g) 2 V2 Minute Cream of Wheat 940 150 Instant Cream of Wheat 640 112 Malt-o-Meal 765 1 30 One Minute Oat 825 190 Instant Oatmeal 600 140 Water was heated in a three quart pot until the temperature of the water reached the boiling point. At this point, the cereal was added while vigorously stirring (approximately 175 strokes per minute) with a wooden spoon to prevent clumping. COW 33 and MOM are stirred for 2 1/2 minutes and lMIN was stirred for 1 minute. For the two instant cereals, ICOW and IOAT, once the water reached the boiling point (100°C), it was removed from the heat and poured into a bowl containing the dry cereal. The resulting instant cereal mix was stirred with a wooden spoon for 1 minute. After stirring for the specified time, the cereal was immediately poured into square Plexiglas containers. These particular containers were designed for several reasons. Plexiglas was used because the material minimizes heat loss as compared to a metal container. This material is also clear, which allows sample observation as the spindle rotates. In addition, square sample containers eliminate slip of the sample against the container wall. Initial tests were conducted in cylindrical cups, but these tests were unsuccessful due to the slip observed. Sample containers met the required dimensions (Steffe 1996) for vessel size according to the dimensions of the vanes used: Z 2/d 3 0.5; Z; = 0.0 or Z [/d _>_ 1.0 if the vane is completely immersed in the sample, where d is the diameter of the vane, 22 is the distance between the bottom of the vane and the bottom of the sample cup, and Z1 is the distance between the top of the vane and the top surface of the sample; D/d 3 2.0 where D is the diameter of the container if circular, or the minimum crossectional dimension if some other shape is used. During testing, a thermocouple was inserted into one container to monitor temperature changes while the vane was submerged in an identical sample. Additional containers holding prepared hot cereal were held at room temperature with a layer of plastic wrap on top of the sample to prevent film formation and minimize evaporative cooling. Samples were held for a series of durations at room temperature before the yield 34 stress was measured. A particular sample was used for only one measurement. Duplicate measurements were taken on identically prepared samples. Nine hold times, covering the period of typical product use, were used: 3, 5, 8, 10, 12, 14, 15, 18, and 20 minutes. These hold times refer to how long the hot sample sat in the measurement container, at room temperature, before the yield stress measurement was initiated. Three different vane sizes (Figure 3.1) and two different container sizes (Figure 3.2) were used to take measurements. The container used depended on the size of the vane necessary based on the torque requirements of the sample. Samples with low torque responses used a larger vane for the measurement, and vice versa. Table 3.2 displays the vane used for each sample. Table 3.2. Spindle size used for measurement of each sample. Spindle 71 (large)* 3min 5min 8min 10min 12min 14min 15min 18min 20min COW X MOM X X X X Spindle 72 (medium)* COW X X X X X X X X MOM X X X X X ICOW X X X X X X X X X lMIN X Spindle 73 (small)* lMIN X X X X X IOAT X X X X X X X X X *Dimensions of spindles are given in Figure 3.1. Spindles 72 and 73 are used in the smaller container, and spindle 71 is used in the larger container. The lMIN sample does not have measurements for 5, 8, and 10 minute hold times When the medium vane was used, the torque was “over-range,” and when the small vane was used, the torque was “under-range.” An over-range reading indicates that the torque response of the sample is too great to be measured by the particular vane. A smaller vane 35 I+~ ”—rl |‘“°"“"| n+— =—>| |¢d+| [ed—5| |4_d__;| Spindle 73 Spindle 72 Spindle 71 h = 2.6cm h = 4.4cm h = 7.0 cm d = 1.0cm d = 2.2cm d = 3.50m Figure 3.1. Dimensions of vanes used with Brookfield Yield Rheometer. I '2... ' ¢ D D N 1. ¢ 1) D H = 12.5cm H = 9.0cm D = 7.0cm D = 5.5cm L = 7.0cm L = 5.5cm Figure 3.2. Dimensions of the sample containers used for measurements. 36 must be used because the blades have less surface area in contact with the sample, therefore the vane encounters less torque. Under-range readings mean that the vane is too small to read the torque response of the sample. By using a larger vane, more surface area is in contact with the sample and a larger torque reading can be obtained. There is an overlap in the ranges of the spindles; however, the cereal may have been changing too quickly to be measured by the instrument between the times of 5 minutes and 10 minutes. The experimental protocol programmed using the Yield Rheometer software was as follows: zero the torque response at 0.1 rpm, wait 5 seconds, rotate the vane at 0.1 rpm, end test when torque reduction is equal to 115%. Hence, the test will only end automatically when one torque value is reduced 15% from the previous torque value. Using the torque reduction of 115% insures that the test will continue until the user chooses to terminate it. The spindle was changed as appropriate. The output was transformed into an Excel worksheet and by plotting torque versus time and observing the data points, a peak torque was extracted. Since many of the torque-time curves do not display a clear peak torque (such as the one illustrated in Figure 2.4) with a well-defined yield stress, the value obtained is referred to as the “apparent yield stress (o.).” Two types of tests resulted from the Brookfield YR-l: passed and terminated. The tests that the instrument defined as “passed” met the criteria of a 15% torque reduction between subsequent data points. The maximum torque value from the passed tests was used as the peak torque. The terminated tests did not demonstrate a torque reduction of that magnitude. Although the torque values leveled out or decreased after reaching a peak torque, they did not decrease steeply enough to register a yield stress on 37 the instrument. The peak torque on the terminated tests was defined as the point at which delta torque equaled zero, or the equilibrium stress. 38 3.2. Descriptive Analysis Measurements 3.2.1. Screening Panel Screening panels are used to find panelists who are suitable to take part in a more extensive trained panel. An appropriate panelist is one who is interested in participating, available for 80% of the time, is prompt, has general good health, is articulate, and is not averse to the test product (Meilgaard et al 1991). The objective of this test was two-fold: l) to screen for individuals who were able to discriminate between large differences in thickness in hot cereal and 2) to screen for individuals who were willing to participate in a trained panel based on this exercise. One hundred one subjects participated in this test. Volunteers were students, faculty, and staff of Michigan State University. Subjects were compensated for participation with ice cream coupons. The two samples used in the study were one pouch of Instant Original Cream of Wheat prepared with l60mL Nanopure water, and one pouch of Instant Original Cream of Wheat prepared with l90mL Nanopure water. Water was boiled and added to Styrofoam bowls each containing one pouch of Instant Original Cream of Wheat. One bowl of each sample was served immediately to the panelist. The subjects participated in a paired comparison test in which they were asked to signify which sample was thicker. The test question was followed by an option to indicate interest and availability in participating in a trained panel based on the exercise. Results from this panel showed that 91% of the participants were able to detect the difference in thickness between two samples. Initially, seventeen of these panelists agreed to take part in the trained descriptive analysis panel. 39 3.2.2. Trained Descriptive Analysis Panel Descriptive analysis panel training was divided into three sessions: Session 1, Session II, and Test Sessions. Sessions I and H were each offered on two different days to accommodate panelists and encourage participation. Each of the days lasted approximately one hour. After completing Session II, panelists were asked to Sign up for three test sessions. The samples used in the test sessions were Nabisco 2 '/2 Minute Cream of Wheat (COW), Nabisco Instant Original Cream of Wheat (ICOW), Malt-o- Meal Co. Malt-o-Meal (MOM), Quaker Oats Co. Quick One Minute Oats (lMIN), and Quaker Oats Co. Instant Regular Oatmeal (IOAT). All samples were prepared according to package instructions and then held at room temperature for five minutes before being served. For Session I, panelists gathered as a group and were presented with one sample each ICOW and IOAT. Five people attended the first day and twelve people attended the second day. They were asked to write down a list of words that described the cereal samples. Table 3.3 lists the complete catalog of terms generated between the two days of Session 1. Each descriptor was discussed within the group and the list of words was narrowed down to those that could logically be correlated with instrumental measures of textural properties. The remaining words were discussed until a definition and a test method was determined for each word. Table 3.4 shows the final list of words from day one and day two of Session I. 40 Table 3.3. Comprehensive list of descriptors for Nabisco Instant Original Cream of Wheat and Quaker Oats Instant Regular Oatmeal. ICOW [OAT Gritty Lumpy Constant consistency Sticky Sticks to spoon Watery Grainy Thick Mounds Viscous Little or fine pieces Slimy Off white with dark Little lumps in thick brown flecks semi-viscous fluid Viscous Grey brown with flecks of dark brown No separation (of Mounds on a spoon water and particles) Thick Viscous strands when dropped from spoon Pasty Fluid Unpleasant odor Coarse Smooth Gooey Flavorless Not uniform (water separates out) Creamy Mild flavor Heavy Clumpy Bland Runny Salty Mushy Fluffy Chunky Do/ Table 3.4. Descriptor with definitions and testing methods from Days 1 and 2 of Session I. Day 1 Day2 Descriptor: Thickness Definition: Resistance to stirring as felt in the hand and arm. Test Method: Hold spoon vertical and move in 5 circles. Descriptor: Stirrability Definition: Amount of force needed to stir cereal at constant speed. Test Method: Hold cup steady, stir around with spoon five times. Descriptor: Viscous Definition: Resistance to flow. Test Method: Lift heaping spoonful, observe degree of mounding on spoon. Descriptor: Stickiness to Self Definition: Degree of adhesion to itself. Test Method: Lift spoon through cereal perpendicularly and observe degree of strands/ropiness. Descriptor: Slimy Definition: Ropelike appearance. Test Method: Tap surface with spoon, lift two centimeters, observe degree of ropiness. Descriptor: Stickiness to Spoon Definition: Degree to which sample sticks to another object. Test Method: Lift spoon vertically out of the cereal and observe the amount of cereal retained on the spoon. Descriptor: Stickiness to Spoon Definition: Degree to which sample sticks to spoon. Test Method: Hold spoon perpendicularly to cup, hold five seconds. Observe amount retained on spoon. Descriptor: Stickiness to Self Definition: Degree of adhesion to itself. Test Method: Lift spoonful of cereal, hold at 45 degrE file, and observe uniformity of flow. 41 Four panelists attended the first day of Session H, and ten attended the second day. The lists in Table 3.4 were merged to create a ballot for the test sessions. The ballot (Figure 3.3) was presented to the panelists during Session II. The line scale used was 3 15cm unstructured line scale. The reference mark is 7.5cm from either end and the reference product was 2 '/2 Minute Cream of Wheat (COW) for rating wheat-based cereals and Quick One Minute Oats (lMIN) for rating oat-based cereal. Panelists were given the reference sample, which was COW prepared in the conventional manner. The test methods were reviewed and practiced to ensure consistency among panelists. Panelists were then given a set of four samples of COW, each of which had increasing amounts of dry cereal with equal amounts of water to create samples which represented the range of yield stresses covered by the actual test samples. Panelists individually marked the intensity of each parameter for each sample on the appropriate line scale on the ballot. Marks were discussed afterwards to ensure that all panelists agreed on the measurement of the intensity. Panelists signed up for three test sessions. They were instructed that each test must be at least one day apart from the prior test. This was done in an effort to prevent panelists from making marks on the line scale simply by remembering where they placed a particular sample in a previous test session. The twelve panelists who completed all sessions were comprised of faculty and students at Michigan State University. Two males and ten females participated. Subjects were compensated with ice cream coupons, candy bars, and chips. Panelists came in singles or in pairs and the tests lasted forty to forty- five minutes. 42 Thickness (StirrabilitY) Definition: Resistance to stirring at a constant speed. Test Method: Hold cup steady with one hand, hold spoon vertical in sample with other hand. Stir spoon in a circle five times at a moderate speed. Evaluate thickness. Indicate thickness relative to the reference on the line scale: I l Low R High Stickiness to spoon Definition: Degree to which sample sticks to spoon. Test Method: Immerse spoon in sample. Lift spoon straight up out of cereal. Hold steady for five seconds. Observe amount of sample retained on spoon. Indicate stickiness to spoon relative to the reference on the line scale: I 1 Low R ' High Stickiness to self Definition: Degree of adhesion of sample to itself. Test Method: Lift heaping spoonful of sample. Tilt spoon towards yourself at a 45 degree angle, allowing sample to flow off spoon. Observe tmiformity of flow. Indicate stickiness to self relative to reference on the line scale: l I Low R High Viscosity Definition: Resistance to flow. Test Method: Lift one heaping spoonful of sample. Drop onto lid of sample container. Observe degree to which sample spreads over the surface of the container. Indicate viscosity relative to reference on the line scale: I 1 Low R High Figure 3.3. Ballot presented to panelists in Session II. The questionnaire (Figure 3.4) was designed using SIMS2000 Version 3.3, and the test was performed in a controlled environment. The first question, which asked for intensity of color, is a “dumping” question. This question allows the panelist to “dump” 43 all other differences noted into a separate category so they could focus their attention on the parameters of interest (Lawless 1991). Figure 3.4. Questionnaire for test sessions. To begin, slide Ready side of card under the window. You will be given 6 different sets of samples one at a time. Each set will consist of 1 reference sample labeled R and l coded test sample. After completion of each set, slide the Finished side of the card under the window and you will be given the next set of samples. Remove lid from reference sample and coded sample. Visually observe overall color of sample as it is presented, without using a spoon. Indicate intensity of color relative to the reference with LOW being very pale and HIGH being very dark. Captions: Intensity of Color + ----------------------------- + ----------------------------- + LOW R HIGH Hold cup steady with one hand, hold spoon vertical in sample with other hand. Stir spoon in a circle 5 times at a moderate speed. Do not scrape sides or bottom of container with the spoon. Indicate thickness relative to the reference with LOW being very little resistance to stirring, HIGH being very large resistance to stirring. Captions: Thickness (stirrability) + ---------------------------- + ----------------------------- + LOW R HIGH Immerse spoon in sample. Lift spoon straight up out of sample. Hold vertically for 5 seconds. Observe amount of sample retained on spoon. Indicate stickiness to spoon relative to the reference with LOW being very small amount retained on spoon and HIGH being very large amount retained on spoon. 44 Stickiness to Spoon Lift heaping spoonful of sample. Tilt spoon towards yourself at a 45 degree angle, allowing sample to flow off spoon. Observe uniformity of flow. Indicate stickiness to self relative to reference with LOW being particles do not stick to each other and HIGH being particles stick strongly to each other. Stickiness to self Lift one heaping spoonful of sample. Drop onto lid of sample container. Observe degree to which sample spreads over the surface of the container. Indicate viscosity relative to reference with LOW being sample spreads widely and HIGH being sample keeps its shape. Viscosity + ---------------------------- + ------------------------------ + LOW R HIGH Figure 3.4 (cont’d). Test sessions were designated Repl , Rep2, and Rep3 in reference to each of the three replications. In each replication, one sample of each cereal was presented sequentially, each time with a freshly prepared reference. Samples were coded using a random numbers table (Table l, Meilgaard 1991). The order of the samples was randomized; however, the order was the same for all panelists within a replication. 45 Samples for each replication are identified in the order presented for each replication in Table 3.5. Reference samples, sample code numbers, and sample identifications for Repl, Rep2, and Rep3. Table 3.5. Repl Reference (R) Sample Sample Code COW COW 339 lMIN IOAT 413 COW MOM 572 COW ICOW 574 lMIN lMIN 973 Repz Reference (R) Sample Sample Code lMIN [CAT 576 COW ICOW 311 lMIN lMIN 975 COW COW 776 COW MOM 583 Repl Reference (R) Sample Sample Code COW ICOW 286 INHN lMIN 235 COW COW 477 COW MOM 782 lMIN IOAT 713 46 4. Results and Discussion 4.1. Rheological Behavior There are two factors present that are responsible for causing the resistance to flow in prepared cereals: 1) thermal effects, and 2) chemical effects. The effects of these phenomena were evaluated through observations of changes in temperature and apparent yield stress values over time. The results for the rheological portion of this work were analyzed graphically and digitally. Temperature at peak torque was plotted versus time (Figures 4.1 and 4.2). Temperatures decreased over time in a similar manner for all five samples. The instant cereals (ICOW and IOAT) both have regression curves that are shifted down from their cook-on-stove counterparts. This is because the instant cereals are not boiled during preparation. The thermal conditions of these tests mimic the effects seen by a consumer; however, there are slight differences. For example, the temperature for the test was measured over twenty minutes, while a consumer may have finished his breakfast before that time lapse. The cooling is caused in part by evaporative heat loss from the surface of the hot cereal. In the test, plastic wrap is used to minimize this loss, and the surface area available for evaporation is smaller than that of a bowl used by the consumer. Additionally, the sample container used in the test is Plexiglas, while a bowl is normally made out of ceramic or glass. These materials are normally better conductors than Plexiglas, causing an increase in heat loss and probably a corresponding increase in yield stress. 47 120.0 —-| o .0 o MOM y = -1.607x + 10227 R2 = 0.8579 COW 8 O 3 E 80.0 1 \ .2 . " = -1.2307x + 94.758 35 R2 = 0.9655 IL 60 0 g 1C0 y = -O.7744x + 79.178 ‘3 40.0 R2 - 0 9909 a . E 20.0 0.0 - ~ . , 0 5 10 15 20 25 Hold Time (mln) Figure 4.1. Temperature at peak torque versus hold time for wheat-based cereals. 100.0 > 90.0 4 . y = -0.9086x + 91.076 3 1 R2 = 0.9846 .. 80.0 ~ 1MIN g . g 70.0 '_; 60.0 4 fl- IOAT 3 y = -0.8695x + 77.996 ‘3 50:0 ' R2 = 0.9928 0 p 40.0 . 3 1! 30.0 , 8 g 20.0 - .— 10.0 . 0.0 ~ . - ,, . l ,, W 0 5 10 15 20 25 Hold Time (min) Figure 4.2. Temperature at peak torque versus hold time for oat-based cereals. 48 An apparent yield stress value was found for each cereal at each hold time. Values from multiple replications (55 n 510) were averaged and a scatterplot of apparent yield stress versus hold time was generated (Figures 4.3 and 4.4). Table 4.1 lists each cereal and its respective average apparent yield stress value for each particular hold time. The trend is for the apparent yield stress to increase as hold time increases. For wheat based cereals, ICOW has the largest average apparent yield stress values and MOM has the smallest. For oat based cereals, IOAT has higher average apparent yield stress values than the lMIN product. Table 4.1. AveraEapparent yield stress values How Time Average 0a :I: STDev How Time Average 021 :l: STDev 3min cowa 21.7 i 2.5 14min cow 40.2 i 3.1 1cowb 51.1 i 7.3 1cow 63.1 i 14.0 MOMc 19.4 i 1.1 MOM 32.6 i 3.2 lMINd 108.3 1 15.5 lMIN 514.2 : 52.9 IOAT‘ 476.2 1 64.8 IOAT 883.9 1 138.4 5min cow 30.3 i 3.7 15min cow 42.4 i 5.0 1cow 52.9 i 8.5 1cow 64.4 _+_ 8.7 MOM 20.5 _+_ 1.5 MOM 30.4 p: 3.4 lMIN lMIN 518.9: 134.2 IOAT 613.4 : 129.8 IOAT 809.1 : 45.5 8min cow 30.4 1 3.8 18min cow 47.3 i 6.1 ICOW 57.1 i 8.2 Icow 63.7 i 13.0 MOM 23.0 i 0.8 MOM 35.8 i 1.4 lMIN lMIN 619.8:702 IOAT 714.0 1 116.0 IOAT 976.3 : 176.4 10min cow 31.0 i 3.4 20min cow 48.7 i 3.0 ICOW 59.0 i 9.0 ICOW 68.1 i 10.9 MOM 24.2 i 1.8 MOM 38.8 i 2.6 lMIN lMIN 658.7 .t 98.6 IOAT 742.3 1 121.1 IOAT 984.0 : 180.9 12min COW 40.2 i 8.6 ICOW 61.9 i 7.3 MOM 30.1 i 1.5 lMIN 513.9 1 35.5 IOAT 776.1 t. 73.4 ‘2 1/2 Minute Cream of Wheat t’Instant Cream of Wheat cMalt-o-Meal dOne Minute Oats cInstant Oatmeal 49 Apparent Yield Stress (Pa) Apparent Yield Stress (Pa) 80.00 . y = 0.9558x + 48.988 2 ICOW 70.00 - R =0.9527 60.00 - 50.00 . Y = 1.52295)! 4’ 19.045 cow R =0.9413 4 . - 000 1 MOM 30.00 ‘ y=1.1758x+14.586 R2=0.9589 20.00 < 10.00 . 0.00 - - . . 0 5 10 15 20 Hold Time (min) Figure 4.3. Average apparent yield stress versus hold time for wheat-based cereals 1200.00 y = 27.981x + 448.57 IOAT 1000.00 R2 = 0.9432 800.00 . 1mm 600.00 A, y = 32.194x + 48.976 2 - 400.00 3 R -0.9524 200.00 1 0.00 - . 0 5 10 15 20 Hold Time (min) Figure 4.4. Average apparent yield stress versus hold time for oat-based cereals. 50 25 25 These measurements cannot be compared with yield stress values available in the literature because most published research focuses on pure starch solutions rather than food systems. Physical and chemical modifications or the addition of foreign materials should lead to a very large diversity (Doublier 1981), as found here. The term “apparent yield stress” is being used here because no definitive yield stress (requiring a clearly defined peak torque) could be identified in all hot cereal tests. In the literature, researchers have been able to find yield stresses in starch dispersions (Lang and Rha 1981, Evans and Haisman 1980, Bagley and Christianson 1983), however these tests have not involved actual food products and the starches have been cooked for times of fifteen minutes and longer. The tests reported in the current study are based on consruner products, which are cooked for a maximum of two and a half minutes. The results clearly indicate that the apparent yield stress measurement can be used to characterize and differentiate the flow behaviors of the hot cereals. Instant Cream of Wheat has a much higher intercept than either MOM or COW (Figure 4.3), meaning that ICOW has the highest torque response of the three wheat cereals. The slopes of the trendlines are positive, meaning that all samples thicken over time. ICOW increased at a slower rate (less slope) than either MOM or COW because it is designed to hydrate very rapidly and has limited ability to thicken later. Over the total range of hold times, apparent yield stress measurements for ICOW only increases in value by a third as compared to COW, which increases by 130% and MOM which doubles from three to twenty minutes. The R2 values signify the goodness-of-fit of the trendline in regards to the individual data points. All the R2 values are close to 1 indicating that all trendlines fit the 51 data points very well. It should be noted that ICOW has a larger standard deviation than the other wheat-based cereals. This may be because ICOW tended to clump more than the others; and if the vane contacted a clump, a sharp increase in measured torque may have occurred. A higher torque response for ICOW over either of the cook-on-stove varieties is expected for two reasons: 1) ICOW is an instant cereal, and as such is designed to hydrate and thicken very rapidly through modification (pregelatinization) of the starch, and 2) ICOW contains guar gum which is often used in food products as a thickening agent. Instant Oatmeal has a much higher intercept than lMIN (Figure 4.4), meaning that it is much thicker. The slopes of each line are similar, although lMIN is slightly steeper. Both slopes are positive, indicating that the oatmeals both thicken over time. The R2 values are near 1, signifying that the trendlines fit the data points very well. The same trend is seen with the oat cereals as with the wheat cereals. The instant oatmeal (IOAT) is designed to hydrate and thicken rapidly through chemical and physical modifications as well as the addition of guar gum. This could explain why, as time increases, although the apparent yield stress measurements for IOAT are higher than for lMIN, the values for IOAT double while for lMIN the apparent yield stress increases by five times. Stanley (1987) states that both the human reaction to, and the mechanical properties of, food texture are the result of forces acting in an underlying organized structure. Chemical components interact to form a microstructure that contributes to the physical and sensory properties of the material. With a food product, as opposed to a 52 pure solution, the interactions of components becomes much more complex and difficult to describe. Chemical changes in the constituents of the cereal when it is prepared cause an increase in apparent yield stress. Over time, two processes of particular importance are starch hydration and gelatinization. Both wheat and oat starches gelatinize well below the boiling point of water (58-64°C and 53-59°C, respectively). Daniels (1974) stipulates that hydration of the cereals for the final prepared breakfast product must take place with water at a temperature higher than 79.4°C or the starch will not gelatinize. Therefore, it is safe to say that the temperature reached by exposing the cereals to water at its boiling point (100°C) is adequate to initiate gel-forming behavior in the starch molecules. As the cereal is immersed in boiling water, the starch granules imbibe water and swell. As the granules occupy more volume, resistance to flow is expected to increase. Also as the granules absorb water, the amylose molecules solubilize and leach out of the granules. As the material ages and cools, the amylose molecules interact, contributing to structure within the cereal. This structure development also leads to an increase in apparent yield stress. How much of the increase in resistance is due to cooling, and how much is due to hydration is unknown for hot cereals. 53 4.2. Descriptive Analysis Panel training was divided into two sessions, each offered on two days, to accommodate panelists and encourage participation. Dividing a panel in half for separate but identical training does not decrease the value of the panel. Heymann (1994) found that consistent results can be found across two independently trained panels. Therefore, data are reliable from two different well-trained panels as long as the same sample preparations are used. The same result was found here. Each Session I group developed a nearly identical ballot (Table 3.4). During both days of Session H the participants scored practice samples on the line scale in the same area around the references. The panel training was limited to only two hours, while traditional descriptive analysis training takes place over 10 hours or more. The limited training does not qualify the panelists as experts, but should not be a detriment to the test because it has been found that a well instructed panel can perform almost as well as an expert panel (Moskowitz 1996) The sensory results for the trained descriptive analysis panel were analyzed with SAS, using proc glm to perform analysis of variance. Typically proc glm is used for unbalanced designs; however, it was used as a tool in this research because it is more conservative than proc anova for balanced designs. The panelists were presented with a 150m line scale, with a reference sample indicated by a mark at the 7.5cm point. Scores for all cereals were assigned relative to this reference. Each of the twelve panelists performed the sensory evaluation in triplicate, generating thirty-six scores for each sensory parameter of each cereal. Results of the scoring are displayed in Tables 4.2-4.5 and Figures 4.5-4.8. 54 Table 4.2. Results of descriptive analysis panel scores for wheat cereals and oat cereals for the parameter “thickness (stirrabilgy)? Thickness (stirrability) Definition: Resistance to stirring at a constant speed. Test Method: Hold cup steady with one hand, hold spoon vertical in sample with other hand. Stir spoon in a circle five times at a moderate speed. Average Standard Deviation Variation cow 7.1 1.3 1.7 ICOW 9.2 2.4 5.7 MOM 11.6 1.9 3.7 lMIN 7.0 2.1 4.4 IOAT 7.9 2.4 5.5 (a) cow ICOW MOM | l | 1 LOW I R I I HIGH (b) lMIN IOAT l I | LOW I it I HIGH Figure 4.5. Scalar illustration of results of descriptive analysis scores for wheat cereals (a) and oat cereals (b) for the sensory parameter “thickness (stirrability).” 55 Table 4.3. Results of descriptive analysis panel scores for wheat cereals and oat cereals for the parameter “stickiness to spoon.” Stickiness to spoon Definition: Degree to which sample sticks to spoon. Test Method: Immerse spoon in sample. Lift spoon straight up out of cereal. Hold steady for five seconds. Observe amount of sample retained on spoon. Average Standard Deviation Variation COW 6.7 1.7 2.9 ICOW 8.0 2.7 7.0 MOM 1 1 .5 1 .9 3.4 lMIN 6.9 1.6 2.6 IOAT 7.4 1.9 3.8 COW ICOW MOM (a) l l l 1 LOW I k I I HIGH 0,) 1 MIN IOAT l I LOW I R HIGH Figure 4.6. Scalar illustration of results of descriptive analysis scores for wheat cereals (a) and oat cereals (b) for the sensory parameter “stickiness to spoon.” 56 Table 4.4. Results of descriptive analysis panel scores for wheat cereals for the parameter “stickiness to self.” Stickiness to self Definition: Degree of adhesion of sample to itself. Test Method: Lift heaping spoonful of sample. Tilt spoon towards yourself at a 45-degree angle, allowing sample to flow off spoon. Observe uniformity of flow. Average Standard Deviation Variation COW 6.7 2.0 4.1 ICOW 9.3 3.1 2.6 MOM 10.8 2.6 6.6 lMIN 7.2 1.6 2.4 IOAT 7.1 2.4 5.7 (a) cow ICOW MOM | l I 1 Low I R I I HIGH lMIN IOAT 0’) I | LOW I R HIGH Figure 4.7. Scalar illustration of results of descriptive analysis scores for wheat cereals (a) and oat cereals (b) for the sensory parameter “stickiness to self.” 57 Table 4.5. Results of descriptive analysis panel scores for wheat cereals for thgparameter “viscosity.” Viscosity Definition: Resistance to flow. Test Method: Lift heaping spoonful of sample. Drop onto lid of sample container. Observe degree to which sample spreads over the surface of the container. Average Standard Deviation Variation cow 6.5 1.8 3.2 1cow 10.1 2.6 6.5 MOM 10.2 2.3 5.5 lMIN 7.3 1.5 2.4 IOAT 5.6 2.4 5.8 (a) cow ICOW MOM I I I Low I R I HIGH IOAT lMIN 0” I I I LOW I I R HIGH Figure 4.8. Scalar illustration of results of descriptive analysis scores for wheat cereals (a) and oat cereals (b) for the sensory parameter “viscosity.” 58 Analysis of variance was performed to indicate where the panelists found significant differences between cereals. The ANOVA model used took into account the variations due to differences in scores between subjects. This helps eliminate the effect caused by panelists using different parts of the scale. For example, Panelist A may use a limited range of the line scale focused around the reference point while Panelist B may make use of the entire line. Without accounting for this factor, the results would be skewed and we would be unable to determine if the difference in cereals was due to the samples themselves being different or to the variation in how the panelists used the line scales. Some degree of variation is inevitable since each unit consists of one assessor. One source of variation is the order in which the samples are presented (Piggott, et al 1998) Analysis of variance results for thickness for wheat cereals yields a p of <0.0001 indicating that there is a significant difference between at least two of the cereals (Appendix Table 7.14). Some of the variation observed is due to the subject (p = 0.0010) (Appendix Table 7.15). The results of the statistical analysis also indicates the least significant means, which determine where the difference between the cereals exists due to treatment. For wheat cereals and the thickness parameter, at 01 = 0.01, there is a significant difference in scores between COW, ICOW, and MOM (Appendix Table 7.16). The ANOVA results for stickiness to spoon for wheat cereals gives a p of <0.0001 indicating that there is a significant difference between at least two of the cereals (Appendix Table 7.17). The variation observed is not due to the subject (p = 0.0531) (Appendix Table 7.18). For wheat cereals and the stickiness to spoon parameter, at 01 = 0.01, there is a significant difference in scores between COW, and MOM and also 59 between ICOW and MOM. At 01 = 0.05, there is a significant difference between COW and ICOW (Appendix Table 7.19). The AN OVA result for stickiness to self for wheat cereals indicates that there is a significant difference between at least two of the cereals (p of < 0.0001) (Appendix Table 7.20). The variation observed is not due to the subject (p = 0.2277) (Appendix Table 7.21). For wheat cereals and the stickiness to self parameter, at 01 = 0.01, there is a significant difference in scores between COW, and ICOW and between COW and MOM. At a = 0.05 there is a significant difference between ICOW and MOM (Appendix Table 7.22). For viscosity for wheat cereals, the ANOVA results give a p of <0.0001 indicating that there is a significant difference between at least two of the cereals (Appendix Table 7.23). The variation observed is not due to the subject (p = 0.3313) (Appendix Table 7.24). For wheat cereals and the viscosity parameter, at 01 = 0.01 , there is a significant difference in scores between COW and ICOW and between COW and MOM. There is no significant difference between ICOW and MOM (p = 0.9576) (Appendix Table 7.25). For the wheat-based cereals, the trained panel found that for thickness, stickiness to spoon, and stickiness to self, there is a significant difference between COW, ICOW, and MOM. For all three parameters, COW had the lowest intensity and MOM had the highest intensity of the particular attribute. The last parameter, viscosity, was not able to distinguish between ICOW and MOM; however, this attribute still followed the trend of COW having the least intensity of the three samples. 60 Analysis of variance was also used on the oat cereals. Even though there were only two samples, it was necessary to use ANOVA to determine the variation in scores due to the subject. For thickness for oat cereals, a p of0.3153 indicates that there is not a significant difference between the cereals (Appendix Table 7.26). The AN OVA results for stickiness to spoon for oat cereals give a p of 0.2400 indicates that there is not a significant difference between the cereals (Appendix Table 7.27). The ANOVA results for stickiness to self are a p of 0.4319 indicating that there is no significant difference between the two cereals (Appendix Table 7.28). For viscosity, the p of 0.0348 indicates that there is a significant difference between cereals at 01 = 0.05 (Appendix Table 7.29). A p of 0.3982 for the subject and 0.0007 for the treatment indicates that all variation between cereals is due to the samples, not the variation among subjects (Appendix Table 7.30). Panelists had a difficult time distinguishing between lMIN and IOAT. Viscosity was the only successful method for determining differences. For this attribute, IOAT was rated at a lower intensity than lMIN. Instant cereals are designed to hydrate and thicken more rapidly than the cook-on -stove cereals. To this end, the starch is pregelatinized by way of steam treatments and the cereal flakes are rolled thin to allow them to absorb water more quickly. Thickeners are also added: both ICOW and IOAT contain guar gum, which hydrates rapidly in hot water. Therefore, it would be expected for the sensory panel and the yield stress values to indicate that the instant cereals are higher in all categories than the cook-on-stove cereals, as is the case for the wheat cereals. During the time period during which the sensory test occurred, the instant cereals may reach higher values because they thicken 61 more rapidly. Cook-on-stove varieties may reach the same thickness later as the native starch sets up unaided. 62 4.3. Correlation of Rheological and Sensory Parameters For the best correlations between instrumental and sensory measurements, the sensory evaluation methods should mimic the instrumental tests. Non-oral sensory methods were emphasized in this test to imitate the use of the vane, which rotates in the sample. Samples were prepared and allowed to rest for five minutes before being given to the panelists. This served two purposes: 1) Consumers who eat hot cereal usually wait for it to cool down before consumption. Szczesniak (1987) reminds us that tests should be designed with the consumer in mind, so we tried to mimic the conditions that the consmner would experience; 2) The materials and methods for the rheological measurements stipulate a hold time for a minimum of three minutes before the first data are collected to allow the structure of the sample to develop. Since the R2 values were near 1 for all the apparent yield stress versus hold time curves, it was possible to use the lines themselves to generate the data points for use in the correlation. The time used in the equation was selected as the average age of the sample that the panelists evaluated. Panelists evaluated the cereal when the samples were between five and ten minutes old. The average time was found and used in the line equation from Figures 4.3 and 4.4 to find the estimated apparent yield stress value at that particular time. This is approximately the apparent yield stress value that the panelist was evaluating. If a different time had been picked, the results would have been similar. Table 4.6 displays the times and equations used for each cereal as well as the resulting apparent yield stress values that were calculated from the regression analysis. The results for the correlation analysis for the sensory data and the apparent yield stress values were analyzed using proc glm in SAS to perform regression analysis (Table 4.7). 63 It should be kept in mind that correlations do not reflect a cause-and-effect relationship, merely that two variables change in unison (Szczesniak 1987). All correlation coefficients obtained in this test are 0.40 and below meaning that there is no correlation between instrumental values and the sensory scores. Figures 4.9-4.16 display sensory parameter versus apparent yield stress value. Table 4.6. Equations and cereal ages used to find apparent yield stress value evaluated by sensory panel. Equation t = age (minutes) 5,, 2 1/2 Minute Cream of Wheat o.=1.5296t + 19.045 8.5 31.6 Instant Cream of Wheat 0,, =0.9558t + 48.988 8 57.5 Malt-o-Meal o, =1.1758t + 14.586 7.75 23.5 One Minute Oats o, =32.194t + 48.976 8 306.5 Instant Oatmeal o, =27.981t + 448.57 8.5 686.4 Table 4.7. Correlation coefficient for regression analysis for wheat- and oat-based cereals for four sensory parameters. Thickness Stickiness to Stickiness to Viscosity (Stirrability) spoon self Correlation coefficient of wheat-based 0.20 0.3 0.04 0.2 cereals Correlation coefi'rcient of oat- 0.20 0.1 0.01 0.4 based cereals 64 16 14 < 12 ~ 5 10 * o ‘0 MOM Z: 8 « § 0 i . A ' COW ICOW 4 2 . 0 10 20 30 40 50 60 70 Apparent Yield Stress (Pa) Figure 4.9. Average sensory score 1 one standard deviation for thickness of wheat-based cereals versus apparent yield stress measurements. 14 12 - 10 8 MOM # COW ICOW Sensory Score 0 10 20 30 40 50 60 70 Apparent Yleld Stress (Pa) Figure 4.10. Average sensory score : one standard deviation for stickiness to spoon 0f wheat-based cereals versus apparent yield stress measurements. 65 12 ‘ 'I a 10 A 8 8 I * ”é, » MOM O 6- .L 2 , ICOW '3 4 cow 2, 0 10 20 30 40 50 60 70 Apparent Yield Stress (Pa) Figure 4.11. Sensory scores for stickiness to self for wheat-based cereals versus apparent yield stress measurements. 14- 12- 101 l I MOM Ic'OW Sensory Scores 0 10 20 30 40 50 60 70 Apparent Yield Stress (Pa) Figure 4.12. Average sensory score : one standard deviation for viscosity for wheat-based cereals versus apparent yield stress measurements. 66 121 10 T g 8 + I. o 0 m z. 6 o ‘ j .t g - IOAT 8 4 . 1MIN 2 . I 0 100 200 300 400 500 600 700 800 Apparent Yield Stress (Pa) Figure 4.13. Average sensory score 1 one standard deviation for thickness for oat-based cereals versus apparent yield stress measurements. 10 9 . I 8 . I. 7 '. g | 6 .. 8 3 .. g 5 1MIN IOAT 2 43 8 3 i 2 ,g 1 . 0 . 0 100 200 300 400 500 600 700 800 Apparent Yield Stress (Pa) Figure 4.14. Average sensory score 1 one standard deviation for stickiness to spoon for oat-based cereals versus apparent yield stress measurements. 67 10’ 9 .. 8 7 .1- § 6 — «3 9 1qu E 5 4 g 4 . IOAT 8 3 I 2 i I 1 . o " 1 - T " T " ' " ' _ Tfi‘TT—_ ' ' ' T "_ "I 0 100 200 300 400 500 600 700 800 Apparent Yield Stress (Pa) Figure 4.15. Average sensory score : one standard deviation for stickiness to self for oat-based cereals versus apparent yield stress measurements. 10 9 8 .. 7 e | a 5 1MIN O 2 4 ° I (D 3 . IOAT 2 1 0- -~ — ._ - , - 1 0 100 200 300 400 500 600 700 800 Apparent Yield Stress (Pa) Figure 4.16. Average sensory score 1: one standard deviation for viscosity for oat-based cereals versus apparent yield stress measurements. 68 There are explanations for why the sensory scores did not correlate with the apparent yield stress. Sensory terms are very complex and often reflect several sensations, contributing to the difficulty in finding meaningful correlations (Szczesniak 1987). It may be necessary to use more than one physical measurement to describe a particular sensory perception, rather than just relying on an apparent yield stress. The conclusion to be drawn is that the sensory parameters chosen do not closely represent the same characteristic that the vane method measures. Perhaps oral judgments could be explored to correlate with apparent yield stress measurements. The lack of correlation does not mean that instrumental analysis is a poor method for characterizing hot cereals. On the contrary, there is a clear difference in hot cereal flow behavior based on the rheological method. Hence, the vane method to calculate an apparent yield stress can be used effectively for quality control and product development applications. For example, if a company is interested in introducing a new product as a line extension, and they would like to retain the same flow properties in the new product that are present in the existing product, this test allows the characteristics of the new product to be objectively evaluated as the ingredients are modified. In this way, the behavior of the new product can be monitored to insure that it has the same flow behavior as the traditional product. It would be possible to implement a quality control test based on the vane method. Although the correlation results were not effective for either wheat-based or oat-based cereals, the rheological data is still valuable. The hold time recommended for evaluation would be between five and ten minutes for wheat cereals, and at twelve minutes for oat 69 cereals, as these times are short enough to be convenient in industry and also produce consistent results. 70 5. Future Research Attempting to characterize the flow behavior of these hot cereals gave rise to several issues that were not addressed in this work. A couple of recommendations for future work can be made: 1. Changes in both hydration and temperature clearly affect the thickening behavior of the hot cereals. As a result of observations during this research, the hypothesis is that time allowed for hydration has the more significant impact; however, the effects of each factor could be separated for a clearer picture. 2. The sensory parameters used in this work did not correlate with the instrumental measurements. Additional sensory parameters should be explored in the form of both oral and non-oral judgments. 71 6. Summary Three wheat-based hot cereals (2 V2 Minute Cream of Wheat, Instant Original Cream of Wheat, and Malt—o-Meal) and two oat-based hot cereals (Regular Oatmeal and Instant Oatmeal) were evaluated using the vane method in an attempt to characterize the flow behavior of the cereals. The vane method was chosen as a simple, objective method of measurement to determine apparent yield stress values. The hot cereals were also presented to a semi-trained descriptive analysis panel for sensory characterization. The panelists were asked to evaluate the cereals against a defined reference sample in four categories: thickness (stirrability), stickiness to spoon, stickiness to self, and viscosity. These parameters were suggested and defined by the panelists as a group. The test method used was a 15cm line scale with a reference sample at 7.50m. Results from the rheological tests were evaluated to determine the change in apparent yield stress overtime (hydration and cooling) and also to compare the difference in apparent yield stress between the cereals in each category (i.e., wheat-based and oat- based). The results indicate that apparent yield stress is an effective parameter to characterize the flow behavior of the cereals. Each cereal had a characteristic value at any particular hold time relative to the other cereals in the category. The sensory scores were analyzed by AN OVA to determine if panelists could distinguish between the different types of cereal. Panelists were able to tell the difference between at least two of the three wheat-based cereals for all four sensory parameters. Conversely, the panelists were only able to distinguish between the oat-based cereals for the parameter of viscosity. 72 Sensory scores for each parameter and the instrumentally determined apparent yield stress values were compared to determine if a particular sensory score could be predicted by the instrumental measurement. Regression analysis indicates that there is poor correlation between apparent yield stress and all hot cereal sensory parameters considered in this study. The results from this work, although lacking correlation between sensory and instrumental parameters, are valuable. Data from the rheological measurements can be used to characterize the differences between each cereal. By choosing one hold time by which to compare the cereals, a quality control test can be designed for evaluation between batches of cereal as it is processed. Also, it is possible to use these apparent yield stress values to aid in product development through evaluation of alternative ingredients in a traditional product. 73 7. Appendix 7.1. Apparent Yield Stress Measurements for Wheat Based Cereals Table 7.1. Measurements for apparent yield stress for each hold time for 2 1/2 Minute Cream of Wheat. min 5min 8min 10min 12min 14min 15min 18min 20min 26.0 26.0 26.8 28.3 55.5 46.8 42.1 35.0 47.1 21.3 30.2 38.7 29.6 37.7 41.5 43.7 58.6 47.3 20.8 30.2 28.4 32.7 35.3 42.5 50.3 48.4 54.0 20.3 34.7 27.3 37.0 36.1 40.4 42.9 47.7 49.9 19.9 25.1 31.6 30.0 36.5 37.7 45.4 45.3 48.0 29.8 29.8 28.1 38.6 46.3 50.9 50.6 27.8 28.8 34.8 36.1 46.5 46.4 36.0 32.0 40.6 39.8 47.8 51.8 32.5 40.3 34.5 45.1 48.6 37.5 43.4 40.9 Average 21.7 30.3 30.4 30.9 40.2 40.2 42.3 47.2 48.7 STDev 2.5 3.7 3.8 3.4 8.6 3.1 5.0 6.1 3.0 Var 6.2 13.5 14.6 11.7 73.6 9.7 24.7 37.6 8.9 Table 7.2. Measurements for apparent yield stress for each hold time for Instant Cream of Wheat. 3min 5min 8min 10min 1 2min 14min 1 5min 1 8min 20min 45.1 64.5 64.8 72.0 60.5 48.7 53.1 80.1 52.7 41.2 45.1 69.7 62.6 69.1 75.8 60.6 67.1 68.9 50.8 58.6 49.1 53.7 62.9 61.8 67.5 74.4 90.1 55.4 57.4 61.9 44.4 54.5 77.6 55.3 59.8 62.2 61.5 61.5 51.5 65.6 62.1 67.7 80.5 68.3 60.9 52.6 43.6 52.5 65.9 49.9 59.2 69.7 70.0 70.3 46.6 50.2 52.7 72.9 75.6 65.5 44.7 66.7 45.6 54.8 63.0 38.6 63.3 45.1 72.8 Average 51.1 52.9 57.1 59.0 61.9 63.1 64.4 63.7 68.1 STDev 7.3 8.5 8.2 9.0 7.3 14.0 8.7 13.0 10.9 Var 52.7 71.5 67.6 81.4 53.6 196.1 75.0 168.5 119.5 74 Table 7.3. Measurements for apparent yield stress for each hold time for Malt-o-Meal. Average STDev Var 3min 19.5 19.6 18.0 18.0 17.8 20.0 21.2 19.8 19.3 20.5 19.4 1.1 1.2 5min 21.2 20.1 19.5 18.6 20.8 23.4 21.4 18.4 20.5 20.8 20.5 1.4 2.1 8min 23.4 22.1 22.7 24.8 22.1 23.6 22.4 22.5 23.6 23.1 23.0 0.8 0.7 10min 24.0 26.9 25.9 24.5 23.8 22.6 24.4 26.3 22.1 21.4 24.2 1.8 3.3 75 12min 31.5 29.0 28.7 29.5 31.9 30.1 1.5 2.1 14min 37.3 32.6 33.6 30.1 29.4 32.6 3.2 10.0 15min 35.7 28.2 27.6 28.6 31.8 30.4 3.4 11.5 18min 36.2 35.2 33.6 36.6 37.4 35.8 1.5 2.2 20min 37.5 40.9 40.8 34.7 40.0 38.8 2.6 6.9 7.2. Apparent Yield Stress Measurements for Oat Based Cereals Table 7 .4. Measurements for apparent yield stress for each hold time for One Minute Oatmeal. 3min 12min 14min 1 5min 1 8min 20min 1 10.4 554.4 569.3 480.0 639.0 749.2 133.7 527.8 524.8 431.8 670.1 570.1 83.2 491.9 483.8 756.8 580.2 799.4 100.9 456.2 553.7 462.2 506.1 656.0 106.1 513.2 439.5 463.4 618.1 627.7 115.5 539.6 705.1 550.0 100.8 716.3 638.6 Average 107.2 513.9 514.2 518.8 619.8 658.7 STDev 15.5 35.5 52.9 134.2 70.2 98.6 Var 239.8 1261.4 2802.3 17999.3 4931.3 9719.3 Table 7.5. Measurements for apparent yield stress for each hold time for Instant Regular Oatmeal. 76 3min 5min 8min l 0min 12min 14min 1 5min l 8min 20min 573.0 742.1 537.7 641.8 679.7 1048.7 780.1 846.0 984.4 410.7 644.8 774.3 880.3 737.5 702.2 748.7 1079.0 709.4 518.5 443.9 828.7 746.0 874.4 987.8 866.1 788.1 1004.6 488.8 559.5 795.2 543.3 780.0 809.5 826.8 945.9 862.5 461.8 791.0 604.8 736.0 808.8 871.6 823.9 1222.4 1207.8 404.2 473.1 743.1 770.3 1135.4 639.1 878.1 Average 476.2 613.3 714.0 742.2 776.1 883.9 809.1 976.3 984.0 STDev 64.8 129.8 116.0 121.1 73.4 138.4 45.5 176.4 180.9 Var 4201.3 16848.9 13447.5 14673.6 5382.1 19166.6 2068.0 31101.5 32719.1 7.3. Sensory Scores Based on 15cm Line Scale for Wheat-Based Cereals Table 7.6. Sensory scores for thickness (stirrability) for wheat—based cereals. COW ICOW MOM Panelist 1 8.53 8.29 11.29 2 8.29 4.78 11.88 3 6.59 11.52 8.09 4 7.97 4.23 8.45 5 10.62 10.97 13.38 6 8.72 12.91 13.42 7 6.75 12.43 12.59 8 6.43 8.21 10.18 9 6.63 9.87 10.22 10 6.24 10.38 11.88 11 8.72 9.28 9.71 12 8.33 8.8 9.83 1 7.3 7.18 10.74 2 7.62 8.8 9.99 3 7.66 8.09 12.31 4 7.3 7.86 9.47 5 8.21 12.2 14.37 6 8.17 11.45 14.52 7 7.62 13.06 13.97 8 7.46 8.96 13.22 9 6.87 6.79 13.58 10 6.67 10.93 11.21 11 7.46 9.47 14.64 12 7.03 8.6 10.81 1 6.51 6.59 12.59 2 6.16 7.86 10.22 3 7.58 8.33 12.39 4 6.99 6.99 12.43 5 5.01 12.31 14.33 6 6.99 8.13 9.43 7 6.3 10.93 10.97 8 4.5 10.58 13.62 9 5.09 4.66 9.08 10 6.51 5.61 8.57 11 3.71 11.76 13.97 12 6.47 10.62 11.56 Average 7.08 9.15 l 1.64 STDev 1.29 2.38 1.92 Var 1.67 5.66 3.69 77 Table 7.7. Sensory scores for stickiness to spoon for wheat-based cereals. COW ICOW MOM Panelist l 7.82 4.78 10.58 2 10.93 3.87 12.63 3 6.47 10.74 6.75 4 8.17 4.11 12.16 5 10.3 6.87 12.31 6 8.29 14.44 13.66 7 4.34 8.68 13.5 8 6.75 6.63 10.03 9 6.16 9.79 9.04 10 6.32 9.71 11.09 11 9.95 8.64 11.13 12 6.83 7.9 l 1.13 l 9.24 7.1 1 10.1 2 8.13 9.08 10.14 3 7.26 8.09 10.42 4 5.84 6.59 9.28 5 4.97 12.98 13.81 6 5.88 11.68 15 7 4.82 8.76 13.3 8 7.58 8.25 13.38 9 7.22 5.45 11.92 10 5.29 6.91 11.01 11 6.04 7.18 12.59 12 6.99 8.49 10.66 1 6.2 6.51 12.47 2 6.4 6.95 9.51 3 6.51 9.55 12.31 4 7.5 7.93 1 1.09 5 4.66 11.49 14.25 6 8.09 5.84 10.66 7 6.25 8.29 11.17 8 4.26 8.68 13.38 9 4.34 0.52 8.96 10 7.11 5.21 7.86 11 3.51 10.54 13.38 12 6.28 8.21 11.68 Average 6.74 7.96 1 1.45 STDev 1.72 2.65 1.86 Var 2.94 7.04 3.44 78 7.8. Sensory scores for stickiness to self for wheat-based cereals. COW ICOW MOM Panelist 1 7.7 5.61 9.47 2 9.75 8.13 10.06 3 6.28 13.26 7.11 4 10.93 2.84 8.76 5 8.72 12.35 12.51 6 6.59 13.81 13.46 7 7.54 13.38 12.12 8 7.5 9.39 9.91 9 6.87 10.74 9.28 10 7.38 10.97 12.23 11 7.54 5.92 12.16 12 6.83 9.75 10.38 1 8.13 7.89 10.18 2 10.06 1.7 4.54 3 7.22 8.8 9.99 4 9.47 5.84 12 5 4.97 11.52 12.43 6 2.69 12.63 15 7 2.65 13.1 13.38 8 6.36 8.25 10.7 9 7.38 7.89 13.3 10 3.79 11.72 10.14 11 4.07 11.29 13.38 12 6.71 10.5 11.09 1 5.13 6.83 12.16 2 5.68 8.41 9.91 3 8.57 4.3 10.42 4 7.46 8.57 13.18 5 6.55 13.18 12.75 6 6.71 11.21 10.18 7 5.95 11.76 11.96 8 4.23 10.77 13.54 9 2.96 7.14 6.36 10 7.11 6.43 7.42 11 4.11 12.08 13.93 12 7.93 5.49 3.99 Average 6.65 9.26 10.82 STDev 2.03 3.15 2.56 Var 4.12 9.90 6.57 79 Table 7.9. Sensory scores for viscosity for wheat-based cereals. COW ICOW MOM Panelist 1 8.01 7.7 11.6 2 10.5 10.7 12.12 3 6.55 13.34 6.43 4 8.88 2.13 8.6 5 7.42 11.37 11.8 6 8.92 13.02 11.76 7 5.21 13.81 9.75 8 6.36 10.54 9.04 9 6.63 9.43 11.72 10 6.79 11.56 12.98 11 9.16 10.38 9.39 12 6.47 9.28 9.67 1 7.89 7.86 9.75 2 4.97 12.04 5.41 3 6.04 8.21 7.7 4 6.59 8.37 9.99 5 5.21 14.4 10.42 6 5.05 14.33 14.48 7 3.48 13.46 11.72 8 6.08 8.49 11.92 9 7.14 7.93 13.46 10 5.25 8.09 7.18 11 3.71 11.68 12.23 12 7.07 8.76 10.62 1 4.94 8.41 11.96 2 5.76 9.51 10.89 3 6 9.71 4.42 4 8.29 9.04 11.8 5 7.26 13.66 9.63 6 6.71 7.9 9.16 7 6.35 9.32 8.64 8 10.58 10.03 13.93 9 2.65 11.21 7.26 10 5.96 6.32 7.42 11 3.99 11.96 10.03 12 6.24 10.89 10.97 Average 6.50 10.13 10.16 STDev 1.79 2.55 2.34 Var 3.21 6.53 5.47 80 7.4. Sensory Scores Based on 15cm Line Scale for Oat-Based Cereals Table 7.10. Sensory scores for thickness (stirrability) for oat-based cereals. lMIN IOAT Panelist 1 10.77 8.53 2 10.81 8.29 3 6.67 6.59 4 4.7 7.97 5 7.38 10.62 6 6.83 8.72 7 7.5 6.75 8 5.41 6.43 9 7.07 6.63 10 8.84 6.24 11 7.22 8.72 12 7.42 8.33 l 1.62 4.78 2 4.3 3.91 3 2.17 7.74 4 8.76 6.63 5 7.86 12.31 6 8.41 10.03 7 4.15 10.85 8 8.17 6.59 9 2.73 6.4 10 6.2 9.75 11 6.67 8.01 12 7.7 7.26 1 10.66 12.51 2 7.54 7.89 3 8.37 2.57 4 6.47 8.84 5 8.13 10.58 6 6.63 6.99 7 8.01 7.86 8 7.42 8.72 9 8.17 9.95 10 7.07 7.54 l l 7.82 1.82 12 6.71 10.46 Average 7.01 7.88 STDev 2.1 1 2.35 Var 4.45 5.54 81 Table 7.11. Sensory scores for stickiness to spoon for oat-based cereals. lMIN IOAT Panelist 1 8.25 6.75 2 8.37 6.99 3 6.16 4.66 4 0.4 0.56 5 7.38 6.28 6 6.95 8.33 7 7.5 8.8 8 6.95 8.17 9 7.22 8.17 10 7.26 7.93 11 6.91 8.53 12 7.5 7.82 1 4.58 7.18 2 7.54 0.4 3 5.21 7.5 4 7.62 8.41 5 7.66 7.7 6 7.62 7.82 7 6.75 7.9 8 7.46 7.58 9 2.17 6.67 10 7.38 8.53 11 6.95 7.89 12 7.5 7.5 1 9.67 9.04 2 7.42 7.38 3 7.46 7.62 4 6.95 6.99 5 7.42 8.88 6 6.71 7.82 7 7.5 7.58 8 7 .3 7.78 9 7.5 8.88 10 7.3 7.42 11 7.82 7.5 12 7.11 10.62 Average 6.93 7.38 STDev 1.62 1.95 Var 2.63 3.79 82 Table 7.12. Sensory scores for stickiness to self for oat-based cereals. lMIN IOAT Panelist l 9 5.76 2 8.6 2.57 3 6.99 6.43 4 2.49 2.49 5 6.59 4.9 6 7.54 10.14 7 7.5 9.59 8 6.87 7.97 9 7.38 8.84 10 7.54 9.16 1 1 5.88 7.78 12 7.5 7.5 l 6.47 6.2 2 8.72 9 3 6.12 7.38 4 8.57 9.12 5 8.57 11.6 6 7.5 9.12 7 6.04 8.8 8 8.8 6.4 9 2.41 5.76 10 7.42 6 1 1 6.24 7.97 12 7.5 7.03 1 9.83 9.16 2 7.46 4.38 3 9.55 8.09 4 5.25 2.13 5 7.5 10.85 6 7.5 3.12 7 7.62 5.57 8 7.54 6.47 9 6.87 6.24 10 7.18 7.42 1 l 8.68 3.75 12 7.03 9.51 Average 7.23 7.06 STDev 1.56 2.40 Var 2.43 5.74 83 Table 7.13. Sensory scores for viscosity for oat-based cereals. lMIN IOAT Panelist 1 9.59 5.61 2 10.26 1.94 3 6.39 3.36 4 6.63 4.66 5 7.07 3.67 6 7.14 6.36 7 7.5 0.83 8 6.75 5.45 9 7.42 7.93 10 8.17 6.2 11 7.34 5.88 12 6.91 9.12 l 3.04 2.88 2 4.86 3.16 3 4.19 6.79 4 10.1 6 5 7.5 8.72 6 7.11 9.35 7 5.05 7.58 8 8.45 4.03 9 4.82 7.14 10 8.45 8.49 11 6.91 6.63 12 7.93 6.4 1 8.72 7.89 2 7.42 4.42 3 8.84 2.21 4 6.51 3.63 5 7.5 10.7 6 7.5 6.24 7 9.43 2.02 8 7.54 5.57 9 7.22 5.17 10 7.46 6.83 1 1 7.97 1.62 12 6.87 6.99 Average 7.29 5.60 STDev 1.54 2.41 Var 2.37 5.80 84 7.5. ANOVA Results for Descriptive Analysis for Wheat-Based Cereals Table 7.14. ANOVA table for wheat cereals for the sensory parameter “thickness.” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 13 479.26 36.87 12.34 <.0001 Error 94 280.89 2.99 Corrected total 107 760. l 5 Table 7.15. Effect of subject on variation between wheat cereals for the sensory parameter “thickness.” Source Degrees of Type 111 SS Mean Square F Value Pr>F freedom Subject 11 105.1 1 9.56 3.20 0.0010 Treatment 2 374.15 187.08 62.61 <.0001 Table 7.16. Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “thickness.” 1 2 <.0001 <.0001 <.0001 3 85 <.0001 Table 7.17. ANOVA table for wheat cereals for the sensory parameter “stickiness to spoon.” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 13 515.34 39.64 9.66 <.0001 Error 94 385.64 4.10 Corrected total 107 900.99 Table 7.18. Effect of subject on variation between wheat cereals for the sensory parameter “stickiness t0 Spoon.” Source Degrees of Type 111 SS Mean Square F Value Pr>F fi'eedom Subject 1 l 84.42 7.67 1.87 0.0531 Treatment 2 430.93 215.46 52.52 <.0001 Table 7.19. Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “stickiness to spoon.” 1 2 3 l 86 2 0.0125 <.0001 3 Table 7.20. ANOVA table for wheat cereals for the sensory parameter “stickiness to self.” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 13 414.77 31.91 4.80 <.0001 Error 94 624.39 6.64 Corrected total 1 07 1039. 1 6 Table 7.21. Effect of subject on variation between wheat cereals for the sensory parameter “stickiness to self.” Source Degrees of Type 111 SS Mean Square F Value Pr>F freedom Subject 11 96.17 8.74 1.32 .2277 Treatment 2 318.60 159.30 23.98 <.0001 Table 7.22. Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “stickiness to self.” 1 2 3 1 <.0001 <.0001 2 .0121 3 .0121 87 Table 7.23. ANOVA table for wheat cereals for the sensory parameter “viscositL” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 13 382.15 29.40 5.89 <.0001 Error 94 468.84 4.99 Corrected total 107 850.99 Table 7.24. Effect of subject on variation between wheat cereals for the sensory parameter “viscosity.” Source Degrees of Type 111 SS Mean Square F Value Pr>F freedom Subject 11 63.20 5.74 1.15 .3313 Treatment 2 318.95 159.48 31.97 <.0001 Table 7.25. Least Squares Means for effect of treatment for wheat cereals and the sensory parameter “thickness.” 1 2 <.0001 3 <.0001 .9576 .9576 88 7.6. AN OVA Results for Descriptive Analysis for Oat-Based Cereals Table 7.26. ANOVA table for oat cereals for the sensory parameter “thickness (stirrability).” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 12 70.530 5.877 1.18 0.3153 Error 59 292.749 4.962 Corrected total 71 363.278 Table 7.27. ANOVA table for oat cereals for the sensory parameter “stickiness to spoon.” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 12 47.874 3.989 1.31 0.2400 Error 59 180.310 3.056 Corrected total 71 228.183 Table 7.28. ANOVA table for oat cereals for the sensory parameter “stickiness to self.” Source Degrees of Stun of squares Mean square F value Pr>F freedom Model 12 49.697 4.141 1.03 0.4319 Error 59 236.537 4.010 Corrected total 71 286.234 89 Table 7.29. ANOVA table for oat cereals for the sensory parameter “viscosity.” Source Degrees of Sum of squares Mean square F value Pr>F freedom Model 12 99.500 8.292 2.05 0.0348 Error 59 238.307 4.039 Corrected total 71 337.807 Table 7.30. Effect of subject on variation between oat cereals for the sensory parameter “viscosity.” Source Degrees of Type 111 SS Mean Square F Value Pr>F freedom Subject 11 47.667 4.333 1.07 0.3982 Treatment 1 51.833 51.833 12.83 0.0007 90 REFERENCES BAGLEY, EB. and CHRISTIANSON, DD. 1983. Yield stresses in cooked wheat starch dispersions. Starch 35, 81-86. BARNES, HA. 1999. The yield stress — a review or ‘navw. pet’ - everything flows? J. Non-Newtonian Fluid Mech. 81, 133-178. BARNES, HA. 1999. A brief history of the yield stress. Applied Rheology 9, 262-266. BARNES, HA. 2001. An examination of the use of rotational viscometers for the quality control of non-newtonian liquid products in factories. Applied Rheology 11, 89-101. BARNES, HA. and NGUYEN, Q.D. 2001. Rotating vane rheometry — a review. J. Non-Newtonian Fluid Mech. 98, 1-14. BOURNE, M.C. 1979. Theory and application of the puncture test in food texture measurement. In Food Texture and Rheology (P. Sherman, ed.) pp. 95-142, Academic Press, New York. BRANDT, M.A., E.Z. SKINNER, and J .A. COLEMAN. 1963. Texture profile method. J. Food Sci. 28, 404-409. BREIDINGER, S.L. and STEFFE, J .F . 2001. Texture map of cream cheese. J. Food Sci. 66 (3), 453-456. BRIGGS, J .L., STEFFE, J .F ., and USTUNOL, Z. 1996. Vane method to evaluate the yield stress of frozen ice cream. J. Dairy Sci. 79, 527-531. } CALDWELL, E.F. 1973. Industrial Uses of Cereals: Oats. 58th Annual Meeting of American Association of Cereal Chemists. CALDWELL, E.F., FAST, R.B, SALISBURY, K., SEIBERT, 8.13., and SLIMMON, I. 2000. Hot Cereals. Breakfast Cereals and How They Are Made (R.B. Fast and E.F. Caldwell, ed.) 2nd Ed., pp 315-342, American Association of Cereal Chemists, Inc., St. Paul, Minnesota. CANTOR, S.M., FALLER, L., and WOLNAK, B. John S. Campbell, assignee. 1959 June 9. US Patent 2,890,118. CANTU-LOZANO, D., RAO, M.A., and GASPARETTO, CA. 2000. Rheological properties of noncohesive apple dispersion with helical and vane impellers: effect of concentration and particle size. J. Food Process Engineering 23, 373-3 85. 91 DANIEL, J .R. and WEAVER, C. 2000. Carbohydrates: functional properties. Food Chemistry: Principles and Applications. (G.L. Christen and J .S. Smith, eds.) pp. 55-78, Science Technology System, California. DANIELS, R. 1974. Breakfast Cereal Technology. Noyes Data Corporation, New Jersey. DAUBERT, C.R., TKACHUK, J.A., and TRUONG, V.D. 1998. Quantitative measurement of food spreadability using the vane method. J. Texture Studies 29, 427-435. DOUBLIER, J .L. 1981. Rheological studies on starch — flow behaviour of wheat starch pastes. Starch 33, 415-420. DOUBLIER, J .L. 1987. A rheological comparison of wheat, maize, faba bean and smooth pea starches. J. of Cereal Sci. 5, 247-262. DOUBLIER, J .L., PATON, D. and LLAMAS, G. 1987. A rheological investigation of oat starch pastes. Cereal Chem 64 (1), 21-26. DRAKE, M.A., GERARD, RD, and CIVILLE, G.V. 1999. Ability of hand evaluation versus mouth evaluation to differentiate texture of cheese. Journal of Sensory Studies 14, 425-441. DZUY, N.Q. and BOGER, D.V. 1985. Direct yield stress measurement with the vane method. J. Rheol. 29, 335-347. EINSTEIN, MA. 1991. Descriptive techniques and their hybridization. Sensory 1 Science Theory and Applications in Foods. (H.T. Lawless and BF. Klein, eds.) pp. 317-338, Marcel Dekker, Inc., New York City, NY. ELLIS, HS. and RING, S.G. 1985. A study of some factors influencing amylose gelation. Carbohydrate Polymers 5, 201-213. EVANS, ID. and HAISMAN, DR. 1979. Rheology of gelatinized starch suspensions. J. of Texture Studies 10, 347-370. FRENCH, D. 1984. Organization of starch granules. Starch Chemistry and Technology. 2“d Ed. (R.L. Whistler, J.N. BeMiller, and E.F. Paschall, eds.), pp. 184-242, Academic Press, Inc., New York. GUINARD, J .X. and MAZZUCCHELLI, R. 1996. The sensory perception of texture and mouthfeel. Trends in Food Sci. and Tech. 7 (7), 213-219. GWARTNEY, E. and H. HEYMANN. 1996. Profiling to describe the sensory characteristics of a simple model menthol solution. Journal of Sensory Studies 11, 39-48. 92 HANSER, W.F. and MARTIN, L.J. National Oats Company, assignee. 1970 July 14. US. Patent 3,520, 695. HEYMANN, H. 1994. A comparison of descriptive analysis of vanilla by two independently trained panels. Journal of Sensory Studies 9, 21. HUFFMAN, G.W. and MOORE, J .W. Quaker Oats Company, assignee. 1961 Sept. 5. US. Patent 2,999,018. HYLDON, R.G. Quaker Oats Company, assignee. 1970 Sept. 1. US. Patent 3,526,513. JANE, J. 2000. Carbohydrates: basic concepts. Food Chemistry: Principles and Applications. (G.L. Christen and J .S. Smith, eds.), pp. 35-53, Science Technology System, California. JANE, J. and CHEN, J. 1992. Effect of amylose molecular size and amylopectin branch chain length on paste properties of starch. Cereal Chem. 69, 60-65. KARWOWSKI, J. Nabisco Brands, Inc., assignee. 1985 Nov. 5. Process of preparing instant, flaked, wheat farina. US. Patent 04551347. KARWOWSKI, J. Nabisco Brands, Inc., assignee. 1986 May 20. Method of preparing instant, flaked, wheat farina. US. Patent 661383. KARWOWSKI, J. Nabisco Brands, Inc., assignee. 1987 May 12. Method of preparing instant, flaked, wheat farina. US. Patent 04664931. KOKINI, J .L. 1985. Fluid and semi-solid food texture and texture-taste interactions. Food Technology 11, 86-92. KOVALENKO, I.V. and BRIGGS, J .L. 2002. Textural charcterization of soy-based yogurt by the vane method. J. of Texture Studies 33, 105-118. KULKARNI, S. 1986. Inter-reltionship among rheological characteristics of butter measured by objective and subjective methods. Indian J. Dairy Sci. 39(1), 65-68. LANG, ER. and RHA, C. 1981. Determination of the yield stress of hydrocolloid dispersions. J. Texture Studies 12, 47-62. LAWLESS, HT. 1991. Bridging the gap between sensory science and product evaluation. Sensory Science Theory and Applications in Foods. (H.T. Lawless and BF. Klein, eds.) pp. 1-36, Marcel Dekker, Inc. New York City, NY. MARR, B.U. and PEDERSEN, GP. 1999. Texturizing mayonnaise with custom-made pectin product. International Food Marketing and Technology 35, 86-89. 93 MEILGAARD, M., G.V. CIVILLE, and B.T. CARR. 1991. Sensory Evaluation Techniques, 2nd Ed., CRC Press, Boca Raton, FL. MEULLENET, J .F ., B.G. LYON, J .A. CARPENTER, and CE. LYON. 1998. Relationship between sensory and instrumental texture profile attributes. Journal of Sensory Studies. 13, 77-93. MILLER, B.S., DERBY, R.I., and TRIMBO, HE. 1973. A pictorial explanation for the increase in viscosity of a heated wheat starch-water suspension. 1973. Cereal Chemistry 50, 271-280. MISSAIRE, F ., QIU, CG, and RAO, MA. 1990. Yield stress of structured and unstructured food suspensions. J. Food Texture 21, 479—490. MORRIS, E.R., R.K. RICHARDSON, and L.J. TAYLOR. 1984. Correlation of the perceived texture of random coil polysaccharide solutions with objective parameters. Carbohydrate Polymers 4, 175-191. MORTENSEN, B.K., H. DANMARK. 1982. Consistency characteristics of butter. Milchwissenschafi 3 7 (9), 530—532. MOSKOWITZ, HR. 1983. Product Testing and Sensory Evaluation of Foods. Food & Nutrition Press, Inc., Westport, CT. MOSKOWITZ, HR. 1996. Experts versus consumers: a comparison. Journal of Sensory Studies 1 I, 19-37. MUNOZ, A.M. 1986. Development and application of texture reference scales. Journal of Sensory Studies I, 55-83. MUNOZ, A.M., and G.V. CIVILLE. 1998. Universal, product and attribute specific scaling and the development of common lexicons in descriptive analysis. Journal of Sensory Studies 13, 57-75. MUNOZ, A.M., A.S. SCZCESNTAK, M.A. EINSTEIN, and NO. SCHWARTZ. 1992. The texture profile. Descriptive Analysis Testing for Sensory Evaluation. (R.C. Hootman, ed.), pp. 35-51, American Society for Testing and Materials, PA. MURRAY, J .M., C.M. DELAHUNTY, and LA. BAXTER. 2001. Descriptive sensory analysis: past, present, and future. Food Research International 34, 461-471. OMURA, AP. and STEFFE, J .F . 2001. Centrifugal slump test to measure yield stress. J. Food Sci. 66, 137-139. 94 O’MAHONEY, M. 1991. Descriptive analysis and concept alignment. Sensory Science Theory and Applications in Foods. (H.T. Lawless and B.P. Klein, eds.), pp. 223- 267, Marcel Dekker, Inc., New York City, NY. PELEG, M. 1983. The semantics of rheology and texture. Food Technology 3 7 (l 1), 54-61. PIGGOTT, J .R. 1991. Selection of terms for descriptive analysis. Sensory Science Theory and Applications in Foods. (H.T. Lawless and HP. Klein, eds.), pp. 339- 351, Marcel Dekker, Inc., New York City, NY. PIGGOTT, J .R., Simpson, S.J., and Williams, S.A.R. 1998. Sensory Analysis. International Journal of Food Science and Technology 33, 7-18. QIU, CG. and RAO, MA. 1988. Role of pulp content and particle size in yield stress of apple sauce. J. Food Sci. 53, 1165-1170. RAO, MA. and STEFFE, J .F . 1997. Measuring yield stress of fluid foods. Food Tech. 51 (2), 50-52. RICHARDSON, R.K., E.R. MORRIS, S.B. ROSS-MURPHY, L.J. TAYLOR, and I.C.M. DEA. 1989. Characterization of the perceived texture of thickened systems by dynamic viscosity measurements. Food Hydrocolloids 3 (3), 175-191. ROBERTS, A.K., and VICKERS, Z.M. 1994. A comparison of trained and untrained judges’ evaluation of sensory attribute intensities and liking of cheddar cheeses. Journal of Sensory Studies 9, 1-20. RONAI, KS. and SPANIER, H.C. Nabisco, Inc., assignee. 1972a Feb. 8. US. Patent 3,640,729. RONAI, KS. and SPANIER, H.C. Nabisco, Inc., assignee. 1972b Feb. 8. US. Patent 3,640,279. RONAI, KS. and SPANIER, H.C. Nabisco, Inc., assignee. 1972c Feb. 8U.S. Patent 3,640,728. SHANNON, J .C. and GARWOOD, D.L. 1984. Genetics and physiology of starch development. Starch Chemistry and Technology. 2“‘1 Ed. Ed. (R.L. Whistler, J.N. BeMiller, and E.F. Paschal], eds.), pp. 25-86, Academic Press, Inc., New York. SHERMAN, P. 1988. Rheological evaluation of the textural properties of foods. Progress and Trends in Rheology 2, 44-53. 95 SPRING, Jr. F.E. Ralston Purina Company, assignee. 1971 Nov. 16. Cook-in-the-bowl cereal. US. Patent 3,620, 761. STANLEY, D.W. 1987. Food texture and microstructure. Food Texture (H.R. Moskowitz, ed.) pp. 35-64, Marcel Dekker, Inc., New York. STANLEY, N.L. and TAYLOR, L.J. 1993. Rheological basis of oral characteristics of fluid and semi-solid foods: A review. Acta Psychologica 84, 79-92. STEP FE, J .F. 1996. Rheological Methods in Food Process Engineering. 2nd ed. Freeman Press, East Lansing, MI. SWANSON, RB. 2000. Grains and Legumes. Food Chemistry: Principles and Applications. (G.L. Christen and J .S. Smith, eds.), pp. 311-350, Science and Technology System, California. SZCZESNIAK, AS. 1963. Classification of textural characteristics. J. Food Sci. 28, 385-389. SZCZESNIAK, AS. 1987. Correlating sensory with instrumental texture measurements — an overview of recent developments. Journal of Texture Studies 18, 1-15. SZCZESNLAK, A.S., M.A. BRANDT, and H.H. FREIDMAN. 1963. Development of standard rating scales for mechanical parameters of texture and correlation between the objective and the sensory methods of texture evaluation. J. Food Sci. 28, 397-403. TANG, C., F. I-ISIEH, H. HEMANN, and HE. HUFF. 1999. Analyzing and correlating instrumental and sensory data: a multivariate study of physical properties of cooked wheat noodles. Journal of Food Quality 22, 193-211. TIMBERS, GE. and VOISEY, P.W. 1987. The Ottawa Pea Tenderometer and the Ottawa Texture Measuring System. The Evolution of a System for Quality Control and Food Texture Research. Food Texture (H.R. Moskowitz, ed.) pp. 65— 93, Marcel Dekker, Inc., New York. TRESSLER, D.K., assignee. 1970 Jan. 20. US. Patent 3,490,915. WALTON, A. 2000. Modern rheometry in characterising the behaviour of foods. Food Sci. and Tech. Today 14, 144-146. WENDIN, K., K. AABY, A. EDRIS, M.R. ELLEKJAER, R. ALBIN, B. BERGENSTAHL, L. JOHANSSEN, E.P. WILLERS, and R. SOLHEIM. 1997. Low-fat mayonnaise: influences of fat content, aroma compounds and thickeners. Food Hydrocolloids 11, 87-99. 96 WENDIN, K. and HALL, G. 2001. Influences of fat, thickener, and emulsifier contents on salad dressing: static and dynamic sensory and rheological analyses. Lebensm.-Wiss. u.-Technol. 34, 222-233. WILLIAMS, MR. and BOWLER, P. 1982. Starch gelatinization: a morphological study of triticeae and other starches. Starch 34 , 221-223. WESON, L.L., SPEERS, R.A., and TUNG, MA. 1993. Yield stresses in molten chocolates. J. Texture Studies 24, 269-286. YOO, B., RAO, M.A., and STEFFE, J .F . 1995. Yield stress of food dispersions with the vane method at controlled shear rate and shear stress. J. Texture Studies 26, 1-10. YOUNG, AH. 1984. Fractionation of starch. Starch Chemistry and Technology. 2nd Ed. (R.L. Whistler, J .N. BeMiller, and E.F. Paschall, eds.), pp. 249-274, Academic Press, Inc., New York. ZOBEL, HF. 1984. Gelatinization of starch and mechanical properties of starch pastes. Starch Chemistry and Technology. 2nd Ed. (R.L. Whistler, J .N. BeMiller, and E.F. Paschall, eds.), pp. 285-305, Academic Press, Inc., New York. 97