| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Nutrition and Food Science Department, American University of Beirut, Riad El Solh 1107 2020, Beirut, Lebanon
2 Corresponding author: ammar.olabi{at}aub.edu.lb
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Key Words: Halloumi cheese • bovine milk • ovine milk • sensory properties
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Halloumi cheese is a semi-hard to hard unripened cheese and has no obvious skin or rind. Its texture is compact and unyielding to applied pressure. Halloumi cheese is elastic and compact with no holes. If holes are present they are usually scarce and irregular. The cheese can be easily sliced into generous portions. The color varies from white when ovine or caprine milk is used to yellowish when bovine milk is the main ingredient. Upon heating, the texture is similar to that of the raw cheese but the stretch and melt properties are altered: the sliced cheese melts evenly, has the capacity to stretch but is not tough or chewy, and behaves like a concentrated viscoelastic polymer solution when molten (Lelievre et al., 1990; Robinson, 1991; Papademas and Robinson, 1998).
Fresh Halloumi cheese is consumed directly after manufacture. Fresh Halloumi cheese has a distinct aroma and a mild milky and creamy flavor. On the other hand, mature Halloumi cheese is matured for at least 40 d after production before being introduced to the market. Alterations in the texture and flavor of the cheese take place during maturation (Papademas, 2006).
Traditionally, Halloumi cheese was produced from raw ovine or caprine milks, alone or as blends. However, because of the growing market demand, new regulations were created to permit the use of bovine milk if the origin of the cheese milk is declared on the label of the product (Moatsou et al., 2004). Halloumi cheese has a unique flavor and taste and a connoisseur can easily distinguish Halloumi made from different milk mixtures (Papademas, 2006). However, these unique sensory characteristics are more pronounced in the matured cheese compared with fresh cheese (Papademas and Robinson, 2000). The manufacturing of Halloumi cheese has been described previously (Papademas, 2006).
According to the Cypriot Standards (Papademas et al., 2000), Halloumi cheese should contain a maximum of 3% sodium chloride, have a minimum fat in dry matter content of 43%, and a maximum moisture content of 46%. However, these figures apply mainly to traditional Halloumi cheese produced from ovine and caprine milks and studies have shown that Halloumi cheese produced from bovine milk can have a different chemical composition (Papademas et al., 2000). Several studies have revealed changes in the cheese composition between cheeses produced from bovine versus ovine and caprine milks (Papademas and Robinson, 2000; Milci et al., 2005). Moreover, several studies have compared the acceptance of Halloumi cheese produced from bovine, ovine, or caprine milk (Kaminarides et al., 2000; Papademas and Robinson, 2000, 2001; Milci et al., 2005). Papademas and Robinson (1998) believed the transition toward bovine milk had an impact on the sensory quality of Halloumi cheese, because the "old" aroma and flavor had been lost.
Consumption of saturated fat has been shown to be highly correlated with an increased risk for obesity, atherosclerosis, coronary artery disease, and elevated blood pressure (Watts et al., 1996; Van Horn and Ernest, 2002). This has created an increased consumer awareness and increase in the demand for and supply of low-fat foods including cheese (Sloan, 2001; Kavas et al., 2004). The production of low-fat cheese has been regulated by many agencies (Shank and Carson, 1990). In Cyprus, 2 categories of low-fat cheese have been recognized: less-fat cheese containing 25% less fat than the regular cheese and reduced-fat cheese containing 50% less fat than the regular cheese (Papademas et al., 2000; Punidadas et al., 2000). In a similar manner to the type of milk differences, compositional differences were obtained for full-fat, less-fat, and reduced-fat Halloumi cheese produced from bovine milk (Papademas et al., 2000).
Papademas et al. (2000) found that the increase in protein content affected some of the basic properties of Halloumi cheese such as chewiness and hardness. Textural analysis showed that low-fat and reduced-fat samples were significantly harder and chewier than the full-fat samples. Protein content also affected the fracturability of the samples, as fracturability was significantly lower in the full-fat cheese. Other studies have demonstrated the effect of Halloumi’s fat content on the textural properties of the cheese (Theophilou and Wilbey, 2007). The sensory evaluation scores obtained from their study showed a significant difference in the overall preference for the full-fat, low-fat, and reduced-fat bovine Halloumi cheese. Preference increased as fat in dry matter increased and decreased as protein content increased. Therefore, to improve reduced-fat products, several alternatives have been proposed in recent years. These alternatives are usually either a modification of the cheese-making procedure or the introduction of new technologies (Rodriguez, 1998). Theophilou and Wilbey (2007) succeeded in improving the meltability and decreasing the hardness of low- or reduced-fat Halloumi cheese by lowering the pH during production.
This study reports on the characteristics of Halloumi cheese, a necessary step for optimizing its production and for creating adequate processing changes, especially for the manufacture of high quality reduced- and low-fat Halloumi. Considerable investments have been made in the Middle Eastern dairy sector over the last few years, and product development activities geared toward the production of healthier food products for export markets can be of considerable economical benefit to Lebanese and Middle Eastern food companies, especially at a time when competing in the regional and global market is becoming more demanding and more challenging. The objectives of this work were to characterize the physicochemical and sensory properties of bovine and ovine Halloumi of different fat levels and to provide a baseline sensory profile for the different types of samples.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Processing.
For each batch, 175 L of raw milk was batch pasteurized at 70°C for 20 min in a cheese vat. To produce low-fat cheese, the milk was first heated to 35°C and then skimmed using a cream separator (Warenhandels GmbH, Treibach, Austria). Whole milk and skim milk were then recombined to provide milk with varying fat contents (0.9 to 2.5%) to produce the reduced- and low-fat cheeses, and the mixture obtained was pasteurized at 70°C for 20 min. To produce full-fat, reduced-fat, and low-fat cheeses from bovine milk, milk fat was standardized at 4, 1.7, and 0.9%, respectively. The 2 batches of cheese obtained had fat contents of 20.88% (SD = 0.31) and 19.94% (SD = 0.54) for full-fat cheese, 12.99% (SD = 0.77) and 13.38% (SD = 0.56) for reduced-fat, and 6.02% (SD = 0.78) and 4.81% (SD = 0.39) for low-fat, respectively. To produce full-fat, reduced-fat, and low-fat ovine cheese, milk fat was standardized at 6 and 6.6% for the 2 batches of full-fat, 2.5% for reduced-fat, and 1.7% for low-fat cheese. The 2 batches of cheese obtained had fat contents of 23.10% (SD = 0.61) and 21.01% (SD = 0.47) for full-fat cheese, 12.99% (SD = 1.12) and 11.96% (SD = 0.39) for reduced-fat, and 6.83% (SD = 0.61) and 7.61% (SD = 0.71) for low-fat cheese, respectively. The milk fat levels for the 2 types of milks were selected based on preliminary trials to produce Halloumi samples that conformed to the Cypriot standards for full-fat (25.75 ± 3.14%), reduced-fat (at least 25% fat reduction from full-fat), and low-fat (at least 50% fat reduction from full-fat) Halloumi cheese (Papademas et al., 2000). In addition, the protein and casein contents of the milks were tested in preliminary tests a few days before the experiment and, accordingly, the milk fat levels were adjusted to produce consistent casein:fat ratios (±0.2) in the milk for a specific fat level for both types of milk (Haddadin et al., 1995).
After pasteurization, milk was cooled to 37 ± 1°C and CaCl2 was added at 0.15%. After stirring, rennet (Chy-Max, Chr. Hansen, Hørsholm, Denmark) was added at 0.04% and the milk was stirred again for 2 min. The curd was left to form for 35 min, and then the coagulum was cut into pieces in the form of a 1-cm3 dice. The curd was left to rest for 5 min, and then it was transferred to square metallic molds lined with cheesecloth. The cloth was folded to cover the upper surface of the curd. The curd was then pressed for 60 min. A pressure of 300 kPa was applied for the first 15 min then the pressure was increased to 400 kPa for the remaining time. In the meantime, the whey was heated to around 95°C to allow the whey proteins to aggregate. Citric acid was added at 0.6% to facilitate the coagulation and floating of the whey protein. The protein aggregates were removed using a perforated metallic scoop. The pressed curd was cut into approximately 10- x 15-cm rectangular pieces and placed in the hot whey. The blocks were cooked in the whey for 20 to 40 min and were removed once they had floated to the surface. The cooked blocks were allowed to cool before being folded. The folded cheese blocks were then placed in plastic containers where pasteurized brine (9%) was added and the containers were transferred to a walk-in cooler (4°C).
Storage.
The 12 batches of cheese were stored in plastic containers filled with pasteurized brine (9%) before sensory evaluation. The cheese stored in brine was also used for the physicochemical analyses and texture measurements. In addition, representative samples were selected from the different batches, ground using a Moulinex food grinder (Food Chopper, Moulinex, Ecully, France) and either analyzed (moisture, ash, calcium, and sodium) or placed in glass jars before being stored in the freezer (–20°C).
Chemical Analyses
Representative samples were prepared according to (AOAC, 2000; method 920.11). The methods of AOAC (2000) were used to determine moisture content (method 948.12), ash content (method 935.42), calcium and sodium contents of the cheese (method 991.25), fat content of the cheese (method 933.05), and protein content (method 920.125). The percentage of protein was calculated using a conversion factor of 6.38. The fat content of the milk was determined according to the Gerber method (IDF, 1991). All determinations were carried out in triplicate.
Instrumental Texture Analysis
The evaluation of the cheeses’ textural properties was carried out using 3 samples from each batch. Therefore, a total of 36 samples were tested (3 replicates x 2 types of milk x 3 fat levels x 2 batches). A texture analyzer (QTS25, Brookfield Engineering Labs, Middleboro, MA) equipped with a 38.1-mm cylindrical probe was used. The speed of the cross-head was 60 mm/min and the time elapsed between the first and second bite was 8 s after compressing the sample by 25% of its height (0.5 cm). Sample cubes (2 cm3) were taken from the center of the cheese blocks. After being cut, the samples were allowed to temper for 20 min and reach room temperature before testing. Texture profile analysis was performed on the samples with adhesiveness, chewiness, cohesiveness, hardness, and springiness being the tested parameters.
Descriptive Analysis
The panel consisted of 11 members. The panelists were 8 females and 3 males (mean age = 25, range = 22–34, SD = 3.59) who were selected based on their willingness to participate and time availability. The panelists were mainly graduate students in the Nutrition and Food Science Department at the American University of Beirut. They were not informed about the true nature of the project or that of the cheese; however, they were all familiar with Halloumi cheese and its characteristics. The study was approved by the Institutional Review Board of the American University of Beirut.
The panelists were trained over six 1-h sessions. During the training sessions, the panelists were provided with different commercial and experimental samples. They were asked to examine the samples and describe the appearance, aroma, flavor, texture, and residual taste of the cheeses. Once the panelists had tasted the different samples, they were asked to provide a list of attributes individually covering appearance, aroma, flavor, texture, and residual. A round-table discussion was then conducted to reach an agreement among the panelists. The panelists were also asked to define the attributes. Reference standards were chosen for some attributes and were used to anchor the descriptors. Panelists’ performance was assessed during the training sessions by conducting analyses of variance on their training ratings to check for replicate effect and 2-way interactions (panelist x sample, panelist x replicate, and sample x replicate). In addition, ranking tests were performed and the performance of the panelists was communicated to them during the training sessions. The final list of 17 attributes with definitions, anchor words, and standards is presented in Table 1.
|
The sensory evaluation was performed by rating the intensity of the attributes generated during the training sessions on a 15-cm line scale (Guinard and Cliff, 1987; Guinard et al., 1999). Panelists were compensated with a small monetary sum as compensation for their participation.
Hedonic Evaluation
Eighty-four panelists participated in the acceptability test (mean age = 24.3, range = 18–57, SD = 8.1). The panelists were mainly students, staff, and faculty members at the American University of Beirut. They were selected based on their willingness to participate and time availability. The participants were given a small monetary sum at the end of the study as compensation for their participation.
The panelists assessed the 12 samples in 2 sessions over 2 consecutive days in a similar setting to the one described above. Six different samples were served at each session and the panelists were instructed to rinse their mouths with water between samples. The cheese samples were cut into 1.5-cm3 cubes and each panelist was provided 2 cheese cubes. The samples were presented in 150-mL covered plastic cups coded with 3-digit random numbers. The samples were prepared 1 d before serving and were stored in the refrigerator (4°C) until 15 min before sensory evaluation. The order of presentation of the samples was counterbalanced based on the design for 6 samples as suggested by Macfie and Bratchell (1989).
At the beginning of each session, the panelists were asked to rate their mood on a 7 point labeled categories smiley scale (Chen et al., 1996; Morien et al., 2008) and their hunger on a 15-cm line scale. The panelists were then asked to taste the samples in the order provided and to rate the samples for overall acceptability and for the acceptability of texture on a 9-point hedonic scale (1 = dislike extremely to 9 = like extremely; Peryam and Pilgrim, 1957). The questionnaire also included a food action rating scale (FACT; Schutz, 1964). The FACT scale consists of 9 categories assigned to statements listed from the most positive attitude (category 1, "I would eat this every opportunity I have") to the most negative food attitude (category 9, "I would eat this only if I were forced to").
Statistical Analyses
Analysis of variance using the Mixed procedure of SAS (version 8.02, SAS Institute Inc., Cary, NC) was performed to assess panelist performance (ability to discriminate among the samples, reproducibility, and concept alignment) during panel training, and to assess the significance of the chemical, physical, and sensory differences among the experimental samples in the evaluation sessions. In the statistical model for the descriptive sensory data, the response variable was the sensory attribute of the samples. The factors in the model were milk type (ovine or bovine), fat level (full-fat, reduced-fat, or low-fat), panelist, batch (nested within milk type and fat level), replicate (nested within batch), and their 2-way interactions. Panelist and batch were included as random effects and milk type, fat level, and replicate were fixed effects in the model. Each of the main effects was tested as well as their interactions. Panelist was not included in the chemical or physical analyses models. In addition, the sensory acceptability model included mood and hunger as covariates and did not include replicate. Significant means for the sensory analyses were separated by Tukey’s honestly significant difference test. Significance was established at 
< 0.05. Moreover, principal components analysis was performed to extract the main factors that summarize several sensory attributes, using the 24 means (2 types of milk x 3 fat levels x 2 batches x 2 replicates) obtained from descriptive analysis. These analyses were also performed using the SAS statistical software.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
Low-fat cheese had the highest moisture level whereas full-fat cheese had the lowest level. The ash level was significantly lower (P < 0.05) in full-fat cheese compared with low-fat cheese, probably caused by the moisture and differences. The fat levels for the different samples were consistent with the experimental plan, as shown by the means for fat levels (Table 2). The fat levels for the samples in this work were also consistent with the Cypriot standards for full-fat, reduced-fat (25% fat reduction compared with full-fat), and low-fat (50% fat reduction) cheeses.
The results of the texture analyses are summarized in Table 3. Analyses of variance revealed that the type of milk significantly (P < 0.01) affected the adhesiveness, chewiness, and hardness of the cheese, whereas fat levels significantly affected chewiness, hardness (P < 0.01), adhesiveness, and cohesiveness (P < 0.05). No significant effects were obtained for type of milk on cohesiveness or springiness and the same applied for fat level on springiness. A significant type of milk x fat level interaction was obtained only for chewiness but not for the other variables.
|
Reduced-fat and low-fat cheeses were significantly chewier and harder than full-fat cheese. However, low-fat and reduced-fat cheeses did not differ significantly in the previously mentioned parameters. Papademas et al. (2000) and Theophilou and Wilbey (2007) measured some physical properties of full-fat, low-fat, and reduced-fat Halloumi using a texture profile analyzer and reported a similar increase in chewiness and hardness with a lower fat content. Papademas et al. (2000) noted that differences in chewiness and hardness between full-fat and less-fat or reduced-fat varieties is a reflection of the higher protein content rather than changes in moisture or fat. As for the significant type of milk x fat level interaction for chewiness, the cheeses had the following means for chewiness: 7,507.73 g·mm for reduced-fat ovine, 6,110.09 g·mm for low-fat ovine, 4,271.79 g·mm for low-fat bovine, 4,009.20 g·mm for reduced-fat bovine, 2,111.08 g·mm for full-fat ovine, and 1,958.29 g·mm for full-fat bovine. It is clear from these means that ovine samples were chewier than bovine samples for the same fat level, although this trend was only significant for the reduced-fat level. In addition, reduced-fat cheese was less adhesive than full-fat cheese, but reduced-fat cheese was not significantly different from the 2 other types of cheese for adhesiveness. A decrease in the adhesiveness of Cheddar cheese with a reduction in its fat content was also reported by Bryant et al. (1995). Low-fat cheese was more cohesive than full-fat cheese but reduced-fat cheese was not significantly different from the 2 other types of cheese for cohesiveness. The lack of differences on springiness has been reported in previous work (Papademas et al., 2000)
Descriptive Analysis
ANOVA and Tukey’s Test.
The results of the analyses of variance for the sensory data and the least squares means of all 17 sensory attributes for each of the 3 fat levels (full-fat, low-fat, and reduced-fat) and for both types of milks (ovine and bovine) are summarized in Table 4. Analyses of variance revealed significant differences between the ovine and bovine cheeses on color, hardness using knife, hardness by mouth, and squeakiness (P < 0.05). Differences were also obtained on fat level for color, hardness by mouth, moisture release (P < 0.05), and hardness using knife (P < 0.01).
|
Full-fat cheese was significantly (P < 0.05) whiter, less hard (by knife and by mouth), and more moist than the reduced-fat and low-fat cheeses. Low-fat cheese had higher mean ratings for yellow, hardness using knife, and hardness by mouth, and lower ratings for moisture release than reduced-fat cheese, although these differences were not statistically significant.
Previous studies (Papademas et al., 2000) showed similar trends for color, moisture release, and hardness in full-fat, low-fat, and reduced-fat cheeses. Hardness in low-fat cheeses correlates most closely with protein content but not with fat content (Olson and Johnson, 1990). Full-fat cheese was also found to be more moist than the reduced-fat and low-fat cheeses although moisture content increased as the fat content decreased (Table 2). In fact, cheeses with reduced fat levels are usually described as excessively dry and possibly grainy, because of the greater structural matrix per unit cross-sectional area. (Rodriguez, 1998). The trained panel failed to identify statistically significant differences in flavor between the different samples. Similar results have been reported by Papademas and Robinson (2000) and Milci et al. (2005). Significant differences have been reported in matured cheese as flavor notes tend to increase in intensity at maturity (Papademas and Robinson, 2000). Although the texture differences on hardness were consistent between the instrumental and sensory descriptive measurements, in addition to similar trends for differences between fat levels on cohesiveness and chewiness (though not significantly different for sensory measurements), it is interesting to note that the instrumental assessment of texture revealed more subtle differences than the subjective panelists’ assessment. Texture profile analysis showed significant differences between the different types of cheeses on all of the tested parameters except springiness, whereas panelists’ sensory assessments revealed differences on hardness and squeakiness only. Romeih et al. (2002) reported similar differences between the instrumental texture assessment and panelists’ evaluation. Their instrumental assessment of texture revealed significant differences on hardness, cohesiveness, springiness, and chewiness between full-fat and low-fat cheeses, whereas panelists’ evaluation showed slightly different scores on hardness and crumbliness although these differences were not statistically significant.
There were no significant fat x type of milk interactions. There were no major inconsistencies in the panelists’ ratings, as shown by the absence of replicate effect, with the exception of porosity of surface (P < 0.01) and saltiness (P < 0.05). The fat x replicate interaction was significant for porosity of surface (P < 0.001), hardness by mouth (P < 0.01), and cohesiveness of mass (P < 0.05), whereas a type of milk x replicate interaction was obtained for porosity of surface (P < 0.05) and whey odor (P < 0.05).
Principal Component Analysis.
Figure 1 illustrates the principal components (PC) analysis for the first 2 PC (denoted PC1 and PC2), which accounted for 58.2% (45.1% for PC1 and 13.1% for PC2) of the variation in the sensory attributes’ scores. There are 4 clusters of attributes in the first 2 PC.
|
As for PC2, it separated attributes based on type of milk. The positive side of PC2 (third cluster, upper quadrants) was characterized by attributes that described ovine cheese. Full-fat and reduced-fat ovine cheeses belonged to this cluster. Full-fat, reduced-fat, and low-fat bovine cheese as well as low-fat ovine cheese belonged to the negative side of PC2 (fourth cluster, lower quadrants). The fourth cluster was the opposite of the third cluster, mostly characterized by the absence of the previously mentioned attributes.
Hedonic Evaluation
The least squares means of the acceptability tests are summarized in Table 5. Type of milk had a significant effect (P < 0.01) on overall acceptability, texture acceptability, and FACT. Analysis of variance also revealed that overall acceptability, FACT (P < 0.01), and texture acceptability (P < 0.001) were affected by fat level. There was no significant fat x type of milk interaction.
|
Full-fat cheese received significantly higher scores than the reduced-fat and low-fat varieties on overall acceptability and texture acceptability and lower scores on FACT. There were no significant differences between the low-fat and reduced-fat samples except for texture acceptability. Overall acceptability and texture acceptability decreased with fat level. It is well established that levels of fat in cheese have a direct effect on acceptability (Aziznia et al., 2008; Koca and Metin, 2004). However, results obtained by Papademas et al. (2000) showed that occasional consumers of Halloumi cheese did not show discrimination against the low-fat and reduced-fat varieties and low-fat versions were accepted without complaints. In their study, tasters from England, who were not regular consumers of Halloumi cheese, gave full-fat, less-fat, and reduced-fat Halloumi hedonic sensory scores (%) of 46, 42, and 47, respectively.
Surprisingly, although the trained panelists did not perceive significant textural differences between the low-fat and reduced-fat samples, acceptability scores indicated that reduced-fat cheese scored significantly higher than low-fat cheese on texture acceptability. It is possible that the combination of hardness and moisture release and their additive effect, in addition to other possible texture parameters, affected the texture acceptability of the cheeses of different fat levels. This phenomenon could also be due to the unlikely possibility of the hedonic evaluation panelists noticing a difference on a textural attribute that was missed by the trained panelists of the descriptive analysis.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Sensory evaluation revealed that panelists were sensitive to changes in color and hardness because of significant differences on these attributes but were less perceptive of other textural and flavor changes. Ovine cheese was found to be more yellow and harder than bovine cheese. Full-fat cheese was also found to be significantly whiter and less hard than the reduced-fat and low-fat cheeses. In addition, full-fat cheese was found to be more moist than the other 2 cheese varieties.
Full-fat cheese received significantly higher scores than the reduced-fat and low-fat varieties on overall acceptability and texture acceptability and lower scores on FACT, and panelists showed higher acceptability to bovine cheese.
Moreover, it is advised to change the conventional Halloumi cheese-making procedure to improve the quality of low-fat varieties and reduce their hardness. Commonly proposed procedures are increasing the moisture content or using protein-based and carbohydrate-based fat replacers.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication January 23, 2009. Accepted for publication May 12, 2009.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSFOOTNOTESACKNOWLEDGMENTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
