La teoría del desarrollo endógeno

The breast portion of each bird was placed inside in an oven bag (Glad®). No salt (NaCl) or any other seasoning was added to any of the meat treatments throughout the sensory analyses. The oven bags and meat samples were then placed on stainless steel grids which were fitted on an oven roasting pan. Thermocouple probes attached to a handheld digital temperature monitor (Hanna Instruments, South Africa) were placed in the centre of each of the meat samples (AMSA, 1995). The prepared samples were then placed in two conventional ovens (Defy, Model 835), pre-heated to 160

°C (AMSA, 1995). The ovens were connected to a computerized monitoring system responsible for regulation of the temperature (Viljoen et al., 2001). The meat samples were removed from the oven when a core temperature of 75 °C was reached (AMSA, 1995). The samples were cooled for 15 min where after they were cut into 1 cm x 1 cm cubes, individually wrapped in aluminium foil and placed into glass ramekins coded with randomized three-digit codes. The coded ramekins, each containing two wrapped meat cubes, were then placed in a preheated industrial oven (Hobart, France) at 100 °C for 10 min after which they were removed and immediately served to the sensory panel for analysis.

2.3 Descriptive sensory analysis

Descriptive sensory analysis (DSA) was performed on the four meat treatments (2 genders x 2 periods). A panel of eight judges, based upon previous experience with sensory analysis of meat, was selected. The panellists were trained according to the guidelines for sensory analysis of meat by the American Meat Science Association (AMSA, 1995) and the generic descriptive sensory analysis technique as described by Lawless and Heymann (2010).

The panel undertook four training sessions and during each of these training sessions the panellists received 1 cm x 1 cm cubes of meat from four reference standards, as well as the four meat treatments. Reference standards were chosen to illustrate the respective aroma and flavour attributes associated with Egyptian geese, as well as the other five treatments. The reference standards included commercial free range chicken, beef sirloin, beef rump, as well as the Longissimus thorasicus et lomborum muscle of locally harvested blesbok (Damaliscus pygargus phillipsi - a free ranging wild ungulate) (Geldenhuys et al., 2014) [Chapter 4]. The reference samples

enabled the panellists to calibrate their sensory perception during the training sessions, thereby allowing them to recognise and score all of the attributes tested in the respective meat samples.

During the training phase the panel used the descriptors obtained from Geldenhuys et al.

(2014). These included the game, chicken, ostrich and beef aromas and flavours, as well as metallic flavour, initial and sustained juiciness, tenderness (first bite and residue). The panel also decided to add fish and sweet-oily-duck aromas and flavours to the final list of sensory descriptors. The definition for each of the attributes is described in Table 2. The test re-test method was used for DSA.

The panellists received the four treatments in a complete randomized order, while seated in individual tasting booths fitted with the software programme Compusense® five (Compusense, Guelph, Canada). The samples were analysed for the respective sensory attributes using an unstructured line scale anchored to zero (indicating “low intensity”) and 100 (indicating “high intensity”) (AMSA, 1995).

The sensory analysis sessions took place inside a temperature-controlled (21 °C) and light-controlled (artificial daylight) room (AMSA, 1995). In order to cleanse and refresh their palates between samples, the panellists received distilled water (21 °C), apple quarters and water biscuits (Carr, UK).

Table 2 Definition and scale of each attribute used for the descriptive sensory analysis

Sensory attribute Description Scale

Game aroma¹ Aroma associated with game meat experienced as soon as the aluminium foil is removed

0 = Extremely bland 100 = Extremely intense Chicken aroma¹ Aroma associated with chicken experienced as

soon as the aluminium foil is removed

0 = Extremely bland 100 = Extremely intense Ostrich aroma¹ Aroma associated with ostrich experienced as

soon as the aluminium foil is removed

0 = Extremely bland 100 = Extremely intense Beef aroma¹ Aroma associated with beef experienced as soon

as the aluminium foil is removed

0 = Extremely bland 100 = Extremely intense Fish aroma¹ Aroma associated with fish experienced as soon

as the aluminium foil is removed

Game flavour¹ Flavour associated with game meat prior to swallowing

0 = Extremely bland 100 = Extremely intense Chicken flavour¹ Flavour associated with chicken prior to

swallowing

0 = Extremely bland 100 = Extremely intense Ostrich flavour¹ Flavour associated with ostrich prior to swallowing 0 = Extremely bland

100 = Extremely intense Beef flavour¹ Flavour associated with beef prior to swallowing 0 = Extremely bland

100 = Extremely intense Fish flavour¹ Flavour associated with fish prior to swallowing 0 = Extremely bland

100 = Extremely intense Metallic flavour¹ Flavour associated with metal/liver prior to

swallowing

0 = Extremely bland 100 = Extremely intense Initial juiciness The amount of fluid exuded from the cut surface

when pressed between the thumb and forefinger

0 = Extremely dry 100 = Extremely juicy Sustained juiciness The level of juiciness perceived after the first 5

chews using the molar teeth

0 = Extremely dry 100 = Extremely juicy First bite The impression of tenderness perceived after the

first 5 chews using the molar teeth

0 = Extremely tough 100 = Extremely tender Residue The amount of residue left inside the mouth after

the first 10 chews

0 = None 100 = Abundant

1 Aroma and flavour were analysed orthonasally and retronasally, respectively.

2.4 Physical measurements 2.4.1 pH

The pH of the four meat treatments, of each replication, was measured after thawing for 36 h, immediately after removal from the packaging and before the start of the cooking process of every sensory analysis session. The pH was measured by means of a Crison pH25 handheld portable pH meter (Lasec (Pty) Ltd, South Africa) calibrated before each set of readings with the standard buffers (pH 4.0 and pH 7.0) provided by the manufacturer.

2.4.2 Cooking loss

The cooking loss of the four meat treatments from each of the six replications were determined according to the method described by AMSA (1995). The weight (Radwag PS 750/C/2, Lasec SA, Cape Town, South Africa) of each treatment, as recorded before the cooking process, was used as the initial, raw weight of the samples. Following the cooking process the cooked meat samples were removed from the cooking bag and allowed to equilibrate to ambient temperature (±15 min) where after the samples were blotted dry with blotting paper and weighed. The difference in the weight of each of the samples was calculated as the percentage of cooking loss.

2.4.3 Water holding capacity

The water holding capacity was determined according to the method described by Trout (1988). A cooked meat sample from each of the four treatments and six replications were used. The sample was cut finely and 0.50 g weighed off and placed on top of a filter paper (Lasec, Paper Filter, grade 292, diameter 90 mm, part nr. FLAS3205090). The filter paper containing the meat sample was placed between two Perspex plates at a standard pressure of 588 N for 60 s. after which a photograph was taken of the filter paper showing the expelled liquid and meat areas. Using Image J Software (Version 1.41, 2009, http://rsbweb.nih.gov/ij/) the ratio between the outer (liquid) and inner (meat) purge area was calculated to indicate the water holding capacity of the 0.50 g meat sample.

2.4.4 Shear force

The shear force test (SF), as described by Honikel (1998), was used to analyse the instrumental shear force of the cooked meat samples. Each of the four treatments (six replications) was evaluated for instrumental tenderness. Two adjacent 1 x 1 cm meat strips were cut parallel to the muscle fibre direction from the centre of the cooked meat samples, wrapped in aluminium foil and placed in the refrigerator (4 °C) for 24 h. The meat strips were then cut to produce a total of five rectangular cubes each with a length of two centimetres and stored at 4 °C until the testing commenced. An Instron Universal Testing Machine (UTM) (Model 2519-107, Advanced Laboratory Solutions, United

Kingdom) attached with a Warner-Bratzler (WB) fitting was used in order to determine the force necessary to shear the cooked rectangular meat cube perpendicular to the muscle fibre direction.

The WB fitting was a 1 mm thick triangular (V-notch) blade with a semi-circular cutting edge (radius of 0.508 mm). The UTM operated with a 2 KN compression load and compression extension of 20 mm.

The shear test was performed at a compression speed of 200 mm/min. The shear force value of each of the samples was recorded in Newton (N).

2.5 Chemical analysis 2.5.1 Sample preparation

The proximate analysis was performed on the uncooked, left breast muscle of the geese used for the six replications of the descriptive sensory analysis. After the breast muscles were thawed (4 °C) homogenisation of the muscles followed. The samples were re-vacuum packed and frozen (-18 °C) until the proximate analysis commenced when the samples were thawed (4 °C) once more.

2.5.2 Proximate analysis

The proximate analyses were performed as described in Geldenhuys et al. (2013c) [Chapter 3]. The moisture content (%) was determined by the use of 2.5 g homogenised meat sample according to the Association of Official Analytical Chemist’s Standard Techniques (AOAC) method 934.01 (AOAC, 2002a). The ash content (%) of the moisture free sample was determined by the official AOAC method 942.05 (AOAC, 2002b). The chloroform/methanol (1:2 v/v) extraction method stipulated by Lee et al. (1996) was used to determine the total lipid (%) (intramuscular fat) of a 5 g homogenised raw meat sample. To establish the total crude protein content (%) the Dumas combustion method 992.15 (AOAC, 2002c) was applied. A 0.15 g defatted, dried and finely grounded meat sample was analysed using a Leco Nitrogen/Protein Analyser (FP – 528, Leco Corporation). The Leco was calibrated with EDTA calibration samples (Leco corporation, 3000 Lakeview Avenue, St. Joseph, MI 49085-2396, USA, Part no. 502-092, Lot no. 1055) before each of the analysis sessions. The results were expressed in % Nitrogen (N). The nitrogen (%) was multiplied with a conversion factor (6.25) in order to determine the crude protein (%) present in the meat sample. All proximate analyses are controlled by a National inter-laboratory scheme (AgriLASA: Agricultural Laboratory Association of South Africa) where blind samples are analysed once every three months in order to control and ensure the accuracy and repeatability of the procedures used.

2.6 Statistical analysis

Experimentally, the study consisted of a randomised factorial block design with 4 treatments (2 seasons X 2 genders) and six replications. The study tested the effects of season and gender, as well as the interaction between the main effects. The trained panel consisted of eight judges and the six treatments were evaluated for the 17 sensory attributes established during the training sessions.

The model for the experimental design is indicated by the following equation:

уij = μ + βj + si + gk + (sg)ik + εijk

The terms within the model are defined as; the overall mean (μ), the effect of the block (βj), the effect of season (si), the effect of gender (gk), the effect of the interaction between season and gender (sg)ik and εijk is the error associated with the effect of the block, season, gender and interaction of the former and latter.

The sensory, physical and proximate data were subjected to an analysis of variance (ANOVA).

The Shapiro-Wilk test was performed to test for normality (Shapiro & Wilk, 1965). All of the outliers were identified and removed before final analysis of the ANOVA’s. The t-Least Significant Differences (LSD) were calculated at a 5% significance level to compare the treatment means. Results were defined as being not significant at P>0.05 and significant at P≤0.05. Correlations were made between the sensory attributes, physical characteristics and proximate composition by means of the Pearson’s correlation coefficient. Principal Component Analysis (PCA) using a correlation matrix and Discriminant Analyses (DA) were performed to illustrate the relationships between the sensory, physical and proximate data (Naes et al., 2010). SAS™ statistical software (Statistical Analysis System, Version 9.2, 2006, SAS Institute Inc., Cary, NC, USA) was used for the analyses of variance (ANOVA) while the multivariate statistical analyses were performed using XL STAT™ statistical software (Version 2011, Addinsoft, New York, USA).