Effects of Intramuscular Fat Levels on SensoryCharacteristics of Duck Breast Meat

P. Chartrin,*1 K. Meteau,† H. Juin,† M. D. Bernadet,‡ G. Guy,‡ C. Larzul,§ H. Remignon,#J. Mourot,� M. J. Duclos,* and E. Baeza*

*Station de Recherches Avicoles, INRA Tours, 37380 Nouzilly, France; †Unite Experimentale sur l’Elevage Alternatifet la Sante des Monogastriques, INRA Le Magneraud, 17700 Surgeres, France; ‡Unite Experimentale sur les Palmipedes

a Foie Gras, INRA Artigueres, 40280 Benquet, France; §Station de Genetique Quantitative et Appliquee, INRA,Domaine de Vilvert, 78352 Jouy en Josas, France; #ENSAT-INP, 31326 Castanet-Tolosan, France;

and �Unite Systemes d’Elevage Nutrition Animale et Humaine, INRA, 35590 St-Gilles, France

ABSTRACT We conducted a study to evaluate the ef-fects of intramuscular fat levels on the sensory characteris-tics of duck breast meat. Combining duck genotypes(Muscovy, Pekin, and their crossbreed hinny and muleducks) and feeding levels (overfeeding between 12 and14 wk of age vs. ad libitum feeding) enabled us to obtaina wide range of lipid levels in breast muscle. The averagevalues were between 2.55 and 6.40 g per 100 g of muscle.Breast muscle from overfed ducks showed higher lipidand lower water levels than breast muscle from ducksfed ad libitum. Muscle from the overfed ducks was alsopaler in color and exhibited greater yellowness and cook-ing loss values. Juiciness was judged lower and flavormore pronounced in overfed ducks. Muscovy ducks ex-hibited higher breast weight and lower lipid levels thanthe other genotypes. At the other extreme, Pekin ducksexhibited the highest lipid levels and the lowest breastweights; values for these criteria were intermediate in

Key words: duck, genotype, lipid, meat, quality

2006 Poultry Science 85:914–922

INTRODUCTION

Intramuscular fat (IMF) is involved in determiningmeat quality, particularly nutritional and sensory charac-teristics, and conservation ability (Ruiz et al., 2001). Inpoultry meat, it is easy to modulate lipid quality, particu-larly fatty acid profiles, using different lipid sources infeeding diets (Ruiz et al., 2001; Cortinas et al., 2004). Manystudies have analyzed the effects of increased levels ofpolyunsaturated fatty acids on meat quality, mainly fo-cusing on sensory characteristics and consumer accept-ability of cooked meat and processed products and onlipid oxidation during storage of fresh and frozen meat

2006 Poultry Science Association, Inc.Received October 20, 2005.Accepted January 9, 2006.1Corresponding author: chartrin@tours.inra.fr

914

hinny and mule ducks. Breast muscle of Muscovy duckswas paler, less red, and more yellow than that of othergenotypes. Breast muscle of Pekin ducks exhibited thelowest values for lightness, yellowness, and energy neces-sary to shear meat, as well as the highest cooking lossvalues, and was judged more tender, juicy and less stringythan that of other genotypes. In contrast, scores for breastmuscle of Muscovy ducks were the lowest for tenderness,juiciness, and flavor, and the highest for stringiness.Breast muscle of hinny and mule ducks scored the highestvalues for redness. Hinny ducks also scored the highestvalues for flavor. Genotype exerted a higher effect on thesensory quality of breast muscle than did feeding levels.Finally, increasing lipid levels in breast muscle increasedlightness, yellowness, cooking loss, tenderness, and fla-vor, with correlation coefficients of 0.49, 0.47, 0.54, 0.43,and 0.28, respectively. However, breast meat color andtenderness were mainly influenced by genotype.

and processed products. Because lipid levels in poultrymeat are low (about 1 to 2% in breast meat of turkeyand chicken; Rabot, 1998), the relationship between lipidlevels and sensory characteristics has received little atten-tion. However, lipid levels are higher in duck meat thanin chicken and turkey meat (Baeza et al., 2000, 2002).

Different genotypes of ducks are used to produce meat,including common ducks such as Pekin ducks (Anas platy-rhynchos), Muscovy ducks (Cairina moschata), and cross-bred ducks such as mule ducks (crossbreed between acommon female and a Muscovy male; hinny ducks beingthe reverse crossbreed). Overfeeding significantly in-creases lipid levels in duck meat (Zanusso et al., 2003).Using different duck genotypes (Muscovy, Pekin, mule,and hinny) combined with 2 feeding levels (ad libitumvs. overfeeding) Chartrin et al. (2003) were able to obtaina wide range of lipid levels in breast muscle (from 2.26to 7.57%). In the present study, this same model was used

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

LIPID CONTENT AND DUCK MEAT QUALITY 915

Table 1. Composition and main characteristics of feed used during rearing and overfeeding periods

Starting Growing Overfeeding1

Composition (g/kg) (0 to 4 wk) (4 to 12 wk) (12 to 14 wk)

Wheat 200.00 254.50 —Wheat starch — — 2.90Corn 357.02 370.48 988.49Sorghum — 80.00 —Triticale 100.00 — —Extruded soybean seeds 40.00 15.00 —Rapeseed oil meal 30.00 50.00 —Soybean meal 184.75 138.75 —Sunflower meal 29.00 44.50 —Sugarcane molasses 20.00 15.00 —Calcium carbonate 13.50 10.00 —Dicalcium phosphate 17.75 15.00 —Sodium chloride 1.00 1.75 1.45Sodium bicarbonate 2.50 1.50 1.45DL-Methionine 1.88 1.12 —Choline-HCl 75% 0.60 0.40 —Vitamin and mineral supplement 2.00 2.00 5.71Characteristics2 (g/kg)

ME (MJ/kg) 11.83 11.68 13.92CP 175.10 160.00 82.46Total lipids 30.40 27.40 37.36Lysine 9.20 7.80 —Sulfur amino acids 7.70 7.10 —Calcium 11.0 9.00 —Available phosphorus 4.50 4.00 —

1Preparation for overfeeding contained corn (25%), corn meal (35%), and water (40%).2Calculated values.

to analyze the influence of IMF levels on the sensoryattributes of duck breast meat. To avoid sex and ageeffects, we reared only males and slaughtered all birdsat 14 wk of age, which is the usual slaughter age foroverfed birds in France. To complete our study, ducklingswere produced using the same sires and dams to evaluategenetic effects on the sensory characteristics of each geno-type. There are few reports to date on genetic effects onmeat quality parameters in poultry. Le Bihan-Duval et

Table 2. Effects of genotype and feeding levels on weight, color, water, and lipid content in breast muscle from 14-wk-old ducks (mean ± SE)

Color of breast muscle1 Water levels Lipid levels WeightFeeding (g per 100 g (g per 100 g of breast

Genotype level n L* a* b* of meat) of meat) muscle (g)

Overfed 68 42.30 ± 3.23a 13.79 ± 1.32 14.52 ± 2.01a 72.52 ± 1.24b 4.54 ± 1.37a 296 ± 63Ad libitum 77 32.37 ± 2.30b 13.74 ± 1.24 9.37 ± 1.93b 73.84 ± 1.13a 3.00 ± 0.70b 286 ± 73

Feeding level effect P < 0.0001 0.9137 P < 0.0001 P < 0.0001 P < 0.0001 0.9505Pekin 31 34.63 ± 5.85d 13.71 ± 1.17ab 10.13 ± 3.07c 72.39 ± 1.13c 4.81 ± 1.44a 196 ± 19d

Mule 41 36.65 ± 5.50c 14.00 ± 1.08a 11.69 ± 3.23b 73.34 ± 1.05b 3.53 ± 0.98b 303 ± 28b

Hinny 41 37.65 ± 5.71b 13.97 ± 1.60a 12.21 ± 3.56ab 72.99 ± 1.41b 3.80 ± 1.35b 276 ± 21c

Muscovy 32 39.03 ± 5.07a 13.24 ± 0.99b 12.98 ± 2.30a 74.15 ± 1.26a 2.80 ± 0.56c 386 ± 33a

Genotype effect P < 0.0001 0.0363 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001Pekin Overfed 11 41.51 ± 4.30 13.79 ± 0.99 13.63 ± 2.02b 72.08 ± 1.60cd 6.40 ± 1.21a 201 ± 26

Ad libitum 20 30.85 ± 1.35 13.66 ± 1.28 8.20 ± 1.31d 72.56 ± 0.76cd 3.93 ± 0.48c 193 ± 13Mule Overfed 21 41.18 ± 3.65 14.21 ± 1.16 14.33 ± 2.02ab 72.72 ± 1.03c 4.28 ± 0.78c 306 ± 31

Ad libitum 20 31.90 ± 1.81 13.78 ± 0.97 8.93 ± 1.39d 74.00 ± 0.58b 2.75 ± 0.38de 299 ± 24Hinny Overfed 21 42.89 ± 1.58 14.06 ± 1.55 15.03 ± 1.93a 72.05 ± 1.23d 4.86 ± 1.04b 276 ± 20

Ad libitum 20 32.15 ± 1.93 13.88 ± 1.69 9.24 ± 2.16d 73.97 ± 0.75b 2.69 ± 0.39de 277 ± 24Muscovy Overfed 15 43.63 ± 3.07 12.80 ± 0.92 14.73 ± 2.02ab 73.20 ± 0.90c 3.09 ± 0.53d 380 ± 37

Ad libitum 17 34.97 ± 2.05 13.63 ± 0.89 11.43 ± 1.11c 74.99 ± 0.88a 2.55 ± 0.48e 392 ± 30Feeding level

× genotype effect 0.2304 0.1690 0.0190 0.0157 P < 0.0001 0.3689

a–eSignificant difference between groups for a single criterion.1Color: L* = lightness (values between 0 = black and 100 = white), a* = positive values for redness, b* = positive values for yellowness.

al. (2002) only reported heritability and genetic correla-tions for color parameters (lightness, redness, and yel-lowness), pH values at 15 min and 24 h postmortem,and drip loss in chicken and turkey meat. They obtainedmoderate to high heritability values for meat qualitytraits, demonstrating the applicability of a genetic ap-proach to improve meat quality in poultry. Comparingoverfed Pekin, Muscovy, hinny, and mule ducks, Larzulet al. (2002) found significant additive effects in breast

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

CHARTRIN ET AL.916

Figure 1. Frequency distribution of lipid levels (g/100 g of muscle) in breast muscle of 14-wk-old ducks.

meat for postmortem pH values, color parameters, andtoughness after cooking, as well as significant heterosiseffects for color parameters.

The aim of this study was to determine the effects oflipid levels in breast muscle on flavor, color, and tender-ness evaluated by sensory analysis and objective mea-surement of shear force values, cooking loss, and color.Heterosis and maternal effects on these parameters werealso calculated.

MATERIALS AND METHODS

Animals and Rearing Conditions

All experimental procedures with birds were in accor-dance with the French National Guidelines for the careand use of animals for research purposes. Male ducksfrom 4 different genotypes were used: Pekin and Mus-

Table 3. Effects of genotype and feeding levels on cooking loss, shear force and working values in cookedbreast muscle from 14-wk-old ducks (mean ± SE)

Cooking loss ShearingFeeding (% of raw Shear force energy

Genotype level n meat weight) value (N) value (J)

Overfed 68 17.18 ± 2.63a 52.69 ± 10.95 179.44 ± 33.12Ad libitum 77 15.21 ± 1.97b 49.76 ± 10.07 168.44 ± 45.77

Feeding level effect P < 0.0001 0.1892 0.1772Pekin 31 18.38 ± 2.95a 46.88 ± 7.45 150.70 ± 30.47b

Mule 41 15.20 ± 1.68b 52.43 ± 11.98 175.66 ± 38.75a

Hinny 41 15.68 ± 2.18b 53.09 ± 10.89 181.53 ± 47.84a

Muscovy 32 15.73 ± 2.04b 51.07 ± 10.03 182.98 ± 33.98a

Genotype effect P < 0.0001 0.1124 0.0071Pekin Overfed 11 21.08 ± 2.18a 48.24 ± 7.57 155.44 ± 27.39

Ad libitum 20 16.90 ± 2.18b 46.13 ± 7.47 148.10 ± 32.42Mule Overfed 21 15.86 ± 1.52b 51.86 ± 11.97 175.74 ± 29.02

Ad libitum 20 14.51 ± 1.60c 53.04 ± 12.28 175.57 ± 47.69Hinny Overfed 21 16.61 ± 2.31b 56.41 ± 11.54 185.10 ± 34.84

Ad libitum 20 14.70 ± 1.56c 49.60 ± 9.19 177.79 ± 59.26Muscovy Overfed 15 16.98 ± 1.92b 51.90 ± 10.01 194.31 ± 31.92

Ad libitum 17 14.63 ± 1.45c 50.34 ± 10.30 172.99 ± 33.46Feeding level

× genotype effect 0.0171 0.3748 0.7223

a–cSignificant difference between groups for a single criterion.

covy ducks and their crossbred mule and hinny ducks.The ducks (42 per genotype) originated from the samesires and dams and were provided by Grimaud (Roussay,France). They were reared under usual conditions of lightand temperature at the Experimental Station for Water-fowl Breeding (INRA, Artigueres, France). They were fedad libitum from hatching to 6 wk of age. From 6 to 12wk of age, they were fed a restricted diet at levels appro-priate to the ingestion ability of each genotype (200 to250 g/duck at the beginning, increasing to 360 to 380 g/duck at the end of this period). At 12 wk of age, 21 ducksper genotype were overfed for 14 d at the maximum oftheir ingestion potential with corn and corn meal (Char-trin et al., 2003). Contemporaneously, 21 ducks per geno-type were maintained on the growing diet ad libitum. Themain characteristics and composition of diets (starting,growing, and overfeeding) are shown in Table 1.

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

LIPID CONTENT AND DUCK MEAT QUALITY 917

Table 4. Effects of feeding levels and genotype on subjective color of raw breast muscle and sensory attributes of grilled breast muscle from 14-wk-old ducks (mean ± SE)

OverallFeeding color

Genotype level n Tenderness Juiciness Stringiness Overall flavor Duck flavor score1

Overfed 68 4.85 ± 1.22 4.18 ± 1.09b 4.09 ± 1.46 4.68 ± 1.23a 5.12 ± 1.18a 2.61 ± 0.38a

Ad libitum 77 4.90 ± 1.28 4.37 ± 1.08a 4.07 ± 1.48 4.46 ± 1.28b 4.96 ± 1.19b 1.18 ± 0.23b

Feeding level effect NS P = 0.001 NS P = 0.001 P = 0.01 P < 0.0001Pekin 31 5.44 ± 1.24a 4.45 ± 1.02a 3.71 ± 1.40c 4.49 ± 1.27ab 5.03 ± 1.21ab 1.70 ± 0.80b

Mule 41 4.76 ± 1.19b 4.22 ± 1.11b 4.18 ± 1.52ab 4.58 ± 1.24ab 5.03 ± 1.18ab 1.80 ± 0.78b

Hinny 41 4.95 ± 1.21b 4.32 ± 1.01ab 4.05 ± 1.44b 4.70 ± 1.27a 5.16 ± 1.16a 1.95 ± 0.82a

Muscovy 32 4.39 ± 1.17c 4.13 ± 1.19b 4.35 ± 1.46a 4.44 ± 1.27b 4.87 ± 1.21b 1.94 ± 0.69a

Genotype effect P < 0.001 P = 0.001 P < 0.001 P < 0.05 P < 0.05 0.0229Pekin Overfed 11 5.32 ± 1.25ab 4.33 ± 1.04abc 3.78 ± 1.37bc 4.59 ± 1.15ab 5.13 ± 1.11ab 2.67 ± 0.52

Ad libitum 20 5.51 ± 1.23a 4.52 ± 1.01a 3.67 ± 1.42c 4.45 ± 1.33b 4.97 ± 1.26ab 1.17 ± 0.17Mule Overfed 21 4.80 ± 1.17c 4.19 ± 1.14bc 4.19 ± 1.55ab 4.69 ± 1.24ab 5.11 ± 1.17ab 2.48 ± 0.42

Ad libitum 20 4.72 ± 1.21cd 4.26 ± 1.09abc 4.17 ± 1.49ab 4.47 ± 1.23b 4.96 ± 1.17ab 1.07 ± 0.13Hinny Overfed 21 4.99 ± 1.18bc 4.26 ± 1.00abc 3.99 ± 1.40abc 4.83 ± 1.24a 5.25 ± 1.17a 2.71 ± 0.24

Ad libitum 20 4.89 ± 1.24bc 4.39 ± 1.01ab 4.12 ± 1.48ab 4.55 ± 1.28ab 5.07 ± 1.16ab 1.14 ± 0.15Muscovy Overfed 15 4.39 ± 1.17d 3.95 ± 1.14c 4.33 ± 1.47a 4.54 ± 1.26ab 4.93 ± 1.25ab 2.59 ± 0.31

Ad libitum 17 4.38 ± 1.18d 4.28 ± 1.21abc 4.36 ± 1.47a 4.35 ± 1.27b 4.83 ± 1.17b 1.37 ± 0.33Feeding level

× genotype effect P < 0.001 P < 0.001 P < 0.001 P < 0.01 P < 0.05 0.0689

a–dSignificant difference between groups for a single criterion.1Overall color score was obtained by adding each score (1 = dark, 2 = intermediate, 3 = light) weighted with its frequency.

Slaughtering Conditionsand Sample Preparation

Ducks were weighed and slaughtered under commer-cial conditions 12 h after their last meal at 14 wk of age.The right breast muscle was excised and weighed imme-diately after plucking and 2 samples were removed. Onesample was frozen and stored for 1 mo at −20°C beforechemical analysis. The second sample was vacuumpacked, stored at 4°C for 24 h, frozen in ethanol, andstored for 1 mo at −20°C until texture analysis. After 24h of storage at 4°C, the left breast muscle was excised fromthe carcass. Meat color (CIE L*, a*, b*) was determined inthe middle part of the inner side with a Miniscan TMspectrocolorimeter (Hunter Lab, Noisy le Grand, France).This breast was then vacuum packed, frozen, and storedat −20°C until sensory analysis.

Laboratory Measurements

Water content was determined by weighing about 5 gof ground meat before and after storage at 105°C for 24h (AOAC, 1984). Total lipids were extracted by homoge-nizing minced tissue in chloroform:methanol (2:1 vol/vol), and collected by gravimetry (Folch et al., 1957).

Table 5. Correlations between physicochemical characteristics and sensory attributes of breast muscle from 14-wk-old ducks

Overall DuckMeasures Tenderness Juiciness Stringiness flavor flavor

Shearing energy value (J) −0.326*** −0.168* 0.321*** 0.083NS −0.01NS

Cooking loss (% of raw meat weight) 0.211* 0.018NS −0.168* 0.259* 0.188*Lipid levels (g per 100 g of meat) 0.429*** 0.063NS −0.262** 0.257*** 0.279**Water levels (g per 100 g of meat) −0.047NS 0.207* 0.125NS 0.020NS −0.176*Breast weight (g) −0.602*** −0.267** 0.467*** −0.207* −0.219**

*P < 0.05; **P < 0.01; ***P < 0.001.

The samples for texture analysis were thawed at 4°Covernight. They were weighed, individually vacuumpacked, and cooked in a water bath at 80°C to reach anend-point temperature of 75°C in the center of each sam-ple. After cooking, the samples were allowed to equili-brate at room temperature for 1 h. They were then quicklywiped and weighed again to determine cooking loss. Tex-ture was determined objectively by using a single-blade(Warner-Bratzler) shear test performed with a universaltesting machine (MTS Synergie 200, MTS Systems, Ivrysur Seine, France). Adjacent strips (1.0 cm wide) were cutfrom the medial portion of the muscle (parallel to thelongitudinal axis of the muscle fibers), and sheared ac-cording to the method of Honikel (1998).

For each sensory analysis session, 8 breast muscles (1per treatment) were thawed at 4°C for 24 h. They werethen distributed on white plates and covered with trans-parent plastic cling film to judge the color intensity ofthe inner side: darkness, intermediate color, and lightness(scored 1, 2, 3, respectively). An overall score was calcu-lated by adding each score weighted with its frequency.Breast meat was then grilled (2 to 4 min on each sideat 170°C depending on weight), cut into 12 pieces, andpresented to 12 trained panelists. Five sensory attributes

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

CHARTRIN ET AL.918

Table 6. Effects of lipid levels on physicochemical characteristics and sensory attributes of breast muscle from14-wk-old ducks (mean ± SE)

Lipid class (g per 100 g of meat)

1.7 to 3.05 3.06 to 4.37 >4.37 P-value

Number of samples 56 54 35Breast weight (g) 329 ± 58a 275 ± 72b 254 ± 50b <0.0001Lipid level (g per 100 g of meat) 2.58 ± 0.35c 3.71 ± 0.45b 5.56 ± 1.10a <0.0001Water level (g per 100 g of meat) 74.29 ± 0.87a 73.01 ± 0.90b 71.84 ± 1.16c <0.0001Color score1

L* 34.38 ± 3.91c 36.71 ± 6.32b 41.75 ± 3.94a <0.0001a* 13.59 ± 1.11 13.76 ± 1.48 14.05 ± 1.16 0.2367b* 10.42 ± 2.41c 11.64 ± 3.53b 14.21 ± 2.54a <0.0001

Cooking loss (%) 14.95 ± 1.63a 16.29 ± 2.31b 17.77 ± 2.95c <0.0001Shear force value (N) 52.14 ± 11.65 49.64 ± 9.03 51.83 ± 10.94 0.4215Working value (J) 180.64 ± 47.17 169.88 ± 37.18 168.06 ± 32.92 0.2489Tenderness 4.56 ± 0.48b 5.04 ± 0.51a 5.14 ± 0.45a <0.0001Juiciness 4.20 ± 0.38b 4.38 ± 0.35a 4.28 ± 0.39ab 0.0413Stringiness 4.23 ± 0.36a 4.02 ± 0.48b 3.94 ± 0.30b 0.0012Overall flavor 4.41 ± 0.43b 4.64 ± 0.49a 4.72 ± 0.37a 0.0018Duck flavor 4.96 ± 0.32b 5.03 ± 0.33ab 5.15 ± 0.28a 0.0170Overall color score2 1.15 ± 0.53c 1.81 ± 0.77b 2.60 ± 0.54a <0.0001

a–cSignificant difference between groups for a single criterion.1L* = lightness (values between 0 = black and 100 = white), a* = positive values for redness, b* = positive

values for yellowness.2Overall color score was obtained by adding each score (1 = dark, 2 = intermediate, 3 = light) weighted with

its frequency.

were judged on a continuous scale (0 to 10) for juiciness,tenderness, stringiness, duck flavor, and overall flavor.

Statistical Analyses

For sensory analysis, the effects of genotype, overfeed-ing levels, and their interaction were analyzed using theGLM procedure of Minitab software (Minitab Inc., StateCollege, PA). Means were compared using Tukey’s test.For all other criteria measured, we used the GLM proce-dure of SAS (SAS Institute, 1989). The model of varianceanalysis included the main effects of genotype, feedinglevels, and their interaction. Significant differences be-tween means for the different groups were tested ac-cording to the Newman-Keuls test. We also calculatedPearson correlations between physicochemical character-istics and the sensory attributes of the meat. The evalua-tion of genetic effects (maternal, additive, and heterosis)was performed for ducks overfed and for those fed adlibitum according to Dickerson (1969) using the GLMprocedure of SAS (SAS Institute, 1989).

RESULTS

Physicochemical Characteristicsof Breast Muscle

At 14 wk of age, Muscovy ducks exhibited the highestweights of breast muscle (+97% compared with Pekinducks; Table 2), lower lipid levels, and higher water levelsin breast muscle than the other genotypes (Table 2). Thereverse was observed for Pekin ducks, whereas mule andhinny ducks presented intermediate values for these 3criteria. Overfeeding had no effect on breast muscle

weight, but it induced a significant increase in lipid levels(1.51-fold) at the expense of water levels. This increasein lipid levels depended on genotype (1.63-, 1.56-, 1.81-,and 1.21-fold for Pekin, mule, hinny, and Muscovy ducks,respectively). The correlation between water and lipidcontent in breast muscle was −0.78 (P < 0.0001). The fre-quency distribution of lipid levels is presented in Figure1. The minimum and maximum values were 1.72 and8.35, respectively, with values ranging mostly (87%) be-tween 1.72 and 5.04%.

We observed a significant effect of genotype on all colorparameters. Feeding levels influenced lightness (L*) andyellowness (b*, Table 2); the interaction of both factorswas only significant for b*. Lightness and yellowness val-ues were highest and redness (a*) values were lowest inbreast muscle of Muscovy ducks (Table 2). In contrast,L* and b* values were lowest in Pekin ducks. Mule andhinny ducks had the highest a* values. The greatest differ-ences between extremes were in L* and b* values. Breastmuscle of overfed ducks was paler and had higher b*values than that of ducks fed ad libitum (31 and 55%for L* and b*, respectively). The difference in b* valuesobserved between overfed and ad libitum-fed ducks waslower in Muscovy ducks than in the other genotypes.Overfeeding had no effect on a* values. Lightness, red-ness, and yellowness values were significantly correlatedwith lipid levels in breast muscle (+0.49, +0.20 and +0.47,respectively). Lightness and yellowness values were alsosignificantly correlated with breast muscle weight (+0.25and +0.27, respectively). Correlation between redness andbreast muscle weight was not significant (+0.15), but L*and b* values were strongly correlated (+0.87; P < 0.0001).

Cooking loss was greater in overfed ducks than in adlibitum-fed ducks (Table 3). Cooking loss was greater in

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

LIPID CONTENT AND DUCK MEAT QUALITY 919

Table 7. Evaluation of genetic effects (additive, heterosis, and maternal) on physicochemical characteristics andsensory attributes of breast muscle from 14-wk-old ducks considering overfed and ad libitum-fed ducks separatelyand keeping only genotype effects in the model1

Overfed ducks Ad libitum ducks

Parameter Additive Heterosis Maternal Additive Heterosis Maternal

Weight of breast muscle 104.417 *** 0.436 NS −15.071 ** 110.75 *** −4.25 NS −11.125 **Lipid content −1.946 *** −0.174 NS 0.291 * −0.66 *** −0.52 *** −0.029 NSWater content 0.112 NS 0.261 NS −0.025 NS 0.24 NS 0.45 + −0.060 NSLightness (L*) 0.200 NS −0.539 NS 0.856 + 1.94 *** −0.89 * 0.126 NSRedness (a*) −0.421 NS 0.844 ** −0.075 NS −0.05 NS 0.17 NS 0.035 NSYellowness (b*) 0.198 NS 0.501 NS 0.353 NS 1.46 *** −0.74 * 0.155 NSCooking loss −2.429 *** −2.798 *** 0.376 NS −1.23 ** −1.16 ** 0.095 NSShear force value −0.445 NS 4.062 NS 2.278 NS 3.82 + 3.09 NS −1.719 NSWorking value 14.753 + 5.543 NS 4.682 NS 11.34 NS 16.14 NS 1.107 NSTenderness −0.567 *** 0.037 NS 0.100 NS −0.65 *** −0.13 NS 0.083 NSJuiciness −0.243 ** 0.076 NS 0.049 NS −0.18 * −0.08 NS 0.066 NSStringiness 0.364 *** 0.007 NS −0.100 + 0.39 *** 0.11 NS −0.036 NSOverall flavor −0.093 NS 0.202 * 0.071 NS −0.09 NS 0.12 NS 0.038 NSDuck flavor −0.167 * 0.135 + 0.067 NS −0.12 + 0.11 NS 0.049 NSColor score −0.158 + −0.027 NS 0.115 * 0.068 NS −0.16 *** 0.036 NS

1A negative value means that additive and maternal effects were lower for Muscovy ducks than for Pekinducks and that the mean value for heterosis effects was lower for crossbred ducks than the mean value of theparental species.

*P < 0.05; **P < 0.01; ***P < 0.001 (significant effects); NS = nonsignificant; and + = 0.05 > P < 0.10.

Pekin ducks than the other genotypes, particularly afteroverfeeding. Cooking losses were significantly correlatedwith lipid and water levels and breast muscle weight(+0.54, −0.31, and −0.43, respectively).

Neither genotype nor feeding levels affected shear forcevalues (Table 3). Nevertheless, the amount of energy nec-essary to shear the samples was significantly lower forbreast muscle of Pekin ducks than for breast muscle ofother genotypes (−18% compared with Muscovy ducks).Feeding levels had no significant effect on energy values,which were significantly correlated with breast muscleweight (+0.25). Correlations between energy values andlipid and water levels in breast muscle were not sig-nificant.

Sensory Characteristics of Breast Meat

Raw breast meat of overfed ducks was judged palerthan that of ducks fed ad libitum (Table 4), and raw breastmeat of Muscovy and hinny ducks was judged paler thanthat of Pekin and mule ducks. The overall color scorewas significantly correlated with lightness and yel-lowness (+0.92 and +0.84, respectively).

Feeding levels had no effect on the tenderness orstringiness of grilled breast meat (Table 4), but the breastmeat of overfed ducks was judged less juicy and exhibitedmore pronounced duck flavor and overall flavor than thatof ducks fed ad libitum. Genotype had a significant effecton all criteria measured. Breast meat of Muscovy duckshad higher scores for stringiness but it was judged lesstender and less juicy, with less flavor than that of othergenotypes (Table 4). In contrast, the breast meat of Pekinducks obtained the highest scores for juiciness and tender-ness, and the lowest scores for stringiness, particularlyfrom birds fed ad libitum. Breast meat of hinny ducksexhibited the more pronounced flavor, particularly from

overfed birds. Tenderness was positively correlated withlipid levels and cooking loss, and negatively correlatedwith energy values and breast weight (Table 5). Juicinesswas positively correlated with water levels and negativelycorrelated with energy values and breast weight. String-iness was positively correlated with energy value andbreast weight, and negatively correlated with cookingloss and lipid levels. Flavor was positively correlated withcooking loss and lipid levels, and negatively correlatedwith breast weight. Distributing our results into 3 classesaccording to lipid levels in breast muscle confirmed thesecorrelations (Table 6). Significant differences in sensoryattributes were mainly found between the first and secondclasses (1.7 to 3.05 and 3.06 to 4.37% of lipid levels, respec-tively).

Evaluation of Genetic Parameters

The different genetic parameters are presented in Table7. The Muscovy genotype had positive additive effectson breast muscle weight and stringiness, and negativeadditive effects on lipid levels in breast muscle and oncooking loss, tenderness, juiciness, and flavor. Except forcolor parameters (L*, b*, and overall score), there wasgood consistency between effects in overfed and ad libi-tum-fed ducks. Heterosis effects were higher in ducksfed ad libitum, except for a* and flavor, for which effectswere only significant in overfed birds. The sensory attri-butes of breast meat were little influenced by heterosiseffects. Maternal effects were only significant for breastmuscle weight (unfavorable for Muscovy genotype).

DISCUSSION

Combining the effects of genotype and different feed-ing levels enabled us to obtain a wide range of lipid levels

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

CHARTRIN ET AL.920

in breast muscles (from 2.55 to 6.40 g per 100 g of muscle inad libitum-fed Muscovy ducks and overfed Pekin ducks,respectively). The frequency of lipid levels higher than5% was low, which might explain why the correlationsbetween lipid levels and other criteria measured in thisstudy were lower than expected. However, significanteffects of lipid levels on color, cooking loss, tenderness,stringiness, and flavor of breast meat were found.

Meat Color

Breast muscle containing high lipid levels was palerand had greater intensity of yellowness. Objective mea-surement of lightness and yellowness yielded the sameresults as visual scoring for color in raw breast meat.Salichon et al. (1998) reported that breast muscle of ov-erfed mule ducks was paler and had higher b* valuesthan breast muscle of ducks fed ad libitum. Fernandezet al. (2003, 2004) also reported that paler breast muscleof overfed mule ducks and geese contained higher lipidlevels. In general, during overfeeding, ducks ingest highamounts of corn containing liposoluble xanthophylls thataccumulate in tissues and contribute to their color. Otherfactors also influence meat color. Overfed and ad libitum-fed Muscovy ducks had paler breast muscle, higher mus-cle weight and b* values, and lower a* values, despitelower lipid levels than the other genotypes. In this study,it can be hypothesized that the greater muscle develop-ment in Muscovy ducks induced dilution of heminic pig-ments (myoglobin and hemoglobin). This phenomenonhas been reported in breast muscle of broiler chickens(Berri et al., 2001). However, in this study we found nosignificant correlation between a* values and breast mus-cle weight. Fernandez et al. (2003) reported that heavierbreast muscle of overfed mule ducks was also paler andhad higher a* and b* values. Comparing the 4 overfedgenotypes, Larzul et al. (2002) also reported that breastmuscle of Muscovy ducks had the highest L* value. Mus-covy ducks reach maximum growth rate later than theother genotypes (Leclercq, 1990). Because heminic pig-ment levels in duck muscles increase with age (Baeza etal., 2002) it can be hypothesized that these levels mighthave not reached their maximum in Muscovy ducks, un-like other duck genotypes.

Meat Juiciness

Cooking loss was greater in breast muscle containinghigh lipid levels but these criteria were not correlated withmeat juiciness. In fact, breast muscle of overfed Muscovyducks obtained the lowest juiciness scores and that ofPekin ducks fed ad libitum obtained the highest juicinessscores. In our study, measurements of cooking loss andjuiciness scores were not obtained with the same cookingprocedures, which might partly explain the differencesfound between criteria. Girard et al. (1993) also reportedthat breast muscle from ad libitum-fed ducks was judgedjuicier than breast muscle of overfed ducks. In contrast,breast muscle containing greater lipid levels was also

judged juicier in overfed Landaise geese (Baeza et al.,1998a), and high IMF levels were associated with highjuiciness scores in pork (Fernandez et al., 1999). In ourstudy, juiciness was positively correlated with water lev-els in breast muscle, and negatively correlated withweight. Other factors might control meat juiciness.

Meat Tenderness

Tenderness measured by trained panelists was corre-lated with shearing energy value. In our study, highertenderness scores were obtained for breast muscle con-taining higher lipid levels, but overfeeding that induced asignificant increase in breast lipid levels had no significanteffect on this sensory attribute. Girard et al. (1993) alsoreported that breast muscle of ad libitum-fed ducks wasjudged more tender than that of overfed ducks. In con-trast, in overfed Landaise geese, breast muscle containinghigher lipid levels had the highest tenderness scores(Baeza et al., 1998a) and high IMF levels were associatedwith high tenderness scores in pork (Fernandez et al.,1999). In fact, in our study, tenderness was strongly andnegatively correlated with muscle weight, with breastmuscle from Muscovy and Pekin ducks obtaining thelowest and the highest tenderness scores, respectively.Larzul et al. (2002) showed that breast muscle from ov-erfed Muscovy ducks had higher collagen levels (4.82 vs.4.30 mg/g of muscle), and lower collagen solubility (13vs. 19%) than breast muscle from Pekin ducks. Chartrin etal. (2005) reported that the cross-sectional area of musclefibers was greater in breast muscle from Muscovy ducksthan in breast muscle from Pekin ducks (929 vs. 683 �m2

for red oxidoglycolytic fibers and 2,028 vs. 1,608 �m2 forwhite glycolytic fibers); this criterion is generally nega-tively correlated with tenderness (Crouse et al., 1991).Muscovy ducks also had the highest scores for stringiness,confirming a possible effect of fiber size in the determina-tion of these criteria. In this species, Baeza et al. (1998b,1999, 2002) reported that tenderness of breast muscle de-creased with increasing age between 8 and 12 wk, al-though collagen levels decreased, collagen solubility re-mained at the same levels, and fiber size increased.

Meat Flavor

Flavor was more pronounced with high lipid levels inbreast meat. Girard et al. (1993) reported that breast mus-cle from overfed ducks had a more pronounced flavorthan that from ducks fed ad libitum. In overfed Landaisegeese, breast muscle exhibiting greater lipid levels ob-tained the highest flavor scores (Baeza et al., 1998a) andhigh lipid levels were associated with high flavor scoresin pork (Fernandez et al., 1999). The correlation betweenlipid levels and flavor obtained in our study was low butonly IMF were taken into account, whereas under usualcooking conditions, breast muscle retains its skin andsubcutaneous lipids, the levels of which are quite highin overfed ducks. Subcutaneous lipids migrate to poultrymeat during cooking (Rabot, 1998). The lowest flavor

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

LIPID CONTENT AND DUCK MEAT QUALITY 921

scores occurred in Muscovy ducks. This could be ex-plained by lower lipid and phospholipid levels in theirbreast muscle, as previously demonstrated by Chartrin etal. (2003) under similar experimental conditions, becausephospholipids mainly contribute to the aroma of cookedmeat (Mottram and Edwards, 1983; Meynier and Gande-mer, 1994; Gandemer 1997).

Conclusions

Genotype had a greater effect on the sensory qualityof breast muscle than did feeding levels, because thedifferential in lipid levels observed between the extremegenotypes (Muscovy and Pekin) was higher than thedifferential observed between breast muscles of overfedand ad libitum-fed ducks. Moreover, it was difficult todiscriminate between lipid and genotype effects. How-ever, the major changes in scores for sensory attributesrecorded by trained panelists were observed for lipidlevels around 3%. In pork meat, Fernandez et al. (1999)reported that flavor and juiciness were significantly en-hanced when IMF levels increased above approximately2.5%. Barton-Gade and Bejerholm (1985) also reportedthat IMF levels in pork meat had to reach values above2% before any noticeable effects on sensory qualitiescould be detected.

Concerning the genetic determination of criteria mea-sured in our study, we did not confirm the previousadditive and heterosis effects on color parameters de-scribed by Larzul et al. (2002) in breast muscle fromoverfed ducks. In our study, these effects were onlysignificant in breast muscle from ducks fed ad libitum.Cooking loss was also related to additive and heterosiseffects in ad libitum-fed and overfed ducks. In the pres-ent study, we found evidence for additive effects onlipid content in both overfed and ad libitum-fed ducks,as well as a heterosis effect in ad libitum-fed ducks. Thediscrepancy between these results and those of Larzulet al. (2002) may be related to the different choice ofduck strain. Moreover, only additive effects were sig-nificant for tenderness, juiciness, and stringiness, withan unfavorable effect of Muscovy genotype for the first2 criteria, and favorable effect for the third criterion.It could be hypothesized that improvement in sensoryquality of breast meat from overfed mule ducks wouldbe more efficient if selection was undertaken in Muscovyducks. Genetic variance components should thus beevaluated before conducting selection on such criteria.

ACKNOWLEDGMENTS

This study was supported by a grant from the Frenchprofessional organization CIFOG. We thank Grimaud(B. Retailleau and M. Blanchet) for producing and pro-viding the animals. We are very grateful to T. Bordeaufor expert technical assistance, and to the technical staffof INRA Artigueres and Magneraud for rearing andslaughtering birds and performing sensory analysis.

REFERENCES

AOAC. 1984. Official Methods of Analysis. 14th ed. Associationof Official Chemists, Arlington, VA.

Baeza, E., C. Dessay, N. Wacrenier, G. Marche, and A. Listrat.2002. Effect of selection for improved body weight and com-position on muscle and meat characteristics in Muscovyduck. Br. Poult. Sci. 43:560–568.

Baeza, E., G. Guy, M. R. Salichon, D. Rousselot-Pailley, D. Klo-sowska, G. Elminowska-Wenda, M. Srutek, and A. Rosinski.1998a. Influence of feeding systems, extensive vs. intensive,on fatty liver and meat production in geese. Arch. Geflu-gelkd. 62:169–175.

Baeza, E., G. Marche, and N. Wacrenier. 1999. Effect of sex onmuscular development of Muscovy ducks. Reprod. Nutr.Dev. 39:675–682.

Baeza, E., M. R. Salichon, G. Marche, and H. Juin. 1998b. Effectof sex on growth, technological and organoleptic characteris-tics of the Muscovy duck breast muscle. Br. Poult. Sci.39:398–403.

Baeza, E., M. R. Salichon, G. Marche, N. Wacrenier, B. Domin-guez, and J. Culioli. 2000. Effects of age and sex on thestructural, chemical and technological characteristics of muleduck meat. Br. Poult. Sci. 41:300–307.

Barton-Gade, P., and A. C. Bejerholm. 1985. Eating quality inpork. Pig Farming 33:56.

Berri, C., N. Wacrenier, N. Millet, and E. Lebihan-Duval. 2001.Effect of selection for improved body composition on muscleand meat characteristics of broilers from experimental andcommercial lines. Poult. Sci. 80:833–838.

Chartrin, P., M. D. Bernadet, G. Guy, J. Mourot, M. J. Duclos,and E. Baeza. 2005. Effect of genotype and overfeeding onlipid deposition in myofibers and intramuscular adipocytesof breast and thigh muscles of ducks. Reprod. Nutr. Dev.45:87–99.

Chartrin, P., J. Mourot, M. D. Bernadet, G. Guy, M. J. Duclos,and E. Baeza. 2003. Effect of genotype and force feeding onthe intramuscular fat deposition in duck. Pages 224–230 inProc. 16th Eur. Symp. Qual. Poult. Meat, Saint-Brieuc, France.ISPAIA, Ploufragan, France.

Cortinas, L., C. Villaverde, J. Galobart, M. D. Baucells, R. Co-dony, and A. C. Barroeta. 2004. Fatty acid content in chickenthigh and breast as affected by dietary polyunsaturation lev-els. Poult. Sci. 83:1155–1164.

Crouse, J. D., M. Koohamaraie, and S. D. Seideman. 1991. Therelationships of muscle fibre size to tenderness of beef. MeatSci. 30:295–302.

Dickerson, G. 1969. Experimental approaches in utilizing breedresources. Anim. Breed. Abstr. 37:191–202.

Fernandez, X., A. Auvergne, M. Renerre, P. Gatellier, H. Manse,and R. Babile. 2003. Preliminary observations on the colourvariability of breast meat (“magrets”) in force-fed ducks.Anim. Res. 52:567–574.

Fernandez, X., G. Monin, A. Talmant, J. Mourot, and B. Lebret.1999. Influence of intramuscular fat content on the qualityof pig meat. 1. Composition of lipid fraction and sensorycharacteristics of m. longissimus lumborum. Meat Sci.53:59–65.

Fernandez, X., S. Leprettre, J. P. Dubois, and R. Babile. 2004.Observations preliminaires sur la variabilite de la couleurdes magrets d’oies. Pages 199–202 in the Proc. 6emes Journeesde la Recherche sur les Palmipedes a Foie Gras, Arcachon,France. ITAVI, Paris, France.

Folch, J., M. Lees, and G. H. Sloane Stanley. 1957. A simplemethod for the isolation and purification of total lipids fromanimal tissues. J. Biol. Chem. 226:497–509.

Gandemer, G. 1997. Lipides du muscle et qualite de la viande.Phospholipides et flaveur. OCL -Ol. Corps Gras Lipides41:19–25.

Girard, J. P., J. Culioli, C. Denoyer, J. L. Berdague, and C. Toura-ille. 1993. Discrimination de deux populations chez deux

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from

CHARTRIN ET AL.922

especes de volaille sur la base de leur composition en lipides.Arch. Geflugelkd. 57:9–15.

Honikel, K. O. 1998. Reference methods for the assessment ofphysical characteristics of meat. Meat Sci. 49:447–457.

Larzul, C., B. Imbert, M. D. Bernadet, G. Guy, and H. Remignon.2002. Qualite du magret dans un croisement factoriel Barb-arie X INRA44. Pages 29–32 in Proc. 5emes Journees de laRecherche sur les Palmipedes a Foie Gras, Pau, France.ITAVI, Paris, France.

Le Bihan-Duval, E., C. Berri, E. Baeza, M. Duclos, V. Sante, H.Remignon, G. Le Pottier, J. Bentley, and X. Fernandez. 2002.Selection on the technological quality of the meat in poultry.Pages 11–14 in Proc. 7th World Congr. Genet. Appl. Livest.Prod., Montpellier, France. INRA, Paris, France.

Leclercq, B. 1990. Croissance et composition corporelle du ca-nard de Barbarie. Pages 23–39 in Le canard de Barbarie. B.Sauveur and H. De Carville, ed. INRA, Paris, France.

Meynier, A., and G. Gandemer. 1994. La flaveur des viandescuites: Relations avec l’oxydation des phospholipides. Vian-des et Produits Carnes 15:179–182.

Mottram, D. S., and R. A. Edwards. 1983. The role of triglyceridesand phospho-lipids in the aroma of cooked beef. J. Sci. FoodAgric. 34:517–522.

Rabot, C. 1998. Vitesse de croissance et caracteristiques lip-idiques et sensorielles des muscles de poulet. PhD Thesis.Inst. Natl. Agron. Paris–Grignon, France.

Ruiz, J. A., L. Guerrero, J. Arnau, M. D. Guardia, and E. Esteve-Garcia. 2001. Descriptive sensory analysis of meat from broil-ers fed diets containing vitamin E or β-carotene as antioxi-dants and different supplemental fats. Poult. Sci. 80:976–982.

Salichon, M. R., G. Guy, D. Rousselot-Pailley, N. Wacrenier, J.C. Blum, and E. Baeza. 1998. Infuence d’une supplementationen vitamine sur la qualite de la viande (filet et magret) etdu foie gras chez le canard mulard. Pages 127–131 in Proc.3emes Journees de la Recherche sur les Palmipedes a FoieGras, Bordeaux, France. ITAVI, Paris, France.

SAS Institute. 1989. SAS/STAT User’s Guide. SAS Institute Inc.,Cary, NC.

Zanusso, J., H. Remignon, G. Guy, H. Manse, and R. Babile.2003. The effects of overfeeding on myofiber characteristicsand metabolic traits of the breast muscle in Muscovy ducks(Caırina moschata). Reprod. Nutr. Dev. 43:105–115.

at National C

hung Hsing U

niversity Library on A

pril 10, 2014http://ps.oxfordjournals.org/

Dow

nloaded from