Influence of production systems and sex on nutritional value and meat quality of native Malawian Muscovy ducks

This study was conducted to assess effects of production systems and sex on nutritional value and meat quality of native Malawian Muscovy ducks. One hundred twenty ducks were randomly assigned to either an intensive (IS), duck-rice integration (DR) or free-range (FR) production system. A starter ration containing 20% crude protein and a finisher containing 17% crude protein were fed to ducks in IS (1 to 4 weeks), and provided as a supplement to ducks in DR and FR (5 to 10 weeks). Feed and water were offered ad libitum. At 10 weeks of age, 16 ducks per treatment were selected randomly, slaughtered and chilled at 4 °C for 24 hours. Carcass temperature, pH and meat colour were measured at 45 min, and at 3, 6, 12 and 24 hours post mortem. Tenderness, cooking loss, proximate and mineral composition were determined 24 hours post mortem. Production system and sex had no effect on carcass temperature, pH and proximate composition of duck breast meat. However, production system affected tenderness and mineral composition of the meat and sex influenced moisture and tenderness. Males were moister and had less tender meat than females. Carcasses from ducks in the FR system contained more zinc, copper, manganese, and potassium, but less iron while those in IS had the lowest mineral content of the three production systems. Thus, DR can be adopted to improve the current FR system of native Malawian Muscovy duck production with supplementation to produce duck with acceptable mineral composition and better meat quality.


Introduction
The duck population is increasing in Malawi with about 2 390 304 ducks, which represent about 2.1% of the total poultry population (Department of Animal Health and Livestock Development, 2017). Duck meat is the third most widely produced poultry meat in the world after chicken and turkey (Uhlířov et al., 2018). Duck meat comprises 2900 tonnes of meat production in Malawi, which represents 0.4% of its total meat production (Department of Animal Health and Livestock Development, 2017). The most common production system of ducks is the traditional free-range system (FAO, 2008;Gondwe, 2004) in which ducks are mostly kept together with indigenous chickens with limited provision of supplementary feed, especially during periods of feed shortage and reduced scavenging (FAO, 2016). It is estimated that the consumption of meat in Malawi was 8.3 kg per person per year in 2013, which is below the 10 kg norm (FAO, 2013). This level of meat consumption is considered insufficient, since such low intake often leads to under-nourishment and malnutrition. Meat from conventional livestock species such as goats, cattle, pigs, sheep and chicken is not sufficient, and ducks could be an alternative since they are hardier and more tolerant of viral diseases than chickens, survive in free-range environments, and contribute significantly to the food security and protein intake of people in rural communities (FAO, 2017).
According to Ramadhan et al. (2010), the diets fed to ducks affect their chemical composition. Duck meat has low cholesterol and high unsaturated fatty acid, and is recognised as a healthy alternative source of animal protein, which is gaining popularity (Hong et al., 2014). Nowadays, consumers are health conscious and prefer food and meat products that are low in fat, cholesterol and caloric content (Yang et al.,  At the end of the experiment, feed was withdrawn 12 hours before slaughter (Kwon et al., 2014) and 48 ducks (16 from each treatment) were selected randomly, slaughtered and kept in the chiller at 4 °C for 24 hrs for further analysis such as cooking loss, tenderness, and proximate and mineral composition. Duck breasts (Pectoralis muscle) were analysed for meat quality attributes since the breast is one of the crucial parts for the culinary value and technological properties of duck meat (Witkiewicz, 2000;Wołoszyn et al., 2002;Galal et al., 2011), the other part being the thigh muscle.
Carcass pH and temperature profiles were measured by a pH meter (Kim et al., 2012) at 45 minutes, 3 hours, 6 hours, 12 hours, and 24 hours post mortem and recorded in tripplicate (Ronald et al., 2006). The pH meter, which also measured carcass temperature, was dipped into the pectoralis muscle to at least a depth of 2 cm at a similar place in each pectoralis muscle.
Meat colour was determined by means of a Konica Minolta CR400 chroma meter (Mussah & Phoya, 2017), whereby 30 x 30 mm samples from the pectoralis muscle were prepared, labelled and used to capture colour scores of lightness (L*), redness (a*), and yellowness (b*). Three random readings were taken from each meat sample from different angles.
Meat tenderness was determined by an Instron universal testing machine (TA.XT plus texture analyser). The machine was calibrated to have a crosshead speed of 100 mm/min at the full-scale load of 50 kg (500 N). From each pectoralis muscle, a sample of 5 × 40 mm (ca. 2.0 cm 2 ) cross section was cut for shear force measurement (Ali et al., 2007). The meat samples were placed at right angles to a plunger-type blade. Nine samples from each pectoralis muscle were tested and readings recorded. Data were collected and analysed based on the shear force values to obtain the maximum force required to shear through each muscle sample.
Cooking loss was determined by cutting and weighing 20 g of the pectoralis muscle. The meat samples were put in the plastic tubes and boiled in a water bath to an internal temperature of 90 °C for 30 minutes. After boiling, the samples were surface dried, and weighed. Cooking loss was determined by expressing weight of cooked sample (B) as a percentage of the precooked sample (A): ( Khan et al., 2016) Proximate composition was determined after removing the skin and external fat from the breast (pectoralis major) meats according to procedures of the AOAC International (2005). Total fat was determined by the AOCS standard procedure Am 5-04 (Ankom, 2001) in which the Ankom XT10 extraction system was used. Micro-minerals that were analysed in the duck meat were potassium (K), zinc (Zn), copper (Cu), iron (Fe), and manganese (Mn).The analyses of micro-mineral content were conducted at the Agricultural Research and Extension Trust (ARET) Laboratories in Malawi. Five grams of defatted and dry samples, which were ashed at 550 °C for six hours, were solubilized in concentrated nitric acid. Then, five ml of 6 molar hydrochloric acid were added to the samples before they were assayed with a Varian atomic absorption spectrometer (Spectra AA20, Agilent Technologies, Australia). Any silica compounds that were present were removed by filtration. The filtrate was then diluted to the desired volume and aspirated into the air-acetylene flame of the atomic absorption spectrometer. The absorbance of each element was measured by comparison with the absorbance of calibration solutions for the same element. The adjusted wavelengths were copper (Cu): 324.8 nm, iron (Fe): 248.3 nm, manganese (Mn): 279.5 nm, zinc (Zn): 213.8 nm, and potassium (K): 766.5 nm.
Data were analysed using the general linear model of IBM SPSS software version 22 (IBM Inc., Armonk, New York, USA). Multivariate analysis of variance was used to test the effects of production system, sex and the interactions among these factors. Differences between treatment means were tested using Duncan's multiple range test (P <0.05). Each duck was regarded as an experimental unit and the statistical significance was set at P-value < 0.05. Means and standard errors (SE) are reported.

Results and Discussion
There were no significant differences (P >0.05) in the conversion of duck breast muscle to meat over the 24-hour post-mortem period, although numerically the carcass pH of ducks in IS decreased faster than those in other treatments (Figure 1), but retained similar values after 24 hours. The same trend was observed for carcass temperature (Figure 2). Pooled carcass temperature dropped from 27.45 °C to 5.47 °C and pooled carcass pH dropped from as high as 6.68 to 5.86, with the highest drop at six hours post mortem for both quality factors. Similarly, the study found no significant differences between production systems and sex for the initial (45 minutes) and final (24 hrs) carcass temperatures and pH of Muscovy duck meat (Table  2). Generally, no interactions between production systems and sex were found, except on ultimate pH.   Female 6.42 ± 0.14 6.57 ± 0.14 6.80 ± 0.14 6.60 ± 0.79 Male 6.55 ± 0.14 6.60 ± 0.14 6.59 ± 0.14 6.58 ± 0.79 1 P: production system, S: sex, P*S: interaction between production system and sex Muscle pH is important in the preservation and stability of meat, because high muscle pH results in shorter shelf life, especially as it pertains to microbial growth (Li et al., 2017). [not in refs] Muscle pH depends on the biochemical state of the muscle at the time of slaughter and the development of rigor mortis (Singh & Pathak, 2017). Post-mortem muscle pH decline is important in terms of meat tenderness, colour, and waterholding capacity {WHC) (Aberle et al., 2001). which depends mainly on a decrease in pH and on protein degradation under the action of proteases (Baeza, 2006). It is a direct indication of meat quality in which glycogen reserves in muscles affect the post-mortem pH level. Duck muscle is regarded as red meat because it has 16% white fibres and 84% red fibres, compared with 100% white fibres in chicken breast (Smith et al., 1993), thus differences post mortem are expected (Ali et al., 2007). Generally, the mean ultimate pH value of duck meat is between 5.90 (Fernandez et al., 2003) and 5.65 (Raj et al., 1998). The current results are in agreement with those of Li et al. (2017), in which production system had no effect on pH level of breast meat and the lack of interaction between production systems and sex is similar to what Wang et al. (2009) andFanatico et al. (2007) reported on ducks in a free-range system versus those in an intensive system. However, the values for ultimate pH in this study were higher than those reported by Erisir et al. (2009) and lower than reports by Ismed et al. (2009) and Khan et al. (2016). On the contrary, Raj et al. (1998), Mallia et al. (2000), Fernandez et al. (2003), and Lacin et al. (2008) reported that treatment had an effect on pH values, and found comparable ultimate pH values for ducks in IS environments, which were lower than FR and duck-fish rearing systems. The differences could be due to age of slaughter, which causes glycolytic metabolism in muscle to increase with age, inducing a decrease in ultimate pH value (Baéza et al., 1998) and glycogen levels during slaughter, which can be altered by many factors ante mortem and post mortem (Braden, 2013).
Lack of difference in pH between sexes in this study is similar to the findings of Liao et al. (2016) andUhlířov et al. (2018) in Pekin ducks and Geldenhuys et al. (2015) in Egyptian geese. However, Kirmizibayrak et al. (2011) reported significantly higher ultimate pH of the breast muscles in females than in males. In the current results at 24 hours post mortem, the pH values on breast meat between sex showed female ducks were slightly lower (5.74) than males (5.84). It is reported that lower pH (5.7) at 24 hours post mortem indicates poor meat quality (Van Laack et al., 2000). All the production systems presented overall pH24 values between 5.72 and 5.87, an indication of good meat quality. The higher ultimate pH above 5.7 at 24 hours in this study could be due to the fasting period and exposure to high temperatures before slaughter (Baéza, 1995). However, the pooled ultimate pH in this study was lower than that reported by Liao et al. (2016). Differences in pH values could be attributed to differences in the glycogen reserves at slaughter, responses to pre-slaughter stress and slaughter weight (Uhlířov et al., 2018).
The colour of the pectoralis muscle was affected by production system only in terms of lightness (L*) at 45 min, and by the sex of the duck on redness (a*) at 24 hours (P <0.5). Interactions between production systems and sex were observed on lightness (L*) and yellowness (b*) at 45 min (Table 3). Generally, the colour of pectoralis muscle at 24 hours post mortem was not affected by production systems. x,y Column means with different superscripts differ significantly at P <0.05 1 P: production system, S: sex, P*S: interaction between production system and sex Meat colour is the most important sensory attribute that affects consumer purchasing decisions, and is affected predominantly by myoglobin, which is the primary red pigment in meat (del Puerto et al., 2016) and influences consumer preferences (Pripwai et al., 2014;Pathare & Roskilly, 2016), since it indicates meat freshness and the level of doneness (Chang et al., 2011). Fibre type and myoglobin content have a strong effect on meat colour. Like other eating quality attributes, meat colour is affected by factors such as species, animal genetics, rearing systems, post-mortem changes in muscle (especially the dynamics of pH and carcass temperature decline). Poultry meat colour is also affected by sex, age, processing method, exposure to chemicals; cooking method, irradiation and freezing (Huda et al. 2009 found a* = 13.86 and b* = 12.18. However, ducks from FR and DR integration had brighter pectoralis muscle than those from IS. This finding is supported by Ponte et al. (2008), who reported that birds from a freerange system were characterized by a brighter colour of meat and a higher contribution of yellow colour in muscles, owing to their diet, which included green forage rich in natural carotenoids. These green forages cause the deposition of melanin and xanthophyll into the skin and give its colour. Ducks on free range and duck-rice integration were more active than those in the intensive system, thus they used more oxygen in their muscles, which requires more myoglobin and a greater supply of oxygen from blood to the muscle cells. Because the iron content of duck meat is high, the meat from free range and duck-rice integration appeared darker than those from the intensive system. In the previous discussion, it was observed that duck meat pH from duck-rice and free range had slightly higher pH and some studies reported a high correlation between ultimate muscle pH and meat colour, especially lightness. Fletcher et al. (2000) reported that muscles with a high pH have a darker colour than those with a low pH. In the current study, meat from duck-rice and freerange systems, which had numerically higher pH values, was darker than that from the intensive system. Hence, rearing systems affect the physical parameters of muscles and may significantly affect consumers' choice and preferences (Fanatico et al., 2007). On the other hand, sex had a significant effect on lightness and redness and females had higher values of redness (a*) than males (P <0.05). No differences between females and males were observed for lightness and yellowness (P >0.05). The results in muscle redness and yellowness at 24 hours from this study between the sexes agree with those of Uhlířov et al. (2018) in geese.
The effects of production system and sex of ducks on tenderness and cooking losses of duck meat are presented in Table 4. The results confirmed that the tenderness of duck breast meat was affected by both production system and sex (P <0.05), while cooking loss was not influenced by either of these factors. As indicated by shear force, ducks kept in IS had more tender breast meat (1.56 kg), followed by ducks in DR integration (1.95 kg), and the least tender meat was from ducks in the FR system (2.28 kg). However, further comparisons showed that tenderness of duck breast for ducks in DR was not different from that of ducks in the FR system. In addition, there was an interaction between production system and sex on tenderness of duck breast meat (P =0.009) x,y Column means with different superscripts differ significantly at P <0.05 1 P: production system, S: sex, P*S: interaction between production system and sex Meat palatability is the primary sensory trait of consumer acceptance (Choi et al., 2016), and the main attribute of palatability that influences consumer acceptance is tenderness (Mennecke et al., 2007;Piao et al., 2015), especially for consumer decisions on repeat purchases (Grunert et al., 2004). It is dictated by the presence of several factors, including the amount of fat, water-holding capacity (WHC) and the actomyosin complex (Singh & Pathak, 2017) and processing techniques (Adam & Abugroun, 2015). The results from the current study on the effect of production systems on tenderness are in line with those of Michalczuk et al. (2014), who reported that meat from chickens on a FR system was characterised by higher shear force, and therefore was tougher than the meat of chickens from IS. In their studies Chen et al. (2006) andFarmer et al. (1997) reported significant production system effects, with higher shear force values of drumstick and breast meat of chickens raised organically, which is similar to the FR system in the current study. However, it was not surprising to have duck meat from IS that was more tender than meat from DR and FR systems, since ducks under IS had restricted space allowances in the house, so the muscles were not under heavy exercise as a result of walking longer distances (Mutibvu et al., 2017). However, shear force values in this study were lower than those that Ali et al. (2007) reported for Chungdong ori ducks. Numerically, meat from ducks in DR was more tender than that from the FR system, because ducks in the DR had some restricted movement in the rice field, which was fenced, as opposed to ducks in FR, where there was no limitation of movement, resulting in greater muscle activity. The effects of production systems on carcass and meat quality, and in particular meat tenderness have been confirmed in other livestock (Webb & Erasmus, 2013). The results of the effect of sex on meat tenderness in this study are in line with those of Chen et al. (2006), who found higher shear force values in males than females in two chicken breeds (Anka and Rugao). However, the results are contrary to the findings by Smith et al. (2015), who did not find differences in tenderness between male and female Pekin ducks, and to those of Uhlířov et al. (2018) in geese and to Huff-Lonergan et al. (1995) in the bovine longissimus muscle. The use of Muscovy ducks in the present study may explain the difference between sexes.
Cooking loss is a critical factor in the meat industry because it determines the technological yield of the cooking process (Kondjoyan, 2013) and serves as an indicator of WHC, which affects meat quality. From a nutritional point of view, cooking loss brings about loss of soluble proteins, vitamins and different supplements (Yarmand et al., 2013) and depends on cooking time and cooking temperatures (Adam & Abugroun, 2015). In this study, cooking loss was not affected by production system or sex, which agrees with the findings of other studies (Lopez et al., 2011;Uhlířov et al., 2018). The values in the present study were lower than those reported by Alvarado and Sams (2000) and Ali et al. (2007). However, Musundire et al. (2017) reported that cooking loss was affected (P <0.05) by sex in chickens and guinea fowl managed under village FR systems. The differences are due to species, age at slaughter, and conditions in the management system.
The nutritional composition of breast muscle from Muscovy ducks is presented in Table 5 and shows that production system had no effect on any of the nutrients that were analysed. However, sex had an effect (P <0.05) on moisture, with males containing a higher percentage of moisture than females. The proximate composition in breast was not affected by interaction between production system and sex. The nutritional composition of duck meat varies, depending on the breed and strain of duck and diet (Ramadhan et al., 2010). The present results agree with those of Holcman et al. (2003) and Michalczuk et al. (2014), who found no significant differences in the chemical composition of muscle due to rearing system. In their studies, Holcman et al. (2003) and Lorenzo et al. (2011) reported that sex did not influence the proximate analysis of breast meat, which may indicate similar skeletal muscle composition between sexes, though not in moisture for this study, which was influenced by sex.
The overall moisture content of duck meat in the current study was lower than that reported by Chartin et al. (2006), Ali et al. (2007, Lorenzo et al. (2011), Galal et al. (2011 and Ismed et al. (2013) in Muscovy ducks, and Baeza et al. (2000) in mule ducks. These differences could be due to age at slaughter, which affects the proximate composition and especially moisture content (Wattanachant, 2008) as it decreases with age. The effect of sex on moisture, whereby males had significantly higher values than females in this study, is comparable with findings by Baeza et al. (2000) and Zangh et al. (2010).
Levels of protein found here in duck meat are in line with studies by Ali et al. (2007), who found 20.1% crude protein (CP), Ensminger et al. (1993), who found 21.4% CP, Rahman (2014), who found 20. 3% CP, and Cobos et al. (2000), who found 20.08% CP. However, the current results are lower than values reported by Lorenzo et al. (2011), who recorded a protein level of 24.29% CP. On the other hand, the current protein value is higher than the 18.3% reported by the USDA (2015), and the 17.9% reported by Ismed et al. (2013). These differences could be as a result of age at slaughter, breed and management system, since the other authors did not specify the breeds of duck. Age at slaughter also affects the protein content of meat, whereby an increase in age results in a decrease in protein content.
The fat content of meat is affected by breed, feeding and age at slaughter. In free-ranging Guinea fowls and chickens fat has been observed to increase with age and females have more fat than males (Musundire et al., 2017). The pooled fat composition found in this study, ranging from 4.18 to 5.37%, is in agreement with results of Chartin et al. (2006), Ismed et al. (2013) andUSDA (2015), but higher than the results of Mazanowski et al. (2003), who reported fat content of 1.7% in specialized sire lines. Though not statistically different numerically, it was observed that ducks kept in free range and duck-rice integration were slightly fatter that those in intensive systems. This may be due to the wide range of feedstuffs that are available to the ducks in the vicinity. Feedstuffs such as grass, insects, frogs and small animals that are eaten by the scavenging ducks could have contributed to their high fat and ash contents, since grass is digested by the ducks in the colon and thereby supplies energy, which is usually stored as fat (Tougan et al., 2003). However, the tendency for birds to deposit fat is genetically driven, indicating that fatness and leanness have to be modified at breeding level for sustainable improvement (Grashorn, 2010). On the other hand, the relationship between fat and moisture in this study, whereby the pooled moisture content for males was slightly higher than for females and affected the fat content negatively, agrees with works of Stankov et al (2002), Chowdhury and Mutalib (2003) and Rahman (2012), who reported that the decrease in moisture content in meat is due to the increase in fat content.
While the ash content of duck breast meat in this study was higher than values reported by Zhou et al. (2000) and Baeza et al. (2002), the current results on ash content are in line with those of Ali et al. (2007), Ismed et al. (2013) andUSDA (2015). The significantly high ash of the ducks on free range might be attributed to the season of production. The study was conducted at the start of the rainy season in Malawi, when the grasses were young and more nutritious, and had a relatively high mineral value. These grasses are liked by scavenging birds, and contribute to the high mineral level. The season is also characterized by snails, worms, and small insects, which contain high levels of minerals, and might have contributed to this high ash content (Tougan et al., 2003).
The minerals of the duck breast muscle were affected by the production system (P <0.05) except for potassium. No effects of sex or interactions between sex and production systems were observed on the minerals of duck breast muscle (Table 6). The general trend in this study was that free-range ducks had the highest values of all minerals, followed by those in duck-rice integration, and the intensively managed ducks had the lowest minerals. However, differences between production systems were observed for manganese, zinc and copper (P <0.05).
Minerals are present in the animal body in concentrations not greater than 50 mg/kg (0.05 mg/g) (McDonald et al., 2010). Ash is an indication of the total amount of minerals in the muscle and the quantity can vary as a result of species differences, hormones, age, sex, region and diet (Keeton & Eddy, 2004). An increase in mineral levels in the current study was as a result of higher ash values (Table 5), which agrees with what Musundire at al. (2017) reported. The high mineral levels of the ducks in the FR and DR systems than ducks in the intensive system are explained by the high values of ash in Table 5, which are attributed to availability of feed resources for scavenging birds, which influence their ash content (Tougan et al., 2013), and consequently the high micro-minerals. a,b,c Row means with different superscripts differ significantly at P <0.05 1 P: production system, S: sex, P*S: interaction between production system and sex

Conclusion
The duck-rice integrated system is a new farming system in Malawi, but has proved to be a useful alternative alongside free-range and intensive systems. The results from this study indicate that the use of duck-rice integration improves the nutritive value of duck meat to levels that comparable with free-range systems because of the freedom of birds to scavenge. The use of the duck-rice integration system produced more tender breast meat than from free-range systems, but less so than from intensive systems. The study also concurred with other studies on the effect of sex on meat tenderness, whereby females produced more tender duck breast meat than males. On the other hand, production systems had no effect on the proximate composition of Muscovy duck breast meat, while sex affected only the moisture content, with males containing more moisture than females. Therefore, farmers could adopt the duck-rice integrated system as an alternative to the current free-range system with supplementation to improve the nutritional value and meat quality of native Malawian Muscovy ducks. However, further on-farm studies need to be conducted before implementation on a full scale, since this was an on-station study with most of the factors under control.