INTRODUCTION

In the Poultry Industry, ducks have the peculiar feature to provide a range of final products that goes from the meat and egg production, feathers for ornamental purposes, liver with fat for production of pates to many other products for a market increasingly on the rise, but little explored in Latin America (Avicultura industrial, 2005). In Brazil, duck meat consumption is restricted to 13 grams/habitant/year. In China, for example, this consumption is 1.5 kg/habitant/year and in Europe remains at 1 kg/habitant/ year. Duck meat is also very consumed in the United States and in countries with Arabic ethnicity, such as Egypt and Saudi Arabia (^{Wawro et al., 2004}; Avicultura Industrial, 2005).

There are few researches regarding on how to manage ducks, most of them are related to the nutritional aspect, and some have suggested only technical recommendations. For example, the current recommendation (^{Coates & Ernst, 2000}) suggested that the density of ducklings in the first two-weeks should be ½ m2/bird, and increased up to at least 1 m2/bird in the first four-weeks, and if birds remain in the housing after one month, give them at least 2 m2/bird. ^{Moreira (1993}) recommended setting the free-range stocking of 1 bird/m² when their weight reaches 3.0 to 3.5 kg (females) and 4.5 to 5.5 kg (males) with slaughter age ranging from 3 to 6 months. And together with the density, the physiological and ethological features should be taken into consideration. ^{O’Driscoll & Broom (2011}) reported that more water was needed in the housing ambient for improvement of the duck`s health aspects.

Moreover, the quality of poultry meat has become increasingly important, since sensory features as appearance and meat tenderness are required by the consumer (^{Beraquet, 1999}), and these are strongly related to all of the the phases of bird management.

The nutritional factor is important not only in raising ducks system, but also for all poultry production. The feed cost is one of the limiting factors in animal production, and only in the poultry sector, it represents approximately 70% of the total production cost (^{Cruz, 2016}).

According to ^{Togashi (2000}), the poultry production in some regions of Brazil is limited due to low grain availability. This means that the nutritional study is very essential for poultry production to reduce the feed cost and, consequently, the cost of production. The increasing demand for a better feed control in poultry, added to the high cost and increase of bird consumption in the world market, are factors that motivated researchers to seek alternatives for a conventional concept used in modern poultry management system, especially in poultry feed.

Considering the above, the present study aimed to evaluate the performance, carcass traits and economic availability of ducks on different phases of nutritional plans and housing densities.

MATERIAL AND METHODS

The study was conducted in the facility of Poultry Sector, Department of Animal and Plant Production (DPAV), Faculty of Agrarian Sciences (FCA), Federal University of Amazonas (UFAM), located in the south sector of the university campus, Manaus, in the State of Amazonas, Brazil.

Two hundred and forty Muscovy ducks (Cairina Moscharadomesticus) of creole lineage were used distributed in boxes with water and food ad libitum. The experimental design was completely randomized in a 3x2 factorial design with three nutritional plans (P1 with 3 phases (1-35 days; 36-70 days and 71-90 days), P2 with 4 phases (1-28 days; 29-49 days, 50-72 days and 73-90 days) and P3 with 5 phases (1-14 days; 14-28 days, 29-63 days, 64-76 days and 77-90 days)), and two housing densities (2 birds/m² and 3 birds/m²) with 4 replicates (8 ducks in boxes with 2 birds/m² and 12 ducks in boxes with 3 birds/m²).

The experimental diets were formulated according the production stages and nutritional plans (Table 1), according to the nutritional requirements and reference values for broilers (^{Rostagno et al., 2011}) adapted to ducks.

The birds started the experimental period with 1-day of age and were evaluated at 90 days. Weekly, in the experimental phase, the birds were weighed to obtain the performance variables. The feed intake was determined by the quotient between the total feed intake and the quantity of poultry. The weight gain was determined by the total weight of each plot divided by the number of birds plot, and feed conversion was determined by relation between the amount of feed consumed and the weight gain in addition to considering the final weight of the experiment.

At 90 days of age, after 12 hours of fasting, four ducks of each treatment were randomly selected, identified and weighed. Next, these were electrically stunned (40 V; 50 Hz), with the birds slaughtered by a cut in the jugular vein. The carcasses were immersed into hot water (60ºC for 62s), plucked and eviscerated according Mendes & Patricio`s (2004) recommendations, and the carcass yield was determined in relation to live weight. Edible viscera (heart, gizzard, pro-ventricle and liver) were separated and individually weighed to measure the yields.

In economic analysis, the fixed costs consisted of labor, facilities and equipment depreciation (^{Martins et al., 2006}). The feed cost was considered only as the variable cost. For analysis of the production cost per kilogram of meat the feed intake was considered and the production per treatment. The feed cost considered the nutritional plans cost per kg (P1 = US$ 0.35/kg; P2 = US$ 0.44/kg; e P3 = US$ 0.52/kg). For live weight of duck in Manaus/AM considering a cost of US$ 3.71/kg (according to the current value of the dollar in R$ 2,69). Crude Income (CI) and Operating Profit (OP) were used with economic indicators according to Martin et al. (1998).

Data were submitted to analysis of variance and means compared by Tukey test at 5% of significance using the statistical program ^{SAS (2008}).

RESULTS AND DISCUSSION

The results of performance of ducks are show in Table 2. Differences were observed in feed intake, weight gain and feed conversion (p<0.05), among nutritional plans and feed intake and weight gain (p<0.05) between housing densities. However, no interaction (p>0.05) between nutritional plans and housing densities could be observed. Nutritional plans with three phases and higher densities showed a positive influence on duck growth, with a direct relationship between the lower feed intake and reduction on feed conversion, with similar results observed by ^{Feijó et al. (2016}).

^{Graças et al. (1990}), affirms that a reduction in feed intake can be caused by increased in housing density, with less physical space, the birds have difficult access to feeders for these. However, according to the results of ^{Garcia et al. (2002}) and ^{Cruz et al. (2013}), increasing housing density can promote greater results of meat/m², therefore, an alternative to increase the productive and economic performance of poultry production. Moreover, to formulate an ideal diet, that is, to present maximum performance and economic results, it’s necessary to deeply understand the nutritional poultry requirements (^{Trindade Neto et al., 2009}), as proposed in this study, from the results that demonstrated the better nutritional energy-protein relations for ducks.

In the poultry industry, the cost of feeding represented 75% of total production costs and, with 40% to 45% of protein composed this cost (^{Sakomura & Silva, 1998}). In general, the formulation of poultry diets in thermo neutral environment seeks to meet the requirements of crude protein (CP), metabolizable energy (ME), vitamins and minerals. However, this may contain excess of essential amino acids (^{Cella, 2001}), and there may be potential unbalances in feed composition.

The results for the carcass traits are shown in Table 3. Differences weren’t observed for carcass yield, feathers, legs and fat (p>0.05) between nutritional plans, housing densities and interaction. It was observed that nutritional plans with more phases, regardless of the density used, showed better results for carcass traits, with similar results observed by ^{Lisboa et al. (1999}), ^{Figueiredo et al. (1999}), ^{Araújo et al. (1999}), Takahashi (2006) and ^{Santos et al. (2012}) that studied different commercial lineages of broilers and didn’t observed significant differences in relationship of management and carcass traits.

Researchers affirm that the management is directly related to the carcass results for broilers, but, that doesn’t mean that it will always influence the carcass results, cuts or visceras of the birds. ^{Hellmeister Filho et al. (2004}) didn`t observe differences in carcass traits of free-range broilers when compared to management with or without access to picket. <sup>Almeida & Zuber (2000</sup>) who also study free-range broilers didn`t observe the effect of management system on carcass traits.

The results of edible visceras are show in Table 4. Differences weren’t observed for liver, gizzard, pro-ventricle and heart weights (p>0.05) between nutritional plans, housing densities and interaction. However, it was observed that lower densities provided better numerical results of edible viscera (gizzard, liver and heart), and didn’t negatively affect these carcass traits.

Despite of peculiar characteristics between species, slow-growing broilers tend to have characteristics very similar to those observed in the management systems currently used for ducks. But, it is very important to highlight the poor literature about the management systems for ducks, with appropriate technical and informative recommendations.

There’s need to work on a management of animals that can meet the growing commercial interest that exists in duck meat, that can be obtained by producing birds with slow development and management in higher input systems, with the objective of attending the niche market made up for a range of consumers, according to the affirmatives of ^{Lewis et al. (1997}) and ^{Carrijo et al. (2002}).

The results of economic analysis are shown in Table 5. Differences were observed for total production cost and operating profit (p<0.05) among nutritional plans, and for total meat production, crude income and operating profit (p<0.05) between housing densities. However, no interaction (p>0.05) between nutritional plans and housing densities could be observed for economic analysis.

Nutritional plans with reduced phases showed better results in the evaluation of economic performance of ducks, that show how feeding has a significant influence on the financial aspect in poultry production, mainly in the cutting segment, where food cost becomes extremely essential to the development of birds.

The total meat production linear growth with increase of housing density, for example, significantly influenced the meat production by square meter, with similar results observed by ^{Goldflus et al. (1997}), who talks about the direct influence of housing densities on performance, carcass traits and financial poultry production.

CONCLUSIONS

In summary, nutritional plans with reduced phases and extensions of energy-protein relationships showed better results for ducks on performance and carcass traits in densities of 3 birds/m². More studies are necessary to determine other nutritional requirements for ducks in housing for a better nutritional and management control.