Multivariate Analysis of the Morphological Traits of Female Duck, Muscovy-duck and Mule-duck

The objective of this study was to evaluate the characteristics of morphometrical measurements inthe female waterfowls. The animals used in this research were 90 ducks , 90 muscovy-ducks and 90mule-ducks in Bulukumba district of Brebes regency, Central Java, Indonesia. Parameters measuredwere maxilla length, neck length, body length, chest circumstance, wing length, chest length, femurlength and tibia length. The data were analyzed using the Statistical Analysis System ver. 9.1. Muscovyduckgenerally had the largest of size, followed by mule-duck and then duck. The most discriminantvariables were showed by chest length and chest circumstance. Muscovy-duck and mule-duck hadclosest genetic distance (3.974870) than both of the distance between duck and mule duck (14.10), andmuscovy-duck and duck (24.73). The smallest errorness level in grouping was showed in duck 1%followed by 2% in mule-duck and 3% in muscovy-duck.


INTRODUCTION
Consumer demand for livestock products increase from year to year, in which those demand is not yet met by the production. This situation is supported by increased public awareness of the importance of animal protein needs. Duck (Anas plathyrhynchos), muscovy-duck (Cairina moschata) and mule-duck are the sources of animal protein which are very potential to be developed and are expected to meet the needs of poultry meat.
Duck, muscovy-duck and mule-duck are waterfowl and as a source of eggs and meat product. Indonesia has a variety of local ducks that are spread almost in all parts of Indonesia which adjusted its name with the name of the area and geography. According to Brahmantiyo et al. (2003), Indonesia has many kinds of local ducks such as Mojosari-duck , Tegal-duck , Bali-duck and Alabio-ducks, that each has advantages. Muscovy-duck or Manila-duck has a body size larger than ducks . Muscovy-duck has a horizontal body shape and good nature (Srigandono, 1997;Widodo and Purnama, 2004;and Sutiyono, 2012). Mule-duck is a hybrid of the duck with muscovyduck that has higher growth rate than broiler (Widodo and Purnama, 2004;Marie-Etancelin, 2008;and Sutiyono, 2012).
Morphometric is a morphological measure-ments on the body length of certain bones in living things. Body size is a good indicator of value and has a fairly close correlation with the live weight trait (Suparyanto et al., 2004). Morphometric measurements can also help to facilitate the selection and crossing between breeds and types of livestock (Sumantri et al., 2007;Setiaji et al., 2012;Kurnianto et al., 2013). The knowledge about the body sizes in ducks, muscovy-duck, and mule-duck is very important, therefore the research on it needs to be done. The purposes of this study were to evaluate morphometric characteristics of females duck, muscovy-duck, and mule-duck. The benefits of this research are to be used as baseline information for the females' development of duck , muscovy-duck and mule-duck as a source of meat and eggs by the farmer community.

Materials and Data Collection
A total of 270 birds of the females sex comprising 90 ducks, 90 muscovy-ducks and 90 mule-ducks were used in this study. These animals were reared by farmers under the traditional system at Bulukumba District of Brebes Regency in Central Java, Indonesia. Purposive sampling method was applied to determine the location based on population density of the waterfowl breeds.
The data were collected by measuring from each part of the females' body of duck, muscovyduck and mule-duck. The following traits that were measured in all animals were Maxillary length (ML), Neck length (NL), Body length (BL), Chest circumference (CC), Wing length (WL), Chest length (CL), Femur length (FL) and Tibia length (TL).

Data analysis
The data obtained were compiled and analyzed using the Statistical Analysis System program 9.1 (SAS, 1990). Univariate analysis covering the proc means and the General Linear Model (GLM) followed by a multivariate analysis including Canonical Discriminant Analysis (Candisc) were used to analyze and to compute squared Mahalonobis distance. Phenogram illustrating distance among breeds was constructed by UPMA (Unweighted Pair Group Method with Arithmetic Mean) of MEGA-5 (Tamura et al., 2011). Principle Component Analysis (PCA) is a data reduction technique to examine the modes of variation of a multivariate random variable in high dimension. Discriminant Component Analysis (DISCRIM) procedure of SAS (1990) was performed to determine percentage assignment of individuals into their own population.
The longest chest circumference was found in muscovy-duck (34.649±1.801) cm, followed by mule-duck (32.406±5.835) cm, and duck (30.783±2.435) cm. The longest wing was found in muscovy-duck (29.113±2.710) cm, followed by mule-duck (28.257±2.091) cm, and duck (26.477±1.665) cm. Wing length of muscovyduck is longer than the mule-duck and duck, because the body size and body weight of muscovy-duck are greater. In accordance with Supriyanto (2003) that increasing the length of the wing is followed by an increase in the body size which indirectly will increase the body weight. The sizes of these variables showed a significant difference (P<0.05).
Body length in the muscovy-duck (25.181±1.514) cm did not show significant difference from mule-duck (25.223±1.378) cm but significantly different from body length of duck (24.472±1.257) cm. The femur length of muleduck (7.543±2.244) cm was not significantly different from muscovy-duck (7.299±0.944) cm, but significantly different from femur length of duck (6.301±0.623) cm. The length of the tibia in duck (10.189±1.092) cm was not significantly different from the mule-duck (10.403±0.924) cm but significantly different from muscovy-duck (9.673±1.411) cm. According to Brahmantiyo et al. (2003) distinguishing variable on body morphology of the ducks that the most influential was the length of the femur and tibia. Table 2 shows eigenvectors of PC1 and PC2 for female waterfowl breeds. In morphometric application of PCAs, PC1 was acceptable as a "size" vector and PC2 as a "shape" vector as reported in livestock (Sumantri et al., 2007;Setiaji et al., 2012 andKurnianto et al., 2013) and waterfowl (Widodo and Purnama, 2004;Marie-Etancelin, 2008;and Sutiyono, 2012). Based on the results of principal component analysis note that the chest circumference showed a large positive number and followed by the length of femur, wing and chest. The chest circumference on Principal Component 1 (0.893771) provide a large positive number and on Principal Component 2 (-0.320260) has a relatively large negative value, so the chest circumference can be used as a differentiating variables for the body sizes and body shapes among waterfowl breeds. This result is a correlation or a combination of all the variables measured on the body sizes from the maxilla length to the tibia length.

Distinguishing on Female Waterfowl Breeds
According to Ismoyowati et al. ( 2006) the chest circumference is one of the characteristics that are closely related to meat production, so layer duck has the chest circumference that is relatively smaller. While the length variables of the maxilla, neck, wing and tibia can not be used as a differentiating variable between duck, muscovy-duck and mule-duck. The assumption is based on the results of the analysis of PC1 and PC2 which shows a negative value.

Grouping Map on Female Waterfowl Breeds
The results of principal component analysis are shown in Figure 1. It can be seen a dividing line among the duck, muscovy-duck and muleduck observed, in which PC1 and PC2 indicated the body size and the body shape . The grouping of the muscovy-duck was seen in the positive both in PC1 and PC2, it indicated that the size and shape of muscovy-duck has a great body . In   Srigandono (1997) that muscovyduck has a large body with almost standing position (horizontal ) which is one of the traits of poultry meat. It was stated by Widodo and Purnama (2004) and Sutiyono (2012 ) that the muscovy-duck was poultry meat. Grouping the duck was seen in the negative PC1 and positive PC2, it can be interpreted that the ducks have small phenotypic size, but the body shape is quite large. Grouping the mule-duck was seen positive in PC1 and in PC2 was negative, it can be interpreted that the mule-duck has a fairly large body size but small body shape.

Genetic Distance among Female Waterfowl Breeds
Based on the analysis of genetic distances between duck, muscovy-duck, and mule-duck (Table 3), the genetic distance between the duck and muscovy-duck was 24.73030, the duck and mule-duck was 14.10275, and for muscovy-duck and mule-duck was 3.974870. Phylogeny tree is presented in Figure 2. According to Wu et al. (2008), based on genetic distance of Peking duck, Cherry valley and Aobaixing located in one cluster, while muscovy-duck is beyond the cluster. It was stated by Su and Chen (2009) that the genetic distance is influenced by the breeds and the level of gene heterozygosity. Sutiyono (2012) reported the closeness of the genetic distance between muscovy-duck and mule-duck, because both display the same type of waterfowl, that is the body tends to flat or horizontal. Futhermore, Marie-Etancelin (2008) stated that genetic distance between duck and muscovy-duck is due to the size and shape of the body, in which the duck is greater than the muscovy-duck . Ducks have a body shape like bottles upright or vertical and slender (Srigandono, 1997), while Sopiyana et al. (2006) and Raji et al., 2009) stated that Tegal-duck has body characteristic like a bottle upright, not horizontal like muscovy-duck Data of the body size can be used to estimate the genetic distance (Brahmantiyo et al., 2003;Muzani et al., 2005;Sumantri et al., 2007) using discriminant analysis through statistical Mahalanobis distance .

Grouping of Errornes Level
Grouping error rate (errorness level) in duck, muscovy-duck and mule-duck can be seen in Table 4. The smallest error rate is in the grouping of ducks (1%) , followed by mule-duck (2%) and muscovy-duck (3%). This suggests that the presence of misclassification rate of morphometric size resemble duck like mule-duck is about 1%, the rate of misclassification morphometric size resembling mule-duck like muscovy-duck is 2%, and a misclassification rate of morphometric size resembling muscovy-duck like mule-duck is 3%. Observation on duck, muscovy-duck and mule-duck indicated the small  possibility of a mixture of other breeds because the level of error in grouping was relatively small and a large percentage of similarity in the types of ducks, muscovy-duck and mule-duck. According to Brahmantiyo et al. (2003) and Muzani et al. (2005), the great similarities in the body characterics show the small mixture of characteristics with other breeds. Indonesian native ducks have common characteristics that are the body in a standing position looks like a bottle, slim, having small head, and small neck (Srigandono, 1997;Muzani et al., 2005;Johari et al., 2013). Poultry research center has developed the performance of local duck by crossing with other breeds to get the ideal duck with superior quality and good production for egg and meat (Brahmantiyo et al., 2003;Supriyanto et al., 2003;Sulaiman and Rahmatullah, 2011;Johari et al., 2013).

CONCLUSION
Muscovy-duck generally has the largest average characteristics of body size, followed by mule-duck, and then duck. The chest circumference, femur length, wing length and chest length are good differentiator variables used to differentiate females body morphometric of waterfowl breeds. Duck has small phenotypic size and enough large of shape, muscovy-duck has the size and shape of a large body, while mule-duck has a fairly large body size but small body shape. The genetic distance between muscovy-duck and mule-duck are closer compared to duck and muleduck, also duck and muscovy-duck. The smallest error rate in grouping was duck than followed by mule-duck and muscovy-duck.