The efficient production of agricultural resources in tropical and subtropical areas, and the necessity to find, in a sustainable way, alternative sources for animal feeding are conditions that facilitate the use of silages in animal feeding (Guzmán 2010).By means of a simple and appropriate procedure, silages allows to preserve citrus residues (Llano et al. 2008), pineapple (Herrera et al. 2009) and mango (Rego et al. 2010),between other wastes, useful for animals intake.

In Ecuador, according to Caicedo et al. (2013a) there are a wide variety of feasible resources to feeding pigs, between them it is the taro (Colocasia esculenta (L.) Schott) .Tubers are recognized as a cheap carbohydrates source of low cost, regarding to cereals and other crops. They also present, high starch digestibility, that can reach up to 98% (Ezedinma 1987).

The use of in vivo conventional methods to measure of foods digestibility and ingredients that are used in diets formulation is too expensive and requires long periods to obtain evaluative results. Consequently, have been developed in vitro methods that have the advantage of being simple, fast and allows evaluating a great number of samples at the same time and at low cost (Boisen and Fernández 1995 and Ramos 1995).

The objective of this study was to evaluate the chemical composition and in vitro digestibility of silages of taro (Colocasia esculenta (L.) Schott) tubers for feeding pigs.

To formulate the silos, Caicedo et al. (2013b) recommendations were fallowed (table1).The micro - silos were prepared with tubers waste that, for its physical appearance, they do not fulfill the requirements established by national and international markets to use them in human feeding. Tubers were washed and milled in fresh way, in mix mill with blades and a sieve of 2.5cm, with the purpose of obtaining uniform particles.

For silages formulation, raw matters were weighed in a CAMRY digital balance, and were placed in four clean plastic tanks, with 400kg each. Ingredients were added in the fallowing order: silage1) cut tubers, natural yogurt and drinkable water for human consumption (1_NY) 2) cut tubers and whey (2_W); 3) cut tubers, molasses B (83º Brix) 5% and whey (3_MB5) and cut tubers, molasses B (83º Brix) 10 % and whey (4_MB10).

The ingredients were mixed in a homogeneous way, manually, with a wooden spatula, during15 min, at room temperature of 24 ºC. Then, the silages mixtures were introduced in polyethylene bags, at a rate of 5kg.The bags were compacted and closed hermetically with a vacuum pump to guarantee their preservation. The micro-silos were stored under roof and protected of the sun light. They were opened at 0, 15, 30, 60, 90 and 180 d after silage.

The material was analyzed for DM, CF, ash, CP (N x 6.25) content, EE and NFE, according to AOAC (2005) .The levels of NDF, ADF and lignin were measured according to van Soest et al. (1991).It was considered that the OM concentration was equal to the difference 100- ash percent. However, the hemicellulose was the result of the subtraction NDF-ADF, as well as the cellulose (ADF-lignin). Both expressed in percent (van Soest and Robertson 1975). The GE was determined in a Parr adiabatic calorimetric pump, model 1241.The in vitro digestibility of the DM (IVDMD) and OM (IVOMD) was carried out by pepsin-pancreatin-viscozime, according to Boisen and Fernández (1991).

A completely randomized design with a 4x6 factorial arrangement for chemical characterization was applied, corresponding to four silage types (table1) and six conservations time (0, 15, 30, 60, 90 and180 d). In vitro digestibility was asses by means of a completely randomized design and analysis of simple variance was applied, with silage of 15 elaboration days. Three replications per each treatment were carried out. All determinations were made in triplicate. The means were contrasted by the analysis of variance technique, according to the recommendations of Steel et al. (1997).Where significant differences (P < 0.05) were found, they were compared with Duncan (1995) test. The analyses were carried out with the application of the Infostat statistical program (Di Rienzo et al. 2012).

During the conservation process, the average temperature was of 22 ºC. When opening the micro-silos, the presence of alcohol or any symptom of ensilaged materials decomposition was never found. All of them had a sweet smell.

The result of the interactions studied for chemical indexes is shown in table 2 and 3.The interactions treatments per day were significant (P < 0.05) for all chemical indexes evaluated:DM,OM,CP,CF, ash, EE,NFE,GE,NDF,ADF,lignin,hemicelluloses and cellulose.

Silage four in the 0 day (4.67 %) showed the lower ash concentration. The higher EE content (P < 0.05) showed silage one, in the 30th (4.95 %) and 180th (4.95 %) days. The higher NFE concentration was evident in silage two, in the 30th (84.09 %) and180th (84.09 %) days. The GE content was shown high in all the studied variants. However, was significantly (P < 0.05) high in silage one, in the 0 day (18.30 kJ gDM^-1).

The NDF (table 3) was high (P < 0.05) in silage one in the 30th (18.21 %) and 180th (18.19 %) days. The lower ADF concentration (P < 0.05) have it silage two, in the 0 (2.55 %) and 60th (2.54 %) days. The higher lignin content (P < 0.05) was determined in silage four, in the 30th (2.46 %) day. The hemicelluloses decreased (P < 0.05) in silage four, in the 15th (9.48%) and 60th (9.48 %) days, and the cellulose increased (P < 0.05) in silage one, in the 30th (6.20%) and 180 th (6.19 %) days.

The results of the digestibility study, of the four taro silages are showed in table 4.In the evaluation of DM and OM in vitro digestibility was showed that there were significant differences (P <0.05).

The higher IVDMD (P < 0.05) coefficient had it silage two (74.65 %).In silage one, the IVDMD was lower (63.83%). Relate to IVOMD, silage two showed the higher digestibility (76.76 %) coefficient (P < 0.05). Similarly, silage one had the lower digestibility coefficient (65.08 %).

Chemical composition. The concentration of DM in the silages was in the recommended level (25-35 %) to characterized a silage of appropriate quality (McCullough 1975).In this regard, McDonald et al. (1981) suggested that 30% of DM was a minimum level to reduce the undesirable growth of clostridia that worse the product. The DM content of silages, until the 180th day of evaluation, were between 29.10 and 33.61%.Consequently, they can be considered good quality products for their use in animal feeding.

Regarding the CP content, the data of this research were in a concentration from 8.13 to 8.83%. Similar CP values informed Marrero et al. (1984) in silages of taro tubers of six months (8.5 %). Ogunlakin et al. (2012) referred higher results in tubers in natural way (4.93 to 5.17 %), Fetuga and Oluyemi (1976) found 3.1% in studies with cooked tubers . Apparently, by means of silage process the CP content is increased regarding to tubers in natural and cooked way. The silages CF in the four variants showed low content (2.84 to 6.23%).In researches carried out with ensiled tubers, Marrero et al. (1984) determined 13.4% of CF. This wide variation in CF content could be due to the used ingredients in silages formulation. Also, the nutritional composition of roots and tubers vary according to climatic conditions, cultured varieties, soil conditions and time in which the crop is made (FAO 1990 and Onwueme 1999).

The NFE (78.55 to 84.09 %) and ash (4.67 to 6.88 %) concentration in silages showed higher content, regarding to that referred by Olajide et al. (2011) in fermented and dryings tubers. These authors informed 73.50 % for NFE and 0.88 % for ashes. Nevertheless, in relation to the EE content (2.63 %) referred by these researches, silages showed a similar performance to this nutrient (1.93 to 4.95%).The variation in NFE values could be due to the amount of fermented carbohydrates that the silages had in this experiment.

In this study, the GE content in silages was high, with values between 17.47 and 18.39 kJ gDM^-1 Abdulrashid and Agwunobi (2009), in tubers meal, informed 3.21 kcal g DM^-1 (13.42 kJ.gDM^-1),lower figures to those verified in this research. Apparently, the caloric content of ensiled tubers was increased by the energy contribution of natural yogurt, whey and molasses B.

In general, silages showed variation in NDF content (13.57 to 18.21%), ADF (2.55 to7.56%), lignin (0.81 to 2.46%), hemicellulose (9.48 to13.16%) and cellulose (1.42 to 6.19%). These variations can be associated to the lack or presence of molasses in the fermented material (Guzmán et al. 2012).

It has been carried out several in vitro digestibility researches of DM and OM in other types of tubers in natural way. Ly et al. (2010) informed in cassava tubers values of 66 % of DM and 68.7 % of OM. In studies made with sweet potato, Ly et al. (1999) informed contents of 54.5 % of DM and 62.5 % of OM. These results were lower to those obtained in the study of the four silages variants. Apparently, the ensiled process and the digestibility method applied increased the in vitro digestibility coefficients of DM (63.83 to 74.65 %) and OM (65.08 to 76.76 %).

From the point of view of the OM in vitro digestibility, the ensiled taro tubers seem to be slightly better than other tropical tubers; due to in the ensiled process by means of anaerobic fermentation, lactic acid is produced and, in presence of this one, can be recover some wastes components, like protein, minerals and lipids (López et al. 2006).

The conservation of taro tubers, between 0 and 180d, by means of the addition of variables levels of natural yogurt, whey and sugar cane molasses originates different products that have good content of DM, CP, GE and low concentration of CF, suitable for it use in feeding pigs.

Abdulrashid M., Agwunobi L. N. & others. 2009. ``Taro cocoyam (Colocasia esculenta) meal as feed ingredient in poultry''. Pakistan Journal of Nutrition, 8 (5), pp. 668-673.

AOAC. 2005. Official Methods of Analysis. 18th ed., Gaithersburg, MD., USA: Association of Official Analytical Chemists Inc.

Boisen S., Ferna J. A. & others. 1995. ``Prediction of the apparent ileal digestibility of protein and amino acids in feedstuffs and feed mixtures for pigs by in vitro analyses''. Animal Feed Science and Technology, 51 (1), pp. 29-43.

Boisen S. & Fernández J. A. 1991. ``In vitro digestibility of energy and amino acids in pig feeds''. In: Verstegen M. W. A., Huisman J. & den Hartzog L. A. (eds.), Digestive Physiology in Pigs, Netherlands: EAAP Publication, p. 23l-236, Available: <http://agris.fao.org/agris-search/search.do?recordID=NL9200846>, [Accessed: November 14, 2014].

Caicedo S. W., Rodríguez B. R., Lezcano P., Vargas B. J. C., Ly J. & Valle S. 2013a. ``Efecto de inocuidad del ensilado biológico de tubérculos de papa China (Colocasia esculenta (L.) Schott) para la alimentación de cerdos''. Revista Amazónica Ciencia y Tecnología, 2 (3), pp. 162-171.

Caicedo W. O., Rodríguez R., Lezcano P. & Ly J. 2013b. ``Estudios de composición química de ensilados de papa china (Colocasia esculenta L. Schott) destinados a la alimentación porcina''. In: XXIII Reunión de la Asociación Latinoamericana de Producción Animal, La Habana, Cuba, ISBN: 978-950-7171-492.

Di Rienzo J. A., Casanoves F., Balzarini M. G., González L. & Robledo C. W. 2012. InfoStat. version 2012, [Windows], Universidad Nacional de Córdoba, Argentina: Grupo InfoStat, Available: <http://www.infostat.com.ar/> .

Duncan D. B. 1955. ``Multiple range and multiple F tests''. Biometrics, 11 (1), pp. 142.

Ezedinma F. O. 1987. ``Response of Taro (Colocasia esculenta) to water management, plot preparation and population. 3rd Intl''. In: III Symp. Trop. Root Crops. Ibadan, Nigeria.

FAO. 1990. Roots, tubers, plantain and bananas in human nutrition. Effect of processing on nutritive values. Rome: Food and Agriculture Organization of the United Nations.

Fetuga B. L. & Oluyemi J. A. 1976. ``The metabolizable energy of some tropical tuber meals for chicks''. Poultry Science, 55 (3), pp. 868-873.

Guzmán O. 2010. Estudio de conservación de desechos de mango (Mangifera indica L.) para la alimentación de ovinos en el Estado de Nayarit. Master Thesis, Universidad Autónoma de Nayarit, Tepic. México, 87 p.

Guzmán O., Lemus C., Martínez S., Bonilla J., Plasencia A. & Ly J. 2012. ``Chemical characteristics of silages of mango (Mangifera indica L.) by products for animal feeding''. Cuban J. Agric. Sci, 46, p. 369.

Herrera M. L., WingChing-Jones R. & Rojas-Bourrillón A. 2008. ``Características fermentativas y nutricionales del ensilaje de rastrojo de piña (Ananas comosus)''. Agronomía Costarricense, 33 (1), Available: <http://revistas.ucr.ac.cr/index.php/agrocost/article/view/6731>, [Accessed: November 14, 2014].

INAMHI. 2013. Reportes Climáticos del Ecuador. Instituto Nacional Meteorológico Hídrico, pp. 20-25.

Llano D., López D. & Mora F. 2008. ``Potencial del ensilaje de desechos de naranja (Citrus sinensis)''. Rev. Ciencias Técnicas Agropecuarias, 17, p. 41.

López J., Sánchez D. I. & Rosas J. A. 2006. ``Analysis of free amino acids in fermented shrimp waste by high-performance liquid chromatography''. Journal of Chromatography A, 1105 (1), pp. 106-110.

Ly J., Almaguel R., Delgado E., Carón M. & Cruz E. 2010. ``Estudios de digestibilidad in vitro (pepsina/pancreatina) de raíces de yuca para alimentar cerdos''. Revista Computadorizada de Producción Porcina Volumen, 17 (4), Available: <http://www.iip.co.cu/rcpp/174/174_06artjly.pdf>, [Accessed: November 14, 2014].

Ly J., Caron M. & Delgado E. 1999. ``A note on in vitro digestibility of sweet potato tubers (Ipomoea batatas (Lam) L) for pigs''. Cuban Journal of Agricultural Science, 33, p. 179.

Ly J. & Delgado E. 2005. ``A note on in vitro (pepsin/ pancreatine) digestibility of taro (Xanthosoma sagitifolia spp) and cocoyam (Colocasia esculenta spp) for pigs''. Revista Computadorizada de Producción Porcina, 12, p. 90.

Marrero L., Vargas S. & Contreras F. 1984. ``Ensilado de malanga japonesa (Colocasia esculenta) en la alimentación de cerdos en ceba''. Cienc. Téc. Agric. Ganado Porcino, 7 (4), pp. 85-93.

McCullough M. E. 1975. ``Nuevas tendencias en ensilaje de forrajes''. Rev. Mundial de Zootecnia, 15, p. 44.

McDonald P., Henderson A. R. & Heron S. J. E. 1981. The biochemistry of silage. 2nd ed., Marlow, Chalcombe: John Wiley & Sons, Ltd., 340 p., Available: <http://www.cabdirect.org/abstracts/19811424455.html>, [Accessed: November 14, 2014].

Ogunlakin G. O., Oke M. O., Babarinde G. O. & Olatunbosun D. G. 2012. ``Effect of drying methods on proximate composition and physico-chemical properties of cocoyam flour''. Am. J. Food Technol, 7 (4), pp. 245-250.

Olajide R., Akinsoyinu A. O., Babayemi O. J., Omojola A. B., Abu A. O. & Afolabi K. D. 2011. ``Effect of processing on energy values, nutrient and anti-nutrient components of wild cocoyam (Colocasia esculenta (L.) Schott) corm''. Pakistan journal of nutrition, 10 (1), pp. 29-34.

Onwueme I. 1999. ``Taro cultivation in Asia and the Pacific''. Rap Publication, 16, pp. 19.

Ramos M. 1995. Aplicación de técnicas enzimáticas de digestión in vitro a la valoración nutritiva de piensos de conejos. PhD Thesis, Facultad de Veterinaria, Departamento de Nutrición y Bromatología, Universidad Complutense de Madrid, Madrid, España.

Rêgo M. M. T., Neiva J. N. M., Rêgo A. C. do, Cândido M. J. D., Carneiro M. S. de S. & Lôbo R. N. B. 2010. ``Chemical and bromatological characteristics of elephant grass silages with the addition of dried cashew stalk''. Revista Brasileira de Zootecnia, 39 (2), pp. 255-261.

Steel R. G., Dickey M. & Torrie J. H. 1997. Principles and Procedures of Statistics. A Biometrical Approach. 3rd ed., New York, USA: MacGraw-Hill Book Company In Company, 666 p., Available: <http://www.sidalc.net/cgi-bin/wxis.exe/?IsisScript=ORTON.xis&B1=Buscar&formato=1&cantidad=50&expresion=Torrie,%20J.H.>, [Accessed: November 14, 2014].

Van Soest P. J. & Robertson J. B. 1975. Analysis of forages and fibrous foods. New York, USA: Cornell University Press, 165 p.

Van Soest P. van, Robertson J. B. & Lewis B. A. 1991. ``Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition''. Journal of dairy science, 74 (10), pp. 3583-3597.