Location of the experiment

The work was carried out between 2014-2016, at the facilities of the Autonomous University of Aguascalientes, the Llano Aguascalientes Technological Institute and the Agricultural Technological Baccalaureate Center num. 61 [Aguascalientes, Mexico (Central-North Region)].

Experimental material

The BM was used, by-product of the juice extraction of the company “Juice extractor Valle Redondo SA of CV de Aguascalientes”. The BM silage lasted 45 d, for which: a base mixture was prepared containing a proportion of 80 g 100 g^-1 MS of BM (with 18% MS) and 20 g 100 g^-1 MS of Chloris gayana hay (with 90% MS) as moisture adherent material. Nitrogenous sources (FN) were added to the base mixture: 1) urea, 3.75% of the total MS of the mixture, and 2) sowing, with 23.5% of PC (from fattening animals) in proportion equivalent to the nitrogen contributed for urea. In addition, were included 10 g t^-1 MS of a commercial inoculum (Sil-All 4x4^®) containing lactic acid bacteria (Streptococcus faecium, Lactobacillus plantarum, Pediococcus acidilactici and Lactobacillus salivarius) and exogenous enzymes (cellulases, hemicellulases, pentosanas and amylase).

An electromechanical mixer was used. The mixtures were deposited in microsilos of polyvinyl chloride (PVC) pipe 5.08 cm in diameter and 30 cm long, compacting with a metallic tamper to eliminate oxygen and closing with a bell-type plug reinforced on both sides; the microsomes were stored at room temperature of 20 °C.

Treatments and response variables

The treatments were: apple bagasse (BM), apple bagasse with inoculum (BM+I), apple bagasse with urea (BM+U), apple bagasse with urea and inoculum (BM+U+I), bagasse apple with sowing (BM+C), and apple bagasse with sowing and inoculum (BM+C+I). Samples collected at 0 and 45 d were dried at 60 °C in a forced air oven until constant weight was obtained and processed in a Thomas Wiley mill with a 1 mm sieve. The MS was obtained by difference between the initial and final weight [ ^{AOAC (1999)} protocol 930.15-1930], the PC was determined through protocol 990.03-2002 of the ^{AOAC (2002)}, the pH was measured in a homogeneous aqueous solution composed of 10 g of silage and 100 ml of distilled water (one hour after its preparation) using a digital pH-meter (^{Cherney and Cherney, 2003}), lactic acid was determined by the colorimetric method (^{Madrid et al., 1999}) and AGV by gas chromatography (Perkin Elmer^® Co., Clarus 560 D Gas Chromatograph) (^{Erwin et al., 1961}).

Statistical analysis

We used a completely randomized design (DCA) with a factorial arrangement (3×2) and four repetitions per treatment according to the model (1). The treatment means were compared with the Tukey test (p< 0.05). In addition, a test of orthogonal contrasts was applied to compare: 1) BM vs. BM with FN o/e I; 2) BM vs. BM with FN; 3) BM vs. BM with I;4) FN vs. treatments without FN; 5) I vs. treatments without I. The general linear procedure (Proc GLM) of the SAS package (Statistics Analysis System V. 9.2) was used.

Where: Y_ij= response variable; FN_i= effect of the ith nitrogen source; I_j= effect of the jth inoculant; (FN*I)_ij = interaction between the ith nitrogen source and the jth inoculant; E_ij= random error.

The MS content was similar in all the treatments (p> 0.05, Table 1) and adequate for the release of lactic acid and pH decrease during silage fermentation (^{Schroeder, 2013}).

There was FN×I interaction for PC content (p< 0.001). The treatments with C had better PC increases (BM+C without I and BM+C + I: 8.00 and 7.44 g 100 g^-1 MS) than the treatments with urea (BM+U without I and BM+U + I: 3.2 and 1.6 g 100 g^-1 MS) or with BM without FN (BM and BM+I: 0.92 and 2.02 g 100 g^-1 MS), the final PC contents of some treatments were comparable to those of corn silage (8-10 g 100 g^-1 of MS, ^{Guedes et al., 2012}). Considering the PC, the silage BM could be used to replace some types of grain and reduce feed costs. ^{Mirzaei-Aghsaghali et al. (2011)} analyzed in vitro gas production (contents of FDN, FDA and non-fibrous carbohydrates of 61.2, 46.7 and 23.8%, respectively) of BM with PC similar to that found in the present study and considered it viable to be used in the feeding of ruminants. ^{Tiwari et al. (2008)} found no negative effects on the production of milk and its fat and protein contents by supplying up to 33% of corn grain with BM (per 300 d).

Silages with I or some FN had higher initial and final pH values than the rest of the treatments (p< 0.002), although at the end of the fermentation all the treatments had pH values within the desirable ranges for the variations of microorganisms and production of fatty acids (^{Weinberg and Ashbell, 2003}; ^{Skladanka et al., 2012}). The treatments with C recorded the highest pH (3.91-4.33 vs 3.38-3.44), but similar to those desirable in corn silage whit pH= 3.8-4.5 (^{Kolver et al., 2001}; Skladanka et al., 2012; ^{Schroeder, 2013}).

The treatments with C had higher final concentration of lactic acid (p< 0.0001, Table 2), which represents better fermentation and stability of the silage and lower probability of contamination and degradation of the proteins (^{Bautista et al., 2007}; ^{Nkosi et al., 2011}; ^{Huntanen et al., 2013}). Although the proportions of lactic acid of all the treatments of the present study were within the minimum and optimum ranges for the high-quality silages (^{Kolver et al., 2001}; ^{Guedes et al., 2012}).

During fermentation, the increase in acetic acid was greater than that of the butyric and propionic acids (p< 0.05), which could help reduce the aerobic deterioration of the silage at the time of opening the silo (^{Weinberg et al., 2002}) since acetic and propionic acid concentrations are indicators of silage stability (^{Huntanen et al., 2013}). The increase in acetic acid in the treatments of the present study was greater than that published by ^{Guedes et al. (2012)} and ^{Skladanka et al. (2012}), especially in the BM+C treatment, followed by the other treatments with C (p< 0.0001). In addition, treatments with C and I had propionic acid at the end of fermentation (p< 0.001).

On the other hand, the increase of butyric acid in BM+U+I and BM+I during fermentation (of 0.60 and 0.63 mol/100 mol), could represent some contamination by bacteria Clostridium spp. in these treatments (^{Leupp et al., 2006}; ^{Bautista et al., 2007}). On the contrary, the proportions of butyric acid of BM, BM+U, BM+C and BM+C+I did not change through the fermentation time, suggesting that the final contents of butyric acid of BM+C and BM+C+I higher than desirable, ^{Guedes et al. (2012)}, were related to the initial content of the acid in C more than to the growth of Clostridium spp.

Conclusiones

In the present study the BM silage process increased the PC and in some cases the pH decreased to values comparable to those of other more conventional silage types such as maize, in addition, the increases of lactic acid and acetic acid of silage BM indicate which has the necessary stability to be included as an ingredient in ruminant diets. Mainly, adding C to the ensile BM could be an alternative to improve its quality, in order to be used in diets without adversely affecting the health or productive behavior of dairy cattle and meat.