Strain and cultivation condition
L. plantarum BCC 65951 were grown for 16-18 h at 37 °C in MRS (Difco, Fisher Scientific, Pittsburgh, PA, USA) broth harvested by centrifugation (Beckman Coulter Avanti J-E, Beckman Coulter Life Sciences, Indianapolis, IN, USA) at 10,000 x G for 10 minutes, washed and resuspended with 0.85% saline solution. The cultures were diluted to supply a starting concentration of approximately 107cfu/g fresh forage in the silage fermentations.
Napier Pakchong 1 grass was harvested at 60 days of maturity and ensiled into two treatments: no inoculants (control) and inoculated with L. plantarum BCC65951 at 107cfu/g fresh weight. Silages were prepared using a small scale system, approximately 20 kg portions of forage material chopped into 1-3 cm. length, which was packed tightly in layered plastic bags and vacuum seal. The plastic bags were stored at ambient temperature for 180 days. Thereafter approximately 1 kg of each replicate was collected as the representative of the experimental silage, the samples were kept at minus 200C before subjected to chemical and rumen degradability analyses.
Analysis of fermentation quality
The silage samples were collected at different positions including at the top, in the middle and in the bottom of the stack and finally mixed together. For chemical analysis, 12.5 g of the shredded Napier grass samples were homogenized in a 0.02 ml N H2SO4 in a stomacher (Seward,West Sussex, UK) for 4 min at 200 rpm. The extract was centrifuged (10,000 rpm, 40C, 10 min) and filtered through 0.45 µm membranes (Minisart RC4) prior to be analyzed for fermentation quality . The extract was measured for pH value immediately with a pH meter (Sartorius PB-20, Goettingen, Germany). The volatile fatty acids (VFAs) were analyzed by gas chromatography (GC)  . The lactic acid contents were measured by HPLC-UV using a Bio-Rad Aminex HPX-87H column (Bio-Rad Lab., Hercules, CA, USA) with a 0.6 ml/min flow rate of 0.02 N H2SO4 at 600C  .
Chemical Composition Analysis
The fresh samples were dried at 600C for 48 h. After that, the sample was ground through a 1 mm of sieve size with a Wiley Mill. Dry matter (DM; Method 934.01, ), crude protein (CP; Method 968.06, ), ether extract (EE; Method 920.39, ) and crude fiber (CF; Method 962.06, ) were analyzed. Neutral detergent fiber (NDF), acid detergent fiber (ADF) and acid detergent lignin (ADL) were analyzed using Detergent methods .
Determination of rumen degradability by in vitro gas-production technique
Rumen fluid was collected from four fistulated Thai native bulls. About 230 mg of feed was weighed into 100 ml calibrated glass syringes. Rumen fluid was added to the buffered mineral solution (9.8 g NaHCO3 + 2.77 g Na2HPO4 + 0.57g KCl + 0.47g NaCl + 0.12g MgSO4. 7H20 + 0.16 g CaCI2.2H20), and maintained in a water bath at 39°C under continuous flushing with CO2. About 30 ml of buffered rumen fluid was dispensed into syringes containing the samples and all syringes were incubated in a water bath maintained at 39°C. Gas production was determined at 4, 8, 12, 24, 48, 72 and 96 h of incubation.
Net gas productions (ml/200mg, DM) after 24 h incubation was calculated using the equations as follows :
(ml /200 mg DM)=
GPt = Net gas production (ml/200mg, DM) after 24 h incubation
Vt = Gas volume at time t
Vt = Gas volume at t = 0
GP0 = Blank value at time t
W = Weight in mg DM of sample in each syringe
FH = Correction factor of hay
FC = Correction factor of concentrate
The incubation residue was distilled by neutral detergent solution to remove the microbial biomass from undegraded substrate . The microbial biomass yield (MBY) was calculated after 24 h incubation according to the following formula:
MBY = truly digested sample − apparently digested sample
truly digested sample
Measurement of growth performance
Ten Brahman bull (10.3±0.8 month) with average205 ± 12.8 kg of body weight were housed in individual pens. Vaccination against epidemic diseases and anthelmintic were applied for parasites elimination 10 days before the experiment started and all the experimental animals were allowed to have the adaptation period with the experimental diet for 15 days. The experiment was conducted with a cross-over design. The cattle were fed 1.81 kg. DM/head/d of concentrate and ad libitum silage. Clean drinking water was available at all time. To assure that the animals received feed ad libitum, silage was offered until at least 5% refusal was obtained. Feed offered and refused were weighted every day to calculate voluntary feed intake. To weight the animal, all feed was taken away from feed bunk at 8.00 PM the day before, and all animals were weighted at 8.00 AM on the following day and fresh feed was provided. This weighting procedure was repeated for 2 consecutive days at the beginning of the experiment after adaptation period and also at the end of each particular period. Average live weights of each individual animal were recorded for statistical analysis.
The fermentation quality of silage, chemical composition and in vitro gas production technique was tested for the equality of variance then statistically analyzed by t-test while, growth performance was statistically analyzed according to the cross-over design using SPSS program. Differences among means were tested using the adjusted Duncan’s test, the significance declared at P<0.05.