**4. Discussion**

Using untreated and treated discarded durian peel at the same amount in the diet mainly caused a change in fiber content such as NDF and ADF content. Using untreated discarded durian peel in the diet (FDP) had a higher NDF and ADF content compared with diets containing discarded durian peel treated with molasses (FDPM), discarded durian peel treated with cellulase (FDPC), discarded durian peel treated with *L. casei* TH14 (FDPL) and discarded durian peel treated with molasses and *L. casei* TH14 (FDPML). This was caused by the higher NDF and ADF content presented in untreated discarded durian peel than treated discarded durian peel. A lower NDF and ADF content in treated discarded durian peel is due to acid hydrolysis action during fermentation and cellulase activity. So et al. [7] used sugarcane bagasse treated with molasses in combination with cellulase or *L. casei* TH14 in TMR for dairy cows had a lower NDF content compared with the TMR diet containing untreated sugarcane bagasse. Oba and Allen [24] stated that fiber intake, ruminal fermentation, and production efficiency could be influenced by dietary NDF content and digestibility.

In this study, the intake of DM, expressed as either kg/d or %BW or g/kg BW0.75 , was not significant among treatments. Thus, using untreated or treated discarded durian peel with additives at 25% DM in the diet did not affect the daily DM intake of growing goats. Similarly, So et al. [7] found that 50% DM of untreated and treated sugarcane bagasse combined with additives in TMR diets did not affect the daily DM intake of dairy cows. This suggests that any roughage feeds treated with or without silage additives does not affect the DM intake in ruminants. Intake of ADF (g/d) was significantly different between the FDP and FDPML treatments. The reason for this effect is not clear, although DM, OM, and NDF intake were not different among these treatments. So et al. [7] found nutrient intake was unchanged in dairy cows fed TMR containing untreated and treated sugarcane bagasse with additives. Cherdthong et al. [13] fed untreated and treated rice straw with

molasses in combination with cellulase or *L. casei* TH14 to Thai Native beef cattle and found no effect on nutrient intake. Dietary treatments did not affect weight gain or BW change in growing goats. The effect of additives in combination on performance is small and unclear [25]. Addah et al. [26] compared untreated and treated whole-crop barley with inoculant combination (*L. buchneri*, *L. plantarum*, and *L. casei*) and found no change in weight gain in growing feedlot steers.

Using pretreatment roughage feeds with additives such as LAB, molasses, and fibrolytic enzymes in the diet has been reported to improve nutrient digestibility in the rumen [6,7,13,27]. Additives contribute two mechanisms during fermentation: (1) acid hydrolysis reaction and (2) direct effect of fibrolytic enzymes on polysaccharide structure, which increases feed digestion efficiency of ruminal microbes [10,28]. This study showed that FDPM, FDPC, FDPL, and FDPML were significantly better in terms of DM, OM, CP, NDF and ADF digestibility compared with FDP. This could be due to the pretreatment effect of molasses, cellulase, and *L. casei* TH14 during fermentation of carbohydrate structures, subsequently resulting in nutrient digestibility improvement. So et al. [7] reported that *L. casei* TH14, cellulase, and molasses in combination with treated sugarcane bagasse significantly increased OM, CP, NDF, and ADF digestibility in dairy cows fed TMR diets compared with untreated sugarcane bagasse. Zhao et al. [29] evaluated the in vitro degradability of untreated rice straw and rice straw treated with *L. plantarum* and molasses and found a significant improvement in DM and NDF degradability by *L. plantarum* and molasses in combination. FDPML significantly increased estimated ME intake expressed per kg DM intake compared with FDP. This could be due to a significantly higher OM digestibility found in FDPML than FDP. So et al. [7] similarly found that a combination of sugarcane bagasse treated with molasses and *L. casei* TH14 in TMR-fed dairy cows increased their estimated ME intake compared untreated sugarcane bagasse.

Ruminal pH significantly determines the normal function of microbes in the rumen [30–33]. The normal ruminal pH ranges from 5.5 to 7.0 [34]. This study showed that FDPML significantly decreased the mean ruminal pH by 0.27 compared with FDP; however, the pH was in a normal range (6.4 to 6.7). This could be due to the higher lactic acid and LAB population found in FDPML than FDP. In addition, a significantly higher propionate concentration in FDPML compared with FDP could contribute to a lower pH in FDPML compared with FDP. pHs ranging from 6.4 to 6.7 showed improved fiber digestibility (Table 3) as activity of cellulolytic bacteria slows down at a pH less than 6 [35]. Similarly, So et al. [7] showed that a combination of sugarcane bagasse treated with molasses and *L. casei* TH14 in TMR fed to dairy cows significantly decreased their ruminal pH by 0.07 after 4 h of feeding. Zhang et al. [36] revealed that whole-plant corn ensiled with complex inoculants (*L. plantarum* L28, *Enterococcus faecium* EF08, and *Lactobacillus buchneri* LBC136) significantly decreased ruminal pH by 0.21 compared with ensiled whole-plant corn without inoculants in growing-finishing cattle. Lower ruminal pH leads to lactic-acid-dependent acid production in the rumen and is achieved approximately 2 to 6 h after feeding [30]. Time after offering feed and lactic acid supply rate mainly determine the change in ruminal pH [2,37]. NH3–N is a main nitrogen source for microbial synthesis in the rumen (5 mg/dL minimum and 30 mg/dL maximum requirement) [38]. Additive-treated discarded durian peel significantly decreased mean NH3–N concentration. This could be due to the activity of LAB present in fermented discarded durian peel that affected deamination, resulting in less ruminal protein degradation and enhancing nitrogen utilization in the lower digestive tract. So et al. [7] similarly revealed that additives combined with treated sugarcane bagasse significantly decreased NH3–N concentration from 22 to 20 mg/dL after 4 h of feeding compared with untreated sugarcane bagasse in dairy cows. FDPML significantly increased the propionate concentration, resulting in lowering the acetate-to-propionate ratio and acetate—butyrate-to-propionate ratio compared with other treatments. Increasing the propionate concentration could explain the high nutrient digestibility (Table 3) found in FDPML. In addition, a high lactic acid concentration in FDPML may contribute to an increase in the propionate concentration as lactic acid is

biologically converted into propionate by ruminal microbes in the rumen. Similarly, So et al. [7] showed that *L. casei* TH14 combined with cellulase- and molasses-treated sugarcane bagasse in TMR significantly increased the propionate concentration compared with untreated sugarcane bagasse in dairy cows. Zhang et al. [36] revealed that whole-corn plant treated with complex inoculants significantly increased the propionate concentration from 6.40 mmol/L to 8.98 mmol/L in growing-finishing cattle. Cherdthong et al. [13] similarly found that rice straw treated with *L. casei* TH14 and molasses fed to Thai Native beef cattle significantly increased their propionate concentration from 20.3 mol/100 mol to 23.2 mol/100 mol compared with untreated rice straw. Estimated CH<sup>4</sup> production was significantly lower in FDPML. The reason could be explained by the increase in propionate concentration found in FDPML as hydrogen was used for propionate synthesis, resulting in less hydrogen available in the methanogenesis pathway of methanogen bacteria to produce CH<sup>4</sup> as the main end-product [39]. Similarly, So et al. [7] showed that sugarcane bagasse treated with *L. casei* TH14 combined with cellulase and molasses in TMR fed to dairy cows significantly decreased CH<sup>4</sup> production by 4%. Monteiro et al. [40] tested *L. plantarum* as direct-fed microbial in high-producing cows and similarly found decreased in CH<sup>4</sup> production compared with no additive treatment.

Dietary treatments did not affect bacteria, fungal zoospore, total protozoa, *Holotrich* sp., or *Entodiniomorphs* sp. populations. Ruminal bacteria favor a pH around 7 for optimum growth [35]; this study found the pH ranged from 6.44 to 6.71, which may have contributed to the unchanged bacteria population. Similarly, Bureenok et al. [41] showed that ruzi grass ensiled with molasses or fermented juice of epiphytic lactic acid bacteria fed to cows separately or as a combination did not change ruminal bacteria. However, not all previous studies found unchanged ruminal bacteria when inoculants wer eused. Cherdthong et al. [13] found that rice straw treated with molasses and *L. casei* TH14 combined fed to Thai Native beef cattle significantly increased ruminal bacteria population; the changed in bacteria population may have been due to the optimum ruminal pH ranging from 7.0 to 7.1 for bacteria growth.

Blood metabolites including glucose, pack cell volume, blood urea nitrogen, total protein, albumin, globulin, the albumin-to-globulin ratio, red blood cell, hemoglobin, mean corpuscular volume, mean corpuscular hemoglobin concentration, RBC distribution width, white blood cells, and lymphocytes were similar among dietary treatments. This suggests that the goats were in good health and had a normal metabolism status. Blood metabolites are usually used to evaluate the nutritious plane and health status in ruminants [42,43]. As well as blood metabolites, glucose, blood urea nitrogen, total protein, and albumin concentration were commonly used to evaluate protein and carbohydrate metabolism, where the higher mean value suggests a better nutrient metabolism when these metabolites changed within a normal range [44]. Liver is the main hub where glucose, albumin, and blood urea nitrogen are synthesized [45,46], and glucose and albumin concentrations were greater in goats fed FDPML. Mean glucose concentration ranged from 68.70 to 70.20 mg/dL, which varied in a normal range of 50 to 75 mg/dL [47]. Blood urea nitrogen is the product of NH3–N recycling and is produced from protein degradation by ruminal microbes [48,49]. The lower blood urea nitrogen paralleled the lower NH3–N concentration found in goats fed FDPML.

Dietary treatments did not influence nitrogen intake, fecal nitrogen, or apparent nitrogen retention, expressed as g/d or% of nitrogen intake in goats. Urine nitrogen was significantly lower, at 11%, in FDPC and FDPML when compared with FDP. The effect of cellulase or a combination of molasses and *L. casei* TH14 on urinary nitrogen reduction was unknown; it may possibly be inconsistent with retained nitrogen, as it was found to be the highest in FDPC and FDPML. Cherdthong et al. [13] showed that rice straw treated with cellulase or *L. casei* TH14 separately or as a combination fed to Thai Native beef cattle reduced nitrogen loss both in the urine and feces but failed to reach statistical significance.
