MATERIALS AND METHODS

Ethical approval

This study was conducted as part of a Ph.D. research project initiated in 2015 at the Dairy Farming Promotion Organization of Thailand, Muak Lek, Saraburi Province, Thailand. The experimental protocol was reviewed and approved by the Thesis Proposal Committee of the Graduate School, Kasetsart University, Thailand. The animal experiment was performed in 2016, before the establishment of the Institutional Animal Care and Use Committee (IACUC) at the institution; therefore, an IACUC approval number was not available. Nevertheless, all animal handling, management, and sample collection procedures were performed by trained personnel in accordance with accepted animal welfare standards for dairy cattle and institutional guidelines applicable at the time of the study. Throughout the experimental period, the animals were maintained under routine farm management conditions, provided free access to clean drinking water, and monitored daily to ensure their health and well-being.

Study period and location

The study was conducted from January to April 2016 at the Dairy Farming Promotion Organization of Thailand, Muak Lek District, Saraburi Province, Thailand.

Study design

The feeding trial was conducted at the Dairy Farming Promotion Organization of Thailand, Muak Lek District, Saraburi Province, Thailand. Eighteen multiparous Holstein Friesian crossbred cows averaging 25 ± 15 days in milk and an initial BW of 429 ± 29 kg were enrolled in the study. The animals were blocked according to parity and days in milk and randomly allocated within each block to one of three dietary treatments using a randomized complete block design (RCBD), with six cows per treatment (n = 6). Each cow was housed individually in a well-ventilated pen equipped with a permanent roof and had unrestricted access to clean drinking water and mineral supplements.

Three dietary treatments were evaluated. The control group received freshly harvested guinea grass, whereas the two experimental groups received ensiled TMFc or ensiled TMFe. Unlike previous studies that evaluated cassava-derived byproducts as feed supplements, protein sources, or concentrate ingredients, the present study investigated their use as primary roughage sources within ensiled TMF systems.

The guinea grass used in the control treatment was harvested and offered fresh daily. In contrast, the TMFc and TMFe diets were thoroughly mixed, ensiled under anaerobic conditions for 7 days to facilitate lactic acid fermentation, and subsequently stored in 25-kg plastic bags until feeding. During storage, the ensiled diets were routinely inspected, and no visible mold growth, abnormal odor, or signs of spoilage were observed, indicating acceptable preservation quality.

Although peNDF was not quantified, the experimental diets were formulated to provide adequate physically effective fiber based on the physical characteristics and particle size of the roughage ingredients. Both guinea grass and the ensiled TMF diets contained sufficient structural fiber to stimulate chewing activity and saliva secretion.

All cows received their assigned roughage ad libitum and were supplemented with a concentrate containing 20% CP. Total feed allowance was calculated at 3.5% of BW while maintaining a roughage-to-concentrate ratio of 40:60. Before data collection, all animals underwent a 14-day adaptation period to acclimate to the experimental diets and management conditions.

Roughage was offered separately from the concentrate twice daily on an ad libitum basis. Feed allowances were adjusted daily to maintain approximately 10% refusals and ensure unrestricted voluntary intake. Concentrate was offered individually during the morning (06:00 h) and evening (16:00 h) milking sessions.

The concentrate was obtained from a commercial feed manufacturer routinely supplying the Dairy Farming Promotion Organization of Thailand during the experimental period. Because the formulation was proprietary, detailed ingredient composition was unavailable. Nevertheless, the chemical composition of the concentrate, including DM, CP, EE, ash, NDF, and acid detergent fiber (ADF), was determined by laboratory analysis and is presented in Table 1.

Feed intake and milk yield were recorded throughout the 90-day feeding period. Nutrient requirements and feed intake were estimated according to NRC [22]. Although the diets were not formulated to be strictly isocaloric or isonitrogenous, all treatments were designed to meet or exceed the minimum nutrient requirements recommended by NRC [22] for early-lactating dairy cows. Based on the observed milk yield and stage of lactation, the nutrient composition of all diets was considered adequate to support maintenance and milk production under the conditions of the present study.

Feed intake and BW measurement

Daily feed intake was calculated as the difference between the amount of feed offered and the refusals recorded for each cow. BW was measured individually at the beginning and end of the experiment using a calibrated livestock digital scale to monitor BW changes and ensure accurate adjustment of feed allowances.

Feed and fecal sampling and chemical analysis

Representative samples of feed and feed refusals were collected throughout the experimental period for chemical analyses. Fecal samples were collected directly from the rectum by grab sampling during the final 7 consecutive days of the feeding trial. Samples were pooled by animal, dried at 60°C for 72 h, ground to pass through a 1-mm screen, and analyzed for acid-insoluble ash (AIA).

Apparent nutrient digestibility was estimated using AIA as an internal marker according to Van Keulen and Young [23] using the following equation:

Apparent digestibility (%) = 100 − [100 × (% AIA in feed / % AIA in feces) × (% nutrient in feces / % nutrient in feed)]

where AIA represents the acid-insoluble ash concentration (% DM basis) and nutrient represents the concentration of the respective nutrient in feed and feces.

DM, CP, EE, and total ash were analyzed according to AOAC International [24]. NDF, ADF, and acid detergent lignin (ADL) were determined using the method described by Van Soest et al. [25].

Rumen fluid and blood sampling

On the final day of the experiment, rumen fluid samples were collected using a stomach tube immediately before the morning feeding (0 h) and 4 h after feeding to determine rumen fermentation characteristics. To minimize saliva contamination, the initial portion of rumen fluid was discarded before sample collection. Ruminal pH was measured immediately using a digital pH meter. The rumen fluid was subsequently filtered through four layers of cheesecloth and centrifuged at 16,000 × g for 15 min. The resulting supernatant was stored at −20°C until determination of NH₃–N and volatile fatty acid (VFA) concentrations.

Ammonia nitrogen concentration was determined using the colorimetric method described by Chaney and Marbach [26]. Individual VFA concentrations were quantified by high-performance liquid chromatography using a Waters 600E system equipped with a Waters 484 ultraviolet detector and a Novapak C18 column (3.9 × 300 mm) (Waters Corporation, Milford, MA, USA). The mobile phase consisted of 10 mmol/L H₂PO₄ adjusted to pH 2.5 according to the method of Samuel et al. [27].

Simultaneously, blood samples were collected from the jugular vein at 0 and 4 h after feeding. Samples were immediately placed on ice, maintained at 4°C for 1 h, and centrifuged at 3,500 × g for 20 min to obtain plasma. Plasma samples were stored at −20°C until blood urea nitrogen (BUN) analysis according to Crocker [28].

Milk yield and composition

Milk yield was recorded individually at each milking session conducted twice daily. Milk samples were collected before the start of the experiment and subsequently at 2-week intervals. On each sampling day, milk from both morning and afternoon milking sessions was thoroughly mixed, and approximately 100 mL was collected for analysis. Milk composition, including total solids, fat, protein, lactose, and solids-not-fat, was determined using an Electric Milk Tester (FOSS Analytical A/S, Hillerød, Denmark).

Statistical analysis

Data were analyzed using the general linear model procedure implemented in SAS software (SAS Institute Inc., Cary, NC, USA) [29] according to an RCBD. The statistical model included treatment as a fixed effect and block (parity and days in milk) as a blocking factor. Treatment means were compared using Duncan's multiple range test, and statistical significance was declared at p < 0.05 [30]. Results with 0.05 ≤ p < 0.10 were interpreted as tendencies.

The sample size was determined based on animal availability and practical constraints under field conditions. Because an a priori statistical power analysis was not performed, findings showing tendency-level significance, such as milk yield, should be interpreted with appropriate caution.