Data were analyzed using the general linear model (GLM) procedure of the Statistical Analysis System Institute, Inc. (SAS, 1995). Differences among means were tested for significance using Duncan’s multiple range test of SAS (1995). Pearson’s correlation coefficients between dietary PAPV and lactating performance parameters were evaluated.

PAPV (mg Ca/g feed DM), pH and gas production by different feedstuffs after 24 h incubation are shown in Table 3. In general, feedstuffs having higher PAPV produced more total gas with lower pH, as expected. The relationship between PAPV and total gas production or pH has been observed in a previous study (Song, 2005) with a similar trend. Of the tested feedstuffs, the grains and corn germ meal showed the highest PAPVs. There was a large variation in PAPV among oilseed meal samples. Corn germ meal and coconut meal were high in PAPV, while corn gluten meal, palm oil meal and rapeseed meal had lower PAPV. Bran and hull samples were also variable in PAPV. Protein sources, except corn germ meal, had a much lower PAPV than the starch feeds, which is similar to the results of Wadhwa et al. (2001a). The low PAPV of protein sources can be explained by higher buffer capacity of dietary proteins. For instance, Jasaitis et al. (1987) reported a high pH and buffering capacity of protein feeds due to the buffering ability of amino groups, and Crawford et al. (1983) also showed that a smaller amount of base was required to maintain the pH in a continuous culture system when high protein feedstuffs were used. Buffering in the rumen is mainly due to ammonia from dietary proteins with plant proteins providing between 10 and 20% of the total buffering capacity (Playne et al., 1966). Although differences in PAPV among various feed ingredients seemed to reflect the fermentable and buffering components of the feeds, the values may not be used as an absolute value for determination of acid production capability. This method was very sensitive to variation in microbial activity between batches of rumen fluid, which resulted in high variation in PAPV (data not shown). Standardization of rumen fluid and inclusion of reference feeds may reduce variation among incubation batches. Before a more reliable method for accurate estimation, PAPV can be useful in ranking of feeds, rather than using them as absolute values, when test feeds are evaluated in several runs. Therefore, PAPVs of total feeds from dairy farms were estimated to determine the relationship between dietary PAPV and production traits of dairy cows in a subsequent experiment.

Highly fermentable carbohydrates are often necessary to meet the energy requirements of high producing dairy cows. However, this practice may cause a reduced rumen pH, low productivity and health problem like metabolic diseases. Generally, there is a close relationship among feeding, health and production; hence, metabolic disease like ruminal acidosis can be prevented by proper feeding management. In this experiment, the PAPV (mg Ca/g feed DM) of TMR diets at five dairy farms was estimated based on values determined in the previous in vitro experiment, and the relationship between dietary PAPV and production traits was evaluated by Pearson’s correlation. Field data were collected from 744 dairy cows, and the average performance is shown in Table 4, while the intake, dietary composition and nutrient concentration are presented in Table 5. Average body weight and milk yield in this in vivo experiment were approximately 620 kg and 31 kg, respectively. Milk fat and protein were 4% and 3.2%, respectively. As shown in Table 6, Pearson’s correlations between PAPV and milk yield and FCM yield showed a positive correlation. Dietary PAPV also had a positive correlation with nutrient intake of dairy cows. However, PAPV was negatively correlated with milk fat and protein concentrations. From data obtained in the present study, it is apparent that cows receiving high PAPV consumed more nutrients and produced more milk although the concentration of milk fat was reduced. This is somewhat different from previous results. For instance, increasing concentrate acidogenicity was negatively related to dry matter intake (Wadhwa et al., 2001b) and milk yield (Dewhurst et al., 2001), with variable effects on milk composition. The PAPV of TMR used in the experimental farms ranged from 5 to 15; therefore, the PAPV was not high enough to yield a negative impact on production traits. Welch et al. (1969) indicated that NDF is the nutritional component of roughage that is connected with chewing activity, which is now frequently suggested in the literature as minutes per kilogram of NDF intake. However, chemical benefits of fiber alone are not suitable for optimal diets of high producing dairy cows due to the variation of fiber effectiveness in stimulating chewing, resulting from differences in particle length. Nocek et al. (1988) reported that a proper ratio of non-structural carbohydrate to NDF be used to formulate diets to maximize milk yield. Poore et al. (1991) insisted that the optimal ratio of forage NDF to ruminally degraded starch be maintained ≥1:1 (wt/wt) to prevent a decrease in milk fat (%) when high NDF forage replaced low NDF forage. In general, structural NDF is important for cud chewing and good rumen health, and chemical NDF provides energy for milk and milk fat production. Some nutritional factors could be considered to influence fiber balance in dairy cow diets. Many nutritionists have used forage NDF as a guideline for optimal rumen health and DM intake. The NDF fraction, generally a low rate of digestion, is considered the primary dietary constituent related with rumen fill effect (NRC, 2001). Mertens (1994) reported that ruminal expansion limits DM intake due to the low flow capacity of undigested feed residues; forage NDF intake and fecal output are increased through pelleted high-fiber forages. Optimal forage NDF should be 20% to 21% of dry matter in diets. Total NDF should be increased when feeding less than 20% forage NDF. In this study, the level of forage NDF was about 13.7% in diets and lower than the optimal level. However, the level of total NDF intake in the diets was not low. The minimum level of structural fiber (NDF) was required to prevent acidosis and other metabolic upsets of the rumen connected with a lack of buffering capacity due to low structural fiber in diets. A negative correlation between dietary PAPV and milk protein was noticed, although the relation was not significant in the present study. Dewhurst et al. (2001) reported that milk protein concentration was not influenced by concentrate type when using corn silage as a roughage source but tended to be higher when high acid load concentrates were fed with grass silage. However, the relationship was not sound, which indicates that milk protein did not fluctuate to a great extent based on diet. Also, the present results showed a negative relationship between PAPV and milk fat content. Researchers (Clark et al., 1993; Depies et al., 1995; Armentano et al., 1997) suggested previously that milk fat response be used to determine NDF effectiveness for non-forage fiber sources as a method of uniting these complex interactions. However, there might be limited application for this technique because effectiveness values have not been repeatable across different types of diets (Clark et al., 1993). As milk yield increases, dietary PAPV also increased (Figure 1), which was reflected in an increased energy requirement for higher milk production. Our result also showed that PAPV for dairy TMR greater than 13.5 did not increase milk yield.

In conclusion, PAPV, although variable, can provide information on acid production capacity of feedstuffs and may be applied to ration formulation for ruminant animals.