Nutrient composition, particle size distribution, and physical effectiveness
The pef>8 values were lowest for TMRs, although the pefm and pef>1.18 values were higher and very close. In contrast to the peNDFm and peNDF>1.18, the peNDF>8 values were similar among treatments. In this experiment, peNDFm and peNDF>1.18 were higher than 21% of the DM that Mertens  suggested to adequately stimulate chewing activity that was required to maintain an average ruminal pH of greater than 6.0. The peNDF content of experimental rations was adequate and the differences among treatments were attributable to the effects of NFC sources or its nature. The concentration of starch was similar in the B and BC treatments and higher than that observed in the BB treatment, but the NDSC significantly varied among treatments. These results thus confirm that intake of starch and NDSC varied with different NFC treatments. We hypothesized that the use different digestion products, which are the NFC complement of the diet, would alter feed intake, ruminal digestion kinetic, chewing activity, and the production and composition of milk.
Intake and apparent total-tract nutrient digestibility
The intake of peNDFm and peNDF>1.18 did not follow the same trend as the pef content of the diets, although this was similar to that of NDF; however, this increased in treatment BB. Treatment BB, which contained more NSDC, resulted in a significant increase in DMI. These results have been confirmed by several previous studies. Substitution of sugars and NDSC with starch has increased [19–22] or did not change  the intake in dairy cattle. Increases in dietary concentrations of NDSC at the expense of starch also increased [20, 24], decreased [22, 25], or did not affect DMI . Changes in the passage rate or fiber digestibility using various NFC sources may partial explain these changes . A higher intake in treatment BB could have been expected based on the high NDSC and elevated starch content of barley and corn compared with treatments B and BC.
In this experiment, intake might have been regulated by the metabolism of propionate in the liver, and ruminal propionate production generally increases with a higher rate of starch fermentation. The amount of starch intake was 4.69, 4.25, and 3.24 kg/d, and ruminal molar percent of propionate in total VFA was 26.75, 25.45, and 23.04 mol/100 mol in treatments B, BC, and BB, respectively, showing that starch intake decreased with a higher rate of starch intake (Table 4). Therefore, it seems that propionate metabolism in the liver caused the reduction in DMI in treatments B and BC compared with BB because hepatic propionate oxidation probably decreased with the supplementation of BP . In addition, ruminal pH might be implicated in the control of feed intake, particularly in satiety, because VFAs were more rapidly absorbed in the undissociated form at lower pH. In addition, substituting a high-fiber byproduct such as BP for grain may have resulted in the reduction of energy content of the ration or the limitation of intake by the physical filling effects of the diet.
The NFC sources such as BP have a similar range of NDF content as forages, although its particle length was generally smaller. In several experiments, these sources were substituted with forage or grain in diets and then the DMI was measured [3, 5, 6, 22]. When DMI is regulated by ruminal distension, its substitution with forage in diets results in an increase in DMI, whereas its substitution with grain decreases DMI [2, 4, 13]. However, non-forage fiber sources showed inconsistent effects on DMI when substituted with either forage or grain in the diets of lactating cows. In the current experiment, the physical characteristics of the diets were similar; therefore, it seems that these do not differ in their relative filling effects. In treatment BB, the DMI was not regulated by ruminal distension [3, 25]. The greater concentration of propionate in the ruminal liquid in treatment B compared with treatments BC and BB suggests that stimulated propionate receptors in the ruminal region may have controlled satiety and DMI, because propionate is more hypophagic than acetate, resulting in a lower DMI in treatment B versus treatments BC and BB [19, 26].
Partially replacing barley with BP in the dairy cow diet resulted in an increase in total digestibility of DM, NDF, and EE, and a decrease in the digestibility of NFC. However, partially replacing barley with corn in the dairy cow diets showed minimal effects on total digestibility. Unfortunately, total digestibility of starch was not measured in this study. In the present study, decreased fiber digestibility in treatment B using barley was consistent with the reduction in rumination time, ruminal pH, and acetate to propionate ratio. The consumption of diets rich in NDSC imparted positive effects on ruminal digestion. In addition, the results of this experiment might be influenced by differences in the digestible NDF of the treatment diet and not simply the source of NFC, as implied in the objectives of this study.
Differences in digestible NDF rations or starch, NDSC, and sugar content were associated with several parameters, especially acetate and propionate concentrations and their ratio in the rumen, as well as milk fat and protein content. These were significantly different among treatments (P < 0.001) and treatment BB showed the highest intake of potentially digestible fiber or NDSC. Previous studies have confirmed that animal intake response to supplemented starch- or pectin-rich feeds remained the same, although changes in digestibilities have been observed. Valk et al.  reported that cows fed on corn consumed more DM and produced more milk than cows fed on BP. However, DMI and digestibility of NDF were higher in cows fed on BP compared with those fed on corn meal.
The extent and rate of degradation is extremely variable among grains, byproducts, and different sources of NFC [6, 27]. Variation in ruminal degradation of DM and OM across feeds has been reported to range from 29% to 90%, and 29% to 67%, respectively . Total tract digestibility of corn ranges from 91.2% to 98.9%, depending on the processing method and grain type, with ground corn averaging 93.5% . Ruminal degradation of starch in ground corn and ground barley ranged from 51% to 93%, and 42% to 91%, respectively . In addition, ruminal degradation of DM and NDF in BP was 78%  and 94.1% , respectively. The potentially digestible fraction leaves the rumen either by degradation or by passage to the lower tract. Furthermore, feed characteristics and ruminal fractional passage rates would most likely vary; these tend to be higher with fine, non-forage fiber sources due to differences in ruminal retention time, which may be influenced by ruminal consistency . The inclusion of BP in treatment BB resulted in a decrease in particulate passage rates, an increase in ruminal and total mean retention time, and a positive effect on ruminal and total digestion.
Rumen fermentation and kinetics
Variables of ruminal pH such as the outflow rate of digesta from the reticulorumen or ruminal indicate that retention time influences total fiber degradation. Rapidly fermenting carbohydrates such as sugars, NDSC, and some starches have the potential to rapidly decrease ruminal pH by virtue of the sheer mass of organic acids produced in this digestive region within a relatively short period of time. The ruminal pH corresponded to a higher rate of degradation and fermentation of barley grain that produced the highest concentration of total VFA and propionate, lower acetate, and acetate: propionate ratio in treatment B compared with the other treatments (Table 4).
Beet pulp, which is rich in pectin or NDSC, has been shown to yield relatively high amounts of acetic acid, thus resulting in an increase in the acetate to propionate ratio and less butyrate and lactate compared with other carbohydrate sources . In addition, using twelve Friesian cows and sixteen heifers, Lees et al.  reported that in wk 9 and 16, the ruminal VFA proportions of cows fed on BP were higher for acetate and butyrate and lower for propionate compared with cows fed on corn. The molar proportion of acetate in the current study and in previous investigations was probably altered by the tendency of BP to increase fiber intake and digestion. In contrast, Van Vuuren et al.  did not detect any differences in ruminal fluid pH or VFA concentrations in either experiment involving cows fed on corn products or BP and soybean hulls.
Changing the dietary source of NFC or NDSC and starch content of diet rations of lactating dairy cows has been shown to alter the ruminal environment and fermentation products. The effects, although not always consistent, have been observed on the proportions of VFA, the concentration of rumen NH3-N, and the extent of fiber digestion with feeding starch, sugars, or NDSC-rich feeds . The application of treatment BC resulted in a higher acidic pH, lower total VFA and proportion of propionate, and increased acetate concentration and acetate:propionate ratio in the ruminal fluid. Replacing BP with barley resulted in a higher acidic pH, increased acetate concentration in ruminal fluid, without any change in total VFA, and decreased proportion of propionate, valerate, isobutyrate, and isovalerate. Therefore, compared with barley, BP created more favorable conditions for microbial utilization of other carbohydrates in the rumen, at least in part due to the neutral pH. Decreasing daily mean ruminal pH, which was associated with an increase in ruminal outflow rate of digesta, resulted in a decrease in ruminal NDF digestibility and total tract digestibility [2, 8, 9, 33]. Zebeli et al.  showed a decrease in total tract NDF digestibility by increasing the outflow rate of particulate digesta, which was associated with a decrease in ruminal mat consistency. It might be possible that a low digesta passage rate and high consistency in the ruminal mat due to inclusion of BP in the treatment BB was associated with a higher water holding capacity, greater entrapment of small particles, decreased outflow rate of solid digesta, and increased ruminal NDF digestibility in dairy cows.
The ruminal particulate passage rate was highest and lowest in treatments BC and BB, respectively. Because corn has a higher specific gravity than barley and BB, inclusion of corn in the diet resulted in an increase in ruminal passage rate and a decrease in rumen mean retention time (Table 4), and consequently, it decreased rumination time and total chewing activity. In contrast, the higher total chewing activity in treatment BB might have resulted from the lower pH that was attributable to a high degradability of barley, which may have decreased the rate and extent of NDF digestion. In treatment BB, the BP did not result in a reduction in ruminal pH despite its high rate of NDF digestion because of the production of a higher concentration of acetate and lower propionate. However, it has a low functional specific gravity and a large particle size , which could decrease its passage rate and increase its ruminal mean retention time and consequently, rumination and total chewing activity.
Two characteristics that affect the passage of digesta are density and particle size. In addition, there is a clear negative relationship between particle density and ruminal mean retention time. Particles with a specific gravity of lower than the critical size (1.18 mm) quickly pass out of the reticulorumen . The specific gravity for BP was 1.152, whereas that of barley and corn grain were higher than 1.237 . Therefore, inclusion of BP in treatment BB decreased particulate passage rates and increased ruminal and total mean retention time. Particles with a density range of 1.2 to 1.5 showed the highest rate of passage in cattle and sheep [3, 8, 9]. Grain and its byproducts (except for BP and soy hulls] showed a higher rate of passage from the rumen than forages because of the higher functional specific gravity and a particle size much smaller than that associated with stimulation of rumination. Therefore, most grains would have a faster rate of passage from the rumen than forages and other byproducts.
In treatment B, negative associative effects resulting from a lower pH due to a high degradability of barley would probably decrease the rate of digestion. Beet pulp did not reduce ruminal pH despite its high rate of NDF digestion ; however, it has a low functional specific gravity and a large particle size , which could decrease its passage rate and increase its ruminal mean retention time. Therefore, BP should have a very high ruminal digestibility , potentially supporting levels of ruminal microbial protein synthesis and milk production, similar to that observed in corn . However, generalization of these results under different feeding conditions would be an oversimplification because different dietary treatments can interact with each other, thus affecting passage rate.
Eating, ruminating, and chewing times
Although the effect of forage particle size on chewing behavior has been substantially investigated, information on the effects of NFC, starch, or NDSC content of ration, especially when barley grain partially replaced with corn and BP, is limited. Despite the absence of differences in eating time, the rumination time and total chewing activity significantly varied among treatments. Prolonged total chewing time due to the increase in dietary peNDF resulted from an increase in ruminating time rather than an increase in eating time.
Treatment BB showed the highest rumination time and total chewing activity. Total daily chewing time varied from 8.6, 9.9, and 10.1 h, with ruminating time at 5.6, 4.8, and 5.9 h in treatments B, BC, and BB, respectively. These values were within limits previously reported for dairy cows [9, 13], although the mean chewing time per kg DM ingested was less than 30 min, contrary to the recommendations of Mertens , who proposed values equal to or exceeding 30 min/kg DM intake as suitable for limiting the risk of digestive disorders. Eating time per day and per kilogram of DMI increased with the supplementation of BP. As diet NDF intake increased with added BP in the current study, a decrease in time eating per kilogram of NDF was observed. However, no effect of BP substitution for corn on time eating per day or per kilogram of DMI was reported [2, 13]. Tafaj et al.  and Zebeli et al.  have previously shown that the fiber content (NDF or forage NDF) and non-fiber carbohydrates in the diet affected daily chewing and rumination activities. Forage content remained the same across treatments; therefore, the nature of NFC or starch and NDSC content impacted the rate of feed intake, ruminal mean retention time, rumination time, and total chewing activity.
Although milk yield was not altered, replacing corn and BP with barley in diets resulted in an increase in milk fat (yield and percentage), and a decrease in protein concentration (yield and percentage; Table 6). Dietary BP can increase , decrease , or support milk yield, which have also been observed in cows fed on corn or barley [22, 35]. In addition, intake may be improved , decreased [22, 35], or remain the same [27, 35] using BP. Various dietary carbohydrates can result in changes in ruminal fermentation and generate VFAs patterns that influence the hormonal status of the animal and potentially affect milk production and composition.
The increase in fat secretion in treatments BC and BB has been shown to elevate both acetate and acetate: propionate ratio in the rumen. However, treatments BC and BB did not alter ruminal concentrations of either the total VFA or butyrate. Mansfield et al.  previously showed a higher milk fat concentration (3.82 vs. 3.64), whereas milk protein (3.01% vs. 2.90%) and yield were lower in cows fed on the BP diet compared with the corn diet. The fat content of milk was enhanced in treatment BB, which can be directly attributed to the increase in fiber intake and digestion in cows and the molar proportion of acetate in the rumen. Consistent with the current experiment, milk fat increased in cows fed on the BP diet as a replacement for grain [22, 31] due to the higher fiber intake and greater acetate concentrations in the rumen. It seems that the higher milk fat observed in treatment BB was due to the increase in intake of digestible NDF and NDSC, ruminal mean retention time, rumination time, total chewing activity, ruminal acetate concentration, and ruminal pH.
A decrease in milk CP content and yield was observed when corn and BP were substituted with barley. This finding was confirmed by Mansfield et al. , but was in contrast to the findings of Valk et al. . In contrast to the current experiment, a lower intake of CP, lower NH3-N concentrations in the rumen, and lower CP degradability for cows fed on BP was observed, compared with that in cows fed on corn  or barley . The consumption of carbohydrates such as NDSC that are more extensively fermented in the rumen may improve the utilization of the NPN through the stimulation of microbial protein synthesis. In contrast, when NH3-N concentrations in the rumen cannot meet microbial requirements, the fermentation of fiber is reduced and the animal responds with a lower intake and digestibility. Total VFA concentration in treatments BC and BB were observed to decrease. The lower total rumen VFA is an indicator of lower ruminal fermentation of fiber that limited milk protein yield. Huhtanen , Rooke et al. , and Chcster-Jones et al.  previously showed that BP supports VFA concentrations in the rumen similar to those for corn and barley grain, whereas VFA concentrations decreased in some situations . In addition, NFC sources that were used in the current experiment varied in their DM rate and extent degradability, support of microbial yield, and supply of protein to the small intestine because microbial CP yield is influenced by carbohydrate and nitrogen sources, rate of carbohydrate fermentation, bacterial growth rate, ruminal dilution rate, and pH .