High grain diets for dairy cattle.

Abstract

19 High grain diets for dairy cattle K.F. Knowlton Depar tment of Dairy Science, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA knowlton@vt.edu Summary Continued improvements in the genetic potential of dairy cattle for milk production increase the need for maximal energy and protein availability from the diet. Grain is fed to increase the nutrient density of the diet, meet these increased needs for protein and energy, and increase the efficiency of milk production. Starch is the primary energy component of grains, and is considered the primary driver of microbial protein synthesis in the rumen. Understanding starch digestion is the key to optimizing protein and energy supply to the cow, and to improving the ef ficiency and effectiveness of high grain diets. This paper reviews studies on starch digestion in the rumen, small intestine, and large intestine of ruminants; the effect of starch digestibility on performance of dairy cows; and the animal health implications and some practical aspects of feeding high grain diets to dairy cows. Keywords: dairy cows, grain diets, starch digestion, rumen, intestines, milk Herrera_Saldana et al. (1990b) compared rates of starch digestion of five cereal grains with in vitro and in situ methods. When processed similarly, oats had the fastest rate of starch digestion, followed by wheat, then barley, then corn, and finally sorghum. In vivo rumen digestion of starch in various grains follows the patterns indicated by in vitro and in situ methods. In a comparison of sorghum_, corn_, and barley_based diets containing approximately 82% grain (Spicer et al. 1986), the rumen digestion of sorghum (75.2%) was significantly less than that of corn (83.7%) or barley (87.7%). In lactating dairy cow rations, in vivo digestion was much greater with barley diets than ground corn diets (77.4 vs 48.6%, P<0.01; McCarthy et al. 1989), and with barley than sorghum diets (80 vs 49%, P<0.05; Herrera_Saldana et al. 1989). Sorghum (milo) generally has the lowest whole tract starch digestibility among cereal grains due to a hard peripheral endosperm layer which is resistant to digestion (Rooney et al. 1986). Waxy sorghums (cultivars whose starch is primarily amylopectin with little amylose) have consistently higher feeding value than normal sorghums and amylopectin content is positively related to starch digestibility (Rooney et al. 1986), but higher digestibilities of waxy cultivars may also be due to other characteristics such as less peripheral endosperm, larger starch granules within the peripheral endosperm, and different structure of the protein coat (Kotarski et al. 1992). Within a grain type, dry and wet processing methods are used to alter digestibility of grains fed to ruminants. Most methods increase both rate of starch fermentation and rumen starch digestibility. Grains with lower rumen digestibilities when unprocessed, such as sorghum, seem to respond to processing to the greatest extent (Theurer, 1986). Dry processing methods include grinding, dry rolling or cracking, popping, extruding, micronizing, roasting, and pelleting. These break the outer coat of the grain to allow access of rumen microorganisms and enzymes. Reducing particle size by cracking or grinding significantly increases the rate of starch digestion Starch digestion in the rumen In the ruminant, the majority of starch in the diet is fermented to volatile fatty acids in the rumen. Rumen fermentation varies with type of grain as well as conservation or processing method, and this variation can greatly affect animal performance. Optimal levels of starch in the diet, and rumen fermentability of that starch depend on animal and dietary factors. Methods of increasing rumen starch fermentation The extent of rumen fermentation of starch is determined by the rate at which starch is fermented and its retention time in the rumen, both of which vary with species and physiological status of animal, grain type, grain genotype, growing conditions, and physical and chemical processing method (Table 1; Herrera_Saldana et al. 1990b; Nocek et al. 1991; Huntington, 1997). Recent Advances in Animal Nutrition in Australia, Volume 13 (2001) 20 Knowlton, K.F. (McAllister et al. 1993), but also increases rate of passage from the rumen (Galyean et al. 1981; Ewing et al. 1986). Rate of passage is affected by animal characteristics and level of intake, and increased rate of passage decreases extent of digestion in the rumen. Fine grinding may, therefore, have less effect on in vivo rumen starch digestibility for animals with higher levels of intake, such as dairy cattle in early lactation, than for dry cows or dairy cattle in late lactation. In beef steers on high corn diets, increased rate of passage has been observed with steam flaked grains as well as with finely ground corn (Turner et al. 1995). The effect of grain processing on rate of passage in lactating cows has not been evaluated. Wet processing methods include the application of heat and/or moisture, such as soaking, steam rolling, steam processing and flaking, exploding, pressure cooking, and high moisture fermentation of early harvested or reconstituted grains. Interactions of heat, moisture, and pressure increase susceptibility of the starch to enzymatic attack and digestion by breaking down the endosperm structure, disrupting the protein matrix encapsulating starch granules in the endosperm, and gelatinizing the normally semi_crystalline starch granules (Kotarski et al. 1992). High moisture ensiling of early harvested or reconstituted grains increases the proportion of starch digested in the rumen of dairy cows (68 vs 78%; P<0.05; Aldrich et al. 1993). Steam flaking sorghum for dairy rations also consistently increases rumen starch digestion compared to dry ground or dry rolled sorghum (Poore et al. 1993a; Oliveira et al. 1995). Consistent increases in rumen starch digestion in lactating cows have also been observed for steam flaked corn compared to dry rolled corn (Plascencia et al. 1996). In steam flaked cereal grains, decreased density (incremental processing) of the flake increases rate of starch digestion (Nocek et al. 1991; Plascencia et al. 1996). Other factors that may influence rumen starch digestion include rumen protozoa (Mendoza et al. 1993), level of dry matter and starch intake (Russell et al. 1981b; Brink et al. 1985), and grain cultivar (Wester et al. 1992; Ladely et al. 1995). Protozoa reduce rate of starch digestion in the rumen through ingestion of starch_digesting bacteria, and ingestion of starch granules and sugars, and defaunation of sheep has been shown to increase rumen starch digestibility (Mendoza et al. 1993). Level of starch intake affects starch digestion: two studies demonstrated that increased corn in steer diets increased its digestion in the rumen (Russell et al. 1981b; Brink et al. 1985). Repeatable differences in in vitro starch disappearance have been observed among sorghum (Wester et al. 1992) and among corn grain (Ladely et al. 1995). Starch digestion in the small and large intestine While most starch in ruminant diets is fermented to VFA in the rumen, in high producing dairy cattle fed high grain diets, large quantities of dietary starch may escape rumen digestion and become available for digestion, fermentation, and absorption from the small Table 1 Starch content and rumen starch digestibility of cereal grains commonly fed to dair y cows (Herrera_Saldana et al. 1990b; Huntington1997). Percent starch, � SD 70.3 � 2.9 Rumen digestibility, percent of intake 88.3 88.1 80.7 � 3.9 84.6 92.7 94.0 76.2 � 7.9 84.8 � 4.1 72.1 89.9 49.5 59.8 � 12.0 78.4 73.2 70.0 Grain Wheat Processing method Dr y rolled Steam rolled Barley 64.3 � 3.3 Dr y rolled Steam rolled Oats 58.1 � 4.3 Dr y rolled Steam rolled Corn 76 � 1.8 Dr y rolled Steam flaked Steam rolled High moisture Ground Sorghum 71.3� 2.7 Dr y rolled Steam flaked High moisture ground Ground High grain diets for dairy cattle 21 and large intestine. Owens et al. (1986) observed that in 40 experiments with cattle, between 18 and 42% of dietary starch from corn and sorghum passed from the rumen intact. Higher starch intakes may promote rapid rates of passage from the rumen, causing a `washout' of grain from the rumen to the lower tract (Huntington, 1994). Recent studies with lactating dairy cows indicate that from 1 kg to nearly 5 kg of starch may disappear post_ruminally in cows on high starch diets (McCarthy et al. 1989; Herrera_Saldana et al. 1990a; Aldrich et al. 1993; Poore et al. 1993a; Oliveira et al. 1995; Plascencia et al. 1996; Knowlton et al. 1998). Given these significant flows of starch past the rumen, understanding the potential for starch digestion in the small and large intestine is important in predicting nutrient availability to the animal. Few studies with ruminants report starch flow at both the duodenum and the ileum, so partitioning starch digestion in the small intestine from fermentation in the large intestine is generally not possible. Small intestinal starch digestion Starch that escapes the rumen must be digested and absorbed in the lower tract to be of benefit to the cow, and the capacity of the ruminant small intestine for digestion of starch and the absorption of glucose may be limited. In general, increased flow of starch to the small intestine increases the quantity of starch digested, but decreases its digestibility there (Karr et al. 1966; Little et al. 1968; Kreikemeier et al. 1991). In steers fed wheat or sorghum based diets (Axe et al. 1987) duodenal starch flows were 518 and 1387 g/d, respectively. Small intestinal starch disappearance was 493 and 867 g/d, and small intestinal starch digestibility was 95% and 63% of that flowing into the small intestine. When Kreikemeier et al. (1991) infused raw corn starch to the abomasum of ilealy cannulated steers at 0, 20, 40 and 60 g/h into the abomasum of steers cannulated at the ileum, intestinal starch disappearance increased quadratically, with disappearance beginning to level off and starch flow past the ileum accelerating at the highest level of infusion (Kreikemeier et al. 1991). Starch in grains treated with heat and moisture to gelatinize the starch may be more digestible in the small intestine than that in dry grains. In lactating cows fed dry or high moisture corn, duodenal starch flows averaged 3108 and 1358 g/d respectively (Figure 1; Knowlton et al. 1998). Despite the much lower duodenal starch flow, disappearance of starch in the small intestine was much greater on the high moisture corn diets (836 g/d; 58% of duodenal starch flow) than on dry ground corn diets (412 g/d; 9% of duodenal flow). The infusion study of Kreikemeier et al. (1991) supports the suggestion that gelatinization of starch with heat and moisture increases its susceptibility to digestion in the small intestine. Infusion of similar levels of corn dextrin (starch partially hydrolyzed to destroy the granular structure) caused a linear increase in starch disappearance, suggesting that the crystalline structure of the raw corn starch inhibits small intestinal digestion in some way. At the highest levels of infusion, 55% of the infused raw corn starch and 79% of the infused dextrin disappeared (Kreikemeier et al. 1991). Starch granules in grain have a protective protein matrix surrounding them, so their digestibility in the small intestine is usually lower than the digestibility of purified starch or dextrin. In addition to the amount of starch flowing past the duodenum and its structure, protein status of the animal may affect small intestinal starch digestion. Increased flow of undegraded intake protein (UIP) to the small intestine appears to increase the digestion of starch there. In steers fed barley or corn based diets, small intestinal starch disappearance was 92 g/d for barley and 328 g/d for corn. Within the corn based diets, supplementation with corn gluten meal instead of soybean meal, which increased UIP, increased starch disappearance (Taniguchi et al. 1994). Digestibility as a percent of duodenal flow was numerically but not significantly higher for the corn_corn gluten meal treatment than for the other three. One interpretation is 100% 90% 80% % of starch intake 70% 60% 50% 40% 30% 20% 10% 0% Dry ground corn High moisture corn Excreted in feces Large intestine Small intestine Rumen Figure 1 Site of starch digestion in six lactating cows fed dry or high moisture cor n (Knowlton et al. 1998). 22 Knowlton, K.F. that protein source interacts with carbohydrate supply to influence small intestinal starch digestibility. Supporting this interpretation is the observation that net glucose absorption from the digestive tract was greater in steers with starch and casein infused to the abomasum than in steers with starch infused to the abomasum and casein infused to the rumen (33 vs 16 mmol/h; P<0.05; Taniguchi et al. 1994). Starch digestibility and animal performance Milk yield Although we can manipulate rumen starch digestion fairly easily, the effect on milk yield varies. Increased milk yield with increased rumen available starch was observed in early lactation cows fed barley in place of sorghum (Herrera_Saldana et al. 1989), fed steam flaked sorghum instead of dry rolled (Moore et al. 1992; Poore et al. 1993b), or fed ground corn instead of cracked corn (Knowlton et al. 1996a). Similar increases in milk yield were observed in mid_lactation cows fed corn ground more finely (Moe et al. 1976), fed steam flaked sorghum in place of dry rolled sorghum (Chen et al. 1994; Chen et al. 1995), fed high moisture shell corn in place of dry shell corn (McCaffree et al. 1968), and fed steam flaked corn instead of dry rolled (Plascencia et al. 1996). However diets with increased ruminally degraded starch did not affect milk or FCM yields in other studies (Clark et al. 1972; Clark et al. 1973; Bilodeau et al. 1989; Grings et al. 1992; Oliveira et al. 1993; Mitzner et al. 1994; Oliveira et al. 1995). Milk and FCM yields were decreased with higher ruminally degraded starch in early lactation cows fed barley instead of corn (McCarthy et al. 1989), fed dry rolled barley versus dry ground corn (Casper et al. 1990), fed high moisture shell corn in place of dry ear corn (Aldrich et al. 1993), and in mid_lactation cows fed untreated barley instead of the more slowly digested ammoniated barley (Robinson et al. 1989). Starch fermentation in the large intestine With large quantities of starch bypassing the rumen in lactating cows on high starch diets, and some limitation of small intestinal starch digestion, there may be a significant amount of starch being fermented in the large intestine. The pathways of digestion in the large intestine are similar to those in the rumen, with VFA, methane, CO2, and microbial biomass as endproducts; volatile fatty acids can be absorbed and utilized, but microbial protein produced there is not absorbed (Armstrong et al. 1979). �rskov et al. (1970) reported that up to 138 g of the 300 g of starch infused to the caecum of sheep fed grass hay disappeared, indicating limited but substantial capacity for fermentation in the large intestine. Increased viable counts of bacteria in caecal fluid collected from the infused sheep, and increased nitrogen excretion with increasing starch infused to the caecum, clearly indicated increased microbial nitrogen synthesis and subsequent loss of that nitrogen to the faeces. Steers fed increasing levels of corn had an increasing proportion of starch disappear from the large intestine (Karr et al. 1966). Up to 11% of total starch digested on the 80% corn diet was digested in the large intestine. In steers on sorghum diets 280 g starch/d (>10% of starch intake) disappeard from the large intestine (Axe et al. 1987), and steers on corn based diets 109 g starch/d was digested in the large intestine. This is nearly 10% of starch intake, and apparent total tract nitrogen digestibility was depressed (Taniguchi et al. 1994) indicating increased microbial fermentation in the large intestine. In high producing dairy cows, few data on disappearance of starch from the large intestine are available, but starch intakes and passage rates are high. Given the apparent limits to small intestinal starch digestion, starch flow to and fermentation in the large intestine is likely significant, increasing nitrogen losses on these diets. In one study, lactating cows fed dry ground corn had much greater duodenal and ileal starch flows (3109 and 2696 g/d) than cows fed high moisture corn (1358 and 523 g/d, Figure 1; Knowlton et al. 1998). Small intestinal disappearance of starch in cows fed the dry ground corn was low, but in the large intestine, more than 1800 g/d of starch disappeared; 23% of the starch in the dry ground corn diet disappeared from the large intestine. Faecal nitrogen excretion increased by 11% in cows fed dry ground corn compared with those fed high moisture corn. Feed intake Decreased dry matter intake (DMI) with more rapidly fermented starch sources has been observed in early lactation cows fed barley instead of corn, (McCarthy et al. 1989; Casper et al. 1990), fed steam flaked sorghum in place of dry rolled sorghum (Moore et al. 1992), and fed high moisture shell corn in place of dry ear corn (Aldrich et al. 1993) or dry shell corn (McCaffree et al. 1968; Clark et al. 1972). Similar reductions in DMI with increased rumen starch digestion were observed in mid_lactation cows fed dry ground corn versus dry ground sorghum (Mitzner et al. 1994) and fed steam flaked sorghum in place of dry rolled sorghum (Oliveira et al. 1993). Other studies, however, showed no change in intake with altered rumen digestion of starch (Clark et al. 1973; Bilodeau et al. 1989; Herrera_Saldana et al. 1989; Grings et al. 1992; Poore et al. 1993b; Chen et al. 1995; Oliveira et al. 1995; Knowlton et al. 1996a). In contrast, in one study with sorghum and one with corn, steam flaking has been shown to improve intake relative to dry rolling (Chen et al. 1994; Plascencia et al. 1996), and grinding corn increased intake in mid lactation cows relative to whole corn (Moe et al. 1976). High grain diets for dairy cattle 23 The improvement in DMI observed in some studies with diets with less ruminally degraded starch may be due to a number of factors. One possible mechanism is less rapid post_prandial accumulation of acid in the rumen and a minimization of rumen pH fluctuation. Excess fermentation of starch to volatile fatty acids in the rumen may overwhelm the buffering and absorptive capacity of the rumen, leading to fluctuations in rumen pH that may decrease DMI in the high producing dairy cow (Robinson et al. 1988; McCarthy et al. 1989). If fluctuation in rumen pH is an appropriate explanation for the decreased DMI with increased rumen digestion of starch observed in some studies, it is more likely to be a factor with higher starch intakes, lower fibre concentrations, or more variation in forage quality. Six studies have been reported with lactating cows where both DMI and rumen starch digestion have been measured. Three studies indicate no effect of rumen starch digestion on DMI, and the quantity of starch digested in the rumen on the high digestibility diet ranged from 4700 to 5400 g/d (Herrera_Saldana et al. 1990a; Poore et al. 1993a; Oliveira et al. 1995). The two studies indicating depression in intakes had rumen starch digestion in excess of 6500 g/d. One study (Plascencia et al. 1996) indicated an increase in DMI with steam flaked corn compared to dry rolled corn when rumen starch digestion was relatively low (2700 vs 1900 g/d). The increase in DMI in this and other studies with ground or steam flaked corn (Moe et al. 1976; Chen et al. 1994) may also be explained by changes in rate of passage observed with these processing methods (Galyean et al. 1981; Ewing et al. 1986; Turner et al. 1995). If bulk fill is limiting intake in these cows, increased rates of digestion and passage with ground or steam flaked grain may allow higher intakes. Another reason for the disparity in research results is the type of grain that is processed. As previously observed, when cows are fed grains with lower intrinsic rumen fermentability (e.g. sorghum), responses to processing are greater. Of the studies indicating improvements in milk yield with increased rumen degradable starch, most are with sorghum grain that compare steam flaking to dry rolling (Moore et al. 1992; Poore et al. 1993b; Chen et al. 1994; Chen et al. 1995). glucose precursors; increased synthesis of glucose elicits an insulin response, and both effects can increase milk protein concentration. Animal health and productivity Increasing grain content of diets increases concerns about the health and productivity of animals fed these diets. Rumen acidosis (acute and sub_acute), impaired fibre digestion, and milk fat depression dominate any discussion of the effect of high grain diets on the health of dairy cows. Rumen acidosis Acute acidosis is primarily associated with the accumulation of lactate, and is more commonly observed in beef cattle on 80_100% grain diets than in lactating dairy cows. Lactate is a major fermentation end product of starch on these very high starch diets. Lactate has a pKa of 3.86 while acetate, propionate and butyrate have pKa's around 4.8. When rumen pH declines, lactate may cause a more dramatic decline in pH than the VFA due to this lower pKa. Cattle abruptly switched to high concentrate diets often accumulate lactate in the rumen, which decreases rumen pH and causes acute acidosis (Slyter, 1976). Lactate production increases because large quantities of readily fermentable carbohydrate are available. Normally, lactate production is not advantageous in the rumen, because each sugar fermented to lactate yields only 2 ATP. When carbohydrate is limiting, production of other VFA is more efficient. When carbohydrate is plentiful, however, microbes capable of rapid growth which often produce lactate are favoured (Russell et al. 1981a). The decline in rumen pH to below 5 caused by lactate accumulation then creates a niche for Lactobacillus species, lactate producers which are resistant to low rumen pH (Slyter, 1976; Russell et al. 1983; Russell et al. 1985). The extent of lactate accumulation equation is also affected by the activity of lactate utilizers which may take time to adjust to changes in diet and become less competitive as pH declines. That these will adjust to high grain diets is supported by feedlot data that show cattle developing mild acidosis and going off feed before adapting to a high concentrate diet (Kunkle et al. 1976). A decline in rumen pH to below 5, or an increase in liquid dilution rate makes most lactate utilizers uncompetitive and creates a niche for Lactobacillus (Slyter, 1976; Russell et al. 1983; Russell et al. 1985). Lactate, with pKa 3.86, is likely to affect rumen pH when it accumulates, and this spiraling effect of increasing lactate production and decreasing pH contributes to the development of acute acidosis. Symptoms and consequences of acute acidosis may include dullness, diarrhoea, hyperventilation, dehydration, inhibited muscular activity of the rumen wall, parakeratosis (a thickening of the rumen wall Milk protein production Increasing the energy content of the diet by increasing grain feeding generally results in an increase in milk protein concentration (Poore et al. 1993b; Chen et al. 1994) that can be attributed to greater microbial protein synthesis promoted by an increased supply of ruminally fermented carbohydrates. Recent research suggests that increased feeding of fermentable starch may also increase milk protein concentration through changes in whole body glucose and insulin response (McGuire et al. 1995). Increased digestion of starch in the rumen generally increases supplies of propionate and other 24 Knowlton, K.F. which decreases VFA absorption), reduced salivation, reduced intestinal motility, damage to the rumen, damage to the liver, and death. Sub_acute acidosis is more common in lactating cows than is acute acidosis, and is primarily associated with production of VFA in excess of the rumen's capacity for absorption and buffering. Consequences of sub_acute acidosis are reduced feed intake, laminitis, and reduced milk yield (Slyter, 1976; Nagaraja et al. 1985). Diets high in ruminally degraded starch increase total ruminal VFA concentrations (McCarthy et al. 1989; Knowlton et al. 1998). Cows fed ground corn diets had mean rumen pH similar to cows fed cracked corn diets (Knowlton et al. 1996b), but the range in pH within a day was increased with ground corn. Reductions in rumen pH have been demonstrated in animals fed ground and pelleted barley as compared to those fed whole barley (�rskov et al. 1975; �rskov et al. 1978) or NaOH treated barley (�rskov et al. 1978). Treatment of high moisture barley with increasing levels of ammonia to decrease digestibility reduced the rate of pH decline following meals in lactating cows (Robinson et al. 1989). A reduction in feed intake is often observed with increased rumen starch digestion and increased VFA concentration. One explanation is that low rumen pH reduces fibre digestibility (see below), which would increase the retention of fibre in the rumen and affect feed intake because of limitations associated with rumen fill. Alternatively, the rapid accumulation of acid, per se, may decrease intake. Treatment of grass silage with sodium bicarbonate increased silage pH from 4.0 to 5.4, and increased subsequent DMI of sheep and cattle by 9.7 to 20.7% (McLeod et al. 1970). Addition of lactic acid to grass silage reduced pH from 5.4 to 3.8, and reduced subsequent DMI of sheep by 22%. Receptors sensitive to changes in pH have been isolated in the rumen epithelium (Baile and Della_Fera, 1981). As rumen pH falls below 5 or 5.5, feed intake is most likely depressed because of rumen stasis associated with acute acidosis. Rumen pH is not likely to be a physiological controller of intake, but rather a primary cause of reduced intake in pathological situations. Chemical treatments of grain have been used to reduce rumen starch digestion in an attempt to avoid acidosis while retaining the benefits of high grain diets. Treatment of whole barley with anhydrous ammonia decreased rumen organic matter digestion in beef cattle by 11% but increased small intestinal digestion with no effect of treatment on total tract digestibility (Mandell et al. 1988). Similar ef fects were found with formaldehyde treatment of ground corn which reduced rumen starch digestion by up to 41% with no reduction in whole tract starch digestibility (Fluharty et al. 1989; Oke et al. 1991). Adoption of these practices in the field has been limited because of human health and safety concerns. Fibre digestion on high grain diets Decreased fibre digestibility is typically observed with diets containing rapidly digested starch. The rapid accumulation of acids reduces rumen pH which impairs the growth of cellulolytic bacteria and slows forage cell wall digestion (McCarthy et al. 1989; Grant, 1994). Impaired fibre digestion may lead to reductions in feed intake. Exceptions have been noted: in studies where grains with low digestibility (e.g. sorghum) were processed, increases in rumen fibre digestibility were reported (Poore et al. 1993a; Chen et al. 1994). These apparently contradictory observations are best explained as the result of making poorly available starch more available, consequently improving overall microbial activity in the rumen. Depression in fibre digestibility is the more common result of increasing ruminally fermented starch, especially when relatively available starch sources (dried ground corn or barley) are processed to increase further the availability (by steam flaking or steam rolling). Milk fat depression on high grain diets Increasing grain content of the diet usually decreases milk fat concentration, a change commonly explained by changes in rumen fermentation (McClymont 1995). Diets containing high amounts of ruminally degraded starch typically increase rumen propionate concentrations and decrease acetate to propionate ratio (McCarthy et al. 1989; Moore et al. 1992; Aldrich et al. 1993; Oliveira et al. 1993; Knowlton et al. 1996b; Plascencia et al. 1996). Propionate is a precursor for the synthesis of glucose by the liver, while acetate is a precursor for the synthesis of fat by the mammary gland and body tissues. The decrease in milk fat with high grain diets has, therefore, been classically explained by a general shift from lipogenesis to gluconeogenesis. Recent research, however, suggests that the depression in milk fat may also be caused by an accumulation of trans_fatty acids in the rumen that is a consequence of a low rumen pH with high grain diets (Gaynor et al. 1995; Kalscheur et al. 1997b; Kalscheur et al. 1997a). Practical aspects of feeding high grain diets Total mixed rations vs meal feeding and frequency of feeding The intake of concentrates reduces the intake of forage and the forage:concentrate ratio. This reduced intake of forages may be due to limited capacity for feed intake, limitations on total energy required, and/or effects of grain feeding on rumen fermentation. Despite decreases in forage intake, total feed intake and intake of energy usually increases with grain supplementation. The response to starch supplementation depends on High grain diets for dairy cattle 25 the physiological state of the cow, the type of forage and grains fed, and how the grain is provided (Reynolds et al. 1997). Addition of grain may be through a total mixed ration (TMR), or by allowing ad libitum intake of forages with a fixed quantity of grain. When starchy grains (corn, oats) are added to the diet using the latter approach, forage intake is typically reduced by 0.4_0.5 kg/kg of grain added (Sutton et al. 1996; Reynolds et al. 1997). More frequent feeding of silage minimizes this reduction. When addition of grain was through a TMR instead of being fed separately, silage consumption decreased by just 0.3 kg/kg concentrate fed (Reynolds et al. 1997). Several aspects of TMR feeding explain this difference in response to grain addition. When grain is included in a TMR, cows are less able to discriminate between feed ingredients, preventing selective consumption of grain in preference to forage. In addition, patterns of rumen fermentation are different in cows fed a TMR from those in cows fed grain and forage separately. When grain is given separately, the fermentation of large quantities of starch in a short period of time can lead to production of VFA that exceeds the capacity of the rumen for absorption and buffering. Sutton et al. (1986) demonstrated that feeding a TMR in two meals caused much greater fluctuation in rumen pH, and decreased minimum rumen pH, compared with feeding the same TMR in 24 hourly meals. When hay and grain were given separately in this series of low_forage experiments (Sutton et al. 1986), more frequent feeding of grain (5_6 times a day) resulted in a smaller depression in milk fat than when it was provided twice daily, and reduced circulating NEFA. In another series of experiments, lactate concentrations peaked after feeding, and these concentrations were higher in cows fed 12 kg of grain once a day than in cows fed the same total amount of grain in two or four feeds (Malestein et al. 1981). and is, therefore, equally effective in alleviating digestive upsets. Conclusions Grains vary in starch conten