HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bargo, F. Articles by Delahoy, J. E. Search for Related Content PubMed PubMed Citation Articles by Bargo, F. Articles by Delahoy, J. E.

Invited Review: Production and Digestion of Supplemented Dairy Cows on Pasture

F. Bargo^*,1, L. D. Muller^*, E. S. Kolver and J. E. Delahoy^*

^* Department of Dairy and Animal Science, The Pennsylvania State University, University Park, PA 16802
Dexcel Ltd., Private Bag 3221, Hamilton, New Zealand

Corresponding author:
L. D. Muller; e-mail:
lmuller{at}psu.edu.

	ABSTRACT

TOP ABSTRACT INTRODUCTION DRY MATTER INTAKE OF... GRAZING BEHAVIOR SUBSTITUTION RATE AND MILK... EFFECT OF SUPPLEMENTATION ON... RUMINAL DIGESTION AND... PASTURE IN SITU DIGESTION POSTRUMINAL DIGESTION OF GRAZING... CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Literature with data from dairy cows on pasture was reviewed to evaluate the effects of supplementation on intake, milk production and composition, and ruminal and postruminal digestion. Low dry matter intake (DMI) of pasture has been identified as a major factor limiting milk production by high producing dairy cows. Pasture DMI in grazing cows is a function of grazing time, biting rate, and bite mass. Concentrate supplementation did not affect biting rate (58 bites/min) or bite mass (0.47 g of DM/bite) but reduced grazing time 12 min/d per kilogram of concentrate compared with unsupplemented cows (574 min/d). Substitution rate, or the reduction in pasture DMI per kilogram of concentrate, is a factor which may explain the variation in milk response to supplementation. A negative relationship exists between substitution rate and milk response; the lower the substitution rate the higher the milk response to supplements. Milk production increases linearly as the amount of concentrate increases from 1.2 to 10 kg DM/d, with an overall milk response of 1 kg milk/kg concentrate. Compared with pasture-only diets, increasing the amount of concentrate supplementation up to 10 kg DM/d increased total DMI 24%, milk production 22%, and milk protein percentage 4%, but reduced milk fat percentage 6%. Compared with dry ground corn, supplementation with nonforage fiber sources or processed corn did not affect total DMI, milk production, or milk composition. Replacing ruminal degradable protein sources with ruminal undegradable protein sources in concentrates did not consistently affect milk production or composition. Forage supplementation did not affect production when substitution rate was high. Fat supplementation increased milk production by 6%, without affecting milk fat and protein content. Increasing concentrate from 1.1 to 10 kg DM/d reduced ruminal pH 0.08 and NH₃-N concentration 6.59 mg/dl, compared with pasture-only diets. Replacing dry corn by high moisture corn, steam-flaked or steam-rolled corn, barley, or fiber-based concentrates reduced ruminal NH₃-N concentration 4.36 mg/dl. Supplementation did not affect in situ pasture digestion, except for a reduction in rate of degradation when high amounts of concentrate were supplemented. Supplementation with energy concentrates reduced digestibility of neutral detergent fiber and intake of N but did not affect digestibility of organic matter or flow of microbial N.

Key Words: high producing dairy cow • pasture • supplementation • milk production and composition

Abbreviation key: CDMI = concentrate DMI, CBW = change in BW, ED = effective degradability, FPr = fat percentage reduction, FYi = fat yield increase, HM = herbage mass, LEG = percentage of legumes in pasture, MN = microbial nitrogen, MPi = milk production increase, MR = milk response, MY = milk yield, NANMN = nonammonia, nonmicrobial nitrogen, NDFp = NDF in pasture available, NDFs = NDF in pasture selected, PA = pasture allowance, PASUP = pasture allowance and total supplementation interaction, PD = potentially degradable fraction, PDMI = pasture DMI, PDMIr = pasture DMI reduction, PPi = protein percentage increase, PYi = protein yield increase, RAD = ruminal apparent digestibility as proportion of intake, RADD = ruminal apparent digestibility as proportion of total tract apparent digestibility, RUPI = RUP intake, SHT = sward height, SR = substitution rate, SUP = total supplementation, TB = number of bites per d, TDMI = total DMI, TDMIi = total DMI increase, TOMI = total OM intake, TTAD = total tract apparent digestibility, WOL = week of lactation

	INTRODUCTION

TOP ABSTRACT INTRODUCTION DRY MATTER INTAKE OF... GRAZING BEHAVIOR SUBSTITUTION RATE AND MILK... EFFECT OF SUPPLEMENTATION ON... RUMINAL DIGESTION AND... PASTURE IN SITU DIGESTION POSTRUMINAL DIGESTION OF GRAZING... CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
The use of pasture for dairy cows results in lower-cost feeding systems because grazed forage is the cheapest source of nutrients (Clark and Kanneganti, 1998; Peyraud and Delaby, 2001). Efficient pasture-based systems are characterized by high milk output per unit of land, while confinement systems are characterized by high milk output per cow (Clark and Kanneganti, 1998). Pastures used for dairy cows are commonly based on temperate species, and are described as high quality or young and leafy pastures with 18 to 24% DM, 18 to 25% CP, 40 to 50% NDF, and 1.53 to 1.67 Mcal/kg DM of NE_L (Clark and Kanneganti, 1998). Muller and Fales (1998) reported a range of 18 to 25% CP, 40 to 55% NDF, and 1.55 to 1.70 Mcal/kg DM of NE_L for well-managed grass pastures typically grazed in the Northeastern United States.

In the United States, the dairy industry over the last 50 yr has been characterized by a favorable milk price:feed cost ratio; therefore, dairy systems have tended to focus on high milk production per cow (Clark and Kanneganti, 1998; Muller and Fales, 1998) and less use of pasture-based systems until recent years (Muller and Fales, 1998). Average milk production per cow in the United States increased from 3191 kg in 1960 to 8263 kg in 2000 (NRC, 2001). Besides the high milk production per cow, climatic conditions (i.e., cold temperatures and snow cover during 4 or 5 mo/yr) in many important dairy areas in the United States (e.g., Midwest and Northeast) do not permit year-round grazing systems (Clark and Kanneganti, 1998). Therefore, the use of feeding systems combining pasture plus additional feed supplements such as concentrates and conserved forage are required.

The main objective of supplementation of grazing dairy cows is to increase total DMI and energy intake relative to that achieved with pasture-only diets (Peyraud and Delaby, 2001; Stockdale, 2000b). For the production system, a primary goal of supplementation is to optimize profit per cow and per unit of land (Kellaway and Porta, 1993; Fales et al., 1995). The objectives of supplementation include (Kellaway and Porta, 1993): 1) increase milk production per cow, 2) increase stocking rate and milk production per unit of land, 3) improve the use of pasture with the higher stocking rate, 4) maintain or improve BCS to improve reproduction during pasture shortage, 5) increase length of lactation during periods of pasture shortage, and 6) increase milk protein content by energy supplementation.

Previous reviews of grazing research have focused on animal production and digestion aspects (Leaver, 1985; Kellaway and Porta, 1993; Doyle et al., 1996; Stockdale, 2000b). However, most of these reviews have focused on research with relatively low producing dairy cows. Appropriate strategies for supplementation of high producing dairy cows requires an understanding of the effect of different types of supplements on DMI, animal performance, and digestion, and of providing nutrients that complement the nutrient content of pasture and meet the nutrient requirements of dairy cows.

The objective of this review is to summarize the effect of supplementation on grazing behavior, pasture and total DMI, milk production, milk composition, and ruminal and postruminal digestion of high producing dairy cows on pasture. For the purpose of this review, high producing dairy cows are defined as those producing more than 25 kg/d of milk in early lactation or about 20 kg/d in late lactation. However, research data from low producing dairy cows are included in those areas where information with high producing dairy cows is not available. This review focused on research data from the United States, but because of limited published research in many areas, information was included from other countries where grazing systems are important (e.g., Argentina, Australia, France, Ireland, Netherlands, New Zealand, and United Kingdom).

	DRY MATTER INTAKE OF GRAZING COWS

TOP ABSTRACT INTRODUCTION DRY MATTER INTAKE OF... GRAZING BEHAVIOR SUBSTITUTION RATE AND MILK... EFFECT OF SUPPLEMENTATION ON... RUMINAL DIGESTION AND... PASTURE IN SITU DIGESTION POSTRUMINAL DIGESTION OF GRAZING... CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Dry Matter Intake of Grazing Cows on Pasture-Only Diets
Different theories about the control of DMI in ruminants have been presented (Forbes, 1995); however, detailed description of these theories exceeds the objectives of this review. Hodgson and Brookes (1999) described three factors affecting pasture DMI of grazing cows: 1) "feeding drive" or nutrient requirements of the cow; 2) "physical satiety" or factors associated with distension of the alimentary tract; and 3) "behavioral constraints" or limits to the potential pasture DMI resulting from the combination of pasture and animal factors affecting grazing behavior.

Low pasture DMI has been identified as a major factor limiting milk production of high producing cows with a grazing system (Leaver, 1985; McGilloway and Mayne, 1996; Kolver and Muller, 1998). Leaver (1985) suggested that high producing dairy cows fed pasture-only diets could reach a total DMI of 3.25% of BW. Mayne and Wright (1988) estimated that with no pasture quantity and quality restrictions, pasture DMI of high yielding dairy cows might reach 3.5% of BW. Beever and Thorp (1997) proposed that total DMI of high producing cows fed pasture-only diets is lower than for cows fed pasture diets plus concentrates. This may be explained by physical constraints, rate of forage removal from the rumen, and water consumption associated with pasture.

Studies in the United States with high producing cows fed pasture-only diets are limited (Kolver and Muller, 1998; Reis and Combs, 2000b; Bargo et al., 2002a). Kolver and Muller (1998) reported that early-lactation cows grazing high quality grass pasture in the spring had a pasture DMI of 19.0 kg/d, or 3.4% of BW. However, when compared with cows fed a nutritionally balanced TMR ration, grazing cows consumed 4.4 kg less DM. The intakes of DM and NE_L were lower on the pasture-only diet; however, intakes of CP and NDF did not differ between the pasture-only diet and TMR. The difference in DMI, rather than energy content of pasture, appeared to be the major factor responsible for the lower total energy intake and milk production (Kolver and Muller, 1998). Pasture DMI of unsupplemented dairy cows increased from 17.7 kg/d or 2.9% of BW to 20.5 kg/d or 3.4% of BW as pasture allowance (PA) increased from 25 to 40 kg DM/cow per day (Bargo et al., 2002a). Although Dalley et al. (2001) suggested that pasture DMI may be increased by more frequent allocation of new pasture, they reported no difference in DMI (15.8 kg/d) or milk production (25.3 kg/d) of early lactation cows grazing a ryegrass pasture when offered one or six times per day.

Effect of Pasture Allowance
Many pasture factors affect DMI (Poppi et al., 1987; Hodgson and Brookes, 1999) including pregrazing pasture mass (amount of pasture per unit area; kg DM/ha) and PA (amount of pasture offered per cow; kg DM/cow per day). Several researchers (Stockdale, 1985; Dalley et al., 1999) have reported that pasture DMI is closely related to PA. The relationships between pasture DMI and PA have been described as asymptotic (Poppi et al., 1987; Peyraud et al., 1996; Dalley et al., 1999). However, it is unclear what PA is required to maximize DMI. In a review, Leaver (1985) proposed a maximum DMI at a PA between 45 to 55 g DM/kg of BW or 27 to 33 kg DM/cow per day for a 600-kg cow. Pasture DMI increased as PA increased, but at a declining rate with a plateau when PA was 10 to 12% of BW or 60 to 72 kg DM/cow per day for a 600-kg BW cow (Hodgson and Brookes, 1999). Data from Australia (Doyle et al., 1996) showed that pasture DMI continues to increase as PA increases up to 15 kg DM/100 kg of BW or 90 kg DM/cow per day for a 600-kg BW dairy cow. Pasture DMI increased curvilinearly from 11.2 to 18.5 kg DM/cow per day as PA increased from 20 to 70 kg DM/cow per day, with a plateau at a PA of 55.2 kg DM/cow per day (Dalley et al., 1999). Peyraud et al. (1996) reported a curvilinear relationship between pasture DMI and PA from 20 to 40 kg DM/cow per day, with pasture DMI reaching a plateau at a PA of 32.6 kg DM/cow per day. Wales et al. (1999) reported that as PA increased from 20 to 70 kg DM/cow per day, pasture DMI increased linearly from 7.1 to 16.2 kg DM/cow per day with a pregrazing pasture mass of 3100 kg DM/ha, and from 9.9 to 19.3 kg DM/cow per day with a pregrazing pasture mass of 4900 kg DM/ha.

Recent research has studied the effect of PA on pasture DMI of high producing dairy cows with no supplementation. Pasture DMI of dairy cows grazing an orchardgrass pasture was 17.5 and 20.6 kg DM/cow per day at low (25 kg DM/cow per day) and high (40 kg DM/cow per day) PA, respectively (Bargo et al., 2002a). In two experiments that measured PA at a 5-cm cutting height (Delaby et al., 2001), pasture DMI increased from 11.3 to 13.0 kg/cow per day as PA increased from 12.1 to 15.8 kg of DM/cow per day, and from 12.9 to 15.0 kg/cow per day as PA increased from 16.5 to 21.0 kg DM/cow per day. Stockdale (2000a) reported an increase in pasture DMI from 14.3 to 19.3 kg/d when PA of a ryegrass pasture was increased from 26.7 to 53.5 kg DM/cow per day. Dalley et al. (2001) also reported an increase of pasture DMI from 13.6 to 17.9 kg/d as the PA of a ryegrass pasture increased from 40 to 65 kg DM/cow per day. Pasture DMI by high producing dairy cows in early lactation increased from 11.2 to 15.6 kg/d when PA of a ryegrass pasture was increased from 19 to 37 kg DM/cow per day (Wales et al., 2001).

In summary, over a range of PA from 20 to 70 kg DM/cow per day, pasture DMI increased 0.19 kg/kg of increased PA (range: 0.17 to 0.24 kg/kg). Data from seven studies (Peyraud et al., 1996; Dalley et al., 1999; 2001; Stockdale, 2000a; Delaby et al., 2001; Wales et al., 2001; Bargo et al., 2002a) were used to describe the relationship between pasture DMI and PA for dairy cows on pasture-only diets. In those studies, cows ranged from 19 to 182 DIM and produced from 23.0 to 45.8 kg/d of milk at the start of the experiment, grazed at a PA from 12.1 to 70 kg DM/cow per day, and consumed 6.7 to 20.5 kg DM/cow per day of pasture. Observations were weighted as described by St-Pierre (2001) to account for unequal replications and variances of the means across studies. Pasture allowance and its quadratic term were considered as independent variables. Parameter estimates for the final equation were obtained using a mixed model approach (i.e., trial was considered a random effect) using the MIXED procedure of SAS (1999). The regression analysis for pasture DMI (PDMI, kg/d) resulted in a best-fit model that included terms for PA (kg DM/cow per day) and its quadratic term: PDMI = 7.79 (SE 1.49) + 0.26 (SE 0.06) PA - 0.0012 (SE 0.0007) PA²; R² = 0.95. Based on this equation, the optimum PA to maximize pasture DMI (21.9 kg/d) is reached at 110 kg DM/cow per day, and pasture DMI increased 0.26 kg/kg of increase in PA up to 110 kg DM/cow per day.

If the goal is to maximize pasture DMI of high producing dairy cows, management must ensure unrestricted pasture quality and quantity, which is only found for short periods of time during the spring. Unrestricted pasture conditions (i.e., high PA) also implies low pasture utilization (pasture DMI/PA < 50%; McGilloway and Mayne, 1996). The use of very high PA might also result in deterioration of pasture quality as the season progresses because of the increase in residual pasture height (Peyraud and Delaby, 2001). The studies reviewed indicate that maximum pasture DMI is achieved when PA is between 3 to 5 times the DMI, which is in agreement with the regression described above. However, even under unrestrictive pasture conditions, total DMI amounts achieved by high producing dairy cows are lower than those by cows consuming TMR (Kolver and Muller, 1998) or pasture plus supplements (Bargo et al., 2002aBargo et al., 2002b). Because of low pasture utilization and deterioration of pasture quality at high PA, a practical recommendation is to provide a PA of 2 times the expected pasture DMI or 25 kg DM/cow per day of PA when cows are also fed supplements (Bargo et al., 2002a).

Methods and Equations to Estimate DMI in Grazing Cows
Estimation of DMI in grazing cows is more difficult and less accurate when compared with the determination of DMI by cows on confinement systems. An extensive review of the different methods and techniques to estimate DMI in grazing cows was published by Leaver (1982). Techniques may be classified as either pasture- or animal-based (Meijs et al., 1982). An extensive review of pasture measurement techniques can be found in Mannetje (2000). The main disadvantage of pasture-based techniques is that pasture DMI is estimated as a group and not individually. The most common animal-based technique used is based on the estimation of fecal production and diet digestibility (Le Du and Penning, 1982; Peyraud, 1998): DMI = fecal production/(1 - digestibility of the diet). Fecal production is estimated using markers such as chromium oxide (Peyraud, 1998) and alkanes (Dove and Mayes, 1991; Dove and Mayes, 1996). A comparison between those two methods has been reported by Mallossini et al. (1996), who concluded that estimation of pasture DMI was similar if a 95.5% recovery is assumed for chromium oxide.

Because estimation of DMI by grazing cows demands the use of labor-intensive and indirect techniques that have several sources of error, equations based on animal and pasture characteristics have been developed to predict DMI of grazing cows (Caird and Holmes, 1986; Vazquez and Smith, 2000). Caird and Holmes (1986) used data from nine experiments conducted with cows grazing ryegrass, consuming 1.2 kg/d of concentrate, and producing 21.5 kg/d of milk on average to predict total DMI. Animal variables included total OM intake (TOMI, kg/d), herbage OM intake, concentrate DMI (CDMI, kg/d), BW (kg), milk yield (MY, kg/d), herbage OM digestibility, and week of lactation. Pasture variables included herbage mass (HM, tonne of OM/ha), PA (kg OM/cow per day), and sward height (SHT, cm). For rotationally grazed cows the best equation (R² = 0.68) was: TOMI = 0.323 + 0.177MY + 0.010BW + 1.636CDMI - 1.008HM + 0.540PA - 0.006PA² - 0.048PA x CMDI.

Vazquez and Smith (2000) used data from 27 grazing studies with dairy cows to obtain regression equations to predict total and pasture DMI. Mean milk production and supplementation amount were 15.9 and 1.9 kg/d, respectively. Independent variables included 4% FCM (kg/d), days since calving, PA (kg DM), NDF in pasture available (NDFp, % DM), NDF in pasture selected (NDFs, % DM), percentage of legumes in pasture (LEG, %), amount of concentrate supplemented (kg DM), amount of forage supplemented (kg DM), total supplementation (SUP, kg DM), PA and total supplementation interaction (PASUP), BW (kg), and change in BW (CBW, kg/d). The best equation (R² = 0.95) for total DMI (TDMI) estimation was: TDMI = 4.47 + 0.14FCM + 0.024BW + 2.00CBW + 0.04PA + 0.022PASUP + 0.10SUP - 0.13NDFp - 0.037LEG. The best equation to estimate PDMI (R² = 0.91) was: PDMI = 4.47 + 0.14FCM + 0.024BW + 2.00CBW + 0.04PA + 0.022PASUP - 0.90SUP - 0.13NDFp - 0.037LEG.

Equations developed by Caird and Holmes (1986) and Vazquez and Smith (2000) differ from the equation presented by NRC (2001) to estimate DMI. While those equations included pasture and supplement variables, the NRC (2001) equation is based only on animal variables such as FCM (kg/d), BW (kg), and week of lactation (WOL): DMI = (0.372 x FCM + 0.0968 x BW^0.75) x (1 - e ^{(-0.192 x (WOL + 3.67))}). We used a dataset of 56 measures from Bargo et al. (2002b), who measured DMI four times during the grazing season using Cr₂O₃ as a fecal marker in dairy cows that grazed an orchardgrass pasture and were supplemented with 8.7 kg/d of a corn-based concentrate. Cows, pasture, and supplement information reported in that study (Bargo et al., 2002b) were used to estimate DMI with the equations of Caird and Holmes (1986), Vazquez and Smith (2000), and NRC (2001). Total DMI estimated by the equations of NRC (2001) (21.9 kg/d) or Caird and Holmes (1986) (21.2 kg/d) did not differ from DMI measured by Cr₂O₃ (21.6 kg/d) (P > 0.05), but estimation of DMI by the equation of Vazquez and Smith (2000) (24.4 kg/d) was higher than measured DMI (P < 0.05). This indicates that estimation of DMI using the equations of Caird and Holmes (1986) and NRC (2001) was accurate for this particular dataset with high producing dairy cows, with the advantage that the NRC (2001) equation is simpler and requires only animal factors.

	GRAZING BEHAVIOR

TOP ABSTRACT INTRODUCTION DRY MATTER INTAKE OF... GRAZING BEHAVIOR SUBSTITUTION RATE AND MILK... EFFECT OF SUPPLEMENTATION ON... RUMINAL DIGESTION AND... PASTURE IN SITU DIGESTION POSTRUMINAL DIGESTION OF GRAZING... CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Methodology
Pasture DMI of grazing cows can be expressed as the product of grazing time (min/d), biting rate (bites/min), and bite mass (g DM/bite; Hodgson and Brookes, 1999; Rook, 2000). Few grazing behavior studies have been conducted with high producing dairy cows under practical or production conditions, probably because of methodological limitations. Bite mass can be measured directly by using esophageally fistulated animals or indirectly dividing pasture DMI by the total bites/d (Forbes, 1988; Rook, 2000). Because use of esophageally fistulated animals is expensive, and may compromise animal welfare and normal behavior, bite mass is often calculated indirectly (Rook, 2000). Both biting rate and grazing time can be measured visually or automatically (Forbes, 1988). Visual estimation of biting rate requires the recording of head movements and sound associated with pasture prehension. Automatic methods for biting rate are based on recording jaw and sometimes head movements. Visual grazing time measurements are based on recording grazing activity at different intervals (e.g., 5 to 10 min) with the disadvantages of being labor intensive and limited by daylight (Rook, 2000). Traditionally, automatic grazing time measurements are conducted using vibracorders (Forbes, 1988). Recently, an automatic method for grazing behavior estimation was developed (Rutter et al., 1997), which has several advantages such as fewer people required, less operator-associated errors, and more detailed behavior information (Champion et al., 1998).

Grazing Behavior on Pasture-Only Diets
Among the three grazing behavior variables, bite mass has the greatest influence on pasture DMI (Forbes, 1988; McGilloway and Mayne, 1996). Although bite mass is also affected by the animal’s anatomy characteristics (e.g., mouth; Rook, 2000), it is principally determined by pasture-related characteristics (Hodgson and Brookes, 1999), such as pasture height (Phillips, 1993; McGilloway et al., 1999) and density (Rook, 2000). Pasture height is the major constraint on bite mass in temperate pastures, with the effect primarily on bite depth rather than on bite area (Rook, 2000). Dairy cows consistently remove around one-third of the height of pasture, regardless of pasture height (Wade et al., 1989). Bite mass decreases with a reduction in pasture height both in unsupplemented (Gibb et al., 1997; McGilloway et al., 1999) and supplemented (Rook et al., 1994) dairy cows.

In many grazing behavior studies, pasture height is expressed as sward surface height, which refers to the height of the top surface of the leaf canopy on an undisturbed sward (Hodgson and Brookes, 1999). Gibb et al. (1997) reported that for dairy cows continuously grazing ryegrass, bite mass decreased from 0.31 g OM/bite at 7 or 9 cm to 0.23 g OM/bite at 5 cm, whereas neither biting rate (76 bites/min) nor grazing time (604 min/d) were affected by sward surface height. McGilloway et al. (1999) found that bite mass decreased from 1.28 to 0.85 g DM/bite in one experiment with reductions in sward surface height (from 21 to 7 cm) and from 1 to 0.66 g DM/bite in a second experiment with reductions in sward surface height (from 11 to 6 cm), while biting rate was not affected (56 bites/min in experiment 1; 62 bites/min in experiment 2). In a third experiment, an interaction was found between sward surface height and density; bite mass was reduced with reductions in sward surface height more at low pasture density (from 1.02 to 0.47 g DM/bite) than at high pasture density (from 0.97 to 0.63 g DM/bite; McGilloway et al., 1999).

Grazing time and biting rate are influenced by animal-related characteristics such as genetic merit and milk production. Both grazing time and biting rate act as compensatory mechanisms to avoid reductions in pasture DMI when bite mass decreases. However, these compensatory mechanisms have a limit. The upper limit of grazing time to compensate for a reduction in bite mass is determined for the time required for other activities such as ruminating (Rook, 2000). Under poor pasture conditions (e.g., very short pasture), all three variables decline (Hodgson and Brookes, 1999). High genetic cows had higher grazing time and biting rate than low genetic cows supplemented with concentrate (Bao et al., 1992). High genetic cows grazed a ryegrass pasture for longer time (218 vs. 204 min, measured visually for a period of 7 h) and at a higher biting rate (64 vs. 61 bites/min) than low genetic cows. Two recent studies (Pulido and Leaver, 2001; Bargo et al., 2002b) reported that high producing cows had greater grazing time, number of bites per day, and rate of intake than low producing cows. Bargo et al. (2002b) found a positive relationship between MY (kg/d) and the number of bites per day (TB, bites/d) for cows producing more than 25 kg/d of milk grazing a orchardgrass pasture and supplemented with 8.7 kg/d of concentrate: MY = 14.1 + 0.0005 TB (R² = 0.74), which indicates an increase of 5 kg/d of milk for every 10,000 bites/d (Bargo et al., 2002b).

Effect of Supplementation on Grazing Behavior
Studies evaluating the effect of supplementation on grazing behavior of dairy cows are presented in Table 1. Increasing the amount of concentrate reduced grazing time but did not affect biting rate (Arriaga-Jordan and Holmes, 1986; Kibon and Holmes, 1987; Rook et al., 1994; Bargo et al., 2002a; Gibb et al. 2002). Arriaga-Jordan and Holmes (1986) reported that grazing time was reduced 11 min/kg of concentrate in continuous grazing and 8 min/kg of concentrate in rotational grazing, while biting rate was not affected by the amount of supplementation. Rook et al. (1994) reported that concentrate supplementation, but not pasture height, reduced grazing time 20 min/kg of concentrate. Bite mass decreased as pasture height decreased, while the supplementation amount had no effect on bite mass (Rook et al., 1994). Amount but not type of energy supplement (cereal vs. beet pulp) reduced grazing time 8 to 12 min/kg of concentrate by dairy cows grazing ryegrass at two pasture heights (Kibon and Holmes, 1987). Biting rate was not affected by supplementation amount, type of supplement or pasture height, while bite mass was lower at the low pasture height (Kibon and Holmes, 1987). Recently, Sayers (1999) found that total bites per day and grazing time were higher when cows were supplemented with a fiber-based concentrate than when cows were supplemented with a starch-based concentrate, whereas bite mass was not affected. When the amount of concentrate was increased from 5 to 10 kg/d, total bites/d decreased from 22,023 to 16,933 and grazing time decreased 16 and 20 min/kg of fiber-based or starch-based concentrate, respectively (Sayers, 1999).

Data from Table 1 show that concentrate supplementation (mean: 4.1 kg/d; range: 2 to 8 kg/d) did not affect biting rate (mean: 58 bites/min; range: 45 to 78 bites/min) or bite mass (mean: 0.46 g of DM/bite; range: 0.27 to 0.64 g of DM/bite) but reduced grazing time by 34 min/d (SE 9 min/d, range: -212 to 25 min/d compared with controls; Student’s t-test, significantly different from zero, P < 0.01) and total bites per day by 2291 (SE 534, range: -6672 to 2833 compared with controls; Student’s t-test, significantly different from zero, P < 0.01). Regression analysis, accounting for the random effect of each study (St-Pierre, 2001), resulted in a negative relationship between grazing time (GT, min/d) and CDMI (kg/d): GT = 578 (SE 23) - 12 (SE 2) CDMI (R² = 0.88). Average grazing time for unsupplemented cows is 578 min/d and grazing time is reduced by 12 min/d for every kilogram of concentrate.

	SUBSTITUTION RATE AND MILK RESPONSE TO SUPPLEMENTATION

TOP ABSTRACT INTRODUCTION DRY MATTER INTAKE OF... GRAZING BEHAVIOR SUBSTITUTION RATE AND MILK... EFFECT OF SUPPLEMENTATION ON... RUMINAL DIGESTION AND... PASTURE IN SITU DIGESTION POSTRUMINAL DIGESTION OF GRAZING... CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
When grazing cows are fed supplements, pasture DMI usually decreases, which is known as substitution rate (SR; Kellaway and Porta, 1993). Substitution rate is calculated as: SR (kg/kg) = (pasture DMI in unsupplemented treatment - pasture DMI in supplemented treatment)/supplement DMI. A SR < 1 kg/kg means that total DMI on the supplemented treatment is higher than total DMI on the unsupplemented treatment. A SR = 1 kg/kg means that total DMI on the supplemented treatment is the same than total DMI on the unsupplemented treatment.

Substitution rate is one of the main factors explaining the variation observed in milk response (MR) to supplementation (Kellaway and Porta, 1993; Stockdale, 2000a). Milk response to supplementation is expressed as kg milk/kg supplement, but it can be defined as: 1) overall MR or the increase in kilograms of milk per kilogram of supplement DMI calculated relative to an unsupplemented treatment; and 2) marginal MR or the increase in kilograms of milk per kilogram of incremental increase in supplement DMI calculated for different amounts of supplement. There is usually a negative relationship between SR and MR. When SR is large, resulting in a small increase in total DMI, the MR is low. Milk response in the short-term determines whether supplementation is profitable based on milk and concentrate prices. However, additional long-term factors should also be considered in any economic evaluation, including increase in stocking rates on the farm, improvement in pasture utilization, positive effects on BCS and reproduction, increase in lactation length, and positive effects on milk composition (Kellaway and Porta, 1993).

Because SR and MR are closely related, factors affecting these two variables are discussed together. Substitution rate and MR to supplementation are affected by several pasture, animal, and supplement factors (Stockdale, 2000a; 2000b). The most important pasture-related factors are PA, pasture height, pasture species, pasture mass, and pasture quality. The most important supplement-related factors are amount and type of supplementation, and the most important animal-related factors are genetic merit of cows, production level, and stage of lactation.

Pasture Allowance
Many studies have reported that SR increases as PA increases (Meijs and Hoekstra, 1984; Stockdale and Trigg, 1985; Stakelum, 1986a, 1986b; Grainger and Mathews, 1989; Robaina et al., 1998; Bargo et al., 2002a). Many of these studies were conducted with low producing cows supplemented with less than 5 kg DM/d of concentrate; only the study of Bargo et al. (2002a) reported high producing cows fed more than 7 kg DM/d of concentrate. When stratifying the treatments in those studies as either low PA (<25 kg DM/cow per day; range: 7.6 to 25 kg DM/cow per day) or high PA (>25 kg DM/cow per day; range: 25 to 42.3 kg DM/cow per day), SR averaged 0.20 kg pasture/kg concentrate (range: 0 to 0.31 kg pasture/kg concentrate) at low PA, and 0.62 kg pasture/kg concentrate (range: 0.55 to 0.69 kg pasture/kg concentrate) at high PA. Considering the study effect as random (St-Pierre, 2001), a significant regression was found between SR (kg pasture/kg concentrate) and PA (kg DM/cow per day): SR = -0.55 (SE 0.13) + 0.05 (SE 0.009) PA -0.0006 (SE 0.0002) PA² (R² = 0.94).

Grazing studies evaluating the effect of PA on SR and MR of high producing dairy cows reported that SR increased and MR decreased as PA increased (Table 2). All those studies showed a negative relationship between MR and SR (Figure 1). Considering the variation due to each study (St-Pierre, 2001), the data from those experiments showed a negative relationship between MR (kg milk/kg concentrate) and SR (kg pasture/kg concentrate): MR = 1.71 (SE 0.29) - 2.01 (SE 0.66) SR (R² = 0.43), indicating that the lower the SR the higher the MR expected. This is in agreement with Stockdale (2000b), who summarized data from 20 grazing experiments and reported that MR was negatively related with SR. Higher SR observed when cows grazed at high PA may be partially explained by the higher quality of pasture actually consumed (Dixon and Stockdale, 1999). Because cows grazing at high PA have the opportunity to be more selective, pasture actually eaten has higher digestibility than at low PA (Mayne and Wright, 1988).

View larger version (14K): [in this window] [in a new window]   Figure 1. Relationship between milk response (MR) and substitution rate (SR) by grazing dairy cows supplemented with concentrate on studies evaluating the effect of pasture allowance (• Bargo et al., 2002a; Robaina et al., 1998; Stockdale, 1999a).

Accounting for the random effect of study (St-Pierre, 2001), analysis of the data from the four experiments resulted in no relationship between MR and SR, as indicated by the following nonsignificant linear equation: MR = 0.95 (SE 0.25) - 0.28 (SE 0.51) SR (R² = 0.02). Based on those studies, the type of relationship between MR and SR when high producing dairy cows on pasture are supplemented with increasing amounts of concentrate is not clear. The lack of a consistent relationship could be attributed to the fact that only few grazing studies have focused on MR with different amounts of concentrate (Peyraud and Delaby, 2001).

Level of Supplementation: Marginal and Overall Milk Response
The marginal MR to increasing amounts of concentrate has been described as curvilinear; i.e., the marginal increase in milk per kilogram of concentrate decreases as the amount of concentrate increases (Kellaway and Porta, 1993). Marginal MR decreased above 3 to 4 kg DM/d of concentrate in some studies, but this is not consistent and occurred primarily when pasture quality and quantity were not limiting and with cows of moderate genetic merit (Peyraud and Delaby, 2001). The response in milk production of high producing dairy cows grazing pasture and supplemented with different amounts of concentrate is shown in Figure 2. The studies were grouped into two categories: 1) those with cows producing more than 28 kg/d at the beginning of the experiment regardless of stage of lactation or with less than 90 DIM, and 2) those with cows producing less than 23 kg of milk/d and more than 160 DIM.

View larger version (14K): [in this window] [in a new window]   Figure 2. Relationship between milk production (MP) and concentrate DMI (CDMI) by grazing dairy cows supplemented with different amounts of concentrate ( studies with cows < 90 DIM or > 28 kg milk/d at the beginning of the experiment; Bargo et al., 2002a; Delaby et al., 2001; Gibb et al., 2002; Reis and Combs, 2000b; Sayers, 1999; studies with cows > 160 DIM or < 23 kg milk/d at the beginning of the experiment; Robaina et al., 1998; Sayers, 1999; Walker et al., 2001). Close and open squares correspond to actual values, regression lines correspond to parameters estimated after adjusting for study effect (St-Pierre, 2001).

The intercepts (12.9 vs. 22.4 kg/d) in Figure 2 show the differences in stage of lactation and may also indicate differences in genetic merit between the two groups; however, these two factors are confounded. From this figure it can be concluded that milk production of high producing dairy cows in early lactation increases linearly as the amount of concentrate increases from 1.8 to 10 kg DM/d with an overall MR of 1 kg milk/kg concentrate. Milk production of high producing dairy cows in late lactation, however, increases as the amount of concentrate increase cows but with a lower marginal MR per kilogram of concentrate. To avoid metabolic health problems such as acidosis or subclinical acidosis, it is not recommended to supplement more than about 10 kg DM/d (or >50% of the total diet DMI). At that limit, decreased marginal MR traditionally observed when supplementation is increased did not occur with high producing cows. Another factor that needs to be considered is the pasture quality. The NDF was >50% in several studies, suggesting that high fiber intake may allow for feeding high amounts of concentrate.

Previous reviews (Journet and Demarquilly, 1979) reported average MR from 0.4 to 0.6 kg milk/kg concentrate. However, recently Peyraud and Delaby (2001) reported that the MR to concentrate was higher than previously reported research in the literature published after 1990, which can be attributed to the increase in genetic merit of cows. A greater response to supplementation may be expected in high genetic merit cows because they partition more nutrients to milk production and lose more BW in early lactation than low genetic merit cows (Kellaway and Porta, 1993). Stage of lactation also influences lactational responses to concentrate supplements (Dixon and Stockdale, 1999). In early lactation, cows partition more nutrients toward milk production, thus MR to supplementation may be higher than in late lactation, when more nutrients are directed to BW (Kellaway and Porta, 1993). The average milk yield response to concentrate supplementation of grazing cows supplemented with 3 kg DM/d concentrate was 0.7, 0.4, 0.5, and 0 kg milk/kg concentrate when they were between 86 to 114, 115 to 133, 134 to 187, and 188 to 243 DIM, respectively (O’Brien et al., 1999). Summarizing five experiments with supplement DMI from 0 to 7 kg/d, the marginal MR was 1.3, 1.1, and 0.7 kg milk/kg supplement in early, mid, and late lactation (Stockdale et al., 1987).

Type of Supplement
The type of supplement influences SR and animal performance (Stockdale, 2000b). Forage supplementation decreases pasture DMI more than concentrates (Mayne and Wright, 1988). Including both low and high PA, SR ranged from 0.84 to 1.02 kg/kg for grass silage supplementation and from 0.11 to 0.50 kg/kg for concentrate supplementation (Mayne and Wright, 1988). Recently, Stockdale (2000b) reviewed 39 datasets from grazing studies and concluded that supplementation with forages, such as hay or corn silage, resulted in higher SR than supplementation with concentrates. Stockdale (1999b) found, however, that SR were similar when dairy cows grazing a ryegrass/white clover pasture were supplemented with grain or hay, which shows that the effects of supplement type on SR may not occur if both PA and amount of supplementation are low.

Meijs (1986) reported that SR was reduced from 0.45 kg pasture/kg high-starch concentrate to 0.21 kg pasture/kg fiber-based concentrate when cows grazed a ryegrass pasture. Studies with high producing dairy cows grazing or fed fresh-cut forage in confinement evaluated the effect of type of concentrate (starch vs. fiber-based) on SR and MR (Table 2). Two studies (Schwarz et al., 1995; Sayers, 1999) had a negative relationship between MR and SR, while the third study (Spörndly, 1991) had a positive of relationship. In the studies of Schwarz et al. (1995) and Spörndly (1991), however, the effect of type of concentrate is confounded with the effect of the amount fed because two amounts of each type of concentrate were fed. They were also conducted in confinement, a condition in which SR may differ from grazing studies. Combining the data from these three experiments and accounting for the random effect of each study (St-Pierre, 2001), no general relationship was found between MR and SR, as indicated by the following nonsignificant linear equation: MR = 0.75 (SE 0.36) - 0.43 (SE 0.64) SR (R² = 0.53). Based on those three studies, there is not enough information in the literature to conclude the type of relationship between MR and SR when high producing dairy cows on pasture are supplemented with starch or fiber-based concentrates.

Causes of Substitution Rate
It has been hypothesized that SR is caused by negative associative effects in the rumen (Dixon and Stockdale, 1999), or reduction in grazing time (McGilloway and Mayne, 1996). When concentrate supplements are included in pasture diets, associative effects may occur if digestive and metabolic interactions between them change the intake of energy (Dixon and Stockdale, 1999). An increase in total digestibility may be expected with the inclusion of concentrates in the diet because they are usually higher in digestibility than pasture. However, interactions between the digestion of concentrates and pasture may reduce fiber digestion (Dixon and Stockdale, 1999). The energy provided by the concentrate (fermentable carbohydrates) may lead to reductions in ruminal pH, which may decrease the activity or number of cellulolytic bacteria, reduce the rate of fiber digestion of pasture, and therefore pasture DMI (Dixon and Stockdale, 1999). Based on this hypothesis, small amounts of concentrate supplementation or supplementation with concentrates with a slow rate of degradation would result in lower SR (Kellaway and Porta, 1993). However, studies on the effect of amount of concentrate supplementation on SR (Table 2) did not show a clear trend to lower SR with small amounts of concentrate, probably because of the small number of studies. Under the same hypothesis, concentrates that are more slowly degraded in the rumen (e.g., fiber-based concentrates) would minimize SR compared with concentrates that are more rapidly degraded in the rumen (e.g., starch-based concentrates) because ruminal pH would be higher with fiber-based concentrates. However, the effect of the type of concentrate (Table 2) showed inconsistent results, which may be related to the different sources and proportion of starch and fiber that determines the rate of ruminal degradation. More information is needed related to the type and amount of concentrate supplementation and their interaction on SR of high producing dairy cows on pasture.

The second hypothesis proposed to explain SR is related to grazing time. It has been suggested that reductions in grazing time by supplementation would explain SR (Mayne and Wright, 1988; McGilloway and Mayne, 1996). A significant negative relationship was presented and discussed above between grazing time and concentrate DMI, which indicates a reduction of 12 min/d per kilogram of concentrate.

Bargo et al. (2002a) studied ruminal digestion and grazing time of high producing dairy cows grazing at low and high PA to test both hypotheses on SR. Substitution rate was higher (0.55 vs. 0.26 kg pasture/kg concentrate) when supplemented cows grazed at higher PA (40 vs. 25 kg of DM/cow per day), and it was related to both negative associative effects in the rumen and reduction in grazing time. Supplementation with 7.9 kg/d of a corn-based concentrate reduced ruminal pH, ruminal degradation rate of pasture, and fiber digestibility at both PA (Bargo et al., 2002a). Grazing time was reduced 75 min/d, with supplementation at the low PA, which explained nearly all of the 2.0 kg/d reduction in pasture DMI measured by Cr₂O₃ (75 min/d x 55 bites/min x 0.55 g of DM/bite = 2.3 kg/d). At the high PA, concentrate supplementation reduced grazing time 104 min/d and explained 80% of the 4.4 kg/d reduction in pasture DMI (104 min/d x 56 bites/min x 0.60 g of DM/bite = 3.5 kg). The remaining 20% may be related to negative associative effects in the rumen; for example, the decrease in apparent digestibility of NDF by concentrate supplementation was greater at the high PA than at the low PA (4.3 vs. 1.1 percentage points, respectively; Bargo et al., 2002a).

	EFFECT OF SUPPLEMENTATION ON DRY MATTER INTAKE, MILK PRODUCTION, AND MILK COMPOSITION

TOP ABSTRACT INTRODUCTION DRY MATTER INTAKE OF... GRAZING BEHAVIOR SUBSTITUTION RATE AND MILK... EFFECT OF SUPPLEMENTATION ON... RUMINAL DIGESTION AND... PASTURE IN SITU DIGESTION POSTRUMINAL DIGESTION OF GRAZING... CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Energy Supplementation
Level of supplementation.
Studies conducted with high producing dairy cows on pasture that have evaluated the effect of amount of concentrate supplementation on DMI, and milk production and composition are presented in Table 2. Overall, pasture DMI decreased and total DMI increased by increasing the amount of concentrate. Pasture DMI was numerically (Arriaga-Jordan and Holmes, 1986; Dillon et al., 1997) or significantly (Spörndly, 1991; Robaina et al., 1998; Sayers, 1999; Reis and Combs, 2000b; Walker et al., 2001; Bargo et al., 2002a) reduced when amount of concentrate increased, which is related to the SR. For the range of concentrate supplementation (1.8 to 10.4 kg DM/cow per day), pasture DMI decreased 1.9 kg/d (SE 0.3 kg/d, range: -0.1 to -4.4 kg/d; Student’s t-test, significantly different from zero, P < 0.01) or 13% compared with pasture DMI of pasture-only diet treatments (14.8 kg/d). When corrected for the random effect of study (St-Pierre, 2001), a significant negative relationship was found between pasture DMI reduction (PDMIr, kg/d) and CDMI (kg/d): PDMIr = 0.26 (SE 0.54) - 0.39 (SE 0.07) CDMI (R² = 0.82). Total DMI was numerically (Arriaga-Jordan and Holmes, 1986) or significantly (Spörndly, 1991; Dillon et al., 1994; Robaina et al., 1998; Sayers, 1999; Reis and Combs, 2000b; Walker et al., 2001; Bargo et al., 2002a) increased 3.6 kg/d (SE 0.5 kg/d, range: 1.0 to 7.5 kg/d; Student’s t-test, significantly different from zero, P < 0.01) or 24% compared with total DMI of pasture-only diet treatments. When corrected for the random effect of study (St-Pierre, 2001), a significant positive relationship was found between total DMI increase (TDMIi, kg/d) and CDMI (kg/d): TDMIi = 0.08 (SE 0.73) + 0.52 (SE 0.08) CDMI (R² = 0.91).

The studies presented in Table 3 reported that milk production increase averaged 4.4 kg/d (SE 0.6 kg/d, range: 0.8 to 10.6 kg/d; Student’s t-test, significantly different from zero, P < 0.01) with the amount of supplementation or 22% compared with the pasture-only diet treatments (19.7 kg/d). When corrected for the random effect of study (St-Pierre, 2001), a significant positive relationship was found between milk production increase (MPi, kg/d) and CDMI (kg/d) and TDMIi (kg/d): MPi = 0.09 (SE 0.65) + 0.26 (SE 0.12) CDMI + 0.89 (SE 0.19) TDMIi (R² = 0.98).

Several authors reported that increasing the amount of concentrate supplementation increased milk protein percentage (Hoden et al., 1991; Spörndly, 1991; Wilkins et al., 1994; Sayers, 1999; Reis and Combs, 2000b; Valentine et al., 2000; Bargo et al., 2002a). Linear increases in milk protein percentage have been reported for a wide range of amounts of supplementation, including 0 to 4 kg DM/d (Hoden et al., 1991; Wilkins et al., 1994), 0 to 10 kg DM/d (Reis and Combs, 2000b), and 7 to 13 kg DM/d (Valentine et al., 2000). Other studies, however, found no changes in milk protein percentage within a range of supplementation from 0 to 3.6 kg DM/d (Dillon et al., 1997) and 0 to 10.4 kg DM/d (Walker et al., 2001). Overall, the increase in protein percentage averaged 0.13 percentage units (SE 0.01 percentage units, range: 0.01 to 0.25 percentage units; Student’s t-test, significantly different from zero, P < 0.01) or 4% compared to the pasture-only diet treatments (3.06%). When corrected for the random effect of study (St-Pierre, 2001), a significant positive relationship was found between protein percentage increase (PPi) and CDMI (kg/d): PPi = 0.05 (SE 0.03) + 0.01 (SE 0.003) CDMI (R² = 0.72). Protein yield was increased numerically (Spörndly, 1991; Robaina et al., 1998; Walker et al., 2001) or significantly (Hoden et al., 1991; Wilkins et al., 1994; Dillon et al., 1997; Reis and Combs, 2000b; Valentine et al., 2000; Bargo et al., 2002a) with supplementation: 0.17 kg/d (SE 0.02 kg/d, range: 0.04 to 0.36 kg/d; Student’s t-test, significantly different from zero, P < 0.01) or 30% compared with the pasture-only diet treatments (0.56 kg/d). When corrected for the random effect of study (St-Pierre, 2001), a significant positive relationship was found between protein yield increase (PYi, kg/d) and CDMI (kg/d) and TDMIi (kg/d): PYi = 0.01 (SE 0.02) + 0.01 (SE 0.004) CDMI + 0.03 (SE 0.006) TDMIi (R² = 0.96).

Starch vs. fiber-based concentrates.
Studies comparing starch or fiber-based concentrates for high producing dairy cows on pasture are presented in Table 4. Some of those were grazing studies (Meijs, 1986; Sayers, 1999; Delahoy et al., 2003), and others were confinement studies with cows fed fresh-cut forage (Garnsworthy, 1990; Valk et al., 1990; Spörndly, 1991; Schwarz et al., 1995). Sources of starch included corn (Schwarz et al., 1995; Valk et al., 2000; Delahoy et al., 2003), barley (Spörndly, 1991), cassava (Meijs, 1986), or the combination of barley, wheat, and corn (Garnsworthy, 1990; Sayers, 1999). Sources of fiber included oatfeed (Garnsworthy, 1990), and beet pulp either alone (Spörndly, 1991; Schwarz et al., 1995; Valk et al., 2000) or combined with soy hulls (Meijs, 1986; Delahoy et al., 2003) or citrus pulp (Sayers, 1999). Because starch sources are more commonly used to supplement dairy cows on pasture than fiber sources, results of these studies are summarized as the effect of fiber-based concentrate compared to starch-based concentrates.

Milk production was increased when fiber-based concentrates replaced starch-based concentrates in only one grazing study (Meijs, 1986), while two grazing studies reported similar milk production (Sayers, 1999; Delahoy et al., 2003). In the confinement studies, milk production was not affected (Garnsworthy, 1990; Spörndly, 1991; Schwarz et al., 1995) or was reduced (Valk et al., 1990) by fiber-based concentrates compared to starch-based concentrates. The higher milk production with starch-based concentrates in the study of Valk et al. (1990) could be attributed to the higher total DM and energy intake in that treatment. Overall, milk production was slightly reduced (-0.46 kg/d) when fiber-based concentrates replaced starch-based concentrates, but the range of variation is large (-2.6 to 1.3 kg/d).

Most of the studies (Meijs, 1986; Garnsworthy, 1990; Valk et al., 1990; Schwarz et al., 1995; Delahoy et al., 2003) did not report changes in milk fat percentage. However, Sayers (1999) reported higher fat percentage with fiber-based than with starch-based concentrates, particularly when large amounts of concentrate (10 kg DM/d) were supplemented to cows grazing a ryegrass pasture. In contrast, Spörndly (1991) found that in confinement, fiber-based concentrates reduced milk fat content. However, in that study, cows consumed 2 kg DM/d of hay, which makes comparison between the studies difficult. Replacing starch-based by fiber-based concentrates reduced significantly (Spörndly, 1991; Sayers, 1999; Delahoy et al., 2003) and numerically (Meijs, 1986; Valk et al., 1990; Schwarz et al., 1995) milk protein percentage. Overall, milk protein percentage was reduced -0.06 percentage units (range: -0.21 to 0.05 percentage units) with fiber-based concentrates compared with starch-based concentrates.

The number of studies in which fiber-based concentrates replaced starch-based concentrates is too small to make strong conclusions, and half of the studies were conducted in confinement. Inconsistency in the results can also be attributed to differences in the source of starch or fiber used in the concentrate, type of pasture, and other components in the diet, all factors that may affect the rate of degradation of concentrates in the rumen. Meijs (1986) suggested that supplementing a highly degradable pasture with a starch-based concentrate might reduce ruminal pH and pasture ruminal digestion, increase retention time of feed in the rumen, and decrease pasture DMI. Replacing starch-based concentrates with fiber-based concentrates would maintain higher pH in the rumen, enhance pasture digestion, and result in higher DMI. In the studies of Meijs (1986) and Sayers (1999) both pasture (ryegrass) and starch (cassava or barley plus wheat) were highly degradable in the rumen and could explain the response in DMI with fiber-based concentrate. In the study of Spörndly (1991), the inclusion of hay in the diet may have maintained a higher pH in the rumen and therefore explain similar DMI even with highly degradable starch source such as barley. The use of a starch source with lower degradability than barley such as corn (Schwarz et al., 1995; Delahoy et al., 2003) may not be as detrimental to ruminal pH and may explain similar pasture DMI observed in those studies with cows grazing a slowly degradable pasture of orchardgrass (Delahoy et al., 2003).

Processed grain.
Processing methods for grains used for dairy cows have been extensively reviewed (Theurer et al., 1999); however, that review focused on cows fed TMR diets in confinement. Studies evaluating the effect of processed grains such as corn or sorghum on DMI, and milk production and composition of dairy cows on pasture are presented in Table 5. Seven of those eight studies were grazing studies (Bargo et al., 1998; Pieroni et al., 1999; Reis and Combs, 2000a; Soriano et al., 2000; Alvarez et al., 2001; Wu et al., 2001; Delahoy et al., 2003), and one was conducted in confinement with fresh-cut forage (Reis et al., 2001). Results are summarized as the effect of forms of processing compared with unprocessed (dry) forms. Forms of processing included high moisture corn (Soriano et al., 2000; Alvarez et al., 2001; Reis et al., 2001; Wu et al., 2001), steam-flaked corn with a density of 290 (Bargo et al., 1998) or 360 g/L (Delahoy et al., 2003), steam-rolled corn with a density of 591 g/L (Reis and Combs, 2000a), and steam-flaked sorghum with a density of 480 g/L (Pieroni et al., 1999).

Compared with unprocessed grains, supplementation with processed grains did not change milk fat percentage in seven of eight studies (Bargo et al., 1998; Pieroni et al., 1999; Reis and Combs, 2000a; Soriano et al., 2000; Alvarez et al., 2001; Delahoy et al., 2003; Reis et al., 2001); only Wu et al. (2001) reported a reduction in milk fat content. Only two of the eight studies (Alvarez et al., 2001; Wu et al., 2001) found higher milk protein percentage with high moisture corn than with dry corn. Increase in milk protein content averaged 3% (0.09 to 0.11 percentage units), which may suggest an increase in ruminal available energy with processed corn. Overall, the replacement of unprocessed by processed grains resulted in small changes in milk fat (mean: -0.06 percentage units, range: -0.39 to 0.16 percentage units) and protein (mean: 0.04 percentage units, range: -0.03 to 0.11 percentage units) content in milk.

Although the number of studies is not large enough to make strong conclusions, the lack of response to processed grains can be related to changes only in site of digestion (i.e., more energy available in the rumen with processed grains vs. more energy available postruminally with unprocessed grains) without affecting the total energy intake. Another factor is that almost all the studies were conducted with dairy cows after the peak of lactation or with relatively low producing cows resulting in cows in positive energy balance.

Rumen Undegradable Protein Supplementation
The use of RUP sources for dairy cows has been extensively reviewed by Santos et al. (1998); however, that review focused on TMR fed in confinement. Supplementation with RUP might be necessary for high producing dairy cows on pasture because the basal diet of pasture has a high ruminal CP degradability (>70%), and therefore provides smaller amounts of RUP compared with cows on TMR diets. The effect of supplementation with isonitrogenous concentrates based on various sources of RDP or RUP on DMI and milk production and composition of high producing dairy cows on pasture is shown in Table 6. All the studies were conducted with cows in early lactation (<75 DIM) supplemented with isonitrogenous concentrates that ranged from 14 to 24% CP, where RDP sources such as soybean meal (Hongerholt and Muller, 1998; McCormick et al., 1999, 2001a, 2001b; Schor and Gagliostro, 2001), sunflower meal (Schroeder and Gagliostro, 2000; Bargo et al., 2001), and urea or rapessed meal (Tesfa et al., 1995) were replaced by RUP sources such as animal protein blend (Hongerholt and Muller, 1998), corn gluten meal (McCormick et al., 1999; 2001a), expeller soybean meal (McCormick et al., 2001b), blood meal (McCormick et al., 2001a; Schor and Gagliostro, 2001), feather meal (Bargo et al., 2001), heat-treated rapeseed meal (Tesfa et al., 1995), and fish meal (Schroeder and Gagliostro, 2000).