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

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal 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 Chapman, D. F. Articles by Thompson, A. N. Search for Related Content PubMed Articles by Chapman, D. F. Articles by Thompson, A. N. Agricola Articles by Chapman, D. F. Articles by Thompson, A. N. Related Collections Forage Management Grazing Management

SYMPOSIA

Impacts of Spatial Patterns in Pasture on Animal Grazing Behavior, Intake, and Performance

^a School of Agriculture and Food Systems, Univ. of Melbourne, Victoria 3010, Australia
^b AgResearch Grasslands, Private Bag 11008, Palmerston North, New Zealand
^c Dep. of Horticulture and Crop Science, Ohio State Univ., 2021 Coffey Rd., Columbus, OH 43210, USA
^d Meat and Livestock Australia, 165 Walker Street, North Sydney, NSW 2060, Australia
^e Primary Industries Research Victoria, Dep. of Primary Industries, Private Bag 105, Hamilton, Victoria 3300, Australia
^f Institute of Grassland and Environmental Research, North Wyke, Okehampton, Devon, EX20 2SB, UK

^* Corresponding author (d.chapman{at}unimelb.edu.au)

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION BIOPHYSICAL AND BEHAVIORAL BASES HETEROGENEITY, GRAZING... ANIMAL PRODUCTION RESPONSES TO... CONCLUSIONS REFERENCES

 
Control over the quantity and quality of food ingested by grazing ruminants in temperate pasture systems remains elusive. This is due in part to the foraging choices that animals make when grazing from communities of mixed plant species. Grazing behavior and intake interact strongly with the feed supply–demand balance, pasture composition, and grazing method. These interactions are not completely understood, even for relatively simple pasture communities such as a perennial ryegrass (Lolium perenne L.)–white clover (Trifolium repens L.) mixture. When offered a free choice between these species, ruminants exhibit a partial preference for clover compared to grass (about 0.7:0.3) and have a higher intake rate from clover but do not graze to maximize their daily intake of dry matter (DM). When monocultures of grass and clover are offered as a free choice in 50:50 area ratio, animal performance is no different than from a clover monoculture alone. Thus, all of the feeding value benefits of clover are available when only 0.5 of the grazing area is sown to clover. These observations accord with the satiety theory and imply that there are constraints to eating pure clover that animals can overcome by adding grass to their diet, provided their ability to locate and ingest each food is not seriously limited. The challenge for grassland management is to present feed to animals at pasture in ways that allow them to meet their dietary preferences, while also allowing high rates of animal production per hectare.

Abbreviations: DM, dry matter • DMI, dry matter intake • WSC, water soluble carbohydrate

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION BIOPHYSICAL AND BEHAVIORAL BASES HETEROGENEITY, GRAZING... ANIMAL PRODUCTION RESPONSES TO... CONCLUSIONS REFERENCES

 
GRAZED VEGETATION COMMUNITIES are heterogeneous in space and time. Many of the variables that influence the interaction between soils, plants, and animals are subject to fine-scale spatial pattern, plus systematic (for example, intra-annual seasonal cycles) and stochastic (for example, interannual variability) temporal influences. If we do not account for this inherent feature of grazed grasslands, then our ability to predict the system level outcomes (primary and secondary productivity, environmental resource condition) of changes to management policies and/or environmental conditions must surely be limited.

Importantly, our account of heterogeneity must be leveled at the appropriate spatial and temporal scale. This poses the challenge of determining the minimum spatial dimension that must be described and manipulated to achieve certain goals. Recent advances in modeling the dynamics of spatial heterogeneity in grazing systems (e.g., Parsons and Dumont, 2003; Marion et al., 2004) have helped define the core principles, but these have not yet been translated into knowledge or tools that bring the necessary predictive power to grassland management decisions or policies. Valuable new information from large-scale, mechanistic studies of the relationships between the spatial distribution of plant species and the grazing behavior and performance of domestic livestock is now available. This approach offers some promising leads as to how grazing systems could be redesigned for improved animal production outcomes.

In this paper, we review information on the grazing behavior and intake of domestic ruminant livestock obtained from research into the effects of controlled spatial variability in the composition of pastures. We concentrate principally on temperate grassland species and the application of the information gained to intensive pasture-based animal production systems. The focus is mainly on field scales of spatial variability, as opposed to patch or bite scale, though we do consider the consequences of bite scale, and hence patch, phenomena for primary and secondary productivity of grazed pastures. More generic descriptions of animal foraging strategies and ingestive behavior can be found in the comprehensive reviews of Laca and Demment (1996) and Ungar (1996).

Our aim in the first part of the paper is to draw insights into the factors that drive the selection of certain feed components from a mixed pasture community. There are good practical reasons for pursuing this knowledge. Spatial variability and selective grazing introduce inefficiencies to intensive pasture-based production systems. These include an increase in foraging costs and an increase in grazing duration required to meet daily nutrient requirements (Parsons et al., 1994a). Selective grazing for a preferred component within a pasture mixture can decrease the presence of that component in the feed, since the preferred species must compete for growth resources against other species that are not incurring the same defoliation costs. This is one potential explanation for the low clover content typically found in mixed grass–clover pastures in temperate areas of the world (e.g., Frame and Newbould, 1986; Caradus et al., 1996; Chapman et al., 1996). Selective grazing may therefore be counterproductive if it increases foraging costs and decreases the presence of preferred feed to the extent that the additional nutritive value of the preferred feed is unable to compensate for losses in nutrient ingestion rate. Better knowledge of the factors that are motivating animals to select for certain dietary components should allow us to consider how pastures could be presented to grazing livestock so that they can acquire their preferred diet. The latter part of this review considers this possibility, after first analyzing the basis for controlling the yield obtained from grazed pastures, and the role of spatial variability in helping or hindering the achievement of goals in this important area. We also consider the theoretical and practical issues that still require resolution in order for the use of controlled spatial variation in the way we present pastures to animals to become a reality.

	BIOPHYSICAL AND BEHAVIORAL BASES

TOP NOTES ABSTRACT INTRODUCTION BIOPHYSICAL AND BEHAVIORAL BASES HETEROGENEITY, GRAZING... ANIMAL PRODUCTION RESPONSES TO... CONCLUSIONS REFERENCES

 
Preference
Our insights begin with the phenomenon of preference, which we define (after Parsons et al., 1994a) as what animals select in their diet when the constraints to locating and ingesting their food are minimal. The phenomenon of "selection" has been defined as "preference modified by environmental constraints" (Hodgson, 1979) and so is a function of an animal's preferred diet as influenced by environmental and management factors. Examples of selection abound in intensive grazing systems, for instance the observation that domestic ruminants select green leaf material in much higher proportions in their diet compared to the total herbage on offer (e.g., L'Hullier et al., 1984; Grant et al., 1985). Equally clear preference for leaf versus stem material in pastures is widely observed and incorporated as a key principle behind certain grassland management practices (e.g., Hodgson, 1990).

These behaviors can be readily understood and explained in terms of the differences between preferred and nonpreferred feed components in their nutritive and feeding value. Thus, grazing management practices in intensively managed pasture systems strive to maintain high green/dead and leaf/stem ratios in the available feed, in the sure knowledge that animal intake and growth responses to pasture presented this way will be greater than to pasture comprising large proportions of dead material and/or stem (Hodgson et al., 1994). Achieving these management goals is not a trivial exercise, since the rates of gross herbage production and senescence vary greatly in space and time, and the transition from vegetative to reproductive growth in grasses differs between species and is strongly related to seasonal conditions, which are also highly variable.

Further insights into the control of grazing behavior in ruminants are revealed when we consider another common grazing behavior phenomenon in temperate pastures: the apparent selection for legume species such as clover growing in mixture with grass species such as perennial ryegrass (Curll and Wilkins, 1982). This too can be thought of as a feeding value response, since it is well documented that animal intake on clover is greater than on grass (e.g., Kenney and Black, 1984) and that the nutritive value of clover commonly exceeds that of grass, often by a wide margin (see for example Waghorn and Barry, 1987). But, well-fertilized grass pastures can readily supply feed of sufficient nutritive value to meet animal requirements in many situations. Therefore, selection for clover instead of grass is not explained as satisfactorily by simple nutritive value differences as, for example, selection for green leaf versus dead material.

Nevertheless, the notion that a high proportion of clover is desirable in mixed pastures is also incorporated into pasture management objectives for optimizing animal production in intensively managed, temperate grasslands (see for example Frame and Newbould, 1986), and we do not dispute the value of such practices. Rather, we have used this observation as a jumping-off point to consider in more depth what is controlling the apparent selection for clover, and thereby identify factors that might be controlling animal grazing behavior and intake in general.

Spatially separated grass–clover monocultures have been a valuable experimental model system for addressing this, because such swards remove the constraints to locating and ingesting preferred dietary components that are inevitable in mixed pastures. Allowing animals a free choice between adjacent monocultures of grass and clover allows a true test of their preference for the dietary components on offer. The results of several studies of the responses of sheep and cattle to controlled spatial variability in pasture have been instructive when considering the complex interactions between domestic ruminants and their feed sources.

Mixed Diets and Partial Preference
The first notable outcome of such studies is that sheep and cattle consume a mixed diet, with a partial preference for clover of about 70% (Rutter, 2006). This is observed consistently, despite the fact that the animals could freely choose a diet that comprised only clover in amounts sufficient to meet their daily energy and protein requirements. The observation of partial preference holds for different livestock classes within sheep (Parsons et al., 1994a; Harvey and Orr, 1996; Cosgrove et al., 1999) and cattle (Penning et al., 1995b; Orr et al., 1996; Rutter et al., 1999, 2004a; Marotti, 2004) in different environments (northern Europe, continental USA, Australia, New Zealand), for different species contrasts (Torres-Rodriquez et al., 1997; Cosgrove et al., 1997; Poli et al., 1997; Cortes et al., 2005; Rutter, 2006), and for different sward states, for example, different area ratios of grass and clover (Parsons et al., 1994a; Rutter et al., 2004a, 2004b) and different height of the clover versus grass component (Torres-Rodriquez et al., 1997; Harvey et al., 2000). The consistency of the result is such that we can rule out randomness and conclude that animals are actively selecting for a mixed diet containing a sizeable grass component when the conventional, but perhaps overly simplistic, expectation is that they would select much more strongly, if not solely, for clover. Several possible explanations for the result have been ruled out by follow-up studies. These include propositions that animals are responding to novelty (Harvey and Orr, 1996; Rutter et al., 1999; and Nuthall et al., 2000 have all shown partial reference continuing for periods of up to 12 wk), that animals have poor ability to visually discriminate between feeds (discounted by the results of Edwards et al., 1997), and that animals have poor ability to remember where preferred foods are located (discounted by spatial memory studies conducted by Edwards et al., 1996). Explanations that have received further attention, and which are still in contention, include the maintenance of a diverse rumen microflora and therefore broad function in terms of handling different feedstuffs, maintenance of some "optimal" C/N balance in the diet, avoidance of toxic consequences of ingestion of one dietary component to excess, and avoidance of grazing at night due to a perceived risk of predation. We return to some of these issues later, and note for the moment only that they are often closely linked, making it difficult to disentangle cause and effect.

Grazing Behavior and Intake
A second notable and consistent outcome of pasture manipulation studies where the full range of treatments is included (grass monoculture, clover monoculture, grass–clover mixture, and grass–clover choice) is that animal grazing behavior and intake on these treatments are quite different and not always explained by conventional knowledge of the relative nutritive value and feeding value of the dietary components on offer. Examples of some of the more significant findings in the context of this review include:

1. Animals grazing pure clover swards typically achieve higher intake rates (generally measured in terms of grams DM per minute) than animals grazing pure grass (Table 1). While the absolute intake rates of grass and clover differ between the studies included for the sheep and dairy cow categories in Table 1, the clover intake rates were consistently in the range 45 to 60% higher than the grass intake rates. The observation that sheep, but not heifers, were able to eat clover faster than grass was predicted by the mechanistic intake model described by Parsons et al. (1994b) because sheep require a greater number of mastication bites per unit feed prehended than cattle to be able to ingest their feed, allied with the fact that clover has a lower mastication cost per unit DM compared to grass. However, the observation by Rutter et al. (2004a) and Marotti (2004) that lactating cows can also eat clover more rapidly than grass means that other factors are involved in the total response. Despite these uncertainties, we can conclude that, because animals consistently include up to 30% grass in their diet when they could eat a 100% clover diet, they are not attempting to maximize intake rate—an outcome that would be achieved if they were to eat only clover and exclude grass altogether from their diet.
   
2. Animals show great elasticity in their grazing behavior and are able to achieve similar intakes on quite different pasture types by increasing grazing time, presumably to compensate for differences in intake rate. For example, K.J. Venning (unpublished data, 2003) observed that lactating ewes on a grass monoculture grazed for 60% longer each day compared to ewes on a clover monoculture (Table 2). Intake rates measured on the grass and clover swards used in the 2002 experiment were 5.8 and 3.5 g DM min^–1 for clover and ryegrass respectively, and estimated daily intakes were 2.9, 2.5, 2.3 and 2.4 kg DM animal^–1 d^–1 for the clover monoculture, grass monoculture, grass–clover mixture and choice treatments respectively (K.J. Venning, unpublished, 2003). Penning et al. (1995a) also observed differences in grazing time of this magnitude between lactating and nonlactating ewes grazing either pure grass or pure clover, although sheep grazing pure clover showed greater capacity to increase intake through a combination of increasing intake rate and grazing time compared to sheep grazing grass.
   
3. Animals grazing a free choice arrangement of adjacent monocultures of grass and clover typically achieve similar daily intakes to those grazing pure clover, even though choice animals are including up to 30% grass in their diet (Table 3). This observation at first seems at odds with the intake rate results described previously. The obvious expectation would be that, in the absence of compensating changes in the time spent grazing, animals eating from the choice treatment would suffer a fall in daily intake in proportion to the difference between grass and clover intake rate weighted for the time spent grazing on grass. Tables 2 and 3 indicate that there are small compensatory increases in grazing time when animals are offered the free choice compared to animals offered pure clover only, and this may be sufficient to ensure their daily intake is not adversely affected. Two data sets (Cosgrove et al., 2001; Champion et al., 2004) suggest that grazing time and intake of animals on the choice treatment may in fact exceed that of animals grazing on the pure clover diet, raising the possibility of a boost to intake from inclusion of grass in the diet, presumably because addition of the grass allowed animals to overcome some constraint to eating which limited intake of the animals on clover alone. However, subsequent experiments have not reproduced this result, and the possibility of boosting intake through increasing dietary diversity remains elusive in intensive pasture-based production systems.
   

In answer to the question "what are animals doing when they select a mixed diet?", we have already noted that several possible explanations such as sampling for novelty, poor visual discrimination, and poor spatial memory have been ruled out. This leads us to consider in more detail, the potential interactions between ingestion and digestion, particularly the nutritional consequences of consuming a diet dominated by a feed characterized by high DM and protein digestibility, such as clover.

Nutritional Physiology and Satiety
The digestion of pasture herbage in the rumen releases a complex array of metabolites and further products, which are then subsequently absorbed across the rumen wall, passed to the small intestine and absorbed there, or voided via excretion in urine and dung. Both beneficial and deleterious metabolites and compounds are produced by the digestion process. It is well established that plant species such as white clover and Persian (T. resupinatum L.) clover have high concentrations of rumen degradable protein (Beever et al., 1985, 1986; Williams et al., 2005). Rapid degradation of this protein in the rumen can lead to high concentrations of ammonia in the rumen fluid, linked to high blood ammonia levels (Parker et al., 1995) that pose a toxicity threat to the animal (Symonds et al., 1981).

One proposition that can be drawn from the observation of partial preference for clover and mixed diets is that ruminants are responding to rumen conditions in ways consistent with the satiety theory of Provenza (1995, 1996). The satiety theory predicts that animals will select a diverse diet even when toxins are absent and one food alone could meet their nutritional needs (Provenza 1996)—precisely the outcome observed in the controlled spatial variability experiments reviewed above. A further important consideration in the satiety theory is the definition of a toxin. For instance, the thesis that rapid degradation of protein leads to elevated concentrations of ammonia in the blood is a good example of identification of a compound that is important in intermediary metabolism when rates of formation are controlled, but a "toxin" when rates of formation are not controlled.

In examining this proposition, we start from the position that there are factors that constrain the length of time for which ruminants can eat from a pure clover pasture and that adding grass to the diet helps overcome this constraint. On the surface, the data in Table 3 do not show strong evidence for this, since the total time spent grazing in both experiments was not significantly different between the clover monoculture and the choice treatments. However, when the daily grazing profile was broken down in more detail in the experiment conducted by Marotti (2004), clear differences in meal patterns were observed (Table 4). Consistent with the findings of Penning et al. (1991), lactating cows grazing the clover monoculture and choice treatments took more, shorter meals within each day compared to cows grazing the grass monoculture. Each meal on pure clover lasted 60 min, whereas each meal on grass monoculture lasted nearly twice as long. The question then arises, if animals are able to eat for nearly 2 h continuously on a grass diet before some constraint or cue causes them to stop eating (for example, rumen distention or fill; Mertens, 1994), why do they stop eating on clover within half that time? In Marotti's experiment, the estimated total intake per meal was 3.23 kg DM on the clover monoculture, but 3.77 kg DM on the grass monoculture, indicating that cows on the pure clover were foregoing about 0.5 kg DM intake per meal compared to cows on the pure grass. Therefore it appears that, in this case, satiety was not a direct function of rumen fill.

Several studies using confined feeding treatments have observed that animals alter diet selection to balance out their protein intake (e.g., Kyriazakis and Oldham, 1993; Tolkamp et al., 1998; Villalba and Provenza, 1997), and it seems reasonable to propose that similar responses could occur in free-grazing ruminants. Grazing animals are able to discriminate between pastures with different N content. Cosgrove et al. (2002) offered growing cattle a free choice between low N ryegrass (mean tissue N concentration = 18 g N kg^–1 DM) and high N ryegrass (mean tissue N concentration = 34 g N kg^–1 DM) and found that they showed a partial preference for the high N grass compared to the low N grass (0.73:0.27). A similar result was obtained for sheep, with tissue N concentrations of 45 and 32 g N kg^–1 DM, respectively. The question is, does this discriminatory ability allow ruminants to mix their feeds to ameliorate the effects of high concentrations of rumen-degradable protein in their preferred diet, clover?

Testing this proposition in a grazing situation poses significant logistical challenges, since it is necessary to collect real time information on rumen condition and grazing behavior and then link the two while accounting for lag times between digestion outcomes and ingestion responses. Cosgrove et al. (1999) tackled this idea by infusing urea directly into the rumen of sheep when they grazed on the clover component of a mixed clover–grass choice (thereby increasing the ammonia load associated with clover). In both indoor and outdoor experiments, the strongest response of animals to infusion was to reduce the time spent grazing: there was no evidence of the infusion inducing a switch to grazing grass in an attempt to ameliorate the ammonia challenge.

Marotti (2004) compared the rumen ammonia profiles of sheep fed pure white clover, pure grass, pure birdsfoot trefoil (Lotus corniculatus L.), or mixtures of white clover and perennial ryegrass (one following the other) in an indoor feeding trial lasting 20 d. Animals were fed twice per day, and the rumen metabolite profile was monitored over an 8-h period, which included the 2-h feeding period. When the data were normalized for crude protein intake, sheep grazing pure clover showed a higher peak rumen ammonia concentration than sheep feeding on all other diets, and a faster rise to the peak from the time when eating began (Fig. 1 ). Notably, the daily dry matter intake (DMI) of sheep on the pure clover diet fell to well below maintenance energy levels over the course of the experiment (average daily DMI = 0.175 kg over the last 4 d of the experiment), whereas sheep grazing grass or birdsfoot trefoil maintained their DMI at above 0.5 kg. Sheep grazing pure clover were clearly experiencing a conditioned aversion to the clover diet, which they were only able to tolerate by reducing their daily intake. However, when sheep of similar physiological state to those used in the indoor experiment were allowed a free choice of grass and clover in the field, and received a urea infusion when consuming a meal of either grass or clover, the added ammonia load did not induce dietary switching (Marotti 2004). As observed by Cosgrove et al. (1999), sheep simply reduced their intake of clover when the ammonia challenge was increased by infusion, and did not compensate for the loss of intake by eating more grass. The experiment was inconclusive with regard to rumen ammonia as a cause of increased dietary diversity. It is possible that, in both the studies cited here, the ammonia dosage rate was too high causing malaise and loss of appetite. Definitive data sets that establish, or disprove, the link between rumen state and diet selection remain to be collected.

View larger version (13K): [in this window] [in a new window]   Fig. 1. Fitted curves for rumen ammonia concentration of previously-fasted sheep grazing white clover (——), perennial ryegrass (– - – -), Lotus corniculatus (— - - —), perennial ryegrass followed by white clover (. . . . .), or white clover followed by perennial ryegrass (– – – –) for 2 h (0–120 min). Data are adjusted for the crude protein intake (CPI) of individual sheep. From Marotti (2004).

The possible role of degradable or fermentable C/N balance in feedstuffs has come under scrutiny as an explanation because of the large impact that variability in the C/N ratio has on rumen digestion and metabolism (Dove 1996). The potential effect of excess ammonia on grazing behavior and intake has been discussed above. Recent unpublished data (K.J. Venning, J. Hill, 2005) has demonstrated in vitro the potential of the C/N synchronization theory in controlling rumen ammonia concentrations and provided further evidence to support the observations of Cosgrove et al. (1999) and Marotti (2004). In a series of experiments, sheep were acclimatized to different ratios of subterranean clover (T. subterraneum L.) and perennial ryegrass (from 100% inclusion of subterranean clover to 100% inclusion of perennial ryegrass). Rumen samples were collected and incubated with different ratios of perennial ryegrass and subterranean clover to determine total gas production and changes in concentration of ammonia and volatile fatty acids. Total gas production and rate of production was reduced substantially in the 100% subterranean clover treatment compared to the 100% ryegrass treatment. The most rapid increase in and greatest total gas volumes were recorded in treatments with a ratio of subterranean clover to ryegrass of 0.67:0.33. This observation aligns with Merry et al. (2002) who added perennial ryegrass and red clover silage to an artificial rumen in different proportions and found that the optimal level of microbial protein synthesis occurred with 70% clover and 30% grass.

Another angle on the possible role of the C/N ratio in feed on grazing behavior is to investigate the effect of variability in water soluble carbohydrate (WSC) concentrations in dietary components. Siever-Kelly et al. (1999) used herbicides to temporarily manipulate the WSC concentration of pasture and found that sheep strongly preferred pasture with elevated WSC (mean concentration in total herbage = 112 g kg^–1 DM) compared to untreated pasture (mean WSC concentration in total herbage = 45 g kg^–1 DM). Ciavarella et al. (2000) used shading to manipulate the WSC concentration of phalaris (Phalaris aquatica L.) pastures and observed that sheep consumed 2.6 times more herbage from the unshaded pasture (mean WSC concentration = 126 g kg^–1 DM) than from the shaded pasture (mean WSC concentration = 62 g kg^–1 DM).

Cosgrove et al. (2001) also investigated grazing behavior and intake responses of sheep grazing perennial ryegrass cultivars bred for either "normal" (cv. Fennema, 98 g kg^–1 DM WSC) or high (cv. AberDove, 138 g kg^–1 DM WSC) soluble carbohydrate concentrations. There were no significant differences between pasture treatments in grazing time, intake rate, or daily intake. Francis (2002) offered lactating dairy cows a free choice (50:50 area ratio) of high WSC concentration (cv. AberDawn) and "normal" WSC concentration (cv. AberElan) perennial ryegrass, as well as a free choice between white clover and each of the ryegrass treatments. Unfortunately, the high WSC trait was not expressed in AberDawn (Francis et al., 2002), and animals chose a 50:50 mixture of the two grasses in their diet. The expected partial preference for clover (0.7:0.3) was observed when animals were offered a free choice between white clover and either grass cultivar. This study at least confirms that no other factors are causing discrimination between grasses bred for different carbohydrate concentrations.

Legumes such as white clover can cause bloat in ruminants when eaten quickly and as a high proportion of the diet. Partial preference for clover could be explained by a learned aversion to eating foods that can cause bloating (Provenza et al., 1992). Although Rutter et al. (2004b) found that an antibloat treatment did not affect the diet preference of cattle, circumstantial evidence for this could be drawn from the experiment reported by Cosgrove et al. (1996) and Cosgrove et al. (1997) who compared preferences of heifers for 50:50 area choices of white clover/perennial ryegrass or birdsfoot trefoil/perennial ryegrass. The animals previously had no exposure to birdsfoot trefoil in their diets. When offered a free choice between white clover and ryegrass in summer, animals showed a partial preference for the legume component of 0.65, whereas when offered a choice between birdsfoot trefoil and ryegrass the partial preference for the legume component was 0.75. Birdsfoot trefoil contains condensed tannin at a concentration of about 2% DM, which can prevent bloating. Thus, it could be inferred that the animals showed a stronger partial preference for birdsfoot trefoil because subclinical bloat limited their intake of clover. However, condensed tannin also reduces the amount of protein degradation, and hence ammonia release, in the rumen (Barry and McNabb, 1999, and see Fig. 1), so the higher partial preference for birdsfoot trefoil as the legume component of the choice diet may have been related to lower rumen ammonia concentrations and therefore removal of a constraint to legume intake.

Finally, several species of clover contain secondary compounds such as cyanogenic glycosides which can limit intake at less-than-acute dosages and cause animals to eat other foods (Burritt and Provenza, 2000). These compounds may also cause satiety in some circumstances and this possibility is worthy of consideration in future research.

	HETEROGENEITY, GRAZING PROCESSES, AND PASTURE HARVEST EFFICIENCY

TOP NOTES ABSTRACT INTRODUCTION BIOPHYSICAL AND BEHAVIORAL BASES HETEROGENEITY, GRAZING... ANIMAL PRODUCTION RESPONSES TO... CONCLUSIONS REFERENCES

 
Spatial variability is present in all grazed pastures. The act of grazing (prehension and removal of herbage) is a spatially discrete process subject to variance (Parsons and Dumont, 2003) so grazed pastures are, by definition, heterogeneous with respect to biomass (e.g., Hirata, 2000). Dung and urine are returned to pastures in distinct patches, which have markedly different nutrient concentration compared to unaffected areas and so create differences in plant growth rate. Furthermore, animals avoid areas affected by excreta when grazing from pastures, leaving these areas to grow taller (e.g., Tayler and Rudman, 1966; Ogura and Hirata, 2001) which has subsequent effects on sward structure and composition within, and at the margins of, the patch. Finally, underlying variation in edaphic factors such as nutrient availability, pH, profile depth, and texture interacts with the physiology and environmental tolerances of species present within the pasture mixture to influence the outcomes of interspecific competition for growth resources.

These outcomes can be further affected by active selection by animals for preferred feeds in their diets, as discussed in the preceding sections. Furthermore, the animal has a very powerful effect on the redistribution and fate of nutrients within the grazing area through patterns of dung and urine deposition. This poses challenges for the management of "leaky" nutrients such as N (Whitehead, 1995), which are returned in highly aggregated fashion in urine patches where the N concentration far exceeds the capacity of the soil–plant system to retain the N within the organic cycle (Whitehead, 2000). Throw in the often serious impacts of unpredictable disturbances such as drought events, and it is little wonder that the efficient management of grassland systems for sustainable production and environmental objectives still poses a serious challenge for scientists and farmers alike (e.g., Illius and Hodgson, 1996).

The challenge is too great to tackle in a review such as this, so we choose instead to focus on interactions between the grazing process, heterogeneity in pastures (principally in biomass, but also considering species diversity), and the efficiency with which pastures are harvested for animal feed. In this analysis, we consider two spatial scales: the first, equivalent to the area of individual bite of the grazing animal; and the second, equivalent to the area of paddocks, or discrete, manageable units into which the whole grazing area is subdivided. These can be analyzed using the same principles of pasture regrowth dynamics, the balance between feed supply and demand, and animal preference and selection, and the results applied to the identification of "optimal" management approaches to controlling these factors. We aim to show that the outcomes of the analyses are internally consistent, despite the difference in scale. The key to this is to apply the same set of principles at both scales.

Heterogeneity at the Bite Scale and Implications for Harvest Efficiency
Foraging animals are continuously making choices about where and what they graze within a mixed pasture community. The opportunity for them to make these choices depends on the daily feed demand of the herd or mob relative to the amount of feed offered on a daily basis, and the method by which animals are confined on pasture and moved from one paddock to another. The first of these is under some degree of management control but is also strongly affected by conditions for pasture growth (soil water availability, temperature, radiation) and the rate of accumulation of herbage across the grazing area. Herbage accumulation rate is rarely, if ever, constant in temperate environments, even with the use of irrigation to supplement summer rainfall. The second factor influencing the opportunity for foraging choices, the grazing method, is entirely under management control. Technological advances such as electric fencing have given pasture managers almost complete control over the spatial allocation of feed to mobs or herds of animals and considerable power to manipulate feeding choices, feeding rates, and rates of herbage accumulation. We address these issues in more depth in the following subsection.

It is a simple matter to show that, at typical commercial stocking rates (defined as numbers of animals per hectare, calculated as an average for the total available grazing area), it is impossible for animals to completely defoliate the entire area of pasture available to them every day. Tharmaraj et al. (2003) observed that a lactating dairy cow grazed a mean surface area of pasture of 257 m² d^–1. Greater surface areas were defoliated when the sward was offered at a height of 14 cm compared to 28 cm. Assuming a stocking rate of 2.5 cows ha^–1, then the proportion of the total area available that is grazed per day is in the order of 0.06. Parsons and Chapman (2000) took a different approach to calculating the proportion of total area that is grazed when animals are consuming 50 kg pasture DM ha^–1 d^–1 (equivalent to about 2.5 cows per hectare) and arrived at a figure of 0.07.

Hence, at the individual plant level, defoliation is a discrete event occurring at variable intervals, commonly between 14 and 60 d (Morris, 1969; Chapman and Clark, 1984; Clark et al., 1984). Importantly, even when grazing animals have continuous access to pasture, plants are not grazed continuously (i.e., every day, as the word "continuous" might imply), but sporadically at intervals determined by biting frequency and bite placement (Parsons and Dumont, 2003). This intermittent process is occurring at the bite scale. Therefore, the grazed pasture is an aggregation of bite-sized patches, each with a characteristic residual biomass and regrowth trajectory of the sort shown in Fig. 2 . The regrowth rate of the whole pasture is then the mean regrowth rate of the individual, bite-sized (or whatever scale is deemed appropriate) patches, not the regrowth rate of the mean biomass of the pasture at any given point in time. Thus, "the mean growth rate is not the same as the growth rate of the mean" and to assume otherwise is to introduce significant error to the prediction of pasture growth outcomes (Parsons et al., 2001).

View larger version (17K): [in this window] [in a new window]   Fig. 2. The regrowth dynamics of temperate, grass-based pastures. Unique regrowth curves (only two shown here, Curve 1 from a low residual state and Curve 2 from a greater residual state) depend on the initial state of each patch, Wi, and there is an optimal time of harvest for that patch (denoted t*def) where a straight line drawn from the intercept with the x axis first touches the curve (Parsons and Chapman, 2000). At even greater residual patch states, regrowth is monotonic, implying the patch should be regrazed immediately.

View larger version (152K): [in this window] [in a new window]   Fig. 3. Example of spatial heterogeneity in pasture mass, in a pasture continuously grazed by sheep.

The point of this analysis is not to claim that animals graze sequentially or randomly, but rather to pull out key principles that describe the interaction between pastures and the grazing animal and explain the total intake achieved. The principle of patchiness and heterogeneity is critical in this regard, for it leads us to quite different conclusions compared to those reached if we consider the sward and the grazing process as homogeneous entities or processes (Parsons et al., 2001).

Once we have made the advance to thinking and analyzing at this spatial scale, we can move further by considering animal selection at the patch level, a phenomenon which occurs for much of the annual cycle in temperate environments and grazing systems. Parsons and Dumont (2003) showed that if animals were to graze with a preference for short or medium length or biomass patches over tall or dense patches, then a bimodal distribution of patch states would be created. This outcome is seen in field studies (e.g., Gibb et al., 1997) and results even if animals show only partial rejection of tall patches. Under these circumstances, Parsons and Dumont (2003) predicted that total animal intake would be reduced compared to the situation where animals do not reject tall patches and eat equally from all patch states and that this reduction would occur even at low stocking rates. Thus, the very process of placing bites and exercising selectivity reduces the harvest efficiency in grazed pasture systems, an inefficiency which is conventionally controlled by the use of subdivision and rotational grazing approaches. We return to this point in the next subsection.

The spatially explicit approach allows one further dimension to be added to the analysis: the foraging costs of searching for preferred feeds that are aggregated within the available pasture area and the implications of this for intake. Foraging costs are incurred when animals pass by sites where they could feed to encounter sites where their preferred feed is located. The costs are high when the preference for one food item is high, and/or when the relative abundance of that feed is low (Parsons and Dumont, 2003). Figure 4 illustrates that, when searching costs are low (for example, when the preferred feed is abundant and can be encountered frequently), the animal should be able to achieve a higher proportion of the preferred feed in the diet compared to the proportion of the preferred feed in the pasture on offer. This outcome is supported by experimental manipulation of the spatial aggregation of, for example, grass and clover. Rutter et al. (2005) found that the critical spatial scale of separation of grass and clover required to achieve the reduction in selection costs in beef heifers lies between 12 and 36 cm. On the other hand, if search costs are very high, animals would be expected to graze from all of the food available to them. With intermediate costs, the animals must make some trade-off between eating the preferred diet versus accepting a lower intake. Again, there is empirical evidence to support these predictions (Dumont et al., 2002).

View larger version (18K): [in this window] [in a new window]   Fig. 4. The complex effects of partial preference and of low, medium, and high (specific) foraging costs on the proportion of preferred food in the diet, in relation to its abundance in the vegetation, as predicted by a model that seeks the optimal trade-off between the benefits and costs of foraging selectively (Thornley et al., 1994). Solid diagonal line: high foraging costs; dashed curved line: moderate foraging costs; solid curved line: low foraging costs.

Continuous grazing methods commonly lead to patch grazing and fine-scale heterogeneity (e.g., Fig. 3) which introduces a limitation for meeting high per-hectare productivity targets. This limitation can be overcome using some form of rotational grazing, where animals are moved from paddock to paddock according to some general plan, which can be adjusted as the weather and other conditions dictate. In keeping with the principles adopted for our analysis of heterogeneity, above, we can include this form of grazing in the same conceptual framework as continuous stocking by recognizing that rotational grazing systems effectively confine the variability to discrete paddocks. In other words, the variability is still present under rotational grazing but at a much different and more manageable spatial scale.

Critically, by constraining the variability to the paddock scale, it is possible to exert control over the key factors that determine the regrowth of the "patch" of pasture: the regrowth interval (optimized at point t*_def in Fig. 2) and the patch residual state. Parsons and Dumont (2003) refer to these variables as controlling the "rate of replacement of the resource." For every possible residual patch state, there are a theoretical maximum sustainable yield that can be achieved during regrowth from that patch and a characteristic optimum regrowth interval.

Parsons and Chapman (2000) modeled these dynamic interactions for continuous and rotational grazing systems. Figure 5 shows the outcomes of this analysis. The solid line is the "global optimum" for the range of range of patch states while, in Fig. 5b, the symbols show the predicted outcomes for continuous grazing (solid symbols) and rotational grazing (open symbols). Neither grazing method achieves the optimum for any residual patch state, because there is always variability in the timing and intensity of grazing and therefore spatial variability in stage of regrowth, irrespective of method of grazing. Only mechanical cutting can achieve truly homogeneous defoliation and regrowth. However, at high stocking rates (moving from right to left relative to the x axis in Fig. 5), rotational grazing does not fall as steeply below the optimum in terms of yield as does continuous grazing (Fig. 5b). The explanation for this is seen in the bottom graph in Fig. 5b, where the defoliation interval moves horizontally to the y axis at 30 d because this is the rotation length prescribed in the model, and therefore it is not possible for animals to defoliate the pasture more frequently than every 30 d. Thus, the pasture patches (in this case, paddocks) have the opportunity to regrow for at least 30 d at all stocking rates, and the rate of herbage accumulation at high stocking rates can exceed that in continuous grazing systems where the defoliation interval progressively declines to as low as 5 d. Such short grazing intervals have been measured in grazing studies (e.g., Hodgson and Ollerenshaw, 1969). Model analyses therefore predict that rotational grazing should lead to higher per-hectare production than continuous grazing at higher stocking rates but may be no different at lower stocking rates. This has indeed been observed in empirical studies (for example McMeekan and Walshe, 1963).

View larger version (23K): [in this window] [in a new window]   Fig. 5. The effect of residual patch state on the maximum sustainable yield that may be achieved and the defoliation interval required to achieve this (solid lines) as predicted by a model that seeks optimal solution for all combination of residual patch state and the timing of harvest. (a) the combinations of residual patch state and defoliation intervals that emerge (separate point for each of a range of stocking rates) when animals are assumed to graze patches either at random (solid dots), or in strict sequence (open circles) and in (b) the same for rotational grazing (open symbols, 30 d with animals not present: 1 d grazing) compared to continuous grazing (solid symbols). Arrows show the direction of increasing stocking rate. After Parsons and Chapman (2000).

In essence, the principle emerging from the modeling and empirical studies can be stated thus: when the balance between feed supply and demand is in deficit, then rotational grazing will lead to better per-hectare yield compared to continuous grazing because it allows greater control over defoliation frequency and the regrowth interval of pasture. A feed deficit may arise in a relatively uniform climatic environment as a result of a high stocking rate compared to herbage accumulation rate. This is a common situation on intensively managed, pasture-based livestock farms in cool-temperate New Zealand (Milligan, 1984). In an environment characterized by interannual variation in climate, a feed deficit may occur as a result of poor seasonal conditions for pasture growth, despite the stocking rate being set conservatively at a level based on the expectation of a "worse than average" outcome. Such situations are typical in the Mediterranean-type climate of southern Australia and are unpredictable in their timing and duration.

The real strength, then, of rotational grazing systems in improving the efficiency of pasture harvest is in the power it gives for helping control the balance between feed supply and demand. This is possible because the variability in "patch" state is confined at the paddock level, and paddocks can be managed as separate entities. Thus, the defoliation interval can be adjusted as the herbage accumulation rate changes, to achieve as close as possible to the optimum regrowth interval (t* in Fig. 2) following each defoliation event. This is a key feature of intensive, rotational grazing systems (Saul and Chapman, 2002), and outcomes such as the amount of surplus feed that can be conserved as hay or silage are emergent properties of the management system. A further strength of rotational grazing systems is the power they give in being able to allocate pasture to animals in amounts that are expected to meet their nutritional requirements, and not in excess of that, which would lead to wastage. The principles described here are put into practice in well-managed, pasture-based livestock production systems in New Zealand, Australia, and other parts of the world with suitable climatic conditions and adaptable pasture species using a variety of different grazing methods.

Heterogeneity at the Paddock Scale and Implications for Foraging
Provenza et al. (2003, 2006) consider the implications of different grazing methods for the way in which animals learn to forage. They challenge the traditional concept of "proper" grazing management based on low stocking rates and either continuous or rotational grazing on the basis that it trains animals to "eat the best and leave the rest," leading to increased abundance of the less desirable plant species. The theoretical and empirical accounts of foraging behavior and intake in intensively managed pastures discussed above support this notion. Provenza et al. (2003, 2006) present the opposite case as being learning to "mix the best with the rest." This is clearly more likely to happen in intensive, rotational grazing systems where high stocking densities are applied to paddocks for short periods, forcing animals to either eat quickly from the total feed resource available (limited or zero searching time) or suffer serious losses in intake. Under these conditions, there is a near-complete trade-off between selection and intake.

Such approaches to grazing get closest to "optimizing" the system in terms of per-hectare productivity. However, they also incur penalties as a result of the dynamic interaction between grazing method and interspecific competition. At this point, we must be careful not to ascribe too much importance to the grazing (bite placement and prehension) process when it comes to understanding and explaining the species composition of grazed pasture communities. As noted above, in the experiment reported by Chapman et al. (2003), rotational grazing often led to better per-hectare production outcomes, but nearly always led to poorer per animal outcomes. The latter result was because the clover content of the mixed grass–clover pastures that were used was much lower under rotational grazing than under continuous grazing, and the perennial grass (phalaris) came to dominate the mixture under rotational grazing. A similar result has been noted by Morley et al. (1969) and Culvenor (2000) using similar pasture types, and was also obtained by Hay and Baxter (1989) in New Zealand, using perennial ryegrass–white clover pastures. In both cases, this result was explained by shading from the grass in the tall swards that accumulated with long defoliation intervals under rotational grazing, forcing clover to invest a higher marginal fraction of photosynthetic energy into lamina and petiole to compete for light. The phenomenon of competition between grass and clover for light has been well documented (Woledge, 1978; Woledge and Dennis, 1982; Davidson and Robson, 1985), as has the difference between grass and clover in fractional allocation of biomass to lamina, petiole/sheath, and stem and the implications of this for the relative competitive ability of grass and clover in mixed swards subject to defoliation (Parsons et al., 1991).

That the clover content in the continuously grazed pastures of Chapman et al. (2003) was higher than in the rotationally grazed pastures runs counter to the predictions based on foraging behavior alone. Foraging theory would lead us to expect that the clover would be selectively grazed, leading to its diminution in the mixture. What is important here is the relativity between foraging behavior and selective grazing, feed supply and demand balance, and grazing method. It is not hard to see how very many different results may emerge from grazing experiments comparing, say, rotational grazing and continuous grazing. Indeed, early research comparing different grazing methods created precisely the sort of confusion we might now anticipate with the benefit of hindsight. It was not until the principles of grazing as a spatially discrete process with characteristic defoliation intervals and variance around those intervals (e.g., Clark et al., 1984), grass–clover competition for light (Woledge, 1978 and others), the dynamics of pasture regrowth (Parsons et al., 1988), and factors controlling selection in grazed pastures (e.g., Thornley et al., 1994) were described mechanistically that we gained the ability to explain the range of results. Modeling (e.g., Parsons et al., 1988) has been critically important in drawing the information together and establishing unifying principles which help resolve the confusion.

Which leaves us with the dilemma that we know how to manage grass–clover pastures to optimize control of pasture regrowth for high per-hectare production, but not how to sustain high proportions of the preferred feed of ruminants (the clover) in the mixture at the same time. It so happens that the preferred feed is generally not able to sustain competition with grass when swards are allowed to accumulate a high biomass. So how could we deal with this dilemma? To answer this, we return to the idea developed in the early part of this paper (with another purpose in mind), the spatial separation of grass and clover within the same paddock.

	ANIMAL PRODUCTION RESPONSES TO CONTROLLED SPATIAL VARIABILITY IN PASTURE COMPOSITION

TOP NOTES ABSTRACT INTRODUCTION BIOPHYSICAL AND BEHAVIORAL BASES HETEROGENEITY, GRAZING... ANIMAL PRODUCTION RESPONSES TO... CONCLUSIONS REFERENCES

 
Solving for high per-hectare production, high per-animal production, and stable pasture composition comprising a high proportion of the preferred feed of domestic ruminant livestock appears to require a different approach to the conventional practice of combining preferred and less-preferred feeds in an intermingled species mixture. Interspecific competition for growth resources, active selection by grazing animals, spatial variability in the food resource, and grazing method interact in complex ways which are difficult to predict and control. We have described earlier in this review how spatially separated grass and clover moncultures presented at the paddock scale in 50:50 area ratio allow animals to select a diet that matches their preference for clover versus grass. Simultaneously, the enhanced dietary diversity supports daily intakes similar to those achieved on clover monocultures and appears to allow animals some capacity to overcome constraints to eating that might be associated with a pure clover diet.

Several of the studies that have been conducted to disentangle the factors contributing to the animal grazing behavior response using spatially separated swards have also measured animal production responses to the pasture treatments. These data are summarized in Table 5 and show that per-animal production from the choice treatment matches that from pure clover, which is generally considered the "benchmark" for individual animal performance (e.g., Ulyatt, 1981). For example, the mean average daily liveweight change of sheep for the eight observations in Table 5 where the experiment continued for 30 d or more was exactly the same for the clover monoculture and choice treatments: 259.6 g head^–1 d^–1. Thus, it appears that all the nutritional benefits of clover could be reaped when only 0.5 of the effective grazing area is sown to clover, the preferred species.

View this table: [in this window] [in a new window]   Table 5. Animal production measured on different pasture types.

These results collectively show that spatially separated grass and clover monocultures meet two of the four criteria listed in the first paragraph of this subsection: high per-animal production and a high proportion of the preferred feed in the pasture mixture. Notably, the high per-animal production was achieved with about 0.3 of the diet comprising grass and 0.5 of the grazed area comprising grass. Harris et al. (1997) used chemical manipulation of ryegrass–white clover mixtures to present pastures containing different proportions of white clover (0, 0.25, 0.5, and 0.75, DM basis) to dairy cows. They found that that cows grazing pasture containing 0.5 white clover produced 33% more milk (liters per cow per day) than cows fed pure grass and concluded that a clover proportion of 0.5 would result in 95% of maximum yield achievable on pure clover. These results accord closely with the data in Table 5 and raise the possibility that spatial separation per se may not be important in the animal production response. However, because the grass and clover are separated in space in the choice treatment, we can be confident that a third criterion listed in the opening paragraph, stable pasture composition, is met better by the choice treatment, since interspecific competition between the grass and the clover is eliminated in the arrangement. The clover content of mixed pastures is notoriously difficult to maintain at proportions higher than 0.1 to 0.2 (Caradus et al., 1996), which is clearly too low to reap the benefits of clover for improved animal performance.

An alternative approach to grazing grass and clover as spatially separate swards has been developed by Rutter et al. (2001, 2003) for use with dairy cows. This is called "temporal allocation," and involves allocating dairy cattle to a clover monoculture (only) following morning milking and a grass monoculture (only) following afternoon milking (i.e., roughly mimicking their diurnal pattern of preference for grass and clover). Rutter et al. (2001) demonstrated that the production benefits of grazing grass and clover as separate swards could be achieved using this "temporal allocation" approach and that continuous free choice between adjacent grass and clover monocultures was not necessary. Under rotational grazing, cattle receiving a fresh allocation of clover following morning milking and a fresh allocation of grass following afternoon milking produced over 14% more milk than cows receiving a twice-a-day fresh allocation of a mixed grass–clover sward (Rutter et al., 2003).

The spatial and temporal separation approaches also allow us to entertain the possibility of applying species-specific management to the individual components to help control botanical composition. Examples could include targeting N fertilizer just to the grass monoculture component (and therefore achieving more efficient uptake of this mobile nutrient) and targeting herbicide use for weed control to a broadleaf or grass species background. It is also conceivable that grass and clover could be grown alternately on each half of individual paddocks on a rotation of, say, 4 to 5 yr. This would allow the new grass to benefit from the build-up of N under the previous clover monoculture while the clover should prosper in the low-N environment created by the previous grass monoculture and be better able to out-compete grass weeds because of its ability to fix atmospheric N. It should be noted, however, that clover monocultures can result in similar potential for N leaching losses to heavily fertilized perennial ryegrass pastures after a period of time (MacDuff et al., 1990), presumably because of the build-up of organic and inorganic N produced by N₂ fixation. What is not clear is, how long is the lag-time between clover establishment and the increase in N leaching losses? Or, in other words, what is the interaction between spatial and temporal scales in N leaching?

There is no evidence available currently to judge whether or not the spatially or temporally separated monocultures would also meet the final criterion in the opening paragraph, high per-hectare production. One drawback of clover is that its total annual yield is generally lower (by about 25%) than the yield of well-fertilized perennial ryegrass (Harris and Hoglund, 1977), for reasons that include the energetic costs of maintaining N₂ fixation. Thus, growing clover on half of the available grazing area of a farm would potentially reduce total annual herbage accumulation and total feed availability. Some of this loss in forage supply may be compensated for by better seasonal distribution of forage availability. Clover is more summer active than grass, and its annual growth pattern is not dominated by the spring peak that is typical of perennial ryegrass (Harris and Hoglund, 1977). Chapman and Kenny (2005) used a farm systems modeling approach to show that, for pasture-based dairy production systems in southern Australia, the economic benefit (in terms of extra operating profit) of growing additional feed in summer was nearly double that of growing the same amount of additional feed in spring. Certainly, a 50:50 area ratio mixture of grass and clover would not impact so severely on total annual herbage accumulation as a 100% clover monoculture and so should allow the animal production potential of pure clover to be achieved without the same magnitude of loss in herbage accumulation potential.

Cosgrove et al. (2003) used a farm system model to determine if the greater nutrient intake possible from a spatially separated clover–grass mixture would translate to better whole-farm performance outcomes. They compared scenarios where proportions of 0, 0.05, 0.1, 0.15, or 0.2 of the available grazing area were sown to spatially separated swards and predicted that growing this pasture type on 0.15 to 0.2 of a typical sheep breeding and finishing farm in North Island, New Zealand, could increase total lamb carcass output and net farm profit. If a higher proportion of available area was used for this pasture type, the distribution of feed supply shifted strongly toward summer, which would be appropriate for specialist lamb finishing systems in New Zealand, but incurs penalties for sheep breeding operations based on spring lambing. The concept has been proved, at least in a theoretical sense. The results warrant further investigation if we are to find new solutions to old problems of controlling the performance of high-producing grass–clover pastures for both animal production and environmental management objectives.

	CONCLUSIONS

TOP NOTES ABSTRACT INTRODUCTION BIOPHYSICAL AND BEHAVIORAL BASES HETEROGENEITY, GRAZING... ANIMAL PRODUCTION RESPONSES TO... CONCLUSIONS REFERENCES

 
Considerable progress has been made in recent years in understanding the grazing behavior and intake responses of domestic ruminant livestock to spatial variability in temperate pastures. It is now well established that animals exhibit a partial preference for clover compared to gr