An additional concern when matching forages with the needs of livestock is that of matching forage allowance to forage consumption by grazing animals. To supply adequate amounts of forage to meet animal nutrient needs while preventing the adverse effects of overgrazing, it is crucial that factors affecting forage consumption be understood. Dry matter intake is responsible for 60 to 90% of the variation in digestible dry matter consumption (Mertens, 1994) making it the most important animal or feed characteristic influencing animal performance. Furthermore, accurate estimates of feed intake are essential in the prediction of rate of gain and in the utilization of equation- predicting nutrient requirements of beef cattle (NRC, 1996). Unfortunately, the factors that regulate dry matter intake by ruminants are complex and not fully understood (NRC, 1996).

In ruminants fed forage-based diets. Van Soest (1965) found voluntary feed intake to be related to the concentration of neutral detergent fiber (NDF, r = -.65) across grasses and legumes. Mertens (1987) further refined the use of NDF to predict dry matter intake, determining that dairy cows consume 1.2% of their body weight per day as NDF if intake was limited by ruminal capacity. This relationship seems to be caused by the filling properties of fiber which is more slowly fermented in the reticulorumen or passes from the reticulorumen at a slower rate than nonfibrous feed components (Jung and Alien, 1995). The effect of fiber on ruminal fill, however, is not only related to fiber concentration, but also to the rate and extent of fiber digestion and the rate of fiber passage (Dado and Alien, 1995). However, digestibilities of the forage NDF varies considerably with plant species and maturity. In an evaluation of 304 samples of temperate and tropical forages, Traxler et al. (1998) found that a mean NDF digestibility of 34.1% with a standard deviation of 16.0. Thus, total NDF concentration of a forage may not consistently predict forage intake in animals fed high roughage diets.

Because it is a more uniform constituent and may only disappear from the rumen by passage into the lower gastrointestinal tract, NDF that is indigestible over an unlimited period of fermentation, has been suggested as a more consistent determinant of rumen fill, and thus forage intake, than total NDF (Llamas-Lamas and Combs, 1990). However, results of experiments evaluating this relationship have been variable. Muller et al. (1972) found sheep fed corn stover silage from a brown midrib hybrid consumed 29% more dry matter than those fed corn stover silage from a normal hybrid with an equal concentration of NDF, but a 20% lower fiber digestibility. Similarly, Dado and Alien (1996) found that dairy cows consumed greater quantities of ensiled alfalfa selected for higher NDF digestibility, although the results of this experiment were confounded with NDF concentration. Robinson and McQueen (1997) also observed that dry matter intake and milk production of dairy cows increased with increasing fermentability of the dietary forage. In a recent study, Karsli and Russell (unpublished data) found that while the intake of total NDF by sheep fed chopped corn stalks with varying proportions of hay had a standard deviation of 47.5 g/d, the amount of undigested NDF excreted in feces had a standard deviation of 13 g/d. In contrast, dry matter intake has been found to be more related to NDF concentration than fiber digestibility in other studies (Varga et al., 1984; Miller et al., 1990). However, the lack of the effect of NDF digestibility on intake may have resulted from the relatively low NDF concentrations or high NDF digestibilities of the forages used in these experiments, or the confounding effects of forage-to-concentrate ratio. Thus, indigestible NDF concentration may significantly affect intake in animals that consume diets with high forage concentrations as is more common in beef cow-calf production than dairy cow production.

Although indigestible NDF seems related to forage intake at various forage allowances, 24 factors associated with forage intake have been identified (Minson and Wilson, 1994). When forage mass, whether measured as density (Allden and Whittaker, 1970) or sward height (Penning et al., 1984), is low enough to limit animal bite size, the feed intake is limited by the number of bites that an animal can take each day (Hendrickson and Minson, 1980). In addition, the types and concentration of chemical constituents present in individual plant species in a mixed sward will influence forage selection by grazing animals (Minson and Wilson, 1994). Thus, equations predicting voluntary forage intake by grazing animals will have to consider not only the indigestible NDF concentration of the feed, but also the physical characteristics of the sward. Furthermore, because deficiencies of other nutrients such as N (Adamu et al., 1989) P, S or Co (NRC, 1996) also affect forage intake, equations predicting feed intake will either have to assume that these nutrients are present in optimal concentrations, or these would have to be included in predictive equations as well.

Economic and risk implications.

Several studies have examined retained ownership for beef cattle production. Kearl (1972), Hewlett and Workman (1978), and Pfeiffer (1986) showed that carrying calves over winter, summering them on grass, and then selling as long yearlings was more profitable and had less income variation than selling calves at fall weaning. Lambert (1989) used a stochastic programming model to examine retained ownership decisions by a hypothetical ranch. He concluded that it was optimal to sell spring-born calves the following spring (May) after wintering the calves to gain about 2.4 to 2.75 pounds per day. His model was constrained such that selling yearling cattle after May and prior to October was not possible. It may be that an intermediate date would have been better. Garoian et al. (1990) used dynamic programming to examine the same issues. They concluded that selling spring-born calves as yearling after grazing grass in the summer had higher expected net returns than selling calves at weaning for most beginning states of nature (different range conditions). Their model made adjustments in numbers of cows that could be carried (fixed grazing resources) when yearlings were in the system. Watt et al. (1987) used prices and costs for 1958-1984 to examine various retained ownership strategies. They concluded that m most years some form of retained ownership was profitable. Few studies have examined the economics of ownership from calving to the rail and the impacts of high-forage systems on beef quality.

Risk analysis has been studied as part of cow-calf and retained ownership systems. Schroeder and Featherstone (1990) used a discrete stochastic programming model determine optimal retention and marketing decisions for cow-calf producers. They concluded that it generally was more attractive to retain heifers than steer calves. The desirability of calf retention was inversely related to the risk aversion of the producer. Rawlins and Bernardo (1991) concluded that efficient ranch plans were sensitive both to the specification of the risk criteria and to the producer's willingness to accept risk. They used minimization of total absolute deviations (MOTAD) in a linear programming model and target MOTAD to evaluate risk-return tradeoffs for various beef-forage systems available to Oklahoma producers. McNeley (1996) used a discrete stochastic programming (DSP) model to study steer retention decisions for cow-calf producers. The more risk-averse producers would keep few steers beyond weaning. Risk-neutral and slightly risk-averse producers were predicted to retain ownership on the steers through finishing. High growth potential steers were more often retained for a longer time frame compared to medium growth potential. Little et al. (1994) studied ten beef cattle bred" systems involving different breeds. Retained ownership decisions also were examined for each breeding system. Their analysis used linear programming and simulation techniques to generate functions that were ranked with stochastic dominance criteria. They found negative returns for all systems when selling weaned calves. The breeding system with medium-frame, medium-milk breeds in rotation with large-frame terminal sires was the most profitable. The optimal strategy for this system was to retain ownership on the calves through the finishing period. Kolajo and Martin (1994) concluded that longer grazing periods for stocker steers and heifers in Alabama were most profitable and least risky. Shorter grain-finishing periods were better, but grain finishing reduced overall profitability. Their study utilized modeled production coupled with MOTAD for the economic analysis.