Pasture intake of grazing livestock needs to be estimated to allow determination of energy and nutrient intakes. It is commonly estimated by difference, subtracting intakes of other feeds from estimated needs for dry matter or energy. However, these estimates are often erroneous, because they do not take individual animal variation for growth, reproductive status or activity level into account. One method that has had success in grazing ruminants has been the use of markers, or tracers, to estimate fecal output and nutrient digestibility. External markers are dosed to the animal and can be used to determine fecal output. Internal markers are an inherent part of the diet in question and can be used to determine dry matter and nutrient digestibilities. These estimates can then be used to give estimates of intake. These studies were conducted to evaluate the effectiveness of traditional marker methods in determining fecal output, digestibility, and thus intake in grazing horses. The first trial was conducted on 8 mature mares and geldings, housed in stalls, to determine if a common external marker, Cr, could be used to determine fecal output. Horses were dosed once daily with a molasses, Cr, and hay mixture for 12 d. Feces were collected throughout the day into individual tubs so that total fecal output (TC) could be measured. Daily fecal Cr excretion values (Ct, mg/kg DM) were fit to a monoexponential equation with one rate constant (k), rising to an asymptote (Ca): Ct = Ca - Ca.e-kt. Superior fits were found when a delay (d) was incorporated into the equation, estimating the time required for Cr to enter the prefecal pool: Ct = Ca - Ca.e-k(t-d). Estimates of fecal output (FO) were calculated using the equation: FO = Cr dose-d / Ca and provided good estimates when compared to TC values. Subsequent trials evaluated to use of internal markers and more frequent dosing of Cr to improve estimates of intake. Eight mature geldings were housed in stalls and were fed 2 hays in a replicated Latin Square design. The monoexponential equation with the delay continued to fit the data well. Thrice daily dosing of Cr improved the predictions of FO, when dosing was every 8 h. The internal marker, yttrium (Y) consistently overestimated digestibility (D). The internal markers, n-alkanes, gave a better estimate of digestibility. When the digestibility estimates were combined with the FO estimates to estimate dry matter intake (DMI, kg/d): DMI = [FO / (1-D)]*100, the combination including n-alkanes gave better estimates. Further studies found that dosing Cr for 12 d did not improve the fit of the monoexponential equation compared to dosing for only 8 d. Marker methods that had been developed in stalls were applied to grazing horses, and results continued to be promising. / Ph. D.

A marker model of fecal kinetics using chromic oxide (Cr) or ytterbium chloride (Yb) is being developed for grazing horses. The model consists of removal of feces at a constant rate from a single compartment, the prefecal mass. It was tested in experiments on stall-fed horses in the context of digestion balance trials. Following the preliminary work of Holland et al., (1998), three improvements in experimental design were tested. First, the rate constants were determined from both the administration and post-administration curve of the one-compartment model. Second, markers were administered three times a day to reduce diurnal variation in fecal marker concentration. Third, yttrium (Y) and Yb were tested were tested as internal markers, for the estimation of digestibility of hay and supplements, respectively. Eight horses were fed Diet 1 (orchardgrass/alfalfa mixed, OG) or Diet 2 (tall fescue/alfalfa mixed, TF) in Exp.1, and Diet 3 (OG plus fat-and-fiber supplement, OGFF) or Diet 4 (OG plus sugar-and-starch supplement, OGSS) in Exp.2. Balance-marker experiments were conducted for 17 and 20 d, with 7 and 10 d of dietary accommodation in Exp.1 and 2, respectively. Chromic oxide and Yb were administered orally and fecal samples were collected every 8 h for 8 d. Dry matter, Cr, Yb and Y were measured in feeds and feces. In balance experiments, estimates of DMD (D_E) using Y, were determined precisely (SE 1 to 3 %) for hay and hay and supplement diets. Linear relationships, correlations and calibration curves were determined, validating Y as a marker. Mean daily fecal Cr data (C_t) at time t (days) including a delay (d) were fitted to a single exponential, with one rate constant (k), rising to an asymptote (C_a): C_t = C_a - C_a·e^-k(t-d) Diets 1 and 2 had two sets of C_tdata, total collection (a) and fecal grab data (b), and each set was used in model development. Diets 3 and 4 had two sets of C_t data (both using fecal grab data), Cr marker dilution (3Cr and 4Cr) and Yb marker dilution (3Yb and 4Yb). For pooled data, delays of 3 to 6 h (Diets 1a, 1b, 2a and 2b) and delays of 5 to 7 h (Diets 3Cr, 4Cr, 3Yb and 4Yb) gave best fits (highest estimates of R²). The delays introduced to the Cr model for both 3Cr and 4Cr diets did not correspond to the preliminary study (Holland et al., 1998), where a 2 h delay gave the best fit in the model for horses fed hay and supplement. The present estimates may more realistically relate to mouth-to-cecum transport times, because the marker was administered three times a day instead of once, and the initial part of the tracer curve was more precisely defined. The results showed that fecal Cr kinetics could be calibrated precisely (SE 1 to 3 %) to predict fecal DM output of horses fed Diets 1b, 2b, 3a but not 4a. Similarly, fecal Yb kinetics could be calibrated to predict fecal DM output of horses fed Diet 3b but not 4b. The rate constants yielded turnover times (TT) that were longer with hay and supplement diets, than with hay alone, and which contrast with previous findings in the horse. However, the longer TT were similar to slower rates of marker excretion in sheep fed concentrates instead of all-roughage diets, suggesting that the lower fiber content retarded the rate of propulsion of digesta through the digestive tract. For two of the eight models of fecal kinetics, the rate constants of the post-administration curve were not well determined by the data, and rate constants from the administration curve were used. In future experiments, more frequent fecal sample collection during the post-administration period may improve rate constant determination. Improvements in diurnal variation of fecal marker concentration were obtained by dosing three times a day. But discrepancies between Cr and Yb concentration means of diurnal samples and combined samples showed incomplete mixing, the major source of tracer error. Therefore more frequent marker administration and fecal samples should be tested in future experiments to achieve more thorough mixing in the prefecal mass for modeling fecal kinetics, and in the small intestine for estimating digestibility. / Master of Science

Determining intake of feeds in horses is an important factor in incorporating supplements in their diets. Fecal recoveries (R), fecal output (FO), dry matter digestibility (DMD) and dry matter intake (DMI) were estimated using alkanes as markers in 8 thoroughbred geldings. The experiment compared two diets in a 2 X 2 latin square experiment. The diets were mixed grass hay only (H) and the same hay plus a cereal by-product (H + CBP). The cereal by-product (CBP) was the high fat component added to feeds at Virginia Tech's Middleburg Agricultural Research and Extension Center. The apparent digestibility of ether extract (EE) and other nutrients in the H and H + CBP, as well as the partial digestibility of CBP were also determined. The periods were 21 d each with a dietary accommodation period followed by eight days of dosing the even chain alkanes dotriacontaine (C32) and hexatriacontane (C36) as external markers. Total collection (TC) was performed the last 4 d of dosing. The results show that mean recoveries of alkanes were close to 100%, but the range for individual alkanes was wide, and the pattern of recoveries for alkanes of different chain length was inconsistent from feed to feed. The results also indicate that mean estimates of the DMI, DMD and FO of a feed, such as H or H + CBP, are determined with reasonable accuracy by means of alkane markers. In contrast, alkane estimates of DMI and DMD in an individual horse fail to predict corresponding TC estimates. The alkane estimate of FO in an individual horse predicts a TC value with error of 16.4%. The CBP was found to be an excellent source of EE, CP and fiber but a poor source of Ca. / Master of Science