Search

The curve number (CN) method for estimating direct runoff from a rainfall event is commonly used in hydrologic investigations, but currently accepted CNs for no-till cropping might need revising. We derived CNs from ranked rainfall-runoff data gathered from 1972 to 2010 from a 2.7 ha, zero-order watershed (P1) near Athens, Georgia. The watershed was initially managed under conventional tillage for 2.5 years, followed by double-cropped continuous no-till rotations. During conventional tillage, on average 19% of the rainfall was partitioned into runoff (median 13%). The mean and median CN was 82 and 85, respectively. Least square fitting of CN versus rainfall produced an asymptotic CN of 81 (R 2 = 0.61). In contrast, the mean and median runoff during no-till was 7% and 0.6% of the rainfall, respectively. The mean CN equaled the median CN of 62. The asymptotic CN was 58 (R 2 = 0.78), compared with 72 from standard tables. When estimated using CN of 72, the mean runoff during no-till was 142% greater than the measured mean runoff. These results support the hypothesis that CNs for no-till cropping systems should be smaller than currently accepted values for comparable watersheds in the region. This has implications for hydrologic modeling for water quantity and quality where the use of the CN method is ubiquitous.

Storm events that increase flow rates can disturb sediments and produce overland runoff in watersheds with animal agriculture, and, thus, can increase surface water concentrations of fecal bacteria and risk to public health. We tested the hypothesis that strategically designed and placed ponds in watersheds with animal agriculture would attenuate downstream fluxes of fecal bacteria. We measured concentrations and fluxes of fecal indicator bacteria (commensal Escherichia coli and fecal enterococci) and manure pathogens (Salmonella and E. coli 0157:H7) in in- and outflows of Bishop Pond in the Southern Piedmont of Georgia during three storm events and in- and outflow concentrations and fluxes of fecal indicator bacteria at Ponds A and C in the Coastal Plain of Georgia during two storm events. Mean concentrations and fluxes of fecal indicator bacteria associated with pond in- and outflow during hydrograph rise, peak, fall, and 5-days after peak flow at Bishop Pond were significantly greater than their mean base flow concentrations and fluxes. In storm flow Bishop Pond significantly reduced the outflow concentrations and fluxes of fecal indicator bacteria compared with corresponding inflow measurements. Unlike fecal indicator bacteria, Bishop Pond appeared not to reduce outflow concentrations and fluxes of Salmonella or E. coli 0157:H7. At Ponds A and C in the Coastal Plain mean in- and outflow concentrations and fluxes of the fecal indicator bacteria associated with the hydrograph rise and peak flows of the storms were not different. Bishop Pond, with a length to width ratio of 3.3, attenuated downstream fluxes of fecal bacteria. In contrast, Ponds A and C were not effective at reducing downstream fluxes of fecal bacteria under storm flow conditions. The ineffectiveness of Ponds A and C may be attributed to their having length to width ratios of 1.2 and 2.5, respectively, both of which are below the minimum for effective pond performance. Our results indicated that in the humid Southeast an appropriately placed and configured pond in watersheds with animal agriculture can reduce storm flow loads of fecal indicator bacteria but not necessarily pathogenic E. coli 0157:H7.

Gypsum has been shown to reduce runoff on soils prone to crust formation in the southeastern United States. Increased infiltration from gypsum applications could therefore help reduce runoff P and other nutrient losses from application of broiler litter (BL), a nutrient-rich fertilizer. In rainfall simulation experiments in June 2009 and May 2011, runoff and nutrient (N, P, Ca, Mg) losses in runoff were compared among treatments consisting of 0, 2.2, 4.5, and 9.0 Mg ha⁻¹ flue gas desulfurization gypsum (FGDG) combined with 13.5 Mg ha⁻¹ of BL and two control treatments of (0–0) and (9.0–0.0) Mg ha⁻¹ (FGDG-BL). A randomized block design with three replications was set up on a Cecil (Typic Kanhapludult) soil growing Coastal bermudagrass (Cynodon dactylon L.) for hay near Watkinsville, Georgia. Amendments were applied each April from 2009 to 2011. A reduction in runoff of 30% each year from the (9.0–13.5) compared with the (0–0) treatment was not statistically significant. Gypsum was effective in reducing concentration and load in one of the two study years: P and NH₄–N in 2009 (up to 83%) and NO₃–N in 2011 (up to 73%). A combination of factors related to weather extremes, timing of FGDG and BL applications, and their implication on soil and vegetation responses at different landscape positions might have caused the different responses between years. Additional studies focused on isolating the impacts of such factors would be helpful to ascertain the effectiveness of multiyear applications of FGDG as a best management practice to reduce P and other nutrient losses in soils of the southeastern United States.

Switchgrass (Panicum virgatum L.; SG) may have CP below recommended concentrations for feeding ruminants that limit animal performance when cut as hay and fed as the sole diet. Three experiments were conducted to determine the influence of white clover (Trifolium repens. L.) at 0, 15, and 30% of DM (Exp. 1), or alfalfa (Medicago sativa L.) at 0, 25, and 50% of DM with either headed SG (Exp. 2) or vegetative SG (Exp. 3) on steer DMI (kg/100 kg of BW) and DM digestion (g/kg). Eating behavior (Exp. 1) and masticated feed characteristics (Exp. 2 and 3) were also examined. Steer (Bos taurus L.) DM digestion of SG (Exp. 1) increased linearly (P < 0.01) from 544 g/kg on SG to 584 g/kg when clover composed 30% of the offered diet. However, steer DMI of headed SG (Exp. 2) was 1.61 kg/100 kg of BW with a DM digestion of 480 g/kg when fed alone, and both increased linearly (P < 0.01) to 2.68 kg/100 kg of BW and 578 g/kg, respectively, when alfalfa composed 50% of the offered diet. The DM digestion of vegetative SG (Exp. 3) increased linearly (P = 0.09) from 541 when fed alone to 575 g/kg when alfalfa composed 50% of the offered diet. The CP of masticated feed (Exp. 2 and 3) increased linearly (P < 0.01) and NDF decreased linearly (P < 0.01) with increasing alfalfa. Perennial legumes can be a useful source of CP in SG hay diets.

Year-round grazing in animal production systems has potential in the mid-Atlantic United States. Tall fescue [Lolium arundinaceum (Schreb.) S. J. Darbysh.] is generally the cool-season perennial and bermudagrass [Cynodon dactylon (L.) Pers.] is the warm-season perennial in these systems. This study evaluated yellow bluestem (Bothriochloa ischemum L.; YBS) and Coastal panicgrass (Panicum amarum var. amarulum; CPG) for their quality and productivity with a tall fescue-bermudagrass system (BGS) as a control. Steer ADG (kg/d) was less (P = 0.05) in the BGS (0.66) than the mean of 0.77 from YBS (0.68) and CPG (0.86), which differed (P = 0.02). Steer ADG (kg/d) in spring (0.95) was greater (P < 0.01) than summer (0.53), but pasture species interacted with season (P = 0.02). The ADG of CPG was stable across season (from 0.97 to 0.76 kg/d), whereas BGS or YBS declined from 0.97 and 0.92, respectively, to 0.37 and 0.49 kg/d. Dietary in vitro true OM disappearance was similar among species (P = 0.29) and seasons (P = 0.62), averaging 803 g/kg; however, DMI (kg/d per 100 kg of BW) was least (P = 0.05) from the BGS (2.64) compared with YBS (2.97) and CPG (3.40) (P = 0.09). The BGS was stocked greater (P < 0.01) at 12.3 compared with 6.2 and 4.9 steers/ha for YBS and CPG, respectively. Adjusting for days tall fescue was grazed (23%) still favored the BGS. No difference in steer gain (P = 0.95) and effective feed units (P = 0.54) occurred between YBS and CPG (473 and 2,754 kg/ha).

Animal agriculture in watersheds produces manure bacteria that may contaminate surface waters and put public health at risk. We measured fecal indicator bacteria (commensal Escherichia coli and fecal enterococci) and manure pathogens (Salmonella and E. coli 0157:H7), and physical-chemical parameters in pond inflow, within pond, pond outflow, and pond sediments in three ponds in agricultural watersheds. Bishop Pond with perennial inflow and outflow is located in the Piedmont, and Ponds A and C with ephemeral inflow and outflow in the Coastal Plain of Georgia. Bromide and chloride tracer experiments at Bishop Pond reflected a residence time much greater than that estimated by two models, and indicated that complete mixing within Bishop Pond was never obtained. The long residence time meant that fecal bacteria were exposed to solar UV-radiation and microbial predation. At Bishop Pond outflow concentrations of fecal indicator bacteria were significantly less than inflow concentrations; such was not observed at Ponds A and C. Both Salmonella and E. coli 0157:H7 were measured when concomitant concentrations of commensal E. coli were below the criterion for surface water impairment indicating problems with the effectiveness of indicator organisms. Bishop Pond improved down stream water quality; whereas, Ponds A and C with ephemeral inflow and outflow and possibly greater nutrient concentrations within the two ponds appeared to be less effective in improving down stream water quality.

Tall fescue (Festuca arundinacea Schreb.) is a well-adapted perennial pasture species utilized across the north-south transition zone of the United States. This 3-yr trial evaluated the influence of three forage masses (FM) on steer and pasture responses for spring (April to July) and fall (October) growths. Spring FM averaged 1450 kg ha ―1 (Short), 3032 kg ha ―1 (Medium), and 3812 kg ha ―1 (Tall) for the three treatments. Canopy in vitro true organic matter disappearance (ITOD) decreased linearly from Short (797 g kg ―1 ) to Tall (771 g kg ―1 ). Steer daily gains (0.88 kg) and gains per hectare (392 kg ha ―1 yr ―1 ) were not altered by FM, but the mean number of steers per hectare decreased linearly from Short (8.2 steers ha ―1 ) to Tall (6.1 steers ha ―1 ) as did effective feed units per hectare (2256 to 1712 kg ha ―1 yr ―1 ). Canopies were predominately leaf (53%) and averaged 847 g kg ―1 ITOD with estimates similar among FM treatments. Diet ITOD declined linearly from Short (870 g kg ―1 ) to Tall (839 g kg ―1 ) as did crude protein (CP) (from 235 to 192 g kg ―1 ). In the fall, FM averaged 1556, 3228, and 3494 kg ha ―1 for the three FM treatments. Canopy ITOD averaged 774 g kg ―1 and was not altered by FM. Canopy leaves averaged 60% with an ITOD of 864 g kg ―1 and were similar among FM treatments. Diets selected averaged 834 g kg ―1 ITOD and 233 g kg ―1 CP and were not altered by FM. Tall fescue pasture free of toxic endophyte needs to be effectively utilized to enhance production and can be managed flexibly in grazing systems.

Poultry (Gallus gallus domesticus L.) litter (PL) is a readily available nutrient source for crop production in the Southeast USA. Long-term PL application may alter availability of N and the effect may be dependent on tillage practice. Tillage [no till (NT) vs. conventional (CT)] and N source (PL vs. commercial fertilizer CF) effects on N availability and plant uptake were evaluated in years 9, 10, and 11 of a long-term cropping systems study at the United States Department of Agriculture, Agricultural Research Service, J. Phil Campbell Sr. Natural Resource Conservation Center, Watkinsville, GA, USA. Mineral N in the top 10 cm, measured in situ, varied each year and was influenced by time, tillage, and N source. In 2003 (year 9), soil mineral N content was greater in CT–CF (100 kg ha−1) than in NT–PL (95 kg ha−1) but in 2004 (year 10) and 2005 (year 11) it was lower in CT–CF (93 and 60 kg ha−1) compared to NT–PL (140 and 71 kg ha−1). Nitrogen mineralization rates were generally greater for PL than for CF treatments with the difference being almost 1 kg ha−1 day−1 in 2003. Mineralization rates were greater for NT–PL compared to CT–CF in 2004 and 2005. Across the three growing seasons, corn (Zea mays L.) aboveground biomass was consistently greater in the NT–PL treatment than in the NT–CF and CT–CF treatments. Correlation between aboveground biomass and N mineralization was greater for PL than for CF (0.75 vs. 0.48). Patterns of N mineralization and total soil mineral N indicated that the distribution of N through the growing season more closely matched corn N demand in PL treatments. Results indicate that improved N availability through the growing season, by combining NT and PL, can result in more profitable corn production in the southeast.

In situ and in vitro experiments were performed to assess the effects of solar radiation and predation by indigenous microflora on the relative die-off rates of a toxigenic strain of Escherichia coli 0157:H7, commensal E. coli, and fecal enterococci in surface waters from ponds in agricultural watersheds. The objective of these experiments was to discern a mechanism of persistence of E. coli 0157:H7 in surface waters compared to fecal indicator bacteria. Results of these experiments indicated that E. coli and fecal enterococci were affected by both insolation and apparent predation; whereas E. coli 0157:H7 appeared to be resistant to both of these environmental stressors. The number of days to reach 99% die-off (T99-values) for E. coli 0157:H7 was significantly greater than that for the indicator bacteria. The capacity to prolong die-off may be connected to the apparent persistence of E. coli 0157:H7 in surface waters.

Management strategies to accommodate switchgrass (Panicum virgatum L.) in production systems have been reported but impact of variation in forage mass (FM) on animal responses and pasture stands are lacking. This 4-yr study assessed canopy and stand characteristic and steer (Bos taurus) responses when 'Kanlow' switchgrass was continuously and variably stocked to produce Short (S) (10―16 cm), Medium (M) (20―26 cm), and Tall (T) (35―41 cm) canopy heights. From April to mid-July (Period 1) FM for S, M, and T average 818, 2164, and 3294 kg ha ―1 , respectively. Canopies had similar in vitro organic matter disappearance (IVOMD) (626 g kg ―1 ), but increased in neutral detergent fiber (NDF) from S (645 g kg ―1 ) to T (665 g kg ―1 ). Diets decreased in IVOMD from S (744 g kg ―1 ) to T (706 g kg ―1 ). Digesta kinetics data were not altered and average daily gain was similar among FM (1.11 kg d ―1 ). From mid-July to late September (Period 2), FM averaged 465, 1282, and 2388 kg ha ―1 , respectively, for S, M, and T. Whole canopy IVOMD increased from S (484 g kg ―1 ) to T (538 g kg ―1 ), whereas NDF was not altered (696 g kg ―1 ). Steers selected a diet similar in IVOMD (685 g kg ―1 ) and NDF (667 g kg ―1 ) among FM. Estimates of dry matter intake (kg 100 ―1 kg body weight) increased from S (1.75) to T (2.54). Tiller density decreased from T to S. Continuously stocked switchgrass at 20 to 40 cm remained vegetative with 1279 tillers m ―2 after 3 yr of grazing.

Pyrolysis is the anaerobic thermal conversion of biomass for energy production. It offers an option of returning carbon and nutrients to the soil while producing energy. The Ultisols in the southeastern United States have inherently low soil organic carbon and fertility, and may benefit from the addition ofbiochar from pyrolysis. Our objectives were to evaluate the effect ofpeanut hull and pine chip biochars on soil nutrients, corn (Zea mays L.) nutrient status and yield in a Kandiudult for two growing seasons (2006 and 2007). Experiments for each biochar source were conducted as completely randomized designs with the biochar applied at 0, 11, and 22 Mg ha -1 with and without N fertilizer. Nitrogen in the peanut hull biochar (209 kg ha -1 at 11 Mg ha -1 rate) was not available during the study based on corn tissue concentrations. The peanut hull biochar linearly increased Mehlich I K, Ca, and Mg in the surface soil (0-15 cm). The increased available K was reflected in the plant tissue analysis at corn stage R1 in 2006, but not in 2007. Pine chip biochar decreased soil pH, but had no effect on other nutrients except Mehlich I Ca. In the peanut hull biochar experiment, grain yields decreased at the 22 Mg ha -1 rate in the fertilized treatments. In the pine chip biochar experiment, grain yields decreased linearly with application rate in 2006, but this did not persist in 2007. Overall yield responses to biochar were smaller than expected based on previous studies.

Effective laboratory methods for identifying avian influenza virus (AIV) in wild bird populations are crucial to understanding the ecology of this pathogen. The standard method has been AIV isolation in chorioallantoic sac (CAS) of specific-pathogen-free embryonating chicken eggs (ECE), but in one study, combined use of yolk-sac (YS) and chorioallantoic membrane inoculation routes increased the number of virus isolations. In addition, cell culture for AIV isolation has been used. Most recently, real-time reverse transcriptase (RRT)-PCR has been used to detect AIV genome in surveillance samples. The purpose of this study was to develop a diagnostic decision tree that would increase AIV isolations from wild bird surveillance samples when using combinations of detection and isolation methods under our laboratory conditions. Attempts to identify AIV for 50 wild bird surveillance samples were accomplished via isolation in ECE using CAS and YS routes of inoculation, and in Madin-Darby canine kidney (MDCK) cells, and by AIV matrix gene detection using RRT-PCR. AIV was isolated from 36% of samples by CAS inoculation and 46% samples by YS inoculation using ECE, isolated from 20% of samples in MDCK cells, and detected in 54% of the samples by RRT-PCR. The AIV was isolated in ECE in 13 samples by both inoculation routes, five additional samples by allantoic, and 10 additional samples by yolk-sac inoculation, increasing the positive isolation of AIV in ECE to 56%. Allantoic inoculation and RRT-PCR detected AIV in 14 samples, with four additional samples by allantoic route alone and 13 additional samples by RRT-PCR. Our data indicate that addition of YS inoculation of ECE will increase isolation of AIV from wild bird surveillance samples. If we exclude the confirmation RT-PCR test, cost analysis for our laboratory indicates that RRT-PCR is an economical choice for screening samples before doing virus isolation in ECE if the AIV frequency is low in the samples. In contrast, isolation in ECE via CAS and YS inoculation routes without prescreening by RRT-PCR was most efficient and cost-effective if the samples had an expected high frequency of AIV.

Improved cool-season grass cultivars may add production potential to ruminant enterprises across the North–South transition zone. Quality among hays of ‘MaxQ’ (‘Jesup’ with novel endophyte), HM4 (‘HiMag’ with novel endophyte No. 4) and ‘Cajun’ (without endophyte) tall fescues [Lolium arundinaceum (Schreb.) Darbysh.] and ‘Persist’ orchardgrass (Dactylis glomerata L.) was evaluated. Forage was harvested in the fl ag-leaf stage in three of 4 yr and a regrowth (late fl ag-leaf to heads-emerging stage) in 1 yr. Goats (four trials) consumed MaxQ, HM4, and Persist similarly (P = 0.12; mean = 2.49 kg 100 –1 kg body weight [BW]) and Cajun least (P

Eastern gamagrass [Tripsacum dactyloides (L.) L.] (EG), a native warm-season perennial grass, lacks evaluation for use in grazing systems. Our objective was to test EG in a 4-yr experiment to estimate forage mass (FM) that maximizes steer (Bos taurus) performance and pasture productivity. Pasture canopy characteristics, diet selection by grazing steers, and stand persistence from EG continuously grazed at Short, Medium, and Tall heights and two rotationally stocked treatments were compared with continuously stocked ‘Coastal’ bermudagrass [Cynodon dactylon (L.) Pers.] (BG). The EG had proportionally more green leaf (78.2 vs. 24.5%), less stem (4.4 vs. 45.2%), less heads (0.4 vs. 5.6%), and similar dead tissue (17.0 vs. 24.8%). Plant fractions were similar in proportion among EG treatments. Digestible leaf mass was greater in EG than in BG (P < 0.01; 713 vs. 292 kg ha–1) and dominated the EG canopies with a linear (P < 0.01) increase in leaf mass with increasing FM (194–922 kg ha–1). Diets were similar in IVTOD (738 g kg–1), CP (151 g kg–1), and NDF (654 g kg–1) when continuously stocked but greater (P < 0.01) in IVTOD (791 g kg–1), and least in NDF (624 g kg–1) from the rotation. Stand declined linearly (P = 0.09) with decreasing FM (Tall = 34.5 and Short = 11.7% basal cover) but was similar between the two rotational systems (35.5%). Continuously stocking EG at about 38 cm gave greatest steer daily gain but rotations may prevent declines in stands with some sacrifice in gains.

Conservation tillage and judicious use of animal manures as fertilizers can make significant contributions for sustainable food production in the twenty-first century. Identifying and understanding the many interactions occurring within agricultural systems is fundamental for accomplishing this feat. This paper synthesizes 14 years of research results from a study that began in the early 1990s in which researchers from USDA-ARS and the University of Georgia investigated cropping system influences on nutrient management under natural rainfall. Increases in C and N with no-till resulted in improved soil structure that increased infiltration rate and soil water availability. Biological activity as indicated by earthworms was greater with no-till and poultry litter (PL). In all but the very driest year, yields of cotton and corn increased on average 10–27% with no-till and 32–42% with combination of no-till and PL. On the other hand soil nutrient accumulation, particularly P and Zn from long-term use of poultry litter in corn production, reached excessive levels and could present environmental risks. Drainage increased in no-till raising the risk of leaching of nutrients into the soil profile. However, runoff decreased in no-till and the presence of a rye cover crop during the winter reduced the leaching losses of N compared to no cover crop. During cotton production under relative drought, no-till and poultry litter led to somewhat elevated dissolved phosphorus concentration in runoff, and fluometuron was detected in runoff and drainage while pendimethalin was not. Fecal indicator bacteria (Escherichia coli and fecal enterococci), and the hormones estradiol and testosterone were observed in drainage and runoff but concentrations were similar across all treatments. By conducting the study for an extended period under natural environmental conditions, we were able to highlight real risks and potentials of the contrasting cropping systems. While 6 out of 14 years of relative drought might have limited the water quality response of treatments, such droughts are common features of the weather pattern in the region. Even then, use of no-till as the predominant tillage system was supported by improved yields. Fertilizer management, especially crop N need-based use of PL, requires closer monitoring to insure that production advantages of no-till and poultry litter are not offset by concerns with environmental risks. Long-term research requires sustained resource inputs to answer critical questions of environmental risk and emerging unknown issues.

A reliable laboratory index ofN availability would be useful for making N recommendations, but no single approach has received broad acceptance across a wide range of soils. We compared several indices over a range of soil conditions to test the possibility of combining indices for predicting potentially mineralizable N (N 0 ). Soils (0-5 and 5-15 cm) from nine tillage studies across the southern USA were used in the evaluations. Long-term incubation data were fit to a first-order exponential equation to determine N 0 , k (mineralization rate), and N 0 * (N 0 estimated with a fixed k equal to 0.054 wk -1 ). Out of 13 indices, five [total C (TC), total N (TN), N mineralized by hot KCI (Hot_N), anaerobic N (Ana_N), and N mineralized in 24 d (Nmin_24)] were strongly correlated to N 0 (r > 0.85) and had linear regressions with r 2 > 0.60. None of the indices were good predictors ofk. Correlations between indices and N 0 * improved compared with N 0 , ranging from r = 0.90 to 0.95. Total N and Hush of CO 2 determined after 3 d (Fl_CO2) produced the best multiple regression for predicting N 0 (R 2 = 0.85) while the best combination for predicting N 0 * (R 2 = 0.94) included TN, Fl_CO2 Cold_N, and NaOH_N. Combining indices appears promising for predicting potentially mineralizable N, and because TN and Fl_CO2 are rapid and simple, this approach could be easily adopted by soil testing laboratories.

Waste handling systems for confined swine production in the upper South (approximately 32-37 degrees N and 79-93 degrees W) depend mainly on anaerobic lagoons and application of the waste effluent to cropland. The main objective of this study was to evaluate the quality of 'Coastal' bermudagrass [Cynodon dactylon (L.) Pers.] hay receiving effluent generated from a raw swine waste treatment system designed to reduce P and K concentrations and delivered by subsurface drip irrigation (SDI) compared with hay produced from commercial N fertilizer. Eight treatments, consisting of commercial N fertilizer or effluent, each irrigated at two irrigation rates (75 and 100% of estimated evapotranspiration) and two lateral spacings (0.6 and 1.2 m), were compared with a control treatment of commercial N fertilizer without irrigation. Three harvests were taken in each of 2 yr and five of the six evaluated using wether sheep (30-45 kg). Greatest dry matter intake (DMI) per unit body weight occurred for the control vs. all irrigated treatments (1.94 vs. 1.77 kg 100(-1) kg; P = 0.02; SEM = 0.11). Among irrigated treatments, DMI was greatest from commercial N vs. effluent (1.81 vs. 1.71 kg 100(-1) kg; P = 0.05; SEM = 0.11). Dry matter intake was similar for the 75% rate treatments and the non-irrigated treatment (mean, 1.87 kg 100(-1) kg) but was reduced for the 100% rate (1.94 vs. 1.72 kg 100(-1) kg; P = 0.03; SEM = 0.11). Hay from the 75% rate was more digestible than hay from the 100% rate (527 vs. 508 g kg(-1); P = 0.03; SEM = 21). The SDI system functioned well, and lateral spacing did not alter hay quality. Treated waste from a raw waste treatment system was readily delivered by SDI at the recommended rate to produce bermudagrass hay of adequate quality for ruminant production systems.

Poultry litter (PL) can supply N, P, K, and other plant nutrients; however, excessive application may cause environmental problems, depending on management and crop nutrient demand. Changes in soil test (ST) nutrient content in a Cecil soil (a fine, kaolinitic, thermic Typic Kanhapludult) during a 10-yr period of PL use was evaluated at the USDA-ARS J. Phil Campbell, Sr., Natural Resource Conservation Center, Watkinsville, GA. During the cotton (Gossypium hirsutum L.) cropping phase (1995-2000), 4.4 Mg PL ha -1 yr -1 resulted in small changes in ST nutrient content in the surface 15 cm. Differences were observed between tillage treatments, with less accumulation of Ca, Mg, and Mn and greater accumulation of Zn for no-till (NT) than conventional tillage (CT). During the corn (Zea mays L.) cropping phase (2001-2005), average annual PL inputs (11.2 Mg ha -1 ) increased P and Zn contents, with changes being similar for CT and NT. After 10 yr, ST nutrient contents in the surface 15 cm reflected 25, 4, 45, 26, 17, and 97% of the input from PL for P K, Ca, Mg, Mn, and Zn, respectively. Changes in soil profile nutrient content (to a depth of 60 cm) from 1997 to 2005 were predominantly at 0 to 15 cm, where P and Zn increased >200%. Accumulation of Ca, K, P, and Zn at lower depths was also observed. Strategies for increasing nutrient removal following repeated long-term application of PL should be considered to avoid excessive levels of nutrients.

'Coastal' and 'Tifton 44' (T44) Bermudagrass [Cynodon dactylon (L.) Pers.] are well adapted across the lower southern United States, but the grazing response of (T44) to N application in the Piedmont of the upper South warrants further evaluation. This 3-yr experiment compared animal and pasture productivity of Coastal and T44 with three annual N rates of 101, 202, and 303 kg of N ha -1 on a Cecil clay loam (fine, kaolinitic thermic Typic Kanhapludult) soil typical of the Piedmont. Herbage mass differed for Coastal and T44 (3.5 and 3.0 Mg ha -1 respectively, P < 0.01), but not among N rates. The canopy of T44 was leafier (20.6 vs. 14.5% of dry matter) than Coastal and greater for in vitro true organic matter disappearance (IVTOD) (522 vs. 498 g kg -1 ) and CP (107 vs. 84 g kg -1 ) and lesser in NDF (596 vs. 605 g kg -1 ). The diet selected from T44 was greater in IVTOD (764 vs. 743 g kg -1 ) and lesser in NDF (596 vs. 605 g kg -1 ) giving greater steer average daily gain (0.63 kg vs. 0.57 kg; P < 0.01) which increased (P = 0.05) with N rate. Weight gain ha -1 (884 kg) and effective feed units (EFU) (4735 kg ha -1 ) were similar, and N rate linearly increased gain from 723 to 1073 kg ha -1 and EFU from 3978 to 5523 kg ha -1 . Soil inorganic N was similar between cultivars but differed among soil depths. Tifton 44 pasture was greater in nutritive value, hence steer performance, and as productive as Coastal in the Piedmont.

Corn (Zea mays L.) producers in the southeastern United States must overcome soil and water limitations to take advantage of the expanding corn market. In this 2001 to 2005 study on a Cecil sandy loam (fine, kaolinitic, thermic Typic Kanhapludult) near Watkinsville, GA, we compared dry land corn biomass and yield under conventional tillage (CT) vs. no-tillage (NT) with ammonium nitrate or sulfate (based on availability) as conventional fertilizer (CF) vs. poultry litter (PL). In a randomized complete block split plot design with three replications, main plots were under tillage and subplots under fertilizer treatments. The cover crop was rye (Secale cereale L.). Over 5 yr, NT and PL increased grain yield by 11 and 18%, respectively, compared with CT and CF. Combined, NT and PL increased grain yield by 31% compared with conventionally tilled and fertilized corn. Similarly, soil water was 18% greater in NT than CT in the 0- to 10-cm depth. In 2 yr of measurements, dry matter of stalks and leaves and leaf area index under PL were an average of 39 and 22% greater, respectively, than under CF during reproduction. Values were 21 and 6% greater, respectively, under NT than CT but during tasseling. Analysis of 70 yr of daily rainfall records showed that supplemental irrigation is needed to meet optimal water requirement. Our results indicate that corn growers can use rainfall more efficiently, reduce yield losses to drought, and expect increased corn yields with a combination of no-tillage management and long-term use of poultry litter.

Tall fescue [ Schedonorus phoenix (Scop.) Holub.] is an important source of forage but often con-tains a fungal endophyte [ Neotyphodium coeno-phialum (Morgan-Jones and Gams) Glenn, Bacon, and Hanlin] that can produce alkaloids and decrease anmi al perf ormance. Removing the endophyte can reduce agronomic performance and persistence. An improved cultivar (‘Jesup’) with a novel (nontoxic) endophyte (trademark “MaxQ”) has improved agronomic performance. Persistence may be related to agronomic traits and animal feeding preferences. We tested for variation in animal preference among two har-vests by testing each of three hays made from Jesup tall fescue that varied in endophyte sta-tus (no endophyte, containing the novel endo-phyte, or containing a wild-type endophyte that produces alkaloids associated with toxicosis). We tested for preference using goats ( Capra hir-cus ), sheep ( Ovis aries ), and cattle ( Bos taurus ). All possible pairs were tested using traditional analysis of variance as well as multidimensional scaling. Multidimensional scaling of the animal preferences indicated that two dimensions were being used to rank the hays. The fi rst dimension for all three animal species was related to varia-tion in forage nutritive value. The second dimen-sion was not correlated with any of the measured variables. Endophyte status was not a factor in the preferences expressed by the three animal species. It appears that no effective feedback mechanism exists related to endophyte status.D.S. Fisher, USDA-ARS, 1420 Experiment Station Rd., Watkinsville, GA 30677; J.C. Burns, USDA-ARS and Dep. Crop Science and Dep. Animal Science, North Carolina State Univ., Raleigh, NC 27695. Cooperative investigation of the USDA-ARS and the North Carolina ARS, Raleigh, NC 27695-7643. The use of trade names does not imply endorsements by USDA-ARS or by the North Carolina ARS of the products named or criticism of similar ones not mentioned. Received 3 Oct. 2007. *Corresponding author (Dwight.Fisher@ars.usda.gov).

Hybrid cultivars of bermudagrass [Cynodon dactylon (L.) Pers.] are a major feed source for ruminants across the southeastern United States. This 4-yr experiment compared animal and pasture performance of 'Coastal' and 'Tifton 44' bermudagrasses over three canopy heights designated as short (5.6 cm), medium (10.1 cm), and tall (13.1 cm). The relationship of canopy height to ingestive mastication and canopy characteristics was also studied. Soil was a Cecil clay loam (clayey, Kaolinitic thermic Typic Hapludult). Pastures were continuously stocked using variable stocking to maintain the targeted canopy heights. Herbage mass (to soil surface) was similar between Coastal and Tifton 44 (T44) within each canopy height averaging 2.36, 4.08, and 5.25 Mg ha -1 . Steer average daily gain (ADG) was greater (P = 0.09) from T44 than Coastal (0.58 vs. 0.51 kg) but no differences were noted in pasture productivity. Increasing herbage mass linearly increased (P < 0.01) ADG (0.40-0.59 kg) but reduced (P < 0.01) stocking rate (16.1-11.2 steers ha -1 ), which influenced animal days (1810-1079 d ha -1 ), weight gain (1057-786 kg ha -1 ), and effective feed units (6392-4452 kg ha -1 ). Steer ADG increased (P = 0.01) from short to medium canopy height (0.40-0.64 kg) with little change between medium and tall canopy height (0.64-0.59kg). Tifton 44 pasture is ofgreater quality than Coastal giving greater ADG but both were productive producing about 1100 kg of gain ha -1 when effectively managed and utilized.

Aims: To better understand and manage the fate and transport of Salmonella in agricultural watersheds, we developed a culture-based, five tube‐four dilution most probable number (MPN) method for enumerating dilute densities of Salmonella in environmental waters. Methods and Results: The MPN method was a combination of a filtration technique for large sample volumes of environmental water, standard selective media for Salmonella and a TaqMan confirmation step. This method has determined the density of Salmonella in 20-l samples of pond inflow and outflow streams as low as 0AE1 MPN l )1 and a low 95% confidence level 0AE015 MPN l )1 . Salmonella densities ranged from not detectable to 0AE55 MPN l )1 for pond inflow samples and from not detectable to 3AE4 MPN l )1 for pond outflow samples. Salmonella densities of pond inflow samples were associated with densities of Escherichia coli and faecal enterococci that indicated stream contamination with faeces and with nondetectable pond outflow densities of the faecal indicator bacteria. The MPN methodology was extended to flux determinations by integrating with volumetric measurements of pond inflow (mean flux of 2AE 5ls )1 ) and outflow (mean flux of 5AE 6ls )1 ). Fluxes of Salmonella ranged from 100 to greater than 10 4 MPN h )1 . Conclusions: This is a culture-based method that can detect small numbers of Salmonella in environmental waters of watersheds containing animal husbandry and wildlife. Significance and Impact of the Study: Applying this method to environmental waters will improve our understanding of the transport and fate of Salmonella in agricultural watersheds, and can be the basis of valuable collections of environmental Salmonella.

Forages accumulate nonstructural carbohydrates during the day, with animals showing preference and improved daily responses from afternoon compared with morning cut hays. This study evaluated alfalfa (Medicago sativa L.) hay harvested at 0700, 1000, 1300, 1600, and 1900 h to determine how nutritive value changes during the day and to assess the impact of these changes on animal preference using cattle (Bos taurus L.), sheep (Ovis aries L.), and goat (Capra hircus L.) responses. Total nonstructural carbohydrates were altered by time of cut (cubic contrast, P < 0.01) ranging from 85 g kg −1 at 0700 h to 83 g kg −1 at 1000 h, then increasing to 97 g kg −1 by 1600 h with little change at 1900 h (96 g kg −1 ). Fiber fractions also varied diurnally, with a quadratic decrease from 418 g kg −1 at 0700 h to

Flue gas desulfurization (FGD) gypsum is a byproduct of coal-fired power plants. Its application to agricultural fields may increase water infiltration, reduce soil erosion, and decrease nutrient losses from applications of animal manures. It may also reduce fecal bacterial contamination of surface waters. We tested the hypothesis that FGD gypsum applications would decrease the load of and the fecal indicator bacterium from poultry litter applications. Two rainfall simulation experiments were undertaken: one in spring 2009 and one in spring 2011. Six treatments consisted of four rates of FGD gypsum (0, 2.2, 4.5, and 9.0 Mg ha) with poultry litter (13.5 Mg ha and two controls) in a randomized, complete-block design with three replications. Each replicate 4- × 6-m plot contained a single 1- × 2-m subplot that was delineated by metal plates and a flume that captured total overland flow or runoff. Rainfall was applied at ∼64 mm h. Volume of overland runoff was measured and subsampled for analysis every 10 min for 1 h. Flow-weighted concentrations, total loads, and soil concentrations of were determined. was not detected in runoff. No significant differences between treatments were observed for the 2009 rainfall simulation. However, after 3 yr of FGD gypsum applications, the highest rate of FGD gypsum resulted in decreased flow-weighted concentrations and total loads of . Flue gas desulfurization gypsum applications may be a management practice that reduces microbial contamination of surface waters from manure applied to agricultural fields in the southeastern United States.

Understanding cover crop and tillage system interactions within specific environments can help maximize productivity and economic returns of cotton (Gossypium hirsutum L.) produced on sandy coastal plain soils of the southeastern USA. A strip-plot design with three replications was used to evaluate the cover crops Austrian winter pea [Pisum sativum L. ssp. arvense (L.)], balansa clover (Trifolium michelianum Savi), crimson clover (Trifolium incarnatum L.), hairy vetch (Vicia villosa Roth subsp. villosa), oil seed radish (Raphanus sativus L.), black oat (Avena strigosa Schreb.), and rye (Secale cereale L.) and tillage (strip and none) influences on cotton grown on a Bonifay fine sand (loamy, siliceous, subactive, thermic Grossarenic PlinthicPaleudults)nearWaynesboro,GAfrom1999to2003.Drought influenced production 3 of 4 yr. Cover crop biomass was greatest from rye, intermediate from black oat, oilseed radish, hairy vetch, and Austrian winter pea. Hairy vetch and Austrian winter pea contained more than 80 kg N ha 21 while other cover crops averaged ,40 kg N ha 21 . Cotton yields following black oat and rye had returns above variable costs ha 21 $461 and $406, respectively. Strip-tillage increased yields by 192 kg ha 21 and annual returns by $112 ha 21 over no-tillage, most likely due to improved available water. Combining strip-tillage with black oat was the best combination for maximizing profit. Using black oat with strip-tillage could increase cotton profit by $50 to $75 ha 21 compared to systems using rye on the 1.45 million ha of cotton where conservation systems have been adopted.

Introducing novel endophytes into tall fescue (Festuca arundinacea Schreb.) that produce no ergot alkaloids could prevent negative impacts on animal performance while improving plant persistence. This 3-yr study evaluated ‘Jesup’ tall fescue (TF) for forage mass, nutritivevalue, and stand persistence when containing no endophyte, a novel endophyte (no ergot alkaloids), or a wild-type endophyte (ergot alkaloids). Forage was accumulated from mid-August and treatments consisted of (i) a grazed control (grazed when growth approximated 10 to 15 cm), or forage accumulated and grazed in (ii) mid-November, (iii) mid-December, (iv) mid-January, and (v) mid-February. Endophyte status had no influence on total forage mass; forage removed by grazing; proportion of leaf, stem, and dead fractions; or on nutritive value (except ergovaline which was greatest in the wild type). Delaying defoliation linearly reduced forage mass, ergovaline concentration, and nutritive value. All stands of TF declined with losses similar (P 5 0.37) for wild-type and novel stands (29 vs. 42%) but were greatest for the TF without an endophyte (29 vs. 75%; P 5 0.01 and 42 vs. 75%; P 5 0.04). These data support the use of novel endophytes in TF for animal production and caution against the use of endophyte-free TF because of decreased stand longevity. The presence of ergovaline can be minimized by stockpiling TF with utilization after late autumn but occurs with a sacrifice in forage mass and nutritive value.

Few studies have reported runoff from small agricultural watersheds over sufficiently long period so that the effect of different cover types on runoff can be examined. We analyzed 45-yrs of monthly and annual rainfall-runoff characteristics of a small (7.8 ha) zero-order typical Southern Piedmont watershed in southeastern United States. Agricultural land use varied as follows: 1. Row cropping (5-yrs); 2. Kudzu (Pueraria lobata; 5-yrs); 3. Grazed kudzu and rescuegrass (Bromus catharticus; 7-yrs); and 4. Grazed bermudagrass and winter annuals (Cynodon dactylon; 28-yrs). Land use and rainfall variability influenced runoff characteristics. Row cropping produced the largest runoff amount, percentage of the rainfall partitioned into runoff, and peak flow rates. Kudzu reduced spring runoff and almost eliminated summer runoff, as did a mixture of kudzu and rescuegrass (KR) compared to row cropping. Peak flow rates were also reduced during the kudzu and KR. Peak flow rates increased under bermudagrass but were lower than during row cropping. A simple process-based ‘tanh’ model modified to take the previous month's rainfall into account produced monthly rainfall and runoff correlations with coefficient of determination (R2) of 0.74. The model was tested on independent data collected during drought. Mean monthly runoff was 1.65 times the observed runoff. Sustained hydrologic monitoring is essential to understanding long-term rainfall-runoff relationships in agricultural watersheds.

Soil water influences hydrological, biological, and biogeochemical processes that determine on- and off-site response of landscapes under different agricultural uses. There are relatively little detailed spatial and temporal soil water measurements to validate current representations of spatial and temporal soil water variability. Soil water was measured over 3 yr at 12 sites to a 1.2-m depth in a 7.8-ha pasture catchment in the Georgia Piedmont in southeastern USA. The Mahalanobis statistical difference was estimated between all pairs of measurement sites for soil water. Multidimensional scaling of the Mahalanobis differences showed that only a single statistical dimension was separating the observed soil water variations at measurement sites (r 5 0.99). This statistical dimension was then found to be most closely correlated with the depth to the Bt and the depth of the Ap horizons for explaining the observed variation in soil water between sampling sites (r 5 0.69 and r 5 0.8, respectively). Those sites where the Bt, an argillic horizon, was close to the surface, even when higher up the landscape, were generally wetter than those in which the Bt was deeper, even in the lower part of the landscape. The depth to the Bt horizon may serve as an indicator of the portions of the watershed most likely to be primary sources of runoff in association with the depth of overlying coarse-texturedsoil.Volumetricsoilwatercontentwasgenerallygreatest

Calibration procedures and data used to parameterize a model, including model components that may or may not have been addressed, are generally not well documented in modeling studies. A comprehensive description of the process and parameters used for calibrating the Root Zone Water Quality Model, v. 1.3.2004.213, is presented in this article. The model was calibrated to simulate tile drainage and leached nitrate under conventional tillage management practices for maize (Zea mays L.) production followed by a rye (Secale cereale L.) cover crop in Cecil soils (kaolinitic, thermic, Typic Kanhapludults), and for cotton (Gossypium hirsutum L.) development in the Georgia Piedmont. Tile drainage and nitrate leaching were simulated within 15% of the observed values in the calibrated maize scenarios with and without the soil macroporosity option. Simulated and observed tile drainage and leached nitrate were not significantly different, and the simulated values were not significantly different with and without the macroporosity option. Simulated cotton biomass and leaf area index were well correlated with observed biomass and leaf area index until the last 21 d of the reproductive stage. Simulated and observed cotton water use were different by

test of the ruminant's ability to learn that the nutritive value or taste of one feed varies slightly from another As a result of photosynthesis, plants typically have greater concen- at two levels of nitrogen fertility in 1999 to produce eight hays; and three times in 2000 to produce six Alamo hays. The hays were field- Alamo was produced at two levels of nitrogen fertility. dried, baled, and passed through a hydraulic bale processor. Hays Thisselectionofhaysproducedoverthreeyearsallowed from each year were tested with cattle (Bos taurus L.), goats (Capra forawiderangeofsignificantvariationinforagecompo- hircus L.), and sheep (Ovis aries L.). During an adaptation phase, sition (or nutritive value) and environmental conditions hays were offered individually. In the experimental phase, all possible within which to test for the impact of PM and AM hay pairsofhayswerepresented.Datawereanalyzedbymultidimensional mowing in the humid southeastern USA. scalingand bytraditional analyses.Multidimensional scalingindicated that selection was based on multiple criteria. The suite of improve- ments associated with PM-harvested hays in fiber content, digestibil- MATERIALS AND METHODS ity, and nonstructural carbohydrate observed for alfalfa and fescue Field Procedures hays in the western USA was difficult to reproduce with switchgrass hay in the southeastern USA. This difficulty is likely related to the All hays for the experiments reported were harvested from less favorable environment for haymaking as well as the physiology, well established stands of switchgrass grown on a Cecil clay anatomy, and morphology of this C4 grass. loam (fine, kaolinitic, thermic Typic Kanhapludults) soil near Raleigh, NC. The hay cuts were paired at each harvest so that each cut in the evening (PM) was followed by another cut the next morning (AM) at sunup. P

Forage yields are influenced by plant response to defoliation. We examined the photosynthesis and nutritive value of first (first leaves) and third (third leaves) fully expanded leaves (numbered from the apex) in three warm-season (C4) grasses. Net photosynthetic rates at uniform temperature and light both before and after a 2-wk exposure to full sunlight and the effect of leaf position on nutritive value were determined on vegetative tillers in well-established swards of bermudagrass [Cynodon dactylon L. (Per.) cv. Tifton 44], Caucasian bluestem [Bothriochloa caucasica (Trin.) C.E. Hubb.], and Atlantic coastal panicgrass [Panicum amarum var. amarulum (Hitchcock and Chase) P.G. Palmer] growing on a Cedl clay loam (clayey, kaolinitic, thermic Typic Hapludults). Bermudagrass had the greatest level of crude protein (CP), followed by panicgrass and bluestem. Fiber was greater in the first leaves than in the third leaves for bermudagrass and panicgrass but not for bluestem. Photosynthetic rates of panicgrass and bluestem third leaves estimated 2 wk after defoliation of the surrounding canopy were less than estimates made before defoliation in the first leaves but were similar to the third leaves before canopy defoliation. The third leaves of bermudagrass 2 wk after defoliation had lesser photosynthetic rates per unit chlorophyll than the first or third leaves before defoliation. Photosynthetic rates were correlated with hemicellulose across leaf classes and species (r 2 = 0.93). The photosynthetic decline observed in third leaves of bermudagrass compared with panicgrass and bluestem is evidence of variation in leaf response after defoliation among warm-season grasses.

The objective of this research was to examine the effect of high concentrations of nonprotein nitrogen (NPN) on the voluntary food intake of sheep fed high-quality grasses. Wether lambs (n = 6 per treatment) were fed dried switchgrass (Panicum uirgatum L.; Exp. 1) or dried tall fescue (Festuca arundinacea Schreb.; Exp. 2). In both experiments, urea was added to the dried forage at 0 (control), 12, or 24 g of N/kg of DM to increase the NPN concentration. Acid detergent fiber concentrations were 305 g/kg of DM in both experiments, although DM digestibility was 663 and 618 g/ kg of DM in Exp. 1 and Exp. 2, respectively. Voluntary feed intake of the control forage was 28.2 and 19.1 g/ kg of BW in Exp. 1 and Exp. 2, respectively, and decreased for the high-urea treatments to 25.2 and 16.2 g/kg of BW in Exp. 1 (P = 0.07) and Exp 2 (P = 0.03), respectively. Total feed N concentrations increased from 29.5 g to 45.7 g of N/kg of DM in Exp. 1 (P < 0.01) and from 28.4 to 55.9 g of N/kg of DM in Exp. 2 (P < 0.01). Nonprotein N concentrations increased from 28. 3 to 53.8% of the total N in switchgrass diets (Exp. 1; P < 0.01), and from 26.4 to 64.0% in tall fescue diets (Exp. 2; P < 0.01). Plasma urea concentrations of the lambs increased from 3.1 to 6.6 mM (Exp. 1; P < 0.01) and from 2.9 to 5.8 mM (Exp. 2; P < 0.01) as the amount of urea added to the diets increased. These changes resulted in an increase in plasma osmolality from 298 to 307 mOsm/kg (Exp. 1; P = 0.04), and from 299 to 307 mOsm/kg (Exp. 2; P = 0.06). Increasing feed N and NPN concentrations through the addition of urea caused a significant decrease in the voluntary feed intake of sheep fed tall fescue and switchgrass. These responses showed no significant cause-and-effect relationship between voluntary feed intake, plasma urea concentrations, and plasma osmolality.

During the history of ruminant nutrition many factors have been proposed as regulators of voluntary feed intake. In some cases, the implicit assumption has been that a factor acted independently and exclusively of other mechanisms. A knowledge of ruminant digestive anatomy aids in understanding both the ecological niche and the feeding behavior of the ruminant animal. Factors controlling ruminant intake should be assumed to function with multiple interactions. A number of feedback regulators such as distension, protein, and energy should be considered in the context of their interacting regulatory effects when attempting to predict intake. Behavioral aspects also influence voluntary feed intake through associations formed via post-ingestive feedback. Ruminants can learn to identify particular feeds and alter intake on the basis of past experiences. An integrated approach is proposed as a means of understanding ruminant feed intake regulation and eventually to improving prediction of intake. Empirical mathematical methods are likely to be fundamental to developing understanding and models of feed intake because of the difficulty of studying the central nervous system. In spite of these difficulties, knowledge of theoretical feedbacks has already been used to develop practical ruminant feeding strategies.

Plant growth and yield responses to carbon dioxide (CO2) enrichment are well es- tablished. Much less is known of the response of arthropod pests to CO2 enrichment. Reproductive response of twospotted spider mites (Tetranychus urticae Koch) on white clover (Trifolium repens L.) to a range of CO2 concentrations was measured. The CO2 treatments were applied for 24 h d 1 at395, 484, 570, 657, and 748 LL 1 on the 14 d before and 26 Ð27 d after infestation with mites. Eggs, larvae, nymphs, and adult mites were removed from leaves and counted 27Ð29 d after infestation. Leaf area and weight were measured, and leaves were analyzed to measure structural and nonstructural carbohydrates, N, amino acids and digestibility. Carbon dioxide enrichment caused linear increases in plant growth and foliar nonstructural carbohydrates, but caused linear decreases in foliar N. Carbon dioxide enrichment signiÞcantly increased the rate of mite reproduction on both clover clones. Correlations between mite population increase were signiÞcantly positive for foliar nonstructural carbohydrates and signiÞcantly negative for foliar N. Concentrations of ambient CO2 expected in the 21st century may increase the risk of mite population damage on some plant species.

Diurnal variation in the concentration of total nonstructural carbohydrates (TNC) occurs in plants as a result of photosynthesis. Ruminants have been shown to prefer tall fescue (Festuca arundinacea Schreber) hays cut in the afternoon but the effect of morning vs. evening cutting had not been tested in legumes. To test for diurnal variation in preference for alfalfa (Medicago sativa L.), we harvested six times in the midbud stage. Harvests were paired so that each time a cutting of alfalfa was made at sundown (PM) another was made the next morning at sunup (AM). We harvested in this manner three times resulting in six hays. The hays were field dried, baled, and chopped prior to their use 3 to 6 mo after harvest. Three experiments were conducted [Exp. 1, sheep (Ovis aries); Exp. 2, goats (Capra hircus hircus); and Exp. 3, cattle (Bos taurus)] utilizing six animals in each case. During an adaptation phase, hays were offered alone as meals. In the experimental phase, every possible pair of hays (15 pairs) was presented for a meal. Data were analyzed by multidimensional scaling as well as by traditional analyses. Multidimensional scaling indicated that the animals were basing selection on at least two criteria. Variables associated with preference through multiple regression varied across experiments but significant coefficients were found between preference and nitrate, protein, carbohydrate fractions, lignin, and cellulose. Coefficients varied depending on which other variables were in the model; however, carbohydrates were associated with positive coefficients. Shifting hay mowing from early in the day to late in the day was effective in increasing forage preference as expressed by short-term dry matter intake.

On a watershed scale, geospatial information can be used to identify water resources that are least buffered from contamination. Implementing conservation practices at these locations may accelerate the process of increasing a watershed's ability to support its designated uses. The Upper Oconee Watershed of Georgia contains land areas devoted to poultry, dairy, and beef production. Within these historically agricultural lands, urbanization is proceeding rapidly around existing cities. Agricultural production practices are concentrated in the watershed with poultry in the headwaters area and dairy near a major lake (Lake Oconee). The objective of this research was to relate data sets representing surface water quality at selected sites throughout the watershed to the predominant land use in that portion of the watershed. The location of 550 poultry operations in the headwaters of the Upper Oconee Watershed, away from the city of Athens GA, has minimized conflicts between agricultural and urban interests. Phosphorus, nitrogen, and fecal coliform bacteria were high near the poultry production area, but were reduced within the watershed prior to reaching the intake for the municipal water supply. Athens had a large impact on surface water quality and approximately doubled the amount of phosphorus and nitrogen in the Oconee River. The Oconee River contributed approximately 70% of the water flowing to Lake Oconee. The residents of Lake Oconee have noted the 30 dairies located west of the lake impacting two relatively minor creeks flowing to the lake. These two creeks make up approximately 2.5% of the flow to the lake, but the proximity of the dairies to the lake makes losses of phosphorus, nitrogen, and fecal bacteria apparent in water samples. Fecal coliform numbers were elevated in some creeks with little agricultural or urban development. To test alternative microbial assays, surface water from a grazed watershed was compared to water from a wooded watershed. Assays for enterococci and E. coli may provide a better test for fecal contamination and allow differentiation between natural areas and areas impacted by grazing animals. Analysis of the Upper Oconee Watershed identified agricultural impacts and areas that should be priorities for natural resource management to reduce agricultural non-point source pollution. Focusing conservation efforts at these locations may prevent agricultural-urban conflict. However, the data also indicate that municipal sources of nutrients and fecal bacteria must be reduced to make significant progress in the watershed.

Calliandra calothyrsus, Paraserianthes falcataria and Gliricidia sepium leaves were supplemented at both 25 and 50% of dietary crude protein to a basal concentrate and fed to growing ram lambs. A control diet consisted of concentrate and the forage grass Brachiaria brizantha. Dry matter intakes and ADG of tree legume supplemented treatments were lower (P < 0.05) than control lambs, but increasing the amount of tree leaves fed did not result in further decreases in either dry matter intake or ADG. Dry matter digestibility was lower (P < 0.05) in both 25 and 50% dietary treatments compared with the control, while NDF digestibility was lower (P < 0.001) only in the 50% dietary treatments. Increasing the amount of leaves fed from all three tree species, led to decreased NDF digestibility. Fecal NDF-N was higher in tree-supplemented groups (P < 0.01) compared with control and increased (P < 0.05) with increasing amounts of tree leaves fed. Both apparent and true N digestibility were lower (P < 0.05) in C. calothyrsus, than in the other tree legume treatments. True N digestibility of 91% for the control was higher (P < 0.01) than the tree legume treatments supplemented at both 25 (range, 83‐86%) and 50% dietary CP (range, 74‐79%). Increasing the amount of tree leaves fed, regardless of tree species, decreased N digestibility in the diets. Higher fecal NDF-N and lower N digestibility in tree leaf supplemented lambs indicated that condensed tannins present in these tree species were binding protein and rendering it less available for digestion. Supplementation with C. calothyrsus, which had the highest levels of SPHE and

'Carostan' flaccidgrass (Pennisetum flaccidum Griseb.), a C 4 perennial grass, has shown high quality potential in animal trials. Its yield response to defoliation frequency and N fertilization, however, has not been reported. We examined the influence of stubble height and of a range of N applications in two defoliation experiments over 4 yr. Experiment 1 (Raleigh, NC) combined two harvest heights, 76 and 38 cm, with two stubble heights, 25 and 15 cm (76-25, 76-15, 38-25, 38-15); all treatments received 380 kg N ha -1 yr -1 . After Year 1, the 76-15 defoliation produced higher yields than 76-25 (Year 2: 11 330 vs. 7930 kg ha -1 , P < 0.001; Year 3: 8950 vs. 6270 g kg -1 , P < 0.001), while in vitro dry matter disappearance (IVDMD) was similar (Year 3) (mean = 582 g kg -1 ). Compared with 38-15, the 38-8 defoliation resulted in lower yields (Year 2: 8638 vs. 9344 kg ha -1 , P < 0.001; Year 3, 6173 vs. 8090 kg ha -1 , P = 0.07) but higher IVDMD (Year 3, 651 vs. 635 g kg -1 , P < 0.01). Experiment 2 (Raleigh, NC, and Watkinsville, GA) had four defoliations: clipping to 5 cm at 20 and 51 cm (vegetative) and to 8 cm at boot and anthesis stages (20-5, 51-5, Boot-8, Anthesis-8). All defoliation treatments were topdressed with 291 kg N ha -1 , and 51-5 and Boot-8 were also evaluated with 179 and 403 kg N ha -1 (eight N-defoliation treatments in all). At both locations, highest yields were obtained with Anthesis-8, averaging 19 010 kg ha -1 at Raleigh and 12 390 kg ha -1 at Watkinsville. The 20-5 defoliation resulted in lowest yields at both locations, averaging 7648 kg ha -1 at Raleigh and 5100 kg ha -1 at Watkinsville. Defoliation frequency altered IVDMD (range = 550 to 789 g kg -1 ), but N application did not; however, increasing N application increased N concentrations linearly (P = 0.05). Flaccidgrass had high yield potential and high nutritive value at the Raleigh location, and warrants further evaluation in the upper South.