Grazing livestock operations in the southern United States

Published January 18, 2018 RESEARCH Overseeding Cool-Season Annual Legumes and Grasses into Dormant Tifton 85 Bermudagrass for Forage and Biomass Joshua A. White, James P. Muir,* and Barry D. Lambert ABSTRACT Overseeding dormant bermudagrass [Cynodon dactylon (L.) Pers.] Tifton 85 with cool-season annuals in the southeastern United States could provide winter and spring forage or bioenergy feedstock. We overseeded cool-season annual legumes and grasses in monocultures or mixtures into dormant Tifton 85 and compared aboveground dry matter (DM) and N yields of single harvests at peak biomass production and a multiple harvest forage system. We also measured first-cut spring Tifton 85 yields. Wheat (Triticum aestivum L.) yielded as much or more (P 0.05) biomass in a single-harvest system than other grasses only during a dry cool season. Well-distributed precipitation, in combination with multiple cool-season grass legume harvests, resulted in lower yields than the single harvest system, whereas the reverse occurred when precipitation was erratic. Hairy vetch (Vicia villosa Roth) matured later in the spring, and monocultures or mixes with it yielded more (P 0.05) DM and N than the more precocious crimson clover (Trifolium incarnatum L.). A drier cool season followed by a spring with good precipitation, as well as repeated cool-season grass and legume harvests, resulted in more (P 0.05) spring first-cut Tifton 85 DM regrowth. Results provide several management options to maximize cool-season, annual legume, and grass bioenergy feedstock or forage while minimizing negative and maximizing positive effects on subsequent early spring Tifton 85 regrowth. J.A. White, Dep. of Plant & Soil Sciences, Mississippi State Univ., Mississippi State, MS; J.P. Muir, Texas A&M AgriLife Research, Soils & Crop Science, 1229 North US Hwy 281, Stephenville, TX 76401; B.D. Lambert, Tarleton State Univ., Stephenville, TX 76401. Received 5 Sept. 2017. Accepted 4 Dec. 2017. *Corresponding author (j-muir@ tamu.edu). Assigned to Associate Editor Jose Dubeux. Abbreviations: DM, dry matter. Grazing livestock operations in the southern United States plant cool-season annual grasses in pure stands to harvest as green chop during the autumn and spring months. Replacing or interseeding these grasses with legumes may help limit the need for N fertilizer application. Under irrigation, dairymen or future bioenergy feedstock producers may also choose to use dormant bermudagrass [Cynodon dactylon (L.) Pers.] sod as a seedbed for their cool-season forages (Muir and Bow, 2011). Overseeding bermudagrass sod with annual cool-season grasses can improve yearly forage production per unit area (Fribourg and Overton, 1973). The main factors contributing to success when overseeding sod grasses with annual cool seasons in the autumn are timing of seeding, precipitation, and temperature (Robinson, 1963). When considering this approach, however, the effect on the bermudagrass stand the following spring growing season must be considered (Muir and Bow, 2011). Tifton 85 is particularly valued for its productivity and nutritive value (Burton, 2001) and could be a valuable bioenergy feedstock (Muir et al., 2010), so yield losses from competition with late spring cool-season annuals could be detrimental to overall production. However, Fribourg and Overton (1973) showed that harvesting annual cool-season grass several times depressed the following season s bermudagrass harvest less than when harvest numbers were reduced. By contrast, overseeding legumes such as arrowleaf (Trifolium vesiculosum Savi) Published in Crop Sci. 58:964 971 (2018). doi: 10.2135/cropsci2017.09.0530 Crop Science Society of America 5585 Guilford Rd., Madison, WI 53711 USA This is an open access article distributed under the CC BY license (https:// creativecommons.org/licenses/by/4.0/). 964 www.crops.org crop science, vol. 58, march april 2018

or crimson clover (Trifolium incarnatum L.) can increase total yearly pasture forage yield from 17 to 22%, respectively, and improve bermudagrass yields by 18% vs. grass without clover treated with 224 kg N ha 1 (Knight, 1970). In addition to green chop or haylage, cool-season forages planted into bermudagrass sod could potentially contribute cellulosic bioenergy and provide an additional option for removing excess dairy manure nutrients from application fields (Muir and Bow, 2011). Biofuel research is gaining momentum, mostly focused on using warmseason (C 4 ) forages such as switchgrass (Panicum virgatum L.; Cassida et al., 2005), with minimal research using cool-season annual plants for eventual biofuel production. Cool-season legumes overseeded on dormant warmseason perennial grasses can also provide N to subsequent warm-season haylage or bioenergy feedstock production systems (Bow et al., 2008), thereby reducing N fertilizer requirements in subsequent bermudagrass crops or by recovering N during the conversion process. However, cool-season grasses or legumes harvested at peak biomass production in the late spring might affect subsequent bermudagrass yield differently when compared with multiple earlier forage harvests that compromise between yield and nutritive value. Most C 3 forages are not typically considered for lignocellulosic biofuel production in warmer climates because of their relatively low efficiency in sequestering C when compared with perennial C 4 grasses under heat and moisture stress (Sánchez et al., 2015). However, in areas where cereal straws are an industry byproduct or annual C 3 grasses grow in the cool season when C 4 grasses cannot, the first may still be a viable biofuel option. Bow et al. (2008) expressed concerns about competition from warm-season perennial grasses to establishing fall-seeded legumes overseeded on those pastures; they also pointed out the potential for legume competition to spring perennial warm-season grass regrowth. They concluded that overseeding cool-season annual legumes did not reduce switchgrass dry matter (DM) yields. Hecht et al. (2009) suggested that the biofuel industry is more likely to succeed if it designs biomass feedstock production systems that fit seamlessly into already established cropping practice. Bermudagrass is already widely grown in the southeastern United States and other subtropical regions and has the potential to produce bioenergy feedstock (Muir et al., 2010). Our objectives were to compare yield differences of cool-season annual grasses and legumes in pure stands or grass-legume mixtures overseeded onto dormant Tifton 85 for biomass and N yield. We also examined the effect of this overseeding on subsequent spring Tifton 85 regrowth. Large differences in rainfall pattern over the 2-yr study allowed us to make these comparisons under very different precipitation patterns. MATERIALS AND METHODS The Tifton 85 bermudagrass field used for this trial was located at the Texas A&M AgriLife Research Center in Stephenville, TX (32 15 N, 98 12 W, 395 m asl) on a Windthorst fine sandy loam (fine, mixed, thermic, Udic Paleustaf; Soil Survey Staff, 1973). The field was situated under an irrigation pivot, and plots encompassed two sections of the pivot to take into account potential soil drainage differences between sections. The trial was initiated after the last Tifton 85 hay cutting was removed in late autumn of 2009 and 2010. Before planting, P was applied at 13 kg ha 1 per soil test recommendations from the Texas A&M University Soil Testing Laboratory (College Station, TX) that showed an average 15 ml P L 1, 111 ml K L 1, 1678 ml Ca L 1, and 302 ml Mg L 1 (Mehlich, 1984) and a ph of 7. Prior to planting, the Tifton 85 field was lightly disked (2.5- to 5.0-cm depth). The main treatment consisted of overseeding cool-season annual grasses the first week of November 2009 and 2010 and a control with no overseeding. Cool-season annual grasses included TAM 401 wheat (Triticum aestivum L.), TAMbar 501 barley (Hordeum vulgare L.), Matton II rye (Secale cereale L.), TAMO 406 oats (Avena sativa L.), TAMcale 6331 triticale ( Tritosecale Wittm. ex A. Camus), and TAMTBO annual ryegrass [Lolium perenne L. ssp. multiflorum (Lam.) Husnot]. Grass seeding rates in pure stands were 112 kg ha 1 for small grains and 35 kg ha 1 for annual ryegrass. Grass plots were either coseeded with hairy vetch and crimson clover or planted in pure grass plots fertilized with N at 0 or 134 kg ha 1 split into equal applications in late November and cool-season grass biomass harvest in the following spring. Hairy vetch and crimson clover were also seeded into control plots containing no cool-season annual grasses. Hairy vetch in pure stands was seeded at 28 kg ha 1 and the clover at 22 kg ha 1. Mixtures were seeded at 80% of the recommended rate for grasses and 66% of the recommended rate of the legumes. A Hege 1000 no-till seeder was used for planting. Plot size was 5 2 m with six rows spaced 30 cm apart. There were four replications arranged in blocks to account for possible soil drainage variation. Plots were planted in a strip-strip plot design. The main strip consisted of cool-season grasses or a no-grass control. Substrips consisted of legume or fertilizer treatments. Harvest timing was determined by the growth and development of each individual grass species. Half of each plot was harvested at 10 cm above the soil once each season for biomass (full bloom or early seed set), whereas the other half was harvested to emulate green chop for forage. Plots representing forage were harvested multiple times, every time 75% of plot growth reached 40-cm height or boot stage (whichever came first) to best emulate green chop systems. Forage plots were no longer harvested when Tifton 85 regrowth height superseded that of the cool-season forage. Grass biomass plots were harvested when the small grains reached soft dough stage or after ryegrass had begun to set seed. Monoculture legume biomass plots were harvested when the greatest plant mass accumulated at full bloom. Mixed species biomass plots were harvested at maximum accumulated DM of each grass regardless of accompanying legume. Emerging Tifton 85 herbage that was part of the sample was included in the DM and N yields. Plots received 30 mm of irrigation immediately after planting to guarantee germination and stand establishment, but crop science, vol. 58, march april 2018 www.crops.org 965

no additional irrigation thereafter. Rainfall between September 2009 and May 2010 totaled 453 mm, whereas between September 2010 and May 2011, it totaled 289 mm with only 42 mm falling from November to March that second cool season (Fig. 1). An Almaco small plot harvester (Allan Machine Company) was used to harvest and determine plot green weight. The width of the Almaco harvested the four middle rows after leaving 1 m at both ends to exclude any border effect of each plot. In plots with a mixture of legumes and grass, the yields were totals of the two. A 200-g subsample was weighed and dried in a forcedair oven at 55 C until weight loss ceased. It was weighed again, and the resulting DM percentage was used to correct plot green weight to a DM basis. This was extrapolated to yield on a perhectare basis. These subsamples (whole plant ground through a 1-mm screen) were analyzed in the laboratory for N and again corrected for DM by drying at 105 C. Nitrogen concentrations were analyzed using an Elementar Vario Macro C-N Analyzer. Dry matter yields were multiplied by N concentration to determine N yields from each subplot. Two Tifton 85 plots in every replication were not overseeded with cool-season species. One plot was fertilized with 112 kg N ha 1 in late April, and the other received no N application. Thirty days after fertilizer was applied or when Tifton 85 in N-treated plots reached 40-cm height (whichever came first), aboveground herbage in all plots was harvested at 5 cm above the soil to determine DM yields and compare the effect of overseeding cool-season forage on the first spring Tifton 85 hay harvest yields. Results were reported as the percentage of control Tifton 85 plots with or without N fertilizer. Analysis of variance (Proc GLM in SAS Institute, 2013) tested year (2009 2010 and 2010 2011), harvest regimen (forage or biomass), cool-season grass (six entries), and legume-n fertilizer (crimson clover, hairy vetch, N fertilizer, or control) as fixed effects on cool-season grass, legume, or mixture DM and N yields. A further ANOVA tested year, harvest regimen, cool-season grass, and legume-n fertilizer effects on Tifton 85 DM percentage change vis-à-vis spring Tifton 85 regrowth receiving 0 or 112 kg N ha 1. Values were considered significant at P 0.05, and multiple means, where appropriate, were separated using Fishers LSD. RESULTS AND DISCUSSION Year interacted with all other factors for all dependent variables. All results are therefore presented by year. In addition, no interaction between the fertilizer-legume treatment and the grass species affected any dependent variable; therefore, grass species and fertilizer-legume are discussed as main effects relative to each other. The following discussion considers the biomass and N production under a forage (multiple) and bioenergy (single) harvest of grass species pooled across fertilizer treatments. Separately, the biomass and N production under a forage and bioenergy harvest of fertilizer treatments are pooled across grass species and discussed. The subsequent effect of fertilizer treatments and grass species treatments on the Tifton 85 are also discussed and analyzed in relation to both a positive and a control plot that were not overseeded. Fig. 1. Monthly precipitation at Stephenville, TX, over the experiment period, as well as the 30-yr average. Dry Matter Yield: Cool-Season Grasses There was a harvest regimen grass species interaction for DM yield both years (Table 1). There were no differences among species either season within the forage harvest regimen. There were, however, differences among species for the biomass harvest in 2010 2011, with wheat and barley yields superior to oats and ryegrass, indicating greater drought tolerance for the first two entries. Wheat, for example, yielded 140% more DM than ryegrass and 127% more than oats. Barley, rye, and wheat DM in the biomass regimen was 83, 159, and 148% greater, respectively, compared with the forage regimen in 2009 2010 (Table 1). Under ideal soil moisture conditions, this is not unusual, as exemplified by Guney et al. (2016) who reported 2860 kg DM ha 1 for barley at heading stage compared with 4720 kg DM ha 1 at seed formation and 7650 kg DM ha 1 (167% greater yield visà-vis heading) at full maturity. The forage regimen resulted in greater barley, oat, rye, and ryegrass yields in 2010 2011, 251% in the case of oats. This indicated that rainfall patterns favored the single biomass harvest in 2009 2010, when precipitation was more evenly distributed throughout the season (Fig. 1). By contrast, rainfall from November 2010 to March 2011, the second cool season, was negligible whereas the total for the season was only 64% that of the first. The 5 mo without effective rainfall apparently favored the forage regimen in 2010 2011 in four of the six species. As a result, differences for the forage regimen between high- and lowrainfall seasons were not as pronounced, whereas the mean DM yield for biomass regimen the drier season was only 32% that of the higher rainfall season. Nitrogen Yield: Grass Species Herbage N concentration averaged across both years ranged from 16 to 19 g kg 1 for material collected from plots containing grass. Plots that contained barley, oats, rye, ryegrass, triticale, and wheat produced forage N concentrations of 19, 19, 16, 17, 19, and 18 g kg 1, respectively. There were no differences in N yield among the grass entries either 966 www.crops.org crop science, vol. 58, march april 2018

Table 1. Cool-season grass dry matter (DM) and N yields when overseeded into a dormant Tifton 85 pasture and harvested as forage or bioenergy feedstock (averaged over N fertilizer or legume treatments). DM yield N yield Year & cultivar Forage Biomass LSD 0.05 Forage Biomass LSD 0.05 k gh a 1 yr 1 k gh a 1 yr 1 2009 2010 Barley 2067 3780 S 25 67 S Oat 2104 3086 NS 26 50 S Rye 1559 4037 S 32 60 S Ryegrass 2407 3224 NS 26 51 S Triticale 1849 2631 NS 25 47 S Wheat 2166 3198 S 23 46 S Mean 2025 3326 26 54 LSD 0.05 NS NS NS 15 2010 2011 Barley 2334 1336 S 18 20 NS Oat 2382 679 S 19 9 S Rye 2255 1145 S 33 14 S Ryegrass 2038 642 S 26 12 NS Triticale 1895 996 NS 16 14 NS Wheat 1778 1541 NS 13 25 S Mean 2114 1056 21 16 LSD 0.05 NS 537 NS 9 NS, not significantly different, P > 0.05; S, different at P 0.05. year under the forage harvest regimen (Table 1). Differences were apparent both years, however, when harvested a single time as biomass at peak accumulation. Barley yielded more N than all entries except rye the 2010 2011 season, 46% more N compared with wheat. However, barley yielded only 30% the N during 2010 2011 that it did the previous year, which was still 122% more productive than oats, indicating greater oat drought adaptation. The single-harvest biomass regimen yielded greater N compared with the forage harvest regimen for all species in 2009 2010 and for wheat in 2010 2011 (Table 1). In the case of barley, in 2009 2010 with greater rainfall, biomass yielded 42 kg more (168% greater) N than the forage harvest. Biomass harvests yielded less N for oats and rye in 2009 2010 than forage harvests, indicating that growth was more responsive to rainfall in the short term than as a cumulative effect over an entire dry season. Dry Matter Yield: Legumes or Fertilizer There was a harvest regimen legume-n fertilizer interaction for DM yield both years (Table 2). Except for N fertilizer in 2009 2010, grass-hairy vetch mixtures yielded greater DM than other treatments when averaged over all grass entries in the forage harvest both years. Other studies reported greater hairy vetch forage yields when overseeded as a monoculture into dormant bermudagrass compared with similar annual cool-season legumes, except during dry spring conditions (Muir and Bow, 2011; Freeman et al., 2016). In our study, plots containing hairy vetch had forage yields 40 and 290% greater than crimson clover and the control, respectively, in 2009 2010. Differences were even greater in 2010 2011, the lower rainfall year: hairy vetch yielded 100 and 451% greater DM, respectively, than crimson clover and the control. Muir and Bow (2011) reported fewer differences among cool-season legume monocultures during a drier year compared with a wetter year. Hairy vetch developed later in the season, flowering in April and May (154 mm total precipitation in 2009 2010), whereas crimson clover was more precocious and flowered in February and March, except in 2010 2011 when precipitation totaled only 11 mm from November through March. Less precipitation may explain the lower DM yields in our study compared with the 4315 kg DM h ha 1 reported for a site with greater rainfall (Freeman et al., 2016). When harvested as biomass, only the control plots consistently yielded less than any legume or N-fertilized treatment in both years (Table 2). Hairy vetch and crimson clover yielded the same DM in both years under the forage harvest regimen. By contrast, under the biomass harvest regime, all yields were consistently lower in 2010 2011 compared with the previous, greater precipitation year. Dry matter yields were consistently greater for all treatments in 2009 2010 when harvested under the biomass regimen (Table 2). The biomass harvest regimen resulted in greater DM yields across all treatments in 2009 2010, >2500 kg ha 1 yr 1 more in the case of N fertilizer. During 2010 2011, the reverse occurred for all treatments except crimson clover; hairy vetch yielded >1600 kg ha 1 yr 1 more as forage than as biomass. Nitrogen Yields: Legumes or Fertilizer Forage N concentration across both years ranged from 14 to 26 g kg 1 in N-fertilized plots. Legume treatments that contained hairy vetch and crimson clover averaged 25 and crop science, vol. 58, march april 2018 www.crops.org 967

Table 2. Cool-season grass dry matter (DM) and N yields when overseeded into a dormant Tifton 85 pasture and harvested as forage or bioenergy feedstock (averaged over N fertilizer or legume treatments). DM yield N yield Year & cultivar Forage Biomass LSD 0.05 Forage Biomass LSD 0.05 k gh a 1 yr 1 k gh a 1 yr 1 2009 2010 Hairy vetch 2584 4277 S 81 76 NS Crimson clover 1847 3738 S 37 47 S N fertilizer 2298 4860 S 47 66 S Control 663 1902 S 11 20 NS Mean 1848 3694 44 52 LSD 0.05 298 619 8 10 2010 2011 Hairy vetch 3003 1314 S 72 24 S Crimson clover 1502 1143 NS 34 16 S N fertilizer 1890 1317 S 36 18 S Control 545 282 S 10 4 S Mean 1735 1014 38 15 LSD 0.05 1320 542 19 9 NS, not significantly different, P > 0.05; S, different at P 0.05. 17 g N kg 1, respectively, across harvest regimens. Likewise, grass receiving N fertilizer and no N averaged 18 and 14 g N kg 1, respectively. In general, plots harvested for forage across all treatments averaged 23 g N kg 1, whereas plots delayed until peak biomass averaged 15 g N kg 1. As a result of greater DM yields and N concentrations, plots with hairy vetch yielded more N than those with crimson or N fertilizer in both years when harvested as forage (Table 2), 636 and 620%, respectively, compared with the control plots. Under the biomass harvest regimen, plots with hairy vetch also yielded more N than crimson clover in 2010 2011 and more than the control both years, 54 kg N ha 1 yr 1 in the case of the control in greater rainfall 2009 2010. Compared with the biomass regimen, the forage harvest regimen resulted in greater N yields in 2009 2010 in plots with crimson clover or N fertilizer (Table 2). In 2010 2011, the opposite was observed: greater N yields were recorded for all legume, N, and control treatments harvested as forage compared with the biomass harvest regimen. This was a 48-kg ha 1 yr 1 difference in the case of hairy vetch, but only 6 kg ha 1 yr 1 for the control. This effect in part was possibly due to the relative maturity of the biomass plots compared with the biomass plots, especially considering that these were annual crops. The differences between seasons were likely attributed to the rainfall variability experienced throughout the trial, which may have caused variable responses among treatments. Nitrogen Yield: Forage vs. Bioenergy Value The extrapolated crude protein value of the positive fertilizer treatments harvested multiple times ranged from 100 to 190 g kg 1 in 2010 and 110 to 140 g kg 1 in 2011, indicating a high-nutritive-value forage that could be used for both lactating and stocker operations. Grass crude protein levels were lower and more variable ranging from 66 to 128 g kg 1 in 2010 and 45 to 79 g kg 1 in 2011, suggesting a lower potential forage nutritive value that could be offset as a bioenergy feedstock in low-rainfall years. Considering the positive fertilizer and grass treatments harvested for peak biomass production, forage N concentration, when extrapolated to percent crude protein, ranged from 80 to 110 g kg 1 (14 18 g N kg 1 ) for both years. These values could still be considered adequate for some ruminant livestock systems (Ball et al., 1998), suggesting that any potential bioenergy value would also have to overcome this base forage value. In addition, most lignocellulosic bioenergy systems function optimally with a feedstock containing <6% g N kg 1 (Kauter et al., 2003), suggesting potential negative issues with using these species as a feedstock. Fertilized Tifton 85 Spring First-Cut Dry Matter Yield: Grass Species Tifton 85 first-cut spring DM yields in control plots receiving 112 kg N ha 1 averaged 2193 kg ha 1 in 2009 2010 and 2163 kg ha 1 in 2010 2011. In control plots that received no added N, the first-cut spring yields averaged 872 kg ha 1 in 2009 2010 and 382 kg ha 1 in 2010 2011. In both years, there was a harvest regimen grass entry interaction for Tifton 85 first spring-cut DM yield as a percentage of first spring-cut control plot Tifton 85 yield, with or without N fertilizer (Table 3). Compared with Tifton 85 control plots receiving 112 kg N ha 1, overseeding barley and triticale onto dormant Tifton 85 reduced spring first-cut DM yields 10 to 11% more than oats and ryegrass when the cool-season grasses were harvested as forage in 2009 2010, but not in 2010 2011, the drier year. When the cool-season grasses were harvested as biomass, barley in 2009 2010 suppressed N-fertilized Tifton 85 regrowth less than triticale or ryegrass. In 2010 2011, N-fertilized spring first-cut Tifton 968 www.crops.org crop science, vol. 58, march april 2018

Table 3. Percentage of Tifton 85 (spring fertilized with 0 or 112 kg N ha 1 ) first-cut regrowth dry matter yield after overseeding with cool-season grasses harvested as forage or bioenergy feedstock (averaged over winter N fertilizer or legume treatments). 0 kg N ha 1 112 kg N ha 1 Percentage of control yield Percentage of control yield Year & cultivar Forage Biomass LSD 0.05 Forage Biomass LSD 0.05 % % 2009 2010 Barley 80 55 S 32 22 S Oat 54 45 NS 21 18 NS Rye 62 39 S 25 16 S Ryegrass 55 31 S 22 12 S Triticale 83 25 S 33 10 S Wheat 75 51 NS 30 20 NS Mean 68 41 27 16 LSD 0.05 22 26 9 10 2010 2011 Barley 201 242 S 36 43 S Oat 183 216 NS 32 38 NS Rye 189 177 NS 33 31 NS Ryegrass 240 171 NS 42 30 NS Triticale 192 192 NS 34 34 NS Wheat 217 247 S 38 44 S Mean 204 208 36 37 LSD 0.05 NS 70 NS 13 NS, not significantly different, P > 0.05; S, different at P 0.05. 85 DM yield was reduced less by barley and wheat than by rye or ryegrass after cool-season grasses were harvested a single time under a biomass harvest regimen. Under these conditions, cool-season grass regrowth in the spring may contribute sufficient biomass to mitigate Tifton 85 DM yield reductions due to shade or soil moisture competition with cool-season grasses. At locations with greater rainfall than ours, Han et al. (2012) reported March to September (spring summer) total DM yields for unfertilized bermudagrass overseeded with crimson clover that were equal to that of bermudagrass monoculture fertilized with 112 kg N yr 1. Their results are not directly comparable with ours because we did not harvest throughout the warm growing season, but they indicate that the positive effects of cool-season legumes may be delayed. Overseeding barley, rye, ryegrass, and triticale reduced fertilized spring first-cut Tifton 85 DM yields more under the biomass harvest regimen in 2009 2010 compared with the forage regimen (Table 3). The pattern was reversed in lower rainfall 2010 2011: overseeding barley and wheat reduced Tifton 85 yields less when harvested under the biomass regimen, rather than as forage. The apparent difference observed between harvest regimens on subsequent Tifton 85 DM yield is possibly due to the cool-season DM yield resulting from varying rainfall amounts. Rainfall seemed to have no effect on the overall DM yields between years; however, DM yields were 364% more in the greater rainfall year, which likely led to greater suppression of Tifton 85 yields. Unfertilized Tifton 85 Spring First-Cut Dry Matter Yield: Grass Species In unfertilized Tifton 85, DM yield losses as a percentage of first-cut spring Tifton 85 DM yields were greater for ryegrass and oats plots harvested as forage than for barley and triticale by 35 to 37% in 2009 2010 (Table 3). These differences were not apparent in 2010 2011, when rainfall was less during the cool season and greater during April and May compared with the year before. In barley and wheat plots harvested under the biomass regimen in 2009 2010, spring first-cut unfertilized Tifton 85 DM yields were greater than in triticale plots vis-à-vis control unfertilized Tifton 85 plots. In 2010 2011, spring first-cut Tifton 85 in wheat and barley plots again yielded as much or more DM, compared to the unfertilized control Tifton 85, than the other cool-season grass entries. In 2009 2010, spring first-cut Tifton 85 DM yields as a percentage of unfertilized, control Tifton 85 plots were greater in the barley, rye, ryegrass and triticale forage-harvest regime plots compared to those in the biomass-harvest regimen (Table 3). The reverse occurred for barley and wheat in 2010 2011: spring first-cut Tifton 85 DM yields as a percentage of unfertilized, control Tifton 85 plots were greater in the biomass compared to the forage regimen. Tifton 85 Spring First-Cut Dry Matter Yield with Spring N Fertilization: Cool-Season Legumes and N Fertilizer Both years there was a harvest regimen cool-season legume-fertilizer interaction for Tifton 85 spring first-cut DM yield as a percentage of spring first-cut control plot crop science, vol. 58, march april 2018 www.crops.org 969

Tifton 85 yield when receiving 112 kg N ha 1 in the spring (Table 4). Nitrogen fertilizer applied to cool-season grasses harvested as forage resulted in the greatest spring first-cut Tifton 85 DM yield. This may indicate residual soil N or less spring competition for soil moisture and solar irradiation in the absence of legumes, as is reported in legume-monoculture studies by Freeman et al. (2016) and Muir and Bow (2011), in which spring first-harvest bermudagrass yields were lower when overseeded with cool-season legumes the preceding winter compared with no-legume controls. Research by Han et al. (2012) indicated that March to September total DM yields (cool-season legume and bermudagrass) were comparable with bermudagrass by itself fertilized with 112 kg N yr 1. Compared with the 0-kg spring-n control Tifton 85 plots, these cool-season N plots yielded only 7% less Tifton 85 in 2009 2010 and 187% more in 2010 2011, the year with a preceding dry cool-season. Plots that had legumes during the cool season, by contrast, were undifferentiated from control plots with neither legumes nor N fertilizer in 2009 2010. During the drier 2010 2011, compared with the cool-season control, plots overseeded with hairy vetch in the forage harvest regimen produced 136% more spring first-cut, spring-fertilized Tifton 85 DM yield. Although there were no differences for the biomass harvest regimen in 2009 2010, cool-season N fertilizer plots had the greatest spring first-cut Tifton 85 DM yields as a percentage of N-fertilized Tifton 85 control in 2010 2011, the drier year. Irrigated cool-season grasses and legume monocultures overseeded onto dormant bermudagrass decreased early-season Tifton 85 despite adequate soil moisture throughout the cool season and spring (Muir and Bow, 2011). Although this was also the case with subterranean clover overseeded onto Coastal bermudagrass (Harris and Zuberer, 1993), by the end of the warm growing season, unfertilized bermudagrass yields can be as much as 50% greater in plots that contained legumes the previous cool season. Our results indicated that soil moisture is likely the most important factor: unlike years with greater coolseason rainfall, less cool-season precipitation combined with more spring moisture favored greater spring first-cut Tifton 85 DM yields, especially when the latter did not receive spring N fertilizer. This was true whether coolseason grass or legume was overseeded onto the dormant Tifton 85. This negative aftereffect of overseeding coolseason legumes onto dormant bermudagrass has also been reported for season-long bermudagrass yields in which hairy vetch reduced yields by 22% (Freeman et al., 2016). Except for the cool-season control, plots harvested as forage during 2009 2010 yielded greater spring first-cut Tifton 85 yields than the biomass regimen plots (Table 4). This was 118% greater in the case of cool-season N plots. By contrast, there were no differences between forage and biomass harvest plots in 2010 2011, when cool-season precipitation was low but spring rainfall was greater than in the previous year (Fig. 1). Tifton 85 Spring First-Cut Dry Matter Yield with No Spring N Fertilization: Cool-Season Legumes and N Fertilizer In both years, there were harvest regimen cool-season legume-fertilizer interactions for Tifton 85 spring first-cut DM yield as a percentage of spring first-cut control plot Tifton 85 yield when the latter received 0 kg N ha 1 in the spring (Table 4). Compared with cool-season legumes or control, N fertilizer applied to cool-season grasses harvested as forage resulted in the greatest spring first-cut Tifton 85 DM yield. Residual soil N from the cool-season application was even more evident in the absence of spring Table 4. Percentage of Tifton 85 (spring fertilized with 0 or 112 kg N ha 1 ) first-cut regrowth dry matter (DM) yield after overseeding with cool-season legumes or fertilized with winter N harvested as forage or bioenergy feedstock (averaged over cool-season annual grasses). 0 kg N ha 1 112 kg N ha 1 Percentage of control yield Percentage of control yield Year & cultivar Forage Biomass LSD 0.05 Forage Biomass LSD 0.05 % % 2009 2010 Hairy vetch 59 36 S 23 14 S Crimson clover 54 35 S 21 14 S N fertilizer 93 42 S 37 17 S Control 63 47 NS 25 19 NS Mean 67 40 27 16 LSD 0.05 9 NS 9 NS 2010 2011 Hairy vetch 186 210 NS 33 37 NS Crimson clover 143 177 NS 25 31 NS N fertilizer 287 321 NS 51 57 NS Control 77 102 NS 14 18 NS Mean 173 203 31 36 LSD 0.05 95 89 16 15 NS, not significantly different, P > 0.05; S, different at P 0.05. 970 www.crops.org crop science, vol. 58, march april 2018

N application (30% difference for 2009 2010, 210% for 2010 2011) than when spring N was applied to Tifton 85 (12% difference for 2009 2010, 37% for 2010 2011) in plots harvested as forage during the cool season. In contrast with 2009 2010, when there were no differences, spring first-cut Tifton 85 yields in 2010 2011 plots that received fall N fertilizer and were harvested as biomass in the coolseason outyielded all other cool-season treatments as a percentage of spring first-cut unfertilized Tifton 85. As with the Tifton 85 that was fertilized with N in the spring, plots harvested as forage during the cool season that subsequently received no spring N yielded more spring first-cut Tifton 85 DM than those harvested only once in a biomass regimen (Table 4). This was 64, 54, and 121% greater in the case of cool-season hairy vetch, crimson clover, and N plots, respectively. This was a similar result to that reported by Han et al. (2012) for spring and summer overseeded crimson clover and bermudagrass, which outyielded no-legume, no-n fertilizer bermudagrass by 11 to 63%, depending on growing conditions, at two Arkansas locations and 2 yr. CONCLUSIONS Because of the differences seen between a cool season with adequate precipitation and one with very low rainfall over a long period, the relative benefits of harvesting as forage vs. as bioenergy feedstock was heavily influenced by weather patterns and the seed set patterns of plant species (precocious vs. late maturing). Likewise, competition of cool-season annual forages with subsequent Tifton 85 early spring regrowth was related to precipitation patterns during the cool season, as well as the early warm season. Considering the relatively low biomass yields in the cool-season species when compared with warm-season grass bioenergy, overseeded cool-season grasses and legumes would provide a poor feedstock for energy conversion, especially when coupled with subsequent Tifton 85 suppression resulting from moisture or cover competition during bermudagrass green-up. Conflict of Interest The authors declare that there is no conflict of interest. References Ball, D.M., C.S. Hoveland, and G.D. Lacefield. 1998. Southern forages. 2nd ed. Potash & Phosphate Inst. Found. Agron. Res., Norcross GA. Bow, J.R., J.P. Muir, D.C. Weindorf, R.E. Rosiere, and T.J. Butler. 2008. Integration of cool-season annual legumes and dairy manure compost with switchgrass. Crop Sci. 48:1621 1628. doi:10.2135/cropsci2007.12.0697 Burton, G.W. 2001. Tifton 85 Bermudagrass: Early history of its creation, selection and evaluation. Crop Sci. 41:5 6. doi:10.2135/cropsci2001.4115 Cassida, K.A., J.P. Muir, M.A. Hussey, J.C. Read, B.C. Venuto, and W.R. Ocumpaugh. 2005. Biofuel component concentrations and yields of switchgrass in south central U.S. environments. Crop Sci. 45:682 692. doi:10.2135/cropsci2005.0682 Freeman, S.R., M.H. Poore, H.M. Glennon, and A.D. Schaeffer. 2016. Winter annual legumes seeded into bermudagrass: Production, nutritive value and animal preference. Crop Forage Turfgrass Manage. 2:1 9. doi:10.2134/cftm2014.0102 Fribourg, H.A., and J.R. Overton. 1973. Forage production on bermudagrass sods overseeded with tall fescue and winter annual grasses. Agron. J. 65:295 298. doi:10.2134/agronj1973.00021962006500020032x Guney, M., C. Kale, D. Bolat, and S. Deniz. 2016. Determination of the yield characteristics and in vitro digestibility of barley forage harvested in different vegetation periods. Indian J. Anim. Res. 50:947 950. Han, K.-J., M.W. Alison, Jr., W.D. Pitman, and M.E. McCormick. 2012. Contributions of overseeded clovers to bermudagrass pastures in several environments. Crop Sci. 52:431 441. doi:10.2135/cropsci2011.08.0401 Harris, P.A., and D.A. Zuberer. 1993. Subterranean clover enhances production of Coastal bermudagrass in the revegetation of lignite mine spoil. Agron. J. 85:236 241. doi:10.2134/agronj1 993.00021962008500020014x Hecht, A. D., D. Shaw, R. Bruins, V. Dale, K. Kline, and A. Chen. 2009. Good policy follows good science: Using criteria and indicators for assessing sustainable biofuel production. Ecotoxicology 18:1 4. doi:10.1007/s10646-008-0293-y Kauter, D., I. Lewandowski, and W. Claupein. 2003. Quantity and quality of harvestable biomass from Populus short rotation coppice for solid fuel use: A review of the physiological basis and management influences. Biomass Bioenergy 24:411 427. doi:10.1016/s0961-9534(02)00177-0 Knight, W.E. 1970. Productivity of crimson and arrowleaf clovers grown in a Coastal Bermuda sod. Agron. J. 62:773 775. doi:10.2134/agronj1970.00021962006200060027x Mehlich, A. 1984. Mehlich-3 soil test extractant: A modification of Mehlich-2 extractant. Commun. Soil Sci. Plant Anal. 15:1409 1416. doi:10.1080/00103628409367568 Muir, J.P., and J.R. Bow. 2011. Yield dynamics of Tifton 85 overseeded with cool-season forages. Agron. J. 103:1019 1025. doi:10.2134/agronj2011.0060 Muir, J.P., B.D. Lambert, A. Greenwood, A. Lee, and A. Riojas. 2010. Comparing repeated forage bermudagrass harvest data to single, accumulated bioenergy feedstock harvest. Bioresour. Technol. 101:200 206. doi:10.1016/j. biortech.2009.07.078 Robinson, R.R. 1963. Rainfall distribution in relation to sodseeding for winter grazing. Agron. J. 55:307 308. doi:10.2134/ agronj1963.00021962005500030032x Sánchez, E., S. Gil, J. Azcón-Bieto, and S. Nogués. 2015. The response of Arundo donax L. (C 3 ) and Panicum virgatum (C 4 ) to different stresses. Biomass Bioenergy 85:335 345. doi:10.1016/j.biombioe.2015.12.021 SAS Institute. 2013. The SAS system for Windows. Release 6.10. SAS Inst., Cary, NC. Soil Survey Staff. 1973. Soil survey of Erath County, Texas. USDA, Soil Conserv. Serv., Texas Agric. Exp. Stn., College Station, TX. crop science, vol. 58, march april 2018 www.crops.org 971