A gronomy J our n al • Volume 10 0 , I s sue 3 • 2 0 0 8 765 ABSTRACT Comparison among warm-season legume forage trials may not be valid if harvest techniques vary. To address this question, herbage dry matter (DM) yields, branching, crude protein (CP), and fi ber concentrations for nine warm-season annual herbaceous legumes were measured by hand-plucking all leaves and pliable tips or clipping at 7.5-or 15-cm height. Th e experiment was conducted in Texas on a Windthorst fi ne sandy loam over 2 yr. Harvest technique did not aff ect DM yield in 2004, but the hand-plucked harvest technique produced 34 to 39% less forage in 2005 (low rainfall year) compared with the clipped plots. Most entries had greater branching on hand-plucked than on clipped plants (entry by harvest P < 0.05). Crude protein concentration was greater (P < 0.05) and fi ber concentrations lower in the hand-plucked compared with the clipped plants. Th ese results suggest that neither yields nor nutritive values of hand-plucked forage trials examining annual warm-season herbaceous legumes should be compared with clipped forage trials, whereas clipping heights may be less problematic. Results support a careful choice of experiment harvest technique based on the future fi eld-scale harvest method or degree of target herbivore selectivity.

MATERIALS AND METHODSAn on-dairy trial was performed from 2005 to 2007 near Dublin, TX (32°15' N, 98°12' W, altitude 395 m) on an established 5-yr-old stand of Tift on 85 with a history of dairy manure application. Th e soil in the experiment area was a Windthorst fi ne sandy loam (pH 7.7, 83 mg P kg -1 , 249 mg K kg -1 , 1770 mg Ca kg -1 , and 147 mg Mg kg -1 using the Mehlich III extractant method; Mehlich, 1984). Th ese values were stable from year to year. Adjacent but distinct experiment sites under the same pivot each year were seeded with 13 annual cool-season ABSTRACT 'Tift on 85' bermudagrass [Cynodon dactylon L. (Pers.) X C. transvaalensis Burt Davy] is a warm-season forage used on southeastern U.S. dairy concentrated animal feeding operations (CAFO). Cool-season annual forages need to be identifi ed that can be overseeded into Tift on 85 and remove the greatest quantity of P or N while minimizing suppression of bermudagrass spring regrowth. During 3 yr we overseeded six grass, fi ve legume, and two Brassica spp. on a dairy CAFO Tift on 85 fi eld on a Windthorst fi ne sandy loam (fi ne, mixed, active, thermic Udic Paleustalf) containing 83 mg plant-available P kg -1 soil. Hairy vetch (Vicia villosa Roth.), barley (Hordium vulgare L.), oat (Avena sativa L.), and rye (Secale cereale L.) yielded the most (entry × year interaction P < 0.05) dry matter (DM) during at least 2 of the 3 yr, with the grasses and some legumes reducing (entry × year interaction P < 0.05) spring fi rst-harvest Tift on 85 DM yields by 41 to 80%. Cumulative cool-season and fi rst-harvest Tift on 85 P yields increased (entry × year interaction P < 0.05) in some combinations and years compared to Tift on 85 alone, while N yields rose up to 173% when legumes were overseeded, depending on year and combination (entry × year interaction P < 0.05). Results indicate that overseeding some cool-season winter annual forage species onto dormant Tift on 85 can increase N for CAFO feed but only occasionally will it improve P removal.

Annual cool‐season legumes can contribute forage or green manure to warm‐season grass pastures or biofuel crops but may interfere with subsequent grass development. Arrowleaf clover (Trifolium vesiculosum Savi cv. Yuchi), common vetch (Vicia sativa L.), and button medic [Medicago orbicularis (L.) Bartal cv. Estes] were oversown on 10‐yr‐old switchgrass (Panicum virgatum L.) in north‐central Texas to evaluate dry matter (DM) yield and nutritive value. Legumes were no‐till drilled into a Windthorst fine sandy loam in autumn of 2000 and 2001; plots either received no dairy manure compost or 30 Mg compost ha−1 Forage was harvested when individual legume species reached flowering, when switchgrass regrowth reached boot stage, and at the end of the summer growing season. Switchgrass DM yields (3022–6630 kg ha−1 yr−1) were not reduced by overseeded legumes. Arrowleaf clover had the greatest production among the legume species and yielded more in a monoculture (1762–1923 kg ha−1 yr−1) than with switchgrass (757–814 kg). Cumulative yields increased the second year as a result of legumes, compost, and combinations of the two. Compost increased phosphorus concentrations in both the grass and legumes, but not crude protein or acid detergent fiber concentrations. Cool‐season annual legumes may be beneficial in switchgrass forage and biofuel systems, especially in combination with manure compost.

electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. L egumes can contribute substantial N fi xation to natural and agricultural systems (Brink and Fairbrother, 1988). Few non-native, perennial, warm-season herbaceous legumes, however, have been identifi ed as widely adapted to the southern Great Plains of North America, so the search is on for natives that will fi ll the gap. Of more than 20 genera of herbaceous perennial legumes native to north-central Texas and south-central Oklahoma (Diggs et al., 1999), only a few have been evaluated for cultivation (Muncrief and Heizer, 1985;Schweitzer et al., 1993;Muir et al., 2005). Th is severely constrains the use of herbaceous legumes for cultivated pastures, green manure, rangeland restoration, right-of-way reseeding, native prairie re-establishment, biofuel production, and wildlife plantings.Native tick-clovers (Desmodium spp.) and bush-clovers (Lespedeza spp.) have been documented as important wildlife vegetation in Texas and Oklahoma (Gee et al., 1994;Stevens et al., 2002;Dillard et al., 2006). Data indicate that they are heavily browsed in late summer and autumn by white-tailed deer (Odocoileus virginianus), and nutritive values (Posler et al., 1993;McGraw et al., 2004;Muir et al., 2005) indicate promise for both extensive and intensive applications. However, little is known of their persistence, forage, C, N, or seed yields under defoliation. With interest in C-negative biofuels from lowinput high-diversity native grasslands on the increase (Tilman et al., 2006) and legume N fi xation sought as an alternative to industrial fertilizers (Sprent, 1999), both C and N yields from native legumes are important to today's agricultural and natural systems.Seeds of 23 native, herbaceous perennial legumes were collected from more than18 rangeland sites in Texas and Oklahoma during 2000. Th ese were planted in small plots the following spring and evaluated for establishment, herbaceous production, and seed yield (Muir et al., 2005). Of the 15 that survived, three were selected for further agronomic evaluation based on herbage nutritive value as well as ease of seed harvest. Th is research evaluated herbage dry matter (DM), C, and N yields of those three (Tweedy's tick-clover, panicled tickclover, and tall bush-clover) under three levels of defoliation. Seed number and winter survival were measured to determine persistence and reproductive potential under four levels of defoliation. MATERIALS AND METHODSIn February 2004, seeds of tall bush-clover, panicled tickclover, and Tweedy's tick-clover were scarifi ed, inoculated with a general "cowpea" Rhizobium inoculant (EMD Crop BioScience, Brookfi eld WI), and germinated in 100-mL containers fi lled with topsoil collected at the Stephenville, TX (N 32° 15´, W 98° 12´, altitude 395 m) fi eld plot site. Th e soil in the experiment area was a Windthorst fi ne sandy loam (fi ne, mixed, active, thermic Udic Paleustalf) (pH 6.6, 11...

Few native herbaceous perennial warm‐season legumes are available for cultivation in the southern United States. These would be useful for pastures, biomass production, wildlife plantings, rangeland reseeding, or native prairie restoration. Three native perennial legumes were selected from an initial screening of north Texas germplasm for potential domestication based on successful establishment, ease of seed harvest, and herbage production. Response to herbage removal was evaluated by measuring forage nutritive value when clipped at 10‐, 20‐ and 40‐cm height (whenever regrowth exceeded 20 cm above cutting height) over three seasons in 2‐ by 2‐m plots with 0.5 m between plants. Tweedy's tick‐clover (Desmodium tweedyi Britton) became reproductive in June so herbage acid detergent fiber (ADF) (338–373 g kg−1) was greater (P < 0.05) than tall bush‐cover (Lespedeza stuevei Nutt.) and panicled tick‐clover [D. paniculatum (L.) DC.] that both flowered primarily in September. Herbage N concentrations were uniformly (P > 0.05) high (31.7 g kg−1) the first year and unaffected by harvest height all 3 yr; by the third year, however, tall bush‐cover had the lowest (P < 0.05) N concentration (22.6 g kg−1) and in vitro dry matter disappearance (IVDMD; 651 g kg−1). Condensed tannin (CT) concentrations exceeded 120 g kg−1 for both the panicled tick‐clover and the tall bush‐clover although fraction distribution varied among entries. These three native North American legumes show potential as nutritious forage components to natural and cultivated systems.