Effects and Sustainability of Clover Inclusion within Warm-Season Turf Swards by James D. McCurdy A dissertation submitted to the Graduate Faculty of Auburn University in partial fulfillment of the requirements for the Degree of Doctor of Philosophy Auburn, Alabama December 14, 2013 Urban Ecology, Sustainable Landscapes, Turfgrass, Legume Inclusion, Nitrogen Fixation, Clover Copyright 2013 by James D. McCurdy Approved by J. Scott McElroy, Chair, Associate Professor of Agronomy and Soils Elizabeth A. Guertal, Professor of Agronomy and Soils C. Wesley Wood, Professor of Agronomy and Soils Amy N. Wright, Professor of Horticulture ii Abstract Efforts to decrease supplemental nitrogen (N) applications to turfgrass justify alternative fertility strategies such as legume inclusion. Legumes such as clovers (Trifolium spp.) are present within many turfgrass scenarios. Legume persistence is partly due to an ability to biologically fix atmospheric N, which is incorporated into the plant as proteins and other compounds. N is subsequently shared with associated grasses through the decomposition of legume -roots and - foliage. For this reason, turf health is often improved rather than diminished. There are very few guidelines for white clover (T. repens) establishment and maintenance within warm-season turfgrasses. In fact, much of what we know is from clover inclusion within forage and pasture scenarios. Research was conducted to answer serious knowledge gaps preventing the implementation of white clover inclusion within warm season turf swards. Four studies were conducted to evaluate seeded white clover establishment within dormant bermudagrass (Cynodon spp.) turf as affected by 1) pre-seeding mechanical surface disruption, 2) establishment timing, 3) seeding rate, and 4) companion grass species. White clover establishment was improved by scalping prior to October seeding, but these effects were not further enhanced by the addition of verticutting or hollow tine aerification. Un-scalped turfgrass yielded nearly 50% lower white clover densities than those scalped prior to seeding, possibly due to decreased seed to soil contact and increased bermudagrass competition. January and February establishment dates generally yielded the lowest spring clover densities, while October timing yielded superior establishment. Clover densities resulting from six seeding rates (0 to 6.0 g live seed m -2 ) were fit to the linear model y = y 0 + ax b , where y equals trifoliate leaves m -2 and x is iii equal to initial seeding rate. An important feature of this model was that it accurately represented the diminishing response of increasing seeding rate. Clover establishment was negatively correlated with companion grass densities, with the largest densities occurring when planted with tall fescue and the smallest when planted with annual ryegrass. Weed control within turf-clover swards is often hampered by the lack of effective herbicides that are safe on clovers. Furthermore, differential tolerance of legume species to common row-crop and pasture herbicides has previously been reported. Field and greenhouse studies conducted in Auburn, AL indicate varying herbicide tolerances of Trifolium species to common turf herbicides. In field experiments, imazaquin controlled hop clover 91% but controlled white clover only 50%. Imazaquin reduced hop clover height 87%, which far exceeded height reductions measured among other clovers (< 46%). Although visual estimates of 2,4-DB control (35%) did not differ due to species, differential height reductions were significant. 2,4-DB failed to reduce the height of crimson and ball clovers, while white clover was almost 50% shorter than the non-treated. In contrast to field experiments, 2,4-DB control during greenhouse experiments was less than all other clovers (3% versus > 40% for other clovers). These differential herbicide tolerances are novel but must be refined in order to be adapted in real-world scenarios. On a practical level, our results demonstrate potential herbicide options for maintaining biodiverse turf-legume swards. Candidate herbicides include bentazon, MCPA, 2,4-DB, imazaquin, and imazethapyr. The relative tolerance of clover species to these candidate herbicides is further evidence of their utility within certain mixed turf scenarios. Little is known of the N contribution and carbon (C) sequestration from decaying clover foliage. An in situ decomposition study was conducted in Auburn, AL to quantify C and N - release from the decomposition of white clover (T. repens L.) foliage within a bermudagrass iv lawn. Fresh white clover was applied during March, June, and December and was retrieved periodically after application. Four parameter double exponential decay models were used to describe clover mass as well as N and C -loss. These models reveal important features of white clover decomposition; mainly that white clover is composed of a quickly decaying labile fraction. White clover litter applied at 0.5 kg fresh weight (FW) m -2 potentially contributed from 2.9 to 4.2 g N m -2 , with more than half available for mineralization between 10 and 73 days after application, depending upon time of year. Given that clover populations are regenerative, litter deposited during mowing events may be considered a viable N source to sustain healthy turf. Knowledge of the decomposition of clover within turf swards will enable turfgrass- researchers and professionals to more accurately predict nutrient contribution to associated grasses and help optimize supplemental fertilizer recommendations. A 3-year study evaluated the effects of white clover inclusion within a hybrid bermudagrass lawn. Supplemental N (0, 0.5, 1, 2, 4, and 8 g N m -2 ) was applied monthly, April to August, in order to evaluate the effects of supplemental N upon biomass composition, N fixation, N transfer, and soil carbon. Mixed grass plus clover swards yielded higher clipping biomass than grass-alone swards, which was evidence of enhanced bermudagrass growth due to biological N fixation. Likewise, grass biomass of mixed swards was increased relative to that of grass-alone swards at supplemental N rates ! 10 g N m -2 year -1 but was decreased at higher supplemental N rates. N fixation was estimated to be 6.6 g m -2 year -1 during the 3-year study, with an apparent increase in fixation as years progressed. Results indicate that N fixation was suppressed at the lower and upper extremes of supplemental N rates. N transfer to the associated bermudagrass sward was estimated to be 24% during the latter two years of the study. Soil carbon levels were similar among treatments. v Acknowledgments To my committee members, thank you for your time and patience. You have been inspiring and helpful throughout this more than 4-year-long process. To the graduate students I have agonized and toiled along side - Hunter Perry, Mark Doroh, Caleb Bristow, Ethan Parker, Ian Turner, Mike Mulvaney, Michael Flessner, Phillipe Aldahir, Phil Bruner, Patrick Connard, Jared Hoyle, Steve Lee, Brandon Green, and so many more - thank you for the time we spent together. To the undergraduate workers, thank you for your honest labor and willingness to do often-underappreciated jobs. To Brenda Wood, thank you for shepherding me through so many processes that I knew nothing about. To Dr. Jorge Mosjidis, thank you for your kindness and support during the preliminary phase of my work. To my wife Vicky, thank you for your love, kindness, and patience during our journey through graduate school and life. Mostly, thank you for your wisdom and for knowing how to encourage me through the tough times. vi Table of Contents Abstract ......................................................................................................................................... ii Acknowledgments ........................................................................................................................ v List of Tables ............................................................................................................................. viii List of Figures ............................................................................................................................... x Literature Review ....................................................................................................................... 1 Turfgrass Sustainability .................................................................................................... 1 Clover Inclusion .............................................................................................................. 3 Clover Habitat ................................................................................................................. 4 Insect Habitat ................................................................................................................... 4 Trifolium Taxonomy ....................................................................................................... 5 Origins and Evolution ..................................................................................................... 7 Nitrogen Fixation ............................................................................................................. 9 Nitrogen Fixation within Mixed Grass Clover Swards ................................................ 11 Decomposition of Clover Biomass within Turf ............................................................ 13 Clover Establishment ..................................................................................................... 14 Weed Control in Mixed Grass Clover Swards .............................................................. 17 White Clover Establishment within Dormant Bermudagrass Turf ........................................... 23 Introduction .................................................................................................................. 23 Materials and Methods ................................................................................................. 27 vii Results .......................................................................................................................... 29 Discussion and Implications ......................................................................................... 34 Differential Response of Four Trifolium Species to Common Broadleaf Herbicides .............. 44 Introduction .................................................................................................................. 44 Materials and Methods ................................................................................................. 45 Results and Discussion ................................................................................................. 48 Implications for Management ....................................................................................... 52 Dynamics of White Clover Decomposition in a Southeastern Bermudagrass Lawn ................ 60 Introduction .................................................................................................................. 60 Materials and Methods ................................................................................................. 63 Results and Discussion ................................................................................................. 65 Conclusions .................................................................................................................. 72 White Clover Inclusion within a Bermudagrass Lawn and the Effects of Supplemental Nitrogen upon Botanical Composition and Nitrogen Dynamics .................................................. 80 Introduction .................................................................................................................. 80 Materials and Methods ................................................................................................. 83 Results and Discussion ................................................................................................. 85 Conclusions .................................................................................................................. 92 Literature Cited ......................................................................................................................... 99 Appendix 1 ............................................................................................................................. 120 Introduction ................................................................................................................ 120 Materials and Methods ............................................................................................... 120 Results and Discussion ............................................................................................... 122 viii List of Tables Table 1. Companion grasses planted 14 October, 2010 and 1 October, 2011 with two commercial white clover varieties (1.5 g live seed m-2). ................................................................. 37 Table 2. Analysis of variance (ANOVA) for 2010-2011 and 2011-2012 white clover establishment trials. Replication year was significant in all studies; therefore, analysis was performed separately. .............................................................................................. 38 Table 3. April observed, spring white clover density as affected by mechanical surface disruption methods. .......................................................................................................................... 39 Table 4. White clover density as affected by seeded establishment timing. .............................. 40 Table 5. Companion grass densities along side affected white clover densities. ....................... 41 Table 6. Four clover (Trifolium) species and their respective harvest and transplant dates. Plants were harvested and allowed to mature in a greenhouse setting. Plants were then subject to selection for uniform size and maturity followed by random assignment to either field or greenhouse experiments. ................................................................................................ 54 Table 7. Herbicide rates and formulations applied in field and greenhouse experiments to four clover (Trifolium) species. All treatments included a 0.25% v v -1 non-ionic surfactant. Herbicides were applied at 280 L ha -1 spray volume. Experimental rates were chosen based upon common labeled rates and unpublished studies where legume tolerance had been observed. ............................................................................................................... 55 Table 8. ANOVA results and source sum of squares (SS) relative to the total SS for field and greenhouse experiments 6 WAT. .................................................................................. 56 Table 9. Control and height reductions of four clover (Trifolium) species measured 6 weeks after treatment (WAT) in field studies. Effects were restricted to P ! 0.05 level of significance. Effects were combined across years. Model validity (P > F) is provided for significant species by herbicide interaction. .................................................................................... 57 Table 10. Herbicide main effects upon clover (Trifolium) biomass reductions measured 6 weeks after treatment (WAT) during greenhouse experiments. ................................................ 59 Table 11. Double exponential decay equations regressed on time (days) for mass, carbon (C), and nitrogen (N) -loss from white clover incubated in litter bags under field conditions. Double exponential decay equations are of the form Y = Ae -k1t + Be -k2t , where Y = response, A approximates the labile portion, B approximates the recalcitrant portion, k1 and k2 are rate constants fitted to the data, and t = time in days after application. Percent remaining data were normalized to initial day 0 applications to facilitate approximations of labile and recalcitrant portions. .................................................................................. 74 Table 12. Persistence of white clover litter based upon predicted 95% confidence bands of double exponential decay equations (Table 1) regressed on time (days) for nitrogen (N), mass, and carbon (C) -remaining. Residue persistence was normalized to 100% ash free dry weight of initial day 0 applications .......................................................................... 79 ix Table 13. Grass alone and grass plus white clover biomass, as well as clover portion of the mixed sward, ? 95% confidence interval (CI) as affected by supplemental N. Means ? CI are presented to allow treatment separation among similar response variables. .................. 95 Table 14. Grass alone and grass plus clover biomass, as well as estimated biological N fixation and N transfer to associated grasses, regressed upon yearly supplemental N levels. Transfer was not calculated in 2010. Data were fit to the quadratic model y = a + bx + cx 2 , where a is the estimated response at 0 g N m -2 year -1 , b is the linear coefficient, c is the quadratic coefficient, and x is g N m -2 year -1 applied over five months of active bermudagrass growth from April to August as CaNO 3 . ................................................. 97 Table 15. Nitrogen Fixation and Nitrogen Transfer ? 95% confidence interval (CI) as affected by supplemental N. Means ? CI are presented to allow treatment separation among similar response variables. .......................................................................................................... 98 x List of Figures Figure 1. Trifolium taxonomy. ................................................................................................... 18 Figure 2. Rosids (or Eurosids) clade according to Wang et al. (2009) and Worberg et al. (2009) ............................................................................................................................ 19 Figure 3. Papilionoideae (Wojciechowski et al., 2000) is the largest of the three subfamilies of Fabaceae. ....................................................................................................................... 20 Figure 4. The inverted repeat-lacking clade. .............................................................................. 21 Figure 5. Legume root nodulation process. ................................................................................ 2 Figure 6. Root hair curling and invasion. ................................................................................... 22 Figure 7. Nodule meristem and zones of infection. .................................................................... 22 Figure 8. April observed white clover density as a function of five rates of October seeded white clover. Error bars represent 95% confidence intervals about the mean. . ...................... 42 Figure 9. Correlation of December grass density and April white clover density during seasons one and two of rate response experiment, where r = Pearson?s Correlation Coefficient, and P = probability that r is different from 0. Correlations were only significant when white clover data were combined across companion grass species. ............................... 43 Figure 10. Average daily soil temperatures at 10 cm depth at the study site and average daily air temperature at 1.5 m near the Auburn, AL study site. ................................................... 75 Figure 11. Percent mass remaining from surface incubated white clover residue. Shapes represent mean ? 95% confidence intervals (CI?s). Residue persistence was normalized to 100% ash free dry weight of initial Day 0 applications. Days to 50% decomposition (D50) values are presented on the horizontal axis with adjusted 95% CI?s as a means of comparing residue persistence across application date. ................................................ 76 Figure 12. Percent carbon (C) remaining from surface incubated white clover residue. Shapes represent mean ? 95% confidence intervals (CI?s). Residue persistence was normalized to 100% ash free dry weight of initial Day 0 applications. Days to 50% decomposition (D50) values are presented on the horizontal axis with adjusted 95% CI?s as a means of comparing residue persistence across application date. ................................................. 77 Figure 13. Percent nitrogen (N) remaining from surface incubated white clover residue. Shapes represent mean ? 95% confidence intervals (CI?s). Residue persistence was normalized to 100% ash free dry weight of initial Day 0 applications. Days to 50% decomposition (D50) values are presented on the horizontal axis with adjusted 95% CI?s as a means of comparing residue persistence across application date. ................................................. 78 Figure 14. Average daily air and soil temperature as well as average daily precipitation for the 2009-2010, 2010-2011, and 2011-2012 bermudagrass-white clover growing seasons. .94 xi Figure 15. Grass alone and grass plus clover biomass, as well as estimated biological N fixation and N transfer to associated grasses, regressed upon yearly supplemental N levels. Transfer was not calculated in 2010. Data were fit to the quadratic model y = a + bx + cx 2 , where a is the estimated response at 0 g N m -2 year -1 , b is the linear coefficient, c is the quadratic coefficient, and x is g N m -2 year -1 applied over five months of active bermudagrass growth from April to August as CaNO 3 . Means ? 95% confidence intervals are presented to allow treatment separation among similar response variables. ............ 96 Figure A1. Root dry weight response to supplemental N applied as CaNO3. Effects of inoculation are displayed, but they are insignificant (P > 0.05). .................................. 104 1 Literature Review Turfgrass Sustainability The sustainability of urban environments is among the foremost issues facing humanity. More than 80% of the United States population resides within urban or suburban environments, and it is estimated that greater than 90% will reside within urban centers by the year 2050 (United Nations, 2009, CIA World Fact Book, 2013). Urban ecology has become a central concern for residents, designers, and ecologists, alike. In much of the U.S., rooftops, parking lots, busy city streets, and home lawns are quickly replacing native flora. Civilization, for better and for worse, has changed the way we interact, build, and perceive our environment. Turfgrass is just one result of these changes. Turfgrass has been a mainstay of U.S. urban ecology since the mid-20 th century, during which large tracts of land were developed to accommodate growing urban populations. Turfgrass comprises 163,800 km 2 (?35,850 km 2 ) of the contiguous United States (Milesi et al. 2005), an area roughly the size of Florida. Turfgrass occupies approximately 1.9% of U.S. surface area and by some estimates is the largest irrigated crop within the contiguous U.S. Turfgrasses and their definitive uses vary around the world. However, in the U.S., turf is frequently utilized for transportation right-of-way, golf courses, sports-pitch, and commercial- and residential- lawns. In fact, it has been estimated that roughly 80% of U.S. cultivated turf inhabits residential lawns (Roberts and Roberts, 1987). Benefits of turf are well documented and include: recreational health, erosion control, increased water infiltration, reduced nutrient leaching, aesthetics, carbon (C) sequestration, and mediation of the ?heat-island? effect (Beard and Green 1994, Qian and Follett, 2002). Yet the ecological impact of turf is often questioned, due in part to nutrient and water requirements as 2 well as its often-unsustainable monoculture cultivation (Milesi et al. 2005, Robbins and Birkenholtz 2003, Robbins et al. 2001). Turfgrass is often managed using repeat applications of synthetic fertilizers and pesticides, which are costly and may be detrimental to the environment (Robbins and Birkenholtz 2003, Robbins et al. 2001). Nitrogen (N) is essential to turf health and quality (Beard 1973, Turgeon 2002). Commercial-lawn N requirements vary with species and environmental conditions, but within the southern U.S. common rates range from less than 5 g m - 2 year -1 for bahiagrass (Paspalum notatum) and centipedegrass (Eremochloa ophiuroides) to almost 30 g m -2 year -1 for hybrid bermudagrass (Cynodon dactylon ! C. transvaalensis; Duble 1996). Improper N fertilization leads to negative environmental effects. Nitrogen loss from turf contributes to surface water eutrophication, leads to elevated nitrate (NO 3 ) levels in drinking water, and contributes to rising global temperatures by emitting the potent greenhouse gas nitrous oxide (N 2 O; Robbins and Birkenholtz 2003, Robbins et al. 2001, Wu and McGechan 1999). In addition, resources currently used to maintain turfgrass would arguably be more efficiently allocated if used in food-production. Nitrogen application often leads to a lush monoculture turfgrass sward that favors plant- feeding arthropods by influencing bottom-up effects on nutritional quality and chemical defenses of their hosts (Busey and Snyder 1993, Davidson and Potter 1995, Salminen et al. 2003) and by reducing harborage and alternative resources for natural enemies (Braman et al. 2002, Frank and Shrewsbury 2004). Furthermore, pesticides required to support these conditions can disrupt ecosystem services, leading to soil compaction and excessive thatch accumulation, pest resurgence, or secondary pest outbreaks (Lopez and Potter 2000, Peck 2009, Potter 1993). 3 An equally important consequence of turfgrass cultivation may be its impact upon insect habitat -loss and -fragmentation (Gels et al. 2002). As urban areas expand, managed landscapes replace natural insect habitat. Furthermore, the aesthetic standards for manicured turfgrass, such as that found upon golf courses and home lawns, result in significant insecticide use to control foliage-feeding insects, which further disrupts ecosystem stability. Clover Inclusion Since the advent of herbicides, efforts in the turf industry have often focused on maintaining monocultures for aesthetics and increased playability. For this reason, a biologically diverse turf sward, with mixed species of grasses and broadleaves, is sometimes classified as weedy and therefore undesirable for home lawns and golf courses. However, for many scenarios, the environmental impact of biodiversity may outweigh those of monoculture. Inclusion of leguminous species, which biologically fix N and provide pollinator habitat, is a proposed means of increasing the sustainability of certain low maintenance turfgrass scenarios. White clover (Trifolium repens L.) is well suited for use within warm-season turfgrasses and is already a common feature within bermudagrass pastures of the southeastern U.S. (Brink and Fairbrother, 1991). White clover increases turfgrass greenness by contributing N to associated grasses and has been reported to increase turfgrass color ratings within cool-season turfgrass (Sincik and Acikgoz, 2007) as well as increase vegetative cover within dormant bermudagrass (Dudeck and Peacock, 1983). Estimates of white clover N fixation within three cool-season turfgrasses are greater than 25 g N m -2 year -1 , with 4.2 to 13.7% of total N contributed to the associated turfgrasses (Sincik and Acikgoz, 2007). Other pertinent research concerning white clover inclusion has been conducted in forage scenarios where white clover was grazed or harvested for animal fodder. Estimates of N fixation 4 for grass-white clover pastures range from nil to 40 g N m -2 year -1 , though most are from 10 to 25 g N m -2 year -1 (Ledgard and Steele, 1992; McNeil and Wood, 1990; Whitehead, 1995). Using the 15 N transfer method, McNeil and Wood (1990) estimated N fixation by white clover within perennial ryegrass was approximately 15.5 g N m -2 year -1 , with 28% of the total fixed N having been transferred to associated ryegrass. Clover Habitat Various clovers (Trifolium spp.) occur throughout the world and are found within a wide range of habitats. They are commonly found within areas that receive high solar irradiation, and they rarely tolerate low light conditions. Clovers are frequently cultivated as livestock forage and as green manure within rotational cropping systems. In fact, it is theorized that certain clovers have co-evolved with foraging animals and relied upon them for their maintenance and transmittance (U.S. Fish and Wildlife, 2007). Clovers may also have been aided in development by pollinators such as bees (Apis, Bombus, and other spp.). Insect Habitat There are many reasons that warrant further research into new, likely biodiverse, turfgrass swards. As urban areas expand, turfgrass continues to supplant and augment natural insect habitat. Incorporating nectar-producing plants, such as legumes, into turf habitats has been shown to attract and sustain pollinating insects, such as Apis spp., and predatory arthropods, such as Tiphia vernalis and Larra bicolor (Abraham et al. 2010, Rogers and Potter 2004). Diversification of turfgrass ecosystems to conserve and augment natural enemies is increasingly recognized as compatible with golf course and home lawn maintenance (Held and Potter 2011). A negative consequence, however, may be that insecticides are applied to turf areas with flowering weeds that attract honeybees and native pollinators. Lawn care professionals, 5 homeowners, and golfcourse superintendants routinely apply insecticides to lawns with flowering weeds (Potter 1998, Racke and Leslie 1993, Racke 2000). Foliar feeding pests are typically controlled with applications of organophosphate, carbamate or pyrethroid insecticides, with residues allowed to dry on stems (Potter 1998). Exposure to these insecticides has been associated with bee poisonings in food crops (Kevan 1975, Johansen 1977, Kearns et al. 1998). Such compounds may intoxicate pollinators through direct contact, exposure to residues, or spray contamination of nectar and pollen (e.g., Burgett and Fisher 1980, Johansen et al. 1983). Trifolium Taxonomy According to the USDA?s Plant Database (2013), the order Fabales contains only one family - Fabaceae. Similarly, the Cronquist System places only the family Fabaceae within the order Fabales (Cronquist, 1981). However, the Angiosperm Phylogeny Group (APG) has rather convincingly listed Fabaceae as well as Quillajaceae, Surianaceae, and Polygalaceae families as part of the order Fabales (Stevens, 2001). Quillajaceae has previously been included within Rosaceae (Takhtajan, 1997) or within Spiraeaoideae as Quillajeae (Robertson, 1974). Members of this family are small evergreen trees that contain saponins within their bark. The only apparent economically important species from this family is the soapbark tree (Quillaja saponaria). A native of the temperate climes of central Chile, this tree has many uses. Most notably, its saponins have application as adjuvant within certain anti-viral medicines (Dalsgaard, 1978; Takahashi et al., 1990). The extracts obtained from its bark are also commonly used as food additives for their foaming characteristic (Eastwood et al., n.d.). Surianaceae has previously been included within Rosales by Cronquist (1981) and in Rutales by Takhtajan (1997). It is synonymous with Stylobasiaceae. Plants range from small 6 shrubs to tall trees (Stevens, 2001). Stevens (2001) also notes that this family is quite variable in vegetative description and that members of this family have not been studied extensively for their chemical properties. Polygalaceae are widely distributed throughout the world. The family has previously been grouped within its own order, Polygalales, by Cronquist (1981) and includes many perennial or annual herbs, shrubs, and trees. Taxonomic features vary. Polygalaceae include the Polygala genus, commonly referred to as milkwarts. These plants have numerous medicinal properties. Fabaceae, sometimes called the pea family or bean family, is the third largest angiosperm family. The family is characterized by compound leaf structure. Its flowers are highly variable; though, fruit of these plants are characteristically legumes. Plants of this family are known for their symbiotic relationships with rhizo-bacteria, an end result of which is fixation of atmospheric nitrogen (Sprent, 2001). Fabaceae has traditionally been divided into three subfamilies, Mimosoideae, Caesalpinioideae, and Papilionoideae (Polhill and Raven, 1981). There has been considerably recent molecular phylogeny (since the mid 1990?s) that has reasoned successfully for grouping these three subfamilies as a monophyletic family (Doyle et al., 2000; Kajita et al., 2001). Yet the placement of several subfamilies is still unresolved (e.g., Cercideae and Detarieae). Evidence of legume nodulation is lacking in the Cassieae sub-family. This includes Detarieae and Cercideae (Sprent, 2006). Trifolium is, by broad taxonomic means, most closely related to other genera within first, its sub-tribe (Trifolieae, somewhat synonymous with Vicioid clade), secondly, its subfamily (Papilionoideae, sometimes called Faboideae), and more generally, within its family (Fabaceae). 7 For brevity, I mention genera that fall within Trifolieae. These include, but are not necessarily limited to: Medicago, Melilotus, Ononis, Parochetus, and Trigonella. This subject is reviewed in-depth within Ellison et al. (2006). Currently accepted taxonomy is shown within the Inset of Figure 4. A phylogenetic approach places Trifolium most closely linked to Trigonella and Melilotus. The genus Melilotus includes many important plants. Yellow sweet clover (Melilotus officinalis) or alfalfa plants within this genus are best known for use in forage production. Like many members of this family, they are a source of nectar for honey bees (USDA, n.d.).There are more than 30 species recognized within the genus Trigonella. Of note is fenugreek. T. foenum- graecum is both an herb and spice, often found in Indian and South Asian cuisine (Katzer, n.d.; USDA, n.d.). It is also a source of animal fodder (USDA, n.d.). Origins and Evolution Estimates place Fabaceae diversification within the Early Tertiary, approximately 60 Mya (Herendeen et al., 1992). The fossil record of Fabaceae is abundant and diverse, according to Wojciechowski et al. (2000). It includes many fossil legume fruits and flowers as well as early indications of nodulation and symbiotic relationships. Legumes first appeared during the late Paleocene, circa 56 Mya (Herendeen, 2001; Herendeen and Wing, 2001; Wing et al., 2004). Diversification into the currently accepted subfamilies, Caesalpiniods, Mimosoids, and Papilionoids began around 50 to 55 Mya (Herendeen et al., 1992). It is interesting to note that a diverse assemblage of taxa were located upon the Mississippi Embayment of North America during the upper Eocene (55 to 34 Mya during the emergence of modern mammals; Herendeen et al., 1992). 8 Lavin et al. (2005) and Schrire et al. (2005) suggests we consider the diversification of Fabaceae in terms of the success of biomes rather than geographic regions, as the North Atlantic land bridge would have been aiding in the trans continent dispersal of early legumes. There are an abundance of trans continent disjunctions within Fabaceae, most of them no older than 22 Mya (Schrire et al. 2005). Fossil evidence indicates diversification of Papilionoids 59 to 39 Mya (Lavin et al., 2005). Based upon multiple sources the Papilionoideae can be divided into several major clades. Hologalegina is the name given to the largest of these well-supported major clades, which, based upon its center of diversity, originated in Eurasia. Wojciechowski et al. (2000) suggests Hologalegina, the major clade containing Trifolium, originated approximately 50 Mya. Hologalegina lacks an early Eocene fossil record; though, its origin is estimate at 51 Mya (Lavin et al., 2005). Based upon phylogenetic analyses, strong evidence has emerged for two subclades of the Hologalegina - that of the Robinioids and the Inverted Repeat-lacking clade (IRLC). Unlike the Robinioid clade, which contains species such as birdsfoot trefoil (Lotus corniculatus), the IRLC lacks one copy of a large inverted repeat (25 kb) that encodes a duplicate set of ribosomal RNA genes. This mutation is remarkable for its rarity - with few exceptions, it is conserved throughout green algae and land plants (Palmer et al. 1988). The IRLC centers of greatest species diversity lie within Eurasia and Northern America (Polhill and Raven, 1981; Polhill, 1994). The clade includes many economically important crops, such as alfalfa (Medicago sativa), garden pea (Pisium sativum), and the genus Trifolium. The IRLC contains several yet unresolved genera, including Afgekia, Calerya, Wisteria, and Glycyrrhiza, as well as three well-supported sub-clades, including the Hedysaroid, Galegeae 9 and Vicioid. The latter contains many of the agriculturally important crops mentioned previously, most notably Trifolium. Trifolium is estimated to have originated from other Fabaceae in the Early Miocene, 16 to 23 Mya (Lavin et al., 2005; Ellison et al., 2006). Current bioinformatics such as mapping of chloroplast DNA and known species diversity lead to a fairly well substantiated center of origin within the Mediterranean basin. Ellison et al. (2006) go into incredible detail during their review of clover phylogenetics. A summary of which is that the dispersal of Trifolium species has led to more than 275 individual species, many of them considered native to North and South America. Rather than displaying a mix of lineages, these groups can be distinguished by their monophyletic qualities, meaning that they are singly cladistic in origin. Numerous accounts of hybridization and reticulate evolution are present. Ellison et al. (2006) highlight the likely introgression of cytoplasmic chloroplast DNA between T. campestre and T. dubium (two very similar hop-clovers endemic to the southeastern U.S.). They also attempt to identify the origins of T. repens (white clover) and come short of concluding that T. occidentale and T. pallescens are likely its diploid progenitors. Nitrogen Fixation Estimates place Fabaceae diversification approximately 60 Mya. (Herendeen et al., 1992). All nitrogen-fixing, flowering plants fall within the Eurosid clade. Scattered throughout this clade are numerous plants that nodulate with filamentous bacterium, such as Frankia. More confined, however, are the plant species that nodulate with unicellular rhizobia. The evolution of symbiotic soil-borne bacteria has paralleled the origins of modern legumes. Early legume nodulation occurred roughly 58 Mya and were inviting habitats for soil-borne life to develop. Many single-celled organisms may not have been beneficial to early plants. Rather, legume- 10 rhizobia relationships developed gradually and resulted in highly specific pairings. In fact, they only occur within the order Fabales. Furthermore, with only one exception, Ulmaceae, these plants fall within the family of Fabaceae (Soltis et al., 2000). Legume root nodulation occurs due to at least three known genera of ?rhizo-bacteria?, including: Rhizobium, Bradyrhizobium, and Azorhizobium. Certain anatomical features typify legume root nodulation, primarily induction of a new plant meristem that develops as an invitation to (or possibly as a result of interspecific signaling from) bacterial infection (Rolfe and Gresshoff, 1988; Schultze and Kondorsori, 1998). This is induced by the initial colonization of the root surface by the bacteria. Prior to infection, lectin-receptors on the host plant must specifically recognize the potential pathogen. Coordination and communication between the symbiotes is required throughout the initial and subsequent stages of infection and are highly specific for both host and pathogen. Rhizobia near the surface of host plants respond to flavonoids, such as luteolin, by expressing nod genes (Brewin, 1991; Schultze and Kondorsori, 1998). Expression leads to the production of return signals, sometimes called Nod -signals or -factors (Schultze and Kondorsori, 1998). For Trifolium spp. these Nod factors are lipochito-oligosaccharides and are specific to Rhizobium species. They initiate root-hair curling and consequently nodule primordia (Figures 5 and 6). Subsequently invasion occurs as the infection thread penetrates the epidermis then moves into the inner cortex. The spread of infection between cells is aided by regular planes of cell division in young meristamatic tissue as well as pre-infection orientation (reviewed by Brewin, 1991; and Buchannan et al. 2000). During this infection, bacteria are engulfed by plant cells forming organelle-like structures that some have termed symbiosomes (Roth and Stacey, 1989; Buchannan et al., 2000). 11 Nodules are populated by roughly dozens of bacteria, which in many cases cease to propagate after two or three rounds. Proteins involved in transport of substrate, as well as metabolism of carbon and nitrogen, are manufactured within these structures. These furnish the machinery necessary to ?fix? atmospheric nitrogen and share that nitrogen with the associated plants as NH 3 . The primary structure responsible for nitrogen fixation is nitrogenase. Nitrogenase enzyme is actually two separate protein structures ? dinitrogenase and dinitrogenase reductase. Dinitrogenase binds N 2 while dinitrogenase reductase provides electrons to reduce N 2 resulting in 2 NH 3 molecules. An important note about dinitrogenase reductase: it not only reduces N but also reduces acetylene to ethylene. This provides a useful assay to assess nitrogenase activity. Nitrogenase activity is inhibited at oxygen concentrations greater than roughly 1% (Brewin, 1991). Therefore, it is important, that uninfected parenchyma cells function as barriers to oxygen. In addition, leguminous plants produce the oxygen-binding protein leghemoglobin, which serves to reduce oxygen concentrations near the site of nitrogenase activity (Buchannan et al., 2000). As an exchange for the nitrogen fixed by rhizobia, plant hosts provide photosynthate. This carbon source enters nodules as sucrose. Evidence suggests that mono- and di-saccharides are not directly transported into bacterioids; rather, the sugars are converted into dicarboxylic acids such as malate and oxaloacetate via a process similar to fermentation (Buchannan et al., 2000). Nitrogen Fixation within Mixed Grass Clover Swards Even in persistent stands of legumes biological N fixation varies, largely due to the relative composition of turfgrass swards and soil N availability (Crush et al., 1982). Fixation is 12 highly dependent upon the relative level of nodulation occurring in root tissues and activity of the bacteria within. Most research indicates that high soil N concentrations inhibit nodule growth and development. Macduff et al. (1996) observed that the ratio of root to nodule dry-weights was 6:1 in white clover without NO 3 treatment but increased with applications of NO 3 . It is well documented that increasing N fertilization decreases clover density and allows the grass portion of the sward to outcompete clover (Frame and Boyd 1987; Pederson 1995; Sincik and Acikgoz 2007). Other factors affecting biological N fixation include absorption of photosynthetically active radiation, C- assimilation rates, and allocation of photosynthate to roots (Lie 1971). White clover leaves have a higher photosynthetic capacity at low N levels than do competing perennial ryegrass; however, at higher N levels the opposite is true (Faurie et al. 1996). Increased light interception at low N levels can be attributed to a greater leaf area index in the upper canopy of the grass-clover sward as well as clover?s ability to avoid shade by increasing petiole length (Davies and Evans 1990; Faurie et al. 1996; Woledge et al. 1992). White clover persistence varies greatly due to soil conditions. In their review of N fixation of grass-legume pastures, Ledgard and Steele (1992) report that fixation is greatly reduced due to dry soil conditions, acid soils, and the ?pest/disease complex.? Another major factor affecting N fixation is soil temperature. Frame and Newbould (1986) found that a minimum temperature of 9?C was necessary for active N fixation by Rhizobium. It has also been reported that temperatures for nitrogenase activity range from 13 to 26?C (Halliday and Pate 1976). 13 Decomposition of Clover Biomass within Turf There is no doubt that root nodule decomposition is a significant source of N for associated grasses, as reported N concentration of root nodules ranges from 4.8 to 9.0% of root dry matter (Chu and Robertson, 1974; Wardle and Greenfield, 1991). However, root nodules are not the sole source of N transfer, as above ground white clover dry matter has been reported to be 9.1 to 24.2% protein, depending upon harvest date (about 1.5 to 4.0% N; Burton and DeVane, 1992). Unlike forage scenarios, turfgrass systems differ in that they are not grazed; rather, they are mown frequently to maintain utility and aesthetics. Mown clippings are returned to the turf surface, potentially contributing a mineralizable source of N. Polyculture lawns of grass and white clover are historically common, yet little is known of the N-contribution and C-flux from decaying clover foliage. The rates of decomposition, N mineralization, and C deposition would be useful information for future research regarding this subject as well as when assigning nutrient credits to white clover-culture in warm and cool season turfgrass. Such information would be highly dependent upon a multitude of factors, including time of year, litter composition, soil and climactic -conditions, as well as soil fauna. For these reasons, it may not be possible to control all factors in situ. Organic residues decompose in two phases. Soil microbes rapidly consume the labile fraction, which is composed of sugars, starches, and proteins, leaving behind a recalcitrant fraction composed of cellulose, fats, waxes, lignin, and tannins (Wieder and Lang, 1982). This slowly decomposing fraction helps to develop soil organic matter. Due to the two-step nature of decomposition, a double exponential decay model is often implemented to describe litter decay (Wieder and Lang, 1982). Double exponential decay equations are of the form Y = Ae -k1t + Be - k2t , where Y = response, A and B are initial concentrations approximating the labile and 14 recalcitrant portions, k1 and k2 are rate constants fitted to the data, and t equals time in days after application (DAA). Such models have been used successfully to describe quickly decaying legume litter in Alabama (Mulvaney et al., 2010) as well as the decomposition and N release of hedgerow species in Haiti (Isaac et al., 2000). Modeling white clover decomposition may enable turfgrass researchers and professionals to more accurately predict nutrient contribution to associated grasses and help optimize supplemental fertilizer recommendations. Clover Establishment Legumes such as clovers are present within many turfgrass scenarios in the temperate climes of the southeastern United States as both weeds and amenity plants. As amenity plants, clover species may provide important ecosystem services, such as nitrogen fixation (Ledgard and Steele 1992, McNeil and Wood 1990, Whitehead 1995) and insect habitat (Abraham et al. 2010, Rogers and Potter 2004). Clover has been, and continues to be, included in grass mixtures for roadsides as well as other maintained turfgrass areas and has proven useful for slope stabilization (Roberts and Bradshaw 1985). In particular, white clover (Trifolium repens L.) thrives within home lawns and golf courses because it can flower and produce seed at mowing heights as low as 6 mm (Sincik, and Acikgoz 2007; Watschke et al. 1995). Other clover species are also common within maintained turf swards. Prominent amongst Auburn, AL flora are small hop clover (T. dubium Sibth.), crimson clover (T. incarnatum L.), and ball clover (T. nigrescens Viv.). Little has been written about the establishment and maintenance of mixed turf-clover swards; though, similarities can be drawn between those of mixed grass-legume forage systems as well as examples provided by overseeding practices common within the transition zone, where warm and cool season grasses grow equally as well. 15 Proper white clover establishment is key to maximizing stand uniformity as well as N contribution to associated grasses (Frame and Newbould, 1986). However, there are currently no guidelines for establishment within warm-season turfgrass scenarios common to the southeastern U.S. Furthermore, unlike pasture systems, managed turfgrass scenarios may offer unique opportunities to manipulate turfgrass height and density, as well as soil characteristics, in favor of white clover establishment. There are several agronomic practices used to improve overseeded grass establishment within maintained turf scenarios. Scalping is among the most common techniques and refers to the excessive removal of living tissue at any one mowing occurrence (Turgeon, 2002). Though scalping often results in turfgrass injury, it is a means of exposing bare soil and eliminating turfgrass competition, which is essential to overseeded grass establishment. Verticutting, or vertical mowing, is another mechanical method often used to remove accumulated thatch or to elevate decumbent turfgrass prior to overseeding. Verticutting is performed by passing a rapidly rotating horizontal shaft with vertically oriented knives over affected turfgrass (Turgeon, 2002). Vertical mowing is often used in addition to scalping in order to prepare warm-season turfgrass for overseeding. Hollow tine aerification is less commonly used for fall overseeding but is an agronomic practice used to improve soil characteristics by removing cores of soil from turfgrass. Core sizes may vary, but the desired result is much the same. That is, the cores are removed to alleviate compaction by decreasing soil bulk density, accelerate drying, and increase infiltration of water and gasses. Once performed, cores are often collected or scattered, and the remaining holes are either filled with sand or left open. Hypothetically, scalping alone or scalping in combination with vertical mowing and aerification may be a means of improving seeded white clover establishment, via improved seed- 16 to-soil contact, and by limiting competition effects from associated turfgrasses. Soil aerification may also alleviate competition but has the added benefit of providing holes in which white clover may find more adequate soil conditions for initial establishment. It is therefore reasonable that it too should be tested as a means of improving white clover establishment. Other variables that affect white clover establishment are establishment timing and seeding rate. Recommended establishment dates for white clover in the southeastern U.S. are largely anecdotal. For instance, establishment timing is often recommend from 2 to 6 weeks prior to historical first frost (approximately November 1 st in Auburn, AL). Previous research in Florida recommends September planting dates (Dudeck and Peacock, 1983), while others have recommended spring seeding to avoid hard freeze in more northern climates (Frame and Newbould, 1986). These dates are highly variable and dependent upon locations and climate. Further, they may not account for nuances of a maintained turf sward, which may insulate young white clover seedlings from effects of frost or hard freeze. Anecdotal to our own research, proper stand density is highly dependent upon seeding rate, yet it does not appear to be a linear response, perhaps due to intra-species competition. White clover establishment within cool-season grass swards has largely been dictated by seed mixtures of cool-season grass blends containing roughly 3 to 10% white clover by weight (Sincik and Acikgoz, 2007). Yet, these rates have not been evaluated in existing warm season turf swards. Likewise, information about interaction effects of white clover and companion grass species is absent from the scientific literature. Alternative, grass-white clover mixtures for turfgrass are commercially available in much of Europe and the United States; however, they have not been evaluated for winter overseeding of dormant warm-season grasses. 17 Weed Control in Mixed Grass-Clover Swards Like forage systems, broadleaf weeds and sedges are problematic during sward establishment. Herbicidal weed control is often required as seedling legumes are not competitive with many weeds and grasses (Carlisle et al. 1980; Evers et al. 1993; Young et al. 1992). In addition to competition effects, weedy species often negatively affect aesthetics and reduce the utility of certain mixed swards. Weed control within turf-clover swards is often hampered by the lack of effective herbicides that are safe on clovers. Few herbicides are labeled for postemergence application to various clover species, and most are restricted to states where the respective species are grown for seed production. Furthermore, differential tolerance of legume and cultivars within species to common row-crop and pasture herbicides has previously been reported (Beran et al. 1999; Bowran 1993; Young et al. 1992). These results have shown that individual species exhibit different reactions to various broad-leaf herbicides, including differential reductions in seed yield, biomass, and nitrogen input for subsequent crops (Bowran 1993). For these reasons, research is needed to evaluate differential herbicide responses of Trifolium species, which are commonly included within mixed turf scenarios. It is also increasingly important to identify common turf herbicides that are tolerated legume plants of biodiverse swards. 18 Kingdom: Plantae ? Plants Subkingdom: Tracheobionta ? Vascular plants Superdivision: Spermatophyta ? Seed plants Division: Magnoliophyta ? Flowering plants Class: Magnoliopsida ? Dicotyledons Subclass: Rosidae (Figure 2) Order: Fabales Family: Fabaceae ? Pea family Subfamily: Papilionoideae Tribe: Trifolieae Genus: Trifolium L. ? Clover Figure 1. Trifolium Taxonomy (Plants Database, 2012) 19 Figure 2. Rosids (or Eurosids) clade according to Wang et al. (2009) and Worberg et al. (2009). 20 Figure 3. Papilionoideae (Wojciechowski et al., 2000) is the largest of the three subfamilies of Fabaceae. 21 Figure 4. The inverted repeat-lacking clade (Wojciechowski et al., 2000) is distinguished from the Robinioid clade due to its loss of one copy of a large inverted repeat. These plants are predominantly herbaceous annuals and perennials that typicaly have compound leaves. Inset. The position of Trifolium amongst the genre of the Vicioid clade. The area of the triangles is proportional to the number of species in each subgenus. Bayesian posterior probabilities are below branches; Parsimony bootstrap values are above. Values below 0.50 (or 50%) are not shown. 22 Figure 5. Legume rot nodulation proces (based upon Buchanan et al., 200). 1. Plant rots release elicitors of Nod gene expresion 2. Bacterium releases Nod factor. 3. Plant rot is infected and undergoes nodule morphogenesis. Figure 6. Root hair curling and invasion. 1. Nod factors initiate rot-hair curling. 2. Invasion ocurs as an infection thread penetrates the epidermis then the iner cortex. Figure 7. Nodule meristem and zones of infection. 1. Nodule meristem 2. Zone of infection thread growth and cel penetration 3. Zones of expanding infected cels 4. Mature bacteroid-containing tisue 5. Senesecent bacteroid-containing tisue 6. Outer cortex 7. Nodule endodermis. 8. Iner cortex 9. Nodule vascular bundle. 10. Rot epidermis 1. Rot cortex 12. Rot endodermis 13. Rot xylem and phloem elements. 1 2 3 1 2 1 2 3 4 5 6 7 8 9 8 10 11 12 13 23 White Clover (Trifolium repens) Establishment within Dormant Bermudagrass (Cynodon dactylon) Turf INTRODUCTION Benefits of turf are well documented and include: recreational health, erosion control, increased water infiltration, reduced nutrient leaching, aesthetics, carbon (C) sequestration, and mediation of the ?heat-island? effect (Beard and Green, 1994; Qian and Follett, 2002). Yet the ecological impact of turf is often questioned, due in part to nutrient and water requirements (Milesi et al., 2005; Robbins et al., 2001; Robbins and Birkenholtz, 2003) as well as often- unsustainable monoculture cultivation, which contributes to insect habitat -loss and - fragmentation (Gels et al., 2002). For these reasons, the turfgrass industry is experiencing new demands for ecologically and economically -sustainable maintenance options. Inclusion of leguminous species, which biologically fix N and provide pollinator habitat, is a proposed means of increasing the sustainability of certain low maintenance turfgrass scenarios. However, little is known about inclusion of legumes in maintained turfgrass. Since the advent of herbicides, efforts in the turfgrass industry have often focused on maintaining monocultures for aesthetics and increased playability. Thus, a biologically diverse turfgrass sward with mixed species of grasses and broadleaf plants is sometimes classified as weedy and therefore undesirable for scenarios such as golf-course and sports pitch. However, for many scenarios, such as home lawns, roadsides, or other ?unimproved? turfgrass areas, the environmental benefits of biodiversity may outweigh those of monoculture. White clover inclusion within maintained turfgrass has mainly been limited to cool-season turfgrass scenarios. Important research by Sincik and Acikgoz (2007) reported increased color 24 ratings in three cool-season turfgrass-white clover (T. repens L.) mixtures and that white clover fixed greater than 25 g N m -2 year -1 and contributed between 4.2 to 13.7% of that total N to the associated turfgrass. Additional information concerning white clover inclusion within maintained turf is absent. However, information about the benefits of white clover inclusion within pasture systems is fairly abundant but mainly focuses on perennial ryegrass (Lolium perenne L.) -white clover pastures. These mixed systems supply high-quality grazing for animals while simultaneously improving soil fertility (Lampkin, 2002). Estimates of N fixation for grass-white clover pastures range from nil to 40 g N m -2 year -1 , though most are roughly 10 to 25 g N m -2 year -1 (Ledgard and Steele, 1992; McNeil and Wood, 1990). Yet white clover is well suited for use within warm-season turfgrasses and is already a common feature within bermudagrass pastures of the southeastern U.S. (Brink and Fairbrother, 1991). Proper white clover establishment is key to maximizing stand uniformity as well as N contribution to associated grasses (Frame and Newbould, 1986). However, there are currently no guidelines for establishment within warm-season turfgrass scenarios common to the southeastern U.S. Furthermore, unlike pasture systems, managed turfgrass scenarios may offer unique opportunities to manipulate turfgrass height and density, as well as soil characteristics, in favor of white clover establishment. Our objectives were to test standard overseeding methods, cultural practices, seeding rates, and companion grass combinations for their effects upon spring white clover establishment within a maintained bermudagrass lawn. We hypothesized that white clover establishment is comparable to overseeding dormant warm-season turfgrass with cool-season grasses such as perennial ryegrass. However, unlike perennial ryegrass, recommended white clover establishment rates are much lower [from 3 to 5 kg white clover seed ha -1 recommended by Frame and Newbould (1986)]. 25 There are several agronomic practices used to improve overseeded grass establishment within maintained turf scenarios. Scalping is among the most common techniques and refers to the excessive removal of living tissue at any one mowing occurrence (Turgeon, 2002). Though scalping often results in turfgrass injury, it is a means of exposing bare soil and eliminating turfgrass competition, which is essential to overseeded grass establishment. Verticutting, or vertical mowing, is another mechanical method often used to remove accumulated thatch or to elevate decumbent turfgrass prior to overseeding. Verticutting is performed by passing a rapidly rotating horizontal shaft with vertically oriented knives over affected turfgrass (Turgeon, 2002). Vertical mowing is often used in addition to scalping in order to prepare warm-season turfgrass for overseeding. Hollow tine aerification is less commonly used for fall overseeding but is an agronomic practice used to improve soil characteristics by removing cores of soil from turfgrass. Core sizes may vary, but the desired result is much the same. That is, the cores are removed to alleviate compaction by decreasing soil bulk density, accelerate drying, and increase infiltration of water and gasses. Once performed, cores are often collected or scattered, and the remaining holes are either filled with sand or left open. Hypothetically, scalping alone or scalping in combination with vertical mowing and aerification may be a means of improving seeded white clover establishment, via improved seed- to-soil contact, and by limiting competition effects from associated turfgrasses. Soil aerification may also alleviate competition but has the added benefit of providing holes in which white clover may find more adequate soil conditions for initial establishment. It is therefore reasonable that it too should be tested as a means of improving white clover establishment. Other variables that affect white clover establishment are establishment timing and seeding rate. Recommended establishment dates for white clover in the southeastern U.S. are largely anecdotal. For instance, establishment timing is often recommend from 2 to 6 weeks prior 26 to historical first frost (November 1 st in Auburn, AL). Previous research in Florida recommends September planting dates (Dudeck and Peacock, 1983), while others have recommended spring seeding to avoid hard freeze in more northern climates (Farme and Newbould, 1986). These dates are highly variable and dependent upon locations and climate. Further, they may not account for nuances of a maintained turf sward, which may insulate young white clover seedlings from effects of frost or hard freeze. Anecdotal to our own research, proper stand density is highly dependent upon seeding rate, yet it does not appear to be a linear response, perhaps due to intra- species competition. White clover establishment within cool-season grass swards has largely been dictated by seed mixtures of cool-season grass blends containing roughly 3 to 10% white clover by weight (Sincik and Acikgoz, 2007). Yet, these rates have not been evaluated in existing warm season turf swards. Likewise, information about interaction effects of white clover and companion grass species is absent from the scientific literature. Alternative, grass-white clover mixtures for turfgrass are commercially available in much of Europe and the United States; however, they have not been evaluated for winter overseeding of dormant warm-season grasses. Due to the many knowledge gaps limiting the utility of white clover inclusion within warm-season scenarios, experiments were conducted to test the effects of pre-seeding mechanical surface disruption, establishment timing, seeding rate, and companion grass species on establishment of two commercially available white clover populations within dormant bermudagrass turfgrass. White clover was chosen as a model species for a variety of reasons, but specifically because turf-compatible white clover varieties are commercially available, and due to white clover prevalence in maintained turfgrass as a weed species (Watschke et al. 1995). Here we present results that may influence future scientific studies and the utility of white clover inclusion within warm and cool season turf scenarios. 27 MATERIALS AND METHODS Studies were designed as randomized complete blocks with four replications. Blocking considerations were mowing direction and return of clippings. Studies were initiated 14 October 2010 and 1 October 2011 at the Auburn University Turfgrass Research Unit (32?34?40? N, 85?29?57? W; elevation 185 m) in Auburn, AL. Research was conducted within a maintained ?Tifway? hybrid bermudagrass [Cynodon dactylon (L.) Pers. ! C. transvaalensis Burtt Davy] lawn on a Marvyn sandy loam (fine-loamy, kaolinitic, thermic Typic Kanhapludult) soil with an average pH of 6.3 (1:1 soil:H 2 O). Turfgrass was maintained at a height of 3.8 cm; all clippings were returned to the turfgrass surface. Plots received 3 cm supplemental irrigation on a weekly basis between March and September of 2011 and 2012. The area was fertilized (5 g N m -2 ) 15 February 2011 and 20 February 2012. Four studies were conducted to evaluate the effects of pre-seeding mechanical surface disruption, establishment timing, seeding rate, and companion grass species on establishment of two commercially available white clover populations, Dutch White (Main Street Seed and Supply, Bay City, MI) and DLF Microclover (DLF-International Seeds, Halsey, Oregon). Seed were drop seeded through a stainless steel device, which contains 5 seed dispersion screens (6.4 mm 2 mesh openings) oriented horizontally to evenly scatter small grass and broadleaf seeds. With the exception of the seeding rate study, all clover were seeded at 1.5 g live seed m -2 . Trifoliate leaves were counted within three 730 cm 2 sub-samples per 1.0 m 2 experimental unit on 20 April, 2011 and 2012 as a means of quantifying spring clover density (trifoliate leaves m -2 ). Companion grass plants were quantified using similar subsampling methods during January of 2011 and 2012 when bermudagrass was completely dormant. Mechanical Disruption Study. 28 This study evaluated common pre-overseeding cultural practices, such as verticutting, aerification, as well as scalping, and their ability to enhance seeded clover establishment relative to normally mown, non-scalped turfgrass. Treatments were arranged as a factorial to test the effects of four common cultural practices upon the establishment of two commercially available clover populations. Treatments were intended to mechanically disrupt the soil surface as well as eliminate bermudagrass competition and included: scalping (6 mm mowing height), scalping plus vertical mowing (6 mm below soil level), and scalping plus hollow tine aerification (6 mm hollow tines; 3.8 cm depth; 15.2 cm spacing). Treatments also included a non-scalped control maintained at 3.8 cm mowing height. Clippings were removed from scalped surfaces, and clover was drop seeded as previously described. Timing Study Treatments were arranged as a factorial to test the effects of seeding time (October through February) upon establishment of two commercially available clover populations. Plots were scalped at each seeding date, as previously described, and were blown free of clippings. Clover was seeded at monthly intervals beginning in October and ending in February. Seeding Rate A seeding rate trial was arranged as a factorial to estimate the effects of seeding rate upon establishment of two commercially available clover populations. Plots were scalped and blown free of clippings. Clover was seeded at 0, 0.4, 0.8, 1.5, 3.0, and 6.0 g live seed m -2 . Companion Grasses Treatments were arranged as a factorial to test the effects of seeding companion grass species in combination with one of two commercially available clover populations. Companion species were: annual ryegrass (Lolium multiflorum Lam.) perennial ryegrass (L. perenne L.), 29 creeping bentgrass (Agrostis stolonifera L), red fescue (Festuca rubra), and Poa trivialis L. (See Table 1 for rates and sources). Statistical Analysis All data were subject to analysis of variance (ANOVA) within SAS procedure GLIMMIX using mixed model methodology (SAS ? Institute v. 9.2, Cary, North Carolina, USA). Treatment was considered a fixed effect in the model. Year, replication (nested within year), and iterations containing these effects were considered random in the model (Carmer et al. 1989; Hager et al. 2003). Basic model assumptions were confirmed. Means were separated based upon adjusted 95% confidence intervals, which allows for multiple comparisons by protecting family-wise error rate (Littell et al. 2006). Least squares estimates for linear models were determined for rate-response studies using the Marquardt-Levenberg algorithm to provide the best fit (SPSS Inc., Sigma Plot v. 11.2, Chicago, Illinois, USA). R 2 values were used to determine ?goodness of fit? for the selected equations. Initial parameter ranges were selected with a maximum of 200 fits and 200 iterations. The relationship of clover density (trifoliate leaves m -2 ) to the clover seeding rates investigated in this trial were described using the linear model y = y 0 + ax b , where y equals trifoliate leaves m -2 , y 0 equals the y-intercept (held constant at 0), a serves as a scaling factor (moving the values of x b up or down), x is equal to initial seeding rate (g live seed m -2 ), and b is the scaling exponent that determines the function?s rate of growth or decay. Correlation between companion grass density and clover establishment were described using Pearson product moment within SigmaPlot 11.2. RESULTS Analysis of variance (Table 2) indicated that results for all studies differed due to replication year. For this reason, 2010-2011 and 2011-2012 (season 1 and 2, respectively) results are presented separately for all studies. However, with few exceptions, treatment separations were 30 similar across years and are used in support of our main conclusions. It is possible that the earlier initiation date of season 2 (October 1 rather than October 14) had some affect upon clover establishment, as bermudagrass dormancy was much more delayed during season 2 relative to season 1. Mechanical canopy disruption methods !"#$%&#?(")"?*"+"),%%-?#./.%,)?,0)1##?#",#1+#2?(.&3?#0,%4.+*?,%1+"?,+5?.+? 01/6.+,&.1+?(.&3?1&3")?/"&315#?3,7.+*?"+3,+0"5?#4).+*?(3.&"?0%17")?"#&,6%.#3/"+&?89,6%"? :;?@?2?#0,%4.+*?,%1+"?,+5?.+?01/6.+,&.1+?(.&3?7")&.0$&&.+*?1)? ,").A.0,&.1+?-."%5"5?BCC2?D?>2?,+5?D@:?&).A1%.,&"?%",7"#?/ E> 2?)"#4"0&.7"%-HC?&).A1%.,&"?%",7"#?/ E> ;?"#&,6%.#3/"+&?5"+#.&."#?(")"?*"+"),%%-?%1(")?&3,+?&31#"?1A?#",#1+?@?89,6%"? :;?B?&).A1%.,&"?%",7"#?/ E> ;?&3,+?&31#"?1A?+1)/,%%-?/,.+&,.+"5?&$)A*),##?8K?&).A1%.,&"?%",7"#?/ E > ;?,+5?#0,%4.+*?,%1+"?&)",&/"+&#?8H:?&).A1%.,&"?%",7"#?/ E> ; 2?(3.03?(,#?*)",&")?&3,+?+1)/,%%-?/,.+&,.+"5? &$)A*),##2?(.&3?&3"?L$&03?7,)."&-?3,7.+*? *)",&")?#4).+*?(3.&"?0%17")?5"+#.&-?&3,+?M.0)10%17")?8@:??,+5?GK?&).A1%.,&"?%",7"#?/ E> 2? )"#4"0&.7"%-; ; ? )"%,&.7"?&1?,%%?1&3")?#""5.+*?5,&"#?89,6%"?B;,7?9+?=#&!/!"8.#?112#@A2#+":#B#&?!()*!+&,#*,+-,.# / 01 2#?,.C,5&!-,*=DE#F+?!,&+*#:!((,?,"5,.#3,?,#:,&,5&,:#:9?!"8#.,+.)"#@2#3!&4#&4,#;9&54# -+?!,&=#4+-!"8#=!,*:,:#",+?*=#:)97*,#&4,#34!&,#5*)-,?#:,".!&=#)(#G!5?)5*)-,?#+5?)..# ,.&+7*!.4/,"&#&!/!"8.#?%H#-,?.9.#BB#&?!()*!+&,#*,+-,.#/ 01 2#?,.C,5&!-,*=DE### I,+.)"#10.C?!"8#34!&,#5*)-,?#,.&+7*!.4/,"&#:!((,?,:#:9,#&)#&!/!"8#7=#-+?!,&=# !"&,?+5&!)"#?!"J#HEHBH@DE#K4!.#!"&,?+5&!)"#3+.#*+?8,*=#:9,#&)#,L+88,?+&,:#-+?!,&+*# :!((,?,"5,.#)(#&4,#;,5,/7,?#&!/!"82#C,?4+C.#:9,#&)#:,*+=,:#.,,:*!"8#,/,?8,"5,#)(# G!5?)5*)-,?#?;+&+#")&#.4)3"DE#M4,"#,.&+7*!.4,:#!"#;,5,/7,?2#&4,#;9&54#-+?!,&=#=!,*:,:# @$$#&?!()*!+&,#*,+-,.#/ 01 2#34,?,+.#G!5?)5*)-,?#=!,*:,:#)"*=#NH#&?!()*!+&,#*,+-,.#/ 01 E#I!/!*+?# &?,":.#3,?,#.,,"#&4?)984)9&#&4,#,LC,?!/,"&O#4)3,-,?2#")#)&4,?#,.&+7*!.4/,"&#&!/!"8# =!,*:,:#.!8"!(!5+"&#-+?!,&+*#:!((,?,"5,.E#K4,.,#&?,":.#+?,#+8+!"#/+"!(,.=#-+?!,&+*#/+!"# ,((,5&.#?;+&+#")&#.4)3"D2#3!&4#&4,#;9&54#-+?!,&=#4+-!"8#=!,*:,:#/)?,#&4+"#:)97*,#&4,#34!&,# 5*)-,?#:,".!&=#)(#G!5?)5*)-,?#+5?)..#,.&+7*!.4/,"&#&!/!"8.#?PH#-,?.9.#1P#&?!()*!+&,#*,+-,.# / 01 2#?,.C,5&!-,*=DE### Q.&+7*!.4/,"&#&!/!"8#/+!"#,((,5&.#3,?,#+8+!"#,-!:,"&#:9?!"8#.,+.)"#1#?K+7*,#BDE# <+"9+?=#+":#>,7?9+?=#&!/!"8#5*,+?*=#:!/!"!.4,:#.C?!"8#34!&,#5*)-,?#,.&+7*!.4/,"&#*,-,*.2# 3!&4#>,7?9+?=#&!/!"8#4+-!"8#=!,*:,:#)"*=#P#&?!()*!+&,#*,+-,.#/ 01 E#R)3,-,?2#9"*!S,#&4,# C?,-!)9.#.,+.)"2#65&)7,?#&!/!"8#:!:#")&#?,.9*&#!"#.9C,?!)?#.C?!"8#34!&,#5*)-,?#:,".!&!,.#?N@# &?!()*!+&,#*,+-,.#/ 01 D#+":#3+.#!"#(+5&#,T9+*#&)#*,-,*.#(?)/#U)-,/7,?2#<+"9+?=#+":#>,7?9+?=# &!/!"8.#?P@2#VP2#+":#P#&?!()*!+&,#*,+-,.#/ 01 2#?,.C,5&!-,*=DE#W+&4,?2#;,5,/7,?#&!/!"8#=!,*:,:# &4,#4!84,.&#.C?!"8#34!&,#5*)-,?#:,".!&=#?@H%#&?!()*!+&,#*,+-,.#/ 01 D2#34!54#3+.#,T9+*,:#)"*=# 7=#U)-,/7,?#,.&+7*!.4/,"&#?P@#&?!()*!+&,#*,+-,.#/ 01 DE## 32 !""#$%&?()*"+? As anticipated, April observed white clover densities increased proportionally to October seeding rate (Figure 8). For this reason, data were fit to the linear model y = y 0 + ax b , where y equals trifoliate leaves m -2 and x is equal to initial seeding rate (g live seed m -2 ). An important feature of this model is the diminished response of increasing seed yield. This characteristic highlights an important feature of white clover overseeding. That is, as white clover-seeding rate increases beyond a certain point, competition effects may begin to reduce yield response. We acknowledge that these functions do not account for seasonable variability. In fact, there are many variables that may affect white clover establishment, including soil and air temperature as well as moisture availability. Ideally, these equations could be used to estimate spring white clover densities and demonstrate the diminishing nature of seeded white clover yields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able 2. Analysis of variance (ANOVA) for 2010-2011 and 2011-2012 white clover establishment trials. Replication year was significant in all studies; therefore, analysis was performed separately. Study Effect F-value P > F F-value P > F 2010-2011 2011-2012 Mechanical disruption study Method 7.91 <0.0001 9.49 <0.0001 Variety 3.30 0.0721 4.00 0.0495 Method ! variety 2.18 0.0766 0.66 0.6191 Establishment timing Time 53.97 <0.0001 8.65 <0.0001 Variety 15.77 0.0001 24.53 <0.0001 Time ! variety 1.60 0.1808 2.66 0.0401 Seeding rate Rate 64.96 <0.0001 13.30 <0.0001 Variety 2.59 0.1111 1.51 0.2227 Rate ! variety 2.16 0.0789 0.31 0.8735 Companion grass Grass 32.37 <0.0001 7.85 <0.0001 Variety 3.43 0.0670 0.01 0.9390 Grass ! variety 0.84 0.5043 0.62 0.6491 All data were subject to analysis of variance (ANOVA) within SAS procedure GLIMMIX using mixed model methodology. Treatment was considered a fixed effect in the model. Year, replication (nested within year), and iterations containing these effects were considered random in the model 39 Table 3. April observed, spring white clover density as affected by mechanical surface disruption methods. Season Method Trifoliate leaves m -2 ? 95% CI 2010-11 Aerification 513 a a 75 Verticut 502 a 75 Scalp 499 a 75 Non-treated 279 b 77 2011-2012 Aerification 204 A 49 Verticut 108 AB 49 Scalp 73 BC 49 Non-treated 6 C 49 a Means were separated by 95% confidence intervals (CI). 40 Table 4. White clover density as affected by seeded establishment timing. Season Time Trifoliate leaves m -2 ? 95% CI 2010-11 October 199 a a 23 November 68 b 20 December 22 c 19 January 17 c 19 February 4 c 19 2011-2012 October 51 BC 27 November 91 AB 28 December 108 A 27 January 39 BC 27 February 9 C 27 a Means were separated by 95% confidence intervals (CI). 41 Table 5. Companion grass densities along side affected white clover densities. Year Grass Trifoliate leaves m -2 ? 95% CI Grass (plants m -2 ) ? 95% CI 2010-2011 Tall fescue 383 a a 47 2648 c 1410 Creeping bentgrass 262 b 42 3810 c 1410 Perennial ryegrass 165 c 46 9429 b 1410 Poa trivialis 109 cd 42 15113 a 1411 Annual ryegrass 68 d 42 11302 b 1410 2011-2012 Tall fescue 65 A 20 2454 C 1452 Creeping bentgrass 49 AB 20 6465 B 1452 Perennial ryegrass 11 BC 21 8491 B 1483 Poa trivialis 7 C 20 15007 A 1516 Annual ryegrass 5 C 20 8626 B 1483 a Means were separated by 95% confidence intervals (CI). 42 Figure 8. April observed white clover density as a function of five rates of October seeded white clover. Error bars represent 95% confidence intervals about the mean. 43 1 2 3 4 Figure 2. Correlation of December grass density and April white clover 5 density during seasons one and two of companion grass experiments, where r 6 = Pearson?s Correlation Coefficient, and P = probability that r is different 7 from 0. Correlations were only significant when white clover data were 8 combined across companion grass species. 9 10 T r i f ol i at e l e ave s m - 2 44 Differential Response of Four Trifolium Species to Common Broadleaf Herbicides: 11 Implications for Mixed Grass-Legume Swards. 12 13 INTRODUCTION 14 Clovers (Trifolium spp.) are routinely included within pastures and low-maintenance turf 15 as utility plants. These legumes provide important ecosystem services, such as nitrogen (N) 16 fixation (Ledgard and Steele 1992, McNeil and Wood 1990, Whitehead 1995) and insect habitat 17 (Abraham et al. 2010, Rogers and Potter 2004). Clovers, like many legumes, increase forage 18 yields and quality as well as decrease N fertilizer requirements (Hoveland 1989; Rao et al. 2007). 19 When included within low maintenance turf, clovers improve sward color by contributing N to 20 associated grasses (Sincik and Acikgoz 2007) and have proven useful for maintaining roadside 21 slopes maintained as turf (Roberts and Bradshaw 1985). 22 Herbicidal weed control is critical to maximizing forage yields (DiTomaso 2000; Seefeldt 23 et al. 2005) and is often required during clover establishment, as seedlings are not competitive 24 with many weeds and grasses (Carlisle et al. 1980; Evers et al. 1993; Young et al. 1992). Weeds 25 compete with desirable species for nutrients and resources and are often toxic to grazing animals 26 (Carlisle et al. 1980; Marten and Andersen 1975; Vengris et al. 1953). 27 Selective weed control in grass-clover swards is hampered by the lack of effective 28 herbicides that are tolerated by clovers. Many effective broadleaf herbicides are reported to 29 control clover, including 2,4-D, carfentrazone, clopyralid, dicamba, and triclopyr (MacRae et al. 30 2005, Neal 1990, Neal and Mascianica 1988, Willis et al. 2007). Yet few herbicides are labeled 31 for postemergence application to various clover species, and most are restricted to states where 32 clovers are cultivated for seed production or forage. 33 45 Furthermore, differential herbicide tolerance of legume -cultivars and -species has 34 previously been reported, including differential reductions in seed yield, biomass, and N input 35 for subsequent crops (Beran et al. 1999; Bowran 1993; Young et al. 1992). Understanding 36 differential herbicide treatment effects upon clover species may advance efforts for selective 37 weed control within grass-clover swards as well as increase clover control options within grass 38 monocultures. 39 Experiments were conducted to identify herbicides tolerated by utility clovers and to 40 evaluate the potential for differential clover response to common herbicide treatments. Due to 41 previous reports of differential herbicide tolerance amongst other legume species, researchers 42 postulated that clover response to herbicides would differ by species. Emphasis was placed upon 43 determining herbicide tolerance of four clover species endemic amongst the local flora, 44 including: white clover (T. repens L.), small hop clover (T. dubium Sibth.), crimson clover (T. 45 incarnatum L.), and ball clover (T. nigrescens Viv.). We report differential responses of these 46 species to a range of broadleaf herbicides. 47 MATERIALS AND METHODS 48 Field and greenhouse experiments were repeated for two years to evaluate clover 49 response to a range of common broadleaf herbicides. Field experiments were conducted during 50 2010 and 2011 at the Auburn University Turfgrass Research Unit (32?34?40? N, 85?29?57? W) 51 in Auburn, AL. 52 Cool-season legumes (Table 6) were collected to a depth of 7.6 cm using a 10.8 cm 53 diameter golf-green cup-cutter (Par Aide Product Company, Lino Lakes, Minnesota, USA) 54 between 19 to 22 January 2010 and 1 to 18 February 2011. Plants were collected at a single site 55 from a Marvyn sandy loam (fine-loamy, Kaolinitic, thermic Typic Kanhapludult) soil with pH 56 46 6.3 (1:1 soil:H 2 O) and were allowed to mature in a greenhouse setting until subject to selection 57 for uniform size and maturity. 58 Plants were transplanted into field conditions 10 February 2010 or 15 to 21 February 59 2011. The transplant site was a hybrid bermudagrass (Cynodon dactylon (L.) Pers. x C. 60 transvaalensis Burtt-Davy) sward maintained at 5 cm mowing height without supplemental 61 fertility. Soil at the transplant site was a Marvyn sandy loam soil similar to that found at the 62 collection site where plants originated. The site was not mown or fertilized during studies, but 63 was hand watered to prevent clover wilt. Plants were clipped with shears to identical height (8 64 cm) and diameter (11 cm) two days prior to treatment. Further information concerning collection 65 date, stage of growth, and transplant date is presented in Table 6. 66 The field study was conducted as a split-plot design with the four clover species as 67 randomized sub-units within herbicide main plots (3 replications). Herbicide treatments and 68 application rates (Table 7) included commonly applied broadleaf herbicides or were chosen 69 based upon labeling for leguminous crops. Treatments included a non-treated control. All 70 treatments included a 0.25% v v -1 non-ionic surfactant (Induce, Helena Chemical Company, 71 Collierville, TN). Herbicides were applied at 280 L ha -1 spray volume on 10 March 2010 or 22 72 February 2011 via a CO 2 pressurized back-pack sprayer equipped with four TeeJet XR8002 flat 73 fan nozzles (Spraying Systems Co., Wheaton, Illinois, USA). 74 During field experiments, clover control was visually assessed 6 weeks after treatment 75 (WAT) relative to the non-treated control, where 100% control equaled complete plant death. 76 Control was based upon a combination of herbicide injury and plant health. Control assessments 77 did not account for height reductions. However, plant height from the soil surface was sampled 78 47 twice by lifting the two tallest foliar meristems, whether inflorescence or leaf, and measuring to 79 the uppermost point. 80 Supplemental greenhouse experiments were conducted during 2011 and 2012 at the 81 Auburn University Weed Science Greenhouse, (32?35?12? N, 88?29?15? W) in order to evaluate 82 herbicide effects upon clover biomass. Plants were collected 1 to 18 February 2011 and 13 to 20 83 January 2012 (Table 6) identically to those of the field experiments. To prevent sample erosion 84 and to facilitate sample randomization, greenhouse plants were placed in pots (11 cm diameter, 85 730 cm 3 volume). Greenhouse air temperature was maintained between 23 and 25 ?C. Plants 86 were subject to normal daytime irradiance (less than 350 !mol m -2 s -1 at foliage height) and were 87 watered via over-head mist irrigation twice daily. Herbicide treatments were identical to those 88 applied in field experiments (Table 7). Treatments were applied in an enclosed research spray 89 cabinet applying 280 L ha -1 through a single TeeJet TP8002EVS nozzle (Spraying Systems Co., 90 Wheaton, Illinois, USA). The study was conducted as a completely randomized design with three 91 replications and one pot per experimental unit. Plants were randomized daily to account for 92 variations within the greenhouse microclimate. Foliage was harvested at the soil surface and 93 oven dried at 50?C for 72 hours to ascertain above ground biomass. 94 Height and biomass responses are based upon percent reduction relative to the non- 95 treated control. All data were subject to analysis of variance (ANOVA) within SAS procedure 96 GLIMMIX using mixed model methodology (SAS ? Institute v. 9.2, Cary, North Carolina, USA). 97 Field and greenhouse data were analyzed separately. Treatment was considered a fixed effect in 98 the model. Year, replication (nested within year), and iterations containing these effects were 99 considered random in the model and were non-significant for all response variables (Carmer et 100 al. 1989; Hager et al. 2003). Basic model assumptions were confirmed. Means were separated 101 48 based upon adjusted 95% confidence intervals, which allows for multiple comparisons by 102 protecting family-wise error rate (Littell et al. 2006). 103 RESULTS AND DISCUSSION 104 Analysis of variance indicated that year and year by treatment interactions were not 105 significant (P > 0.05; Table 8). Therefore, experiments were pooled across years with respect to 106 growing condition (field or greenhouse). Precedence was given to field data, with greenhouse 107 biomass reductions presented separately. Of the field data, priority was given to percent control, 108 with relative height discussed as supporting evidence. Studies indicated varying control and 109 height reductions due to species by herbicide interactions. Interaction effects were given 110 precedence to main effects. 111 Field experiments. ANOVA (Table 8) indicated significant herbicide by species 112 interaction effects upon control and height data of field experiments (Table 9). 2,4-D control did 113 not differ due to species and was ! 88% for all clovers. However, 2,4-D reduced small hop 114 clover height greater than that of white clover (97% versus 41%) and reduced ball and crimson 115 clover heights 64 and 63%, respectively. Herbicide effects on plant height are likely of biological 116 importance to plant survival and stand resilience. However, reductions in size may be linked to 117 more than just herbicide induced plant injury. Fletcher and Raymond (1956) first demonstrated 118 that phenoxy-hebicides, like 2,4-D, reduced the success of Rhizobium trifolii to form symbiotic 119 relationships with white clover, subsequently reducing N fixation. More recent studies have 120 demonstrated that various herbicides directly damage both host plant and symbiotic rhizobium 121 (Clark and Mahanty 1991). Herbicide effects upon rhizobium, nodulation, and N fixation were 122 not examined within these experiments. However, future research should focus upon plant 123 competitiveness, rather than simply plant survival. 124 49 Since the 1950?s legume tolerance to butyric acid compounds, such as 2,4-DB and 125 MCPB, has been linked to reduced beta-oxidation within tolerant species (Wain and Wightman 126 1954). Within our own experiments, 2,4-DB was moderately tolerated by all clovers, and control 127 did not differ due to species (! 58% control; Table 9). However, 2,4-DB did affect clover heights 128 differently. 2,4-DB did not affect crimson and ball clover heights (+2% and 12%, respectively) 129 relative to the non-treated control; however, 2,4-DB did reduce small hop clover height 27%, 130 which was similar to ball and white clover height reductions but greater than that of crimson 131 clover. 2,4-DB reduced white clover height 50%, which was greater than ball and crimson clover 132 height reductions and similar to height reductions observed due to 2,4-D. Differential response to 133 2,4-DB in leguminous pasture species has previously been reported. Mulholland et al. (1989) 134 demonstrated differential Medicago species responses, while Young et al. (1992) reported that 135 M. aculeata and T. subterraneum were more tolerant of 2,4-DB than M. truncatula. 136 MCPA is applied alone and in commercially available herbicide mixtures for pasture and 137 rangeland management but may lack selectivity for many pasture legumes (Conrad and Stritzke 138 1980; Evers et al. 1993). Our experiments demonstrated this lack of tolerance amongst four 139 clover species. MCPA controlled clovers between 56 and 86% and reduced heights between 11 140 and 67%. An alternative to MCPA not included amongst our treatments was the butyric acid 141 compound MCPB, which has utility within leguminous crops (Senseman 2007) and has 142 previously been demonstrated safe upon white clover (Elliot 2006). 143 Clopyralid and dicamba effectively controlled all clovers (" 95%; Table 9) and 144 completely reduced heights across species. Triclopyr control was similar to that of clopyralid (" 145 81%); however, triclopyr affected clover heights differently. Triclopyr failed to reduce ball 146 clover height relative to the non-treated and reduced crimson clover height only 22%. Small hop 147 50 clover height was reduced 61%, which was similar to reductions in crimson clover height but 148 greater than that of ball clover. Triclopyr reduced white clover height 91%, which was greater 149 than ball and crimson clover height reductions. It is noteworthy that herbicides from the same 150 family (e.g., clopyralid and triclopyr) did not exhibit similar efficacy in this experiment. 151 Atrazine effectively controlled all clovers (! 98%) and reduced clover heights ! 86% 152 (Table 9). On the contrary, bentazon was well tolerated by all clover species (" 15% control and 153 " 17% height reduction). In fact, a 30% increase in white clover height was observed due to 154 bentazon application. Other researchers have previously reported similar responses to bentazon. 155 Ceballos et al. (2004) reported increases in red clover (T. pretense) plant height (70 and 48% for 156 12 and 24 g 100 m -2 rates) at the expense of roots, which were reported to have decreased 42% 157 by 20 days after treatment. Root biomass was not measured during our experiments. 158 Only imazaquin resulted in differential clover control. Imazaquin controlled small hop 159 clover greater than white clover (91% versus 50%; Table 9). Ball and crimson clover control (80 160 and 62%, respectively) were similar to that of other clovers. Imazaquin reduced small hop clover 161 height 88%, which exceeded height reductions measured among other clovers (" 47%). 162 Differential soybean-cultivar responses to imazaquin have been reported (Kent et al. 1988). More 163 recently, differential responses to acetolactate synthase (ALS) inhibitors, such as imazaquin, 164 have been attributed to resistance mechanisms (Tranel and Wright, 2002). However, ALS 165 resistance has not been confirmed amongst Trifolium spp. (International Survey of Herbicide 166 Resistant Weeds, 2012). 167 Imazethapyr was well tolerated by all clover species. Imazethapyr controlled clovers " 168 15% (Table 9). Crimson clover height (+9%) did not differ from that of the non-treated. Small 169 hop and white clovers were reduced in height 33 and 45%, respectively, while ball clover height 170 51 was reduced 12%. Previous research has demonstrated imidazolinone herbicides, such as 171 imazethapyr, can be utilized for promoting the establishment of certain legumes within tall-grass 172 prairies (Beran et al. 1999). 173 Metsulfuron and trifloxysulfuron herbicides are highly effective against many broadleaf 174 weeds found within mixed grass swards, yet knowledge of differential tolerance among legume 175 species is limited. Our results did not suggest differential tolerance, with metsulfuron and 176 trifloxysulfuron having controlled and reduced heights similarly across clovers. Metsulfuron 177 controlled all clover species ! 88% and reduced clover heights ! 78% (Table 9). Similarly, 178 trifloxysulfuron controlled clovers ! 80% and reduced clover heights ! 45%. 179 Greenhouse experiments. Supplemental greenhouse experiments evaluated biomass 180 harvests (Table 10). Biomass reductions differed due to herbicide treatment as well as clover 181 species but did not differ due to herbicide by species interaction. Biomass reductions are 182 important considerations when managing mixed grass-clover swards for forage. 183 Clopyralid and atrazine reduced clover biomass 98%, similar to 2,4-D (85%), dicamba 184 (92%), triclopyr (89%), and metsulfuron (84%), but greater than those of all other treatments 185 (Table 10). Imazaquin reduced clover biomass 73%, similar to 2,4-D, dicamba, triclopyr, 186 metsulfuron, and trifloxysulfuron (68%). MCPA reduced clover biomass 50%, similar to 2,4-DB 187 (45%), bentazon (36%), and imazethapyr (28%). 188 White clover biomass was reduced less than crimson and hop clovers (58% versus 72%), 189 but equal to that of ball clover (61%; data not shown). Species main effects are important in 190 several contexts. Foremost, labels do not always clearly define species for which herbicides are 191 tolerated. These results suggest that clovers vary in herbicide susceptibility. Secondly, labels 192 may ambiguously emphasize hop clover control. Yet there are at least three Trifolium spp. that 193 52 are generically called ?hop clovers? (Plants Database, 2013; WSSA, 2013), some of which differ 194 dramatically in phylogeny (Ellison et al. 2006). 195 IMPLICATIONS FOR MANAGEMENT 196 On a practical level, our results demonstrate potential herbicide options for maintaining 197 mixed grass-clover swards. Candidate herbicides include bentazon, 2,4-DB, and imazethapyr. 198 These herbicides are commonly labeled for use within leguminous crops as well as forage and 199 rangeland legumes. Bentazon and 2,4-DB have proven to be moderately tolerated by 200 subterranean- (T. subterranean) and arrowleaf- (T. vesiculosum) clovers (Hawton et al. 1990; 201 Smith and Powell, 1979). The relative tolerance of clover species to these candidate herbicides is 202 further evidence of their value within certain scenarios. Yet, it is difficult to foresee herbicide 203 applicators choosing these herbicides without further evidence of weeds controlled, costs, and 204 effects upon mixed swards. There are undoubtedly many herbicides that are tolerated by clover 205 species, yet questions remain about application rates and timing. 206 Our experiments suggest varying tolerances amongst clover species and common 207 broadleaf herbicides. This agrees with previous research of differential herbicide tolerance 208 amongst other pasture and forage legumes (Bowran 1993; Mulholland et al. 1989; Young et al. 209 1992). However, to our knowledge, this is the first report of differential tolerance solely amongst 210 Trifolium spp. This supposition has broad impacts within agronomic scenarios. Pasture and 211 rangeland managers have long sought herbicidal weed control without harming utility clover 212 species, with limited success. Clover seed producers may benefit from the knowledge that certain 213 clovers may be preferentially favored by differential herbicide responses. Additionally, legumes 214 such as clovers have application within mixed turf swards. Legume species and varieties 215 continue to be developed and improved for various agronomic applications (Rajeev et al. 2009). 216 53 However, herbicide labels often fail to clearly define the clover species for which an herbicide is 217 intended (whether for selective weed control or for tolerance). As the number of species, 218 varieties, and uses of clovers increase, label statements must more precisely scrutinize species 219 tolerance in order to increase the viability and profitability of biodiverse agricultural scenarios. 220 221 222 54 54 223 Table 6. Four clover (Trifolium) species and their respective harvest and transplant dates. Plants were harvested and alowed to mature in a grenhouse seting. Plants were then subject to selection for uniform size and maturity folowed by random asignment to either field or grenhouse experiments. Year Clover Harvest date Growth cycle Transplant date Treatment date Flowering stage at treatment a Leaves per plant at treatment 2010 white (T. repens) 19 January Perenial 10 February 10 March vegetative 10 to 20 smal hop (T. dubium) 19 January Anual 10 February 10 March early-flowering 20 to 30 crimson (T. incarnatum) 20 January Anual 10 February 10 March early-flowering 10 to 20 bal (T. nigrescens) 19 January Anual 10 February 10 March early-Flowering 15 to 25 2011 b white (T. repens) 1 February Perenial 15 February 2 February vegetative 10 to 20 smal hop (T. dubium) 1 February Anual 15 February 2 February early-flowering 20 to 30 crimson (T. incarnatum) 10 February Anual 15 February 2 February early-flowering 10 to 20 bal (T. nigrescens) 10 February Anual 15 February 2 February early-flowering 20 to 30 2012 c white (T. repens) 1 February Perenial 15 February 2 February early-flowering 10 to 20 smal hop (T. dubium) 1 February Anual 15 February 2 February late-flowering 20 to 30 crimson (T. incarnatum) 10 February Anual 15 February 2 February early-flowering 20 to 30 bal (T. nigrescens) 10 February Anual 15 February 2 February mid-flowering 20 to 30 a Flowering stage is indicated as either early (bloms present but remaining un-opened or slightly opened), mid (having blomed but no signs of flower senescence), late (flower kels having more than roughly 25% discoloration due to senescence). b 201 dates refer to both field and grenhouse studies. c 2012 dates refer to grenhouse studies only. 55 224 Table 7. Herbicide rates and formulations aplied in field and grenhouse experiments to four clover (Trifolium) species. Al treatments included a 0.25% v v -1 non-ionic surfactant. Herbicides were aplied at 280 L ha -1 spray volume. Experimental rates were chosen based upon comon labeled rates and unpublished studies where legume tolerance had ben observed. Mechanism of Action a Comon ame Trade name Formulation Rate 10 m -2 Manufacturer City, State Website synthetic auxins 2,4-D Amine 40 dimethyl amine salt 15.8 g ae PBI Gordon Kansas City, MO ww.pbigordon.com 2,4-DB b c Butyrac 20 dimethyl amine salt 15.8 g ae Albaugh Ankeny, IA ww.albaughinc.com dicamba Banvel dimethyl amine salt 1.2 g ae Arysta LifeScience Cary, NC ww.arystalifescience.com MCPA b MCPA Ester 4 ethylhexyl ester 5.2 g ai Albaugh Ankeny, IA ww.albaughinc.com clopyralid Lontrel Turf and Ornamental monoethanola mine salt 4.2 g ai Dow AgroSciences Indianapolis, IN ww.dowagro.com triclopyr Turflon Ester Ultra butoxyethyl ester 5.6 g ai Dow AgroSciences Indianapolis, IN ww.dowagro.com photosystem I inhibitors atrazine Atrex 4L -- 2.4 g ai Syngenta Crop Protection Grensboro, NC ww.syngenta.com bentazon b c Basagran sodium salt 1.2 g ai Arysta LifeScience Cary, NC ww.arystalifescience.com acetolactate synthase inhibitors imazaquin b c Scepter 70 DG fre acid 5.6 g ai BASF Research Triangle Park, NC ww.basf.com imazethapyr b c Pursuit amonium salt 0.7 g ai BASF Research Triangle Park, NC ww.basf.com metsulfuron- methyl MSM Turf -- 0.2 g ai FarmSaver Raleigh, NC ww.farmsaver.com trifloxysulfuron Monument 75 WG sodium salt 0.3 g ai Syngenta Crop Protection Grensboro, NC ww.syngenta.com a Acording to Senseman (207). b Comonly labeled for use within forage and pasture legumes. c Labeled for use within soybean production (Glycine max). 56 Table 8. ANOVA results and source sum of squares (S) relative to the total S for field and grenhouse experiments 6 WAT. Experiment Field a Grenhouse b Source Control c Height d Biomas d Herbicide 0.001 e 0.001 0.028 Species 0.5752 0.001 0.001 Herbicide ! Species 0.081 0.001 0.0695 a Field experiments were conducted during winters 2010 and 201 and did not include biomas analysis. b Suplemental grenhouse experiments were conducted during winters 201 and 2012 and evaluated biomas. c Control was visualy asesed on a percent scale 6 WAT relative to the non-treated control. d Height and biomas responses were calculated based upon percent reduction relative to the non- treated control 6 WAT. e P > F values obtained within SAS Proc MIXED. 57 Table 9.1. Control and height reductions of four clover (Trifolium) species measured 6 weks after treatment (WAT) in field studies. Efects were restricted to P ! 0.05 level of significance. Efects were combined acros years. Model validity (P > F) is provided for significant species by herbicide interaction. % Control a % Height reduction b Herbicide Clover Mean c ? 95% CI d P > F Mean ? 95% CI P > F 2,4-D bal 88 8 NS d -64 ab 34 0.049 crimson 91 8 -63 ab 27 smal hop 95 8 -97 a 27 white 91 8 -41 b 27 2,4-DB bal 18 28 NS -12 bc 12 < 0.01 crimson 30 28 +2 c 12 smal hop 58 26 -27 ab 12 white 28 28 -50 a 12 dicamba bal 99 1 NS -100 0 NS crimson 100 1 -100 0 smal hop 100 1 -100 0 white 100 1 -100 0 MCPA bal 86 31 NS -23 32 NS crimson 58 27 -11 32 smal hop 56 25 -67 32 white 78 25 -51 39 clopyralid bal 100 3 NS -100 0 NS crimson 100 3 -100 0 smal hop 95 3 -100 0 white 100 3 -100 0 triclopyr bal 88 11 NS -17 c 21 < 0.01 crimson 81 12 -2 bc 21 smal hop 92 12 -61 ab 21 white 88 11 -91 a 21 Continued in Table 9.2 on the folowing page. a % Control was visualy asesed 6 WAT relative to the non-treated control. b % Height and biomas reductions are relative to the non-treated control. Negative numbers indicate height reduction. c Mean separations were performed using 95% confidence intervals. Overlaping intervals signify a lack of diference betwen means of the same herbicide treatment. Leters are presented as a method of easily distinguishing significant diferences amongst herbicide treatment. d Abreviations: 95% CI, 95% confidence interval; NS, non-significant. 58 Table 9.2. Continued from Table 9.1: Control and height reductions of four clover (Trifolium) species measured 6 weks after treatment (WAT) in field studies. Efects were restricted to P ! 0.05 level of significance. Efects were combined acros years. Model validity (P > F) is provided for significant species by herbicide interaction. % Control a % Height reduction b Herbicide Clover Mean c ? 95% CI d P > F Mean ? 95% CI P > F atrazine bal 100 1 NS -100 21 NS crimson 100 1 -100 17 smal hop 100 1 -100 17 white 98 1 -86 17 bentazon bal 9 11 NS -17 22 NS crimson 15 10 -2 22 smal hop 4 11 -3 22 white 5 10 +30 27 imazaquin bal 80 ab 19 0.03 -47 b 20 0.012 crimson 62 ab 19 -38 b 24 smal hop 91 a 21 -8 a 20 white 50 b 19 -36 b 20 imazethapyr bal 7 13 NS -12 26 NS crimson 15 13 +9 26 smal hop 10 13 -33 26 white 10 13 -45 32 metsulfuron bal 90 12 NS -79 24 NS crimson 93 11 -78 24 smal hop 93 11 -97 24 white 88 11 -82 24 trifloxysulfuron bal 92 14 NS -84 35 NS crimson 95 14 -70 29 smal hop 80 14 -91 29 white 89 15 -45 35 a % Control was visualy asesed 6 WAT relative to the non-treated control. b % Height and biomas reductions are relative to the non-treated control. Negative numbers indicate height reduction. c Mean separations were performed using 95% confidence intervals. Overlaping intervals signify a lack of diference betwen means of the same herbicide treatment. Leters are presented as a method of easily distinguishing significant diferences amongst herbicide treatment. d Abreviations: 95% CI, 95% confidence interval; NS, non-significant. 59 Table 10. Herbicide main effects upon clover (Trifolium) biomass reductions measured 6 weeks after treatment (WAT) during greenhouse experiments. Greenhouse % Biomass reduction Herbicide Mean a ? 95% CI 2,4-D -85 abc 9 2,4-DB -45 ef 10 MCPA -50 ed 9 dicamba -92 ab 9 clopyralid -98 a 9 triclopyr -89 ab 9 atrazine -98 a 9 bentazon -36 f 9 imazaquin -73 bc 9 imazethapyr -28 f 9 metsulfuron -84 abc 9 trifloxysulfuron -68 cd 9 a Mean separations were performed using 95% confidence intervals. Overlaping intervals signify a lack of diference betwen means of the same herbicide treatment. Leters are presented as a method of easily distinguishing significant diferences amongst herbicide treatment. 60 Dynamics of White Clover (Trifolium repens) Decomposition in a Southeastern Bermudagrass Lawn INTRODUCTION The ecological impact of turfgrass is frequently questioned, due in part to nutrient and water requirements as well as often-unsustainable monoculture cultivation (Milesi et al., 2005; Robbins and Birkenholtz, 2003; Robbins et al., 2001). Nitrogen (N) is essential to turfgrass health and quality (Beard, 1973; Turgeon, 2002). Commercial-lawn N requirements vary with species and environmental conditions, but within the southern United States, common rates range from less than 5 g N m -2 year -1 for bahiagrass (Paspalum notatum Fluegg?) and centipedegrass (Eremochloa ophiuroides (Munro) Hack.) to almost 30 g N m -2 year -1 for bermudagrass (Cynodon dactylon (L.) Pers.; Duble, 2004). Clover (Trifolium spp.) inclusion within maintained turfgrass is a proposed means of increasing turfgrass sustainability (Dudeck and Peacock, 1983; Sincik and Acikgoz, 2007). Clover has been included in grass mixtures for roadsides as well as other maintained turfgrass areas and has proven useful for slope stabilization (Roberts and Bradshaw, 1985). In particular, white clover (T. repens L.) thrives within home lawns and golf courses because it can flower and produce seed at mowing heights as low as 6 mm (Watschke et al., 1995). White clover increases turfgrass greenness by contributing N to associated grasses and has been reported to increase turfgrass color ratings within cool-season turfgrass (Sincik and Acikgoz, 2007) and increase vegetative cover within dormant bermudagrass 61 (Dudeck and Peacock, 1983). Estimates of white clover N fixation within three cool- season turfgrasses are greater than 25 g N m -2 year -1 , with 4.2 to 13.7% of total N contributed to the associated turfgrasses (Sincik and Acikgoz, 2007). Other pertinent research concerning white clover inclusion has been conducted in forage scenarios where white clover was grazed or harvested for animal fodder. Estimates of N fixation for grass-white clover pastures range from nil to 40 g N m -2 year -1 , though most are from 10 to 25 g N m -2 year -1 (Ledgard and Steele, 1992; McNeil and Wood, 1990; Whitehead, 1995). Using the 15 N transfer method, McNeil and Wood (1990) estimated N fixation by white clover within perennial ryegrass was approximately 15.5 g N m -2 year -1 , with 28% of the total fixed N having been transferred to associated ryegrass. Transfer of N from legumes to associated turfgrass occurs indirectly through excreted N and decomposition of nodules, roots, and foliage (Brophy et al., 1987; Dubach and Russelle, 1994; Jensen, 1996; Wardle and Greenfield, 1991). Decomposition of root nodules is a significant source of N. Reported N concentration of root nodules ranges from 4.8 to 9.0% of root dry matter (Chu and Robertson, 1974; Wardle and Greenfield, 1991). However, root nodules are not the sole source of N transfer, as above ground white clover dry matter has been reported to be 9.1 to 24.2% protein, depending upon harvest date (about 1.5 to 4.0% N; Burton and DeVane, 1992). Unlike forage scenarios, turfgrass systems differ in that they are not grazed; rather, they are mown frequently to maintain utility and aesthetics. Mown clippings are returned to the turf surface, potentially contributing a mineralizable source of N. Polyculture lawns of grass and white clover are historically common, yet little is known of the N-contribution and C-flux from decaying clover foliage. The rates of 62 decomposition, N mineralization, and C deposition would be useful information for future research regarding this subject as well as when assigning nutrient credits to white clover-culture in warm and cool season turfgrass. Such information would be highly dependent upon a multitude of factors, including time of year, litter composition, soil and climactic -conditions, as well as soil fauna. For these reasons, it may not be possible to control all factors in situ. Organic residues decompose in two phases. Soil microbes rapidly consume the labile fraction, which is composed of sugars, starches, and proteins, leaving behind a recalcitrant fraction composed of cellulose, fats, waxes, lignin, and tannins (Wieder and Lang, 1982). This slowly decomposing fraction helps to develop soil organic matter. Due to the two-step nature of decomposition, a double exponential decay model is often implemented to describe litter decay (Wieder and Lang, 1982). Double exponential decay equations are of the form Y = Ae -k1t + Be -k2t , where Y = response, A and B are initial concentrations approximating the labile and recalcitrant portions, k1 and k2 are rate constants fitted to the data, and t equals time in days after application (DAA). Such models have been used successfully to describe quickly decaying legume litter in Alabama (Mulvaney et al., 2010) as well as the decomposition and N release of hedgerow species in Haiti (Isaac et al., 2000). Modeling white clover decomposition may enable turfgrass researchers and professionals to more accurately predict nutrient contribution to associated grasses and help optimize supplemental fertilizer recommendations. Our objectives were to 1) explore the use of double exponential decay models as a method to predict C and N 63 contributions of white clover litter applied at different times throughout the year, and 2) quantify white clover litter decomposition, as well as C and N -release rates. MATERIALS AND METHODS An in situ decomposition study was conducted at the Auburn University Turfgrass Research Unit (32?34?40? N, 85?29?57? W; elevation 185 m) in Auburn, AL, on a Marvyn sandy loam (fine-loamy, kaolinitic, thermic Typic Kanhapludult) soil with pH 6.3 (1:1 soil:H 2 O). Treatments (application date by retrieval timing) were arranged in a completely random design with four replicates. White clover litter was applied Mar 1, 2010; Jun 1, 2010; and Dec 1, 2010 (March, June and December -applications, respectively). Retrieval timings were 0, 1, 4, 7, 14, 28, 56, and 112 DAA. Time 0 DAA samples were truly replicated in the field. Samples for decomposition studies were harvested from a stand of commercially available white clover, ?Dutch? white clover (Main Street Seed and Supply, Bay City, MI), which had been established in previous experiments unrelated to this research. The population is an intermediate growth-type marketed for grazing and wildlife habitat. White clover was maintained at 7.6 cm mowing height with supplemental irrigation applied as needed and no supplemental fertilization. Soil moisture was greater than 25% at each harvest date, ensuring that the clover stand was fully turgid prior to harvesting foliage. Leaves of the harvest area were patted dry with paper towels prior to harvest. In order to mimic a standard mowing occurrence, white clover foliage was harvested 4 cm above soil level using hand-held shears. Litter was transported on ice in order to preserve samples during preparation and field placement. Within two hours of harvest, litter was mixed thoroughly, and contaminates such as grass and dead or necrotic tissue were 64 removed. Clover foliage, excluding flowers, was placed into nylon bags measuring 10 ! 20 cm with 50 to 60 ?m openings on a fresh rate (FW) basis at 10.0 g bag -1 (500 g FW m - 2 ). Individual litterbags were placed within a 50 ! 50 cm experimental area to prevent possible bag-to-bag interference. In preparation for litterbag placement, 10 ! 20 cm areas of ?Tifway? hybrid bermudagrass (Cynodon dactylon (L.) Pers. ! C. transvaalensis Burtt Davy) were denuded to the soil level using a gas-powered string trimmer. Steel sod staples 20 cm in length were used to secure each of the four corners of the sealed litterbag to the soil layer. The study area was maintained at a height of 3.8 cm; all clippings were returned to the turfgrass surface. Plots received 3 cm supplemental irrigation on a weekly basis between May 31 and September 10, 2010, and resuming March 14, 2011. Prior to study initiation, the area was fertilized (49 kg N ha -1 ) Feb 15, 2010 and received no supplemental fertility for the duration of the study. Retrieved litter was air-dried at 60?C for 48 hours and weighed for dry-matter determination. Litter was then ground to pass a 16-mesh sieve and analyzed for total C and N by LECO TruSpec CN (Leco Corp, St. Joseph, MI). To account for possible soil contamination of litterbag contents, all data were converted to an ash-free dry weight (AFDW) basis by ashing 5 g of sample in a muffle furnace at 400?C for 12 hours (Cochran, 1991). Soil temperature was recorded via TidbiT ? v2 Temp loggers (Onset Computer Corp, Pocasset, MA) buried 10 cm below soil level. Air temperatures at 1.5 m above ground level were obtained from a nearby weather station (32?36?00? N, 85?30?00? W; elevation 199 m) in Auburn, AL (AWIS, 2013). 65 Analysis of variance was conducted. Means, standard errors, and statistical significance of treatments were determined at the 95% confidence level using mixed models procedures within Proc Glimmix (SAS Institute, 2004). Least squares estimates for nonlinear models were determined within SigmaPlot 11 using Marquardt-Levenberg algorithm to provide the best fit (Systat Software, 2008). Initial parameter ranges were selected with a maximum of 200 fits and 200 iterations. Double exponential decay models served as the basis for comparison of mass, N, and C loss between application dates. In most cases, double exponential decay models minimized residual sums of squares and produced comparatively lower residual mean squares, standard errors, and PRESS statistics as well as better coefficients of determination (R 2 adj ) than single exponential decay models. For brevity, comparisons of single and double exponential models are omitted. However, instances where double exponential decay models could be collapsed into single exponential models are generally indicated by the presence of k2 values close to zero. Days to 50% decomposition (D50) values were estimated based upon double exponential decay equations to compare and contrast regression estimates. RESULTS AND DISCUSSION Double exponential decay models Analysis of variance indicated a significant application-date by retrieval-time interaction for all response variables (discussed separately below). Parameters fit to the double exponential decay curve are shown in Table 1. All regression equations were significant (p < 0.0001) and were good approximations of the data (R 2 adj ). It is convenient to represent decay patterns on a percent of original material basis such that one can 66 extrapolate for hypothetical amounts of residue in field conditions. Table 1 shows residue persistence normalized to 100% of initial AFDW. Normalized equations offer an approximation of labile (A) and recalcitrant (B) litter on a percent basis. Generally, models revealed two mass, C, and N pools for all application dates. Across all response variables, initial decomposition of litter occurred more rapidly during June application relative to March and December applications. These trends are visually evidenced by steeper slopes (Figure 11, 12, 13) as well as greater k1 and k2 values (Table 1) during the decay of labile and recalcitrant portions. Differences in the rate of decay are apparent by comparing k1 and k2 values from each equation. Decay constants are similar, whether presented on a percent remaining or area (data not shown) -basis. Mass remaining. White clover mass decreased fastest when applied in June (Figure 11). This was visually evidenced by steeper slopes and greater decay constants (Table 1). The labile decay constant of June-applied litter (0.1056) was nearly three times greater than that of March-applied litter (0.0367) and more than 6 times that of December-applied litter (0.0166). The effects of application date upon the decay of recalcitrant portions were more pronounced. The recalcitrant decay constant of June-applied litter (0.0043) was more than 10 orders of magnitude greater than that of March and December-applied litter. The relatively quick decay of June-applied litter is typical of warmer soil temperatures (Figure 1) and the increased microbial activity involved in decomposition. For all application dates, labile portions were greater than 80%, and recalcitrant portions were ! 25% (Table 1). Mass predictions based upon % remaining data generally over estimated 0 DAA mass, which is likely due to 3 and 7 DAA data having 67 been abnormally greater than predicted levels (Figure 11). This phenomenon may be in large part due to an initial resistance to decay during application of fresh litter. This lag is not reflected by the double exponential decay curve but may be more adequately considered a sigmoidal response. Others have suggested that litter deposited during climactic periods unfavorable for decomposition, such as winter conditions, may be best explained by a sigmoidal curve (Swift et al., 1979). We suggest that the initial delay is linked to litter having been applied as fresh material, rather than dried material. Yet it could also be argued that fresh litter had the moisture necessary to drive microbial activity. Our goal was simply to simulate actual occurrences under field conditions, and it is not known what effect foliage moisture had on initial decay. Table 12 contains predicted mass persistence expressed as 95% confidence intervals ranging from 0 to 112 DAA. Applications varied upon rapidness to reach 50% of original material (Figure 1). June-applied litter mass halved in 12.1 d, decreasing from an initial equivalent of 104.4 g m -2 to 52.2 g m -2 (data not shown). March-applied litter was slightly slower, taking 29.0 d to halve in remaining mass from 71.9 g m -2 to 36.0 g m -2 , while December-applied litter took an estimated 57.4 d to decrease from 72.1 g m -2 to 36.1 g m -2 . Due to the nature of depositing fresh material rather than dry material, seasonal applications differed in dry matter equivalence placed upon an area basis. These differences may have been due to foliage moisture content despite precautions to minimize differentials. To what extent this may have influenced decay is not explored within this analysis. Carbon remaining. 68 Carbon composition of initial clover foliage differed slightly among application date. June-applied litter C (44.0% ? 95% CI = 0.3%) was slightly greater than that of March (42.7% ? 95% CI = 0.1%) and December (42.1% ? 95% CI = 0.6%) -applied litter. C loss models (Figure 12) were comparable to those of mass loss. This is attributed to mass lost through microbial respiration of C, which is lost as CO 2 to the atmosphere (Wood and Edwards, 1992). White clover C decreased fastest when applied in June (Figure 12). The labile decay constant of June-applied litter (0.1061; Table 1) was three times greater than that of March-applied litter (0.0354) and more than seven times that of December-applied litter (0.0143). The recalcitrant decay constant of June-applied litter (0.0045) was nearly six times greater than that of March-applied litter (0.0008) and nearly eight orders of magnitude greater than December-applied litter (6.7 E-13). When percent remaining data were analyzed, four parameter exponential decay models revealed two C pools for application dates (Table 1). Labile portions of all applications were greater than 79%, and recalcitrant portions were ! 26%. Carbon predictions based upon % remaining data generally over estimated 0 DAA C, which is likely due to 3 and 7 DAA data having been abnormally greater than predicted levels (Figure 12). This phenomenon may be in large part due to an initial resistance to decay during application of fresh litter. Applications varied upon rapidness to reach 50% of original material (Figure 12). June-applied litter C halved in 11.0 d, decreasing from an initial equivalent of 44.6 g C m - 2 to 22.3 g C m -2 . March-applied litter was slightly slower, taking 27.8 d to halve in 69 remaining C from 31.6 g C m -2 to 15.8 g C m -2 , while December-applied litter took an estimated 60.5 d to decrease from 30.3 g C N m -2 to 15.2 g C m -2 . Nitrogen remaining. Nitrogen composition of June-applied litter (4.8% ? 95% CI = 0.4%) was slightly greater than that of March (4.1% ? 95% CI = 0.1%) and December (4.0% ? 95% CI = 0.2%) -applied litter. However, these data agree with the range of 4.1 to 4.9% N reported by Sincik and Acikgoz (2007). Variation of N content within clover stands is likely due to environmental factors affecting symbiotic N fixation within plant roots, which were not accounted for within this study, as well as N availability within cooler soils. In their review of N fixation of grass-legume pastures, Ledgard and Steele (1992) report that fixation is greatly reduced due to dry soil conditions, acidic soils, and the ?pest/disease complex.? Another factor affecting N fixation is soil temperature. Frame and Newbould (1986) found that a minimum temperature of 9?C was necessary for active N fixation by Rhizobium. It has also been reported that temperatures necessary for nitrogenase activity range from 13 to 26?C (Halliday and Pate, 1976). Soil and air temperatures may not have been compatible with active N fixation prior to harvest of March and December plant material. Loss of N from decomposition of white clover was quickest when applied in June (Figure 13). The labile decay constant of June-applied litter (0.0938; Table 1) was more than three times that of March-applied litter (0.0271) and nearly 10 times that of December-applied litter (9.970 E-3) suggesting that temperatures during June were far more conducive to microbial decomposition of litter; however, climactic conditions are not modeled within decomposition equations. Decay constants of the recalcitrant portions 70 (k2) were smaller when litter was applied in March (1.2947E-12; Table 1) rather than June (0.0036) and December (0.0100), indicating that the nature of the recalcitrant decay was slower when applied in June than March or December. It is also possible that the lack of sampling dates beyond 112 DAA did not allow for accurate prediction of recalcitrant decay within March and December applications. Had sampling dates extended further, N remaining would have been more likely to approach 0%. When percent remaining data were analyzed, four parameter exponential decay models revealed two N pools for application dates, each having similar size. The labile portions of all applications were greater than 80% (Table 1), and recalcitrant portions were less than 25%. These sizes are slightly greater for the faster decaying labile portions than the approximately one to one ratios of Lespedeza cuneata (Dum. Cours.) G. Don, Albizia julibrissin Durazz., and Glycine max (L.) Merr. reported by Mulvaney et al. (2010). Predictions generally over estimated 0 DAA N, which is likely due to 3 and 7 DAA data having been abnormally greater than predicted levels (Figure 13). This abnormality is not exceptional, and may be due to N immobilization from surrounding sources. Similar faults in non-linear models fit to litter decay curves have been attributed to N immobilization, though these were for higher C/N ratio wheat straw (Mulvaney et al., 2010). Similarly, rainfall or irrigation during these months could be sources of immobilized N. Application dates varied markedly in their rapidness to reach 50% of original material (Figure 13). June-applied litter N concentrations halved in 10.9 d, decreasing from an initial equivalent of 3.4 g N m -2 to 1.7 g N m -2 . March-applied litter was slightly slower, taking 37.0 d to halve in remaining N from 4.2 g N m -2 to 2.1 g N m -2 , while 71 December-applied litter took an estimated 73.6 d to decrease from 2.9 g N m -2 to 1.5 g N m -2 . After 112 d, predicted N remaining from March, June, and December-applied litter had reduced to 22.7, 14.6, and 34.2% of that applied, respectively (Table 12). Caution should be used when making predictions beyond the length of the study due to the variable nature of litter decay over seasons. Furthermore, predicted decay is rapid given suitable conditions. Therefore, any extrapolations would be minute in comparison to initial nutrient release. C/N ratios. The nature of clover decomposition is similar to that of other legumes. That is, legumes contain a relatively high concentration of N, allowing for very rapid initial decay. Clover samples had C/N ratios (10.1 ? S.D. 0.8), which were similar across all application intervals. C/N composition of remaining litter according to predicted decay equations are shown in Table 14. C/N ratios are frequently used to describe a residue?s propensity to mineralize or immobilize soil inorganic N. However, associated chemical analysis of decomposable fractions (e.g., the labile and recalcitrant fractions) may be a better means of determining a residues? effect upon soil N concentrations (Hadas et al., 2004). What is not well understood is the persistence of recalcitrant fractions beyond initial decay. With C/N ratios of nearly 10 to 1 throughout decomposition (Table 12), these fractions are likely long-term N contributors when surface applied. Also not understood are the effects of multiple applications of litter upon the soil surface. It may be that increased clover litter contributes to residual soil N pools. Though we have not gone so far as to use presented data for such estimations. 72 CONCLUSIONS This research demonstrates important aspects of white clover decomposition, mainly that clovers are composed of a quickly decaying labile fraction. Given that clover populations are regenerative, N from white clover decay may be adequate to maintain associated turfgrasses. Modeling white clover decomposition may enable turfgrass researchers and professionals to more accurately predict nutrient contribution to associated grasses. Such information could be used to formulate an integrated N fertility program that includes both biologically fixed and synthetic N sources. The underpinnings of such a program have largely been overlooked by mainstream turfgrass research. Questions remain as to the consequences of synthetic N applications upon mixed turfgrass-legume swards. For instance, how would N application effect clover populations and levels of biologically fixed N? Future research should evaluate different leguminous species and their inherent decay patterns. White clover is not the only Trifolium species capable of cultivation within maintained turfgrass. Others include T. incarnatum, T. dubium, T. nigrescens, T. campestre, and T. aureum. Likewise, there are many leguminous species already present within low-maintenance turf, including Medicago and Kummerowia spp. Results of future studies may enable more appropriate species selections that sustain associated turfgrasses with much needed N, while simultaneously contributing other ecosystem services, such as pollinator habitat. Furthermore, research should evaluate periodicity of legume N contribution to warm and cool season grasses as well as possible ways to synchronize litter deposition with turfgrass N needs. All of these topics serve to advance 73 the effectiveness of alternative turf scenarios but also apply broadly to other sectors of conservation agriculture. 74 Table 11. Double exponential decay equations regresed on time (days) for mas, carbon (C), and nitrogen (N) -los from white clover incubated in liter bags under field conditions. Double exponential decay equations are of the form Y = Ae -k1t + Be -k2t , where Y = response, A aproximates the labile portion, B aproximates the recalcitrant portion, k1 and k2 are rate constants fited to the data, and t = time in days after aplication. Percent remaining data were normalized to initial day 0 aplications to facilitate aproximations of labile and recalcitrant portions. % Remaining Equation P > F ? R 2 adj Syx ? Mas March Y = 84.2842e -0.0367t + 19.4689e -2.8495E-12t <0.001 0.9641 7.3424 June Y = 80.6503e -0.1056t + 25.0928e -0.043t <0.001 0.9754 5.2408 December Y = 8.1369e -0.016t + 16.0402e -9.901E-13t <0.001 0.9695 4.6932 C March Y = 83.8157e -0.0354t +19.4509e -0.008t <0.001 0.9531 7.082 June Y = 79.3461e -0.1061t + 26.3596e -0.045t <0.001 0.974 5.3148 December Y = 93.8594e -0.0143t + 10.5685e -6.9172E-12t <0.001 0.9704 4.638 N March Y = 84.464e -0.0271t + 18.6349e -9.7045E-14t <0.001 0.968 5.537 June Y = 81.7382e -0.0938t + 21.7853e -0.036t <0.001 0.9807 4.6019 December Y = 52.2190e -9.970E-3t + 52.280e -9.9690E-3t <0.001 0.9596 5.0713 ? Significance of fit. ? Standard error of the estimate of Y on X. 75 Figure 1. Average daily soil temperatures at 10 cm depth at the study site and average daily air temperature at 1.5 m near the Auburn, AL study site. 76 Figure 1. Percent mas remaining from surface incubated white clover residue. Shapes represent mean ? 95% confidence intervals (CI?s) . Residue persistence was normalized to 10% ash fre dry weight of initial Day 0 aplications. Days to 50% decomposition (D50) values are presented on the horizontal axis with adjusted 95% CI?s as a means of comparing residue persistence acros aplication date. Mar: D50 CI = 2.7 d Jun: D50 CI = 10.2 d Dec: D50 CI = 14.5 d 12.1 d 29.0 d 57.4 d 77 Figure 12. Percent carbon (C) remaining from surface incubated white clover residue. Shapes represent mean ? 95% confidence intervals (CI?s) . Residue persistence was normalized to 10% ash fre dry weight of initial Day 0 aplications. Days to 50% decomposition (D50) values are presented on the horizontal axis with adjusted 95% CI?s as a means of comparing residue persistence acros aplication date. Mar: D50 CI = 2.2 d Jun: D50 CI = 9.6 d Dec: D50 CI = 17.4 d 1.0 d 27.8 d 60.5 d 78 Figure 13. Percent nitrogen (N) remaining from surface incubated white clover residue. Shapes represent mean ? 95% confidence intervals (CI?s) . Residue persistence was normalized to 10% ash fre dry weight of initial Day 0 aplications. Days to 50% decomposition (D50) values are presented on the horizontal axis with adjusted 95% CI?s as a means of comparing residue persistence acros aplication date. Mar: D50 CI = 2.3 d Jun: D50 CI = 10.5 d Dec: D50 CI = 17.0 d 10.9 d 37.0 d 74.1 d 79 Table 12. Persistence of white clover liter based upon predicted 95% confidence bands of double exponential decay equations (Table 1) regresed on time (days) for nitrogen (N), mas, and carbon (C) -remaining. Residue persistence was normalized to 10% ash fre dry weight of initial day 0 aplications. Mas Carbon Nitrogen C/N Date ? Days ? % Remaining Remaining Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI 1 Mar. 2010 0 103.8 6.7 103.3 5.0 103.1 4.6 10.4 8 Mar. 2010 7 84.7 3.8 84.8 4.4 8.5 2.9 10.0 15 Mar. 2010 14 69.9 5.6 70.3 4.8 76.4 4.4 9.6 29 Mar. 2010 28 49.6 5.5 50.1 5.5 54.2 4.7 9.6 26 Apr. 2010 56 30.3 7.0 30.1 6.8 37.2 5.4 8.4 21 Jun. 2010 112 20.9 7.7 19.4 7.0 2.7 5.6 8.9 1 Jun. 2010 0 105.7 3.1 105.7 3.5 103.5 2.1 9.4 8 Jun. 2010 7 62.9 3.6 63.3 3.6 63.6 3.1 9.1 15 Jun. 2010 14 42.0 3.3 42.7 3.3 42.7 3.0 9.2 29 Jun. 2010 28 26.4 4.0 27.3 4.3 25.6 3.7 9.8 27 Jul. 2010 56 19.9 4.3 20.7 4.4 18.2 4.2 10.4 21 Sep. 2010 112 15.5 6.1 15.9 6.1 14.6 5.3 10.0 1 Dec. 2010 0 104.2 4.9 104.4 7.0 104.4 3.7 10.5 8 Dec. 2010 7 94.5 3.9 95.5 4.0 97.4 3.5 10.3 15 Dec. 2010 14 85.9 6.4 87.4 5.2 90.8 4.1 10.1 29 Dec. 2010 28 71.4 4.1 73.5 4.8 79.0 4.1 9.8 26 Jan. 2010 56 50.8 4.4 52.7 5.6 59.8 4.4 9.3 23 Mar. 2010 112 29.8 4.9 29.5 4.7 34.2 5.1 9.1 ? Date of liter retrieval from field conditions. Day 0 liter was truly replicated and placed within the field. ? Days after initial aplication. 80 White Clover Inclusion within a Bermudagrass Lawn: Effects of Supplemental Nitrogen upon Botanical Composition and Nitrogen Cycling INTRODUCTION Total turfgrass cultivation within the contiguous United States occupies 163,800 km 2 (?35,850 km 2 ; Milesi et al., 2005), an area nearly the size of the state of Florida. Turfgrass is frequently criticized for its negative environmental effects, due in part to the use of agrichemicals and often-limited natural resources (Milesi et al., 2005; Robbins and Birkenholtz 2003; Robbins et al., 2001). However, there are many turfgrass scenarios, that when managed properly, can be a sustainable asset to modern communities. Benefits of turfgrass are well documented and include: erosion control, increased water infiltration, reduced nutrient leaching, aesthetics, and carbon sequestration (Beard and Green 1994). Nitrogen (N) is essential to turfgrass health and quality (Beard 1973; Turgeon 2002), with commercial-lawn N requirements within the southern United States ranging from less than 5 g m - 2 year -1 for bahiagrass (Paspalum notatum Fluegge) and centipedegrass [Eremochloa ophiuroides (Munro) Hack.] to almost 30 g m -2 year -1 for hybrid bermudagrass [Cynodon dactylon ! C. transvaalensis; Duble 1996). Improper N fertilization can lead to negative environmental effects. N loss from turfgrass may contribute to the eutrophication of aquatic and terrestrial ecosystems as well as lead to elevated nitrate (NO 3 ) levels in drinking water (Robbins and Birkenholtz 2003; Robbins et al., 2001). For these reasons, and due to the economic costs associated with labor and materials, reducing N application to certain turfgrass scenarios is desirable. White clover (Trifolium repens L.) inclusion within turfgrass is a proposed means of increasing the sustainability of certain low maintenance turfgrass scenarios due to its ability to 81 biologically fix N and transfer it to associated grasses (Dudeck and Peacock, 1983; Sincik and Acikgoz, 2007; McCurdy et al., 2013). Little is known about white clover N contribution to maintained turfgrass swards. Sincik and Acikgoz (2007) reported that white clover fixed greater than 25 g N m -2 year -1 and contributed between 4.2 to 13.7% of that total N to three associated cool season turfgrass species, resulting in improved color ratings. Beyond these estimates, most are limited to cool season pasture scenarios. Historically, grass plus clover mixtures have proven to be important pasture systems that supply high-quality grazing for animals while simultaneously improving soil fertility (Lampkin 2002). Estimates of N fixation for grass plus clover pastures are roughly 10 to 25 g N m -2 year -1 (Ledgard and Steele 1992; McNeil and Wood 1990; Whitehead 1995). Using the 15 N transfer method, McNeil and Wood (1990) estimated white clover N fixation to be 15.5 g N m -2 year -1 , with 28% having been transferred to associated ryegrass. N transfer occurs indirectly through excreted N and mineralization of nodules, roots, and foliage (Brophy et al., 1987; Dubach and Russelle 1994; Jensen 1996; Wardle and Greenfield 1991). Previous research indicates that root nodules are a significant N source in mixed swards; with root nodule N concentrations ranging from 4.8 to 9.0% of root dry matter (Chu and Robertson 1974; Wardle and Greenfield 1991). Decomposition of foliage is another means of N transfer, as above ground white clover dry matter has been reported to be 4.0 to 4.9% N (McCurdy et al., 2013; Sincik and Acikgoz 2007). White clover foliage is composed of a quickly decaying labile fraction; however, decay is highly dependent upon time of year, presumably due to soil temperature and microbial decay mechanisms (McCurdy et al., 2013). Even in persistent stands of legumes, biological N fixation varies, largely due to the relative composition of turfgrass swards and soil N availability (Crush et al., 1982). Fixation is 82 highly dependent upon the relative level of nodulation occurring in root tissues and activity of the bacteria within. Most research indicates that high soil N concentrations inhibit nodule growth and development. Macduff et al. (1996) observed that the ratio of nodule to root dry-weights was 1:6 in white clover without NO 3 treatment but decreased with applications of NO 3 . Similarly, it is well documented that increasing N fertilization decreases clover density and allows the grass portion of the sward to outcompete clover (Frame and Boyd 1987; Pederson 1995; Sincik and Acikgoz 2007). Other factors affecting biological N fixation include absorption of photosynthetically active radiation, carbon assimilation rates, and allocation of photosynthate to roots (Lie, 1971). At low N levels, white clover leaves have a higher photosynthetic capacity than do competing perennial ryegrass; however, at higher N levels the opposite is true (Faurie et al., 1996). Increased light interception at low N levels can be attributed to a greater leaf area index in the upper canopy of the grass plus clover sward as well as clover?s ability to avoid shade by increasing petiole length (Davies and Evans 1990; Faurie et al., 1996; Woledge et al., 1992). Another factor affecting N fixation is soil temperature. Frame and Newbould (1986) found that a minimum temperature of 9?C was necessary for active N fixation by Rhizobium. It has also been reported that temperatures required for nitrogenase activity range from 13 to 26?C (Halliday and Pate 1976). Our research was prompted by the numerous knowledge gaps pertaining to white clover inclusion within warm-season turf scenarios. We sought to evaluate the effects of white clover inclusion within a maintained hybrid bermudagrass lawn. Our objectives were 1) quantify the effects of clover inclusion upon sward biomass, 2) evaluate the effects of long term supplemental 83 N upon clover establishment and sward composition, 3) estimate clover N fixation and N transfer to associated turfgrass. MATERIALS AND METHODS Study Design and Field Conditions A 3-year study was conducted to evaluate the effects of white clover (Trifolium repens L.) inclusion within a hybrid bermudagrass (Cynodon dactylon ! C. transvaalensis) lawn. The study was conducted at the Auburn University Turfgrass Research Unit, (32?34?40? N, 85?29?57? W) in Auburn, AL, on a Marvyn sandy loam (fine-loamy, Kaolinitic, thermic Typic Kanhapludult) soil with pH 6.3 (1:1 soil/H 2 O). Treatments were arranged as randomized complete blocks (4 replicates). Treatment factors were clover inclusion and supplemental N rate. Plot size was 2 m 2 . White clover [c.v. Dutch White (Main Street Seed and Supply, Bay City, MI)] was seed-established October 2009, 2010, and 2011 (1.5 g pure live seed m -2 ). Supplemental N (0, 0.5, 1, 2, 4, and 8 g N m -2 ) was applied monthly, April to August, during 2010, 2011, and 2012 as CaNO 3 , with Ca applied to uniformity via CaSO 4 . During 2011 and 2012, the area received 1 cm of supplemental overhead irrigation twice per week from March to September in order to insure adequate clover and bermudagrass growth. Air temperature at a 1.5 m height and soil temperature at a 10 cm depth were obtained from a nearby weather station (32?36?00? N, 85?30?00? W; elevation 199 m) in Auburn, AL (AWIS, 2013). Clipping Biomass and Sampling Trifoliate leaves were counted within three 730 cm 2 sub-samples per 2.0 m 2 experimental unit on 20 April 2010, 2011 and 2012 as a means of quantifying spring clover density (trifoliate leaves m -2 ). Plots were harvested and collected at a 2.5 cm mowing height 1 month after fertilization (May to September) via reel mower. Biomass was air-dried at 60?C for 1 wk and 84 was weighed for dry-matter determination. Botanical composition of mixed swards was determined by partitioning three, 3 g sub-samples into their constituent grass or clover parts. A 20 g sample of whole harvest biomass from each plot was ground to pass a 16-mesh sieve for N analysis by LECO TruSpec CN (Leco Corp, St. Joseph, MI, USA). Remaining biomass was returned to the turfgrass surface in order to mimic normal turfgrass mowing practices. However, due to processing constraints, biomass could not be returned immediately and was instead returned to respective experimental units one month later following the subsequent harvest or one wk after the last harvest of the year. Nitrogen Fixation and Transfer Nitrogen fixation was calculated using the difference method, by subtracting N-yield of grass-alone plots from the total N-yield of grass plus clover mixtures. Furthermore, the apparent N transfer was estimated as the difference between grass-alone N-yield of mixtures and that of grass-alone monocultures. Grass N-yield of mixtures was calculated based upon estimates of grass-clover composition, with error propagated throughout. Statistical Analysis Analysis of variance (ANOVA) was conducted. Means, standard errors, and statistical significance of treatments were determined at the 95% confidence level using mixed models procedures within Proc Glimmix (SAS Institute, Raleigh, NC, USA). Means were separated based upon adjusted 95% confidence intervals, which allowed for multiple comparisons by protecting family-wise error rate (Littell et al., 2006). Overlapping limits indicated lack of significant difference between responses. Least squares estimates for nonlinear models were determined within SigmaPlot 11 (Systat Software, Chicago, IL, USA) using the Marquardt- 85 Levenberg algorithm to provide the best fit. Initial parameter ranges were selected with a maximum of 200 fits and 200 iterations. Data were fit to the quadratic model y = ax 2 + bx + c, where a is the quadratic coefficient, b is the linear coefficient, c is the estimated response at 0 g N m -2 year -1 , and x is g N m -2 year -1 . In most cases, quadratic models minimized residual sums of squares and produced comparatively lower residual mean squares, standard errors, and PRESS statistics as well as better coefficients of determination (R 2 adj ) than linear models. For brevity, comparisons of quadratic and linear exponential models are omitted. However, instances where quadratic models could be collapsed into linear models are generally indicated by the presence of a values near zero. RESULTS AND DISCUSSION Clover Establishment Spring clover densities decreased as the study progressed, with 875 > 605 > 451 trifoliate leaves m -2 for 2010, 2011, and 2012, respectively This apparent trend may be one of chance or may be due to the effects of long term clover cultivation. Previous research indicates that clover persistence varies greatly due to soil conditions. In their review of N fixation of grass-legume pastures, Ledgard and Steele (1992) report that fixation was reduced due to dry soil conditions, acid soils, and the ?pest/disease complex.? October seeding rates were equal throughout the study; however, in simultaneous research at a similar location, we also noticed declining clover establishment during this study. These effects were likely due to delayed bermudagrass dormancy during years two and three. Similarly, a delayed spring in 2011 likely had similar effects (Figure 14). When regressed across supplemental N rates applied during the previous 86 season, 2011 and 2012 clover establishment did not differ. This implies that seeded clover establishment was not suppressed by prior supplemental N applications. Clipping Biomass During 2010, grass biomass (80.3 g m -2 year -1 ) was smaller than that of 2011 or 2012 (484 and 526 g m -2 year -1 , respectively; Table 13). Similarly, 2010 grass plus clover biomass (105 g m -2 year -1 ) was smaller than that of 2011 (690 g m -2 year -1 ) and 2012 (791 g m -2 year -1 ), with 2012 biomass having been the largest of the three years. First year biomass was much smaller than that of subsequent years due to a lack of supplemental water prior to, and during, the 5-month harvest period. Supplemental irrigation was applied twice weekly during the following 2011 and 2012 seasons in order to produce adequate harvests to determine botanical composition. Within years, biomass differed due to date by supplemental N rate interactions (data not shown). However, for simplicity, we present annual effects of supplemental N by sward type (grass alone or grass plus clover). Sward types responded differently to supplemental N (Figure 15) Generally, grass plus clover biomass was greater than that of grass alone plots; however, 2011 and 2012 grass plus clover biomass were equal to that of grass alone swards at the highest supplemental N rate (Figure 15). This trend was consistent across date of harvest (Data not shown). Biomass data for both grass and mixed swards was regressed with supplemental N levels using a quadratic model (Table 14). The quadratic model, y = ax 2 + bx + c, provides an estimate of sward biomass when no supplemental N is applied (c) and estimates the response due to increasing supplemental N (b). Unlike a simple linear model, the quadratic model?s quadratic coefficient (a) may provide modest insight into the diminishing returns due to increasing N 87 levels by way of its parabolic shape. The x-coordinate of the parabola?s vertex is predicted by the equation x = -b / 2a, which estimates the theoretical N level at which maximum biomass may be realized. In cases where the parabola is upward opening, c values are positive. In contrast, where the parabola is downward opening, c values are negative, indicating a plateau effect or deleterious effect of N application beyond a certain point. We present means corresponding to treatment effects as well as theoretical N levels corresponding to maximum biomass. 2010 Biomass The quadratic response of 2010 biomass to increasing N levels is less pronounced than other years, presumably due to the aforementioned lack of supplemental irrigation during the first season of the study. When 0 to 20 g N m -2 year -1 was applied, grass biomass was ! 85 g m -2 year -1 (Figure 15; Table 13). Grass biomass was greater (137 g m -2 year -1 ) when 40 g N m -2 year -1 was applied. A plateau effect was not evident within the quadratic model of grass-alone biomass, as the upward opening parabola implies that N, even at the highest supplemental rate, may have been limiting to bermudagrass growth (Figure 15). In this instance, the relatively low quadratic coefficient (Table 14) indicates that a simple linear model would have been adequate. Grass plus clover biomass generally increased with increasing N levels. However, grass plus clover biomass appeared to plateau (169 g m -2 year -1 ) at the 20 to 40 g N m -2 year -1 level, which was confirmed by the theoretical estimate 33.9 g N m -2 year -1 provided by the quadratic model. 2011 Biomass Grass alone biomass increased from 281 to 948 g m -2 year -1 with supplemental N (Table 13; Figure 15). The quadratic model estimates a maximum biomass at the 64 g N m -2 year -1 rate (Table 14). Grass plus clover biomass was larger than grass alone biomass, presumably due to mixed sward composition and fixed N. Grass plus clover biomass increased from 458 to 965 g 88 m -2 year -1 , with the highest biomass due to the 40 g N m -2 year -1 rate; although this level was equaled by that of the 20 g N m -2 year -1 rate (864 g m -2 year -1 ). Similarly, the quadratic response indicates a theoretical maximum biomass at 35 g N m -2 year -1 . 2012 Biomass Grass alone biomass increased from 271 to 964 g m -2 year -1 with increasing supplemental N (Figure 15; Table 13). The quadratic model estimates a maximum biomass at 48 g N m -2 year -1 (Table 14). Grass plus clover biomass was again larger than that of grass alone swards, increasing from 572 to 990 g m -2 year -1 . The quadratic response indicates a maximum biomass at 33 g N m -2 year -1 . Biomass Composition The composition of grass plus clover swards was quantified throughout 2011 and 2012 seasons by subsampling total clippings. Clover biomass of 2011 (180 g m -2 year -1 ) and 2012 studies (231 g m -2 year -1 ) accounted for 23 and 29% of 2011 and 2012 biomass, respectively (Table 13). Contrary to previous reports, clover biomass was not seriously affected by increasing N rate. Only 2011 clover biomass was affected by supplemental N rate, with the highest rate of supplemental N having reduced clover biomass relative to the 20 g N m -2 year -1 rate. Furthermore, grass biomass of grass plus clover swards (data not shown) was greater than that of grass-alone plots at supplemental N rates ! 10 g N m -2 year -1 . However, grass of mixed swards was reduced relative to grass alone plots at 20 and 40 g N m -2 year -1 rates. Nitrogen Fixation Similar to Elgersma et al. (1998), white clover-derived N was calculated using the difference method, by subtracting N-yield of grass-alone plots from the total N-yield of grass plus clover mixtures. During the 3-year study, N fixation was estimated to be 6.6 g m -2 year -1 89 regardless of supplemental N rate. However, N fixation differed due to study years (0.9, 8.0, 10.9 g N m -2 year -1 for 2010, 2011, and 2012, respectively), with an apparent increase in fixation as years progressed (Table 15). This observation is most likely due to the relatively slow soil organic matter degradation and lagging N availability between years (i.e. clover residue mineralized N at a pace detected only by multiple sampling years). We have previously reported that, when applied at 500 g fresh weight m -2 (equivalent to approximately 75 to 100 g dry weight m -2 ), more than half of available N was mineralized between 10 and 73 days after application, depending upon application timing (McCurdy et al., 2013). However, in the present study, effective clover biomass accumulation exceeded the dry weight equivalents of our previous work and suggests that some residue remains within the system acting as a long term N contributor. The authors acknowledge several limitations pertaining to the N difference method. A basic assumption of the N difference method is that litter mineralization and subsequent N immobilization are the same for all treatment scenarios (Hauck and Bremner, 1976). However, due to the ?priming effect,? it is often reported that fertilized plots (whether through biological fixation or applied N) have increased N availability beyond the levels of that applied (Rao et al., 1992). This may be due to a number of factors, including: increased microbial activity (Westerman and Kurtz, 1973), acid hydrolysis of soil organic matter (Turchin, 1964), and increased root growth in fertilized plots (Olson and Swallow, 1984) possibly increasing nutrient access. In instances where such faults cannot be accounted for, the difference method may significantly over estimate apparent N fixation. On the contrary, the difference method may also underestimate apparent N fixation, in part due to unaccounted loss through volatility and leaching. The method may also underestimate N fixation when soil N is sufficient to meet bermudagrass needs without additional fertilizer 90 (Varvel and Peterson, 1990). During preparations for this study, plots were maintained at considerably lower N levels than recommended for actively growing bermudagrass, yet turfgrass cover was at no time diminished in 0 g N m -2 year -1 treated plots. This may indicate adequate soil N and mineralizable N sources within the turfgrass canopy that, in this instance, led to an under- estimate of N fixation within clover-included plots. The method also assumes that bermudagrass and white clover take up soil N at the same rate, which is unlikely, as bermudagrass is considered an almost voracious N consumer while white clover most often abounds without supplemental N. 2010 N Fixation During 2010, N fixation was 0.9 g m -2 year -1 (Table 15). N fixation was generally suppressed at the high and low extremes of supplemental N rate (Figure 15), which is evidenced by estimates of maximum N fixation at the supplemental N rate of 24 g N m -2 year -1 (Table 14). N fixation generally decreased as the summer progressed (data not shown), presumably because early season harvests measured N fixation that had occurred prior to the five month long harvest season. Another explanation may be decreased clover populations and increased competitiveness of bermudagrass as the harvest season progressed. 2011 N Fixation During 2011, N fixation was 8.0 g m -2 year -1 , which was significantly larger than that of 2010 (Table 15). However, like 2010, N fixation was generally suppressed at the high and low extremes of supplemental N rate (Figure 15). This is again evidenced by estimates of maximum N fixation at the supplemental N rate of 17 g N m -2 year -1 (Table 14). Like that of the previous season, N fixation waned towards the latter summer months; although, May N fixation was equivalent to that of August and September (data not shown). 91 2012 N Fixation During 2012, N fixation was 10.9 g m -2 year -1 , which was significantly larger than that of previous years (Table 15). Like that of previous years, N fixation was suppressed at high supplemental N rates (Figure 15), evidenced by a regression model decreasing across application rates and a maximum estimate of N fixation at the supplemental N rate of -10.0 g N m -2 year -1 (Table 14). Like that of previous seasons, N fixation waned towards the latter summer months. N Transfer Nitrogen transfer from white clover to associated bermudagrass was estimated during 2011 and 2012 harvest periods (Table 15). During the two-year period, N transfer was 2.3 g m -2 year -1 regardless of supplemental N rate, which amounted to 24% of N fixed during those two harvest years. Our estimate appears to be similar to that reported by McNeil and Wood (1990) who estimated 28% N transfer from white clover to associated perennial ryegrass. However, it is slightly higher than the estimated 4.2 to 13.7% N transfer from white clover to three cool season turfgrasses reported by Sincik and Acikgoz (2007). As with fixation estimates, we admit that our results may overestimate N transfer due to a lag between early season N fixation before bermudagrass green-up. Simultaneously, the complex nature of N availability from fixed N in the relatively cool soil temperatures is poorly understood. We have previously reported slower N availability from winter and spring applied clover biomass relative to that of summer-applied biomass (McCurdy et al., 2013). 2011 Nitrogen Transfer Nitrogen transfer to associated bermudagrass was 3.9 g m -2 year -1 during 2011, which is equivalent to 49% of fixed N (Table 15). The quadratic model of 2011 N transfer was upward opening (Figure 15), seemingly due to the somewhat lower N transfer estimate of the 20 g N m -2 92 year -1 rate (Table 15). Despite this phenomenon, we propose that N transfer was less affected by increasing N rate than N fixation during 2012. This is evidenced by the nearly overlapping N fixation and transfer estimates at the highest level of supplemental N during 2011. N transfer did, however, differ due to supplemental N rate within harvest months (Table 15). Yet unlike N fixation, N transfer generally did not decrease throughout the harvest period (data not shown). This may be due to a priming effect from supplemental N and the resulting N mineralization from accumulated biomass, as well as the delayed release of fixed N from organic matter within the gradually warming soils. 2012 Nitrogen Transfer Nitrogen transfer to associated bermudagrass was only 0.6 g m -2 year -1 during 2012, which is equivalent to 6% of fixed N (Table 15). However this estimate was disproportionately affected by a single highly negative estimate (-10.3 g m -2 year -1 ) at the 40 g N m -2 year -1 rate. It is not clear how such a negative estimate can exist. If this estimate was removed, the average transfer amounted to more than 25%. In spite of this, N transfer appears to have been suppressed at higher supplemental N rates (Figure 15). This was further evidenced by the quadratic model, which estimates a maximum N transfer at a supplemental N rate of 1.25 g N m -2 year -1 (Table 14). CONCLUSIONS Year one results were severely impacted by the lack of supplemental irrigation, which suggests the utility of white clover inclusion as a means of sustaining low maintenance turfgrass may be limited in drier climates. However, our results demonstrate that white clover inclusion is a viable option for sustainably supplementing the N requirements of warm-season grass swards. Grass plus clover swards yielded higher clipping biomass than grass-alone swards during 93 irrigated study years, which was evidence of enhanced bermudagrass growth due to biological N fixation. Likewise, grass biomass of mixed swards was increased relative to that of grass-alone plots at supplemental N rates ! 10 g N m -2 year -1 . N fixation was estimated to be 6.6 g m -2 year -1 during the 3-year study, with an apparent increase in fixation as years progressed. This estimate is considerably lower than the roughly 10 to 25 g N m -2 year -1 reported in various cool season scenarios (Ledgard and Steele 1992; McNeil and Wood 1990; Whitehead 1995). N transfer to the associated bermudagrass sward was estimated to be 24% across the latter two years of the study, which is comparable to previous estimates within cool-season pastures and larger than those reported within cool-season turfgrass by Sincik and Acikgoz (2007). Unlike previous research, our results indicate that N fixation was suppressed at low supplemental N rates. These results may indicate the N demands of clover establishment and that small amounts of supplemental fertility are needed to maximize N fixation. However, it is also possible that supplemental N increased soil organic matter decay and N uptake by associated bermudagrass, effectively biasing estimates of N fixation within fertilized plots. This is again a liability associated with the N difference method. Upper extremes of supplemental N were also deleterious to N fixation, which agrees with previous indications of clover decline in the presence of high soil N levels. Our research was limited to white clover inclusion; however, other legumes should be evaluated for their utility within warm-season turfgrasses. Multiple Trifolium species are common amongst pasture and turfgrass scenarios of the southeastern U.S. Alternatives include: T. incarnatum, T. dubium, T. nigrescens, T. campestre, and T. aureum. Likewise, warm-season legumes, such as Kummerowia and Arachis, may provide more timely N release to associated warm-season turfgrasses. 94 Figure 14. Average daily air and soil temperature as well as average daily precipitation for the 2009-2010, 2010-2011, and 2011-2012 bermudagrass-white clover growing seasons. 95 Table 13. Gras alone and gras plus white clover biomas, as wel as clover portion of the mixed sward, ? 95% confidence interval (CI) as afected by suplemental N. Means ? CI are presented to alow treatment separation among similar response variables. Total Biomas (g m -2 year -1 ) Clover portion (% Total Biomas) Suplemental N ? g m -2 year -1 2010 2011 2012 2011 2012 Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI Gras plus Clover 0.0 49 c 20 458 e 48 572 e 56 23 ab 6 31 a 9 2.5 72 bc 20 710 c 53 723 cd 56 25 ab 6 35 a 9 5.0 78 bc 20 587 d 53 682 de 58 24 ab 6 36 a 9 10.0 103 b 20 74 bc 52 836 bc 56 25 ab 6 28 a 9 20.0 169 a 20 864 ab 54 927 ab 56 29 a 6 25 a 9 40.0 169 a 20 965 a 52 90 a 56 13 b 6 19 a 9 Yearly Total 105 C ? 53 690 B 39 791 A 38 23 B 2 29 A 2 Gras alone 0.0 60 b 23 281 d 48 271 e 56 2.5 6 b 23 309 d 48 291 de 56 5.0 56 b 23 347 d 48 364 d 58 10.0 85 b 23 473 c 48 503 c 56 20.0 78 b 23 739 b 53 752 b 56 40.0 137 a 23 948 a 51 964 a 56 Yearly Total 80 B 65 484 A 46 526 A 38 ? Suplemental N was aplied during five consecutive months (April to August). ? Yearly totals are comparable amongst years, respective to response variables. 96 Figure 15. Gras alone and gras plus clover biomas, as wel as estimated biological N fixation and N transfer to asociated grases, regresed upon yearly suplemental N levels. Transfer was not calculated in 2010. Data were fit to the quadratic model y = a + bx + cx 2 , where a is the estimated response at 0 g N m -2 year -1 , b is the linear coeficient, c is the quadratic coeficient, and x is g N m -2 year -1 aplied over five months of active bermudagras growth from April to August as CaNO 3 . Means ? 95% confidence intervals are presented to alow treatment separation among similar response variables. 97 ! ! ! Table 14. Gras alone and gras plus clover biomas, as wel as estimated biological N fixation and N transfer to asociated grases, regresed upon yearly suplemental N levels. Transfer was not calculated in 2010. Data were fit to the quadratic model y = a + bx + cx 2 , where a is the estimated response at 0 g N m -2 year -1 , b is the linear coeficient, c is the quadratic coeficient, and x is g N m -2 year -1 aplied over five months of active bermudagras growth from April to August as CaNO 3 . Response Equation P > F ? R 2 adj Syx ? Maximum response ? 2010 G biomas 61.65 + 0.54x + 0.03x 2 < 0.001 0.193 51.17 -9.0 ? G + C biomas 47.76 + 7.45x - 0.10x 2 < 0.001 0.331 62.15 3.9 N fixation -0.37 + 0.2x - 0.045x 2 < 0.001 0.1394 2.06 24.4 2011 G biomas 257.84 + 24.34x - 0.191x 2 < 0.001 0.3040 38.73 63.7 G + C biomas 517.2 + 25.84x - 0.369x 2 < 0.001 0.157 363.43 35.0 N fixation 9.03 + 0.34x - 0.010x 2 < 0.001 0.0278 9.1 17.0 N transfer 3.80 + 0.09x + 0.02x 2 0.0498 0.019 7.8 -2.5 2012 G biomas 234.51 + 31.67x - 0.33x 2 < 0.001 0.3676 33.21 47.6 G + C biomas 607.9 + 24.25x - 0.370x 2 < 0.001 0.2129 265.29 32.8 N fixation 8.72 - 0.06x - 0.03x 2 < 0.001 0.0819 8.02 -10.0 N transfer 3.64 + 0.02 - 0.08x 2 < 0.001 0.2520 7.72 1.25 ? Significance of regresion fit. ? Standard eror of the estimate of Y on X. ? N rate (g m -2 year -1 ) to obtain theoretical maximum response. ? Negative numbers represent a minimum response due to an upward facing parabola. 98 ! ! Table 15. Nitrogen Fixation and Nitrogen Transfer ? 95% confidence interval (CI) as afected by suplemental N. Means ? CI are presented to alow treatment separation among similar response variables. Nitrogen Fixation (g m -2 year -1 ) Nitrogen Transfer (g m -2 year -1 ) Suplemental N ? g m -2 year -1 2010 2011 2012 2011 2012 Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI Mean ? 95% CI 0.0 -0.2 d 0.4 6.7 b 0.1 10.4 b 1.0 2.8 cd 0.9 2.9 b 0.9 2.5 0.2 cd 0.4 9.7 a 0.1 13.8 a 1.1 5.0 ab 0.9 5.0 a 0.9 5.0 0.7 bcd 0.4 9.5 a 0.1 1.5 b 1.0 4.6 abc 0.9 3.3 ab 0.9 10.0 0.8 bc 0.4 9.6 a 0.2 1.0 b 1.0 6.4 a 0.9 2.7 b 0.9 20.0 2.6 a 0.4 9.0 a 0.1 8.4 c 1.0 1.5 d 0.9 0.2 c 0.9 40.0 1.1 b 0.4 3.5 c 0.2 6.1 d 1.0 3.2 bcd 1.0 -10.3 d 1.0 Yearly Total 0.9 C ? 0.7 8.0 B 0.4 10.9 A 0.4 3.9 A 0.3 0.6 B 0.3 ? Suplemental N was aplied during five consecutive months (April to August). ? Yearly totals are comparable among years, respective to response variable. 99 Literature Cited Abraham, C M., D.W. Held, and C. Wheeler. 2010. Seasonal and diurnal activity of Larra bicolor (Hymenoptera: Crabronidae) and potential ornamental plants as nectar sources. Appl. Turf. Science Accessed September 9, 2013 at http://www.plantmanagementnetwork.org/pub/ats/research/2010/nectar/. AWIS. 2013. Alabama Mesonet, AWIS Weather Services, Inc., Alabama Cooperative Extension Service, Alabama Agricultural Experiment Station, Auburn University, Auburn, AL. Beard, J.B. 1973. Turfgrass: Science and Culture; Prentice-Hall: Englewood Cliffs, NJ. Beard, J.B., and R.L. Green 1994. The roles of turfgrasses in environmental protection and their benefits to humans. Journal of Environ Quality 23:452?460. Beran, D., R.A. Masters, and R.E. Gaussoin. 1999. Grassland legume establishment with imazethapyr and imazapic. Agronomy Journal 91:592-596. Bowran, D.G. 1993. Differential response of four legume species to broad-leaf herbicides. Proceedings of the Tenth Australian Weeds Conference pp. 181-185. Braman, S.K., A.F. Pendley, W. Corley. 2002. Influences of commercially available wildflower mixes on beneficial arthropod abundance and predation in turfgrass. Environmental Entomology 31:564-72. Brewin, J.N. 1991. Development of the legume root nodule. Annual Review of Cell Biology. 7:191-226. 100 Brink, G.E., and T.E. Fairbrother. 1991. Yield and quality of subterranean clover and white clover-bermudagrass and tall fescue associations. Journal of Production Agriculture 4:500-504. Brink, G.E. and D.E. Rowe. 1993. Growth of white clover clones in monoculture and contrasting bermudagrass swards. Crop Science 33:1091-1094. Brophy, L.S., G.H. Heichel, and M. P. Russelle. 1987. Nitrogen transfer from forage legumes to grass in a systematic planting design. Crop Science 27:753-758. Buchanan, B.B., W. Gruissem, and R.L. Jones. 2000. Biochemistry and Molecular Biology of Plants. Sixth Edition. John Wiley and Sons, NJ. Burgett, M., and G.C. Fisher. 1980. Recovery of Penncap-M from foraging honey bees and pollen storage cells. Environmental Entomology 9:430-431. Burns, J.C., L. Good, H.D. Gross, and A.C. Linnerud. 1973. Cow and calf gains on ladino clover-tall fescue, grazed alone and with Coastal bermudagrass. Agronomy Journal 65:877-880. Burton, G.W., and E.H. DeVane. 1992. Growing legumes with Coastal bermudagrass in the lower Coastal Plains. Journals of Production Agriculture. 5:278-281. Busey, P., and G.H. Snyder. 1993. Population outbreak of the southern chinch bug is regulated by fertilization. International Turfgrass Society Journal 7:353-357. Carlisle, R.J., V.H. Watson, and A.W. Cole. 1980. Canopy and chemistry of pasture weeds. Weed Science. 28:139-142. 101 Carmer, S.G., W.E. Nyquist, and W.M. Walker. 1989. Least significant differences for combined analysis of experiments with two- or three-factor treatment designs. Agronomy Journal 81:665-672. Ceballos, R., G. Palma, H. Brevis, F. Ortega, and A. Quiroz. Effect of five postemergence herbicides on red clover shoot and root growth in greenhouse studies. Phytoprotection 85:153-160. Chu, A.C.P., and A.G Robertson. 1974. The effects of shading and defoliation on nodulation and nitrogen fixation by white clover. Plant and Soil 41:509-519. CIA World Factbook. 2013. Accessed September 9, 2013 at https://www.cia.gov/library/publications/the-world-factbook/fields/2212.html. Clark, S.A., and H.K. Mahanty. 1991. Influence of herbicides on growth and nodulation of white clover, Trifolium repens. Soil Biology and Biochemistry 23:725-730. Cochran, V.L. 1991. Decomposition of barley straw in a subarctic soil in the field. Biology and Fertility of Soils 10:227-232. Conrad, J.D., and J. F. Stritzke. 1980. Response of arrowleaf clover to postemergence herbicides. Agronomy Journal 72:670-672. Cronquist, A. 1981. An integrated system of classification of flowering plants. The New York Botanical Garden. Columbia University Press. New York. Crush, J.R., G.P. Cosgrove, and R.W. Broughman. 1982. The effect of nitrogen fertilizer on clover nitrogen fixation in an intensively grazed Manawatu pasture. New Zealand Journal of Experimental Agriculture 10:395-399. 102 Dalsgaard, K. 1978. A study of the isolation and characterization of the saponin Quil A: evaluation of its adjuvant activity with special reference to the application in the vaccination of cattle against foot-and-mouth disease. Acta Veterinaria Scandinavica. 69S:1. Davidson, A.W., and D.A. Potter. 1995. Response of plant-feeding, predatory, and soil- inhabiting invertebrates to Acremonium endophyte and nitrogen fertilization in tall fescue turf. Journal of Economic Entomology 88:367-379. Davies, A. and E. Evans. 1990. Axillary bud development in white clover relation to defoliation and shading treatment. Annals of Botany 55:171-178. DiTomaso, J.M. 2000. Invasive weeds in rangelands: Species, impacts, and management. Weed Science 48:255-265. Doyle, J.J., J.A. Chappill, C.D. Bailey, and T. Kajita. 2000. Towards a comprehensive phylogeny of legumes: evidence from rbcL sequences and non-molecular data. Pages 1 -20 in Advances in legume systematics, part 9, (P. S. Herendeen and A. Bruneau, eds.). Royal Botanic Gardens, Kew, UK. Dubach, M. and M.P. Russelle. 1994. Forage legume roots and nodules and their role in nitrogen transfer. Agronomy Journal 86:259-266. Duble, R.L. 2004. Turfgrasses: their management and use in the southern zone, second ed. Texas A&M University Press, College Station, TX. Dudeck, A.E., and C.H. Peacock. 1983. Rate, method, and time of overseeding white clover on bermudagrass during the winter. Proceedings of the Florida State Horticultural Society 96:159-161. 103 Eastwood, J., E. Vavaour, and J. Baines. No Date. World Health Organization Food Additive Series: 48. Safety evaluation of certain food additives and contaminants: Quillaia extracts. Accessed online September 6, 2013 at http://www.inchem.org/documents/jecfa/jecmono/v48je03.htm Elgersma, A., M. Nassiri, and H. Schlepers. 1998. Competition in perennial ryegrass- white clover mixtures under cutting. 1. Dry-matter yield, species composition and nitrogen fixation. Grass and Forage Science 53:353-366. Elliot, J.G. 2006. The reaction of Trifolium repens (L.) in grassland to 2,4-D, MCPA and mixtures of 2,4-D and MCPB. Weed Research 1:184-195. Ellison, N.W., A. Liston, J.J. Steiner, W.M. Williams, and N.L. Taylor. 2006. Molecular phylogenetics of the clover genus Trifolium (Leguminosae). Molecular Phylogenetics and Evolution 39:688-705. Evers, G.W., W.J. Grichar, C.L. Pohler, and A.M. Schubert. 1993. Tolerance of three annual forage legumes to selected postemergence herbicides. Weed Technology 7:735-739. Faurie, O., J.F. Soussana, H. Sinoquet. 1996. Radiation interception, partitioning and use in grass-clover mixtures. Annals of Botany 77:35-45. Fletcher, W., and J.C. Raymond. 1956. Toxicity and breakdown of ?hormone? herbicides. Nature. 178:151-152. Frame, J., and A.G. Boyd. 1987. The effect of fertilizer nitrogen rate, white clover variety and closeness of cutting on herbage productivity from perennial ryegrass/white clover swards. Grass and Forage Science 45:57-64. 104 Frame, J., and P. Newbould. 1986. Agronomy of white clover. Advances in Agronomy 40:1-88. Gels, J.A., D.W. Held, and D.A. Potter. 2002. Hazards of Insecticides to the Bumble Bees Bombus impatiens (Hymenoptera: Apidae) Foraging on Flowering White Clover in Turf. Ecotoxicology. pp. 722-728. Frank, S.D., and P.M. Shrewsbury. 2004. Effect of conservation strips on the abundance and distribution of natural enemies and predation of Agrotis ipsilon (Lepidoptera: Noctuidae) on golf course fairways. Environmental Entomology 33:1662-1672. Hadas, A., L. Kautsky, M. Goek, and E.E. Kara. 2004. Rates of decomposition of plant residues and available nitrogen in soil, related to residue composition through simulation of carbon and nitrogen turnover. Soil Bio. and Biochem. 36:255-266. Hadden, E.J. 2005. Clover improves your lawn. Accessed September 8, 2013, at http://www.lesslawn.com/articles/article1061.html. Hager, A.G., L.M. Wax, G. A. Bollero, and E. W. Stoller. 2003. Influence of diphenylether herbicide application rate and timing on common waterhemp (Amaranthus rudis) control in soybean (Glycine max). Weed Technology 17:14- 20. Halliday, J., and J.S. Pate. 1976. The acetylence reduction assay as a means of studying nitrogen fixation in white clover under sward and laboratory conditions. Journal of British Grassland Society 31:29-35. Hauck, R.D., and J.M. Bremner. 1976. Use of tracers for soil and fertiizer nitrogen research. Adv. Agronomy 28:219-266. 105 Hawton, D., I.D.G. Johnson, D S. Loch, G L. Harvey, J.M.T. Marley, W.H.L. Hazard, J. Bibo, and S.R. Walker. 1990. A guide to the susceptibility of some tropical crop and pasture weeds and the tolerance of some crop legumes to several herbicides. Trop. Pest Manage. 36:147-150. Held, D.W., and D.A. Potter. 2012. Prospects for managing turfgrass pests with reduced chemical inputs. Annual Review of Entomology 57:329-354. Herendeen, P.S. 1992. The fossil history of Leguminosae from the Eocene of southeastern North America. Pages 85-160 in Advances in Legume Systematics, part 4, the fossil record (P.S. Herendeen and D.L. Dilcher, eds.). Royal Botanic Gardens, Kew, UK. Herendeen, P.S., W.L. Crepet, and D.L. Dilcher. 1992. The fossil history of the Leguminosae: phylogenetic and biogeographic implications. Pages 303 ? 316 in Advances in Legume Systematics, part 4, the fossil record (P. S. Herendeen and D .L. Dilcher, eds). Royal Botanic Gardens, Kew, UK. Herendeen, P.S., and S. Wing. 2001. Papilionoid legume fruits and leaves from the Paleocene of northwestern Wyoming. Botany 2001 Abstracts, published by Botanical Society of America accessed online September 8, 2013 at http://www.botany2001.org/. Hoveland, C.S. 1989. Legume persistence under grazing in stressful environments in the USA. In: G.C. Marten et al., editors, Persistence of forage legumes. ASA, CSSA, and SSSA, Madison, WI. p. 375-385. 106 International Survey of Herbicide Resistant Weeds. 2013. ALS Inhibitors (B/2) Resistant Weeds. Accessed online September 6, 2013 at http://www.weedscience.org. INTX Microbials, LLC, Kentland, IN. N-DURE, A Premium Inoculant for Alfalfa/True Clover Combination Specimen Lable. Accessed online September 8, 2013 at http://www.arbico-organics.com/downloads/n-dure-alfalfa-clover-inoculant- label.pdf. Isaac, L., C.W. Wood, and D.A. Shannon. 2000. Decomposition and nitrogen release of prunings from hedgerow species assessed for alley cropping in Haiti. Agronomy Journal 92:501-511. Jensen, E.S. 1996. Grain-yield, symbiotic N 2 fixation and interspecific competition for inorganic N in pea-barley intercrops. Plant and Soil 182:25-38. Johansen, C.A. 1977. Pesticides and pollinators. Annual Review of Entomology 22:77- 92. Johansen, C.A., D.F. Mayer, J.D. Eves, and C.W. Kious. 1983. Pesticides and bees. Environmental Entomology. 12:1513-1518. Kajita, T., H. Ohashi, Y. Tateishi, C.D. Bailey, and J.J. Doyle. 2001. rbcL and legume phylogeny, with particular reference to Phaseoleae, Millettieae, and allies. Systematic Botany 26: 515-536. Katzer, G. No Date. Gernot Katzer?s Spice Pages. Accessed online September 8, 2013 at http://gernot-katzers-spice-pages.com/engl/Trig_foe.html?style=flow&redirect=1 107 Kearns, C.A., D.W. Inouye, and N.M. Waser. 1998. Endangered mutualisms: the conservation of plant pollinator interactions. Annual Review of Ecology and Systematics 29:83-112. Kent, L.M., W.L. Barrentine, and G.D. Wills. 1988. Response of twenty determinate soybean (Glycine max) cultivars to imazaquin. Proc. South. Weed Science Society 41:50. Kevan, P.G. 1975. Pollination and environmental conservation. Environmental Conservation 2:293-298. Lampkin, N. 2002. Organic Farming. Old Pond Publishing, Ipswich, United Kingdom. Lavin, M., P.S. Herendeen, and M. F. Wojciechowski. 2005. Evolutionary rates analysis of Leguminosae implicates a rapid diversification of lineages during the Tertiary. Systematic Biology 54: 530-549. Lawn Reform Coalition. 2013. Better lawn or near-lawn types. Accessed online September 9, 2013 at http://www.lawnreform.org/better-species-of-lawn/ Ledgard, S.F., and K.W. Steele 1992. Biological nitrogen fixation in mixed legume/grass pastures. Plant Soil 141:137-153. Lie, T.A. 1971. Symbiotic nitrogen fixation under stress conditions. Plant and Soil: Special Volume:117-127. Littell, R.C., G.A. Milliken, W.W. Stroup, R.D. Wolfinger, and O. Schabenberger. 2006. SAS ? for Mixed Models. 2 nd ed. Cary, NC: SAS Institute Inc. Pp.15-56. 108 Lopez, R., and D.A. Potter. 2000. Ant predation on eggs and larvae of the black cutworm and Japnese beetle in turfgrass. Environmental Entomology 29:116-125. Special volume on Biological Nitrogen Fixation in Natural and Agricultural Habitats. Ledgard, S.F., and K.W. Steele. 1992. Biological nitrogen fixation in mixed legume/grass pastures. Plant and Soil 141:137-153. Macduff, J.H., S.C. Jarvis, and I.A. Davidson. 1996. Inhibition of N 2 fixation by white clover (Trifolium repens L.) at low concentrations of NO 3 in flowing solution culture. Plant and Soil 180:287-295. MacRae, A.W., W.E. Mitchem, D. W. Monks, and M. L. Parker. 2005. White clover (Trifolium repens) control and flower head suppression in apple orchards. Weed Tech. 19:219-223. Marten, G.C., and R.N. Andersen. 1975. Forage nutritive value and palatability of 12 common annual weeds. Crop Science 15:821-827. Martinez, J., and G. Guiraud. 1990. A lysimeter study of the effects of a ryegrass catch crop, during a winter wheat/maize rotation, on nitrate leaching and on the following crop. J. Soil Science 41:5-16. McCurdy, J.D., J.S. McElroy, E.A. Guertal, and C.W. Wood. 2013. Dynamics of white clover decomposition in a southeastern bermudagrass lawn. Agronomy Journal 105:1277-1282. McLaren, J.B., R. J. Carlisle, H. A. Fribourg, and J. M. Bryan. 1983. Bermudagrass, tall fescue, and orchardgrass pasture combinations with clover or N fertilization for grazing steers. Agronomy Journal 75:587-592. 109 McNeill, A.M., and M. Wood. 1990. 15 N estimates of nitrogen fixation by white clover (Trifolium repens L.) growing in a mixture with ryegrass (Lolium perenne L.). Plant and Soil 128:265-273. Milesi, C., S.W. Running, C.D. Elvidge, J.B. Dietz, B.T. Tuttle, and R.R. Nemani. 2005. Mapping and modeling the biogeochemical cycling of turgrass in the United States. Environmental Management 36:426?438. Mitchell, R.L., J.B. McLaren, and H.A. Fribourg. 1986. Forage growth, consumption, and performance of steers grazing bermudagrass and fescue mixtures. Agronomy Journal 78:675-680. Mulholland, J.G., B.S. Dear, G.B. Scott, and A.G. Kaiser. 1989. Effect of broad-leaf herbicides on production of annual legumes. Proceedings of the 5 th Australian Agronomy Conference p. 452. Mulvaney, M.J., C.W. Wood, K.S. Balkcom, D.A. Shannon, and J.M. Kemble. 2010. Carbon and nitrogen mineralization and persistence of organic residues under conservation and conventional tillage. Agronomy Journal 102:1424-1433. Neal, J.C. 1990. Non-phenoxy herbicides for perennial broadleaf weed control in cool- season turf. Weed Technology 4:555-559. Neal, J.C., and M.R. Mascianica. 1988. Phenoxy and non-phenoxy herbicides for controlling dandelion, plantain, clover, and veronica. Proceedings of the Northeast Weed Science Society 42:183?184. 110 Nelson, J.A., P.W. Mayer, W.B. DeOreo, E.M. Opitz, J.C. Kiefer, W.Y. Davis, and B. Dzieqielewski. 1999. Residential end uses of water. American Water Works Association. p. 310. Newell, A. 1999. International Turfgrass Bulletin. October. 206:24-25. Olson, R.V., and C.W. Swallow. 1984. Fate of labeled nitrogen fertilizer applied to winter wheat for five years. Soil Science Society of America Journal 48:583-586. Palmer, J.D., R.K. Jansen, H.J. Michaels, M.W. Chase, and J.R. Manhart. 1988. Chloroplast DNA variation and plant phylogeny. Annals of the Missouri Botanical Garden 75:1180-1206. Palmer, J.D., B. Osorio, J. Aldrich, and W.F. Thompson. 1987. Chloroplast DNA evolution among legumes: loss of a large inverted repeat occurred prior to other sequence rearrangements. Current Genetics 11:275-286. Plants Database. 2012. National Plant Data Center, NRCS, USDA. http://plants.usda.gov. Accessed September 13, 2013. Peck, D.C. 2009. Comparative impacts of white grub (Coleoptera: Scarabaeidae) control products on the abundance of non-target soil-active arthropods in turfgrass. Pedobiologia 52:287-99. Pederson, G.A. 1995. White clover and other perennial clovers. In Forages; R. F. Barnes, D.A. Miller, and C. J. Nelson, (eds), Iowa State University Press: Ames, Iowa, 227-236. Plants Database. 2012. National Plant Data Center, NRCS, USDA. http://plants.usda.gov 111 Polhill, R.M., and P.H. Raven. 1981. Advances in legume systematics, parts 1 and 2. Royal Botanic Gardens, Kew, UK. Polhill, R.M. 1994. Classification of the Leguminosae. Pages xxxv?xlviii in Phytochemical Dictionary of the Leguminosae (F. A. Bisby, J. Buckingham, and J.B. Harborne, eds.). Chapman and Hall, New York, NY. Polhill, R.M., and P.H. Raven.1981. Advances in legume systematics, parts 1 and 2. Royal Botanic Gardens, Kew, UK. Potter, D.A. 1993. Pesticide and fertilizer effects on beneficial invertebrates and consequences for thatch degradation and pest oubreaks in turfgrass. In Pesticides in Urban Environments: Fate and Significane, K.D. Racke and A.R. Leslie (eds.) pp. 331-343. ACS Sym. Ser. 522. Washington, DC: American Chemical Society. Potter, D.A. 1998. Destructive turfgrass insects: biology, diagnosis, and control. Ann Arbor Press, Chelsea, MI. Pouyat, R.V., I.D. Yesilonis, N.E. Golubiewski. 2009. A comparison of soil organic carbon stocks between residential turf grass and native soil. Urban Ecosystems 12:45-62. Qian, Y.L., and R.F. Follett. 2002. Assessing soil carbon sequestration in turfgrass systems using long-term soil testing data. Agronomy Journal 94:930?935. Racke, K.D. 2000. Pesticides for turfgrass pest management: Uses and environmental issues, p. 45-63. In J.M. Clark and M.P. Kenna (eds), Fate and management of turfgrass chemicals. ACS symposium series 743:American Chemical Society, Washington, DC. 112 Racke, K.D., and A.R. Leslie (eds.). 1993. Pesticides in urban environments. Fate and significance. ACS Symposium Series 522: American Chemical Society, Washington, DC. Rajeev, K.V., T.J. Close, N.K. Singh, D.A. Hosington, and D.R. Cook. Orphan legume crops enter the genomics era. Current Opinion in Plant Biology 12:202-210. Rao, S.C., B.K. Northup, W. A. Phillips, and H. S. Mayeux. 2007. Interseeding novel cool-season annual legumes to improve bermudagrass paddocks. Crop Science 47:168-173. Rao, A.C.S., J.L. Smith, J.F. Parr, and R.I. Papendick. 1992. Considerations in estimating Nitrogen Recovery Efficiency by the difference and isotopic dilution methods. Fertilizer Research 33: 209-217. Robbins, P., and T. Birkenholtz. 2003. Turfgrass revolution: measuring the expansion of the American lawn. Land Use Policy 20:181-194. Robbins, P., A. Polderman, and T. Birkenholtz. 2001. Lawns and toxins: an ecology of the city. Cities 18:369-380. Roberts, R.D., and A.D. Bradshaw. 1985. The development of a hydraulic seeding technique for unstable sand slopes II. Field Evaluation. Journal of Applied Ecology 22:979-994. Roberts, E.C., and B.C. Roberts. 1987. Lawn and Sports Turf Benefits. Pleasant Hill, TN: The Lawn Institute. p. 31. Robertson, K.R. 1974. The genera of the Rosaceae in the southeastern United States. Journal of the Arnold Arboretum. 55: 303-401, 611-662. 113 Rogers, M.E., and D.A. Potter. 2004. Potential for sugar sprays and flowering plants to increase parasitism of white grubs by tiphiid wasps (Hymenoptera: Tiphiidae). Environmental Entomology 33:520-527. Rolfe, B.G., and P.M. Gresshoff. 1988. Genetic analysis of legume nodule initiation. Annual Review. Plant Physiology and Plant Molecular Biology. 39:297-319. Roth, L.E., and G. Stacey. 1989. Bacterium release into host cells of nitrogen ?fixing soybean nodules: the symbiosome membrane comes from three sources. European Journal of Cell Biology. 49:13-23. Salminen, S.O., P.S. Grewal, and M.F. Quigley. 2003. Does mowing height influence alkaloid production in endophytic tall fescue and perennial ryegrass? Journal of Chemical Ecology. 29:1319-1328. SAS Institute. 2004. SAS system for Windows release 9.1.3. SAS Institute, Cary, NC. Schrire, B.D. 2005. Millettieae. Pages 367-387 in Legumes of the world (Lewis et al., eds.). Royal Botanic Gardens, Kew, UK. Schultze, M., and A. Kondorsori. 1998. Regulation of symbiotic root nodule development. Annual Reviews 32:33-57. Seefeldt, S.S., J.M. C. Stephens, M.L. Verkaaik, and A. Rahman. 2005. Quantifying the impact of a weed in a perennial ryegrass-white clover pasture. Weed Science 53:113-120. Senseman, S.A. (ed.) 2007. Herbicide Handbook. Weed Science Society of America. Lawrence, KS. 114 Sincik, M. and E. Acikgoz. 2007. Effects of white clover inclusion on turf characteristics, nitrogen fixation, and nitrogen transfer from white clover to grass species in turf mixtures. Communications in Soil Science and Plant Analysis 38:1861-1877. Sleugh, B., K.J. Moore, J.R. George, and E.C. Brummer. 2000. Binary legume-grass mixtures improve forage yield, quality and seasonal distribution. Agronomy Journal 92:24-29. Smith, A.E., and J.D. Powell. 1979. Herbicides for weed control during establishment of arrowleaf clover. University of Georgia, Research Report p.324. Soltis, D.E., P.E. Soltis, M.E. Chase, M.E. Mort, D.C. Albach, M. Zanis, V. Savolainen, W.H. Hahn, S.B. Hoot, M.F. Fay, M. Axtell, S.M. Swensen, L.M. Prince, W.J. Kress, K.C. Nixon, and J.S. Farris. 2000. Angiosperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Linnaean Society. 133:381-461. Sprent, J.I. 2001. Nodulation in legumes. Royal Botanic Gardens, Kew, UK. Sprent, J.I. 2006. Evolving ideas of legume evolution and diversity: a taxonomic perspective on the occurrence of nodulation. New Phytologist. 174:11-25 Springer, T.L. 1997. Effect of bermudagrass height on clover establishment. Crop Science 37: 1663-1665. Stevens, P.F. 2013. Angiosperm Phylogeny Website. Version 9, June 2008. Accessed online September 8, 2013 at http://www.mobot.org/MOBOT/research/APweb/. Swift, M.J., O.W. Heal, and J.M. Anderson. 1979. Decomposition in Terrestrial Ecosystems, Volume 5. University of California Press p.266. 115 Systat Software, 2009. Sigma Plot 11.2. Systat Software, Inc. San Jose, CA. Takahashi, H., T. Takashita, B. Morein, S. Putney, R.C. Germain, and J.A. Berzofsky. 1990. Induction of CD*+ cytotoxic T cells by immunization with purified HIV-1 envelope protein in ISCOMs. Nature 344:873. Takhtajan, A. 1997. Diversity and Classification of Flowering Plants. Columbia University Press, New York. Tranel, P. J., and T. R. Wright. 2002. Resistance of weeds to ALS-inhibiting herbicides: what have we learned? Weed Science 50:700-712. Tukey, P. 2009. Six pieces of evidence that clover is not a weed. Accessed online September 8, 2013 at http://www.thedailygreen.com/environmental- news/latest/clover-lawn-weeds-47071703. Turchin, F.V. 1964. Transformation of fertilizer nitrogen in soil and its uptake by plants. Agrokhimiya 3:3-19. Turgeon, A.J. 2002. Turfgrass management, 6 th ed. Prentice-Hall, Upper Saddle River, New Jersey. Turkington, R., and J.J. Burden. 1983. The biology of Canadian weeds. 57. Trifolium repens L . Canadian Journal of Plant Science 63:243-266. USDA. (No Date). Germplasm Resources Information Network. Accessed online September 6, 2013 at http://www.ars-grin.gov. U.S. Fish and Wildlife. 2007. Running buffalo clover (Trifolium stoloniferum) recovery plan. Accessed online September 9, 2013 at http://www.fws.gov/midwest/endangered/plants/pdf/RBC_RPlanFnl.pdf. 116 United Nations World Urbanization Prospects. 2009. Urban Population. Accessed September 8, 2013. http://esa.un.org/unpd/wup/CD-ROM_2009/WUP2009-F03- Urban_Population.xls. Varvel, G.E., and T.A. Peterson. 1990. Nitrogen fertilizer recovery by corn in monoculture and rotational systems. Agronomy Journal 82:935-938. Vengris, J., M. Drake, W.G. Colby, and J. Bart 1953. Chemical composition of weeds and accompanying crop plants. Agronomy Journal 45:213-218. Wain, R.L., and F. Wightman. 1954. The growth-regulating activity of certain !- substituted alkyl carboxylic acids in relation to their b-oxidation within the plant. Proceedings of the National Academy of Sciences (USA) 142: 525?536. Wang, H., M.J. Moore, P.S. Soltis, C.D. Bell, S. F. Brockington, R. Alexandre, C.C. Davis, M. Latvis, S.R. Manchester, and D.E. Soltis. 2009. Rosid radiation and the rapid rise of angiosperm-dominated forests. Proceedings of the National Academy of Sciences (USA) 106:3853?3858. Wardle, D.A., and L.G. Greenfield. 1991. Release of mineral nitrogen from plant root nodules. Soil Biology and Biochemistry 23:827-832. Watschke, T.L., P.H. Dernoeden, and D.J. Shetlar. 1995. Managing turfgrass pests. Lewis, Boca Raton, FL. Westerman, R.L., and L.T. Kurtz. 1973. Priming effect of 15 N labeled fertilizers on soil nitrogen in field experiments. Soil Science Society America Proceedings 37:725- 727. Whitehead, D.C. 1995. Grassland Nitrogen. CAB International, Wallingford. p. 397 117 Wieder, R.K., and G.E. Lang. 1982. A critique of the analytical methods used in examining decomposition data obtained from litter bags. Ecology 63:1636-1642. Willis, J.B., S.D. Askew, and J.S. McElroy. 2007. Improved white clover control with mesotrione by tank-mixing bromoxynil, carfentrazone, and simazine. Weed Technology 21:739-743. Wing, S.L., F. Herrera, and C. Jaramillo. 2004. A Paleocene flora from the Cerraj?n Formation, Guaj?ra Peninsula, northeastern Colombia. Pp. 146-147 in VII International Organization of Paleobotany Conference Abstracts (21-26 March). Museo Egidio Feruglio, Trelew, Argentina. Wojciechowski, M. (No Date). Tree of life website project. Accessed online September 8, 2013 at http://tolweb.org/tree/phylogeny.html. Wojciechowski, M.F., M. Lavin, and M.J. Sanderson. 2004. A phylogeny of legumes (Leguminosae) based on analysis of the plastid matK gene resolves many well- supported subclades within the family. American Journal Botany 91:1846-1862. Wojciechowski, M.F., M.J. Sanderson, K.P. Steele, and A. Liston. 2000. Molecular phylogeny of the ?temperate herbaceous tribes? of papilionoid legumes: a supertree approach. pp. 277-298 in Advances in Legume Systematics, part 9 (P.S. Herendeen and A. Bruneau, eds.). Royal Botanic Gardens, Kew, UK. Woledge, J., K. Davidson, and W.D. Dennis. 1992. Growth and photosynthesis of short cultivars of white clover with tall and short grasses. Grass and Forage Science 47:230-238. 118 Wood, C.W., and J.H. Edwards. 1992. Agroecosystem management effects on soil carbon and nitrogen. Agriculture Ecosystem and Environment 39:123-138. Worberg, A., M.H. Alford, D. Quandt, and T. Borsch. 2009. Huerteales sister to Brassicales plus Malvales, and newly circumscribed to include Dipentodon, Gerrardina, Huertea, Perrottetia, and Tapiscia. Taxonomy 58:468-478. WSSA. 2012. Weed Science Society America http://wssa.net/weed/. Accessed September 9, 2013. Wu, L., and M.B. McGechan. 1999. Simulation of nitrogen uptake, fixation and leaching in a grass/white clover mixture. Grass and Forage Science 54:30-41. Young, R.R., K.J. Morthorpe, P.H. Croft, and H. Nico. 1992. Differential tolerance of annual medics, Nungarin subterranean clover and hedge mustard to broadleaf herbicides. Australian Journal of Experimental Agriculture 32:49-57 119 Appendices 120 Appendix 1: Effects of Nitrogen and Rhizobium Inoculation on White Clover Characteristics and Control INTRODUCTION Understanding the effects of inoculation and nitrogen fertilization upon clover morphology may lead to a more integrated approach of managing clover within maintained turfgrass swards. Most research indicates that high concentrations of soil N inhibit nodule growth and development within leguminous plants. Macduff et al. (1996) observed that the ratio of root to nodule dry-weights was 6:1 in white clover (Trifolium repens L.) without NO 3 treatment, but increased with applications of NO 3 . It is well documented that increasing N fertilization decreases clover density and allows the grass portion of the sward to out compete clover (Frame and Boyd 1987; Pederson 1995; Sincik and Acikgoz 2007). However, little is known about how this interference manifests itself. Still in question are effects of fertilization on clover not in competition with grass. Additionally, little is known about effects of seed inoculation in native soils and whether an interaction is present between inoculation and nitrogen application. MATERIALS AND METHODS A study was conducted at Auburn University, Auburn, AL (32?35?N, 85?29?W, elevation 198 m), in an environmentally controlled greenhouse. Experiments were conducted as a completely random design with a two-by-six factorial treatment arrangement. Factorial levels were seed treatment (inoculated or not) by N-rate (0, 1.8, 3.6, 7.8, 14.4, 28.8 g N m -2 ) applied as CaNO 3 . Experiments were initiated 5 April, 7 121 June, and 9 August of 2010. Temperatures were monitored and maintained between 25 and 32?C. Fungicide-treated seeds of Dutch white clover (T. repens; Main Street Seed and Supply Co., Bay City, MI) were inoculated with N-Dure (INTX Microbials, LLC, Kentland, IN) which contains the clover specific inoculant Rhizobium leguminosarum biovar trifolii. Inoculant was applied dry directly to seeds according to specimen label. Inoculated seeds (approximately 25) were sown into 90 cm 2 plastic pots containing a Wickham sandy loam soil (fine-loamy, siliceous, subactive, thermic Typic Hapludult). It was the goal of the researchers to choose a soil from the Auburn area that represented a new- or newly renovated- turf site that had previously been maintained as a monoculture. The soil for this study was excavated at 5 to 20 cm depth from a centipede grass (Eremochloa ophiuroides (Munro) Hack.) site that had been fumigated with methyl bromide three years prior and had no recent (within three years) history of legume growth. Soil was mixed thoroughly and was screened through a 4.75 mm sieve to remove grass roots. One month after germination, clover seedlings were thinned to five seeds per pot and were fertilized with CaNO 3 at six different N rates: 0, 0.6, 1.2, 2.4, 4.8, 9.6 g N m -2 . All pots were fertilized with a modified 6x, N free, Hoagland?s solution, including minors, to ensure that there would be no nutrient deficiencies. Beginning two weeks after initial fertilization, plants were mown with a rotary mower at 5.1 cm mowing height. Mowing continued on a bi-weekly basis at the same height until two weeks before final harvest. Plants received overhead mist irrigation daily and supplemental irrigation when 122 needed. Greenhouse temperatures were monitored and maintained between 25?C and 35?C. Plants were fertilized monthly for three months. One month after final fertilizer treatment, foliar growth was harvested at soil level, and roots were gently shaken free of soil. Soil of individual pots was dried, sieved at 2 mm particle size, and analyzed for total Carbon (C) and N by LECO (?). Roots were washed free of excess soil and patted dry with paper towels. Only foliar- fresh weights were recorded, as root samples were placed within sealed plastic bags and were frozen for later analysis. Plant foliage was dried in a plant press, and leaf area of pressed and dried foliage was determined using a Licor 3100C leaf area meter (LICOR BioSciences, Lincoln, NE). Foliar-dry weight, total number of trifoliate leaves, and the length of three randomly sub-sampled petioles of individual pots were recorded. Upon thawing, root nodules were removed, counted, and weighed. After removing nodules, root-alone-fresh weight was recorded. Total root weights were the sum of nodule- and root-alone- fresh weight. Data were analyzed using PROC Mixed within SAS. Data were normally distributed. Differences were determined by ?Type 3 Tests of Fixed Effects,? with p-value less than 0.05 indicating a significant effect. RESULTS AND DISCUSSION Neither inoculation by N-rate interaction or inoculation main effect was observed. Only root DW differed due to N-rate, increasing from 250 to nearly 500 mg/pot as rate increased (Figure A1). Foliar DW as well as petiole length, leaf -area, -count, and -size were unaffected by N-rate. Percent C and N of roots, nodules, and foliage were similar to those reported within previous literature. 123 Future research should quantify soil born Rhizobia within common lawn soils and determine if these numbers are adequate for legume nodulation in situ. Our study, however, is reasonable evidence to conclude that clover-seed inoculation may not be necessary upon the planting of new or over-seeded lawns within the immediate geographical region. We feel that the size and length of this sole experiment may limit its conclusiveness; however, it appears that N-rate has a limited effect upon acute clover phenology, including many important leaf characteristics. These and other studies like them will enable more appropriate agronomic decisions for managing clover with maintained lawn scenarios. 124 Figure A1. Root dry weight response to supplemental N applied as CaNO3. Effects of inoculation are displayed, but they are insignificant (P > 0.05). F ol i a r dr y we i ght ( m g pot - 1 ) Supplemental N (g m -2 month -1 ) 100 200 300 400 500 Uninoculated Inoculated 0 0. 6 1. 2 2. 5 4. 9 9. 8