<%BANNER%>

Growth and Nitrogen Fixation of Legumes Native to the Longleaf-Wiregrass Ecosystem


PAGE 1

GROWTH AND N2-FIXATION OF LEGUMES NATIVE TO THE LONGLEAFWIREGRASS ECOSYSTEM By SARAH ELIZABETH CATHEY A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Sarah Elizabeth Cathey

PAGE 3

This thesis is dedicated to my husband, pa rtner and most devoted supporter, Marcus.

PAGE 4

iv ACKNOWLEDGMENTS Financial support for this research was provided through the cooperative graduate research agreement between the Joseph W. J ones Ecological Research Center and the UF College of Natural Resources and Environment. Technical, sample processing and field support were provided by the staff of the J ones Center, especially Scott Taylor, Mary Cobb, and Sarah Becker. A special thankyou goes to Dr. Kay Kirkman of the Jones Center who provided all of the nursery plan ts for my experiments, and to Dr. Ken Quesenberry who provided additional seeds. I would like to thank my committee for th eir advice and revisions regarding my research and the writing of this thesis. Sp ecial thanks go to Tom Sinclair and Lindsay Boring for their assistance in revisions and fo r personal support. I would like to thank my parents for their moral support, as well. I especially need to thank Susan Sorrell, technician at the USDA ARS lab at the Univers ity of Florida, for her tireless efforts in data collection and sample preparatio n, and for sharing her expertise in spectrophotometry analysis. Finally, I would like to thank my husband, Ma rcus, for his tireless dedication to my graduate studies and attendant technical s upport and for his willingness to move all the way to central Florida from Tennessee in order for me to pursue this degree.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................ix Chapter 1 INTRODUCTION........................................................................................................1 2 GROWTH AND PHENOLOGY OF NATIVE LEGUMES IN TWO LIGHT ENVIRONMENTS.......................................................................................................7 Introduction...................................................................................................................7 Methods and Materials.................................................................................................9 Planting..................................................................................................................9 Measurements......................................................................................................13 Data and Statistical Analysis...............................................................................16 Results........................................................................................................................ .16 Survivorship........................................................................................................16 Morphology.........................................................................................................17 Phenology............................................................................................................18 Plant Responses to Light Environment...............................................................20 Discussion...................................................................................................................32 Morphology and Phenology................................................................................32 Growth Patterns...................................................................................................33 3 USE OF CORROBORATIVE ME THODS TO ASSESS THE N2-FIXATION OF NATIVE LEGUMES..................................................................................................37 Introduction.................................................................................................................37 Methods and Materials...............................................................................................39 Planting................................................................................................................39 N2 Fixation Assessment.......................................................................................41 Statistical Analysis..............................................................................................43 Results........................................................................................................................ .44

PAGE 6

vi Survivorship........................................................................................................44 Fixation Assessment............................................................................................45 Discussion...................................................................................................................54 Comparison of Methodology...............................................................................54 Species Differences.............................................................................................59 Summary..............................................................................................................62 4 GROWTH AND N2-FIXATION OF NATIVE LEGUM ES IN LONGLEAF PINE RESTORATION.........................................................................................................64 Introduction.................................................................................................................64 Materials and Methods...............................................................................................65 Site Description...................................................................................................65 Experimental Design and Planting......................................................................66 Statistical Analysis..............................................................................................70 Results........................................................................................................................ .70 Preliminary Results and Survivorship.................................................................70 Growth.................................................................................................................71 N2-Fixation..........................................................................................................78 Discussion...................................................................................................................78 Shading Effects on Species.................................................................................78 Ecological and Management Implications..........................................................83 5 CONCLUSION...........................................................................................................85 Conclusions from the Current Study..........................................................................85 Introduction.........................................................................................................85 Objectives............................................................................................................85 Directions for Future Research...................................................................................88 Further Application of N2-Fixation Assessment Techniques..............................88 Future Research for Native Legume Utilization.................................................89 LIST OF REFERENCES...................................................................................................91 BIOGRAPHICAL SKETCH.............................................................................................96

PAGE 7

vii LIST OF TABLES Table page 2-1. Description of Native Legumes used in study. ......................................................10 2-2. Fruit and nodule descriptions. ...............................................................................11 2-3. Regression equations for stem elongation curves of the form y= Dx3 + Cx2 + Bx + A. .........................................................................................................................22 2-4. Values given are slopes calculated from the derivative of the equations given in 2-3 at the mean for each coefficient.........................................................................23 2-5. Maximum plant heights by sp ecies, regardless of treatment. ................................25 3-1. Nodule mass and number of nodules. ....................................................................47 3-2. N-transport/storage products extracted from stem sections....................................50 3-3. Specific nodule activities of species in this study and other co mparative reports. ............................................................................................................................... ...55 4-1. Analysis of variance results for experimental variables. .......................................73 4-2. Total biomass (aboveand belowg round tissues, including nodules) per plant and aboveground values for %N, 15N, and total N.................................................74

PAGE 8

viii LIST OF FIGURES Figure page 2-1. Weather for Gainesville, FL 8 February to 20 November 2004. ..........................14 2-2. Phenological change by species. ...........................................................................19 2-3. Stem elongation. ....................................................................................................21 2-4. Leaflet counts and plant widths of sun and shade grown plants. A) Leaflet counts for Clitoria mariana Tephrosia virginiana and Lespedeza hirta plants grown in sun and shade. B) Width of Crotalaria rotundifolia plants grown in sun and shade...........................................................................................................26 2-6. Aboveground and belowground harvested biomass. .............................................30 2-7. Root to shoot ratio of harvested biomass. ..............................................................31 3-1. Ethylene production trends for the growing season by species. ............................48 3-2. Maximum ethylene production (C2H4 reduction) peaks. .......................................49 3-3. 15N values by species. ..........................................................................................51 3-4. Mean % Ndfa, %N, and total N by species. A) Percent of total N derived from the atmosphere. B) Percent N in aboveg round tissues. C) Total N content of aboveground tissues.................................................................................................52 4-1. Volumetric soil moisture patterns for all plots. .....................................................68 4-2. Aboveground biomass and ch ange in plant heights from T0 by species in each of the three light treatments. .......................................................................................72 4-3. Root-to-shoot ratios by species in the three light treatments. ................................76 4-4. Percent N concentration in aboveground biomass (stem + leaves) by species in the intermediate and open light environments. ......................................................77 4-5. 15N and %Ndfa values by species for aboveground tissues in the intermediate and open light treatments. ......................................................................................79

PAGE 9

ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science GROWTH AND N2-FIXATION OF LEGUMES NATIVE TO THE LONGLEAFWIREGRASS ECOSYSTEM By Sarah Elizabeth Cathey December 2005 Chair: Thomas R. Sinclair Major Department: Agronomy The longleaf pine( Pinus palustris Mill.) wiregrass ( Aristida stricta Michx.) savanna ecosystem once dominated the southern coastal plain of the United States, but presently less than 1.5 million of the historic 37.2 million ha remain intact. Restoration plantings reclaiming more than 283,000 ha of former agricult ural fields, pulpwood plantations, and other fire-suppressed lands ar e being established in the Southeast. Groundcover reestablishment of native legumes a nd grasses is the key to restoring soil nitrogen levels, wildlif e habitat and continuous fuels for frequent prescribed burning in young longleaf pine plantings. More informa tion is needed about the growth and N2fixation rates of native legumes under shade in order for informed selections to be made for groundcover restoration plantings. A potted plant study was used to assess growth and N2-fixation responses of 10 species under two light regimes. Total biomass accumulation and root-to-shoot ratios were used to examine growth responses to shading, and several corroborative techniques

PAGE 10

x including analysis of nodule biomass, 15N natural abundance, percent N derived from the atmosphere, transport product analysis, and the acetylene reduction assay were used to assess potential N2-fixing capabilities. Overall, N2-fixation rates increased throughout the season. Relative N2-fixation rates, as assessed by the above approaches, indicated that Mimosa quadrivalvis Crotalaria rotundifolia and Centrosema virginianum were quickly developing species and effective N2-fixers, and that Lespedeza hirta and Orbexillum lupinellus showed lower N2-fixation rates in a one-y ear study. Shade did not have a significant effect on N2-fixation in this controlled study. The acetylene reduction assay is best used as a check for nitrogenas e activity and for following seasonal patterns, but is not as useful for a quantitative estimate of N2-fixed. The nitrogen transport product analysis may have limited usefulness for these species due to extremely low nitrate levels, but should also be tested as a field technique. The 15N natural abundance method was a useful technique for estimating N2-fixation inputs over time. A garden plot study, situated in a 14 yea r-old longleaf pine pl antation on an old agricultural site, was used to assess growth and N2-fixation of eight species of native legumes under three levels of canopy openness. Growth and N2-fixation declined rapidly between approximately 60 and 80 percen t canopy openness, indicating that legumes native to the longleaf pine-wiregrass ecosystem have a limited degree of shade tolerance. Root-to-shoot ratios also indi cated that belowground growth was dominant among plants growing under shaded conditions. Groundcover restoration plantings involving native legumes will be most effective when conducte d after an initial thinning after about 15-20 years of tree growth

PAGE 11

1 CHAPTER 1 INTRODUCTION The longleaf pine( Pinus palustris Mill.) wiregrass ( Aristida stricta Michx.) savanna ecosystem dominated th e coastal plain region of the southeastern United States in the pre-European era, covering as much as 37.2 million ha, but less than 1.2 million ha remain intact. Private hunting lands and plan ted stands constitute most of the longleaf pine coverage that still exists in the U.S. (Landers et al., 1995). Over the past 100 years, land managers have used frequent prescrib ed burning to mainta in the longleaf pinewiregrass ecosystem for wildlife habitat (Boring et al., 2004). Although frequent fire disturbance is necessa ry to maintaining the structure of the longleaf-wiregrass ecosystem by suppressing th e oak midstory, some have hypothesized that nutrient losses due to the consumpti on of litter and volatilization may cause a continual decline of overall N in the syst em (Carter and Foster, 2004). Nitrogen and phosphorus are potentially co-limiting to ne t primary productivity in these woodlands, secondary only to water limitations (Hendric ks et al., 2002). Recent studies have shown that periodic mineraliza tion of phosphorus, which occu rs when large amounts of accumulated needle litter is burned, may serve an important role in the nutrition, growth and reproduction of phosphorus-demanding legume s, leading to the eventual replacement of nitrogen lost during burning (Boring et al., 2004). Native herbaceous legumes constitute more than 10 percent of the vascular plants in frequently-burned longleaf pine savannas (Hainds et al., 1999). Due to their high numbers and density as well as their quick regeneration follo wing fire, the potential for

PAGE 12

2 significant input of biologically-f ixed nitrogen is great. Life history adaptations such as high belowground biomass allocation, opportunistic flowering and N2-fixation (in legumes) are important to the success of speci es native to frequen tly burned ecosystems (Knapp et al., 1998; Morgan, 1999; Hiers et al., 2000; Jacobs and Schloeder, 2002). Legumes have been considered to be a si gnificant component of the N-cycles of many ecosystems, but the actual quan tification of N from biological N2-fixation is difficult to estimate. Using a series of field-conducted population surveys and acetylene reduction assays (and corroborative 15N data), Hendricks and Boring (1999) conservatively estimated legume nitrogen input s from biological fixati on in the longleafwiregrass system to range from 7 to 9 kg N ha-1 yr-1. Other studies have used this technique to estimate poten tial contributions of N2 fixed by legumes in other fire affected ecosystems including prairies (Becker and Croc kett, 1976). However, N-input estimates made from the acetylene reduction assay are limited by the ability to recover nodules in the field. Recent studies seeking to quantify biological N2-fixation in natural ecosystems and agroforestry systems often rely more heavily on isotopic methods, such as 15N enrichment or the natural abundance of 15N (Peoples et al., 1996; Hendricks and Boring, 1999; Medina and Izaguirre, 2004). A comparis on of studies using th ese diverse tools, and an understanding of the limitations inherent in each of the assessment techniques reinforce the need to use corrobor ative methods when assessing N2-fixation in legumes. Because of the high energy requirements necessary to maintain nitrogen fixation, legumes outside of the tropics have generally been assumed to be shade intolerant (Sprent, 1999). Sprent (1973) found that Lupinus arboreus plants grown under shade conditions produced less nodule biomass and subsequently lower nitrogenase activity

PAGE 13

3 (acetylene reduction). Alt hough the longleaf-wiregrass ecosystem has a relatively open canopy compared to other forest systems, the numerous legumes in this ecosystem must be able to tolerate a light environment that is variable but averages approximately 40 percent of full sunlight ( Battaglia et al., 2003; Pecot et al., 2005). Light levels in young, planted longleaf pine stands may be substa ntially higher than th e native woodland, but canopy closure occurs only a few years afte r planting. Thus some of the more shade tolerant native legume species may be mo re satisfactorily adapted for restoration plantings. Reforestation initiatives on pub lic and private lands since 1998 have resulted in the planting of approximately 700,000 acres of form er agricultural, pulpwood plantation and fire suppressed land back into longleaf pine stands through the USDA Conservation Reserve Program (CRP) and other independe nt landowner efforts. Due to the long rotation period typical of longleaf pine stands they have the potential to generate income for private landowners over many years th rough timber revenue, hunting leases and government subsidies paid by the CRP (Landers et al., 1995). Groundc over restoration in young longleaf pine stands planted on deplete d, former agricultural soil is important for rebuilding soil organic matter and N (Markewitz et al., 2002), for providing wildlife food and cover (Stoddard, 1931), and for enhanci ng pyrrhic fuel continuity necessary to reintroduce frequent prescribed fires (Mul ligan and Kirkman, 2002) Reintroduction of native legumes for groundcover rest oration rather than exotic or agricultural species should prevent problems of inva siveness with non-native legume species in the past, such as serecia lespedeza ( Lespedeza cuneata ) and L. bicolor (Miller, 2003), and agronomic species lack of adaptability to woodland environments.

PAGE 14

4 The overall objectives of this study were (1) to explore the impact of various degrees of shading on relative growth and N2-fixation rates of legume species native to longleaf pine-wiregrass savanna s, (2) to make initial observations of phenological development and nodule morphology for each species, and (3) to examine the effectiveness of corroborati ve methods for assessing N2-fixation. These objectives have not been previously addressed for most of th ese species. Controlled potted studies and a common garden experiment were used to assess species responses to shading under potted and field grow ing conditions. Chapter 2 examines the affects of shad e on growth of nine species of native legumes and provides descriptions of phe nological development and nodule morphology after one growing season in a pot study. Desc riptive information rega rding the species in this study, Chamaecrista nictitans (L.) Moench, Centrosema virginianum (L.) Benth., Clitoria mariana L., Crotalaria rotundifolia J.F. Gmel., Lespedeza hirta (L.) Hornem., Mimosa quadrivalvis (L.), Orbexillum lupinellus (Michx.) Isley, Rhynchosia reniformis D.C., and Tephrosia virginiana (L.) Pers, is very limited. Phenological development data is helpful for better understandi ng the life history of a species, es pecially in regard to fire adaptation. Nodule morphology can be an important taxonomic tool for the Leguminoseae (Sprent 2002), and relative nodule biom ass is often indica tive of relative N2-fixation capacity. In addition, biomass accumulation and allocation (root-to-shoot ratios) patterns are indicative of shade tolera nce and environmental adaptability. Chapter 3 describes a companion study to Chapter 2 that examines the N2-fixation patterns and capabilities of th e same nine species of native legumes during a growing season, using five corroborative methods of assessment: nodule biomass, the acetylene

PAGE 15

5 reduction assay, nitrogen transpor t and storage product analysis, 15N natural abundance, and total N content. Each of these methods has been independently used to assess relative N2-fixation rates among species in both c ontrolled and field studies, and to determine the affect of shading on N2-fixation rates. However, this study is the first instance in which the nitrogen transport product analysis has been used to assess these species. In this study, the relative effec tiveness of each corroborative assessment was determined for use in a small, controlled study. Chapter 4 further examines the growth and N2-fixation capabilities of eight species of native legumes planted under three canopy openi ng conditions in a 14 year-old longleaf pine plantation. Because of their dominant gr owth form and importance for wildlife cover and food, more semi-woody species were included in this field study than in the potted studies. The species examined were Centrosema virginianum Desmodium ciliare (Muhl. ex Willd.) DC., Lespedeza angustifolia (Pursh.) Ell., Lespedeza hirta Mimosa quadrivalvis Orbexillum lupinellus Pediomelum canescens (Michx.) Rydb., and Tephrosia virginiana As in Chapter 2, biomass accumulation a nd root-to-shoot ratios were used to assess growth responses to shade in each of the species. The 15N natural abundance method was used to compare relative N2-fixation rates among species and across light environments. Together, these studies shoul d provide a greater understand ing of the growth habits, morphology, phenological development, N2-fixation capabilities and shade tolerance characteristics of several species of native le gumes. This information can be applied to further studies of these and similar species and could be used to make preliminary

PAGE 16

6 decisions about which species should be us ed for groundcover restoration plantings in longleaf pine stands.

PAGE 17

7 CHAPTER 2 GROWTH AND PHENOLOGY OF NATIVE LEGUMES IN TWO LIGHT ENVIRONMENTS Introduction N2-fixing legumes are generally considered to be shade intolerant species -due to the high energetic cost of nodule production, ma intenance, and N fixation (Vitousek et al., 2002). Among temperate ecosystems, native legumes are most diverse and abundant in grasslands (Becker and Crockett, 1976) a nd savanna ecosystems (Hainds et al.,1999). The variably-shaded and frequently burn ed environment of the longleaf pine( Pinus palustris Mill.) wiregrass ( Aristida stricta Michx.) ecosystem supports over forty species of native herbaceous legumes. These legumes constitute more than 10 percent of the vascular species in these pine savannas and occur in high densities across a grea t range of site conditions (Hainds et al., 1999). They demonstrate fire tolerance, adaptability to infertile and droughty soils, values for wild life food resources, and may fix varying amounts of N (Hainds et al., 1999; Hendric ks and Boring, 1999; Hiers et al., 2003). Hiers et al. (2000) showed that flowering of native legumes in this system is tied to occurrence of fire, but responses to a specific seasonal burn varied for Tephrosia virginiana (L., Pers.), Centrosema virginianum (L., Benth.) and Rhyncosia reniformis (D.C.). These legumes are found ubiquitously across the long leaf-wiregrass savanna landscape with only the most extreme deep sands or seasona lly-inundated lowlands having a lowered abundance and diversity of species (Hainds et al., 1999). In addition to surviving droughty conditions and frequent fire, legumes are able to overcome the problem of very low-

PAGE 18

8 fertility soils through N2-fixation (Hiers et al., 2003). Ho wever, shade tolerance of these species has not been explored. Scie ntific knowledge of the biology of N2-fixing species is limited to a few ecosystems, and environmental to lerances have not been explored for most of these species (Vitousek et al., 2002). Since 1998, reforestation initiatives on public and private lands in the southeastern U.S. have resulted in the planting of a pproximately 283,000 ha of former agricultural fields, harvested pulpwood plan tations and otherwise fire suppressed land back into longleaf pine stands, with 48,000 ha of margin al coastal plain farm land in Georgia alone under the USDA Conservation Reserve Program (Coffey and Kirkman, 2004). Many of these sites are characterized by coarse sandy soils that are prone to drought and may be highly depleted in C and N from prior agri cultural production (Marke witz et al., 2002). Groundcover restoration was proposed to be vita l in the recovery of soil organic matter and N-availability. N2-fixing legumes could be especi ally valuable. Although longleaf pine overstory has been successf ully established, there is a great need to better determine the compatibility of groundcover species to a range of canopy light conditions so that recommendations may be made to integrate suitable species into habitat restoration projects. This potted-plant study was designed to obser ve the effects of two light conditions on growth responses of ten species of legum es native to the longl eaf-wiregrass ecosystem over the course of a single growing season. The specific ob jectives of the research reported here were: (1) to document the in fluence of shading on growth habits and biomass accumulation; and (2) to make initia l observations of phenological development and nodule morphology for each species. Variou s observations of root, shoot and nodule

PAGE 19

9 growth were all used as indicators of grow th response to light. A companion study also measured N2-fixation responses of these species a nd results are reported in Chapter 3. Methods and Materials Planting Young plants of ten species of native legum es (Table 2-1) were grown outdoors in Gainesville, Florida (82o 20' W, 29o 38' N) between April and November of 2004. Onehalf of the plants from each species was grown in the sun and one-half under a shade cloth that excluded approximately one -half ambient light. Difference in photosynthetically active radiation (PAR) betw een light treatments was determined using a Li-Cor Quantum Sensor, LI-185A. Measur ements were taken on a clear day, 19 March 2004, at approximately 13:00 Eastern Standard Time. Three readings each were taken under the shade cloth and outside adjacent to th e potted plants. PAR in the shaded area (753 82 mol s-1m-2) was 56 percent of full sun (1340 51 mol s-1m-2). The seedlings were initially propagated by Dr. L. Katherine Kirkman at the Joseph W. Jones Ecological Research Center (JWJ ERC) from seeds collected from throughout the native woodland on the 12,500 ha Ichauw ay reserve, located in Baker County, Georgia, USA (31o19'N and 80o20'W). Seeds were scarif ied by physical abrasion and then germinated in a soil mix consisting of 8 parts Fafard 3B soil mix, 2 parts peat (sphagnum), 2 parts sand, and 1 part perlite, co ntained in plug flats. The seeds were sown June/July 2003 and kept in a greenhouse over the winter.

PAGE 20

10Table 2-1. Description of Native Legumes used in study. Nomenclature follows Wunde rlin and Hansen (2003). Descriptions adapt ed from Isley (1990). Species Code Common Name Subfamily Category Growth Habit Mature Size Centrosema virginianum (L.) Benth. CEVI Spurred Butterfly Pea Papilionoideae Vining/Spread ing Perennial Stems 1-1.5m Clitoria mariana (L.) CLMA Butterfly Pea Papilionoideae Vining/Spreading Perennial Stems 30-100cm Chamaecrista nictitans (L.) Moench CANI Sensitive Pea (Partridge Pea) Caesalpiniodeae Erect herb Annual Plant 15-60cm Crotalaria rotundifolia J.F. Gmel. CRRO Rabbitbells Papilionoide ae Vining/Spreading Perennial; prostrate Stems 1-7cm Lespedeza hirta (L.) Hornem. LEHI Hairy Lespedeza Papilionoideae Erect herb Perennial; semi-woody Plant 0.8-1.5m Mimosa quadrivalvis (L.) MIQU Sensitive Briar Mimosoideae Vining/Spreading Perennial; trailing, thorned stems Stems 1-2m Orbexillum lupinellus (Michx.) Isley ORLU Piedmont Leatherroot Papilionoideae Erect herb Perennial; rhizomatous Stems 20-60cm Pediomelum canescens (Michx.) Rydb. -Buckroot Papilionoide ae Erect herb Perennial; diffusely branched to bushy Plant up to 1m Rhynchosia reniformis D.C. RHRE Dollarleaf Papilionoideae Erect herb Perennial; rhizomatous Plant 7-15cm Tephrosia virginiana (L.) Pers. TEVI Goats Rue Papilionoideae Erect herb Perennial; brancing from a central point Plant 30-60cm

PAGE 21

11Table 2-2. Fruit and nodule descriptions. Nodule sizes indicate an average, and ar e represented as follows: --, not nodulate d; +, <1mm; ++, 1-2mm; +++, <2 to 6mm; ++++, >6mm to 10mm. The number of pl ants examined to estimate nodule size is also given. Nodules Species Fruit Shape Size: Sun Size: Shade Centrosema virginianum Legume; Linear, 7-12cm x 3-4 mm, dehiscent Spherical +++ (n=2) ++ (n=5) Clitoria mariana Legume; Oblong, 3-5 cm x 5-7 mm, seeds sticky Spherical +++ (n=2) +++ (n=5) Chamaecrista nictitans Legume; Oblong, flat 2-4 cm x 4-5 cm Spherical -(n=4) ++++ (n=3) Crotalaria rotundifolia Legume; Ellipsoid, inflated, 1.5-2.5 cm x 712 mm Coralloid ++++ (n=8) +++ (n=8) Lespedeza hirta Legume; 5-8 mm long Spherical +++ (n=7) ++++ (n=6) Mimosa quadrivalvis Legume; Oblong to linear, 3-5cm long, prickled Coralloid ++++ (n=4) ++++ (n=2) Orbexillum lupinellus Legume; Obliquely transverse-ridged Spherical +++ (n=3) ++ (n=2) Pediomelum canescens Legume; 8-11mm long Spherical n/a n/a Rhynchosia reniformis Legume; Oblong or elliptically-oblong, 1.21.8 cm x 6-7 mm Spherical -(n=1) ++ (n=2) Tephrosia virginiana Legume; Oblong, flat, 3-5 cm x 4 mm Elongated ++++ (n=4) ++++ (n=7)

PAGE 22

12 The small, 7 month-old seedlings from the JWJERC were transferred into oneo f the following: clear acrylic rooting tube s (90cm long x 3.5cm diameter), cylindrical polyvinyl chloride (PVC) pots, 35cm x 10.5cm diameter, or black plastic tree seedling pots, 35cm tall x 80cm2 (Stuewe and Sons, Corvallis, OR). PVC pots were also used to assess N2-fixation using the acetylene reduction a ssay (Chapter 3). A total of eight rooting tubes, six PVC pots, and 14 black plasti c pots were planted for each species in the experiment. One half of each set of pots wa s grown under the shade treatment. For this experiment, plants in the PVC and black plas tic pots were measured and harvested as a single group since the volume was approxi mately the same. Transplanting was completed on 11 February 2004 (Day 42). Dates are represented in figures and tables as numbered days beginning with 1 January 2004 as Day 1. Seeds for the annual Chamaecrista nictitans (CANI) were obtained from Dr. Ken Quesenberry (Agronomy Department, University of Florida). The seeds were collected from along the roadside in Gainesville, FL. Seeds for (CANI) were scarified with sand paper and then germinated on moist filter paper. Emerged CANI seedlings were planted directly into PVC and black plas tic pots on 21 April (Day 122). Plants were inoculated by introducing na tive soil to each pot. Two topsoils (020cm) were collected at the Jones Center, fr om a fine-loamy, kaolinitic, thermic Typic Kandiudult (Orangeburg Series), and from a loamy, kaolinitic, thermic Arenic Kandiudult (Wagram Series). Collections we re taken from areas with thriving and diverse legume populations. Soils were transp orted to Gainesville, FL where they were stored in a cool, dark room and covered with plastic to maintain moisture. A mixture of equal parts of each native soil type was used as the inoculation source. The pots had been

PAGE 23

13 previously filled with commercia l topsoil (Walmart Corp.) to within 6 cm from the top, then 2 cm of native soil was added to the surface of all pots. Finally, the plant was placed in the pot, and additional topsoil was used to cover the roots as needed. For the rooting tubes, the 2 cm of native so il was added below where the seedling was pressed into the top of the tube. Seedlings in the rooting tubes were inserted so that there was approximately 2-3 cm in the tube abov e the soil to facilitate watering. Water was applied by drip-irrigation ever y 12 hours (5:00 and 17:00 EST) using a battery-operated timer (Rainbird), but was ad justed as needed throughout the experiment to prevent excessive watering durin g periods of heavy rainfall. Aphids were detected on CRR O, beginning around 3 April (Day 94), but the plants were not detrimentally affected by the infestation. Pediomelum canescens began to yellow and to develop brown leaf spots as early as 30 June (Day 182), followed by rapid leaf loss and, consequently, this species was omitted from our results. Gainesville experienced hurricane ac tivity around 14 August (Day 227) and 3 September (Day 247; Figure 2-1). Some plan ts experienced leaf loss due to wind, and LEHI, which was moved indoors, experienced some water stress. However, the reason for loss of leaves at harvest was difficult to distinguish, because senescence had begun by that time for most species. Measurements Height measurements were taken weekly of plants in both pot types and the rooting tubes. Height was determined by measuri ng the distance from the soil surface to the highest point on the plant. The height of vining plants such as MIQU, CEVI, and CLMA was considered to be the length of the longest stem. Each of these vines was measured until the date when the branches became too tangled for a measurement to be plausible.

PAGE 24

14 35 65 95 125 155 185 215 245 275 305 0 20 40 60 80 100 120Calendar DayRainfall (mm) 35 65 95 125 155 185 215 245 275 305 0 10 20 30 Temp. Min Temp. Max Irradiance 0 10 20 30 40 Calendar DayIrradiance (MJ m-2day-1)Temperature (oC) Figure 2-1. Weather for Gaines ville, FL, 8 February to 20 November 2004. Temperature and irradiance values are weekly averages Rainfall values are daily totals. Data from FAWN (2005).

PAGE 25

15 Root elongation was measured along the side of the clear acrylic tubes from the top of the tube to the tip of the longest visible root using a measuring tape. Each tube was placed in a white PVC sleeve. Root depth measurements and notations regarding nodule presence were taken twice weekly until the ro ots reached the bottom of the tube or until there was no increase in rooting de pth for at least three readings. Leaf addition was measured by counting th e number of leaflets on each plant on a weekly basis. This method was continued throughout the experiment for CLMA, LEHI, RHRE, and TEVI. However, due to the large nu mber of leaflets or the indistinguishable nature of the leaflets, CEVI, MIQU, CRR O and ORLU leaf addition was instead determined by measuring the width of the plant. The width of CEVI and MIQU was determined to be the sum of the lengths of the two longest stems. Widths of CRRO and ORLU were determined as the width of the plan t at the widest point, le af tip to leaf tip. Width measurements were taken using a measuring tape or ruler. The presence of flowers and fruits was not ed along with the height measures. A phenological phase was considered to be ini tiated when half of the specimens for each species had expressed the particular characteris tic such as presence of flowers or fruit, or the absence of flowers wh ile fruit remained. At the conclusion of the experiment, all plants were destru ctively harvested. Aboveground material was collected from specime ns grown in root tubes. Both aboveand belowground biomass was collected from all specimens grown in pots. Roots were washed free of soil and nodules were colle cted. Nodules were counted and then individually measured by laying each one alongs ide a millimeter ruler. The diameter of spherical nodules and the longest dimension of elongated nodules was measured in order

PAGE 26

16 to estimate an average nodule size for each pl ant. All tissues were dried to constant weight at 80oC. Data and Statistical Analysis Data were analyzed using analysis of variance (ANOVA) with species and light environment as main effects. If differences existed (p<0.05), Duncans multiple comparison post-test was used to determin e which means differed significantly. The GLM procedure performed in the Statistical Analysis System (SAS, 2003) was used for ANOVA and post-tests. Patterns of stem el ongation were analyzed using a non-linear regression model (third -order polynomial), followed by an analysis of slope using the derivative (dh/dt). The slope of the curve at selected points was calculated and compared using ANOVA to test for species and trea tment effects. Students T test ( =0.05) was used to test for significant differences in pl ant height, number of leaves and plant width response to light treatments within a species. Non-linear regression and t-tests were performed using Prism (Prism, 1996). Results Survivorship Out of the 20 individual plants of each species that were planted in the black plastic and PVC pots, an average of 7 survived. LEHI had the highest survival with 13 remaining, and ORLU and RHRE had the lowest survivorship with only 4 and 3 plants remaining, respectively. The cause of mort ality for most of these young seedlings is unknown. Most of the species were still dormant when transplanted, and many of the individuals never reemerged. Some of the MIQU plants were lost due to desiccation during May during a dry and windy period when conductance and evaporative demand must have exceeded the watering rate for this large species.

PAGE 27

17 Survivorship in the rooting tubes was also low, with only an average of 3 plants surviving out of the 8 transpla nted per species. Lack of r eemergence, inadequate space for watering in the top of the tubes and soil compaction (l eading to inundation of roots) were the reasons for the small success of th ese individuals. Despite low survivorship overall, 6 of the 10 species generally had enough survivorship to allow for treatment analysis, excluding Pediomelum canescens MIQU, ORLU and RHRE. Morphology Although they are quite different in structur e and life history, the ten species in this experiment can be categorized into two ge neral morphological types, a vining/spreading herb or an erect herb (Table 2-1). The vi ning/spreading species, including CEVI, CLMA, CRRO and MIQU, although quite different from each other, all tend to be either prostrate or to climb over adjacent plants. In contra st, the erect herbs are generally upright, but LEHI or ORLU may droop over other plants or onto the ground. CANI, RHRE and TEVI branch from a central stem or root cr own, and the branching stems may be nearly horizontal to the ground. The legume fruits of these species ranged in size from 1.2 to 12 cm long (Table 22). CEVI, CLMA, CRRO, and RHRE pods are dehiscent, but the dense, bract-like, tightly-fitting pods of LEHI ar e persistent (Kirkman et al ., 2004). Dispersion strategies represented besides dehiscing in clude pods that stick to clothing or fur, such as LEHI, individual seeds that are coated with an a dhesive material, such as CLMA, and pods that simply fall to the ground, such as MIQU a nd TEVI. CRRO legumes are inflated, and the seeds will rattle inside of the pod when dry. Inoculation with a mixture of native soils appeared to be eff ective due to nodule development in all species. Nodule morphol ogy differs among species, as well. CEVI,

PAGE 28

18 CLMA, CANI, LEHI, ORLU and RHRE all have similarly shaped, spherical nodules. These nodules vary from 1mm to nearly 10mm in diameter (Table 2-2). CRRO and MIQU have coralloid nodules from >2mm to <10mm in length. Coralloid nodules have a central branching point from which they randomly bifurcate, w ith occasional twicebifurcation. The shape of this type of nodule is highly irregular, and a definitive size is difficult to estimate. TEVI nodules are elongat e, cylindrical and often bifurcated. The length of TEVI nodules average >6mm to 10mm, although largest nodules are >10-mm long. Phenology Shade-grown plants did not demonstrat e significant phenol ogical delay in comparison to plants grown in full sun. Ho wever, weekly observations of phenological change did not allow high resolution for dete rmination of treatment effects on flowering and fruit initiation. Data for the two light treatments within a sp ecies were combined, since there were no significant differences. Da tes given are the days on which one half of the specimens for each species had begun to express the given phenological change, such as flowering. MIQU was the first of the study species to flower (28 May, Day 149; Figure 2-2) and to produce fruit (30 June, Day 182), followed by CRRO, CLMA, CEVI, and LEHI, respectively. Flowering and fruiting con tinued throughout the s eason for MIQU, CRRO and CEVI. The lack of a specific time for flowering and fruiting to occur indicates that these species may be opportunistic in their fr uit production. Both flow ers and fruits were present on MIQU, CRRO and CEVI from the first fruiting until the end of the experiment. In contrast, CLMA, which stopped flowering around 14 September (Day 258), had a much more determinant flowering and fruiting pattern. Flowers and fruits

PAGE 29

19 CEVI CLMA CRRO LEHI MIQU ORLU TEVI 140 160 180 200 220 240 260 280 300Flowering Flowers and Fruits Fruits only SpeciesCalendar Day Figure 2-2. Phenological change by species. Bars represent the time at which over onehalf of specimens for a species had reach ed the prescribed phenological phase, such as the onset of flowering. TEVI did not flower during the experiment.

PAGE 30

20 were only concurrently present for appr oximately 30 days on CLMA before fruit production ceased and the plant retu rned to a vegetative state. All species examined in this study were late spring and summer-flowering and fruiting except for LEHI, which is considered to be fall flowering and did not begin to flower until around 24 September (Day 268) TEVI did not flower during this experiment. ORLU flowered for the shortest duration of any of the species in this study, only 20 days. Although ORLU flow ered, fruits were neither detected nor collected. Plant Responses to Light Environment Overall, the height patterns of all specie s showed rapid increase in stem elongation toward the beginning of the season, with th e exception of CRRO, and LEHI. Growth curves were fitted to a third order polynomi al (Figure 2-3; Table 2-3), but CEVI and MIQU had a poor fit for this equation (r2<0.300) in a particular treatment, making them difficult to analyze. Stem elongation rates for all species were greatest around 14 May (Day 135), followed by decline as the seas on progressed, with the exception of LEHI, which had its strongest elongation rate duri ng the middle portion of the season, between 29 June and 19 August (Day 181 and 232) (Table 2-4). CANI, CRRO and LEHI underwent a slight decline by the final meas urement before harvest, 22 October (Day 296). Stem elongation rates of CRR O, LEHI, ORLU and TEVI diverged according to treatment on 19 August and 22 October (Day 232 and 296 (Table 2-4). Height was also statistically different (unpai red t test, p< 0.05) at dates within the rang e that slope diverged for CLMA, CRRO, and LEHI (Figur e 2-3). For CLMA, stem elongation rates in the shade were approximately twice that of sun-grown plants across all calculated

PAGE 31

21 Chamaecrista nictitans 92 122 152 182 212 242 272 302 0 10 20 30 40 50 60 70 80 90 100Sun Shade Calendar DayHeight (cm)Sun r2 = 0.848 Shade r2 = 0.937 Clitoria mariana 92 122 152 182 212 242 272 302 0 10 20 30 40 50 60 70 80 90 100* * * * *sun r2 = 0.726 shade r2 = 0.374** ** *Calendar DayHeight (cm) Crotalaria rotundifolia 92 122 152 182 212 242 272 302 0 25 50 75 100* * **Sun r2 = 0.555 Shade r2 = 0.694* * *Calendar DayHeight (cm) Mimosa quadrivalvis 92 122 152 182 212 242 272 302 0 25 50 75 100 Sun r2 = 0.221 Shade r2 = 0.618Calendar DayHeight (cm) Orbexilum lupinellus 92 122 152 182 212 242 272 5 30 50 75 100Calendar DayHeight (cm)Sun r2 = 0.832 Shade r2 = 0.645 Tephrosia virginiana 92 122 152 182 212 242 272 302 0 25 50 75 100Calendar DayHeight (cm)Sun r2 = 0.534 Shade r2 = 0.444 Lespedeza hirta 92 122 152 182 212 242 272 302 0 25 50 75 100 125 150* ** *Sun r2 = 0.898 Shade r2 = 0.575Calendar DayHeight (cm) Centrosema virginianum 92 122 152 182 212 242 272 302 0 25 50 75 100 125 150 Sun r2 = 0.625 Shade r2 = 0.262Calendar DayHeight (cm) Figure 2-3. Stem elongation. Values are means +/SE. represents plants grown in the sun, and represents plants grown under the shade treatment. Goodness of fit values (r2; =0.05) are given for each curve th at was fitted with a thirdorder polynomial.

PAGE 32

22 Table 2-3. Regression equations for stem elongation curves of the form y= Dx3 + Cx2 + Bx + A. Coefficients are mean +/SE. Equation Coefficients Species Light A B C D CANI Sun -107 32.34 1.38 0.54 -0.004 -0.002 -3.8x10-6 5.23x10-6 CANI Shade -63.16 28.22 0.74 0.47 -0.001 0.002 -5.18x10-7 4.52x10-6 CLMA Sun -71.01 37.33 1.09 0.59 -0.003 0.002 4.84x10-6 4.87x10-6 CLMA Shade -155.60 129.7 2.13 2.06 -0.007 0.010 8.45x10-6 1.70x10-5 CRRO Sun -48.50 44.59 0.48 0.71 -8.05x10-5 0.003 -2.26x10-6 6.00x10-6 CRRO Shade -50.50 51.62 0.49 0.82 -9.77x10-5 0.004 -1.28x10-6 6.95x10-6 LEHI Sun 186 60.94 3.62 0.97 0.02 0.004 -3.65x10-5 8.14x10-6 LEHI Shade 80.9 120.1 2.06 1.94 0.01 0.01 -2.78x10-5 1.65x10-5 ORLU Sun -69.92 26.73 1.18 0.42 -0.004 0.002 7.10x10-6 3.54x10-6 ORLU Shade -76.76 29.50 1.26 0.46 -0.005 0.002 7.31x10-6 3.90x10-6 TEVI Sun -48.13 58.98 0.65 0.93 -0.001 0.004 1.004x10-6 7.81x10-6 TEVI Shade -120.4 66.14 1.74 1.05 -0.006 0.005 7.88x10-6 8.70x10-6

PAGE 33

23 Table 2-4. Values given are slopes calculated from the derivative of the equations given in Table 2-3 at the mean for each coefficient. dh/dt Species Light Day131 (14 May) Day 181 (29 June) Day 232 (19 August) Day 296 (22 October) CANI Sun 0.462 0.213 0.018 -0.141 CANI Shade 0.384 0.232 0.070 -0.145 CLMA Sun 0.319 -0.160 -0.455 -0.790 CLMA Shade 0.677 0.350 0.148 0.080 CRRO Sun 0.346 0.233 0.081 -0.157 CRRO Shade 0.399 0.329 0.237 0.094 LEHI Sun 0.294 0.799 0.749 -0.120 LEHI Shade 0.497 0.721 0.520 -0.346 ORLU Sun 0.263 0.106 0.569 0.151 ORLU Shade 0.260 0.073 -0.003 0.061 TEVI Sun 0.311 0.208 0.119 0.292 TEVI Shade 0.500 0.239 0.095 0.087

PAGE 34

24 dates. Slopes and heights for CANI were very similar for both sun and shade-grown plants throughout the growing season. Most species reached maximum height near the end of the experiment (9 October to 22 October, Day 283 to 296). However, CANI (15 August, Day 228) and MIQU and ORLU (1 July to 5 September, Day 183 to 249) peaked earlier. CANI and MIQU experienced a slight decline in measured hei ght after peaking due to some defoliation and stem breakage. Differences in maximum pl ant height were due to differences among species, as determined by analysis of vari ance (p < 0.001), and treatment effect was not significant due to the amount of variation between individual plants. Without regard to treatment, LEHI and CEVI were the talle st plants, followed by MIQU, CLMA, TEVI, CRRO, CANI, and ORLU, re spectively (Table 2-5). Leaf addition patterns for CEVI, CRRO (after 15 June, Day 167), and TEVI show that the plants grown in the shade tended to have more leaves on a given date than those grown in the sun. Due to large variability among individual plants, this pattern is only significant on a few dates for CRRO during the season (F igure 2-4A, B). CLMA, CRRO (before 15 June, Day 167), and LEHI patterns reveal the opposite effect, more leaf addition occurring on those plants grown in the sun than thos e in the shade. Again, due to large variability between i ndividual plants, this effect is only significant for CLMA and LEHI on a few dates across the season (Fig ure 2-4A). ORLU, MIQU, and CANI did not show distinct patterns with regard to leaf addition over the season, and no statistical differences according to treatment were found on any dates. Differences in root elonga tion patterns were difficult to detect due to the high mortality rates of specimens grown in the root tubes. In addition, due to the coloration of

PAGE 35

25 Table 2-5. Maximum plant heights by species regardless of treatment. Values are means SE. Different letters indicate signi ficant differences (Duncan's posttest). Species Day Max. Height (cm)n CEVI 289 (15 October) 53.2 +/27.6 a 5 LEHI 296 (22 October) 102.0 +/10.1 a 11 MIQU 225 (12 August) 81.3 +/9.5 ab 6 CLMA 293 (19 October) 53.2 +/9.7 bc 8 TEVI 296 (22 October) 46.6 +/4.6 c 10 CRRO 283 (9 October) 44.5 +/4.1 c 9 CANI 228 (15 August) 33.6 +/1.3 c 9 ORLU 249 (5 October) 31.4 +/3.3 c 4

PAGE 36

26 Clitoria mariana 130 160 190 220 250 280 0 300 600 900 1500Sun Shade Calendar DayNumber of Leaflets Tephrosia virginiana 130 160 190 220 250 280 0 300 600 900 1200 1500Calendar DayNumber of Leaflets Lespedeza hirta 125 155 185 215 245 275 0 300 600 900 1500Calendar DayNumber of Leaflets Figure 2-4. Leaflet counts a nd plant widths of sun and shade grown plants. Values shown are means SE. A) Leaflet counts for Clitoria mariana Tephrosia virginiana and Lespedeza hirta plants grown in sun a nd shade. B) Width of Crotalaria rotundifolia plants grown in sun and shade. A A

PAGE 37

27 Crotalaria rotundifolia 180 210 240 270 300 0 50 100 150Sun Shade Calendar DayWidth of Plant (cm) Figure 2-4. Continued. B

PAGE 38

28 the roots being close to that of the soil and the tendency of roots to not remain along the walls of the tubes, measurement were highl y variable, and no diffe rences according to treatment were significant. The rooting de pth of the specimens in the sun and shade treatment were very similar for the first ha lf of the growing season, although sun-grown plants tended to root slightly deeper in all three species shown (Figure 2-5). Shaded specimens appeared to have rooted slightly d eeper than those in the sun at approximately 10 June (Day 162) for TEVI, and 30 June (Day 182) for LEHI and CRRO. Biomass accumulation was not affected by lig ht treatment for either total biomass, aboveground or belowground accumulation (p = 0.3765), therefore, additional values reported are pooled treatments by species. MIQU accumulated the largest amount of biomass over the season, followed by CANI and LEHI, then CEVI, CRRO, TEVI, and finally CLMA and ORLU (Figure 2-6). Root to shoot ratios (R/S) s howed no significant treatment effect, but species effect was highly significant (p <0.001). MIQU and TE VI showed heavy allocation of biomass belowground (R/S>3, Figure 2-7), followed cl osely by CLMA, which had twice as much allocation belowversus aboveground (R/S>2). CEVI and CRRO had a mostly balanced allocation aboveand belowground (R/S ~ 1). CANI, LEHI and RHRE favored aboveground allocation (R/S < 1). These patt erns of biomass allocation did not reflect the same pattern as that of total biomass, but the different allocation patterns were represented in each of the groupings.

PAGE 39

29 CRRO 92 122 152 182 212 242 272 302 10 35 60 85Calendar DayLength of Longest Visible Root (cm) LEHI 92 122 152 182 212 242 272 302 10 35 60 85Calendar DayLength of Longest Visible Root (cm) MIQU 92 122 152 182 212 242 272 302 10 35 60 85Calendar DayLength of Longest Visible Root (cm) Figure 2-5. Elongation of root s grown in the sun and under shade treatment. Data was composited by species since light treatment effect was not significant. Values given are means SE.

PAGE 40

30 MIQULEHICANICEVITEVICRROCLMAORLU 0 10 20 40 30 20 10 Aboveground Belowground Species b bc c bc b a c bcDry Weight (g) Figure 2-6. Aboveground and belowground harves ted biomass. Values are means +/SE. Data is composited by species b ecause light treatment effect is not significant. Different le tters represent significantly different total biomass (aboveground + belowground) values, and alphabetical order designates order of total biomass values, greates t to least (Duncan's post-test).

PAGE 41

31 MIQULEHICANICEVITEVICRROCLMAORLURHRE 0 1 2 3 4 5 a d d cd ab d bc ab d SpeciesRoot-to-Shoot Ratio Figure 2-7. Root to shoot ratio of harvested bi omass. Values are means +/SE. Data are composited by species because light trea tment effect is not significant. Different letters represen t significantly different r oot-to-shoot ratios, and alphabetical order designates order of ra tios, greatest to l east (Duncan's posttest).

PAGE 42

32 Discussion Morphology and Phenology The diverse native legumes in this study represented all three sub-families of legumes, as well as vines, erect herbs, a nd a semi-woody shrub. The life histories of these species represent a variety of strategi es for overcoming shade, frequent fire and drought conditions. However, of the forty na tive legume species that grow extensively across the Ichauway reserve, only two are annual species: Cassia fasciculata and C. nictitans (Hainds et al.,1999). Perennial growth form is common in many frequentlyburned ecosystems (Knapp et al., 1998; Morg an, 1999; Jacobs and Schloeder, 2002) and is apparently an effective adaptation to the frequently-burned longleaf woodland environment. The nodule morphologies of th e species in this study follow the types described by Sprent (2002) for each of the subfamilies. The majority of the species in this st udy began flowering during the late-spring, early summer. LEHI, like many of the other semi-woody Lespedeza and Desmodium species in the native woodland, flowered during the early fall (Chapter 3). All of the plants that produced fruit continued to flow er and fruit until the time when they were harvested, except for CLMA. The indeterm inate nature of flowering and fruiting represented by most of the speci es in this study is another adaptation to a fire-maintained ecosystem. Hiers et al. (2000) found that ma ny legumes showed little change in duration or timing of peak flowering in response to the season in which they were burned, especially those that were fall-flowering or that matured multiple seed crops each year. Those plants that continuously produced seeds over the course of the season continuously contributed to the seed bank, and thus, did not need to re-grow and mature for a hastened seed crop after a late growing-season fire.

PAGE 43

33 LEHI represented a fall-flowering plant that has determinant seed production in the wild, although it was no t well represented in this study due to the timing of the harvest date. However, the number of seeds produced by LEHI in each c ohort of seeds and the ease and range of its dispersal is much greate r than that of the sp ecies which produce seed throughout the year. Unlike the other species in this study th at rely on dehiscent pods or gravity for localized dispersion of seed, LEHI s eeds adhere to fur (or clothing) of passing animals and are therefore widely dispersed. TEVI, which did not flower during this experiment, is an example of a species that fl owers prolifically in th e field in response to fire (Clark, 1971), although th e flowering may be delaye d by lightning-season fires (Hiers, 2000). Seed production (or lack thereof) in this study may not be indicative of production in the wild, due to unknown pollination factors that may not have been present in urban Gainesville, FL ve rsus in the natural woodlands. Growth Patterns The shading treatment imposed in th is study was 56 percent of full sun. Measurements of light quantity in the na tive woodland, below the tree canopy, but above the wiregrass canopy may be variable but is approximately 40 percent (Pecot et al., 2005), which is substantially higher than that typica lly reported for other forest types, including young pine plantations. Thus, shad e tolerance among speci es in the longleaf woodland understory is not on the same scale of tolerance for understory species in other forest types that may only receive around 3 pe rcent light infiltrati on (Battaglia et al., 2003). Light quality factors such as the red to far-red ratio may also be important to understanding responses of understory species, however, such analysis was outside of the scope of this study.

PAGE 44

34 The species in this study all had variab le growth responses to the two light environments, but shading did not have a si gnificant impact on to tal biomass, or R/S allocation patterns (Figure 2-7). Although the sm all sample sizes due to high mortality in this study reduced our effectiveness to detect biomass shading effects, these effects were manifested in the later measurement dates for stem elongation (Figure 2-1). The effect of shading on height growth of CLMA, CRRO, and LEHI became was significant after 15 June (Day 167). CLMA and CRRO shade-grown plants were significantly taller than the sun-grown plants toward the end of the growing season, and LEHI grew significantly taller in the sun. The fact that tallest plants for individual species we re not all located in the shade indicates that the shade-grown plan ts were not simply etiolated, but rather showed some degree of shade tolerance. CRRO, which is a prostrate spreading plan t, and CLMA, which is semi-erect to vining, both grow below and amongst the bunc hgrasses of the native woodland, thus, their tolerance of shade is not surprising (F igure 2-1). LEHI, a tall semi-woody species, showed a more favorable growth response to sun than shade during the last weeks of the growing season. In the woodland, LEHI quick ly outgrows the surr ounding grasses and avoids mutual shading in the understory mo re aggressively than the smaller-stature species. CANI and TEVI, which showed no si gnificant responses to shading, attain the same height as the surrounding grasses in th e native woodland. The effects of shading on the two large vines, CEVI and MIQU, were not well defined in this study due to high mortality and the difficulties of measuring vines that intertwine and break easily. However, even if these vines have reduced shad e tolerance, they sti ll have the ability to climb over neighboring plants and to grow into sunflecked gaps in the savanna in order to

PAGE 45

35 reach more direct light than smaller pl ants, such as ORLU and RHRE, for which adaptation to some shading would be more advantageous. The variation in reported height among species, from 30cm to 100cm, represents the diversity of sizes and growth forms among these species. However, if the crown width of vining/spreading plants is considered, CEVI is the largest plant, spreading to over 2m, followed by MIQU, which grew to over 1m across, as well. CRRO also spread approximately twice as wide (Figure 4) as it grew tall, ~80cm versus ~45cm, respectively. Keeping in mind the differences in plan t morphology, the most decisive way to address comparisons of size of these species is by using total biomass. In an ecosystem that is frequently-burned and prone to frequent drought, significant allocation to belowground biomass is a positive survival adaptation (Knapp et al., 1998). These perennial species that can re -grow from reserves in the r oots before surrounding plants will have a temporal and li ght-availability advantage ove r neighbors. MIQU and TEVI, both with R/S>3 (Figure 2-7), have also b een shown to be significantly altered in phenology by fire (Hiers et al., 2000). Young MIQU plants excavated in the field had taproots exceeding 2 m long (personal observati on). Plants that ha ve higher R/S ratios and deep, branching root systems are better at enduring droughty c onditions (Kramer and Kozlowski, 1979; Knapp et al., 1998), such as those that develop quickly in the coarsesandy soils present in much of the longleaf-wiregra ss ecosystem (Hainds et al., 1999). For a legume, an extensive root surface ar ea also provides increased opportunity for rhizobial colonization.

PAGE 46

36 The legume species in this study are difficu lt to generalize as a single group due to their diverse growth forms, reproductive phe nologies, and life hist ories. Although the reduced early survivorship of our study population obscured some of our ability to document shade effects for all species, these da ta indicate that exam ples of legumes with some degree of shade tolerance can be found out side of the tropics (Sprent, 1999). Some species exhibited greater height growth with shading, but none of the species in this study demonstrated strong shade intolerance thr ough reduced biomass responses to the 56 percent of ambient light level treatment. This light regime would certainly be representative of conditi ons in open longleaf pine savannas or in thinned young plantations with restoration plantings (C hapter 4), although a more heavily shaded treatment might have provided a greater eff ect upon several species. However, most of the life-history characteristics of the species examined in th is study may be more strongly associated with adaptation to fire, N deficien t soils, and drought than to light environment in longleaf pine savanna ecosystems.

PAGE 47

37 CHAPTER 3 USE OF CORROBORATIVE ME THODS TO ASSESS THE N2-FIXATION OF NATIVE LEGUMES Introduction The fire-dependent longleaf pine( Pinus palustris Mill.) wiregrass ( Aristida stricta Michx.) savanna ecosystem once dominated th e southern coastal plain of the United States, covering as much as 37.2 million ha. Presently, less than 1.5 million ha of these ecosystems remain intact (Landers et al ., 1995). However, restoration plantings reclaiming more than 283,000ha of former ag ricultural fields, pulpwood plantations, and other fire-suppressed lands ar e being established in the S outheast, 48,000ha of which are in USDA Conservation Reserve plantings in Georgia (Coffey and Kirkman, 2004). Groundcover reestablishment is the key to restoring wildlife habitat in these young longleaf pine plantings as well as providing continuity of pyrrhic fuels (Clewell, 1989; Kirkman, 2002). Legumes may also have a ma jor role in maintaining N balance firemaintained restored systems that are being established on depleted, former agricultural soils (Markewitz et al., 2002; Boring et al., 2004). Plan ting of native legumes would permit groundcover restoration without some of the problems that have occurred as a result of introducing non-native and agr onomic legumes, including serecia lespedeza ( Lespedeza cuneata ) and L. bicolor that have little shade tolerance and can be very invasive (Miller, 2003). There is a need for species of legumes native to the longleaf-wiregrass ecosystem to be identified for use in groundcover restora tion plantings that show strong potential for

PAGE 48

38 N2-fixation and that can make large, N-rich biomass contributions to depleted soil organic matter (Markewitz et al., 2002). In addition, adaptation to environmental factors such as nutrient deficiency, drought condi tions, and reduced light under a young closed forest canopy must be considered along with an analysis of N2-fixing potential. After initial screening, these species should be planted in field conditions for a better assessment of their physiol ogical adaptations and N2-fixing potential (Dreyfus et al., 1988). Assessing N2-fixation under field conditions is a difficult task, and most current methods measure fixed-N2 indirectly. Traditional methods of assessment involve nodule excavation and other destructive biomass meas ures to assess N-fixation and cannot be performed repeatedly on the same plants. These methods can also be destructive when used in a woodland ecosystem because of the amount of disturbance caused by excavating root systems. The results of traditional methods of N2-fixation assessment, such as nodule biomass measures, total-N comparison, and the acetylene reduction assay are not readily convertible into actual amounts of N2-fixed. However, more direct methods, including the 15N natural abundance and 15N enrichment methods (Virginia et al., 1989; Hiers et al., 2003), ar e integrated over time and can be used to calculate a quantitative estimate of N2-fixed. These methods may eliminate the problems associated with instantaneous measures of fixation th at may fluctuate over diurnal and seasonal conditions, such as the acetylene reduction assay (Boring and Swank, 1984; Halvorson et al., 1992). By combining methods of assessing N2-fixation that include both instantaneous and cumulative measures and first applying them in a controlled study, a well-informed assessment can be made of the N2-fixing capabilities of a legume species.

PAGE 49

39 This study was designed to compare the N2-fixation capabilities of nine species of legumes native to the longleaf-wiregrass ecosy stem. Two additional objectives were also addressed: (1) to compare estimates of N2-fixation capabilities as derived from five different methods of assessment: nodule bi omass, the acetylene reduction assay, N transport and storage product analysis, 15N natural abundance, a nd total N content; and (2) to examine the effects of shading on their N2-fixation activity. Methods and Materials Planting Seedlings of nine native legume species were used in this experiment. They were propagated as described in Ch apter 2, and were obtained fr om Dr. Kay Kirkman at the Joseph W. Jones Ecological Research Center, Baker County, Georgia: Centrosema virginianum (L.) Benth. (CEVI), Clitoria mariana L. (CLMA), Crotalaria rotundifolia J.F. Gmel. (CRRO), Lespedeza hirta (L.) Hornem. (LEHI), Mimosa quadrivalvis (L.) (MIQU), Orbexillum lupinellus (Michx.) Isley (ORLU), Rhynchosia reniformis D.C. (RHRE) and Tephrosia virginiana (L.) Pers (TEVI). Seeds for the annual Chamaecrista nictitans (L.) Moench (CANI) were obtained from Dr. Ken Quesenberry (Agronomy Department, University of Florida). Six speci mens of each species were planted in pots that were specially designed to allow gas to fl ow through them as part of a closed system. These pots were constructed from a cappe d 35-cm long section of 10.5-cm diameter polyvinyl chloride (PVC) pipe. Fourteen specimens from each species were transplanted into black tree-seedling pots, 35-cm tall x 80-cm2 base (Stuewe and Sons, Corvallis, OR). Transplantation of seedlings was completed 11 February 2004. CANI was planted on 21 April 2004 from seeds that were germinated on moist filter paper. Due to the small numbers of plants,

PAGE 50

40 destructive harvests of seedlings were not possible at the beginning of the experiment (T0), but non-destructive measur ements were taken. Plant he ights at planting represented between 5.3 and 19.5 percent of the final height s. The smallest pl ants, such as ORLU and RHRE had the largest percentage of the final height present at T0, and the larger plants, such as LEHI, were represented by th e lower range of the pe rcentages. Leaves were also counted at T0 to determine percentage of leaves initially present. Percent leaves present at T0 ranged from 1.5 to 8.7 percent for th e smallto large-stature plants, respectively. All plants were grown outdoors in Gainesville, Florida (29o 38' N, 82o 20' W) between April and November of 2004. A shad e cloth enclosure was used to create an environment that provided a 0.54 fraction of tota l ambient light as determined using a LiCor Quantum Sensor, LI-185A (Chapter 2). Plants were inoculated by introducing native soil to th e pots. Two topsoils were collected at the Jones Center, from a fine -loamy, kaolinitic, thermic Typic Kandiudult (Orangeburg Series), and from a loamy, kao linitic, thermic Arenic Kandiudult (Wagram Series). Each collection was made in an ar ea with a diverse, thriving legume population. Soil collections were transported to Gainesvi lle, FL where they were stored in a cool, dark room and covered with plastic to main tain moisture. Equal parts of each soil collection were mixed in order to provid e inoculation for legumes that may be predominantly found in different locations. Pots were filled with purchased topsoil (Walmart Corp.) to within 6 cm from the t op of the pot. Next, 2 cm of the native soil mixture were added, and finally, the plant wa s transplanted, using additional topsoil to

PAGE 51

41 cover the roots as n eeded. Water was applied by drip -irrigation every 12 hours using a battery-operated timer (Rainbird). Plants from acetylene reduction assay pots a nd black plastic pots were destructively harvested on 11 November 2004. Roots and nodules were washed free of soil, nodules counted, and a 2-cm section was collected from the base of each stem for analysis. All tissues were dried to constant weight at 80oC. N2 Fixation Assessment Measurements of ethylene production as a result of exposing legume roots to acetylene (C2H4) gas were taken every three weeks ove r the course of the experiment. Repeated assays using the same plant were ma de possible by the use of a non-destructive, flow-through system. Specially-designed pots were used to allow th e plants to be repeatedly assayed with minimal disturbance. The gas input line wa s attached to the bottom of the pot. A lid matching a 3.6L (20-cm tall x 17-cm diamet er) plastic container (Rubbermaid Inc.) was permanently attached with wing nuts and screws to the flange that formed the top of each pot, and the joints were all sealed air-tight using weather-strip cau lking. The center of each lid was removed to reveal the mouth of the pot. The plastic containers were inverted over the top of the plan t and sealed to the stationary lid. On the bottom of each plastic container (the top of the apparatus when inverted), a port was created over which the gas line would fit tightly. Air-tightness was confirmed for each sealed pot before gas flow was initiated. Ten percent acetylene was flowed into the pots at a rate of 1L min-1. Gas samples for each pot were taken both from the lines running into the pots and those running out. The inflow samples served as a baseline for the amount of ethylene that might be present

PAGE 52

42 in the acetylene source. The outflow samples, which flowed past the plant root systems, contained the ethylene generated from acetyl ene reduction as a result of nitrogenase activity. Duplicates of 1-mL samples were co llected using syringes that were inserted into the tubing. Samples were transported to the laborator y and analyzed immediately. Analyses were performed using two gas chromatographs (Hewlett Packard 5710A and Shimadzu GC-8A). Stem sections were cut from the basal 3cm of plants grown in PVC and black plastic pots as they were harvested (11 Nove mber 2004). Each stem was placed in a 20mL glass vial and covered with a phosphate buffer solution, mixed according to Herridge (1984), except ethanol was used as the solven t rather than water at the suggestion of Izaguirre (personal communicatio n). Vials with stems in solution were stored at 0oC until extracted. Stem sections were placed in a boiling waterbath (100oC) for 25 minutes to extract stem contents. Stems were removed and dried to constant weight at 80oC. Deionized water was added to each extract to a standard volume of 25 mL. Extracts were covered and stored at -30oC between analyses to prevent eva poration. Aliquots of each extract were analyzed for ureide, NO3, and -amino acid concentrations using spectrophotometry (Beckman DU 640). Ureide concentrations were estimated as the phenylhydrazone of glyoxalate using allantoin as the standard, and -amino acid concentrations were determined using a modification of th e ninhydrin method (Yemm and Cocking, 1955), with asparagine as the standard; both anal yses were conducted as described by Herridge (1984). NO3 was extracted using the salicylic acid method as reported by Cataldo et al. (1975). Taking into account that ureides contain 4 N atoms per molecule, an index of the

PAGE 53

43 relative abundance of ureide-N in each extract wa s calculated according to Peoples et al. (1996), where the bracketed variables indicat e the molar concentra tion of each of the extract constituents: RUI = 400[ureides] / (4[ureides] + [nitrates] + [ -amino-acids]) [3-1]. Use of 15N natural abundance tec hnique in the field requires that differences between the soil and atmospheric 15N pools be well establishe d by also ascertaining 15N values of non-fixing reference plants and av ailable soil N (Virgini a et al., 1989). ORLU was selected as a non-fixing reference species for this controlled study was selected using corroborative assessments of N2-fixation. Stems and leaves harvested from plants grown in the PVC and black plastic pots were coarsely ground using a Cyclotec Samp le Mill, and then ground to a fine powder using a Spex CertiPrep Mi xer Mill 8000-D. Roots were not analyzed for 15N due to the amount of organic matter that remained attach ed after thorough clean ing. Finely ground tissues were analyzed for 15N natural abundance and N cont ent at the University of California, Davis (Stable Isotope Facility, Department of Agronom y, Davis, CA) using mass spectrometry. 15N natural abundance is expressed as 15N ( 15N depletion units): 15N = [(atom%15Nsample / atom%15Nstandard) 1] x 1000 [3-2] where the standard is the atom percent 15N concentration of atmospheric N2. Percent of total N derived from the atmosphere (% Ndfa) was calculated according to the equation: % Ndfa = 1 ( 15NN2-fixing plant / 15Nref ) [3-3]. Statistical Analysis Data were analyzed using analysis of variance (ANOVA) with species and light environment as main effects. If differences existed (p<0.05), Duncans multiple

PAGE 54

44 comparison test was used to determine which means differed significantly. The GLM procedure performed in the Statistical An alysis System (SAS, 2003) was used for ANOVA and post-tests. ANOVA analysis did not show any significant treatment effects for any of the datasets in this study. Therefore, differences further discussed rela te only to those among species, and means reflect a composite of both treatments. Results Survivorship Of the six specimens for each species that were planted in the pots for the flowthrough acetylene reduction assay, an averag e of three survived. LEHI and CRRO had the best survivorship, with five plants eac h, followed by CANI and TEVI with four, then CLMA with three. RHRE and ORLU had the worst survivorship, with no RHRE plants surviving, and only two ORLU, which did not permit statistical analysis of treatment effect on the acetylene reduction assay for these two species. An average of four out of fourteen speci mens grown in black tree seedling pots reached final harvest. LEHI, had the best survivorship with eight remaining, followed by TEVI and CLMA with six and five plants surviving until final harvest, respectively. CANI, CEVI, CRRO, RHRE, ORLU, and MIQU, had average or lower survivorship. ORLU and MIQU had only two and one plant surviving, respectivel y. However, since these plants were used in addition to the ones harvested from the PVC pots, statistical analysis was still possible for all species for most fixation indices. Shade treatment effect was not significant for any of the N2-fixation assessments in this study. Small samples and high variability contributed to the inability to detect statistical differences accord ing to treatment. Ambient lig ht for this study was also

PAGE 55

45 impacted toward the end of the season by tropical storm occurrence. Detectable differences in plant height were devel oping around this period, and the lowered irradiance overall may have impeded the maxi mum effect of shading during this critical growth period. However, results from biomass and nodulation data did not show significant responses to shading. Fixation Assessment Species differences in nodule mass were significant (p = 0.0006). MIQU had the greatest nodule mass, but further statistical de lineations were not de tectable due to the large amount of variation betw een individual plants (Table 3-1). None of the small herbaceous species in this study had an average nodule mass of more than 1g. Nitrogenase activity (acetylene reduction) wa s initially very low, overall, however, later in the season there were isolated peak s of substantial ethylen e production from the older plants (Figure 3-1). CANI, LEHI, CEVI, and MIQU all showed an increase after the middle of the season, around 28 June (Day 180). CLMA showed the lowest nitrogenase activity acro ss the season, never producing more than 0.3 mol C2H4 hr-1 at any given date. ORLU also showed low activity, with only a single peak above 0.3 mol C2H4 hr-1. Comparison of late season maximu m ethylene peaks using analysis of variance showed a significant species eff ect (p = 0.0002). The peak for MIQU was significantly greater than all other species maximums, but no other significant differences among the remaining species were detected (Figure 3-2). The N transport and storage products ex tracted from stem sections of native legumes were dominated by ureides and -amino acids. The relative concentration of amino acids (RAC) was approximately equal to that of the relative ureide concentration (RUC) for most species, followed by a very minute NO3 concentration (Table 3-2).

PAGE 56

46 Ureide appears to be an impor tant transport/storage product molecule for this suite of species. The differences among species were not significant for RUC or RAC, but were highly significant for total extracted N and relative NO3 concentration (RNC; p = 0.0006 and 0.0001, respectively). The 15N values of the species were significan tly different and ranged from -3.13 to -1.45 (Figure 3-3). CLMA had the 15N value closest to zero, followed by CANI, CEVI and TEVI, which all had values of approxima tely -2.3. ORLU and RHRE had the most negative values, approximately -3.2. However, only CLMA and CRRO were significantly different from ORLU, which was used as the non-fixing reference. Percent Ndfa was calculated (Equation 3-3) using ORLU as the non-fixing reference. ORLU had very low nitrogenase activity (acetylene reduction) across the season, indicating very little, if any, N2-fixation, in spite of having numerous small, but apparently ineffective nodules. Specie s effect was significant for percent Ndfa (p = 0.0002), and values ranged from 12.0 to 54.9 percent (Figure 3-4A). CLMA was determined to have the highest percent Ndfa, followed by CRRO TEVI CEVI MIQU LEHI RHRE. Percentage of N (gN g-1 tissue) in stem and leaf tissues ranged from nearly 3.0 percent to approximately 1.5 pe rcent (Figure 3-4B). Specie s effect was significant for percent N (p = 0.0001). MIQU had the highest percentage of N (2.83 percent), followed by CANI, then CRRO CLMA CEVI ORLU TEVI, and RHRE. LEHI had the lowest percentage of N in its tissues (1.47%). Total grams of plant accumulated N over the season was significantly different among species (p = 0.0001). CANI and MIQU had the largest amount of accumulated N,

PAGE 57

47 Table 3-1. Nodule mass and number of nodules. Rankings are according to Duncans post-test. RHRE not included in ANOVA (n = 1). Species Mass (g) # Nodules (per plant) MIQU 0.9114 a 218.60 a CEVI 0.3303 b 59.43 b TEVI 0.1316 b 51.40 b CANI 0.1175 b 44.00 b CRRO 0.0864 b 27.5 b CLMA 0.0818 b 32.17 b LEHI 0.0547 b 48.92 b ORLU 0.0229 b 27.25 b RHRE 0.0100 7.00

PAGE 58

48 Chamaecrista nictitans 120 150 180 210 240 270 300 0 1 2 3 4Calendar Day moles ethylene hr1plant-1 Centrosema virginianum 120 150 180 210 240 270 300 0.0 0.2 0.4 0.6 0.8 1.0 2 3 4Calendar Day moles ethylene hr1plant-1 Clitoria mariana 120 150 180 210 240 270 300 0.0 0.2 0.4 0.6 0.8 1.0 2 3 4Calendar Day moles ethylene hr1plant-1 Crotalaria rotundifolia 120 150 180 210 240 270 300 0.0 0.2 0.4 0.6 0.8 1.0 2 3 4 Calendar Day moles ethylene hr1plant-1 Lespedeza hirta 120 150 180 210 240 270 300 0.0 0.2 0.4 0.6 0.8 1.0 2 3 4Calendar Day moles ethylene hr1plant-1 Mimosa quadrivalvis 120 150 180 210 240 270 300 0 1 2 3 4Calendar Day moles ethylene hr-1plant-1 Orbexilum lupinellus 120 150 180 210 240 270 300 0.0 0.2 0.4 0.6 0.8 1.0 2 3 4Calendar Day moles ethylene hr1plant-1 Tephrosia virginiana 120 150 180 210 240 270 300 0.0 0.2 0.4 0.6 0.8 1.0 2 3 4Calendar Day moles ethylene hr1plant-1 Figure 3-1. Ethylene production trends for the growing season by species. Data shown are means SE with treatments combined, a nd lines show mean trends.

PAGE 59

49 Figure 3-2. Maximum ethylene production (C2H4 reduction) peaks. Data shown are means SE. Different letters represent sta tistical differences (Duncans posttest).

PAGE 60

50 Table 3-2. N-transport/storage products extracted from stem sections. RUC, RAC, and RNC represent relative con centrations of ureides, -amino acids, and NO3, respectively. RUI and Total N are rela tive ureide index (Equation 3-1) and total extracted N (mmol N g-1 stem). Letters represen t statistical differences within columns according to Duncans post-test. Species effect was not significant, ns. Species RUC RAC RNC RUI Total N (mmol N g-1 stem) CANI 45.68 54.28 0.032 a 73.01 0.0018 b MIQU 34.26 65.72 0.008 e 62.11 0.0055 a LEHI 27.15 72.82 0.020 bc 56.26 0.0014 b CEVI 46.09 53.89 0.013 de 67.15 0.0049 a CRRO 55.75 65.75 0.013 cd 67.15 0.0035 ab TEVI 24.84 75.14 0.010 e 48.39 0.0035 ab ORLU 51.83 48.14 0.026 b 79.79 0.0056 a CLMA 44.23 55.75 0.013 de 67.15 0.0035 ab

PAGE 61

51 Figure 3-3. 15N values by species. Data shown are means SE. Different letters represent statistical differences (Duncan s post-test). RHRE, (n=1) was not included in the ANOVA.

PAGE 62

52 A B Figure 3-4. Mean % Ndfa, %N, and total N by species. Different letters represent statistical differences (Duncans post-test). A) Percent of total N derived from the atmosphere. CANI and ORLU did not have adequate replication for inclusion in ANOVA, and ORLU was used as a non-fixing reference in the calculation of % Ndfa (Equation 3-3). B) Percent N in aboveground tissues. C) Total N content of aboveground tissues.

PAGE 63

53 C Figure 3-4. Continued

PAGE 64

54 followed by LEHI, CEVI, and CRRO, and finally, TEVI, ORLU, CLMA, and RHRE (Figure 3-4C). Discussion Comparison of Methodology Although there were some diff erences in the way the N2-fixation capacity of the species were ranked by each assessment te chnique, the species could generally be assigned into highand low-fixe r categories. With little discrepancy between assessment results among methods, MIQU, CANI, CRRO, CLMA and CEVI showed higher N2fixation potential, and LEHI, TEVI, and RHRE showed relatively lower potential. The results for ORLU generally corroborated that this species did not form effective nodules in this study and thus had the least effective N2-fixing capacity, approximately none (Figures 3-1, 3; Table 3-3). The acetylene reduction assay, an instantaneous indicator of nitrogenase activity, is strongly affected by environmental stresse s such as drought conditions and light reduction, which reduce photosynthesis rates. Th erefore, seasonal and diurnal patterns in nitrogenase activity can generally be well documented with frequent acetylene reduction assays (Boring and Swank, 1984; Zitzer and Dawson, 1989; Halvorson et al., 1992; Peoples et al., 1996). This method of assessing N2-fixation is also a quick and inexpensive, instantaneous first assessment of nitrogenase activity. However, it is necessary to follow temporal patterns of plant development to adequately index an annual pattern of nitrogenase activity, and even the n, the assay is not a direct measure of N2fixation, since many species have nitrogenase systems with varying efficiencies and hydrogenase activities (Zitzer and Dawson, 1989).

PAGE 65

55 Table 3-3. Specific nodule ac tivities of species in this study and other comparative reports. Specific nodule activities for this study were estimated using the results of the final acetylene reducti on assay (22 October 2004) and biomass of harvested nodules. Data from this study are ranges and means SE. 1Hendricks and Boring, 1999; 2Hogberg and Kvarnstrom, 1982; 3Boring and Swank, 1984; 4Zitzer and Dawson, 1989. Species Specific nodule activity ( mol C2H4 hr-1 mg nodule-1) Centrosema virginianum (CEVI) 0.272 0.272 Mimosa quadrivalvis (MIQU) 0.297 0.116 Other native legumes 0.002 to 0.007 Orbexilum lupinellus (ORLU) 0 Desmodium viridiflorum1 0.08 Lespedeza procumbens1 0.04 Leucaena leucocephala2 0.048 Robinia pseudoacacia3 0.05 Alnus glutinosa4 0.02 to 0.02 Eleaganus angustifolia4 0.002 to 0.01

PAGE 66

56 Specific nodule activity comparisons betw een species in th is study and other similar and larger species of legumes sugge st that CEVI and MIQU have extremely high nodule efficiency (Table 3-3). However, these comparisons may not be completely accurate due to the difference in sampli ng techniques. The flow-through assay to measure nitrogenase activity (acetylene reduc tion) that was used in the current study, avoids many of the problems associated with field sampling, which was used in the comparative studies, such as limited nodule re covery and possible evolution of ethylene from excised plant parts. The dispar ity between values given for strong N2-fixers in this study as compared to others may be part ially due to these sampling differences. The use of the relative ureide a bundance technique for assessing N2-fixation was limited in this study by the extremely low NO3 concentration in the stems which shifted the relative ureide content (27-55%) higher th an would have been expected according to results reported in other st udies (RUC, 1-33%; Izaguirre-Ma yoral et al., 1992; Sicardi de Mallorca and Izaguirre-Mayoral 1993; Medina and Izaguirr e, 2004). As a result, the RUI, which takes into account the 4:1 atomic ratio of ureide to NO3 and is the value usually compared among species, was very high for all species in this study, making the use of delineation between high and low N2-fixers according to RUI values as reported by Izaguirre et al. (1992; RUI > 60, high; RUI < 30, low) difficult to apply to this study. The most valuable comparisons for this study appeared to be between those species that appeared to be using a relatively high or low amount of soil NO3, as distinguished by RNC values (Table 3-2). The 15N relative abundance technique is most us eful in field studies with maturing plants, with adequate sample numbers of legumes, and with a verified non-N2-fixing

PAGE 67

57 species as a reference plant. In this cont rolled study, ORLU was sele cted as a reference species using corroborative methods to determine that it was non-N2-fixing and could represent soil 15N uptake values. The strong 15N signature of CLMA is not easily understood in that this species had one of the lowest nitr ogenase activities across the season. CLMA also had one of the lowest RN C indices, which would indicate that it was using less soil-N than some of the other species. It is possi ble that ephemeral periods of peak nitrogenase activity occurred within the three-week intervals between acetylene reduction assays, and it is also possible that since this species ha d very small biomass (Chapter 2), low levels of N2-fixation could have a relativ ely stronger influence over the 15N value than the same levels of N2-fixation in a larger plant. It appears possible the consistent, low levels of nitrogenase ac tivity could have a strong influence on the 15N values of CRRO and TEVI, as well (Figures 3-1, 3-3). Total N does not follow the same pattern as percent N because it also incorporates the total biomass of the plant which varied dramatically among the species (CLMA, 9 g to MIQU, 40 g). This measure follows th e pattern of aboveground biomass accumulation very closely (Chapter 2). Although not completely applicable as a comparative measure of N2-fixation, an estimate of total N for each of these species can be very useful for considering which should be included in rest oration plantings. Those with the greatest biomass and tissue-N overall will contribute the most to bui lding the N-pool, soil organic matter, and N availability in depleted soils (M arkewitz et al., 2002). This estimate is also useful in determining which of the species could provide the greatest amount of N-rich material available for wildlife to browse (Hendricks and Boring, 1999).

PAGE 68

58 This study points out some differences in N2-fixation capabilities among the species of legumes native to the longleaf-wir egrass woodland ecosystem, as well as differences in results that can occur from usi ng various assessment methods. The merits and problems associated with using a variety of methods to assess N2-fixation in this controlled study were most evident when co mparing results including some immature, small plants. Evidence from shorter-time-based measures such as the acetylene reduction assay and N-transport product analys is did not always co rroborate definitively with cumulative measures for these small plants. Cumulative measures (total N, 15N approaches) may hold more insight for older, established perenni al plants, especially field populations. Nodules are ephemeral and N2-fixation rates will change diurnally and seasonally according to environmental conditio ns. Instantaneous methods may not be the most appropriate measures to estim ate actual or maximum potential N2-fixation, but can serve as a quick and inexpensive way to determine seasonal a nd diurnal-dependent patterns of nitrogenase activit y. Instantaneous measures may be best used to document age of effective symbiosis and to define de velopment of peak nitrogenase activity when the plant has accumulated effective leaf area, and to examine detailed responses of mature plants to varying environmental c onditions and stresses (Dreyfus et al., 1988; Sprent, 1999; Vitousek at al., 2002). When conducting an ecosystem field study, it is often important to use the leastdestructive method possible as well as cumula tive measures over long periods of time. For perennial species, especially in fre quently-burned ecosystems, removing the aboveground portion of the plan t does not totally remove the individual from the ecosystem, is similar to burning, and is much less destructive than the soil disturbance

PAGE 69

59 caused by digging for roots and nodules. Additi onal error can also be introduced due to the fact that full root and nodule recovery is never assure d in field conditions. Thus, methods utilizing aboveground plant tissues for N-transport product analysis, total N content and 15N natural abundance may be highly e ffective for use in long-term field studies when soil available N is decisively di fferent from the atmospheric isotopic value (Virginia et al., 1989; He ndricks and Boring, 1999). Species Differences Species effects were significant for most of the N2-fixation assessment techniques used in this study. This strong species effect echoes the great divers ity of growth forms and life histories among the species in this st udy. Each of the three major subfamilies of Leguminoseae were represented, Caesalpiniodeae Papilionoideae and Mimosoideae Two very different growth forms were also represented: spreadi ng to vining plants (CRRO, CEVI, CLMA, and MIQU) and erect forbs (ORLU, CANI, LEHI, TEVI, and RHRE). LEHI becomes semi-woody by matura tion at the end of a growing season. All of the species, except for CANI, in this study were perennial, as is common for species native to fire-adapted ecosystems (Mor gan, 1999; Jacobs and Schloeder, 2002). Rates of nitrogenase activity (C2H4 reduction) increased to ward the end of the growing season, when most of the plants had accumulated substa ntial photosynthetic tissues, and had formed active nodules for N2-fixation. CEVI and MIQU show exceptional nodule activity levels as compared to other legumes in the longleaf-wiregrass ecosystem, large woody legumes and actinorhyzal N2-fixers (Table 3-3). In contrast, the other native legumes in this study showed much lower nodule act ivity levels than the two leading species, however, th ey are still within the ra nge represented by other N2-fixers. The differences in the groups of native species in this study likely indicate that the large

PAGE 70

60 vines, CEVI and MIQU, reached a level of maturity and nodule development that some of the other species were not able to reach in an initial growing season. The pattern of acceleration in nitrogenase activity toward th e end of the season might begin earlier, and be more rapid with older field populations of perennial plants that would readily establish photosynthetic tissues using car bon and N stored belowground from the previous year (Hendricks and Boring, 1999). The two dominant constituents of the stem extracted-N, ureide and -amino-acids, appear to have an inverse relationship for the species in this study (Table 3-2). The relationship between these two constituents is different from ot her studies where NO3, was a much larger component (3.5 to 35.9%, Sicardi de Mallorca a nd Izaguirre-Mayoral, 1993; 18 to 70%, Peoples et al., 1996). T hose studies had an inverse relationship between RUC and RNC. In this study, RU C and RAC were not si gnificantly different among species, however, patterns among species according to RNC did show statistical differences, and it was elevated for th ose species with relatively low N2-fixation capabilities as suggested by ot her corroborative methods. Thos e species with the highest NO3 concentration, although very low compared to values in the other studies (3.5 versus 61.1%), were also among those with the lowest percent N and 15N values in aboveground tissues in this study indicating lower N2-fixation. Sicardi de Mallorca and Izaguirre-Mayoral (1993) and Izaguirre-Mayoral et al. ( 1992) found a similar pattern when comparing lowto high-N2-fixers. RNC is likely dependent upon soil NO3 availability, which is very low under the ty pical field conditions where these native legumes are found (Wilson et al., 1999), and RNC values sampled in the native woodland could potentially be even lower than those reported in the current study.

PAGE 71

61 The species examined in this study were again divided into two main groups of fixation activity using 15N values (high: CLMA CRRO TEVI CEVI MIQU; low: LEHI CANI RHRE ORLU), with CLMA indicati ng the least depletion of 15N (Figure 3-3). Other reported 15N values for LEHI and CANI also indicate that these two species are relatively lower N2-fixers (Hendricks and Boring, 1999) LEHI was not statistically different from the non-fixing reference in a study conducted by Hendricks and Boring (1999). The annual CANI was also identi fied as a low-fixer by the cumulative, 15N abundance assessment (Figure 3-3). This annua l species would have relied on more soilN uptake to become established than perennial species that were planted at the beginning of the growing season, and the RNC value in Table 3-2 indicated that it may be the species with the greatest rate of NO3 uptake. Percent Ndfa of the species in this study ( 12-55%; Figure 3-4A) were lower than values reported by Hendricks and Boring (1999 ) for other species of mature legumes from a similar ecosystem (54-88%). Species order according to % Ndfa is similar to %N, with the exception of MIQU, which had a large concentration of N, but a lower %Ndfa. Aboveground percent N also divided the specie s in this study into two categories of potential highand low N2-fixation, those species with >2% N (MIQU, CANI and CLMA) and those with <2% (CEVI, CRRO, TEVI, ORLU, LEHI and RHRE). Potential N2-fixation rates indicated by %N agree with the acetylene reduction assay and 15N techniques for MIQU (high) and LEHI (low). However, comparatively large amounts of NO3 uptake by CANI and CLMA compared to ot her species indicate th at high levels of N2-fixation are not related to high N concentra tions in the tissues of these species. TEVI, which had lower percent tissue N in this study than that reported fo r Hiers et al. (2003),

PAGE 72

62 also failed to flower within the growing season examined here (Chapter 2). TEVI has been shown to respond to fire with increased flower production (Hiers et al., 2000) and a significant elevation in percent N (Hendricks and et al., 2003). Thus, for this study, TEVI was probably lacking the maturity and pe rhaps stimulation by fire and subsequent phosphorus enrichment needed for maximum N2-fixation to be detected (Christensen, 1977, Gholz et al., 1985). Summary Estimates of N2-fixation capabilities of native legumes in this study were most conclusive for the two rapidly growing and apparently quickly ma turing vine species CEVI and MIQU. These two species s howed strong indices of fixation by both instantaneous and cumulative measures. Th e annual CANI and perennial LEHI accreted large amounts of N in biomass, but apparently from higher NO3 uptake and low rates of N2 fixation. The data were less conclusive for the remaining, sl ower-growing, smaller species that accreted N more slowly and ha d variable indices of lower but significant rates of fixation among the methods of assessm ent. However, these species represent only a small sample of the 40 species of legumes present in the longleaf wiregrass ecosystem, which likely also have varying capabilities for N2-fixation (Hainds et al., 1999). For example, other large, semi -woody legume species of the genera Lespedeza and Desmodium within this ecosystem have shown higher potential for N2-fixation than LEHI and should be investigated further (H endricks and Boring, 1999). With further investigation of mature plants in the fiel d, using a combination of total N content, 15N, and N transport product analysis would provide for effective delineation of N2-fixation for these species and others. Such an a pproach may provide more information to

PAGE 73

63 facilitate using native legumes in restoration plantings for im proving soil characteristics, wildlife habitat and forest productivity.

PAGE 74

64 CHAPTER 4 GROWTH AND N2-FIXATION OF NATIVE LEGUM ES IN LONGLEAF PINE RESTORATION Introduction Conservative estimates of nitrogen inputs from biological fixation by native legume populations in pine woodland and grassland ecosystems have been grossly estimated from 5.2 to 9 kg N ha -1 yr -1 (Ojima et al., 1990; Hendric ks and Boring, 1999). Native legumes are often found in high density popul ations across highly-variable light and water regimes in the longleaf pine( Pinus palustris Michx.) wiregrass ( Aristida stricta Mill.) ecosystem of Southeastern North America ranging from xeric sandhills to wetmesic sites to edges of depressional wetlands (Hainds et al., 1999). However, species of legumes native to this ecosystem have been reported to have large variability in symbiotic N2-fixation. Many factors may control N2-fixation rates in woodland ecosystems, including water and nutrien t availability, rhizob ium populations, and especially light, given the re quirement for large energetic costs to drive symbiotic N2fixation (Sprent, 1987; V itousek et al., 2002). Recent restoration initiative s on private and public lands in the southeastern U.S. coastal plain have resulted in the plan ting of approximately 283,000ha of former agricultural, pulpwood plantation and fire suppressed land back into longleaf pine stands through the USDA Conservation Reserve Program since 1996 (CRP; Coffey and Kirkman, 2004). These young longleaf stands grow vigorously on sandy, carbonand Ndepleted former agricultural sites (Markewitz et al., 2002).

PAGE 75

65 Restoration of groundcover in young, planted longleaf pine stands is important for rebuilding soil organic matter and N, for pr oviding wildlife food and cover, and for enhancing pyrrhic fuel continuity necessary to reintroduce frequent prescribed fires. Native grasses and legumes should be prefer red over exotics for reintroduction, as has been made evident by the inva sive nature of two species, Lespedeza bicolor Turcz. and L. cuneata (Dum. Cours.) G. Don, that have been previously introduced for soil improvement and wildlife fora ge (Miller, 2003). Field rese arch is needed to help determine desirable native species for reintr oduction on targeted sites. There are also issues related to the timi ng of groundcover species introduction in young pine stands, due to the large differences in light tran smittance through developing longleaf canopies between planting and maturation, and followi ng thinning operations associated with silvicultural practices (Mulligan and Kirkman, 2002). The objectives of this study were to test for the differences in biomass accumulation and distribution, N-content, and N2-fixation potential of six legume species planted in three levels of li ght under longleaf pi ne canopies. This field test is an important step in understanding growth and N2-fixation of native forest legumes under shaded conditions. Materials and Methods Site Description This common garden study was conducted at Ichauway, a property managed by the Joseph W. Jones Ecological Research Cent er (JWJERC), a 12,500 ha reserve located in Baker County, Georgia, USA (31o19'N and 80o20'W). The climate for this region is humid subtropical. Mean da ily temperature during the st udy (10 May-7 November 1004) was 26.9oC, and cumulative rainfall was 595mm. Although most of Ichauway lands were

PAGE 76

66 under longleaf pine savanna when they were acquired in the mid 1930 's, some of the land has been cultivated. Over the years, cultivated areas have been plowe d, fertilized with N, P and K, and planted to crops including cotton ( Gossypium spp.) and sorghum ( Sorghum spp.) (Markewitz et al., 2002). The soil at th e site was a fine-loamy, kaolinitic, thermic Typic Kandiudult (Norfolk Series). The 14 year-old longleaf stand used in this study was planted on a formerly cultivated area according to CRP st ocking recommendations of 1,235 trees ha-1. Three canopy opening levels were initially located using a densiometer. Plots representing intermediate light levels (an average of 48% openness) were established in locations where a single tree had previously died, probabl y due to fire scorching or insect damage, leaving a small gap in the canopy. Closed canopy plots (an average of 9% openness) were located near the intermediate plots in an area with no missing trees. Open canopy plots were established at the edges of the st and. Trees within appr oximately 1 m of the open plots (3 trees per plot) were removed in order to achieve desired openness, and limbs below 2m height were removed adjacent to all plots for ease of access. Canopy openness was more accurately quantifie d using a line quantum sensor on a clear, cloudless day (Li-Cor, Inc., Lincoln, Ne braska). Five light readings, taken after pruning and tree removal, were made on 5 May 2005 within each plot and in an adjacent open field between 12:39 and 15:13 EST. The proportion of light pres ent in plots versus the open field, expressed as photos ynthetically ac tive radiation mol m-2s-1 (PAR) was used to describe canopy openness. Experimental Design and Planting Four replicated 9 m2 plots were arranged according to a completely random statistical design under the thre e light environments. Each plot was planted with five

PAGE 77

67 plants from eight different specie s of 7 month-old native legumes: Centrosema virginianum (L.) Benth. (CEVI), Desmodium ciliare (Muhl. ex Willd.) DC. (DECI), Lespedeza angustifolia (Pursh.) Ell. (LEAN), Lespedeza hirta (L.) Hornem. (LEHI), Mimosa quadrivalvis (L.), Orbexillum lupinelus (Michx.) Isley (ORLU), Pediomelum canescens (Michx.) Rydb., and Tephrosia virginiana (L.) Pers (TEVI). Nomenclature follows Wunderlin and Hansen (2003). S eedlings were propagated at the JWJERC by Dr. K. Kirkman as described in Chapters 2 a nd 3. Plants were randomly arranged in rows that were spaced 60 cm apart, with 30 cm spacing between plants. Planting was completed 10 May 2004. Weeds were suppressed by mulching the plots with on-site pine straw and supplemented with more from the adjacent area. Time Domain Reflectometry (TDR) rods we re used to monitor soil moisture. 30and 90-cm TDR rods were placed in the corn ers of each plot, and readings were taken every two weeks. Volumetric soil moistu re for each plot across the season did not indicate any occurrence of severe drought conditions, nor did the plants undergo any periods of defoliation. Soil moisture was not significantly different among individual plots or between light treatments. Due to a dry period around the time of planting, plants were hand watered to aid root establishment (Figure 4-1). Approximately 1L was applied to each plant on a bi-weekly basis between 18 and 30 May, and once on 11 June. Plants were destructively harvested from each plot at the end of the growing season, between 25 October and 7 November, 2004. Plant heights, except for CEVI, were measured from soil surface to the top of the stem before collection. Approximately 6 L of soil were removed with each plant by digging at a 10 cm radius around each stem

PAGE 78

68 Figure 4-1. Volumetric soil moisture patterns for all plots. Readi ngs were taken every two weeks between 11 May and 8 Octobe r, 2004. Data shown are means SE

PAGE 79

69 to a depth of approximately 20 cm. Roots a nd nodules were taken to a field lab, washed free of soil, and collected. Leaves, stems, roots and nodules were separated, dried to constant weight at 70oC and weighed. Due to the extremely small size of the plants in the closed light environments, the plant parts were weighed and recorded, but no further analysis was conducted on this set of sample s (Figure 4-2). Leaves and stems from the plants harvested in the open light envi ronment plots were composited and ground for further analysis. Plants harvested from the in termediate light environment were separated into leaves, stems and roots and ground for furthe r analysis. Plants in this light treatment were also used for a related retranslocation study. Longleaf pine n eedles were collected from mid-canopy from trees near the e xperimental plots for use as a non-N2-fixing reference for assessment by 15N. Needles were prepared for analysis in the same manner as the legume tissues. The usefulness of the 15N natural abundance tec hnique for assessing N2-fixation under field conditions has been verified th rough numerous agricultural and controlled studies. However, the utility of this technique in the field is dependent on comparison of 15N values of a non-N2-fixing reference plant, to repr esents the ratios of isotopic nitrogen forms present in the soil with those of the legumes. Ground tissues from legumes and pine needles were analyzed for 15N natural abundance and total N content at the University of California, Davis (Stabl e Isotope Facility, Department of Agronomy, Davis, CA) using mass spectrometry. Leaf and stem 15N values of plants harvested from the intermediate light environment were composited so that comp arisons with the aboveground (leaf + stem) 15N values from the open light environment could be made. CEVI and ORLU did not

PAGE 80

70 produce enough samples for this analysis. A weighted 15N value for the plants from the intermediate plots was calculated: 15Naboveground = [( 15Nstem x Total Nstem) + ( 15Nleaves x Total Nleaves)] / (Total Nstem + Total Nleaves) [4-1]. Using aboveground tissue values of 15N from intermediate and open light treatments, an estimate of percent N derived from the atmosphere (%Ndfa) was calculated: % Ndfa = 1 ( 15NN2-fixing plant / 15Nref ) [4-2]. Statistical Analysis Data were analyzed using two-way anal ysis of variance (ANOVA), with species and light environment as main effects. Where significant effects existed (p<0.05), Duncans multiple comparison post-test was used to determine which means differed significantly. The GLM procedure performed in the Statistical Analysis System (SAS, 2003) was used for ANOVA and post-tests. Results Preliminary Results and Survivorship PAR readings in the open field adj acent to the study site were 1938 10.6 mol m2s-1, and plot readings were 370 45, 1214 103, and 1539 61 for closed, intermediate and open plots, respectively. Canopy openness for each plot type, expressed as a fraction of adjacent field light le vels were 0.085, 0.614, and 0.834 for closed, intermediate, and open, respectively. Diffe rences in percent canopy openness among plot types were highly significant (P <0.0001). There was an average survival rate of 29% across species where LEHI > TEVI > DECI > LEAN > ORLU. CEVI had the lowe st survival of only 15%. An average sample size of five for each species in each light environment remained except for CEVI

PAGE 81

71 in the intermediate light environment (n=1). M. quadrivalvis and P. canescens were excluded from further analysis in this study due to poor growth and survivorship. M. quadrivalvis showed very little change in size at mid-season and final harvest date, therefore it was not harvested. Other field observations of M. quadrivalvis suggest that biomass accumulation may have been directed predominantly belowground. Taproots of maturing M. quadrivalvis in the field can be as long as 3m (personal observation). As in a companion study, P. canescens developed brown spots, defo liated and appeared to be dead by mid-season (Chapters 2 and 3). Plants grown in the closed light environment accumulated very little biomass over the season as compared to plants grown in th e other light environments (Figure 4-2). Due to the small growth response of plants in the closed light environment, plants harvested from this set of plots were not analyzed for N content nor assessed for N2-fixation. Growth Total aboveground biomass showed signifi cantly different species and light treatment effects (Table 4-1, 2). Light e ffect on biomass was Open > Intermediate > Closed. However, CEVI growth showed no statistical differences between open and closed treatments and was not adequately represented for intermediate. LEAN did not have significantly different greater biom ass in the open environment versus the intermediate, but both were greater than the closed. In the open light treatment, the biomass was DECI > CEVI LEHI TEVI LEAN ORLU; only DECI had a significantly greater biomass than any other sp ecies (Figure 4-2). Biomass differences by species were not statistically different in th e intermediate or closed light treatments.

PAGE 82

72 Figure 4-2. Aboveground biomass and ch ange in plant heights from T0 by species in each of the three light treatments. Data shown are means SE, and different letters within a species represent statis tically different means (Duncans posttest). Only one plant was harvested for CEVI in the intermediate light environment.

PAGE 83

73 Table 4-1. Analysis of variance results for e xperimental variables. Light treatment and species were main effects. Values for total aboveground biomass (Total AB Biomass), root-to-shoot ratio (R/S) an d nodule biomass (nodulation) represent differences across all three light treatments. Results for %N, 15N, Total N, and %Ndfa represent differences among open and intermediate light treatments for aboveground tissues. Main Effects Variables Species Light Species x Light Total AB Biomass (g) 0.0094 <0.0001 0.0292 R/S 0.2247 <0.0001 0.2652 Nodulation (g) 0.0149 0.0023 0.5694 %N <0.0001 0.0002 0.0421 15N <0.0001 <0.0001 0.9063

PAGE 84

74 Table 4-2. Total biomass (aboveand belowg round tissues, including nodules) per plant and aboveground values for %N, 15N, and total N. Different letters within each column represent statistically different means. Closed Light Treatment Species Total Biomass (g plant-1) R/S Nodule biomass (mg plant-1) CEVI 5.38 0.51 98.2 DECI 2.06 1.93 11.22 LEAN 0.52 2.26 6.22 LEHI 1.64 1.79 11.03 ORLU 0.47 1.83 3.74 Intermediate Light Treatment Species Total Biomass (g plant-1) R/S Nodule biomass (mg plant -1) %N 15N Total N (g plant-1) CEVI 2.50 0.46 22.80 DECI 7.43 0.90 9.93 1.82 b -3.06 0.05 LEAN 2.86 1.07 13.42 1.68 b -2.59 0.03 LEHI 3.29 0.76 17.23 1.96 b -2.80 0.05 ORLU 0.67 1.65 15.40 TEVI 2.25 0.91 29.08 2.79 a -2.20 0.03 Open Light Treatment Species Total Biomass (g plant-1) R/S Nodule biomass (mg plant-1) %N 15N Total N (g plant-1) CEVI 15.59 b 0.89 131.50 a DECI 37.27 a 0.29 74.63 ab 1.33 b -1.06 0.40 a LEAN 8.60 b 0.55 37.57 b 1.76 b -0.67 0.16 b LEHI 13.53 b 0.54 18.77 b 1.21 b -1.25 0.13 b ORLU 1.82 b 1.16 16.00 b TEVI 9.65 b 0.58 55.96 b 2.33 a -0.46 0.15 b

PAGE 85

75 Plants grown in the closed light environmen t were the shortest, overall. The lack of height growth is apparent when the heights of the plants at harvest are compared with the heights of the plants at time of planting (T0). The largest change in height from T0 was found in the open light environment, and the difference between these plant heights and those in the other treatments were significant overall, with few exceptions for individual species (Figure 4-2). Root-to-shoot ratios (R/S) were also signi ficantly different by light environment, but species effect was not si gnificant (Table 4.1, 4.2). Most of the species showed the following pattern of significant differences in R/S across the three light environments: Closed > Intermediate > Open. However, CEVI, a vine, showed an opposite R/S pattern, Open > Closed. (The intermediate light level was not sufficien tly represented.)ORLU showed no differences in R/S across light treatments, and DECI did not show a significant difference between the closed and intermediate treatments (Figure 4-3). Nodulation, as measured by nodule bioma ss (mg), showed significant light treatment and species effects. Due to a larg e degree of variability in nodule biomass, the only species to show the significant light treatment effect was DECI (p = 0.0532) and followed the pattern: Open > Closed > Interm ediate. General trends of nodule biomass accumulation for the other species were Open > Intermediate > Closed, although too variable to show significant differences. In all light treatments, CEVI, TEVI and DECI appeared to have the grea test nodulation (Table 4-2). All species had a higher pe rcentage of N concentra tion in aboveground tissues (stem + leaves) where grown in the intermed iate light environment than in the open. Although species and treatment effects were si gnificant for %N (Table 4-1), statistical

PAGE 86

76 Figure 4-3. Root-to-shoot ratios by species in the three light treatments. Data shown are means SE, and different letters within a species represent statistical differences (Duncans post-test). Only one plant was harvested for CEVI in the intermediate light environment.

PAGE 87

77 Figure 4-4. Percent N concentration in a boveground biomass (stem + leaves) by species in the intermediate and open light envi ronments. Plants grown in the closed light environment were not analyzed for percent N. Data shown are means of four plots for each treatment. Statis tically greater %N is indicated by (Duncans post-test).

PAGE 88

78 differences in the two light treatments we re only represented by DECI (p = 0.0350) and LEHI (p = 0.0002). In both light treatments, TEVI had a statis tically higher %N than the rest of the species followed by LEHI > DECI > LEAN fo r the intermediate treatment and LEAN > DECI > LEHI for the open treatment (Figure 4-4). N2-Fixation Total N in aboveground tissues was significan tly affected by light treatment (Table 4-2). However, this difference was only re presented by DECI and TEVI, where plants grown in the open light treatment had statistica lly greater total N than those grown in the intermediate. Differences among species in th e intermediate light environment were not statistically different, but in the open light environment th ey were ranked: DECI > LEAN TEVI LEHI. Legume 15N values were not statistically di fferent (Figure 4-5) from the non-N2 fixing pine needle reference values in the intermediate light environment (p < 0.1419), but showed a significant difference from the reference in the open (p < 0.0014). 15N values of legumes were statistically affect ed by light environment (p < 0.0001), and this difference was represented in all of the speci es. For all legume species in this study, 15N values indicated that 15N in aboveground tissues was less de pleted (closest to atmospheric value) in those plants grown in the open lig ht treatment than in the intermediate (p = 0.0051 0.0044). Discussion Shading Effects on Species Plant growth was significantly greater, ove rall, in the open light environment than in the intermediate or closed light environments. The reduced level of PAR in the

PAGE 89

79 Figure 4-5. 15N and %Ndfa values by species for aboveground tissues in the intermediate and open light treatme nts. Data shown are means SE. Plants grown in the closed light envi ronment were not analyzed for 15N. Pine needles collected from adjacent to th e plots were used as the non-fixing reference. %Ndfa was calculated using Equation 4-2.

PAGE 90

80 intermediate (61.4% of incident) and closed (8.5% of incident) light environments compared to that of the open (83.4% of in cident) had significant impact on growth and N2-fixation rates of the legume species in th is study. The change in height between T0 and harvest for the species in th is study may be due to etiola tion and/or actual differences in growth. Lesser differences in elongation between intermediate and closed treatments for LEHI and TEVI may be more indicative of etiolation than growth, since biomass was not different between treatments for these sp ecies. DECI and LEAN, which did not show differences between plant heights in open and intermediate treatments suggest that they were less etiolated. LEAN, which had sm aller biomass accumulation differences among treatments showed further evidence that the differences in height between the intermediate and closed treatments reflected actual growth differences rather than etiolation. There were significant differences in R/S between closed and open treatments for all of the species in this study, suggesting a stronger al location of carbon to the roots in low light. Greater allocation to roots due to a variety of envi ronmental stressors, including shade, drought and fire is a we ll-documented physiologica l response (Sprent, 1973; Knapp et al., 1998; Paz, 2003; Fernndez et al., 2004) Allocation to roots may also be a response to increased competition from pine roots (Mulligan and Kirkman, 2002) and is a common adaptation to low-fertility soils (Paz, 2003). In the case of these legumes, a decreased rate of N2-fixation due to shading, and a subsequent increased reliance on soil sources of N could lead to a gr eater allocation of biomass to roots as the plants forage and compete for N resources.

PAGE 91

81 Smaller, non-significant differences between the R/S of plants grown in the open versus intermediate light environments may i ndicate that the threshold light level needed to initiate these allocation changes exist so mewhere between the intermediate and closed light levels used in this study. R/S reported for CEVI, ORLU and TEVI in this study (0.89, open; 1.83, intermediate; a nd 0.91, intermediate, respec tively) were much lower than those reported in a c ontrolled, potted study (approximately 1.5, 2.7, and 3 for CEVI, ORLU and TEVI, respectively; Chapter 3). R/S of LEHI in the controlled study (0.8) was similar to the R/S in the intermediate light environment in the current study (0.76). Sampling differences are the most obvious cause for many of these differences, especially since CEVI and TEVI, have extensiv e, lateral branching root systems that were not harvested in their entirety in this study, but that would have been contained in the pots of the controlled study. Dramatically different levels of N2-fixation (as assessed by 15N) in spite of a lack of difference in nodulation between the open a nd intermediate light environments (Table 4-2) indicates that nodule activity rather than bioma ss is more indicative of N2-fixation in these species. For example, ORLU had a comparable amount of nodulation biomass to other species in a previous, controlled study, but was determin ed to have relatively much lower N2-fixation rates than the other species in the study as assessed by both the acetylene reduction and 15N natural abundance assessments (Figure 4-4; Chapter 3). However, statistical differences in nodulati on between light treatment s were difficult to detect because of high variability between i ndividual plants due to sampling area in the field and variable plant sizes. Nodules that were present under shaded conditions, due to

PAGE 92

82 their relative expense versus effectiveness, may be highly susceptible to being sloughed by the legumes in response to stress (Gadgil, 1971). Legume N2-fixation under shade conditions has rarely been assessed due to the high energetic cost of nodule maintenance and the N2-fixation reaction (Vitousek et al., 2002). A small number of available studies do indicate that nodulat ion and subsequent N2-fixation by legumes is inhibited by sh ading (Gadgil, 1971; Sprent,1973). %Ndfa for the species grown in open light conditions in this study (44.5-79.7%) are similar to values reported for similar species by Hendricks and Boring (1999; 54-88%). The agroforestry legume species Calliandra calothyrsus and Sesbania sesban also have similar reported foliar %Ndfa values, 65-90%, respectiv ely (Stahl et al., 2002). Percent N concentration in leaf and stem tissues was greater in the intermediate light environment than in the open. The in creased concentration of N in the shaded leaves may reflect the reduced amount of carbon assimilation occurr ing under lower light conditions, which is corroborated by the opposite pattern of N accumulation in the open light treatment. Sun leaves tend to be th icker and contain larger amounts of connective and suberized tissue leading to a higher lower N concentratio n than shade leaves (Kramer and Kozlowski, 1979). However, the larger overall size of the plants in the open light environment versus the intermediate explai ns a higher total N value for open-grown versus more shaded plants. Aboveground %N reported in this study for the large, semiwoody, LEHI, LEAN and DECI (1.21 1.76% ) and for the less-woody, TEVI (2.33 0.17%) in this study are within th e same range as previously re ported leaf %N levels for similar, semi-woody (1.6 2.3%) and smaller herbaceous legumes (2.4%; Hendricks and Boring 1992).

PAGE 93

83 Ecological and Management Implications Survivorship was relatively low in this study, across all species. One possible explanation for the plants poor survival is the time of transp lanting. The small seedlings were planted in the spring when temperat ure, photosynthetic rates and associated transpiration rates were high. A similar pl anting of legumes conducted in the fall when temperatures and associated plant processes would have been lowered had a much higher survival rate (Kirkman, unpublished data). Based on this observation, transplantation should be conducted in the fall. 15N values indicated a high level of N2-fixation by most of these legume species in the open light environment. Differe nces between calculated percent Ndfa for the legumes (44-79%; Figure 4-5) were sli ghtly lower than those repor ted by Hendricks and Boring (1999) from an older populati on of similar species in pine woodlands (54-88%). Both studies indicate potentially high growth and N2-fixation rates under relatively open canopy conditions. However, many questions relating to ineffective nodulation and species survivorship highlight the need to examine other environmental and biotic controls on legume N2-fixation, especially drought st ress and rhizobium microsymbionts in older established field popul ations (Dreyfus et al.,1988). Additionally, further field studies should examine the continued development of maturing legumes as this study only followed th e first year of establishment. Further observations of remaining plan ted legumes from this study and from another continuing study under the 14 year-old longleaf pine plantati on indicate that some of the species that were excluded from this study did not die, but did not establish dur ing the first growing season after planting. Specifically Mimosa quadrivalvis Orbexillum lupinellus and

PAGE 94

84 Pediomelum canescens showed significant growth and maturation during the second year after planting (Kirkm an, unpublished data). The native legumes planted under the clos ed canopy of this 14 year-old longleaf pine plantation showed very l ittle capacity for growth, N2-fixation, or nitrogen accretion. Dense litter accumulation under this closed ca nopy can also heavily suppress the growth of other groundcover species, in cluding wiregrass or Rubus sp. (blackberry) (Mulligan and Kirkman, 2002). However, mulching with pine litter could be an important tool to selectively encourage establishment of specific groundcover species under more open canopy conditions. Growth and N2-fixation of the legumes increased dramatically with increased canopy opening, advancing among the closed (8.5 % of incident light), intermediate (61.4% of incident) and open ( 83.4% of incident) treatments. This pattern suggests that substantial legume populations, from the pers pectives of wildlife food production and cover, and N contribution to the soil, will not be likely to grow or persist beneath dense young pine plantations until the tr ees are operationally thinned to a lower density. Even the relatively low stocking of pines in these plantations intended for enhancing wildlife habitat results in dense canopy developmen t and little capacity for supporting native grasses and legumes. The findings from th is study points to a recommendation that landowners plant legumes beneath plantations that have undergone an initial thinning after about 15-20 years of tree growth. At th at point, light conditions should be more favorable for legume populations to grow and contribute a significant amount of fixed-N to the N and C depleted so ils that are typical on conve rted agricultural sites.

PAGE 95

85 CHAPTER 5 CONCLUSION Conclusions from the Current Study Introduction The fire-dependent longleaf pine( Pinus palustris Mill.) wiregrass ( Aristida stricta Michx.) savanna ecosystem once dominated th e southern coastal plain of the United States, covering as much as 37.2 million ha. (L anders et al., 1995). Recent restoration initiatives on private and public lands in the southeastern U.S. coastal plain have resulted in the planting of approximately 283,000ha of former agricultural pulpwood plantation and fire suppressed land back into longleaf pine stands th rough the USDA Conservation Reserve Program since 1996 (CRP; Coffey and Kirkman, 2004). Groundcover reestablishment in these stands is key to improving wildlife habitat and for restoring a continuity of pyrrhic fuels for frequent prescribed burning (Clewell, 1989; Kirkman, 2002). Legumes may also have a major role in maintaining N balance of these restored systems. Shade tolerance of groundcover spec ies to be potentially reintroduced under young pine stands needs to be assessed befo re large-scale operati ons are undertaken (Mulligan and Kirkman, 2002). Objectives The overall objectives of this study were (1) to explore the impact of various degrees of shading on relative growth and N2-fixation rates of legume species native to longleaf pine-wiregrass savanna s, (2) to make initial observations of phenological development and nodule morphology for each species, and (3) to examine the

PAGE 96

86 effectiveness of corroborati ve methods for assessing N2-fixation. These objectives have not been previously addressed for most of th ese species. Controlled potted studies and a common garden experiment were used to assess species responses to shading under potted and field conditions. The species examined in this study repres ented all three common subfamilies of the Leguminoseae Caesalpiniodeae Mimosoideae and Papilionoideae, and represented three distinct growth forms: vines, erect herbs, and semi-woody erect herbs. Nodule morphology differences further confirmed the species subfamily designations. Each species had a unique response to shade, and each one represents a slightly different life history strategy that is appare nt in growth habit and phenol ogical development (Chapter 2). Adaptations to fire, low-fertility soils, droughty conditions, and a highly variable, but relatively open overstory ca nopy structure are manifested in the stress-tolerating and opportunistic nature of most of thes e legume species (Hainds et al., 1999). N2 fixed by legumes, although important to many ecosystems, is difficult to quantify. Due to the specific limitations associated with each assessment technique, corroborative assessments should be made in a ny study that seeks to quantify fixed-N. In this study, relatively higher levels of nitr ogenase activity (acety lene reduction) for a species were corroborated by reduced soil NO3 utilization, 15N signatures nearer atmospheric values, and higher tissue-N concen trations (Chapter 3) Even in a potted study with complete nodule recovery, nodul e biomass was not a good predictor of N2fixing ability, since only one species had significantly diffe rent biomass from all of the others.

PAGE 97

87 N2-fixing capabilities were species dependent. Some of the species in the potted study did not appear to have fully developed during this short-term study. For example, Tephrosia virginiana which was indicated as one of the strongest N2-fixers in the common garden study (Chapter 4), did not re ach a reproductive stage, nor was it very active in N2-fixation in the potted study (Chapter 3). Additionally, Orbexillum lupinellus which appeared to be well-nodulated in the potted study, was not active in N2-fixation. This ineffective nodulation may indicate a slow ly-developing plant, or it may indicate that the preferred symbiont for this species was not present (Chapter 3). Field studies following more controlled st udies can confirm in itial findings and give more realistic estimates of legume N2-fixing capacities (Dreyf us et al., 1988). The common garden experiment, which empl oyed three levels of canopy opening, was designed to make such estimates of N2-fixation capabilities under field conditions. Shade not only impacted the biomass and tissue-N allocation patterns of the legumes in the study, but was effective in produc ing light levels that provid e critical limitations to growth and N2-fixation rates (Chapter 4). This study also showed that Lespedeza spp., Desmodium spp. and Tephrosia virginiana are potentially very good candidates for inclusion in groundcover restorat ion projects due to their hi gh tissue-N concentrations, N2-fixation capabilities, and dom inant growth forms that will provide adequate cover for wildlife. These findings should be consid ered when landowners make decisions about groundcover restoration. The fina l conclusion of this field st udy was that native legumes will not thrive or fix N2 under a closed canopy and should be established in more open stands that have been recently thinned.

PAGE 98

88 Directions for Future Research This study examined only a small percenta ge of the legume population native to the longleaf pine-wiregrass ecosystem. Additi onal species should be considered for groundcover restoration, especially other species of the genera Desmodium and Lespedeza Since the legumes predominantly repr esent a fall food source for bobwhite quail, perhaps other fall-flowering legumes s hould be considered for groundcover, such as Dalea spp. (Stoddard, 1931). Further Application of N2-Fixation Assessment Techniques Of the N2-fixation assessment techniques used in this study, only the N-transport and storage product analysis has not been used for field analys is of species in southern pine ecosystems. The portabili ty and minimally-destructive na ture of this technique may have great potential for field studies in th e longleaf pine-wiregra ss ecosystem, and may be especially valuable for a first-assessment of N2-fixing capabilities of numerous species of legumes under field conditions when used in corroboration with the 15N natural abundance technique. Another interesting appr oach to quantifying N2-fixation using the stem N transport/storage product analysis (relative ureide analysis) involves calibrating ureide indices with fixed-N2 using a series of labeled 15N2 assays and relative ureide analysis (Peoples et al., 1996). This method may have potential for allowing an estimate of N2fixed by a population of legumes with a quick an d easy analysis that is easy to conduct on a variety of soils. Further studies should also investigate th e relative efficiencies of specific hostsymbiont interractions, including tripartite interactions between mycorrhizae, rhizobium, and legumes. Relative efficiencies can be qu ite different between rhizobium strains, and

PAGE 99

89 inoculation with highly effective strains could become import ant to native legume restoration if the land on which they are be ing planted has had heavy herbicide usage. Future Research for Native Legume Utilization Future research regarding groundcover re storation under longleaf pine should involve studies that investig ate physiological properties of these and many other of the legumes native to the longleaf-wiregrass ecosy stem and continue to ask questions relating directly to groundcover restoration objectives such as soil-N and organic material development and wildlife food quality. Add itional studies should look at competitive interactions between native gr asses and legumes planted in a restoration setting to determine when competition for soil and light resources becomes limiting for each type of plant. Fine root turnover, mycorrizh al associations, root exudates, and other belowground processes that contribute dire ctly to building soil organic matter and Navailability should be further researched. Analysis of na tive legume and grass shoot and seed digestibility and nutri tional composition could be important information for land managers seeking to attract and sustai n wildlife populations on their property. Imported legumes have been used for so il stabilization and game management purposes, and many of these sp ecies, including kudzu ( Pueraria lobata ), sericea lespedeza ( Lespedeza cuneata ) and shrubby lespedeza ( L. bicolor ), out-compete native vegetation in natural and roadside areas ac ross the Southeast (Mil ler, 2003). Native legumes are a better choice for many soil pr eservation and land management purposes and are much less likely to become inva sive. Many of the legumes native to the southeast region, such as Centrosema virginianum and Dalea pinnata to name only a few, are also aesthetically pl easing and should be considered for wildflower plantings. The N2-fixation capability of C. virginianum could possibly be marketed as an added

PAGE 100

90 benefit to native plant collectors. More res earch is needed before other species can be designated as N2-fixers. Native legumes are important and poorly expl ored genetic resources. N is typically the most limiting plant nutrient in terrestrial ecosystems, and biological N2-fixation is a free source of N-addition. In a time wh en organic farming practices are gaining popularity in the temperate parts of the world an d agroforestry is increasing in feasibility and scientific recognition in tropical and subtropical parts of the world, legumes are becoming more valuable and marketable. Th e legume species-richness resources that are available in the Southeastern coastal plai n should be actively examined for future development.

PAGE 101

91 LIST OF REFERENCES Battaglia, M. A., R. J. Mitchell, P. P. Mou, and S. D. Pecot. 2003. Light transmittance estimates in a longleaf pine woodland. Forest Science 49: 752-762. Bazzaz, F. A., 1996. Plants in Changing Environments: Linking physiological, population, and community ecology. Cambri dge University Press, Cambridge, Great Britain. Becker, D. A., and J. J. Crockett. 1976. Nitrogen fixation in so me prairie legumes. The American Midland Naturalist 96: 133-143. Boring, L. R., and W. T. Swank. 1984. Symbiotic nitrogen fixation in regenerating black locust stands. Forest Science 30: 528-537. Boring, L. R., C. A. Wilson, J. J. Hendricks, and R. J. Mitchell. 2004. Season of burn and nutrient losses in a l ongleaf pine ecosystem. International Journal of Wildland Fire 13: 443-453. Cataldo, D. A., M. Haroon, L. E. Schrader and V. L. Youngs. 1975. Rapid colormetric determination of nitrate in plant tissue by nitration of salicylic acid. Communications in Soil Science and Plant Analysis 6: 71-80. Carter, M. C., and C. D. Foster. 2004. Pres cribed burning and productivity in southern pine forests: a review. Forest Ecology and Management 191: 93-109. Christensen, N. L. 1977. Fire and soil-plant nutrient relations in a pine-wiregrass savanna on the Coastal Plain of North Carolina. Oecologia 31: 27-44. Clark, G. T. 1971. The ecological life history of Tephrosia virginiana PhD dissertation, University of Arkansas. Clewell, A. F. 1989. Natura l history of wiregrass ( Aristida stricta Michx., Gramineae ). Natural Areas Journal 9: 223-233. Coffey, K. L., and L. K. Kirkman. 2004. Native ground cover partners: reestablishing native ground cover in longleaf pine savannas. Ecological Restoration 22: 288289. Dreyfus, B. L., H. G. Diem, and Y. R. Dommergues. 1988. Future directions for biological nitrogen fixation research. Plant and Soil 108: 191-199.

PAGE 102

92 Fernndez, M. E., J. E. Gyenge, and T. M. Schlichter. 2004. Shade acclimation in the forage grass Festuca pallescens : biomass allocation and foliage orientation. Agroforestry Systems 60: 159-166. Gadgil, R. 1971. The nutritional role of Lupinus arboreus in coastal sand dune forestry. I. The potential influence of undamaged lupin plants on nitrogen uptake by Pinus radiata Plant and Soil 34: 357-367. Gholz, H. L., R. F. Fisher, and W. L. Pr itchett. 1985. Nutrient dynamics in slash pine plantation ecosystems. Ecology 66: 647-659. Hainds, M. J., R. J. Mitchell, B. J. Pa lik, L. R. Boring, and D. H. Gjerstad. 1999. Distribution of native legumes ( Leguminoseae ) in frequently burned longleaf pine ( Pinaceae )-wiregrass ( Poaceae ) ecosystems. American Journal of Botany 86: 1606-1614. Halvorson, J. J., E. H. Franz, J. L. Smit h, and R. A. Black. 1992. Nitrogenase activity, nitrogen fixation, and nitrogen input s by lupines at Mount St. Helens. Ecology 73: 87-98. Hendricks, J. J., and L. R. Boring. 1992. Litter quality of native herbaceous legumes in a burned pine forest of the Georgia Piedmont. Canadian Journal of Forest Research 22: 2007-2010. Hendricks, J. J., and L. R. Boring. 1999. N2-fixation by native legumes in burned pine ecosystems of the southeastern United States. Forest Ecology and Management 113: 167-177. Hendricks, J. J., C. A. Wilson, and L. R. Boring 2002. Foliar litter position and decomposition in a fire-maintained l ongleaf pinewiregrass ecosystem. Canadian Journal of Forest Research 32: 928-941. Herridge, D. F. 1984. Effects of nitrate a nd plant development on the abundance of nitrogenous solutes in root-bleeding a nd vacuum-extracted exudates of soybean. Crop Science 25:173-179. Hiers, J. K., R. Wyatt, and R. J. Mitche ll. 2000. The effects of fire regime on legume reproduction in longleaf pine sava nnas: is season selective? Oecologia 125: 521530. Hiers, J. K., R. J. Mitchell, L. R. Bori ng, J. J. Hendricks, and R. Wyatt. 2003. Legumes native to longleaf pine sava nnas exhibit capacity for high N2-fixation rates and negligible impacts due to timing of fire. New Phytologist 157: 327-338. Hogberg, P., and M. Kvarnstrom. 1982. Nitrogen fixation by the woody legume Leucaena leucocephala in Tanzania. Plant and Soil 66: 21-28.

PAGE 103

93 Isley, D. 1990. Vascular Flora of the Southeastern United States: Leguminoseae (Fabaceae) Vol. 3, part 2. University of Nort h Carolina Press, Chapel Hill and London. Izaguirre-Mayoral, M. L., O. Carballo, S. Flores, M. Sicardi de Mallorca, and T. Oropeza. 1992. Quantitative analysis of the symbiotic N2-fixation, non-structural carbohydrates and chlorophyll content in si xteen native legume species collected in different savanna sites. Symbiosis 12: 293-312. Jacobs, M. J., and C. A. Schloeder. 2002. Fi re frequency and species associations in perennial grasslands of south-west Ethiopia. African Journal of Ecology 40: 1-9. Kirkman, L. K. 2002. A forest gap approach to restoring longleaf pine-wiregrass ecosystems (Florida and Georgia). Ecological Restoration 20: 50-51. Kirkman, L. K. K. L. Coffey, R. J. Mitchell, and E. B. Moser. 2004. Ground cover recovery patterns and life-history traits: implications for restoration obstacles and opportunities in a species-rich savanna. Journal of Ecology 92: 409-421. Knapp, A. K., J. M. Briggs, D. C. Hartnett, and S. L. Collins, eds. 1998. Grassland Dynamics: Long-term ecological research in tallgrass prairie. Oxford University Press, New York. Kramer, P. J., and T. T. Kozlowski. 1979. Physiology of Woody Plants. Academic Press, New York. Landers, J. L., D. H. Van Lear, and W. D. Boyer. 1995. The longleaf pine forests of the Southeast: requiem or renaissance? Journal of Forestry 93: 39-44. Markewitz, D., F. Sartori, and C. Craft. 2002. Soil change and carbon storage in longleaf pine stands planted on marg inal agricultural lands. Ecological Applications 12: 1276-1285. Medina, E., and M. L. Izaguirre. 2004. N2-fixation in tropical American savannas evaluated by the na tural abundance of 15N in plant tissues and soil organic matter. Tropical Ecology 45: 87-95. Miller, J. H. 2003. Nonnative Invasive Plants of Southern Forests: A field guide for identification and control. Southern Resear ch Station, U.S. Forest Service General Technical Report SRS62, Asheville, NC. Morgan, J. W. 1999. Defining grassland fire ev ents and the response of perennial plants to annual fire in temperate grassl ands of south-eastern Australia. Plant Ecology 144: 127.

PAGE 104

94 Mulligan, M. K., and L. K. Kirkman. 2002. Competition effects on wiregrass ( Aristida beyrichana ) growth and survival. Plant Ecology 167: 39-50. Ojima, D. S., W. J. Parton, D. S. Schime l, and C. E. Owensby. 1990. Simulated impacts of annual burning on prairie ecosystems. In S. L. Collins and L. L. Wallace [eds.], Fire in North American Ta llgrass Prairies, 118-132. University of Oklahoma Press, Norman, OK. Paz, H. 2003. Root/shoot allocation and root architecture in seedlings: variation among forest sites, microhabitats, and ecological groups. Biotropica 35: 318. Pecot, S. D., S. B. Horsley, M. A. Battag lia, and R. J. Mitchell. 2005. The influence of canopy, sky condition, and solar angle on light quality in a longleaf pine woodland. Canadian Journal of Forest Research 35: 1356-1366. Peoples, M. B., B. Palmer, D. M. Lill ey, L. Minh Duc, and D. F. Herridge. 1996. Application of 15N and xylem ureide methods for assessing N2 fixation of three shrub legumes periodically pruned for forage. Plant and Soil 182: 125-137. Prism. 1996. GraphPad Prism, release 2.01. Gr aphPad Software, Inc., San Diego, CA. SAS. 2003. SAS/STAT and base SAS, re lease 9.1.3. SAS Institute, Cary, NC. Sicardi de Mallorca, M., and M. L. Izagui rre-Mayoral 1993. A comp arative evaluation of the symbiotic N2-fixation and physiological perfor mance of thirty-six native legume species collected in a tropical savanna during the rainy and dry season. Symbiosis 16: 225-247. Sprent, J. I. 1973. Growth and nitrogen fixation in Lupinus arboreus as affected by shading and water supply. New Phytologist 72: 1005-1022. Sprent, J. I. 1987. The Ecology of the Nitroge n Cycle. Cambridge University Press, Cambridge, Great Britain. Sprent, J. I. 1999. Nitrogen fixation and gr owth of non-crop legume species in diverse environments. Perspectives in Plant Ecology Evolution and Systematics 2: 149162. Sprent, J. I. 2002. Nodulation as a taxonomic tool. In P. S. Herendeen and A. Bruneau, [eds.], Advances in Legume Systematics Part 9, 21-44. Royal Botanical Gardens, Kew, London. Stahl, L, G. Nyberg, P. Hogberg, and R. J. Buresh. 2002. Effects of planted tree fallows on soil nitrogen dynamics, above-ground and root biomass, N2-fixation and subsequent maize crop productivity in Kenya. Plant and Soil 243: 103-117.

PAGE 105

95 Stoddard, H. L. 1931. The Bobwhite Quail: Its habits, preservation and increase. Charles Scribners Sons, Quincy, FL. Virginia, R. A., W. M. Jarrell, P. W. Runde l, G. Shearer, and D. H. Kohl. 1989. The use of variation in the natural abundance of 15N to assess symbiotic nitrogen fixation by woody plants. In P. W. Rundel, J. R. Ehleringer, and K. A. Nagy [eds], Stable Isotopes in Ecological Research, 3 75-394. Springer-Verlag, New York. Vitousek, P. M., K. Cassman, C. Cleveland, T. Crews, C. B. Field, N. B. Grimm, R. W. Howarth, R. Marino, L. Martinelli, E. B. Rastetter, and J. I. Sprent. 2002. Towards and ecological understandi ng of biological nitrogen fixation. Biogeochemistry 57: 1-45. Wilson, C. A., R. J. Mitchell, J. J. He ndricks, and L. R. Boring. 1999. Patterns and controls of ecosystem function in longleaf pine wiregrass savannas. II. Nitrogen dynamics. Canadian Journal of Forest Research 29: 752-760. Wunderlin, R. P., and B. F. Hansen. 2003. Guid e to the Vascular Plants of Florida, 2nd ed. University Press of Florida, Gainesville. Yemm, E.W., and Cocking. 1955. The determination of amino acids with ninhydrin. Analyst 80: 209-213. Zitzer, S. F., and J. O. Dawson. 1989. Seasonal changes in nodular n itrogenase activity of Alnus glutinosa and Eleaganus angustifolia Tree Physiology 5: 189-194.

PAGE 106

96 BIOGRAPHICAL SKETCH Sarah Elizabeth (Wright) Cathey receive d a bachelors degr ee in biology from Lipscomb University, Nashville, TN, in 2001, graduating with honors designation. As an undergraduate student, she served as an intern for the TN Department of Environmental Conservation (TDEC), Division of Natural Heri tage, and assisted Dr. James Carpenter in a study of the endangered crayfish, Orconectes shoupi Cathey also completed a senior honors thesis entitled A n Eradication Study of Vinca minor , which was undertaken in cooperation with the TDEC Divi sion of Natural Heritage an d the Lipscomb University Biology Department. A strong interest in native plants and app lied research led Cathey to take a position as research assistant in the cooperative progr am between the University of Florida and The Joseph W. Jones Ecological Research Center (Newton, GA) that supported the research presented in this thesis. With regard to the future, Cathey has research interests in applied plant and soil sc ience, ecology and ecophysiology, and she has been awarded a fellowship by the College of Agricultural and Life Sciences at the University of Florida to pursue a PhD. Catheys world view has been strongly influenced by her travel to undeveloped and developing nations (Haiti, Ukraine, and Peru), and she intends to pursue research that could have positive environmental and socio-economical implications for people in the U.S. and abroad. A desire to teach and conduct research will probably lead Cathey to pursue a faculty position at a university in the future.

PAGE 107

97


Permanent Link: http://ufdc.ufl.edu/UFE0013280/00001

Material Information

Title: Growth and Nitrogen Fixation of Legumes Native to the Longleaf-Wiregrass Ecosystem
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013280:00001

Permanent Link: http://ufdc.ufl.edu/UFE0013280/00001

Material Information

Title: Growth and Nitrogen Fixation of Legumes Native to the Longleaf-Wiregrass Ecosystem
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013280:00001


This item has the following downloads:


Full Text












GROWTH AND N2-FIXATION OF LEGUMES NATIVE TO THE LONGLEAF-
WIREGRASS ECOSYSTEM
















By

SARAH ELIZABETH CATHEY


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Sarah Elizabeth Cathey





















This thesis is dedicated to my husband, partner and most devoted supporter, Marcus.















ACKNOWLEDGMENTS

Financial support for this research was provided through the cooperative graduate

research agreement between the Joseph W. Jones Ecological Research Center and the UF

College of Natural Resources and Environment. Technical, sample processing and field

support were provided by the staff of the Jones Center, especially Scott Taylor, Mary

Cobb, and Sarah Becker. A special thankyou goes to Dr. Kay Kirkman of the Jones

Center who provided all of the nursery plants for my experiments, and to Dr. Ken

Quesenberry who provided additional seeds.

I would like to thank my committee for their advice and revisions regarding my

research and the writing of this thesis. Special thanks go to Tom Sinclair and Lindsay

Boring for their assistance in revisions and for personal support. I would like to thank

my parents for their moral support, as well. I especially need to thank Susan Sorrell,

technician at the USDA ARS lab at the University of Florida, for her tireless efforts in

data collection and sample preparation, and for sharing her expertise in

spectrophotometry analysis.

Finally, I would like to thank my husband, Marcus, for his tireless dedication to my

graduate studies and attendant technical support and for his willingness to move all the

way to central Florida from Tennessee in order for me to pursue this degree.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ............... ............................... ........... ............ vii

LIST OF FIGURES ......... ...... ......................................... ............. viii

ABSTRACT .............. ......................................... ix

Chapter

1 IN T R O D U C T IO N ............................................................................. .............. ...

2 GROWTH AND PHENOLOGY OF NATIVE LEGUMES IN TWO LIGHT
E N V IR O N M E N T S ............................................................... .......... ...................... 7

Introduction..................................... ........................... ..... ..... ........ 7
M methods and M materials ................................................................. ........................ 9
P planting ...................................................................................................... ....... 9
M easurem ents ................................................................. ............13
D ata and Statistical A analysis ..................................................... ...... ......... 16
R e su lts .........................................................................................................................1 6
S u rv iv o rsh ip ................................................................16
M o rp h o lo g y ................................................................. ................................1 7
Phenology .............................. ...... ... ................................. .......18
Plant Responses to Light Environment ........................................ .....20
D isc u ssio n ............................................................................................................. 3 2
M orphology and Phenology ........................................................ 32
G ro w th P pattern s ............................................................................................. 3 3

3 USE OF CORROBORATIVE METHODS TO ASSESS THE N2-FIXATION OF
N A T IV E L E G U M E S ............................................................................................ 37

Introduction ........................................................................................................ 37
M eth od s an d M materials ......................................................................................... 39
Planting ................................................................ .... ..... ........ 39
N 2 F ix ation A ssessm ent.......................................................................................4 1
Statistical A naly sis ...........................................................43
R e su lts ...........................................................................................4 4









S u rv iv o rsh ip ................................................................4 4
Fixation Assessment............. ......... ........... ...................... 45
D iscu ssion ......... .. ....... ..... ............ ........................................54
Com prison of M ethodology.......... ................. ........................ ............... 54
Species D differences ........................ .. ....................... .... .. ........... 59
Su m m ary .................................................................................................. 62

4 GROWTH AND N2-FIXATION OF NATIVE LEGUMES IN LONGLEAF PINE
R E ST O R A TIO N ........... .................................................................. ................... 64

In tro d u ctio n .......................................................................................6 4
M materials and M methods ....................................................................... ..................65
Site Description ........................... ................ 65
Experimental Design and Planting .......................................... ...............66
Statistical A naly sis ........................................ .. .. .... ........... 70
R e su lts .................. ..... ............. ............. .................... ................7 0
Preliminary Results and Survivorship..... .......... ........................................70
G row th ......... ...................................................... .. ............. .. 7 1
N 2-F ix action .........................................................................7 8
Discussion ............. ................................................78
Shading Effects on Species ........................................... .......................... 78
Ecological and M management Implications .................................. ............... 83

5 C O N C L U SIO N ......... ...................................................................... ......... .. ..... .. 85

Conclusions from the Current Study ........................................ ....... ............... 85
In tro d u ctio n ................................................................................................... 8 5
O bj ectiv es ........... ... ............. ................................................................ 85
D directions for Future Research...................... .. .... ........... .............................88
Further Application of N2-Fixation Assessment Techniques ..............................88
Future Research for Native Legume Utilization .............................................89

LIST OF REFEREN CES ......... .. ................................... ......... ..................... 91

B IO G R A PH IC A L SK E TCH ..................................................................... ..................96
















LIST OF TABLES


Table page

2-1. Description of Native Legumes used in study. ............................................. 10

2-2. Fruit and nodule descriptions. .......................................................................... 11

2-3. Regression equations for stem elongation curves of the form y= Dx3 + Cx2 + Bx
+ A .......................................................................22

2-4. Values given are slopes calculated from the derivative of the equations given in
2-3 at the m ean for each coefficient. ........................................ ...... ............... 23

2-5. Maximum plant heights by species, regardless of treatment. ...............................25

3-1. Nodule mass and number of nodules. ......................................... ............... 47

3-2. N-transport/storage products extracted from stem sections. ..................................50

3-3. Specific nodule activities of species in this study and other comparative reports.
....................................................................................................... . 5 5

4-1. Analysis of variance results for experimental variables. .....................................73

4-2. Total biomass (above- and belowground tissues, including nodules) per plant
and aboveground values for %N, 615N, and total N..............................................74
















LIST OF FIGURES


Figure page

2-1. Weather for Gainesville, FL, 8 February to 20 November 2004 .........................14

2-2. Phenological change by species. ......................................................................19

2-3. Stem elongation. ................................. ..... ......... .. .............21

2-4. Leaflet counts and plant widths of sun and shade grown plants. A) Leaflet
counts for Clitoria mariana, Tephrosia virginiana and Lespedeza hirta plants
grown in sun and shade. B) Width of Crotalaria rotundifolia plants grown in
sun and shade. ........................................................................26

2-6. Aboveground and belowground harvested biomass. ...........................................30

2-7. Root to shoot ratio of harvested biomass. ................................... .................31

3-1. Ethylene production trends for the growing season by species. .......................... 48

3-2. Maximum ethylene production (C2H4 reduction) peaks. .....................................49

3-3. 15N values by species. ................................................ .............................. 51

3-4. Mean % Ndfa, %N, and total N by species. A) Percent of total N derived from
the atmosphere. B) Percent N in aboveground tissues. C) Total N content of
aboveground tissues. ............................ ........... ...... ...... ...... ...... 52

4-1. Volumetric soil moisture patterns for all plots. ............................................. 68

4-2. Aboveground biomass and change in plant heights from To by species in each of
the three light treatm ents. ................................. .......................................72

4-3. Root-to-shoot ratios by species in the three light treatments. ...............................76

4-4. Percent N concentration in aboveground biomass (stem + leaves) by species in
the intermediate and open light environments. ............................................. 77

4-5. 615N and %Ndfa values by species for aboveground tissues in the intermediate
and open light treatm ents. ............................................................................. 79















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

GROWTH AND N2-FIXATION OF LEGUMES NATIVE TO THE LONGLEAF-
WIREGRASS ECOSYSTEM

By

Sarah Elizabeth Cathey

December 2005

Chair: Thomas R. Sinclair
Major Department: Agronomy

The longleaf pine- (Pinuspalustris Mill.) wiregrass (Aristida strict, Michx.)

savanna ecosystem once dominated the southern coastal plain of the United States, but

presently less than 1.5 million of the historic 37.2 million ha remain intact. Restoration

plantings reclaiming more than 283,000 ha of former agricultural fields, pulpwood

plantations, and other fire-suppressed lands are being established in the Southeast.

Groundcover reestablishment of native legumes and grasses is the key to restoring soil

nitrogen levels, wildlife habitat and continuous fuels for frequent prescribed burning in

young longleaf pine plantings. More information is needed about the growth and N2-

fixation rates of native legumes under shade in order for informed selections to be made

for groundcover restoration plantings.

A potted plant study was used to assess growth and N2-fixation responses of 10

species under two light regimes. Total biomass accumulation and root-to-shoot ratios

were used to examine growth responses to shading, and several corroborative techniques









including analysis of nodule biomass, 815N natural abundance, percent N derived from

the atmosphere, transport product analysis, and the acetylene reduction assay were used

to assess potential N2-fixing capabilities. Overall, N2-fixation rates increased throughout

the season. Relative N2-fixation rates, as assessed by the above approaches, indicated

that Mimosa quadrivalvis, Crotalaria rotundifolia and Centrosema virginianum were

quickly developing species and effective N2-fixers, and that Lespedeza hirta and

Orbexillum lupinellus showed lower N2-fixation rates in a one-year study. Shade did not

have a significant effect on N2-fixation in this controlled study. The acetylene reduction

assay is best used as a check for nitrogenase activity and for following seasonal patterns,

but is not as useful for a quantitative estimate of N2-fixed. The nitrogen transport product

analysis may have limited usefulness for these species due to extremely low nitrate

levels, but should also be tested as a field technique. The 615N natural abundance method

was a useful technique for estimating N2-fixation inputs over time.

A garden plot study, situated in a 14 year-old longleaf pine plantation on an old

agricultural site, was used to assess growth and N2-fixation of eight species of native

legumes under three levels of canopy openness. Growth and N2-fixation declined rapidly

between approximately 60 and 80 percent canopy openness, indicating that legumes

native to the longleaf pine-wiregrass ecosystem have a limited degree of shade tolerance.

Root-to-shoot ratios also indicated that belowground growth was dominant among plants

growing under shaded conditions. Groundcover restoration plantings involving native

legumes will be most effective when conducted after an initial thinning after about 15-20

years of tree growth














CHAPTER 1
INTRODUCTION

The longleaf pine- (Pinuspalustris Mill.) wiregrass (Aristida strict, Michx.)

savanna ecosystem dominated the coastal plain region of the southeastern United States

in the pre-European era, covering as much as 37.2 million ha, but less than 1.2 million ha

remain intact. Private hunting lands and planted stands constitute most of the longleaf

pine coverage that still exists in the U.S. (Landers et al., 1995). Over the past 100 years,

land managers have used frequent prescribed burning to maintain the longleaf pine-

wiregrass ecosystem for wildlife habitat (Boring et al., 2004).

Although frequent fire disturbance is necessary to maintaining the structure of the

longleaf-wiregrass ecosystem by suppressing the oak midstory, some have hypothesized

that nutrient losses due to the consumption of litter and volatilization may cause a

continual decline of overall N in the system (Carter and Foster, 2004). Nitrogen and

phosphorus are potentially co-limiting to net primary productivity in these woodlands,

secondary only to water limitations (Hendricks et al., 2002). Recent studies have shown

that periodic mineralization of phosphorus, which occurs when large amounts of

accumulated needle litter is burned, may serve an important role in the nutrition, growth

and reproduction of phosphorus-demanding legumes, leading to the eventual replacement

of nitrogen lost during burning (Boring et al., 2004).

Native herbaceous legumes constitute more than 10 percent of the vascular plants

in frequently-burned longleaf pine savannas (Hainds et al., 1999). Due to their high

numbers and density as well as their quick regeneration following fire, the potential for









significant input of biologically-fixed nitrogen is great. Life history adaptations such as

high belowground biomass allocation, opportunistic flowering and N2-fixation (in

legumes) are important to the success of species native to frequently burned ecosystems

(Knapp et al., 1998; Morgan, 1999; Hiers et al., 2000; Jacobs and Schloeder, 2002).

Legumes have been considered to be a significant component of the N-cycles of

many ecosystems, but the actual quantification ofN from biological N2-fixation is

difficult to estimate. Using a series of field-conducted population surveys and acetylene

reduction assays (and corroborative 615N data), Hendricks and Boring (1999)

conservatively estimated legume nitrogen inputs from biological fixation in the longleaf-

wiregrass system to range from 7 to 9 kg N ha1 yr1. Other studies have used this

technique to estimate potential contributions of N2 fixed by legumes in other fire affected

ecosystems including prairies (Becker and Crockett, 1976). However, N-input estimates

made from the acetylene reduction assay are limited by the ability to recover nodules in

the field. Recent studies seeking to quantify biological N2-fixation in natural ecosystems

and agroforestry systems often rely more heavily on isotopic methods, such as 15N

enrichment or the natural abundance of 15N (Peoples et al., 1996; Hendricks and Boring,

1999; Medina and Izaguirre, 2004). A comparison of studies using these diverse tools,

and an understanding of the limitations inherent in each of the assessment techniques

reinforce the need to use corroborative methods when assessing N2-fixation in legumes.

Because of the high energy requirements necessary to maintain nitrogen fixation,

legumes outside of the tropics have generally been assumed to be shade intolerant

(Sprent, 1999). Sprent (1973) found that Lupinus arboreus plants grown under shade

conditions produced less nodule biomass and subsequently lower nitrogenase activity









(acetylene reduction). Although the longleaf-wiregrass ecosystem has a relatively open

canopy compared to other forest systems, the numerous legumes in this ecosystem must

be able to tolerate a light environment that is variable but averages approximately 40

percent of full sunlight (Battaglia et al., 2003; Pecot et al., 2005). Light levels in young,

planted longleaf pine stands may be substantially higher than the native woodland, but

canopy closure occurs only a few years after planting. Thus some of the more shade

tolerant native legume species may be more satisfactorily adapted for restoration

plantings.

Reforestation initiatives on public and private lands since 1998 have resulted in the

planting of approximately 700,000 acres of former agricultural, pulpwood plantation and

fire suppressed land back into longleaf pine stands through the USDA Conservation

Reserve Program (CRP) and other independent landowner efforts. Due to the long

rotation period typical of longleaf pine stands they have the potential to generate income

for private landowners over many years through timber revenue, hunting leases and

government subsidies paid by the CRP (Landers et al., 1995). Groundcover restoration in

young longleaf pine stands planted on depleted, former agricultural soil is important for

rebuilding soil organic matter and N (Markewitz et al., 2002), for providing wildlife food

and cover (Stoddard, 1931), and for enhancing pyrrhic fuel continuity necessary to

reintroduce frequent prescribed fires (Mulligan and Kirkman, 2002). Reintroduction of

native legumes for groundcover restoration rather than exotic or agricultural species

should prevent problems of invasiveness with non-native legume species in the past, such

as serecia lespedeza (Lespedeza cuneata) and L. bicolor (Miller, 2003), and agronomic

species' lack of adaptability to woodland environments.









The overall objectives of this study were (1) to explore the impact of various

degrees of shading on relative growth and N2-fixation rates of legume species native to

longleaf pine-wiregrass savannas, (2) to make initial observations of phenological

development and nodule morphology for each species, and (3) to examine the

effectiveness of corroborative methods for assessing N2-fixation. These objectives have

not been previously addressed for most of these species. Controlled potted studies and a

common garden experiment were used to assess species responses to shading under

potted and field growing conditions.

Chapter 2 examines the affects of shade on growth of nine species of native

legumes and provides descriptions of phenological development and nodule morphology

after one growing season in a pot study. Descriptive information regarding the species in

this study, Chamaecrista nictitans (L.) Moench, Centrosema virginianum (L.) Benth.,

Clitoria mariana L., Crotalaria rotundifolia J.F. Gmel., Lespedeza hirta (L.) Hornem.,

Mimosa quadrivalvis (L.), Orbexillum lupinellus (Michx.) Isley, Rhynchosia reniformis

D.C., and Tephrosia virginiana (L.) Pers, is very limited. Phenological development data

is helpful for better understanding the life history of a species, especially in regard to fire

adaptation. Nodule morphology can be an important taxonomic tool for the

Leguminoseae (Sprent 2002), and relative nodule biomass is often indicative of relative

N2-fixation capacity. In addition, biomass accumulation and allocation (root-to-shoot

ratios) patterns are indicative of shade tolerance and environmental adaptability.

Chapter 3 describes a companion study to Chapter 2 that examines the N2-fixation

patterns and capabilities of the same nine species of native legumes during a growing

season, using five corroborative methods of assessment: nodule biomass, the acetylene









reduction assay, nitrogen transport and storage product analysis, 15N natural abundance,

and total N content. Each of these methods has been independently used to assess

relative N2-fixation rates among species in both controlled and field studies, and to

determine the affect of shading on N2-fixation rates. However, this study is the first

instance in which the nitrogen transport product analysis has been used to assess these

species. In this study, the relative effectiveness of each corroborative assessment was

determined for use in a small, controlled study.

Chapter 4 further examines the growth and N2-fixation capabilities of eight species of

native legumes planted under three canopy opening conditions in a 14 year-old longleaf

pine plantation. Because of their dominant growth form and importance for wildlife cover

and food, more semi-woody species were included in this field study than in the potted

studies. The species examined were Centrosema virginianum, Desmodium ciliare (Muhl.

ex Willd.) DC., Lespedeza angustifolia (Pursh.) Ell., Lespedeza hirta, Mimosa

quadrivalvis, Orbexillum lupinellus, Pediomelum canescens (Michx.) Rydb., and Tephrosia

virginiana. As in Chapter 2, biomass accumulation and root-to-shoot ratios were used to

assess growth responses to shade in each of the species. The 615N natural abundance

method was used to compare relative N2-fixation rates among species and across light

environments.

Together, these studies should provide a greater understanding of the growth habits,

morphology, phenological development, N2-fixation capabilities and shade tolerance

characteristics of several species of native legumes. This information can be applied to

further studies of these and similar species and could be used to make preliminary






6


decisions about which species should be used for groundcover restoration plantings in

longleaf pine stands.














CHAPTER 2
GROWTH AND PHENOLOGY OF NATIVE LEGUMES IN TWO LIGHT
ENVIRONMENTS

Introduction

N2-fixing legumes are generally considered to be shade intolerant species -due to

the high energetic cost of nodule production, maintenance, and N fixation (Vitousek et

al., 2002). Among temperate ecosystems, native legumes are most diverse and abundant

in grasslands (Becker and Crockett, 1976) and savanna ecosystems (Hainds et al.,1999).

The variably-shaded and frequently burned environment of the longleaf pine- (Pinus

palustris Mill.) wiregrass (Aristida strict, Michx.) ecosystem supports over forty species

of native herbaceous legumes. These legumes constitute more than 10 percent of the

vascular species in these pine savannas and occur in high densities across a great range of

site conditions (Hainds et al., 1999). They demonstrate fire tolerance, adaptability to

infertile and drought soils, values for wildlife food resources, and may fix varying

amounts ofN (Hainds et al., 1999; Hendricks and Boring, 1999; Hiers et al., 2003).

Hiers et al. (2000) showed that flowering of native legumes in this system is tied to

occurrence of fire, but responses to a specific seasonal burn varied for Tephrosia virginiana

(L., Pers.), Centrosema virginianum (L., Benth.) and Rhyncosia reniformis (D.C.). These

legumes are found ubiquitously across the longleaf-wiregrass savanna landscape with only

the most extreme deep sands or seasonally-inundated lowlands having a lowered

abundance and diversity of species (Hainds et al., 1999). In addition to surviving drought

conditions and frequent fire, legumes are able to overcome the problem of very low-









fertility soils through N2-fixation (Hiers et al., 2003). However, shade tolerance of these

species has not been explored. Scientific knowledge of the biology ofN2-fixing species is

limited to a few ecosystems, and environmental tolerances have not been explored for most

of these species (Vitousek et al., 2002).

Since 1998, reforestation initiatives on public and private lands in the southeastern

U.S. have resulted in the planting of approximately 283,000 ha of former agricultural

fields, harvested pulpwood plantations and otherwise fire suppressed land back into

longleaf pine stands, with 48,000 ha of marginal coastal plain farmland in Georgia alone

under the USDA Conservation Reserve Program (Coffey and Kirkman, 2004). Many of

these sites are characterized by coarse sandy soils that are prone to drought and may be

highly depleted in C and N from prior agricultural production (Markewitz et al., 2002).

Groundcover restoration was proposed to be vital in the recovery of soil organic matter

and N-availability. N2-fixing legumes could be especially valuable. Although longleaf

pine overstory has been successfully established, there is a great need to better determine

the compatibility of groundcover species to a range of canopy light conditions so that

recommendations may be made to integrate suitable species into habitat restoration

projects.

This potted-plant study was designed to observe the effects of two light conditions

on growth responses of ten species of legumes native to the longleaf-wiregrass ecosystem

over the course of a single growing season. The specific objectives of the research

reported here were: (1) to document the influence of shading on growth habits and

biomass accumulation; and (2) to make initial observations of phenological development

and nodule morphology for each species. Various observations of root, shoot and nodule









growth were all used as indicators of growth response to light. A companion study also

measured N2-fixation responses of these species and results are reported in Chapter 3.

Methods and Materials

Planting

Young plants of ten species of native legumes (Table 2-1) were grown outdoors in

Gainesville, Florida (82 20' W, 29 38' N) between April and November of 2004. One-

half of the plants from each species was grown in the sun and one-half under a shade

cloth that excluded approximately one-half ambient light. Difference in

photosynthetically active radiation (PAR) between light treatments was determined using

a Li-Cor Quantum Sensor, LI-185A. Measurements were taken on a clear day, 19 March

2004, at approximately 13:00 Eastern Standard Time. Three readings each were taken

under the shade cloth and outside adjacent to the potted plants. PAR in the shaded area

(753 82 [mol s-m-2) was 56 percent of full sun (1340 51 [mol s-m-2).

The seedlings were initially propagated by Dr. L. Katherine Kirkman at the Joseph

W. Jones Ecological Research Center (JWJERC) from seeds collected from throughout

the native woodland on the 12,500 ha Ichauway reserve, located in Baker County,

Georgia, USA (31l19'N and 8020'W). Seeds were scarified by physical abrasion and

then germinated in a soil mix consisting of 8 parts Fafard 3B soil mix, 2 parts peat

(sphagnum), 2 parts sand, and 1 part perlite, contained in plug flats. The seeds were

sown June/July 2003 and kept in a greenhouse over the winter.












Table 2-1. Description of Native Legumes used in study. Nomenclature follows Wunderlin and Hansen (2003). Descriptions adapted


from Isley (199C
Species
Centrosema virginianum
(L.) Benth.
Clitoria mariana (L.)
Chamaecrista nictitans
(L.) Moench
Crotalaria rotundifolia
J.F. Gmel.
Lespedeza hirta
(L.) Hornem.
Mimosa quadrivalvis (L.)


Orbexillum lupinellus
(Michx.) Isley
Pediomelum canescens
(Michx.) Rydb.


Rhynchosia reniformis
D.C.
Tephrosia virginiana
(L.) Pers.


Code Common Name


CEVI

CLMA
CANI

CRRO

LEHI

MIQU


Spurred
Butterfly Pea
Butterfly Pea
Sensitive Pea
(Partridge Pea)
Rabbitbells

Hairy
Lespedeza
Sensitive Briar


ORLU Piedmont
Leatherroot
Buckroot


RHRE

TEVI


Dollarleaf

Goats Rue


Subfamily
Papilionoideae

Papilionoideae
Caesalpiniodeae

Papilionoideae

Papilionoideae

Mimosoideae


Papilionoideae

Papilionoideae



Papilionoideae

Papilionoideae


Category
Vining/Spreading

Vining/Spreading
Erect herb

Vining/Spreading

Erect herb

Vining/Spreading


Erect herb

Erect herb



Erect herb

Erect herb


Growth Habit
Perennial

Perennial
Annual

Perennial;
prostrate
Perennial;
semi-woody
Perennial;
trailing,
thorned stems
Perennial;
rhizomatous
Perennial;
diffusely
branched to
bushy
Perennial;
rhizomatous
Perennial;
brancing from
a central point


Mature Size
Stems 1-1.5m

Stems 30-100cm
Plant 15-60cm

Stems 1-7cm

Plant 0.8-1.5m

Stems 1-2m


Stems 20-60cm

Plant up to lm



Plant 7-15cm

Plant 30-60cm












Table 2-2. Fruit and nodule descriptions. Nodule sizes indicate an average, and are represented as follows: --, not nodulated; +,
6mm to 10mm. The number of plants examined to estimate nodule size is
also given.


Species

Centrosema virginianum

Clitoria mariana

Chamaecrista nictitans

Crotalaria rotundifolia

Lespedeza hirta

Mimosa quadrivalvis

Orbexillum lupinellus

Pediomelum canescens

Rhynchosia reniformis

Tephrosia virginiana


Fruit


Legume; Linear, 7-12cm x 3-4 mm, dehiscent

Legume; Oblong, 3-5 cm x 5-7 mm, seeds
sticky
Legume; Oblong, flat 2-4 cm x 4-5 cm

Legume; Ellipsoid, inflated, 1.5-2.5 cm x 7-
12 mm
Legume; 5-8 mm long

Legume; Oblong to linear, 3-5cm long,
prickled
Legume; Obliquely transverse-ridged

Legume; 8-11mm long

Legume; Oblong or elliptically-oblong, 1.2-
1.8 cm x 6-7 mm
Legume; Oblong, flat, 3-5 cm x 4 mm


Shape

Spherical


Nodules

Size: Sun

+++ (n=2)


Spherical +++ (n=2)


Spherical

Coralloid

Spherical

Coralloid

Spherical

Spherical

Spherical


-- (n=
++++ (n:

+++ (n=

++++ (n:


+++ (n=3)

n/a

-- (n=l)


Elongated ++++ (n=4)


Size: Shade

++ (n=5)

+++ (n=5)

++++ (n=3)

+++ (n=8)

++++ (n=6)

++++ (n=2)

++ (n=2)

n/a

++ (n=2)

++++ (n=7)









The small, 7 month-old seedlings from the JWJERC were transferred into oneo f

the following: clear acrylic rooting tubes (90cm long x 3.5cm diameter), cylindrical

polyvinyl chloride (PVC) pots, 35cm x 10.5cm diameter, or black plastic tree seedling

pots, 35cm tall x 80cm2 (Stuewe and Sons, Corvallis, OR). PVC pots were also used to

assess N2-fixation using the acetylene reduction assay (Chapter 3). A total of eight

rooting tubes, six PVC pots, and 14 black plastic pots were planted for each species in the

experiment. One half of each set of pots was grown under the shade treatment. For this

experiment, plants in the PVC and black plastic pots were measured and harvested as a

single group since the volume was approximately the same. Transplanting was

completed on 11 February 2004 (Day 42). Dates are represented in figures and tables as

numbered days beginning with 1 January 2004 as Day 1.

Seeds for the annual Chamaecrista nictitans (CANI) were obtained from Dr. Ken

Quesenberry (Agronomy Department, University of Florida). The seeds were collected

from along the roadside in Gainesville, FL. Seeds for (CANI) were scarified with sand

paper and then germinated on moist filter paper. Emerged CANI seedlings were planted

directly into PVC and black plastic pots on 21 April (Day 122).

Plants were inoculated by introducing native soil to each pot. Two topsoils (0-

20cm) were collected at the Jones Center, from a fine-loamy, kaolinitic, thermic Typic

Kandiudult (Orangeburg Series), and from a loamy, kaolinitic, thermic Arenic

Kandiudult (Wagram Series). Collections were taken from areas with thriving and

diverse legume populations. Soils were transported to Gainesville, FL where they were

stored in a cool, dark room and covered with plastic to maintain moisture. A mixture of

equal parts of each native soil type was used as the inoculation source. The pots had been









previously filled with commercial topsoil (Walmart Corp.) to within 6 cm from the top,

then 2 cm of native soil was added to the surface of all pots. Finally, the plant was placed

in the pot, and additional topsoil was used to cover the roots as needed. For the rooting

tubes, the 2 cm of native soil was added below where the seedling was pressed into the

top of the tube. Seedlings in the rooting tubes were inserted so that there was

approximately 2-3 cm in the tube above the soil to facilitate watering.

Water was applied by drip-irrigation every 12 hours (5:00 and 17:00 EST) using a

battery-operated timer (Rainbird), but was adjusted as needed throughout the experiment

to prevent excessive watering during periods of heavy rainfall.

Aphids were detected on CRRO, beginning around 3 April (Day 94), but the plants

were not detrimentally affected by the infestation. Pediomelum canescens began to

yellow and to develop brown leaf spots as early as 30 June (Day 182), followed by rapid

leaf loss and, consequently, this species was omitted from our results.

Gainesville experienced hurricane activity around 14 August (Day 227) and 3

September (Day 247; Figure 2-1). Some plants experienced leaf loss due to wind, and

LEHI, which was moved indoors, experienced some water stress. However, the reason

for loss of leaves at harvest was difficult to distinguish, because senescence had begun by

that time for most species.

Measurements

Height measurements were taken weekly of plants in both pot types and the rooting

tubes. Height was determined by measuring the distance from the soil surface to the

highest point on the plant. The height ofvining plants such as MIQU, CEVI, and CLMA

was considered to be the length of the longest stem. Each of these vines was measured

until the date when the branches became too tangled for a measurement to be plausible.











_,na i i i i An


U0 I I
35 65 95 125 155 185 215 245 275 305

Calendar Day


-1
-I
-30 3


-20 -

b
-10 0


--- Temp. Min
- Temp. Max
Irradiance


25 155 185 215 ;
Calendar Day


Figure 2-1. Weather for Gainesville, FL, 8 February to 20 November 2004. Temperature
and irradiance values are weekly averages. Rainfall values are daily totals.
Data from FAWN (2005).


/'-' a'

S /' *
I S
0 14









Root elongation was measured along the side of the clear acrylic tubes from the top

of the tube to the tip of the longest visible root using a measuring tape. Each tube was

placed in a white PVC sleeve. Root depth measurements and notations regarding nodule

presence were taken twice weekly until the roots reached the bottom of the tube or until

there was no increase in rooting depth for at least three readings.

Leaf addition was measured by counting the number of leaflets on each plant on a

weekly basis. This method was continued throughout the experiment for CLMA, LEHI,

RHRE, and TEVI. However, due to the large number of leaflets or the indistinguishable

nature of the leaflets, CEVI, MIQU, CRRO and ORLU leaf addition was instead

determined by measuring the width of the plant. The width of CEVI and MIQU was

determined to be the sum of the lengths of the two longest stems. Widths of CRRO and

ORLU were determined as the width of the plant at the widest point, leaf tip to leaf tip.

Width measurements were taken using a measuring tape or ruler.

The presence of flowers and fruits was noted along with the height measures. A

phenological phase was considered to be initiated when half of the specimens for each

species had expressed the particular characteristic such as presence of flowers or fruit, or

the absence of flowers while fruit remained.

At the conclusion of the experiment, all plants were destructively harvested.

Aboveground material was collected from specimens grown in root tubes. Both above-

and belowground biomass was collected from all specimens grown in pots. Roots were

washed free of soil and nodules were collected. Nodules were counted and then

individually measured by laying each one alongside a millimeter ruler. The diameter of

spherical nodules and the longest dimension of elongated nodules was measured in order









to estimate an average nodule size for each plant. All tissues were dried to constant

weight at 800C.

Data and Statistical Analysis

Data were analyzed using analysis of variance (ANOVA) with species and light

environment as main effects. If differences existed (p<0.05), Duncan's multiple

comparison post-test was used to determine which means differed significantly. The

GLM procedure performed in the Statistical Analysis System (SAS, 2003) was used for

ANOVA and post-tests. Patterns of stem elongation were analyzed using a non-linear

regression model (third-order polynomial), followed by an analysis of slope using the

derivative (dh/dt). The slope of the curve at selected points was calculated and compared

using ANOVA to test for species and treatment effects. Student's T test (a=0.05) was

used to test for significant differences in plant height, number of leaves and plant width

response to light treatments within a species. Non-linear regression and t-tests were

performed using Prism (Prism, 1996).

Results

Survivorship

Out of the 20 individual plants of each species that were planted in the black plastic

and PVC pots, an average of 7 survived. LEHI had the highest survival with 13

remaining, and ORLU and RHRE had the lowest survivorship with only 4 and 3 plants

remaining, respectively. The cause of mortality for most of these young seedlings is

unknown. Most of the species were still dormant when transplanted, and many of the

individuals never reemerged. Some of the MIQU plants were lost due to desiccation

during May during a dry and windy period when conductance and evaporative demand

must have exceeded the watering rate for this large species.









Survivorship in the rooting tubes was also low, with only an average of 3 plants

surviving out of the 8 transplanted per species. Lack of reemergence, inadequate space

for watering in the top of the tubes and soil compaction (leading to inundation of roots)

were the reasons for the small success of these individuals. Despite low survivorship

overall, 6 of the 10 species generally had enough survivorship to allow for treatment

analysis, excluding Pediomelum canescens, MIQU, ORLU and RHRE.

Morphology

Although they are quite different in structure and life history, the ten species in this

experiment can be categorized into two general morphological types, a vining/spreading

herb or an erect herb (Table 2-1). The vining/spreading species, including CEVI, CLMA,

CRRO and MIQU, although quite different from each other, all tend to be either prostrate

or to climb over adjacent plants. In contrast, the erect herbs are generally upright, but

LEHI or ORLU may droop over other plants or onto the ground. CANI, RHRE and

TEVI branch from a central stem or root crown, and the branching stems may be nearly

horizontal to the ground.

The legume fruits of these species ranged in size from 1.2 to 12 cm long (Table 2-

2). CEVI, CLMA, CRRO, and RHRE pods are dehiscent, but the dense, bract-like,

tightly-fitting pods of LEHI are persistent (Kirkman et al., 2004). Dispersion strategies

represented besides dehiscing include pods that stick to clothing or fur, such as LEHI,

individual seeds that are coated with an adhesive material, such as CLMA, and pods that

simply fall to the ground, such as MIQU and TEVI. CRRO legumes are inflated, and the

seeds will rattle inside of the pod when dry.

Inoculation with a mixture of native soils appeared to be effective due to nodule

development in all species. Nodule morphology differs among species, as well. CEVI,









CLMA, CANI, LEHI, ORLU and RHRE all have similarly shaped, spherical nodules.

These nodules vary from 1mm to nearly 10mm in diameter (Table 2-2). CRRO and

MIQU have coralloid nodules from >2mm to <10mm in length. Coralloid nodules have a

central branching point from which they randomly bifurcate, with occasional twice-

bifurcation. The shape of this type of nodule is highly irregular, and a definitive size is

difficult to estimate. TEVI nodules are elongate, cylindrical and often bifurcated. The

length of TEVI nodules average >6mm to 10mm, although largest nodules are >10-mm

long.

Phenology

Shade-grown plants did not demonstrate significant phenological delay in

comparison to plants grown in full sun. However, weekly observations of phenological

change did not allow high resolution for determination of treatment effects on flowering

and fruit initiation. Data for the two light treatments within a species were combined,

since there were no significant differences. Dates given are the days on which one half of

the specimens for each species had begun to express the given phenological change, such

as flowering.

MIQU was the first of the study species to flower (28 May, Day 149; Figure 2-2)

and to produce fruit (30 June, Day 182), followed by CRRO, CLMA, CEVI, and LEHI,

respectively. Flowering and fruiting continued throughout the season for MIQU, CRRO

and CEVI. The lack of a specific time for flowering and fruiting to occur indicates that

these species may be opportunistic in their fruit production. Both flowers and fruits were

present on MIQU, CRRO and CEVI from the first fruiting until the end of the

experiment. In contrast, CLMA, which stopped flowering around 14 September (Day

258), had a much more determinant flowering and fruiting pattern. Flowers and fruits










I I I I I Ia


* Flowering
SFlowers and Fruits
. Fruits only


140 I I -
CEVI CLMACRRO LEHI MIQU ORLU TEVI
Species



Figure 2-2. Phenological change by species. Bars represent the time at which over one-
half of specimens for a species had reached the prescribed phenological phase,
such as the onset of flowering. TEVI did not flower during the experiment.


300-
>280-
o 260-
S240-
5 220-
S200-
I8O-
180-
160-


I |









were only concurrently present for approximately 30 days on CLMA before fruit

production ceased and the plant returned to a vegetative state.

All species examined in this study were late spring and summer-flowering and

fruiting except for LEHI, which is considered to be fall flowering and did not begin to

flower until around 24 September (Day 268). TEVI did not flower during this

experiment. ORLU flowered for the shortest duration of any of the species in this study,

only 20 days. Although ORLU flowered, fruits were neither detected nor collected.

Plant Responses to Light Environment

Overall, the height patterns of all species showed rapid increase in stem elongation

toward the beginning of the season, with the exception of CRRO, and LEHI. Growth

curves were fitted to a third order polynomial (Figure 2-3; Table 2-3), but CEVI and

MIQU had a poor fit for this equation (r2<0.300) in a particular treatment, making them

difficult to analyze. Stem elongation rates for all species were greatest around 14 May

(Day 135), followed by decline as the season progressed, with the exception of LEHI,

which had its strongest elongation rate during the middle portion of the season, between

29 June and 19 August (Day 181 and 232) (Table 2-4). CANI, CRRO and LEHI

underwent a slight decline by the final measurement before harvest, 22 October (Day

296).

Stem elongation rates of CRRO, LEHI, ORLU and TEVI diverged according to

treatment on 19 August and 22 October (Day 232 and 296 (Table 2-4). Height was also

statistically different (unpaired t test, p< 0.05) at dates within the range that slope

diverged for CLMA, CRRO, and LEHI (Figure 2-3). For CLMA, stem elongation rates

in the shade were approximately twice that of sun-grown plants across all calculated














Chamaecrista nictitans


0-

0-
0-
0-
0,



0un r2 = 0 848
0o S Shade r2 = 0 937

92 122 152 182 212 242 272 30
Calendar Day


Crotalaria rotundifolia


25-
Sun r2 = 0 555
Shade r2 = 0 694
92 122 152 182 212 242 272 302
Calendar Day

Orbexilum lupinellus
100 I- i -

75-

50'


'D 30- i


Sun r2 = 0 832
SShade r2 = 0 645
92 122 152 182 212 242 272
Calendar Day


Mimosa quadrivalvis


5- 1 -


50-

5

SaSun r = 0 221
e/- Shade r2 = 0618
92 122 152 182 212 242 272 302
Calendar Day


Clitoria mariana


'0-*

60-
.0-



0- sun r2 = 0 726
0- shade r2 = 0 374

92 122 152 182 212 242 272 3
Calendar Day


Lespedeza hirta


SSun r2 = 0 898
Shade r2 = 0 575
2 122 152 182 212 242 272 302
Calendar Day


Tephrosia virginiana


.5
Sun r2 = 0 534
Shade r2 = 0 444
0-
92 122 152 182 212 242 272 302
Calendar Day


Centrosema virginianum


92 122 152 182 212 242 272 302
Calendar Day


Figure 2-3. Stem elongation. Values are means +/- SE. U represents plants grown in

the sun, and A represents plants grown under the shade treatment. Goodness

of fit values (r2; a=0.05) are given for each curve that was fitted with a third-

order polynomial.









Table 2-3. Regression equations for stem elongation curves of the form y= Dx3 + CX2 +
Bx + A. Coefficients are mean +/- SE.
Equation Coefficients

Species Light A B C D
CANI Sun -107 + 32.34 1.38 + 0.54 -0.004 + -0.002 -3.8x10-6 + 5.23x10-6
CANI Shade -63.16 28.22 0.74 0.47 -0.001 0.002 -5.18x10-7 4.52x10-6
CLMA Sun -71.01 + 37.33 1.09 + 0.59 -0.003 0.002 4.84x10-6 + 4.87x10-6
CLMA Shade -155.60 129.7 2.13 2.06 -0.007 0.010 8.45x10-6 1.70x10-5
CRRO Sun -48.50 + 44.59 0.48 + 0.71 -8.05x10-5 + 0.003 -2.26x10-6 + 6.00x10-6
CRRO Shade -50.50 + 51.62 0.49 + 0.82 -9.77x10-5 + 0.004 -1.28x10-6 6.95x10-6
LEHI Sun 186 60.94 3.62 0.97 0.02 0.004 -3.65x10-5 8.14x10-6
LEHI Shade 80.9 120.1 2.06 + 1.94 0.01 0.01 -2.78x10-5 1.65x10-5
ORLU Sun -69.92 26.73 1.18 0.42 -0.004 0.002 7.10x10-6 3.54x10-6
ORLU Shade -76.76 + 29.50 1.26 0.46 -0.005 0.002 7.31x10-6 3.90x10-6
TEVI Sun -48.13 58.98 0.65 0.93 -0.001 0.004 1.004x10-6 7.81x10-6
TEVI Shade -120.4 66.14 1.74 1.05 -0.006 0.005 7.88x10-6 8.70x10-6









Table 2-4. Values given are slopes calculated from the derivative of the equations given
in Table 2-3 at the mean for each coefficient.
dh/dt

Species Light Dayl31 Day 181 Day 232 Day 296
(14 May) (29 June) (19 August) (22 October)

CANI Sun 0.462 0.213 0.018 -0.141
CANI Shade 0.384 0.232 0.070 -0.145
CLMA Sun 0.319 -0.160 -0.455 -0.790
CLMA Shade 0.677 0.350 0.148 0.080
CRRO Sun 0.346 0.233 0.081 -0.157
CRRO Shade 0.399 0.329 0.237 0.094
LEHI Sun 0.294 0.799 0.749 -0.120
LEHI Shade 0.497 0.721 0.520 -0.346
ORLU Sun 0.263 0.106 0.569 0.151
ORLU Shade 0.260 0.073 -0.003 0.061
TEVI Sun 0.311 0.208 0.119 0.292
TEVI Shade 0.500 0.239 0.095 0.087









dates. Slopes and heights for CANI were very similar for both sun and shade-grown

plants throughout the growing season.

Most species reached maximum height near the end of the experiment (9 October

to 22 October, Day 283 to 296). However, CANI (15 August, Day 228) and MIQU and

ORLU (1 July to 5 September, Day 183 to 249) peaked earlier. CANI and MIQU

experienced a slight decline in measured height after peaking due to some defoliation and

stem breakage. Differences in maximum plant height were due to differences among

species, as determined by analysis of variance (p < 0.001), and treatment effect was not

significant due to the amount of variation between individual plants. Without regard to

treatment, LEHI and CEVI were the "tallest" plants, followed by MIQU, CLMA, TEVI,

CRRO, CANI, and ORLU, respectively (Table 2-5).

Leaf addition patterns for CEVI, CRRO (after 15 June, Day 167), and TEVI show

that the plants grown in the shade tended to have more leaves on a given date than those

grown in the sun. Due to large variability among individual plants, this pattern is only

significant on a few dates for CRRO during the season (Figure 2-4A, B). CLMA, CRRO

(before 15 June, Day 167), and LEHI patterns reveal the opposite effect, more leaf

addition occurring on those plants grown in the sun than those in the shade. Again, due

to large variability between individual plants, this effect is only significant for CLMA

and LEHI on a few dates across the season (Figure 2-4A). ORLU, MIQU, and CANI did

not show distinct patterns with regard to leaf addition over the season, and no statistical

differences according to treatment were found on any dates.

Differences in root elongation patterns were difficult to detect due to the high

mortality rates of specimens grown in the root tubes. In addition, due to the coloration of









Table 2-5. Maximum plant heights by species, regardless of treatment. Values are
means SE. Different letters indicate significant differences (Duncan's post-
test).


Day


289 (15 October)
296 (22 October)
225 (12 August)
293 (19 October)
296 (22 October)
283 (9 October)
228 (15 August)
249 (5 October)


Max. Height (cm)

53.2 +/- 27.6 a
102.0 +/- 10.1 a
81.3 +/- 9.5 ab
53.2 +/- 9.7 bc
46.6 +/- 4.6 c
44.5 +/- 4.1 c
33.6 +/- 1.3 c
31.4 +/-3.3 c


Species


CEVI
LEHI
MIQU
CLMA
TEVI
CRRO
CANI
ORLU











A Clitoria mariana
1500 |
s, Sun
a Shade

9001
S600

300,

0 ---II-
130 160 190 220 250 280
Calendar Day

Tephrosia virginiana
1500- I I i I Ip I

1200-
O 000
-J 900. 0
o 0 [ u 3
600- *
n o3
300- M 2

0 I I II
130 160 190 220 250 280
Calendar Day

Lespedeza hirta
1500 1 1 1 1



jo
900
S600

3 300- M 3D


125 155 185 215 245 275
Calendar Day



Figure 2-4. Leaflet counts and plant widths of sun and shade grown plants. Values
shown are means + SE. A) Leaflet counts for Clitoria mariana, Tephrosia
virginiana and Lespedeza hirta plants grown in sun and shade. B) Width of
Crotalaria rotundifolia plants grown in sun and shade.












Crotalaria rotundifolia


150. I I -


100-



50-


180


210


240


270


Calendar Day


Figure 2-4. Continued.


A Sun
A Shade


A A AA AA
A A
A
AL .


S.


300









the roots being close to that of the soil and the tendency of roots to not remain along the

walls of the tubes, measurement were highly variable, and no differences according to

treatment were significant. The rooting depth of the specimens in the sun and shade

treatment were very similar for the first half of the growing season, although sun-grown

plants tended to root slightly deeper in all three species shown (Figure 2-5). Shaded

specimens appeared to have rooted slightly deeper than those in the sun at approximately

10 June (Day 162) for TEVI, and 30 June (Day 182) for LEHI and CRRO.

Biomass accumulation was not affected by light treatment for either total biomass,

aboveground or belowground accumulation (p = 0.3765), therefore, additional values

reported are pooled treatments by species. MIQU accumulated the largest amount of

biomass over the season, followed by CANI and LEHI, then CEVI, CRRO, TEVI, and

finally CLMA and ORLU (Figure 2-6).

Root to shoot ratios (R/S) showed no significant treatment effect, but species effect

was highly significant (p <0.001). MIQU and TEVI showed heavy allocation of biomass

belowground (R/S>3, Figure 2-7), followed closely by CLMA, which had twice as much

allocation below- versus aboveground (R/S>2). CEVI and CRRO had a mostly balanced

allocation above- and belowground (R/S 1). CANI, LEHI and RHRE favored

aboveground allocation (R/S < 1). These patterns of biomass allocation did not reflect

the same pattern as that of total biomass, but the different allocation patterns were

represented in each of the groupings.

















-8
)E
O)
C 6
0
-j 0 6(
4 '- 30



10
_j


C.-85
0) U










ge
C5o

0


.i 3f

_J>







10






_J 0 60.

0 0

C).- 35-


10-


CRRO






-
5-




*
5.


**
5- 0


0.


92 122 152 182 212 242 272 302

Calendar Day


LEHI


5-
so
















M IQU
sO


O*


0.


92 122 152 182 212 242 272 302

Calendar Day


MIQU
I I I I I





000 0 0

.*.ee.O


0- i i i -


92 122 152 182 212 242 272 302

Calendar Day


Figure 2-5. Elongation of roots grown in the sun and under shade treatment. Data was
composite by species since light treatment effect was not significant. Values
given are means + SE.










20- b b
% 10- bc bc bc c
0--
10-

20- Aboveground
30- M Belowground
40
MIQU LEHI CANI CEVI TEVI CRRO CLMA ORLU

Species
Figure 2-6. Aboveground and belowground harvested biomass. Values are means +/-
SE. Data is composite by species because light treatment effect is not
significant. Different letters represent significantly different total biomass
(aboveground + belowground) values, and alphabetical order designates order
of total biomass values, greatest to least (Duncan's post-test).










* I


O
0

4-
0
o
0 3-




W 1n


0-I-


CANI CEVI TEVI C

Species


I O
4 ORLU RHRE


Figure 2-7. Root to shoot ratio of harvested biomass. Values are means +/- SE. Data are
composite by species because light treatment effect is not significant.
Different letters represent significantly different root-to-shoot ratios, and
alphabetical order designates order of ratios, greatest to least (Duncan's post-
test).


d d









Discussion

Morphology and Phenology

The diverse native legumes in this study represented all three sub-families of

legumes, as well as vines, erect herbs, and a semi-woody shrub. The life histories of

these species represent a variety of strategies for overcoming shade, frequent fire and

drought conditions. However, of the forty native legume species that grow extensively

across the Ichauway reserve, only two are annual species: Cassia fasciculata and C.

nictitans (Hainds et al.,1999). Perennial growth form is common in many frequently-

burned ecosystems (Knapp et al., 1998; Morgan, 1999; Jacobs and Schloeder, 2002) and

is apparently an effective adaptation to the frequently-burned longleaf woodland

environment. The nodule morphologies of the species in this study follow the types

described by Sprent (2002) for each of the subfamilies.

The majority of the species in this study began flowering during the late-spring,

early summer. LEHI, like many of the other semi-woody Lespedeza and Desmodium

species in the native woodland, flowered during the early fall (Chapter 3). All of the

plants that produced fruit continued to flower and fruit until the time when they were

harvested, except for CLMA. The indeterminate nature of flowering and fruiting

represented by most of the species in this study is another adaptation to a fire-maintained

ecosystem. Hiers et al. (2000) found that many legumes showed little change in duration

or timing of peak flowering in response to the season in which they were burned,

especially those that were fall-flowering or that matured multiple seed crops each year.

Those plants that continuously produced seeds over the course of the season continuously

contributed to the seed bank, and thus, did not need to re-grow and mature for a hastened

seed crop after a late growing-season fire.









LEHI represented a fall-flowering plant that has determinant seed production in the

wild, although it was not well represented in this study due to the timing of the harvest

date. However, the number of seeds produced by LEHI in each cohort of seeds and the

ease and range of its dispersal is much greater than that of the species which produce seed

throughout the year. Unlike the other species in this study that rely on dehiscent pods or

gravity for localized dispersion of seed, LEHI seeds adhere to fur (or clothing) of passing

animals and are therefore widely dispersed. TEVI, which did not flower during this

experiment, is an example of a species that flowers prolifically in the field in response to

fire (Clark, 1971), although the flowering may be delayed by lightning-season fires

(Hiers, 2000). Seed production (or lack thereof) in this study may not be indicative of

production in the wild, due to unknown pollination factors that may not have been

present in urban Gainesville, FL versus in the natural woodlands.

Growth Patterns

The shading treatment imposed in this study was 56 percent of full sun.

Measurements of light quantity in the native woodland, below the tree canopy, but above

the wiregrass canopy may be variable but is approximately 40 percent (Pecot et al.,

2005), which is substantially higher than that typically reported for other forest types,

including young pine plantations. Thus, shade tolerance among species in the longleaf

woodland understory is not on the same scale of tolerance for understory species in other

forest types that may only receive around 3 percent light infiltration (Battaglia et al.,

2003). Light quality factors such as the red to far-red ratio may also be important to

understanding responses of understory species, however, such analysis was outside of the

scope of this study.









The species in this study all had variable growth responses to the two light

environments, but shading did not have a significant impact on total biomass, or R/S

allocation patterns (Figure 2-7). Although the small sample sizes due to high mortality in

this study reduced our effectiveness to detect biomass shading effects, these effects were

manifested in the later measurement dates for stem elongation (Figure 2-1). The effect of

shading on height growth of CLMA, CRRO, and LEHI became was significant after 15

June (Day 167). CLMA and CRRO shade-grown plants were significantly taller than the

sun-grown plants toward the end of the growing season, and LEHI grew significantly

taller in the sun. The fact that tallest plants for individual species were not all located in

the shade indicates that the shade-grown plants were not simply etiolated, but rather

showed some degree of shade tolerance.

CRRO, which is a prostrate spreading plant, and CLMA, which is semi-erect to

vining, both grow below and amongst the bunchgrasses of the native woodland, thus,

their tolerance of shade is not surprising (Figure 2-1). LEHI, a tall semi-woody species,

showed a more favorable growth response to sun than shade during the last weeks of the

growing season. In the woodland, LEHI quickly outgrows the surrounding grasses and

avoids mutual shading in the understory more aggressively than the smaller-stature

species. CANI and TEVI, which showed no significant responses to shading, attain the

same height as the surrounding grasses in the native woodland. The effects of shading on

the two large vines, CEVI and MIQU, were not well defined in this study due to high

mortality and the difficulties of measuring vines that intertwine and break easily.

However, even if these vines have reduced shade tolerance, they still have the ability to

climb over neighboring plants and to grow into sunflecked gaps in the savanna in order to









reach more direct light than smaller plants, such as ORLU and RHRE, for which

adaptation to some shading would be more advantageous.

The variation in reported height among species, from 30cm to 100cm, represents

the diversity of sizes and growth forms among these species. However, if the crown

width of vining/spreading plants is considered, CEVI is the largest plant, spreading to

over 2m, followed by MIQU, which grew to over Im across, as well. CRRO also spread

approximately twice as wide (Figure 4) as it grew tall, -80cm versus -45cm,

respectively.

Keeping in mind the differences in plant morphology, the most decisive way to

address comparisons of size of these species is by using total biomass. In an ecosystem

that is frequently-burned and prone to frequent drought, significant allocation to

belowground biomass is a positive survival adaptation (Knapp et al., 1998). These

perennial species that can re-grow from reserves in the roots before surrounding plants

will have a temporal and light-availability advantage over neighbors. MIQU and TEVI,

both with R/S>3 (Figure 2-7), have also been shown to be significantly altered in

phenology by fire (Hiers et al., 2000). Young MIQU plants excavated in the field had

taproots exceeding 2 m long (personal observation). Plants that have higher R/S ratios

and deep, branching root systems are better at enduring drought conditions (Kramer and

Kozlowski, 1979; Knapp et al., 1998), such as those that develop quickly in the coarse-

sandy soils present in much of the longleaf-wiregrass ecosystem (Hainds et al., 1999).

For a legume, an extensive root surface area also provides increased opportunity for

rhizobial colonization.









The legume species in this study are difficult to generalize as a single group due to

their diverse growth forms, reproductive phenologies, and life histories. Although the

reduced early survivorship of our study population obscured some of our ability to

document shade effects for all species, these data indicate that examples of legumes with

some degree of shade tolerance can be found outside of the tropics (Sprent, 1999). Some

species exhibited greater height growth with shading, but none of the species in this study

demonstrated strong shade intolerance through reduced biomass responses to the 56

percent of ambient light level treatment. This light regime would certainly be

representative of conditions in open longleaf pine savannas or in thinned young

plantations with restoration plantings (Chapter 4), although a more heavily shaded

treatment might have provided a greater effect upon several species. However, most of

the life-history characteristics of the species examined in this study may be more strongly

associated with adaptation to fire, N deficient soils, and drought than to light environment

in longleaf pine savanna ecosystems.














CHAPTER 3
USE OF CORROBORATIVE METHODS TO ASSESS THE N2-FIXATION OF
NATIVE LEGUMES

Introduction

The fire-dependent longleaf pine- (Pinuspalustris Mill.) wiregrass (Aristida strict,

Michx.) savanna ecosystem once dominated the southern coastal plain of the United

States, covering as much as 37.2 million ha. Presently, less than 1.5 million ha of these

ecosystems remain intact (Landers et al., 1995). However, restoration plantings

reclaiming more than 283,000ha of former agricultural fields, pulpwood plantations, and

other fire-suppressed lands are being established in the Southeast, 48,000ha of which are

in USDA Conservation Reserve plantings in Georgia (Coffey and Kirkman, 2004).

Groundcover reestablishment is the key to restoring wildlife habitat in these young

longleaf pine plantings as well as providing continuity of pyrrhic fuels (Clewell, 1989;

Kirkman, 2002). Legumes may also have a major role in maintaining N balance fire-

maintained restored systems that are being established on depleted, former agricultural

soils (Markewitz et al., 2002; Boring et al., 2004). Planting of native legumes would

permit groundcover restoration without some of the problems that have occurred as a

result of introducing non-native and agronomic legumes, including serecia lespedeza

(Lespedeza cuneata) and L. bicolor, that have little shade tolerance and can be very

invasive (Miller, 2003).

There is a need for species of legumes native to the longleaf-wiregrass ecosystem

to be identified for use in groundcover restoration plantings that show strong potential for









N2-fixation and that can make large, N-rich biomass contributions to depleted soil

organic matter (Markewitz et al., 2002). In addition, adaptation to environmental factors

such as nutrient deficiency, drought conditions, and reduced light under a young closed

forest canopy must be considered along with an analysis of N2-fixing potential. After

initial screening, these species should be planted in field conditions for a better

assessment of their physiological adaptations and N2-fixing potential (Dreyfus et al.,

1988).

Assessing N2-fixation under field conditions is a difficult task, and most current

methods measure fixed-N2 indirectly. Traditional methods of assessment involve nodule

excavation and other destructive biomass measures to assess N-fixation and cannot be

performed repeatedly on the same plants. These methods can also be destructive when

used in a woodland ecosystem because of the amount of disturbance caused by

excavating root systems. The results of traditional methods of N2-fixation assessment,

such as nodule biomass measures, total-N comparison, and the acetylene reduction assay

are not readily convertible into actual amounts ofN2-fixed. However, more direct

methods, including the 615N natural abundance and 15N enrichment methods (Virginia et

al., 1989; Hiers et al., 2003), are integrated over time and can be used to calculate a

quantitative estimate ofN2-fixed. These methods may eliminate the problems associated

with instantaneous measures of fixation that may fluctuate over diurnal and seasonal

conditions, such as the acetylene reduction assay (Boring and Swank, 1984; Halvorson et

al., 1992). By combining methods of assessing N2-fixation that include both

instantaneous and cumulative measures and first applying them in a controlled study, a

well-informed assessment can be made of the N2-fixing capabilities of a legume species.









This study was designed to compare the N2-fixation capabilities of nine species of

legumes native to the longleaf-wiregrass ecosystem. Two additional objectives were also

addressed: (1) to compare estimates of N2-fixation capabilities as derived from five

different methods of assessment: nodule biomass, the acetylene reduction assay, N

transport and storage product analysis, 615N natural abundance, and total N content; and

(2) to examine the effects of shading on their N2-fixation activity.

Methods and Materials

Planting

Seedlings of nine native legume species were used in this experiment. They were

propagated as described in Chapter 2, and were obtained from Dr. Kay Kirkman at the

Joseph W. Jones Ecological Research Center, Baker County, Georgia: Centrosema

virginianum (L.) Benth. (CEVI), Clitoria mariana L. (CLMA), Crotalaria rotundifolia

J.F. Gmel. (CRRO), Lespedeza hirta (L.) Hornem. (LEHI), Mimosa quadrivalvis (L.)

(MIQU), Orbexillum lupinellus (Michx.) Isley (ORLU), Rhynchosia reniformis D.C.

(RHRE) and Tephrosia virginiana (L.) Pers (TEVI). Seeds for the annual Chamaecrista

nictitans (L.) Moench (CANI) were obtained from Dr. Ken Quesenberry (Agronomy

Department, University of Florida). Six specimens of each species were planted in pots

that were specially designed to allow gas to flow through them as part of a closed system.

These pots were constructed from a capped 35-cm long section of 10.5-cm diameter

polyvinyl chloride (PVC) pipe.

Fourteen specimens from each species were transplanted into black tree-seedling

pots, 35-cm tall x 80-cm2 base (Stuewe and Sons, Corvallis, OR). Transplantation of

seedlings was completed 11 February 2004. CANI was planted on 21 April 2004 from

seeds that were germinated on moist filter paper. Due to the small numbers of plants,









destructive harvests of seedlings were not possible at the beginning of the experiment

(To), but non-destructive measurements were taken. Plant heights at planting represented

between 5.3 and 19.5 percent of the final heights. The smallest plants, such as ORLU

and RHRE had the largest percentage of the final height present at To, and the larger

plants, such as LEHI, were represented by the lower range of the percentages. Leaves

were also counted at To to determine percentage of leaves initially present. Percent

leaves present at To ranged from 1.5 to 8.7 percent for the small- to large-stature plants,

respectively.

All plants were grown outdoors in Gainesville, Florida (29 38' N, 82 20' W)

between April and November of 2004. A shade cloth enclosure was used to create an

environment that provided a 0.54 fraction of total ambient light as determined using a Li-

Cor Quantum Sensor, LI-185A (Chapter 2).

Plants were inoculated by introducing native soil to the pots. Two topsoils were

collected at the Jones Center, from a fine-loamy, kaolinitic, thermic Typic Kandiudult

(Orangeburg Series), and from a loamy, kaolinitic, thermic Arenic Kandiudult (Wagram

Series). Each collection was made in an area with a diverse, thriving legume population.

Soil collections were transported to Gainesville, FL where they were stored in a cool,

dark room and covered with plastic to maintain moisture. Equal parts of each soil

collection were mixed in order to provide inoculation for legumes that may be

predominantly found in different locations. Pots were filled with purchased topsoil

(Walmart Corp.) to within 6 cm from the top of the pot. Next, 2 cm of the native soil

mixture were added, and finally, the plant was transplanted, using additional topsoil to









cover the roots as needed. Water was applied by drip-irrigation every 12 hours using a

battery-operated timer (Rainbird).

Plants from acetylene reduction assay pots and black plastic pots were destructively

harvested on 11 November 2004. Roots and nodules were washed free of soil, nodules

counted, and a 2-cm section was collected from the base of each stem for analysis. All

tissues were dried to constant weight at 800C.

N2 Fixation Assessment

Measurements of ethylene production as a result of exposing legume roots to

acetylene (C2H4) gas were taken every three weeks over the course of the experiment.

Repeated assays using the same plant were made possible by the use of a non-destructive,

flow-through system.

Specially-designed pots were used to allow the plants to be repeatedly assayed with

minimal disturbance. The gas input line was attached to the bottom of the pot. A lid

matching a 3.6L (20-cm tall x 17-cm diameter) plastic container (Rubbermaid Inc.) was

permanently attached with wing nuts and screws to the flange that formed the top of each

pot, and the joints were all sealed air-tight using weather-strip caulking. The center of

each lid was removed to reveal the mouth of the pot. The plastic containers were

inverted over the top of the plant and sealed to the stationary lid. On the bottom of each

plastic container (the top of the apparatus when inverted), a port was created over which

the gas line would fit tightly. Air-tightness was confirmed for each sealed pot before gas

flow was initiated.

Ten percent acetylene was flowed into the pots at a rate of 1L min-. Gas samples

for each pot were taken both from the lines running into the pots and those running out.

The inflow samples served as a baseline for the amount of ethylene that might be present









in the acetylene source. The outflow samples, which flowed past the plant root systems,

contained the ethylene generated from acetylene reduction as a result of nitrogenase

activity. Duplicates of 1-mL samples were collected using syringes that were inserted

into the tubing. Samples were transported to the laboratory and analyzed immediately.

Analyses were performed using two gas chromatographs (Hewlett Packard 5710A and

Shimadzu GC-8A).

Stem sections were cut from the basal 3cm of plants grown in PVC and black

plastic pots as they were harvested (11 November 2004). Each stem was placed in a 20-

mL glass vial and covered with a phosphate buffer solution, mixed according to Herridge

(1984), except ethanol was used as the solvent rather than water at the suggestion of

Izaguirre (personal communication). Vials with stems in solution were stored at 0C until

extracted.

Stem sections were placed in a boiling waterbath (100C) for 25 minutes to extract

stem contents. Stems were removed and dried to constant weight at 800C. Deionized

water was added to each extract to a standard volume of 25 mL. Extracts were covered

and stored at -300C between analyses to prevent evaporation. Aliquots of each extract

were analyzed for ureide, NO3, and a-amino acid concentrations using spectrophotometry

(Beckman DU 640). Ureide concentrations were estimated as the phenylhydrazone of

glyoxalate using allantoin as the standard, and a-amino acid concentrations were

determined using a modification of the ninhydrin method (Yemm and Cocking, 1955),

with asparagine as the standard; both analyses were conducted as described by Herridge

(1984). NO3 was extracted using the salicylic acid method as reported by Cataldo et al.

(1975). Taking into account that ureides contain 4 N atoms per molecule, an index of the









relative abundance of ureide-N in each extract was calculated according to Peoples et al.

(1996), where the bracketed variables indicate the molar concentration of each of the

extract constituents:

RUI = 400[ureides] / (4[ureides] + [nitrates] + [a-amino-acids]) [3-1].

Use of 615N natural abundance technique in the field requires that differences

between the soil and atmospheric 15N pools be well established by also ascertaining 615N

values of non-fixing reference plants and available soil N (Virginia et al., 1989). ORLU

was selected as a non-fixing reference species for this controlled study was selected using

corroborative assessments of N2-fixation.

Stems and leaves harvested from plants grown in the PVC and black plastic pots

were coarsely ground using a Cyclotec Sample Mill, and then ground to a fine powder

using a Spex CertiPrep Mixer Mill 8000-D. Roots were not analyzed for 615N due to the

amount of organic matter that remained attached after thorough cleaning. Finely ground

tissues were analyzed for 15N natural abundance and N content at the University of

California, Davis (Stable Isotope Facility, Department of Agronomy, Davis, CA) using

mass spectrometry. 15N natural abundance is expressed as 615N (%o 15N depletion units):

615N = [(atom%15Nsample / atom%15Nstandard) 1] x 1000 [3-2]

where the standard is the atom percent 15N concentration of atmospheric N2.

Percent of total N derived from the atmosphere (% Ndfa) was calculated according

to the equation:

% Ndfa = 1 (615NN2-fixing plant / 615Nref) [3-3].

Statistical Analysis

Data were analyzed using analysis of variance (ANOVA) with species and light

environment as main effects. If differences existed (p<0.05), Duncan's multiple









comparison test was used to determine which means differed significantly. The GLM

procedure performed in the Statistical Analysis System (SAS, 2003) was used for

ANOVA and post-tests.

ANOVA analysis did not show any significant treatment effects for any of the

datasets in this study. Therefore, differences further discussed relate only to those among

species, and means reflect a composite of both treatments.

Results

Survivorship

Of the six specimens for each species that were planted in the pots for the flow-

through acetylene reduction assay, an average of three survived. LEHI and CRRO had

the best survivorship, with five plants each, followed by CANI and TEVI with four, then

CLMA with three. RHRE and ORLU had the worst survivorship, with no RHRE plants

surviving, and only two ORLU, which did not permit statistical analysis of treatment

effect on the acetylene reduction assay for these two species.

An average of four out of fourteen specimens grown in black tree seedling pots

reached final harvest. LEHI, had the best survivorship with eight remaining, followed by

TEVI and CLMA with six and five plants surviving until final harvest, respectively.

CANI, CEVI, CRRO, RHRE, ORLU, and MIQU, had average or lower survivorship.

ORLU and MIQU had only two and one plant surviving, respectively. However, since

these plants were used in addition to the ones harvested from the PVC pots, statistical

analysis was still possible for all species for most fixation indices.

Shade treatment effect was not significant for any of the N2-fixation assessments in

this study. Small samples and high variability contributed to the inability to detect

statistical differences according to treatment. Ambient light for this study was also









impacted toward the end of the season by tropical storm occurrence. Detectable

differences in plant height were developing around this period, and the lowered

irradiance overall may have impeded the maximum effect of shading during this critical

growth period. However, results from biomass and nodulation data did not show

significant responses to shading.

Fixation Assessment

Species differences in nodule mass were significant (p = 0.0006). MIQU had the

greatest nodule mass, but further statistical delineations were not detectable due to the

large amount of variation between individual plants (Table 3-1). None of the small

herbaceous species in this study had an average nodule mass of more than Ig.

Nitrogenase activity (acetylene reduction) was initially very low, overall, however,

later in the season there were isolated peaks of substantial ethylene production from the

older plants (Figure 3-1). CANI, LEHI, CEVI, and MIQU all showed an increase after

the middle of the season, around 28 June (Day 180). CLMA showed the lowest

nitrogenase activity across the season, never producing more than 0.3 tMmol C2H4 hr-1 at

any given date. ORLU also showed low activity, with only a single peak above 0.3 [tmol

C2H4 hr-1. Comparison of late season maximum ethylene peaks using analysis of

variance showed a significant species effect (p = 0.0002). The peak for MIQU was

significantly greater than all other species maximums, but no other significant differences

among the remaining species were detected (Figure 3-2).

The N transport and storage products extracted from stem sections of native

legumes were dominated by ureides and a-amino acids. The relative concentration of a-

amino acids (RAC) was approximately equal to that of the relative ureide concentration

(RUC) for most species, followed by a very minute NO3 concentration (Table 3-2).









Ureide appears to be an important transport/storage product molecule for this suite of

species. The differences among species were not significant for RUC or RAC, but were

highly significant for total extracted N and relative NO3 concentration (RNC; p = 0.0006

and 0.0001, respectively).

The 615N values of the species were significantly different and ranged from -3.13 to

-1.45 (Figure 3-3). CLMA had the 615N value closest to zero, followed by CANI, CEVI

and TEVI, which all had values of approximately -2.3. ORLU and RHRE had the most

negative values, approximately -3.2. However, only CLMA and CRRO were

significantly different from ORLU, which was used as the non-fixing reference.

Percent Ndfa was calculated (Equation 3-3) using ORLU as the non-fixing

reference. ORLU had very low nitrogenase activity (acetylene reduction) across the

season, indicating very little, if any, N2-fixation, in spite of having numerous small, but

apparently ineffective nodules. Species effect was significant for percent Ndfa (p

0.0002), and values ranged from 12.0 to 54.9 percent (Figure 3-4A). CLMA was

determined to have the highest percent Ndfa, followed by CRRO > TEVI > CEVI > MIQU

> LEHI > RHRE.

Percentage of N (gN g-1 tissue) in stem and leaf tissues ranged from nearly 3.0

percent to approximately 1.5 percent (Figure 3-4B). Species effect was significant for

percent N (p = 0.0001). MIQU had the highest percentage of N (2.83 percent), followed

by CANI, then CRRO > CLMA > CEVI > ORLU > TEVI, and RHRE. LEHI had the

lowest percentage of N in its tissues (1.47%).

Total grams of plant accumulated N over the season was significantly different

among species (p = 0.0001). CANI and MIQU had the largest amount of accumulated N,









Table 3-1. Nodule mass and number of nodules. Rankings are according to Duncan's
post-test. RHRE not included in ANOVA (n = 1).
Mass # Nodules
Species (g) (per plant)
MIQU 0.9114 a 218.60 a
CEVI 0.3303 b 59.43 b
TEVI 0.1316b 51.40 b
CANI 0.1175 b 44.00 b
CRRO 0.0864 b 27.5 b
CLMA 0.0818 b 32.17 b
LEHI 0.0547 b 48.92 b
ORLU 0.0229 b 27.25 b
RHRE 0.0100 7.00













Chamaecrista nictitans


3 3-
C

w 2-

1-i

0-
120 150 180 210 240 270 300
Calendar Day


Lespedeza hirta
411


Clitoria mariana


d 2

S 08-
*i '




06-
S04
0 2-

120 150 180 210 240 270 300
Calendar Day

Crotalaria rotundifolia
4 1 I


3 3
, 2 2

o 08- 8-
S06- 6-


o ----- 00 .. -- -
120 150 180 210 240 270 300 120 150 180 210 240 270 300
Calendar Day Calendar Day


Centrosema virginianum
41


e e 3

1 0- 2
0 8-
IE 0-
:- 02-



120 150 180 210 240 270 300 1
Calendar Day


Orbexilum lupinellus


Mimosa quadrivalvis
4 I


-.h~




20 150 180 210 240 270 3(
Calendar Day


Tephrosia virginiana


S061 0
E. 04 iE 04-

02- 0 -
0 O0-

120 150 180 210 240 270 300 120 150 180 210 240 270 300
Calendar Day Calendar Day



Figure 3-1. Ethylene production trends for the growing season by species. Data shown

are means + SE with treatments combined, and lines show mean trends.







49





4
a


.7 3
3-

C
,v b






-c c
E


C C C C


0 1
MIQU CANI ORLU LEHI CRRO CEVI TEVI CLMA

Species

Figure 3-2. Maximum ethylene production (C2H4 reduction) peaks. Data shown are
means + SE. Different letters represent statistical differences (Duncan's post-
test).









Table 3-2. N-transport/storage products extracted from stem sections. RUC, RAC, and
RNC represent relative concentrations of ureides, a-amino acids, and NO3,
respectively. RUI and Total N are relative ureide index (Equation 3-1) and
total extracted N (mmol N g-1 stem). Letters represent statistical differences
within columns according to Duncan's post-test. Species effect was not
significant, ns.
Species RUC RAC RNC RUI Total N
(mmol Ng1
stem)
CANI 45.68 54.28 0.032 a 73.01 0.0018 b
MIQU 34.26 65.72 0.008 e 62.11 0.0055 a
LEHI 27.15 72.82 0.020 bc 56.26 0.0014 b
CEVI 46.09 53.89 0.013 de 67.15 0.0049 a
CRRO 55.75 65.75 0.013 cd 67.15 0.0035 ab
TEVI 24.84 75.14 0.010 e 48.39 0.0035 ab
ORLU 51.83 48.14 0.026 b 79.79 0.0056 a
CLMA 44.23 55.75 0.013 de 67.15 0.0035 ab











-1.0


-1.5


-2.0


-2.5


-3.0


-3.5


-4.0


I.


ab a abc
acabc

i { -r


bc
bc

*


CLMACRRO TEVI CEVI MIQU LEHI I RHREORLU
CLMACRRO TEVI CEVI MIQU LEHI CANI RHREORLU


Species


Figure 3-3. 615N values by species. Data shown are means + SE. Different letters
represent statistical differences (Duncan's post-test). RHRE, (n=l) was not
included in the ANOVA.







52





A 70

60 a

50

40 abc ab
z abc
30 -

20 bc

10 -


CANI MIQU LEHI CEVI CRRO TEVI ORLUCLMARHRE
Species


3.5
B
a
3.0

2.5 -
Sbc bc
b cd cd bc
im 2.0
z

z 1.5

1.0

0.5

0.0
CANI MIQU LEHI CEVI CRRO TEVI ORLUCLMARHRE
Species
Figure 3-4. Mean % Ndfa, %N, and total N by species. Different letters represent
statistical differences (Duncan's post-test). A) Percent of total N derived from
the atmosphere. CANI and ORLU did not have adequate replication for
inclusion in ANOVA, and ORLU was used as a non-fixing reference in the
calculation of % Ndfa (Equation 3-3). B) Percent N in aboveground tissues.
C) Total N content of aboveground tissues.































CANI MIQU LEHI CEVI CRRO TEVI ORLUCLMARHRE
Species


Figure 3-4. Continued









followed by LEHI, CEVI, and CRRO, and finally, TEVI, ORLU, CLMA, and RHRE

(Figure 3-4C).

Discussion

Comparison of Methodology

Although there were some differences in the way the N2-fixation capacity of the

species were ranked by each assessment technique, the species could generally be

assigned into high- and low-fixer categories. With little discrepancy between assessment

results among methods, MIQU, CANI, CRRO, CLMA and CEVI showed higher N2-

fixation potential, and LEHI, TEVI, and RHRE showed relatively lower potential. The

results for ORLU generally corroborated that this species did not form effective nodules

in this study and thus had the least effective N2-fixing capacity, approximately none

(Figures 3-1, 3; Table 3-3).

The acetylene reduction assay, an instantaneous indicator of nitrogenase activity, is

strongly affected by environmental stresses such as drought conditions and light

reduction, which reduce photosynthesis rates. Therefore, seasonal and diurnal patterns in

nitrogenase activity can generally be well documented with frequent acetylene reduction

assays (Boring and Swank, 1984; Zitzer and Dawson, 1989; Halvorson et al., 1992;

Peoples et al., 1996). This method of assessing N2-fixation is also a quick and

inexpensive, instantaneous first assessment of nitrogenase activity. However, it is

necessary to follow temporal patterns of plant development to adequately index an annual

pattern of nitrogenase activity, and even then, the assay is not a direct measure of N2-

fixation, since many species have nitrogenase systems with varying efficiencies and

hydrogenase activities (Zitzer and Dawson, 1989).









Table 3-3. Specific nodule activities of species in this study and other comparative
reports. Specific nodule activities for this study were estimated using the
results of the final acetylene reduction assay (22 October 2004) and biomass
of harvested nodules. Data from this study are ranges and means SE.
1Hendricks and Boring, 1999; 2Hogberg and Kvarnstrom, 1982; 3Boring and
Swank, 1984; 4Zitzer and Dawson, 1989.
Specific nodule activity
Species
(tumol C2H4 hrf mg nodule- )
Centrosema virginianum (CEVI) 0.272 0.272
Mimosa quadrivalvis (MIQU) 0.297 + 0.116
Other native legumes 0.002 to 0.007
Orbexilum lupinellus (ORLU) 0
Desmodium viridiflorum' 0.08
Lespedeza procumbens' 0.04
Leucaena leucocephala2 0.048
Robinia pseudoacacia3 0.05
Alnus glutinosa4 0.02 to 0.02
Eleaganus angustifolia4 0.002 to 0.01









Specific nodule activity comparisons between species in this study and other

similar and larger species of legumes suggest that CEVI and MIQU have extremely high

nodule efficiency (Table 3-3). However, these comparisons may not be completely

accurate due to the difference in sampling techniques. The flow-through assay to

measure nitrogenase activity (acetylene reduction) that was used in the current study,

avoids many of the problems associated with field sampling, which was used in the

comparative studies, such as limited nodule recovery and possible evolution of ethylene

from excised plant parts. The disparity between values given for strong N2-fixers in this

study as compared to others may be partially due to these sampling differences.

The use of the relative ureide abundance technique for assessing N2-fixation was

limited in this study by the extremely low NO3 concentration in the stems which shifted

the relative ureide content (27-55%) higher than would have been expected according to

results reported in other studies (RUC, 1-33%; Izaguirre-Mayoral et al., 1992; Sicardi de

Mallorca and Izaguirre-Mayoral, 1993; Medina and Izaguirre, 2004). As a result, the

RUI, which takes into account the 4:1 atomic ratio of ureide to NO3 and is the value

usually compared among species, was very high for all species in this study, making the

use of delineation between high and low N2-fixers according to RUI values as reported by

Izaguirre et al. (1992; RUI > 60, high; RUI < 30, low) difficult to apply to this study.

The most valuable comparisons for this study appeared to be between those species that

appeared to be using a relatively high or low amount of soil NO3, as distinguished by

RNC values (Table 3-2).

The 15N relative abundance technique is most useful in field studies with maturing

plants, with adequate sample numbers of legumes, and with a verified non-N2-fixing









species as a reference plant. In this controlled study, ORLU was selected as a reference

species using corroborative methods to determine that it was non-N2-fixing and could

represent soil 815N uptake values. The strong 615N signature of CLMA is not easily

understood in that this species had one of the lowest nitrogenase activities across the

season. CLMA also had one of the lowest RNC indices, which would indicate that it was

using less soil-N than some of the other species. It is possible that ephemeral periods of

peak nitrogenase activity occurred within the three-week intervals between acetylene

reduction assays, and it is also possible that since this species had very small biomass

(Chapter 2), low levels of N2-fixation could have a relatively stronger influence over the

615N value than the same levels of N2-fixation in a larger plant. It appears possible the

consistent, low levels of nitrogenase activity could have a strong influence on the 615N

values of CRRO and TEVI, as well (Figures 3-1, 3-3).

Total N does not follow the same pattern as percent N because it also incorporates

the total biomass of the plant which varied dramatically among the species (CLMA, 9 g

to MIQU, 40 g). This measure follows the pattern of aboveground biomass accumulation

very closely (Chapter 2). Although not completely applicable as a comparative measure

of N2-fixation, an estimate of total N for each of these species can be very useful for

considering which should be included in restoration plantings. Those with the greatest

biomass and tissue-N overall will contribute the most to building the N-pool, soil organic

matter, and N availability in depleted soils (Markewitz et al., 2002). This estimate is also

useful in determining which of the species could provide the greatest amount of N-rich

material available for wildlife to browse (Hendricks and Boring, 1999).









This study points out some differences in N2-fixation capabilities among the

species of legumes native to the longleaf-wiregrass woodland ecosystem, as well as

differences in results that can occur from using various assessment methods. The merits

and problems associated with using a variety of methods to assess N2-fixation in this

controlled study were most evident when comparing results including some immature,

small plants. Evidence from shorter-time-based measures such as the acetylene

reduction assay and N-transport product analysis did not always corroborate definitively

with cumulative measures for these small plants. Cumulative measures (total N, 615N

approaches) may hold more insight for older, established perennial plants, especially field

populations. Nodules are ephemeral and N2-fixation rates will change diurnally and

seasonally according to environmental conditions. Instantaneous methods may not be the

most appropriate measures to estimate actual or maximum potential N2-fixation, but can

serve as a quick and inexpensive way to determine seasonal and diurnal-dependent

patterns of nitrogenase activity. Instantaneous measures may be best used to document

age of effective symbiosis and to define development of peak nitrogenase activity when

the plant has accumulated effective leaf area, and to examine detailed responses of

mature plants to varying environmental conditions and stresses (Dreyfus et al., 1988;

Sprent, 1999; Vitousek at al., 2002).

When conducting an ecosystem field study, it is often important to use the least-

destructive method possible as well as cumulative measures over long periods of time.

For perennial species, especially in frequently-burned ecosystems, removing the

aboveground portion of the plant does not totally remove the individual from the

ecosystem, is similar to burning, and is much less destructive than the soil disturbance









caused by digging for roots and nodules. Additional error can also be introduced due to

the fact that full root and nodule recovery is never assured in field conditions. Thus,

methods utilizing aboveground plant tissues for N-transport product analysis, total N

content and 615N natural abundance may be highly effective for use in long-term field

studies when soil available N is decisively different from the atmospheric isotopic value

(Virginia et al., 1989; Hendricks and Boring, 1999).

Species Differences

Species effects were significant for most of the N2-fixation assessment techniques

used in this study. This strong species effect echoes the great diversity of growth forms

and life histories among the species in this study. Each of the three major subfamilies of

Leguminoseae were represented, Caesalpiniodeae, Papilionoideae, and Mimosoideae.

Two very different growth forms were also represented: spreading to vining plants

(CRRO, CEVI, CLMA, and MIQU) and erect forbs (ORLU, CANI, LEHI, TEVI, and

RHRE). LEHI becomes semi-woody by maturation at the end of a growing season. All

of the species, except for CANI, in this study were perennial, as is common for species

native to fire-adapted ecosystems (Morgan, 1999; Jacobs and Schloeder, 2002).

Rates of nitrogenase activity (C2H4 reduction) increased toward the end of the

growing season, when most of the plants had accumulated substantial photosynthetic

tissues, and had formed active nodules for N2-fixation. CEVI and MIQU show

exceptional nodule activity levels as compared to other legumes in the longleaf-wiregrass

ecosystem, large woody legumes and actinorhyzal N2-fixers (Table 3-3). In contrast, the

other native legumes in this study showed much lower nodule activity levels than the two

leading species, however, they are still within the range represented by other N2-fixers.

The differences in the groups of native species in this study likely indicate that the large









vines, CEVI and MIQU, reached a level of maturity and nodule development that some

of the other species were not able to reach in an initial growing season. The pattern of

acceleration in nitrogenase activity toward the end of the season might begin earlier, and

be more rapid with older field populations of perennial plants that would readily establish

photosynthetic tissues using carbon and N stored belowground from the previous year

(Hendricks and Boring, 1999).

The two dominant constituents of the stem extracted-N, ureide and a-amino-acids,

appear to have an inverse relationship for the species in this study (Table 3-2). The

relationship between these two constituents is different from other studies where NO3,

was a much larger component (3.5 to 35.9%, Sicardi de Mallorca and Izaguirre-Mayoral,

1993; 18 to 70%, Peoples et al., 1996). Those studies had an inverse relationship

between RUC and RNC. In this study, RUC and RAC were not significantly different

among species, however, patterns among species according to RNC did show statistical

differences, and it was elevated for those species with relatively low N2-fixation

capabilities as suggested by other corroborative methods. Those species with the highest

NO3 concentration, although very low compared to values in the other studies (3.5 versus

61.1%), were also among those with the lowest percent N and 615N values in

aboveground tissues in this study indicating lower N2-fixation. Sicardi de Mallorca and

Izaguirre-Mayoral (1993) and Izaguirre-Mayoral et al. (1992) found a similar pattern

when comparing low- to high-N2-fixers. RNC is likely dependent upon soil NO3

availability, which is very low under the typical field conditions where these native

legumes are found (Wilson et al., 1999), and RNC values sampled in the native woodland

could potentially be even lower than those reported in the current study.









The species examined in this study were again divided into two main groups of

fixation activity using 615N values (high: CLMA>CRRO>TEVI>CEVI>MIQU; low:

LEHI>CANI>RHRE>ORLU), with CLMA indicating the least depletion of 15N (Figure

3-3). Other reported 615N values for LEHI and CANI also indicate that these two species

are relatively lower N2-fixers (Hendricks and Boring, 1999). LEHI was not statistically

different from the non-fixing reference in a study conducted by Hendricks and Boring

(1999). The annual CANI was also identified as a low-fixer by the cumulative, 615N

abundance assessment (Figure 3-3). This annual species would have relied on more soil-

N uptake to become established than perennial species that were planted at the beginning

of the growing season, and the RNC value in Table 3-2 indicated that it may be the

species with the greatest rate of NO3 uptake.

Percent Ndfa of the species in this study (12-55%; Figure 3-4A) were lower than

values reported by Hendricks and Boring (1999) for other species of mature legumes

from a similar ecosystem (54-88%). Species order according to % Ndfa is similar to %N,

with the exception of MIQU, which had a large concentration of N, but a lower %Ndfa.

Aboveground percent N also divided the species in this study into two categories of

potential high- and low N2-fixation, those species with >2% N (MIQU, CANI and

CLMA) and those with <2% (CEVI, CRRO, TEVI, ORLU, LEHI and RHRE). Potential

N2-fixation rates indicated by %N agree with the acetylene reduction assay and 615N

techniques for MIQU (high) and LEHI (low). However, comparatively large amounts of

NO3 uptake by CANI and CLMA compared to other species indicate that high levels of

N2-fixation are not related to high N concentrations in the tissues of these species. TEVI,

which had lower percent tissue N in this study than that reported for Hiers et al. (2003),









also failed to flower within the growing season examined here (Chapter 2). TEVI has

been shown to respond to fire with increased flower production (Hiers et al., 2000) and a

significant elevation in percent N (Hendricks and et al., 2003). Thus, for this study,

TEVI was probably lacking the maturity and perhaps stimulation by fire and subsequent

phosphorus enrichment needed for maximum N2-fixation to be detected (Christensen,

1977, Gholz et al., 1985).

Summary

Estimates of N2-fixation capabilities of native legumes in this study were most

conclusive for the two rapidly growing and apparently quickly maturing vine species

CEVI and MIQU. These two species showed strong indices of fixation by both

instantaneous and cumulative measures. The annual CANI and perennial LEHI accreted

large amounts of N in biomass, but apparently from higher NO3 uptake and low rates of

N2 fixation. The data were less conclusive for the remaining, slower-growing, smaller

species that accreted N more slowly and had variable indices of lower but significant

rates of fixation among the methods of assessment. However, these species represent

only a small sample of the 40 species of legumes present in the longleaf wiregrass

ecosystem, which likely also have varying capabilities for N2-fixation (Hainds et al.,

1999). For example, other large, semi-woody legume species of the genera Lespedeza

and Desmodium within this ecosystem have shown higher potential for N2-fixation than

LEHI and should be investigated further (Hendricks and Boring, 1999). With further

investigation of mature plants in the field, using a combination of total N content, 615N,

and N transport product analysis would provide for effective delineation of N2-fixation

for these species and others. Such an approach may provide more information to






63


facilitate using native legumes in restoration plantings for improving soil characteristics,

wildlife habitat and forest productivity.














CHAPTER 4
GROWTH AND N2-FIXATION OF NATIVE LEGUMES IN LONGLEAF PINE
RESTORATION

Introduction

Conservative estimates of nitrogen inputs from biological fixation by native legume

populations in pine woodland and grassland ecosystems have been grossly estimated

from 5.2 to 9 kg N ha-1 yr-1 (Ojima et al., 1990; Hendricks and Boring, 1999). Native

legumes are often found in high density populations across highly-variable light and

water regimes in the longleaf pine- (Pinuspalustris Michx.) wiregrass (Aristida strict

Mill.) ecosystem of Southeastern North America ranging from xeric sandhills to wet-

mesic sites to edges of depressional wetlands (Hainds et al., 1999). However, species of

legumes native to this ecosystem have been reported to have large variability in

symbiotic N2-fixation. Many factors may control N2-fixation rates in woodland

ecosystems, including water and nutrient availability, rhizobium populations, and

especially light, given the requirement for large energetic costs to drive symbiotic N2-

fixation (Sprent, 1987; Vitousek et al., 2002).

Recent restoration initiatives on private and public lands in the southeastern U.S.

coastal plain have resulted in the planting of approximately 283,000ha of former

agricultural, pulpwood plantation and fire suppressed land back into longleaf pine stands

through the USDA Conservation Reserve Program since 1996 (CRP; Coffey and

Kirkman, 2004). These young longleaf stands grow vigorously on sandy, carbon- and N-

depleted former agricultural sites (Markewitz et al., 2002).









Restoration of groundcover in young, planted longleaf pine stands is important for

rebuilding soil organic matter and N, for providing wildlife food and cover, and for

enhancing pyrrhic fuel continuity necessary to reintroduce frequent prescribed fires.

Native grasses and legumes should be preferred over exotics for reintroduction, as has

been made evident by the invasive nature of two species, Lespedeza bicolor, Turcz. and

L. cuneata (Dum. -Cours.) G. Don, that have been previously introduced for soil

improvement and wildlife forage (Miller, 2003). Field research is needed to help

determine desirable native species for reintroduction on targeted sites. There are also

issues related to the timing of groundcover species introduction in young pine stands, due

to the large differences in light transmittance through developing longleaf canopies

between planting and maturation, and following thinning operations associated with

silvicultural practices (Mulligan and Kirkman, 2002).

The objectives of this study were to test for the differences in biomass

accumulation and distribution, N-content, and N2-fixation potential of six legume species

planted in three levels of light under longleaf pine canopies. This field test is an

important step in understanding growth and N2-fixation of native forest legumes under

shaded conditions.

Materials and Methods

Site Description

This common garden study was conducted at Ichauway, a property managed by the

Joseph W. Jones Ecological Research Center (JWJERC), a 12,500 ha reserve located in

Baker County, Georgia, USA (3119'N and 8020'W). The climate for this region is

humid subtropical. Mean daily temperature during the study (10 May-7 November 1004)

was 26.90C, and cumulative rainfall was 595mm. Although most of Ichauway lands were









under longleaf pine savanna when they were acquired in the mid 1930's, some of the land

has been cultivated. Over the years, cultivated areas have been plowed, fertilized with N,

P and K, and planted to crops including cotton (Gossypium spp.) and sorghum (Sorghum

spp.) (Markewitz et al., 2002). The soil at the site was a fine-loamy, kaolinitic, thermic

Typic Kandiudult (Norfolk Series).

The 14 year-old longleaf stand used in this study was planted on a formerly

cultivated area according to CRP stocking recommendations of 1,235 trees ha-1. Three

canopy opening levels were initially located using a densiometer. Plots representing

intermediate light levels (an average of 48% openness) were established in locations

where a single tree had previously died, probably due to fire scorching or insect damage,

leaving a small gap in the canopy. Closed canopy plots (an average of 9% openness)

were located near the intermediate plots in an area with no missing trees. Open canopy

plots were established at the edges of the stand. Trees within approximately 1 m of the

open plots (3 trees per plot) were removed in order to achieve desired openness, and

limbs below 2m height were removed adjacent to all plots for ease of access.

Canopy openness was more accurately quantified using a line quantum sensor on a

clear, cloudless day (Li-Cor, Inc., Lincoln, Nebraska). Five light readings, taken after

pruning and tree removal, were made on 5 May 2005 within each plot and in an adjacent

open field between 12:39 and 15:13 EST. The proportion of light present in plots versus

the open field, expressed as photosynthetically active radiation [amol m-2s1 (PAR) was

used to describe canopy openness.

Experimental Design and Planting

Four replicated 9 m2 plots were arranged according to a completely random

statistical design under the three light environments. Each plot was planted with five









plants from eight different species of 7 month-old native legumes: Centrosema

virginianum (L.) Benth. (CEVI), Desmodium ciliare (Muhl. ex Willd.) DC. (DECI),

Lespedeza angustifolia (Pursh.) Ell. (LEAN), Lespedeza hirta (L.) Hornem. (LEHI),

Mimosa quadrivalvis (L.), Orbexillum lupinelus (Michx.) Isley (ORLU), Pediomelum

canescens (Michx.) Rydb., and Tephrosia virginiana (L.) Pers (TEVI). Nomenclature

follows Wunderlin and Hansen (2003). Seedlings were propagated at the JWJERC by

Dr. K. Kirkman as described in Chapters 2 and 3. Plants were randomly arranged in rows

that were spaced 60 cm apart, with 30 cm spacing between plants. Planting was

completed 10 May 2004. Weeds were suppressed by mulching the plots with on-site

pine straw and supplemented with more from the adjacent area.

Time Domain Reflectometry (TDR) rods were used to monitor soil moisture. 30-

and 90-cm TDR rods were placed in the corners of each plot, and readings were taken

every two weeks. Volumetric soil moisture for each plot across the season did not

indicate any occurrence of severe drought conditions, nor did the plants undergo any

periods of defoliation. Soil moisture was not significantly different among individual

plots or between light treatments. Due to a dry period around the time of planting, plants

were hand watered to aid root establishment (Figure 4-1). Approximately 1L was applied

to each plant on a bi-weekly basis between 18 and 30 May, and once on 11 June.

Plants were destructively harvested from each plot at the end of the growing

season, between 25 October and 7 November, 2004. Plant heights, except for CEVI,

were measured from soil surface to the top of the stem before collection. Approximately

6 L of soil were removed with each plant by digging at a 10 cm radius around each stem







68





-. 0.20

S---- 30cm
S0.18 -- 90cm
E
U
0 0.16-


I-
E 0.14-

S0.12-


2 0.10-
O
0.08


E 0.06


> 0.04-
May Jun Jul Aug Sep Oct

Date
Figure 4-1. Volumetric soil moisture patterns for all plots. Readings were taken every
two weeks between 11 May and 8 October, 2004. Data shown are means + SE









to a depth of approximately 20 cm. Roots and nodules were taken to a field lab, washed

free of soil, and collected. Leaves, stems, roots and nodules were separated, dried to

constant weight at 700C and weighed. Due to the extremely small size of the plants in the

closed light environments, the plant parts were weighed and recorded, but no further

analysis was conducted on this set of samples (Figure 4-2). Leaves and stems from the

plants harvested in the open light environment plots were composite and ground for

further analysis. Plants harvested from the intermediate light environment were separated

into leaves, stems and roots and ground for further analysis. Plants in this light treatment

were also used for a related retranslocation study. Longleaf pine needles were collected

from mid-canopy from trees near the experimental plots for use as a non-N2-fixing

reference for assessment by 615N. Needles were prepared for analysis in the same

manner as the legume tissues.

The usefulness of the 15N natural abundance technique for assessing N2-fixation

under field conditions has been verified through numerous agricultural and controlled

studies. However, the utility of this technique in the field is dependent on comparison of

615N values of a non-N2-fixing reference plant, to represents the ratios of isotopic

nitrogen forms present in the soil with those of the legumes. Ground tissues from

legumes and pine needles were analyzed for 615N natural abundance and total N content

at the University of California, Davis (Stable Isotope Facility, Department of Agronomy,

Davis, CA) using mass spectrometry.

Leaf and stem 615N values of plants harvested from the intermediate light

environment were composite so that comparisons with the aboveground (leaf + stem)

615N values from the open light environment could be made. CEVI and ORLU did not









produce enough samples for this analysis. A weighted 815N value for the plants from the

intermediate plots was calculated:

615Naboveground = [(15Nstem x Total Nstem) + (615N,, x Total N,,, )] / (Total Nstem +
Total N,,, ) [4-1].

Using aboveground tissue values of 615N from intermediate and open light treatments, an

estimate of percent N derived from the atmosphere (%Ndfa) was calculated:

% Ndfa = 1 (615NN2-fixing plant / 615Nref) [4-2].

Statistical Analysis

Data were analyzed using two-way analysis of variance (ANOVA), with species

and light environment as main effects. Where significant effects existed (p<0.05),

Duncan's multiple comparison post-test was used to determine which means differed

significantly. The GLM procedure performed in the Statistical Analysis System (SAS,

2003) was used for ANOVA and post-tests.

Results

Preliminary Results and Survivorship

PAR readings in the open field adjacent to the study site were 1938 10.6 [tmol m

2s1, and plot readings were 370 45, 1214 103, and 1539 61 for closed, intermediate

and open plots, respectively. Canopy openness for each plot type, expressed as a

fraction of adjacent field light levels were 0.085, 0.614, and 0.834 for closed,

intermediate, and open, respectively. Differences in percent canopy openness among plot

types were highly significant (P <0.0001).

There was an average survival rate of 29% across species where LEHI > TEVI >

DECI > LEAN > ORLU. CEVI had the lowest survival of only 15%. An average

sample size of five for each species in each light environment remained except for CEVI









in the intermediate light environment (n=l). M. quadrivalvis and P. canescens were

excluded from further analysis in this study due to poor growth and survivorship. M.

quadrivalvis showed very little change in size at mid-season and final harvest date,

therefore it was not harvested. Other field observations ofM. quadrivalvis suggest that

biomass accumulation may have been directed predominantly belowground. Taproots of

maturing M. quadrivalvis in the field can be as long as 3m (personal observation). As in

a companion study, P. canescens developed brown spots, defoliated and appeared to be

dead by mid-season (Chapters 2 and 3).

Plants grown in the closed light environment accumulated very little biomass over

the season as compared to plants grown in the other light environments (Figure 4-2). Due

to the small growth response of plants in the closed light environment, plants harvested

from this set of plots were not analyzed for N content nor assessed for N2-fixation.

Growth

Total aboveground biomass showed significantly different species and light

treatment effects (Table 4-1, 2). Light effect on biomass was Open > Intermediate >

Closed. However, CEVI growth showed no statistical differences between open and

closed treatments and was not adequately represented for intermediate. LEAN did not

have significantly different greater biomass in the open environment versus the

intermediate, but both were greater than the closed. In the open light treatment, the

biomass was DECI > CEVI > LEHI > TEVI > LEAN > ORLU; only DECI had a

significantly greater biomass than any other species (Figure 4-2). Biomass differences by

species were not statistically different in the intermediate or closed light treatments.

















0)


30
m

g 20


010


0


CEVI DECI LEAN LEHI ORLU TEVI
Species


DECI LEAN LEHI ORLU TEVI


Species


Figure 4-2. Aboveground biomass and change in plant heights from To by species in
each of the three light treatments. Data shown are means + SE, and different
letters within a species represent statistically different means (Duncan's post-
test). Only one plant was harvested for CEVI in the intermediate light
environment.


Open
--- Intermediate
aT Closed
T


a


a b


a


abb b


b

ni






73



Table 4-1. Analysis of variance results for experimental variables. Light treatment and
species were main effects. Values for total aboveground biomass (Total AB
Biomass), root-to-shoot ratio (R/S) and nodule biomass nodulationn) represent
differences across all three light treatments. Results for %N, 615N, Total N,
and %Ndfa represent differences among open and intermediate light
treatments for aboveground tissues.
Main Effects
Variables Species Light Species x Light
Total AB Biomass (g) 0.0094 <0.0001 0.0292
R/S 0.2247 <0.0001 0.2652
Nodulation (g) 0.0149 0.0023 0.5694
%N <0.0001 0.0002 0.0421
6SN <0.0001 <0.0001 0.9063









Table 4-2. Total biomass (above- and belowground tissues, including nodules) per plant
and aboveground values for %N, 615N, and total N. Different letters within
each column represent statistically different means.
Closed Light Treatment
Total Nodule
Biomass biomass
Species (g plant-) R/S (mg plant1)
CEVI 5.38 0.51 98.2
DECI 2.06 1.93 11.22
LEAN 0.52 2.26 6.22
LEHI 1.64 1.79 11.03
ORLU 0.47 1.83 3.74
Intermediate Light Treatment
Total Nodule
Biomass biomass Total N
Species (g plant') R/S (mg plant ) %N '5N (g plant-)
CEVI 2.50 0.46 22.80 -
DECI 7.43 0.90 9.93 1.82b -3.06 0.05
LEAN 2.86 1.07 13.42 1.68b -2.59 0.03
LEHI 3.29 0.76 17.23 1.96b -2.80 0.05
ORLU 0.67 1.65 15.40 -
TEVI 2.25 0.91 29.08 2.79 a -2.20 0.03
Open Light Treatment
Total Nodule
Biomass biomass Total N
Species (g plant) R/S (mg plant1) %N 15N (g plant-)
CEVI 15.59 b 0.89 131.50 a
DECI 37.27 a 0.29 74.63 ab 1.33 b -1.06 0.40 a
LEAN 8.60 b 0.55 37.57 b 1.76b -0.67 0.16 b
LEHI 13.53 b 0.54 18.77 b 1.21 b -1.25 0.13 b
ORLU 1.82 b 1.16 16.00 b -
TEVI 9.65 b 0.58 55.96 b 2.33 a -0.46 0.15 b









Plants grown in the closed light environment were the shortest, overall. The lack of

height growth is apparent when the heights of the plants at harvest are compared with the

heights of the plants at time of planting (To). The largest change in height from To was

found in the open light environment, and the difference between these plant heights and

those in the other treatments were significant overall, with few exceptions for individual

species (Figure 4-2).

Root-to-shoot ratios (R/S) were also significantly different by light environment,

but species effect was not significant (Table 4.1, 4.2). Most of the species showed the

following pattern of significant differences in R/S across the three light environments:

Closed > Intermediate > Open. However, CEVI, a vine, showed an opposite R/S pattern,

Open > Closed. (The intermediate light level was not sufficiently represented.)ORLU

showed no differences in R/S across light treatments, and DECI did not show a

significant difference between the closed and intermediate treatments (Figure 4-3).

Nodulation, as measured by nodule biomass (mg), showed significant light

treatment and species effects. Due to a large degree of variability in nodule biomass, the

only species to show the significant light treatment effect was DECI (p = 0.0532) and

followed the pattern: Open > Closed > Intermediate. General trends of nodule biomass

accumulation for the other species were Open > Intermediate > Closed, although too

variable to show significant differences. In all light treatments, CEVI, TEVI and DECI

appeared to have the greatest nodulation (Table 4-2).

All species had a higher percentage of N concentration in aboveground tissues

(stem + leaves) where grown in the intermediate light environment than in the open.

Although species and treatment effects were significant for %N (Table 4-1), statistical












3.0
Open
IIntermediate a a
2.5 Closed
a
o abb a
o 2.0 -

0
1.5 b a
a ab


bb




0.0 b
b b b b

0.5 b



CEVI DECI LEAN LEHI ORLU TEVI

Species

Figure 4-3. Root-to-shoot ratios by species in the three light treatments. Data shown are
means + SE, and different letters within a species represent statistical
differences (Duncan's post-test). Only one plant was harvested for CEVI in
the intermediate light environment.






77




3.0

Intermediate
2.5 Open


W 2.0

C
c 1.5
E

S 1.0
Z

0.5


0.0
DECI LEAN LEHI TEVI

Species

Figure 4-4. Percent N concentration in aboveground biomass (stem + leaves) by species
in the intermediate and open light environments. Plants grown in the closed
light environment were not analyzed for percent N. Data shown are means of
four plots for each treatment. Statistically greater %N is indicated by *
(Duncan's post-test).









differences in the two light treatments were only represented by DECI (p = 0.0350) and

LEHI (p = 0.0002).

In both light treatments, TEVI had a statistically higher %N than the rest of the

species followed by LEHI > DECI > LEAN for the intermediate treatment and LEAN >

DECI > LEHI for the open treatment (Figure 4-4).

N2-Fixation

Total N in aboveground tissues was significantly affected by light treatment (Table

4-2). However, this difference was only represented by DECI and TEVI, where plants

grown in the open light treatment had statistically greater total N than those grown in the

intermediate. Differences among species in the intermediate light environment were not

statistically different, but in the open light environment they were ranked: DECI >

LEAN > TEVI > LEHI.

Legume 615N values were not statistically different (Figure 4-5) from the non-N2

fixing pine needle reference values in the intermediate light environment (p < 0.1419),

but showed a significant difference from the reference in the open (p < 0.0014). 615N

values of legumes were statistically affected by light environment (p < 0.0001), and this

difference was represented in all of the species. For all legume species in this study, 615N

values indicated that 15N in aboveground tissues was less depleted (closest to atmospheric

value) in those plants grown in the open light treatment than in the intermediate (p =

0.0051 + 0.0044).

Discussion

Shading Effects on Species

Plant growth was significantly greater, overall, in the open light environment than

in the intermediate or closed light environments. The reduced level of PAR in the






















































DECI


LEAN LEHI
Species


TEVI PINE


TEVI


Figure 4-5. 615N and %Ndfa values by species for aboveground tissues in the
intermediate and open light treatments. Data shown are means + SE. Plants
grown in the closed light environment were not analyzed for 615N. Pine
needles collected from adjacent to the plots were used as the non-fixing
reference. %Ndfa was calculated using Equation 4-2.


* Intermediate
o Open


DECI LEAN LEHI
Species


* Intermediate
O Open







0









intermediate (61.4% of incident) and closed (8.5% of incident) light environments

compared to that of the open (83.4% of incident) had significant impact on growth and

N2-fixation rates of the legume species in this study. The change in height between To

and harvest for the species in this study may be due to etiolation and/or actual differences

in growth. Lesser differences in elongation between intermediate and closed treatments

for LEHI and TEVI may be more indicative of etiolation than growth, since biomass was

not different between treatments for these species. DECI and LEAN, which did not show

differences between plant heights in open and intermediate treatments, suggest that they

were less etiolated. LEAN, which had smaller biomass accumulation differences among

treatments showed further evidence that the differences in height between the

intermediate and closed treatments reflected actual growth differences rather than

etiolation.

There were significant differences in R/S between closed and open treatments for

all of the species in this study, suggesting a stronger allocation of carbon to the roots in

low light. Greater allocation to roots due to a variety of environmental stressors,

including shade, drought and fire is a well-documented physiological response (Sprent,

1973; Knapp et al., 1998; Paz, 2003; Fernmandez et al., 2004) Allocation to roots may also

be a response to increased competition from pine roots (Mulligan and Kirkman, 2002)

and is a common adaptation to low-fertility soils (Paz, 2003). In the case of these

legumes, a decreased rate of N2-fixation due to shading, and a subsequent increased

reliance on soil sources of N could lead to a greater allocation of biomass to roots as the

plants forage and compete for N resources.









Smaller, non-significant differences between the R/S of plants grown in the open

versus intermediate light environments may indicate that the threshold light level needed

to initiate these allocation changes exist somewhere between the intermediate and closed

light levels used in this study. R/S reported for CEVI, ORLU and TEVI in this study

(0.89, open; 1.83, intermediate; and 0.91, intermediate, respectively) were much lower

than those reported in a controlled, potted study (approximately 1.5, 2.7, and 3 for CEVI,

ORLU and TEVI, respectively; Chapter 3). R/S of LEHI in the controlled study (0.8)

was similar to the R/S in the intermediate light environment in the current study (0.76).

Sampling differences are the most obvious cause for many of these differences,

especially since CEVI and TEVI, have extensive, lateral branching root systems that were

not harvested in their entirety in this study, but that would have been contained in the

pots of the controlled study.

Dramatically different levels of N2-fixation (as assessed by 615N) in spite of a lack

of difference in nodulation between the open and intermediate light environments (Table

4-2) indicates that nodule activity rather than biomass is more indicative of N2-fixation in

these species. For example, ORLU had a comparable amount of nodulation biomass to

other species in a previous, controlled study, but was determined to have relatively much

lower N2-fixation rates than the other species in the study as assessed by both the

acetylene reduction and 815N natural abundance assessments (Figure 4-4; Chapter 3).

However, statistical differences in nodulation between light treatments were difficult to

detect because of high variability between individual plants due to sampling area in the

field and variable plant sizes. Nodules that were present under shaded conditions, due to









their relative expense versus effectiveness, may be highly susceptible to being sloughed

by the legumes in response to stress (Gadgil, 1971).

Legume N2-fixation under shade conditions has rarely been assessed due to the

high energetic cost of nodule maintenance and the N2-fixation reaction (Vitousek et al.,

2002). A small number of available studies do indicate that nodulation and subsequent

N2-fixation by legumes is inhibited by shading (Gadgil, 1971; Sprent,1973). %Ndfa for

the species grown in open light conditions in this study (44.5-79.7%) are similar to values

reported for similar species by Hendricks and Boring (1999; 54-88%). The agroforestry

legume species Calliandra culhithl, \// and Sesbania sesban also have similar reported

foliar %Ndfa values, 65-90%, respectively (Stahl et al., 2002).

Percent N concentration in leaf and stem tissues was greater in the intermediate

light environment than in the open. The increased concentration of N in the shaded

leaves may reflect the reduced amount of carbon assimilation occurring under lower light

conditions, which is corroborated by the opposite pattern of N accumulation in the open

light treatment. Sun leaves tend to be thicker and contain larger amounts of connective

and suberized tissue leading to a higher lower N concentration than shade leaves (Kramer

and Kozlowski, 1979). However, the larger overall size of the plants in the open light

environment versus the intermediate explains a higher total N value for open-grown

versus more shaded plants. Aboveground %N reported in this study for the large, semi-

woody, LEHI, LEAN and DECI (1.21 1.76%) and for the less-woody, TEVI (2.33 +

0.17%) in this study are within the same range as previously reported leaf %N levels for

similar, semi-woody (1.6 2.3%) and smaller herbaceous legumes (2.4%; Hendricks and

Boring 1992).









Ecological and Management Implications

Survivorship was relatively low in this study, across all species. One possible

explanation for the plants poor survival is the time of transplanting. The small seedlings

were planted in the spring when temperature, photosynthetic rates and associated

transpiration rates were high. A similar planting of legumes conducted in the fall when

temperatures and associated plant processes would have been lowered had a much higher

survival rate (Kirkman, unpublished data). Based on this observation, transplantation

should be conducted in the fall.

815N values indicated a high level of N2-fixation by most of these legume species in

the open light environment. Differences between calculated percent Ndfa for the legumes

(44-79%; Figure 4-5) were slightly lower than those reported by Hendricks and Boring

(1999) from an older population of similar species in pine woodlands (54-88%). Both

studies indicate potentially high growth and N2-fixation rates under relatively open

canopy conditions. However, many questions relating to ineffective nodulation and

species survivorship highlight the need to examine other environmental and biotic

controls on legume N2-fixation, especially drought stress and rhizobium microsymbionts

in older established field populations (Dreyfus et al.,1988).

Additionally, further field studies should examine the continued development of

maturing legumes as this study only followed the first year of establishment. Further

observations of remaining planted legumes from this study and from another continuing

study under the 14 year-old longleaf pine plantation indicate that some of the species that

were excluded from this study did not die, but did not establish during the first growing

season after planting. Specifically Mimosa quadrivalvis, Orbexillum lupinellus and









Pediomelum canescens showed significant growth and maturation during the second year

after planting (Kirkman, unpublished data).

The native legumes planted under the closed canopy of this 14 year-old longleaf

pine plantation showed very little capacity for growth, N2-fixation, or nitrogen accretion.

Dense litter accumulation under this closed canopy can also heavily suppress the growth

of other groundcover species, including wiregrass or Rubus sp. (blackberry) (Mulligan

and Kirkman, 2002). However, mulching with pine litter could be an important tool to

selectively encourage establishment of specific groundcover species under more open

canopy conditions.

Growth and N2-fixation of the legumes increased dramatically with increased

canopy opening, advancing among the closed (8.5% of incident light), intermediate

(61.4% of incident) and open (83.4% of incident) treatments. This pattern suggests that

substantial legume populations, from the perspectives of wildlife food production and

cover, and N contribution to the soil, will not be likely to grow or persist beneath dense

young pine plantations until the trees are operationally thinned to a lower density. Even

the relatively low stocking of pines in these plantations intended for enhancing wildlife

habitat results in dense canopy development and little capacity for supporting native

grasses and legumes. The findings from this study points to a recommendation that

landowners plant legumes beneath plantations that have undergone an initial thinning

after about 15-20 years of tree growth. At that point, light conditions should be more

favorable for legume populations to grow and contribute a significant amount of fixed-N

to the N and C depleted soils that are typical on converted agricultural sites.














CHAPTER 5
CONCLUSION

Conclusions from the Current Study

Introduction

The fire-dependent longleaf pine- (Pinuspalustris Mill.) wiregrass (Aristida strict,

Michx.) savanna ecosystem once dominated the southern coastal plain of the United

States, covering as much as 37.2 million ha. (Landers et al., 1995). Recent restoration

initiatives on private and public lands in the southeastern U.S. coastal plain have resulted

in the planting of approximately 283,000ha of former agricultural, pulpwood plantation

and fire suppressed land back into longleaf pine stands through the USDA Conservation

Reserve Program since 1996 (CRP; Coffey and Kirkman, 2004). Groundcover

reestablishment in these stands is key to improving wildlife habitat and for restoring a

continuity of pyrrhic fuels for frequent prescribed burning (Clewell, 1989; Kirkman,

2002). Legumes may also have a major role in maintaining N balance of these restored

systems. Shade tolerance of groundcover species to be potentially reintroduced under

young pine stands needs to be assessed before large-scale operations are undertaken

(Mulligan and Kirkman, 2002).

Objectives

The overall objectives of this study were (1) to explore the impact of various

degrees of shading on relative growth and N2-fixation rates of legume species native to

longleaf pine-wiregrass savannas, (2) to make initial observations of phenological

development and nodule morphology for each species, and (3) to examine the









effectiveness of corroborative methods for assessing N2-fixation. These objectives have

not been previously addressed for most of these species. Controlled potted studies and a

common garden experiment were used to assess species responses to shading under

potted and field conditions.

The species examined in this study represented all three common subfamilies of the

Leguminoseae, Caesalpiniodeae, Mimosoideae, and Papilionoideae, and represented

three distinct growth forms: vines, erect herbs, and semi-woody erect herbs. Nodule

morphology differences further confirmed the species subfamily designations. Each

species had a unique response to shade, and each one represents a slightly different life

history strategy that is apparent in growth habit and phenological development (Chapter

2). Adaptations to fire, low-fertility soils, drought conditions, and a highly variable, but

relatively open overstory canopy structure are manifested in the stress-tolerating and

opportunistic nature of most of these legume species (Hainds et al., 1999).

N2 fixed by legumes, although important to many ecosystems, is difficult to

quantify. Due to the specific limitations associated with each assessment technique,

corroborative assessments should be made in any study that seeks to quantify fixed-N. In

this study, relatively higher levels of nitrogenase activity (acetylene reduction) for a

species were corroborated by reduced soil NO3 utilization, 615N signatures nearer

atmospheric values, and higher tissue-N concentrations (Chapter 3). Even in a potted

study with complete nodule recovery, nodule biomass was not a good predictor of N2-

fixing ability, since only one species had significantly different biomass from all of the

others.









N2-fixing capabilities were species dependent. Some of the species in the potted

study did not appear to have fully developed during this short-term study. For example,

Tephrosia virginiana, which was indicated as one of the strongest N2-fixers in the

common garden study (Chapter 4), did not reach a reproductive stage, nor was it very

active in N2-fixation in the potted study (Chapter 3). Additionally, Orbexillum lupinellus,

which appeared to be well-nodulated in the potted study, was not active in N2-fixation.

This ineffective nodulation may indicate a slowly-developing plant, or it may indicate

that the preferred symbiont for this species was not present (Chapter 3).

Field studies following more controlled studies can confirm initial findings and

give more realistic estimates of legume N2-fixing capacities (Dreyfus et al., 1988). The

common garden experiment, which employed three levels of canopy opening, was

designed to make such estimates of N2-fixation capabilities under field conditions. Shade

not only impacted the biomass and tissue-N allocation patterns of the legumes in the

study, but was effective in producing light levels that provide critical limitations to

growth and N2-fixation rates (Chapter 4). This study also showed that Lespedeza spp.,

Desmodium spp. and Tephrosia virginiana are potentially very good candidates for

inclusion in groundcover restoration projects due to their high tissue-N concentrations,

N2-fixation capabilities, and dominant growth forms that will provide adequate cover for

wildlife. These findings should be considered when landowners make decisions about

groundcover restoration. The final conclusion of this field study was that native legumes

will not thrive or fix N2 under a closed canopy and should be established in more open

stands that have been recently thinned.









Directions for Future Research

This study examined only a small percentage of the legume population native to the

longleaf pine-wiregrass ecosystem. Additional species should be considered for

groundcover restoration, especially other species of the genera Desmodium and

Lespedeza. Since the legumes predominantly represent a fall food source for bobwhite

quail, perhaps other fall-flowering legumes should be considered for groundcover, such

as Dalea spp. (Stoddard, 1931).

Further Application of N2-Fixation Assessment Techniques

Of the N2-fixation assessment techniques used in this study, only the N-transport

and storage product analysis has not been used for field analysis of species in southern

pine ecosystems. The portability and minimally-destructive nature of this technique may

have great potential for field studies in the longleaf pine-wiregrass ecosystem, and may

be especially valuable for a first-assessment of N2-fixing capabilities of numerous species

of legumes under field conditions when used in corroboration with the 15N natural

abundance technique.

Another interesting approach to quantifying N2-fixation using the stem N

transport/storage product analysis (relative ureide analysis) involves calibrating ureide

indices with fixed-N2 using a series of labeled 15N2 assays and relative ureide analysis

(Peoples et al., 1996). This method may have potential for allowing an estimate of N2-

fixed by a population of legumes with a quick and easy analysis that is easy to conduct on

a variety of soils.

Further studies should also investigate the relative efficiencies of specific host-

symbiont interactions, including tripartite interactions between mycorrhizae, rhizobium,

and legumes. Relative efficiencies can be quite different between rhizobium strains, and









inoculation with highly effective strains could become important to native legume

restoration if the land on which they are being planted has had heavy herbicide usage.

Future Research for Native Legume Utilization

Future research regarding groundcover restoration under longleaf pine should

involve studies that investigate physiological properties of these and many other of the

legumes native to the longleaf-wiregrass ecosystem and continue to ask questions relating

directly to groundcover restoration objectives such as soil-N and organic material

development and wildlife food quality. Additional studies should look at competitive

interactions between native grasses and legumes planted in a restoration setting to

determine when competition for soil and light resources becomes limiting for each type

of plant. Fine root turnover, mycorrizhal associations, root exudates, and other

belowground processes that contribute directly to building soil organic matter and N-

availability should be further researched. Analysis of native legume and grass shoot and

seed digestibility and nutritional composition could be important information for land

managers seeking to attract and sustain wildlife populations on their property.

Imported legumes have been used for soil stabilization and game management

purposes, and many of these species, including kudzu (Pueraria lobata), sericea

lespedeza (Lespedeza cuneata) and shrubby lespedeza (L. bicolor), out-compete native

vegetation in natural and roadside areas across the Southeast (Miller, 2003). Native

legumes are a better choice for many soil preservation and land management purposes

and are much less likely to become invasive. Many of the legumes native to the

southeast region, such as Centrosema virginianum and Daleapinnata, to name only a

few, are also aesthetically pleasing and should be considered for wildflower plantings.

The N2-fixation capability of C. virginianum could possibly be marketed as an added









benefit to native plant collectors. More research is needed before other species can be

designated as N2-fixers.

Native legumes are important and poorly explored genetic resources. N is typically

the most limiting plant nutrient in terrestrial ecosystems, and biological N2-fixation is a

"free" source of N-addition. In a time when organic farming practices are gaining

popularity in the temperate parts of the world and agroforestry is increasing in feasibility

and scientific recognition in tropical and subtropical parts of the world, legumes are

becoming more valuable and marketable. The legume species-richness resources that are

available in the Southeastern coastal plain should be actively examined for future

development.