ground layer carbon and nitrogen cycling and legume nitrogen inputs ...

American Journal of Botany 93(1): 84–93. 2006.

GROUND

LAYER CARBON AND NITROGEN CYCLING AND

LEGUME NITROGEN INPUTS FOLLOWING FIRE IN MIXED PINE FORESTS1

SARA D. LAJEUNESSE,2,5 JOHN J. DILUSTRO,3,6 REBECCA R. SHARITZ,2,3,4 AND BEVERLY S. COLLINS3 Department of Plant Biology, University of Georgia, Athens, Georgia 30602 USA; and 3Savannah River Ecology Laboratory, P. O. Drawer E, Aiken, South Carolina 29802 USA

2

Many mixed pine forests in the southeastern United States undergo prescribed burning to promote open pine savannas. In these systems, soil texture can influence fire’s effect on vegetation and nutrient cycling. Our objectives were to examine fire and soil texture effects on carbon (C) and nitrogen (N) pools in ground layer vegetation. We measured biomass and tissue nutrient concentrations and estimated legume N inputs via N2 fixation in frequently burned sandy and clayey sites that were in the first and second seasons following a prescribed fire in 2002 (B02) or had been unburned since 2000 (B00). Mean belowground biomass was significantly greater on sandy than on clayey sites. Total aboveground mean biomass did not differ significantly between B00 and B02 sites, but grasses had greater aboveground biomass in clayey than in sandy sites. Carbon and N pools (measured in grams per square meter) in grasses were greater in clayey than in sandy sites, yet grasses had greater tissue concentrations of C (as a percentage) in sandy sites. Legumes showed significant interaction effects between soil texture and fire frequency for tissue C and N pools, above- and belowground biomass, and acetylene reduction activities. Results suggest that soil texture can influence fire effects on ground layer vegetation in southeastern mixed pine forests. Key words:

fire frequency; ground layer biomass; herbaceous legumes; mixed pine forests; nitrogen; soil texture.

Prescribed fire has been widely implemented across the southeastern United States (Boyer and Miller, 1994; Provencher et al., 2001; Wilson et al., 2002), especially on public lands undergoing restoration or management of the longleaf pine ecosystem as habitat for the federally endangered redcockaded woodpecker (Picoides borealis Vieillot) and on private lands managed for wildlife and recreation. Fire can, in the short term, increase standing crop biomass of herbaceous species, particularly in grasslands and pine savannas (Raison, 1979; Schoch and Binkley, 1986; Dudley and Lajtha, 1993; Brockway and Lewis, 1997; Wilson et al., 2002). Grasses and legumes, especially, are stimulated by fire and can dominate ground layer communities of longleaf pine forests burned annually or biennially (Brockway and Lewis, 1997; Glitzensein et al., 2003). Grasses, among other plant groups, may respond favorably to fire because, in addition to producing copious seeds, they are capable of storing nutrients in belowground structures and resprouting from protected meristems following fire (Brockway and Lewis, 1997). An increase in biomass and relative abundance of legumes following fire (Cushwa et al., 1966; Van Lear and Johnson,

1983; Hendricks and Boring, 1992, 1999; Dudley and Lajtha, 1993; Hiers et al., 2000; Caldwell et al., 2002) may be due to their ability to fix nitrogen (N2) and thus colonize disturbed and potentially nitrogen-deficient sites (Towne and Knapp, 1996). The removal of ground layer plants by burning is ideal for germination of many legume species (whose seeds often require scarification by fire) because competition for light, nutrients, and water is reduced (Hendricks and Boring, 1999). Chemical cues from smoke may also promote germination (Flematti et al., 2004). In addition, many legumes are perennials that resprout vigorously from underground structures following fire (Towne and Knapp, 1996). In contrast, frequent fire is thought to result in a decrease in biomass of woody ground layer species (Reich et al., 2001; Glitzenstein et al., 2003) that may compete with herbaceous plants. In addition to altering species composition, fire can alter nutrient cycles in both the short and long term and indirectly impact productivity of ground layer vegetation. A short-term nutrient pulse following fire (Reich et al., 2001), as nutrients immobilized in litter and soil organic matter are released (Brockway and Lewis, 1997), may lead to increased productivity. However, following fire, nutrients may be in an organic form and N availability to plants may be controlled by rates of net N mineralization (Christensen, 1977; Liu and Muller, 1993; Robertson et al., 1999). In southeastern pine forests, soil moisture can be inversely related to net N mineralization (Wilson et al., 1999), and ground layer biomass can be greatest where N mineralization is lower, as in wet mesic longleaf pine sites (Mitchell et al., 1999). Thus, short-term pulses of nutrients following fire may or may not result in greater biomass of ground layer vegetation. In a meta-analysis addressing short-term dynamics of N pools and fluxes, Wan et al. (2001) found increases in N availability, but no detectable changes in pool size. Further, the authors suggest the need for long-term evaluation of the effects

Manuscript received 3 September 2004; revision accepted 7 October 2005. The authors thank L. Duncan, A. Boretsky, M. Tomblin, N. White, M. Johnston, J. Hupp, and T. LaJeunesse for field assistance in vegetation plot sampling; P. Hartel and B. Haines for assistance with gas chromatography; H. Westbury for logistical help at Fort Benning; and H. Balbach and two anonymous readers for reviewing the manuscript. Funding was provided by the Strategic Environmental Research and Development Program (SERDP), funded by the Department of Defense, as part of the SERDP Ecosystem Management Project (SEMP) CS-1114E. Manuscript preparation was supported by Financial Assistance Award Number DE-FC09-96SR18546 between the U.S. Department of Energy and the University of Georgia. 4 Author for correspondence (e-mail: [email protected]) 5 Current address: Everglades National Park, South Florida Ecosystem Office, 950 N. Krome Avenue 3rd Floor, Homestead, Florida 33030 USA. 6 Current address: Department of Biology, Chowan College, 200 Jones Drive, Murfreesboro, North Carolina 27855 USA. 1

84

January 2006]

LAJEUNESSE

ET AL.—GROUND LAYER CARBON AND NITROGEN CYCLING

of fire on ecosystem N. Over the longer term (on the scale of decades), frequent fire may result in a loss of N from ecosystems (Raison, 1979; Ojima et al., 1994; Reich et al., 2001). Repeated N loss through combustion, volatilization, and leaching (Van Lear and Johnson, 1983; Hainds et al., 1999; Neary et al., 1999) can exacerbate N limitation, which is common in temperate systems (Vitousek, 1982; Pastor et al., 1984; Reich et al., 1997). Though longer-term experiments evaluating frequent burning are not common, declines in tree productivity (Boyer, 2000) and N cycling (Wright and Hart, 1997) have been observed. At this longer time scale, N2 fixation by native herbaceous legumes may play an important role in the N cycle of burned forests by replenishing or even exceeding N lost due to fire (Hendricks and Boring, 1992, 1999; Johnson and Curtis, 2001; Caldwell et al., 2002; Hiers et al., 2003). However, the extent of N2 fixation by legumes in southeastern forests has been examined primarily in certain forest types such as pine plantations or pine savannas (Hendricks and Boring, 1999; Hiers et al., 2003). Further, interactions between fire effects on legume composition and N inputs via fixation and soil texture have not been addressed over the full range of southeastern forest types currently undergoing prescribed burning. We examined the effects of fire and soil texture on biomass of plants, including legumes, and carbon (C) and N pools in the ground layer vegetation of southeastern mixed pine forests. We measured above- and belowground biomass and tissue C and N concentrations of ground layer plants in sandy and clayey sites that were in the first and second (B02 sites) or third and fourth (B00 sites) seasons following a prescribed fire. In addition, we assayed N2 fixation rates of many herbaceous legume species in the same sites, using a one-time, mid-year acetylene reduction method. MATERIALS AND METHODS Site description—This research was part of a larger investigation focusing on the effects of land use disturbance (by prescribed fire and military training) on mixed pine hardwood forest communities at Fort Benning, Georgia, USA (328339 N, 858 W) (Dilustro et al., 2002). Fort Benning, the U.S. Army Infantry Center, occupies 73 650 ha in the Piedmont-Coastal Plain Fall Line Sandhills ecoregion (Keys, 1995). Historical land use has included row-crop farming and grazing, and military land use began in 1920 (U.S. Army Infantry Center [USAIC], Fort Benning, Integrated natural resources management plan, unpublished manuscript). The climate is humid temperate; most summer days reach .32.28C and mean low winter temperature is 2.88C. Mean annual precipitation is 129.5 cm (USAIC, unpublished manuscript). Precipitation in 2002 was below average for all months, but was average in 2003 (Fig. 1). Soils include Troup sandy loams, Nankin sandy loams, Lakeland sands, and Siley loamy sands (Natural Resources Conservation Service, 1983). Forests at Fort Benning are managed with prescribed fire to promote longleaf pine savannas, preferred habitat of the federally endangered red-cockaded woodpecker, on appropriate sites (USAIC, unpublished manuscript). Most areas have been thinned on a 9-yr rotation and burned on a 3-yr rotation over the last 20 yr. The 16 stands (400 3 400 m) we selected were winter-burned prior to the 2000 growing season. Half the stands (B02 sites) had the fire cycle accelerated to 2 yr and were burned prior to the 2002 growing season. Thus, they were in the first season following fire when our research began in summer 2002. The other half (B00 sites) had burning delayed to 4 yr; thus, they were in the third season following fire at the start of this research. In addition, half the selected stands are on sandy soils (,13% clay) and half are on clayey soils (.19% clay). Most of the 16 stands have overstories dominated by loblolly pine (Pinus taeda L.) and longleaf pine (Pinus palustris Miller), with some sites containing shortleaf pine (Pinus echinata Miller), tulip poplar (Liriodendron tulipi-

85

Fig. 1. Total and 30-yr mean monthly rainfall in mixed pine forest stands in southwest Georgia, USA, from 2001 through 2003. fera L.), oaks (Quercus spp.), and hickories (Carya spp.). Ground layer species include Andropogon spp., Vitis spp., Heterotheca graminifolia (Michaux) Shinners, Pteridium aquilinum (L.) Kuhn, sprouts and seedlings of Liquidambar styraciflua L., and numerous legume species. Study design—A 100 3 100 m plot, with five transects spaced at 20-m intervals to form a sampling grid, was established in the center of each of the 16 400 3 400 m stands. Subplots were established at 12 randomly chosen grid intersections in each plot for the sampling described next. Vegetation sampling—Ground layer vegetation was sampled in July and early August in both 2002 and 2003 because these months represent the peak biomass of herbaceous plants. To obtain estimates of ground layer biomass, a 0.64-m2 circular vegetation plot (0.58 m2 in 2003) was randomly placed near the center of each of the 12 subplots (12 subplots 3 16 sites 5 192 total vegetation plots sampled per year). Vegetation was clipped at ground level and sorted into categories (ferns, grasses, legumes, other forbs, woody plants ,1 cm stem diameter at base, and standing dead biomass). In 2003, the harvested legumes were identified to the species level (nomenclature follows that of Radford et al., 1968). All aboveground plant material was dried to constant mass at 608C and weighed. In 2002, randomly chosen subsamples of the forbs, grasses, and legumes were ground in a Wiley and Spex mill and analyzed on a Carlo Erba NA 1500 elemental analyzer (Carlo Erba, Milan, Italy) for C and N content. In 2003, C and N analyses were repeated, but only for the legumes, which were analyzed by species. Carbon and N tissue concentrations (as percentages) of forbs, grasses, and legumes were converted to stand-level biomass estimates of C and N pools (i.e., standing crop in grams per square meter) for these plant groups by multiplying the percentage concentration by the biomass. Belowground biomass was measured by taking a 5-cm-diameter soil core to a 15-cm depth within each circular vegetation plot to represent belowground biomass at the stand level. In July 2003, roots and root nodules (on the legumes) were removed from the soil by hand and washed with water over a 2-mm sieve. Root and nodule mass were determined by drying them at 708C for 24 h and weighing them. Root samples were then ground and analyzed for C and N content as described. Acetylene reduction—In July 2003, we used a one-time, mid-year acetylene reduction assay (Myrold et al., 1999) to estimate N2 fixation by herbaceous legumes from eight sites (two sites from each of the following: B02/clay, B02/sand, B00/clay, B00/sand). The three most abundant legume species in each site at the time of assay were selected for N2 fixation measurements. These species were Cassia nictitans L., Desmodium marilandicum (L.) DC,

86

AMERICAN JOURNAL

Desmodium paniculatum (L.) DC, Desmodium viridiflorum (L.) DC, Lespedeza hirta (L.) Hornemann, and Tephrosia virginiana (L.) Persoon. In addition to the acetylene reduction assay, aboveground, belowground, and nodule biomass were obtained to estimate ecosystem-level N2 fixation for the month of July by these herbaceous legumes. The acetylene reduction assays were performed from 14 to 18 July 2003. Ten individuals of each species were chosen at random and excavated (30 cm radius 3 15 cm depth) (Hendricks and Boring, 1999). Small (;2.5 cm) fragments of nodulated roots were excised, placed into 10-mL glass test tubes, and capped with rubber serum stoppers. Acetylene generated from calcium carbide and stored in a gas sampling bag was injected into the test tubes (;10% of atmosphere), and the samples were allowed to incubate for 30 min. The reaction was terminated by transferring the samples to 10-mL Vacutainer tubes (Becton, Dickinson and Co., Franklin Lakes, New Jersey, USA). Subsamples (1 mL) from the Vacutainers were analyzed on a gas chromatograph (Varian, Palo Alto, California, USA), equipped with a flame ionization detector. Operating parameters included a detector temperature of 1608C, an injector temperature of 1358C, 18 mL/min helium carrier gas flow rate, and a stainless steel column packed with Porapak N (Alltech Associates, Deerfield, Illinois, USA). An ethylene standard was obtained by injecting 1 mL of pure ethylene into a 1-L glass jar fitted with a rubber septum. The Vacutainers gave off a small amount of ethylene contamination, which was subtracted to obtain the final acetylene reducing activity (ARA; in nanomoles per hour per gram nodule dry mass). Following this assay, the nodules were dried at 708C to a constant mass and weighed. Ten additional individuals of each legume species were randomly chosen from each site and excavated for estimation of biomass. These plants were separated into above- and belowground material and nodules, dried at 708C, and weighed. This biomass (aboveground and nodule) was used to predict the nodule biomass of legumes of the same species from the 2003 ground layer biomass harvests (circular vegetation plots) according to the following: aboveground biomass per plant (g)/nodule biomass (g) 5 aboveground biomass (circular plots, g/m2)/predicted nodule biomass (circular plots, g/m2). Although the ratio of acetylene reduced to N2 fixed can vary depending on the efficiency of the symbiont’s dehydrogenase system, it was not within the scope of this project to determine a more precise ratio. Therefore, the often-used ratio of 3 : 1 was employed to estimate an ecosystem-level contribution of N by legumes (in grams per square meter per hour) for July by multiplying the predicted nodule biomass for each species by its N2 fixation rate (based on ARA values). An overall value for N2 fixation in July was estimated by using the mean ARA value of the three Desmodium species examined to estimate N2 fixation rates for other Desmodium species found in the vegetation plots; the ARA value for L. hirta was used to estimate N2 fixation rates for other Lespedeza species; and the ARA for T. virginiana was used to estimate N2 fixation rates for other Tephrosia species. Finally, the mean ARA value for all six legume species examined was used to assay standlevel N2 fixation rates for the month of July for the remaining eight legume genera found in the vegetation plots. The accuracy of the acetylene reduction assay has been criticized as underestimating nitrogenase activity due to the disturbance of nodules during the assay and an acetylene-induced decline in nitrogenase activity, which occurs within the first 30–60 min of root exposure to acetylene (Minchin, 1983, 1986). In addition, nitrogenase activity can vary throughout the year (Myrold et al., 1999). We assayed N2 fixation during July when many legumes are at peak biomass, and our estimate represents N2 fixation for that month only. In an effort to identify long-term integration of N2 fixation in these stands, we attempted to use a 15N natural abundance technique (Hendricks and Boring, 1999). This technique yielded a high degree of variability in 15N abundance of reference plants (ranging from 20.89 to 23.01 d15N) across stands, which made it impractical for our purposes. This may be due to the varied disturbance history of these forest stands. Thus, we used the acetylene reduction method as a somewhat less accurate alternative to assay N2 fixation. Although the acetylene reduction assay cannot directly measure total nitrogenase activity, it is still considered a valuable measure of relative nitrogenase activity (Vessey, 1994). In the future, additional studies using a direct isotope dilution technique may have greater potential on these sites.

OF

BOTANY

[Vol. 93

Fig. 2. Total aboveground biomass of pooled ground layer plants with both sample years combined (does not include standing dead biomass) on sites with different (A) soil textures and (B) fire frequencies (mean 1 SE). Similar letters above bars within each year indicate no significant differences at a 5 0.05.

Data analysis—Differences in mean aboveground biomass between fire treatments (B02, B00) and soil texture (sandy, clayey) were determined using a mixed model analysis of variance (ANOVA) with year, site, fire, and texture as fixed effects and subplot as a random effect (PROC MIXED; SAS, 2000). Statistically significant differences between plant groups were determined using the Bonferroni test of multiple pairwise comparisons. Differences in belowground biomass, plant C and N tissue concentrations, estimates of legume N2 fixation rates, and above- and belowground biomass were determined using ANOVA (PROC GLM; SAS, 2000). Statistically significant differences were accepted at a # 0.05.

RESULTS Biomass and tissue C and N concentrations—With plant groups and sample years pooled, mean aboveground biomass did not differ between the B02 and B00 sites (F1,501 5 0.12, P 5 0.73 and F1,490 5 0.27, P 5 0.61 for 2002 and 2003, respectively) or between soil textures (F1,501 5 3.34, P 5 0.07 and F1,490 5 0.00, P 5 0.98 for 2002 and 2003, respectively) in either year. However, clayey sites tended to have greater biomass than sandy sites and B02 sites had greater biomass than B00 sites (Fig. 2A, B). On average, for both 2002 and 2003, woody species made up the greatest proportion of the total ground layer biomass (32%), followed by forbs (28%), grasses (21%), standing dead biomass (13%), legumes (4%), and ferns (2%). With plant groups considered separately, grasses had significantly greater biomass in clayey than in sandy sites in both years (F1,174 5 4.77, P 5 0.03; Fig. 3).

January 2006]

LAJEUNESSE


87

Fig. 3. Aboveground biomass for ground layer plant groups in each sample year (mean 1 SE). Means within a plant group that have different letters were significantly different at a 5 0.05.

Biomass of the other plant groups did not differ between fire treatments or soil textures. In contrast to the aboveground biomass measure, which included only ground layer vegetation, belowground biomass included ground layer roots as well as tree and shrub roots and was a measure at the stand level. Mean belowground biomass was significantly greater on sandy than on clayey sites in both

Fig. 4. Belowground biomass on sites with different (A) soil textures and (B) fire frequencies (mean 1 SE). Means for the same year with different letters were significantly different at a 5 0.05.

2002 (F1,187 5 7.59, P 5 0.007) and 2003 (F1,188 5 4.92, P 5 0.03; Fig. 4A), but did not differ statistically between B02 and B00 burn sites (2002, F1,187 5 0.00, P 5 0.97; 2003, F1,188 5 0.28, P 5 0.60; Fig. 4B). In 2002, aboveground tissues of grasses had significantly greater concentrations of C (always reported as a percentage) on sandy than on clayey sites (F1,105 5 12.48, P 5 0.0006; Fig. 5B), but there were no differences in concentrations of N (F1,105 5 2.70, P 5 0.10). However, standing crops of N and C (always reported in grams per square meter) in grasses were greater on clayey than on sandy sites (N, F1,103 5 6.67, P 5 0.0112; C, F1,103 5 6.65, P 5 0.01; Fig. 5C, D). Legume aboveground standing crops of N and C showed significant interaction effects between soil texture and fire (N, F1,124 5 17.41, P , 0.0001; C, F1,124 5 15.10, P 5 0.0002; data not shown). Both legume N and C standing crops were greater in B02/clay sites and B00/sand sites (Fig. 5C, D). Although legume tissue concentrations of N did not differ between fire treatments (F1,127 5 1.66, P 5 0.20) or soil textures (F1,127 5 2.21, P 5 0.14), they were higher than those of grasses and forbs (Fig. 5A). There were no significant differences in C and N tissue concentrations or standing crops in forbs between soil textures or fire treatments (P . 0.05 for all comparisons). Similarly, C and N tissue concentrations in woody leaves and stems did not differ statistically between soil textures or fire treatments (P . 0.05 for all comparisons, data not shown). Finally, roots had significantly greater standing crops of C and N in sandy than in clayey sites (C, F1,96 5 7.83, P 5 0.0062; N, F1,96 5 6.54, P 5 0.01; Table 1). In 2003, legume aboveground tissue N concentrations again showed a significant interaction effect between soil texture and fire (F1,330 5 7.29, P 5 0.007; data not shown) as did standing crops of N and C (N, F1,330 5 18.89, P , 0.0001; C, F1,330 5 21.51, P , 0.0001; data not shown). These interaction patterns were the same as for legumes in 2002, with greater tissue nutrients on B02/clay sites and on B00/sand sites. Cassia nictitans had significantly greater N tissue concentrations in B00 than in B02 burn sites (F1,295 5.60, P 5 0.02; Fig. 6A) and significantly greater standing crops of C and N on sandy than on clayey sites (C, F1,29 5 6.92, P 5 0.01; N, F1,29 5 5.67, P

88

AMERICAN JOURNAL

OF

BOTANY

[Vol. 93

Fig. 5. Aboveground tissue (A) N concentration, (B) C concentration, (C) N standing crop, and (D) C standing crop in forbs, grasses, and legumes in 2002 (mean 1 SE). Means for the same plant groups with different letters were significantly different at a 5 0.05. * designates a significant interaction effect.

5 0.02; Fig. 6C, D). Desmodium marilandicum showed a significant interaction effect between soil texture and fire for C and N standing crops (C, F1,14 5 7.16, P 5 0.02; N, F1,14 5 8.04, P 5 0.01). Carbon and N standing crops were greater on both clayey and sandy sites that burned in 2002; however, the magnitude of the difference was greater on sandy than on clayey sites (Fig. 6C, D). Desmodium paniculatum had significantly greater tissue concentrations of N in B00 than in B02 burn sites (F1,15 5 7.95, P 5 0.01; Fig. 6A) but was not observed in B02/sand sites. Finally, D. viridiflorum and L. hirta did not have any significant differences in tissue C and N concentrations or standing crops between soil textures or fire treatments (D. viridiflorum, F1,2 5 7.27–0.17, P 5 0.11–0.72; L. hirta, F1,30 5 1.87–0.01, P 5 0.29–0.94; Fig. 6). Tephrosia virginiana was not present in B02/clay sites or in B00/sand sites, so no differences between burn treatments or soil textures could be examined (Fig. 6). N2 fixation assay and biomass of legumes—A significant interaction effect between soil texture and fire was found for TABLE 1. C and N concentrations in belowground biomass in 2002, Fort Benning, west-central Georgia, USA. Treatment

Mean N (g/m2)

%N

Mean C (g/m2)

%C

B02 B00 Clay Sand

0.29 0.26 0.23* 0.32*

0.49 0.48 0.52 0.46

22.19 22.96 17.29* 27.45*

35.04 36.74 36.87 34.96

* Significant at a # 0.05.

legume acetylene reduction activity (ARA, in nanomoles per hour per nodule dry mass) (F1,193 5 27.11, P , 0.0001). In clayey sites, ARA (all six species combined) was greater in stands that were burned in 2002 (B02); whereas, in sandy sites, ARA was greater in stands that had not been burned since 2000 (B00) (Table 2). Cassia nictitans also showed a significant interaction effect between fire and soil texture and the pattern was the same as for legumes overall (F1,76 5 11.13, P 5 0.0013; Table 2). None of the other species differed significantly between fire treatments or soil textures. In contrast to the legumes harvested in the vegetation plots in 2002 and 2003 (Fig. 3), aboveground and belowground biomass of these six legume species combined exhibited significant interaction effects between soil texture and fire (aboveground F1,230 5 35.94, P , 0.0001; belowground F1,230 5 14.49, P 5 0.0002), with higher biomass in B02/clay sites and in B00/sand sites. Cassia nictitans had a significantly greater belowground biomass in sandy than in clayey sites (F1,76 5 11.07, P 5 0.0014; Fig. 7B). The sample size of D. marilandicum was not large enough in B00/sand sites to examine interaction effects, but the main effects of soil texture and fire both showed significant differences. Aboveground biomass of D. marilandicum was greater in B02 sites (F1,55 5 7.11, P 5 0.01) and clayey sites (F1,55 5 12.18, P 5 0.001; Fig. 7A), but belowground biomass did not differ significantly (Fig. 7B). Lespedeza hirta had the opposite pattern, with greater aboveground biomass in B00 sites (F1,17 5 10.88, P 5 0.0042; Fig. 7A). Finally, T. virginiana had greater belowground biomass in B02 sites (F1,15 5 10.45, P 5 0.0056; Fig. 7B), but occurred in such small quantities in sandy sites that it was not harvested.

January 2006]

LAJEUNESSE


89

Fig. 6. Aboveground tissue (A) N concentration, (B) C concentration, (C) N standing crop, and (D) C standing crop in legumes in 2003 (mean 1 SE). Means for the same species with different letters were significantly different at a 5 0.05. Missing error bars indicate a sample size of one individual.

The other legume species did not have any significant differences between soil textures or fire treatments. An ecosystem-level estimate of N2 fixation rate for July was determined using the ARA for the six legume species examined and aboveground biomass from the circular vegetation plots (Tables 2, 3). Cassia nictitans had the greatest N2 fixation rate (in grams per square meter per hour) in B00/clay sites (Table 2). Although C. nictitans’ ARA was relatively similar across sites, its nodule to aboveground biomass ratio was quite high in the B00/clay sites, resulting in a large predicted nodule biomass at the stand-level and thus a high estimate for N2 fixation rate. Desmodium marilandicum had the second highest estimated N2 fixation rate in B02/sand sites, followed by L. hirta in B00/sand sites (Table 2). The overall ecosystem-level estimate for N2 fixation rate in July was 2.648 3 1026 g · m22 · h21 (Table 3). DISCUSSION Biomass of herbaceous species in southeastern pine forests is often greater in more frequently burned sites (Raison, 1979; Schoch and Binkley, 1986; Dudley and Lajtha, 1993; Brockway and Lewis, 1997; Wilson et al., 2002; Glitzenstein et al., 2003) and in wet-mesic compared with xeric sites (Kirkman et al., 2001). This occurs, in part, because many herbaceous species resprout or germinate from scarified seeds following fire (Towne and Knapp, 1996; Brockway and Lewis, 1997; Hendricks and Boring, 1999) and because water limits production on xeric sites (Mitchell et al., 1999; Kirkman et al., 2001). In contrast to these expectations, total aboveground bio-

mass of ground layer vegetation in the 16 mixed pine hardwood sites at Fort Benning did not differ between sites burned in 2002 and those last burned in 2000 or with soil texture (sandy vs. clayey soil). However, greater belowground biomass in sandy compared to clayey sites is consistent with the observation that some plants respond to nutrient or water stress by allocating a greater proportion of their biomass to belowground structures (Keyes and Grier, 1981). In general, ground layer biomass in the sampled mixed pine hardwood stands (85 g/m2 in sandy sites, 106 g/m2 in clayey sites) during the very dry summer of 2002 was intermediate to that reported for other southeastern forests. Biomass in Coastal Plain (southern Georgia) longleaf pine forests burned on a 1–3 yr rotation since 1942 was 60.2 g/m2 (Brockway and Lewis, 1997), while xeric and wet-mesic longleaf-dominated sites in southwestern Georgia also burned on a 1–3 yr rotation ranged from 130–145 g/m2 (in xeric sites) to 233–269 g/m2 (in wet-mesic sites) (Kirkman et al., 2001). In contrast to observations that grasses increase or dominate ground layer vegetation in frequently burned sites (Brockway and Lewis, 1997; Glitzenstein et al., 2003), grass aboveground biomass increased more relative to other plant groups in B02 compared to B00 sites only in the sandy mixed pine-hardwood stands at Fort Benning. Further, in both clayey and sandy sites, grasses made up a smaller proportion (21%) of the total ground layer biomass than forbs (28%) and woody species (32%) for both fire treatments. Although grass biomass may have been low as a result of competition from woody species, it also was influenced by soil texture. Grass biomass was lower in sandy than in clayey sites.

90

AMERICAN JOURNAL

OF

BOTANY

[Vol. 93

TABLE 2. Mean aboveground biomass, nodule biomass, acetylene reduction, and estimated N2 fixation rate per plant for the six dominant legume species sampled. N2 fixation rates for Desmodium paniculatum in B02/sand sites and Tephrosia virginiana in B02/clay sites could not be estimated because they were not present in the vegetation plots in those sites. Desmodium marilandicum

Desmodium paniculatum

8.77 22.45 2.11 —

3.00 — 0.86 1.51

1.74 2.34 1.79 3.22

Mean aboveground biomass per plant (g) B02/clay 0.11 5.62 B02/sand 0.15 2.49 B00/clay 0.01 3.86 B00/sand 0.14 n.d.

n.d. 2.05 0.95 n.d.

Mean nodule biomass per plant (mg) B02/clay 2.46 B02/sand 1.41 B00/clay 2.01 B00/sand 0.66

1.40 2.81 3.58 n.d.

n.d. 2.10 1.37 n.d.

Acetylene reduction activity (nmol · h21 B02/clay 8.65 B02/sand 6.77 B00/clay 4.08 B00/sand 11.76

· mg nodule dry mass21) 7.44 n.d. 9.06 3.66 8.39 1.90 n.d. n.d.

Treatment

Cassia nictitans

Mean aboveground biomass (g/m2) B02/clay 0.22 B02/sand 0.55 B00/clay 0.38 B00/sand 0.33

N2 fixation rate B02/clay B02/sand B00/clay B00/sand

(g · m22 · 3.995 3 3.311 3 2.888 3 1.598 3

h21) 1027 1027 1026 1027

1.484 3 1027 2.135 3 1026 1.484 3 1027 —

— — 2.283 3 1028 —

Desmodium viridiflorum

Lespedeza hirta

Tephrosia virginiana

Mean

7.18 3.45 2.21 8.00

— 17.00 18.27 —

3.49 7.63 4.27 2.18

5.62 n.d. n.d. 24.83

5.97 n.d. n.d. 9.52

14.65 n.d. 8.69 n.d.

0.56 0.31 0.36 0.77

4.81 n.d. n.d. 5.81

5.98 n.d. n.d. 26.78

7.63 n.d. 2.90 n.d.

3.71 1.22 1.64 5.54

5.98 n.d. n.d. 12.76

7.65 n.d. n.d. 8.2

6.32 n.d. 4.75 n.d.

4.43 3.47 3.77 5.13

7.991 3 1028 — — 6.849 3 1028

5.137 3 1027 — — 1.724 3 1026

— — 2.740 3 1027 —

1.142 2.466 3.333 1.952

3 3 3 3

1026 1026 1026 1026

Notes: Mean aboveground biomass was taken for legume species in the circular vegetation plots. Dashes represent an absence of the species in our random sampling scheme, but do not necessarily indicate that the species did not occur in the site. Mean aboveground biomass per plant is a separate calculation specifically for the N2 fixation assay, which included only the six dominant legume species sampled for that assay. Here, n.d. indicates that no data were collected. Only the three most abundant legume species were sampled per site. Means are adjusted for treatment species abundances.

In addition to grasses, the other plant groups, including legumes, generally had greater biomass in clayey sites; exceptions were woody species in 2002 and ferns and forbs in 2003. Soil moisture varies with soil texture at Fort Benning (Dilustro et al., 2005) and is likely an additional influence on ground layer development along with the well-studied influence of fire frequency (Glitzenstein et al., 2003). Legumes have been shown to increase in biomass following fire when it occurs on 2–4 yr intervals (Cushwa et al., 1966; Hendricks and Boring, 1999). However, legume populations usually are limited by availability of water (Lauenroth and Dodd, 1979). Results from the mixed pine forests at Fort Benning support these observations. The combined biomass of the six abundant harvested legumes was greater in clayey sites burned in 2002 and sandy sites last burned in 2000. Low soil moisture may have limited legume biomass following fire in the sandy sites, while burning may have stimulated germination and regrowth in the clayey sites. Sandy sites also may contain fewer nutrients, including phosphorous; this may limit the growth of N2-fixing organisms (Vitousek et al., 2002), although many legume species are adapted to low phosphorous environments (Sprent, 1999). In addition, legumes in temperate latitudes are often shade-intolerant (Sprent, 1999; Vitousek et al., 2002). We found reduced legume biomass in B00 sites where canopy openness is also reduced, but this was not statistically significant. Competition from ground layer woody species also may have influenced legume biomass.

Despite their positive response to fire on some sites, legumes did not contribute substantially to stand-level N inputs via aboveground litter. Although they had the greatest tissue C and N concentrations of the plant groups and thus provided high-quality nitrogen-rich litter input to the ecosystem, legumes made up only 4% of the total ground layer biomass. Among the six legume species examined, T. virginiana, which had peak tissue N concentrations in B00/clay sites, contributed the greatest N standing crop inputs via aboveground litter, as it had the greatest biomass in sites in which it was found (Table 2). Cassia nictitans had the second highest tissue N concentrations in B02/clay and B00/sand sites (Fig. 6A). Acetylene reduction activity (ARA) was greatest in B02/ clay and B00/sand sites, primarily because C. nictitan’s activity was high in both sites and D. viridiflorum’s activity was high in B00/sand sites (Table 2). Desmodium viridiflorum also was found to have the highest ARA of three legume species studied (including Lespedeza hirta and L. procumbens) in a frequently-burned loblolly pine-hardwood forest (Piedmont National Wildlife Refuge, Jones County, Georgia; Hendricks and Boring, 1999). However, activity values for these three species (except L. hirta, which had no detectable activity) were higher than found at Fort Benning. For example, activity for D. viridiflorum was 75.4 nmol · h21 · mg nodule dry mass21 in the wildlife refuge forest (Hendricks and Boring, 1999) and 12.8 nmol · h21 · mg nodule dry mass21 at Fort Benning. Although a drought during 2002 may have influenced the dis-

January 2006]

LAJEUNESSE


Fig. 7. Legume (A) aboveground biomass and (B) belowground biomass per plant of plants harvested for the acetylene reduction assay (means 1 SE). Means for the same species with different letters were significantly different at a 5 0.05.

tribution and abundance of legumes at Fort Benning, leading to a long-term species shift, rainfall was normal during 2003 when the ARA assay was performed and probably did not affect ARA. Canopy openness (approximately 30%) did not differ significantly between burn treatments and was slightly TABLE 3.

91

greater on sites burned in 2002 (B02 sites). Thus canopy openness is not responsible for the high ARA measured in B00/ sand sites. Greater ARA, per-plant nodule mass, and per-plant biomass in clayey sites burned after 2 yr and in sandy sites left unburned were offset by higher biomass of legumes at the plot or community level (i.e., on an area basis) and, thus, higher stand-level N fixation rates in B00/clay sites and B02/sand sites. In general, stand-level estimates for N2 fixation rates of the six legume species combined ranged from 1.690 3 1026 to 3.447 3 1026 g N · m22 · h21 at Fort Benning, with an overall average N2 fixation value estimate of 2.648 3 1026 g N · m22 · h21 at the ecosystem level for July. Although our ecosystem-level N2 fixation rate is just an estimate based on a one-time, mid-year, indirect method, it has value for comparisons with similar estimates. Hendricks and Boring (1999) estimated an ecosystem-level N2 fixation rate of 7.991 3 1025 to 1.027 3 1024 g N · m22 · h21 for two loblolly pine-hardwood dominated sites that had been prescribe-burned at 4–5 yr intervals since 1962 and less than 3.425 3 1026 g N · m22 · h21 for two other loblolly pinehardwood dominated sites that had only burned once at the time of sampling 2–3 yr following the fire. In the Hendricks and Boring sites that were burned only once, legume densities were low, as they are in our Fort Benning sites. Thus, it is not surprising that the ecosystem-level N2 fixation rate estimates are also similar. The large difference between our estimate and the Hendricks and Boring estimate from the sites that underwent prescribed burns at 4–5 yr intervals could be due to the fact that their sites have been regularly burned for 20 yr longer than our sites, which has resulted in greater densities of legumes. Differences in methodology between the two studies also could have influenced the stand-level fixation estimates. Hendricks and Boring measured N2 fixation rates for three legume species on a per-plant basis, using nodule biomass, followed by a gross estimation of N2 fixation rates at the ecosystem level. Our estimates are based on measured N2 fixation rates per plant for six legume species and were extrapolated to the stand level using measured nodule and aboveground biomass estimates. The interaction between soil texture and fire on legume above- and belowground biomass, tissue C and N pools in legumes, and on legume acetylene reduction activities dem-

Mean aboveground biomass and N2 fixation rate for legume species used in the acetylene reduction assay plus additional species.

Variable

Six species total (from Table 2)

All other Desmodium spp.

All other Lespedeza spp.

All other Tephrosia spp.

All other legume spp.

Total

20.91 45.79 25.62 13.06

4.09 1.87 0.36 4.46

2.69 2.25 2.36 2.25

3.21 1.06 1.21 1.71

2.90 2.54 1.13 1.97

33.80 53.51 30.68 23.45

1.941 3 1027 — — 4.795 3 1027

1.027 3 1027 — 2.283 3 1028 —

2

Mean aboveground biomass (g/m ) B02/clay B02/sand B00/clay B00/sand N2 fixation rate (g · m22 · h21) B02/clay B02/sand B00/clay B00/sand Mean N2 fixation rate (g · m22 · h21) across treatments

1.142 2.466 3.333 1.952

3 3 3 3

1026 1026 1026 1026

1.027 1.598 2.283 1.256

3 3 3 3

1027 1027 1028 1027

1.484 3.425 6.849 2.283

3 3 3 3

1027 1028 1028 1027

1.690 2.660 3.447 2.785 2.648

3 3 3 3 3

1026 1026 1026 1026 1026

Note: Mean aboveground biomass of the remaining legume species was used in conjunction with the biomass per plant, nodule biomass, and acetylene reduction activities of the six dominant legume species sampled (Table 2) to generate the ecosystem-level N2 fixation rate of 2.648 3 1026 g · m22 · h21 for July.

92

AMERICAN JOURNAL

onstrates the complexity of the effects of disturbance on plant ecological processes (Towne and Knapp, 1996) and suggests that soil texture may be important when choosing to prescribeburn managed southeastern pine forests. However, stand-level N2 fixation estimates were low for all sites at Fort Benning, primarily because legumes were not abundant. Although these stands have been prescribe-burned on a roughly 3-yr rotation over the last 20 yr, low legume biomass is largely responsible for Fort Benning ecosystem-level fixation estimates being more similar to the once-burned loblolly sites than to the frequently burned loblolly-dominated sites examined by Hendricks and Boring (1999). Other researchers have also suggested that N2 fixation inputs in Fort Benning ecosystems are very low (Garten and Ashwood, 2004). One caveat of our research is the short term of the observations, which spanned 2 yr after fire in the B02 sites and the third and fourth years after fire in B00 sites, following the 20yr history of fire at approximately 3-yr intervals in all sites. Although our results are not a precise estimate, they do illustrate the low contribution of fixed N2 following fire in some stands. Increased legume populations may not ensure increased fixation, but management of these stands to maintain or increase legume populations would undoubtedly have a positive effect on N availability and may serve to alleviate any N deficiency over the longer term in these fire-maintained forests. Fire regimes with frequent understory burning should be implemented to inhibit hardwood dominance and favor grass and legume cover that benefit unique pine-grassland wildlife. This can improve soil quality of such sites with historical legacies of agricultural degradation. LITERATURE CITED BOYER, W. D. 2000. Long-term effects of biennial prescribed fires on the growth of longleaf pine. In W. K. Moser and C. F. Moser, [eds.], Fire and forest ecology: innovative silviculture and vegetation management, 18–21. Tall Timbers Fire Ecology Conference Proceedings, 21, 1998. Tall Timbers Research Station, Tallahassee, Florida, USA. BOYER, W. D., AND J. H. MILLER. 1994. Effect of burning and brush treatments on nutrient and soil physical properties in young longleaf pine stands. Forest Ecology and Management 70: 311–318. BROCKWAY, D. G., AND C. E. LEWIS. 1997. Long-term effects of dormantseason prescribed fire on plant community diversity, structure and productivity in a longleaf pine wiregrass ecosystem. Forest Ecology and Management 96: 167–183. CALDWELL, T. G., D. W. JOHNSON, AND W. W. MILLER. 2002. Forest floor carbon and nitrogen losses due to prescription fire. Soil Science Society of America Journal 66: 262–267. CHRISTENSEN, N. L. 1977. Fire and soil-plant nutrient relations in a pinewiregrass savanna on the Coastal Plain of North Carolina. Oecologia 31: 27–44. CUSHWA, C. T., E. V. BRENDER, AND R. W. COOPER. 1966. The response of herbaceous vegetation to prescribed burning. USDA Forest Service Research Note SE-53. USDA Forest Service, Asheville, North Carolina, USA. DILUSTRO, J. J., B. S. COLLINS, L. K. DUNCAN, AND C. CRAWFORD. 2005. Moisture and soil texture effects on soil CO2 efflux components in southeastern mixed pine forests. Forest Ecology and Management 204: 87– 97. DILUSTRO, J. J., B. S. COLLINS, L. K. DUNCAN, AND R. R. SHARITZ. 2002. Soil texture, land-use intensity, and vegetation of Fort Benning upland forest sites. Journal of the Torrey Botanical Society 129: 289–297. DUDLEY, J. L., AND K. LAJTHA. 1993. The effects of prescribed burning on nutrient availability and primary production in sandplain grasslands. American Midland Naturalist 130: 286–298. FLEMATTI, G. R., E. L. GHISALBERTI, K. W. DIXON, AND R. D. TRENGROVE. 2004. A compound from smoke that promotes seed germination. Science 305: 977.

OF

BOTANY

[Vol. 93

GARTEN, C. T., AND T. L. ASHWOOD. 2004. Modeling soil quality thresholds to ecosystem recovery at Fort Benning, Georgia, USA. Oak Ridge National Laboratory Technical Report ORNL/TM-2004/41. Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA. GLITZENSTEIN, J. S., D. R. STRENG, AND D. D. WADE. 2003. Fire frequency effects on longleaf pine (Pinus palustris P. Miller) vegetation in South Carolina and Northeast Florida, USA. Natural Areas Journal 23: 22–37. HAINDS, M. J., R. J. MITCHELL, B. J. PALIK, 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. 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 herbaceous legumes in burned pine ecosystems of the southeastern United States. Forest Ecology and Management 113: 167–177. HIERS, J. K., R. J. MITCHELL, L. R. BORING, J. J. HENDRICKS, AND R. WYATT. 2003. Legumes native to longleaf pine savannas exhibit capacity for high N2-fixation rates and negligible impacts due to timing of fire. New Phytologist 157: 327–338. HIERS, J. K., R. WYATT, AND R. J. MITCHELL. 2000. The effects of fire regime on legume reproduction in longleaf pine savannas: is a season selective? Oecologia 125: 521–530. JOHNSON, D. W., AND P. S. CURTIS. 2001. Effects of forest management on soil C and N storage: meta analysis. Forest Ecology and Management 140: 227–238. KEYES, M. R., AND C. C. GRIER. 1981. Above- and belowground net production in 40-year-old Douglas-fir stands on low and high productivity sites. Canadian Journal of Forest Research 11: 599–605. KEYS, J. A. 1995. Ecological units of the eastern United States—first approximation. [CD-ROM digital data]. USDA Forest Service, Atlanta, Georgia, USA. KIRKMAN, L. K., R. J. MITCHELL, R. C. HELTON, AND M. B. DREW. 2001. Productivity and species richness across an environmental gradient in a fire-dependent ecosystem. American Journal of Botany 88: 2119–2128. LAUENROTH, W. K., AND J. L. DODD. 1979. Response of native grassland legumes to water and nitrogen treatments. Journal of Range Management 32: 292–294. LIU, Y., AND R. N. MULLER. 1993. Above-ground net primary productivity and nitrogen mineralization in a mixed mesophytic forest of eastern Kentucky. Forest Ecology and Management 59: 53–62. MINCHIN, F. R., J. E. SHEEHY, AND J. F. WITTY. 1986. Further errors in the acetylene reduction assay: effects of plant disturbance. Journal of Experimental Botany 37: 1581–1591. MINCHIN, F. R., J. F. WITTY, J. E. SHEEHY, AND M. MULLER. 1983. A major error in the acetylene reduction assay: decreases in nodular nitrogenase activity under assay conditions. Journal of Experimental Botany 34: 641– 649. MITCHELL, R. J., L. K. KIRKMAN, S. D. PECOT, C. A. WILSON, B. J. PALIK, AND L. R. BORING. 1999. Patterns and controls of ecosystem function in longleaf pine-wiregrass savannas. I. Aboveground net primary productivity. Canadian Journal of Forest Research 29: 743–751. MYROLD, D. D., R. W. RUESS, AND M. J. KLUG. 1999. Dinitrogen fixation. In G. P. Robertson, D. C. Coleman, C. S. Bledsoe, and P. S. Sollins [eds.], Standard soil methods for long-term ecological research, 241–257. Oxford University Press, New York, New York, USA. NATURAL RESOURCES CONSERVATION SERVICE. 1983. Soil survey, Muscogee County, Georgia. United States Government Printing Office, Washington, D.C., USA. NEARY, D. G., C. C. KLOPATEK, L. F. DEBANO, AND P. F. FFOLLIOTT. 1999. Fire effects on belowground sustainability: a review and synthesis. Forest Ecology and Management 122: 51–57. OJIMA, D. S., D. S. SCHIMEL, W. J. PARTON, AND C. E. OWENSBY. 1994. Long- and short-term effects of fire on nitrogen cycling in tallgrass prairie. Biogeochemistry 24: 67–84. PALIK, B. J., R. J. MITCHELL, G. HOUSEAL, AND N. PEDERSON. 1997. Effects of canopy structure on resource availability and seedling responses in a longleaf pine ecosystem. Canadian Journal of Forest Research 27: 1458–1464. PASTOR, J., J. D. ABOR, C. A. MCCLAUGHERTY, AND J. M. MELILLO. 1984. Aboveground production and N and P cycling along a nitrogen miner-

January 2006]

LAJEUNESSE


alization gradient on Blackhawk Island, Wisconsin. Ecology 65: 256– 268. PROVENCHER, L., B. J. HERRING, D. R. GORDON, H. L. RODGERS, K. E. M. GALLEY, G. W. TANNER, J. L. HARDESTY, AND L. A. BRENNAN. 2001. Effects of hardwood reduction techniques on longleaf pine sandhill vegetation in northwest Florida. Restoration Ecology 9: 13–27. RADFORD, A. E., H. E. AHLES, AND C. R. BELL. 1968. Manual of the vascular flora of the Carolinas. University of North Carolina Press, Chapel Hill, North Carolina, USA. RAISON, R. J. 1979. Modification of the soil environment by vegetation fires with particular reference to nitrogen transformations: a review. Plant and Soil 51: 73–108. REICH, P. B., D. F. GRIGAL, J. D. ABER, AND S. T. GOWER. 1997. Nitrogen mineralization and productivity in 50 hardwood and conifer stands on diverse soils. Ecology 78: 335–347. REICH, P. B., D. W. PETERSON, D. W. WEDIN, AND K. WRAGE. 2001. Fire and vegetation effects on productivity and nitrogen cycling across a forest-grassland continuum. Ecology 82: 1703–1719. ROBERTSON, G. P., D. W. WEDIN, P. M. GROFFMAN, J. M. BLAIR, E. A. HOLLAND, K. J. NADELHOFFER, AND D. H. HARRIS. 1999. Soil carbon and nitrogen availability. In G. P. Robertson, D. C. Coleman, C. S. Bledsoe, and P. S. Sollins [eds.], Standard soil methods for long-term ecological research, 258–271. Oxford University Press, New York, New York, USA. SAS. 2000. SAS version 8.1. SAS Institute, Cary, North Carolina, USA. SCHOCH, P., AND D. BINKLEY. 1986. Prescribed burning increased nitrogen availability in a mature loblolly pine stand. Forest Ecology and Management 14: 13–22. SPRENT, J. I. 1999. Nitrogen fixation and growth of non-crop legume species

93

in diverse environments. Perspectives in Plant Ecology and Evolutionary Systematics 2: 149–162. TOWNE, E. G., AND A. K. KNAPP. 1996. Biomass and density responses in tallgrass prairie legumes to annual fire and topographic position. American Journal of Botany 83: 175–179. VAN LEAR, D. H., AND V. J. JOHNSON. 1983. Effects of prescribed burning in the southern Appalachian and upper Piedmont forests: a review. Clemson University Forestry Bulletin Number 36. Clemson, South Carolina, USA. VESSEY, J. K. 1994. Measurement of nitrogenase activity in legume root nodules: in defense of the acetylene reduction assay. Plant and Soil 158: 151–162. VITOUSEK, P. M. 1982. Nutrient cycling and nutrient use efficiency. American Naturalist 119: 553–572. VITOUSEK, P. M., K. CASSMAN, C. CLEVELAND, T. CREWS, C. B. FIELD, N. B. GRIMM, R. W. HOWARTH, R. MARINO, L. MARINELLI, E. B. RASTETTER, AND J. I. SPRENT. 2002. Towards an ecological understanding of biological nitrogen fixation. Biogeochemistry 57/58: 1–45. WAN, S., D. HUI, AND Y. LUO. 2001. Fire effects on nitrogen pools and dynamics in terrestrial ecosystems: a meta-analysis. Ecological Applications 11: 1349–1365. WILSON, C. A., R. J. MITCHELL, L. R. BORING, AND J. J. HENDRICKS. 2002. Soil nitrogen dynamics in a fire maintained forest ecosystem: results over 3-year burn interval. Soil Biology and Biochemistry 34: 679–698. WILSON, C. A., R. J. MITCHELL, J. J. HENDRICKS, 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. WRIGHT, R. J., AND S. C. HART. 1997. Nitrogen and phosphorus status in a ponderosa pine forest after 20 years of interval burning. Ecoscience 4: 526–533.