Dinitrogen fixation in Illinois bundleflower.
Illinois bundleflower is compatible in mixtures with a diversity of warm-season grasses. In east-central Texas, established kleingrass (Panicum coloratura L.) stands interseeded with 'Sabine' Illinois bundleflower had greater forage yields than kleingrass monocultures in the second, third, and fourth years after seeding (Dovel et al., 1990). Mixtures of Illinois bundleflower accession PI 434011 with switchgrass (Panicum virgatum L.), sideoats grama (Bouteloua curtipendula Michx.), or indiangrass [Sorghastrum nutans (L.) Nash] in northeastern Kansas had higher forage yield and crude protein concentration than grass monocultures in the fourth and fifth years after establishment (Posler et al., 1993). Springer et al. (2001) showed that Sabine Illinois bundleflower was highly compatible in mixture with indiangrass.
Although Illinois bundleflower is known to contribute N in forage systems (Dovel et al., 1990; Posler et al., 1993; Springer et al., 2001), there is a paucity of information on its [N.sub.2] fixation capability. Nodulation has been observed in 90% of the Mimosoideae (de Faria et al., 1989) including Illinois bundleflower (Allen and Allen, 1981). Kulakow et al. (1990) found that mean acetylene reduction rates of 70-d-old seedlings of Illinois bundleflower were comparable with those of soybean [Glycine max (L.) Merr.] and alfalfa (Medicago sativa L.), but also noted that rates varied significantly among accessions. However, the instantaneous rates of [N.sub.2] fixation measured by the acetylene reduction technique cannot be extrapolated to obtain an accurate estimate of total [N.sub.2] fixed across a growing season.
Integrated estimates of [N.sub.2] fixation can be made by comparing N accumulation with non-[N.sub.2]-fixing plants (the total N difference technique) or by isotopic N methodologies. The latter depends on a [sup.15]N-labeled soil N pool, which is manifested in higher [sup.15]N concentrations in non[N.sub.2]-fixing reference plants dependent on soil N than in [N.sub.2]-fixing plants, which incorporate both soil and unlabeled atmospheric N. In the [sup.15]N enrichment or isotope dilution technique, the soil N pool is labeled with an [sup.15]N-enriched source, whereas the [sup.15]N natural abundance method relies on the small but measurable elevation of [sup.15]N already present in most soils relative to the atmosphere (Chalk, 1985; Shearer and Kohl, 1986). In this study, we compared both isotope methods and the total N difference method as estimates of [N.sub.2] fixation in monocultures of Illinois bundleflower. Our objective was to determine the in-field [N.sub.2]-fixation of three accessions of Illinois bundleflower managed as a forage crop in pure stands during the first 2 yr after seeding.
MATERIALS AND METHODS
Site and Climate Characteristics
The experiment was conducted at the Sand Plain Research Farm, Becket, MN (45024, N, 93[degrees]53' W), the Rosemount Research and Outreach Center, Rosemount, MN (44053, N, 93013' W), and the Southwest Research and Outreach Center, Lamberton, MN (44[degrees]15, N, 95[degrees]19' W). The soil at Becker was a Hubbard loamy sand (sandy, mixed, frigid Entic Hapludoll). At Rosemount, it was a Wankegan silt loam (fine-silty over sandy, mixed, superactive, mesic Typic Hapludoll). At Lamberton, the soil was a Normania clay loam (fine-loamy, mixed, superactive, mesic Aquic Hapludoll). Soil chemical characteristics for the three sites are reported in Table 1. Previous crops at the three sites were rye (Secale cereale L.) at Becker, corn (Zea mays L.), and soybean at Rosemount, and oat (Arena sativa L.) at Lamberton. Manure had not been applied to the sites within the previous 5 yr. On the basis of soil test results, 390 kg [ha.sup.-1] K and 4.5 kg [ha.sup.-1] B were applied at Becker in March 2000 and 100 kg [ha.sup.-1] P was applied at Lamberton in May 2000. The background [delta][sup.15]N natural abundance of total N in the top 15 cm of the soil at Becker, Rosemount, and Lamberton, as determined by a Europa Scientific Hydra 20/20 mass spectrometer (PDZ Europa, Cheshire, UK) (1), was 7.47 [+ or -] 0.12, 7.85 [+ or -] 0.03, and 7.95 [+ or -] 0.16 [per thousand] (means [+ or -] standard error for six replicates), respectively.
Monthly precipitation totals and average air temperatures from May to September for the three sites are shown in Table 2. Precipitation at Lamberton was near the 30-yr average both years. In the 2000 growing season, precipitation was 175 mm below average at Becker. Precipitation was near average in the 2000 season at Rosemount, but 128 mm above average in July. In the 2001 growing season, precipitation was 127 and 96 mm below the 30-yr mean at Becker and Rosemount, respectively. The plots at Becker received 225 mm of irrigation in both seasons. Air temperatures at all three sites were generally near average in the 2000 season and above average in the 2001 season.
Three accessions of Illinois bundleflower from the University of Minnesota Perennial Native Legume (PNL) collection were used. They were chosen for winter survival, maturation, height, and seed yield using the results collected by DeHaan et al. (2003) for plants grown at Becker and Saint Paul, MN. Plants of PNL534 from Lake Hattie, Stevens County, Minnesota (45[degrees]32' N, 96[degrees]05' W) had early to midseason maturity, short height, low to moderate aboveground biomass, and low to moderate seed yield. Accession PNL539 from Spirit Lake, Dickinson County, Iowa (43[degrees]29' N, 95[degrees]05' W) had midseason maturity, tall height, moderate aboveground biomass, and moderate to high seed yield. PNL541 from Cottonwood Lake, Spink County, South Dakota (44[degrees]47' N, 98[degrees]42' W) had early-season maturity, moderate height, high aboveground biomass, and high seed yield. All accessions had good winter survival at Becker and Saint Paul. On the basis of 54 phenotypic traits measured, DeHaan et al. (2003) clustered PNL539 and PNL541 together, whereas PNL534 grouped in a different cluster.
Wild senna [Senna hebecarpa (Fernald) H.S. Irwin & Barneby], a warm-season perennial legume native to the northeastern and east-central USA (Irwin and Barneby, 1982), was grown as the non-[N.sub.2]-fixing reference plant for use in the comparisons involving [sup.15]N. Like many Caesalpinioid legumes, it is not nodulated (Allen and Allen, 1981), but in Minnesota, it generally exhibits similar emergence times and biomass accumulation patterns as Illinois bundleflower (E. Ristau, 1999, personal communication). Detailed descriptions of the root morphology of D. illinoensis and S. hebecarpa are lacking.
The experiment was designed as a split plot at each of three locations with whole plots arranged in randomized complete blocks. There were six replicates per location. The whole-plot treatments were natural abundance and enriched levels of [sup.15]N. The split-plot treatments were the three Illinois bundleflower accessions and wild senna. Each plot measured 6 by 3 m. A 1.5-m-wide alley was maintained around each whole plot to reduce contamination of [sup.15]N natural abundance plots with enriched [sup.15]N.
Seed of the three populations of Illinois bundleflower were from DeHaan et al. (2003). The seed was mechanically scarified with sandpaper and inoculated with 5 g [kg.sup.-1] of a commercial, peat-based inoculant for Desmanthus using 22 mL [kg.sub.-1] seed of a 10% (w/v) sucrose solution as an adhesive. The inoculant contained two strains: Nitragin 43A1 and 43C2 (Liphatec, Milwaukee, WI). In 2000, seed was drilled in rows 15 cm apart on 25 May at Becket, 3 June at Lamberton, and 7 June at Rosemount. Seeding rates were 200 live seeds [m.sup.-2] for Illinois bundleflower and 510 live seeds [m.sup.-2] for wild senna. The higher rate for senna was based on previous data showing lower persistence and a less-open growth form than bundleflower (E. Ristau, 1999, personal communication). Average stand densities of Illinois bundleflower in 2000 were 157, 93, and 72 plants [m.sup.-2] at Becker, Rosemount, and Lamberton, respectively. The corresponding densities of wild senna were 256, 171, and 155 plants [m.sup.-2].
In the seeding year, 52 mL a.i. [ha.sup.-1] of imazapic [Plateau, [+ or -] 2-[4,5-dihydro-4-methyl-4-(1-methylethyl) -5-oxo-1H-imidazol-2-yl]-5-methyl-3-pyridinecarboxylic acid, American Cyanamid Co., Parsippany, NJ] was applied before seedling emergence to control weeds in the Illinois bundleflower plots. In the second year, 68 mL a.i. [ha.sup.-1] imazapic and 0.65 L a.i. [ha.sup.-1] pendimethalin [Prowl, N-(1-ethylpropyl)-3,4-dimethyl-2,6-dinitrobenzenamine, American Cyanamid Co.] with 1.2 L [ha.sup.-1] of an N-surfactant blend (Class Prefer-28, Cenex-Land O'Lakes Agronomy Co., Winona, MN) were applied before seedling emergence in the Illinois bundleflower plots. Pendimethalin inhibited growth of any seedlings from shattered first-year seed, ensuring purely 2-yr-old stands. Because wild senna does not tolerate imazapic (E. Ristau, 1999, personal communication), 0.5 L a.i. [ha.sup.-1] of trifluralin (Treflan, [alpha], [alpha], [alpha]-trifluoro-2,6-dinitro-N,N-dipropyl-p-toluidine, Dow AgroSciences, Indianapolis, IN) was incorporated into the soil before planting the wild senna plots. In both years, hand weeding was also performed.
At 20 to 23 days after planting (DAP), a 0.2 % (w/v) aqueous solution of 98 atom% [sup.15]N [(N[H.sub.4]).sub.2] S[O.sub.4] was applied to a 1.8-[m.sup.2] area in the enriched plots by injection every 7.5 cm between every row for a total application rate of 0.2 g [sup.15]N [m.sup.-2]. Dinitrogen fixation should not have been inhibited by this low rate of N addition (2 kg N [ha.sup.-1]). The injection cannulas had pores at depths of 7.5 cm and 15 cm. In the second year, the solution was applied to a separate 1.8-[m.sup.2] area in the enriched plots within 28 d after new shoots were visible. In both years, central areas in each plot were chosen for treatment to minimize edge effects.
Plants were harvested at maximum aboveground biomass (approximately 10% seedpod fill). In 2000, this occurred on 9 September at Lamberton, 10 September at Becker, and 14 September at Rosemount. In 2001, harvest dates were 14 August at Rosemount, 20 August at Becker, and 25 to 26 August at Lamberton. Aboveground biomass from a 0.8-[m.sup.2] area within the [sup.15]N-treated area in the enriched plots was clipped at ground level by hand. A 0.8-[m.sup.2] area also was harvested from the [sup.15]N natural abundance plots in a similar manner. Standing first-year stems were mowed and removed from the plots by hand-raking in April and May 2001, before emergence of second-year plants.
Samples were oven-dried for 48 h at 60[degrees]C and ground to pass a 0.5-mm screen. Total N concentration was determined using a LECO CN-2000 analyzer (LECO Corporation, St. Joseph, MI). The University of California-Davis Stable Isotope Facility determined isotopic composition of shoots. Shoots with enriched levels of [sup.15]N were analyzed on a Europa Scientific Integra mass spectrometer (PDZ Europa, Cheshire, UK), whereas shoots with natural abundance levels of [sup.15]N were analyzed on a Europa Scientific Hydra 20/20 mass spectrometer. Illinois bundleflower leaf tissue was analyzed for S and Mo concentration using a LECO S-144DR S analyzer (LECO Corporation, St. Joseph, MI) and an Applied Research Laboratories 3560 inductively coupled plasma spectrometer (Thermo ARL, Ecublens, Switzerland), respectively.
The yield-independent measure %Ndfa represents the percentage of plant N derived from the atmosphere through [N.sub.2] fixation. With natural abundance levels of [sup.15]N,
 %Ndfa = 100([delta][sup.15][N.sub.o] - [delta][sup.15][N.sub.t])/ [delta][sup.15][.sub.o] - [delta][sup.15][N.sub.a]),
where [delta][sup.15]N is the per rail ([per thousand]) departure from the [sup.15]N concentration of the atmosphere (which is at a constant 0.3663 atom% [sup.15]N, thus [delta][sup.15][N.sub.atmosphere] = 0 by definition), o is the non-[N.sub.2]-fixing reference plant, t is the [N.sub.2]-fixing plant grown under field conditions in which both soil and atmospheric N are available, and a is the [N.sub.2]-fixing plant grown under conditions in which only atmospheric N is available (Shearer and Kohl, 1986). Discrimination against the heavier [sup.15]N isotope during chemical reactions such as [N.sub.2] fixation and N metabolism can lead to [sup.15]N enrichment in certain tissues (nodules), and [sup.15]N depletion in others (shoots) (Boddey et al., 2000). The resulting isotope fractionation is accounted for by the value [delta][sup.15][N.sub.a].
If [delta][sup.15][N.sub.o], is sufficiently large relative to [delta][sup.15][N.sub.a], or the converse, [delta][sup.15][N.sub.a] can be disregarded (Shearer and Kohl, 1986). Thus, at enriched levels of [sup.15]N, Eq.  reduces to:
 %Ndfa = 100([delta][sup.15][N.sub.p] - [delta] [sup.15][N.sub.t])/[delta][sup.15][N.sub.o] = 100(1 - [delta] [sup.15][N.sub.t]/[delta][sup.15][N.sub.o]).
At enriched levels of [sup.15]N, isotope concentrations are more conventionally expressed as atom% [sup.15]N excess, the departure in atom% [sup.15]N from the atmospheric concentration of 0.3663 atom% [sup.15]N, rather than as per mil departures:
 %Ndfa = 100[1 - atom%[sup.15]N excess (fs)/ atom%[sup.15]N excess (nfs)],
where FS refers to the [N.sub.2]-fixing system and NFS to the non[N.sub.2]-fixing system (Rennie, 1984).
These measures of %Ndfa are multiplied by N yield (which is calculated as dry matter yield per unit area x N concentration in the tissue) of the [N.sub.2]-fixing plant to give the amount of [N.sub.2] fixed per unit area, a yield-dependent measure. Both methods assume that the [N.sub.2]-fixing and reference plants absorb N from the same pool of soil N across time and that the [N.sub.2]-fixing and nonfixing crops absorb [sup.15.N]-labeled N and the plant-available soil N in the same ratio. An additional estimate of fixed [N.sub.2] can be calculated as the difference in N yield between the [N.sub.2]-fixing plant and the reference plant. This total N difference method assumes that the [N.sub.2]-fixing and non[N.sub.2]-fixing crops absorb equal amounts of soil N (Rennie, 1984).
Nodule Occupancy and [sup.15]N Fractionation
Because isotopic fractionation in plant shoots can vary according to rhizobial strain and host cultivar (Ledgard, 1989), a separate [delta][sup.15]N was determined for selected rhizobium from each location x accession combination. For this determination, 10 nodules sampled from each location x accession combination in August 2001 were surface sterilized in 3% (w/v) NaOCl, rinsed repeatedly, and crushed in yeast-mannitol broth (Somasegaran and Hoben, 1994). Nodule preparations were streaked on yeast mannitol agar plates. Plates were incubated for 3 d at 25[degrees]C, then rhizobium were reisolated and resubcultured to ensure purity (Somasegaran and Hoben, 1994). Rep-PCR (polymerase chain reaction) genomic fingerprinting of cultures was performed using the BOXA1 primer 5'-CTACGGCAAGGCGACGCTGACG-Y (Veraslovic et al., 1994; Schneider and de Bruijn, 1996) and the results subjected to cluster analysis using BioNumerics software (BioSystematica, Devon, UK) to produce an unweighted pair group method with arithmetic mean dendrogram with the Pearson correlation as the similarity coefficient.
Isolates were transferred to yeast mannitol broth cultures. Seeds of the three accessions of Illinois bundleflower were chemically scarified and sterilized by soaking in concentrated [H.sub.2]S[O.sub.4] for 5 min and then were washed repeatedly with sterile water. Plants were grown in 15-cm pots previously soaked in 3% NaOCl and filled with silica sand that had been washed with concentrated HCl. Six seeds were planted and then thinned to two per pot. Plants were inoculated at planting with 10 mL of broth culture of either the inoculant strains alone or a mixture of the inoculant strains plus indigenous strains from Lamberton. Plants were watered daily with N-free nutrient solution through capped glass tubes (Somasegaran and Hoben, 1994) and grown at 24[degrees]C during the day and 18[degrees]C at night, with a daylength of 16 h. Shoots were harvested 72 DAP, dried for 48 h at 60[degrees]C, and analyzed for natural abundance of [sup.15]N as above. The values of [delta][sup.15][N.sub.a] from the plants inoculated with the inoculant strain only were used for Becket and Rosemount, and the values of [delta][sup.15][N.sub.a] from the plants inoculated with the mixture of inoculant and indigenous strains were used for Lamberton (Table 3).
Data were analyzed using the PROC MIXED procedure of SAS 8.2 (SAS Institute, 1996) with all factors fixed except for the blocking variable, which was random and nested within location. To compare each [sup.15]N method with the difference method, a separate repeated measures ANOVA was performed for each [sup.15]N method, using the [sup.15]N method and the difference method as repeated measures. Normality of data was confirmed via normal probability plots of residuals (Oelheft, 2000). Constant variance of data was assessed with plots of raw residuals vs. predicted values across all locations (Oelhert, 2000). Since herbage yield and measures derived from this response (aboveground N yield and [N.sub.2] fixed) had nonconstant variance across locations, each location was analyzed separately for these responses. The herbage yield-independent measure %Ndfa had constant variance across all locations and the data from the three locations were combined into a single ANOVA for this response. Pairwise comparison of adjusted (least squares) means was performed using the Tukey-Kramer approximate honestly significant difference at significance level a = 0.05 (Oelhert, 2000).
RESULTS AND DISCUSSION
Nodule Occupancy and [sup.15]N Fractionation
The rep-PCR results revealed that at both Becker and Rosemount, 26 of the 30 rhizobial isolates grouped with the inoculant strains at >70% similarity (Fig. 1). In contrast, only 13 of the 28 isolates from Lamberton grouped with the inoculant strains at this level of similarity; the remainder presumably representing indigenous rhizobial strains. Both rhizobial strain (P < 0.001) and host accession (P < 0.001) affected isotopic fractionation (Table 3). Accession differences were likely due to differences in N partitioning between shoots and roots. Like Ledgard (1989), we observed lower shoot [delta][sup.15][N.sub.a] for plants inoculated with a mixture of indigenous strains than with the inoculant strains. Thus, we would have overestimated [N.sub.2] fixation at Lamberton if we had based estimates for %Ndfa on the fractionation value from plants infected only with the inoculant strains rather than with a mixture of the inoculant strains plus indigenous strains. It should be recognized that the inoculantadjusted values for [delta][sup.15][N.sub.a] may not reflect the relative occupancy of nodules by different rhizobial strains in the field, which was not estimated in this study. There was no rhizobial strain x accession interaction (P > 0.05).
The three accessions of Illinois bundleflower differed in herbage yield at Becker (P < 0.001) and Rosemount (P < 0.05), but not at Lamberton (P > 0.05) (Table 4). Herbage yield across all three accessions was lowest at Becker and highest at Lamberton. At Becker, PNL541 had higher herbage yield than the other accessions, whereas at Rosemount, PNL539 had higher herbage yield than PNL534. At Becker and Rosemount, herbage yield of all accessions was lower than that of wild senna, but at Lamberton, the three accessions yielded more than wild senna.
The accessions differed in aboveground N yield at Becker (P < 0.001), where PNL541 was the sole accession yielding more N than the non-[N.sub.2]-fixing wild senna (Table 4). Therefore, no [N.sub.2] fixation was detected by the total N difference method for PNL534 and PNL539. At Rosemount and Lamberton, accessions had similar N yield (P > 0.05), that averaged across accessions was 53 kg and 80 kg [ha.sup.-1], respectively. Thus, there were also no differences among accessions at these locations in the estimation of [N.sub.2] fixed by the total N difference method, which averaged 7 kg [ha.sup.-1] at Rosemount and 50 kg [ha.sup.-1] at Lamberton (Fig. 2).
There was a location x [sup.15]N method interaction for %Ndfa (P < 0.05) (Fig. 3). Averaged across accessions, the %Ndfa for the two [sup.15]N methods differed at Becker and Rosemount but not at Lamberton. The contrast between methods was particularly striking at Becker, where the [sup.15]N enrichment method estimated %Ndfa to be 52 units higher than for the [sup.15]N natural abundance method. The %Ndfa was similar for the three accessions at all locations (P > 0.05). With the [sup.15]N natural abundance method, none of the accessions at Rosemount fixed [N.sub.2] during the first year of growth, nor did two of the three accessions at Becker.
Dinitrogen fixation was greatest at Lamberton, where 26 and 48 kg N [ha.sup.-1] were fixed as estimated using natural abundance and enriched levels of [sup.15]N, respectively. Because of low N yield, [N.sub.2] fixation at Becker and Rosemount was only 11 and 16 kg [ha.sup.-1], respectively, according to the [.sup.15]N enrichment method, and nearly zero at both locations according to the [sup.15]N natural abundance method. The [sup.15]N methods gave similar results at Lamberton (P > 0.05), but not at Becket (P < 0.05) and Rosemount (P < 0.01). Accessions differed in [N.sub.2] fixation at Becker (P < 0.05) and Lamberton (P < 0.05), but not at Rosemount (P > 0.05). At Becker and Lamberton, PNL541 consistently fixed the most [N.sub.2] using either [sup.15]N method. Dinitrogen fixation was similar for PNL534 and PNL539.
The difference in %Ndfa between the two [sup.15]N methods at Becker and Rosemount may be due to the small differences in [delta][sup.15]N between Illinois bundleflower and wild senna (0.75 [per thousand] at Becker and 0.10 [per thousand] at Rosemount, vs. 2.33 [per thousand] at Lamberton, averaged across accessions). Numerous physical and biological processes other than symbiotic [N.sub.2] fixation can fractionate [sup.15]N, and can obscure the [sup.15]N signature of fixed atmospheric [N.sub.2] (Handley and Scrimgeour, 1997). The contribution of these fractionation processes becomes important at low levels of [sup.15]N enrichment, and leads to invalidation of the assumption that fixed [N.sub.2] and soil-derived N necessarily have unique [sup.15]N signatures (Handley and Scrimgeour, 1997). Therefore, the estimates of %Ndfa by the [sup.15]N natural abundance method probably should be considered less reliable than those given by the [sup.15]N enrichment method in this study.
Estimates of [N.sub.2] fixed were similar for the [sup.15]N natural abundance and total N difference methods at Becker and Rosemount (P > 0.05), but estimates for the methods differed at Lamberton (P < 0.001), where estimates for the total N difference method were higher (Fig. 2). The pattern was reversed when comparing the [sup.15]N enrichment method with the total N difference method:
the estimates for the two methods were similar at Lamberton (P > 0.05), but differed at Becker (P < 0.001) and Rosemount (P < 0.05), where estimates for the total N difference method were lower than for the [sup.15]N enrichment method.
Because of the low levels of [N.sub.2] fixation in the seeding year, Illinois bundleflower may benefit from supplemental N fertilization to improve establishment (Hojjati et al., 1978). Although Dovel et al. (1990) achieved good establishment of Illinois bundleflower in kleingrass without supplemental N, bundleflower grew very slowly in the establishment year. In Minnesota, the northern edge of its range, Illinois bundleflower may be more difficult to establish because significant growth in the first year does not occur until July (DeHaan et al., 2003). The use of supplemental N fertilizer sufficient to increase vigor and establishment, but not in amounts so that nodulation is suppressed (Streeter, 1988), may be advantageous at low-fertility sites like Becker.
Accessions had similar herbage yield, aboveground N yield, and [N.sub.2] fixed estimated by the total N difference or [sup.15]N methods, or %Ndfa (P > 0.05) (Table 5). The lack of accession effects suggests that the first-year effects could have been due in part to factors such as seedling vigor and establishment, or perhaps in speed of nodulation. Second-year differences among the accessions noted by DeHaan et al. (2003), who grew Illinois bundleflower in space-planted nurseries, may have been masked in our experiments when plants were grown in dense stands.
Herbage yield was highest at Lamberton, where the three accessions averaged 8.3 Mg [ha.sup.-1] (Table 5). By comparison, DeHaan et al. (2003) extrapolated a biomass yield of 4.9 Mg [ha.sup.-1] for his best accession of Illinois bundleflower grown at a low planting density of 0.87 plants [m.sup.-2] and cut once in August in the second year. Like DeHaan et al. (2003), we observed the highest biomass yield at a location within the native range of the species and lower yield outside the native range at Becket and Rosemount. At Becket, the accessions averaged 4.6 Mg [ha.sup.-1], whereas at Rosemount, the accessions averaged only 3.0 Mg [ha.sup.-1]. Wild senna yields varied from about 3.0 Mg [ha.sup.-1] at both Becker and Rosemount to 3.7 Mg [ha.sup.-1] at Lamberton.
A legume control with known yield potential was not included in this research. We are limited, therefore, to comparing data from Illinois bundleflower with published data from other trials. Mean yields of established stands of alfalfa, red clover (Trifolium pratense L.), and birdsfoot trefoil (Lotus corniculatus L.) at Rosemount were 13.2, 10.5, and 7.4 Mg [ha.sup.-1], respectively (Minnesota Agricultural Experiment Station, 2002), compared with only 3.0 Mg [ha.sup.-1] for Illinois bundleflower in our trial. In the same report, alfalfa yielded 12.0 Mg [ha.sup.-1] at Lamberton, about 50% more than Illinois bundleflower in our experiment in 2002. Lower yield of Illinois bundleflower compared with these cool-season legumes is likely due to its shorter growing season, the undomesticated status of the accessions used, or to less-than-optimum [N.sub.2] fixation. Aboveground N yield of Illinois bundleflower was lowest at Rosemount, where the accessions averaged only 48 kg [ha.sup.-1], which was less than first-year N yield at that site. Nitrogen yields at Becker and Lamberton were about 100 and 400% greater, respectively, than at Rosemount. Nitrogen yield of wild senna was more consistent than Illinois bundleflower, with low accumulation (<40 kg N [ha.sup.-1]) at all locations. On the basis of yields reported in the Minnesota cultivar trials cited above, and assuming a mean herbage N concentration of 26 g N [kg.sup.-1] dry mass, harvested herbage N of established alfalfa, red clover, and birdsfoot trefoil at Rosemount averaged 340, 270, and 190 kg N [ha.sup.-1], respectively, and alfalfa contained 310 kg N [ha.sup.-1] at Lamberton. Harvested N in Illinois bundleflower at the most productive site (Lamberton) was, therefore, about 40% lower than alfalfa and about equal to birdsfoot trefoil.
As in the first year, the two [sup.15]N methods differed in estimated %Ndfa (P < 0.01), with higher estimates of %Ndfa at enriched than at natural abundance levels of [sup.15]N. The location effect was significant (P < 0.001). Plants at the low-N Becker site were most dependent on fixed [N.sub.2], with %Ndfa ranging from 63 to 83% averaged across accessions. At Rosemount, %Ndfa of PNL534 and PNL541 was zero according to the [sup.15]N natural abundance method, whereas the [sup.15]N enrichment method estimated %Ndfa of 28 and 17% for these accessions, respectively. At Lamberton, %Ndfa across accessions was estimated at only 40% by the [sup.15]N natural abundance method, but 67 to 79% by the [sup.15]N enrichment and total N difference methods, respectively. In contrast to our results, Brandon et al. (1998), using the [sup.15]N natural abundance method, measured %Ndfa in inoculated secondyear stands of the related species Desmanthus virgatus (L.) Willd. of 78, 48, and 26%, in an infertile sandy soil, a moderately fertile black earth, and a highly fertile clay soil, respectively. This response corresponds more closely to what is generally expected; that is, that legume reliance on symbiotic [N.sub.2] fixation declines with increasing N supply (Allos and Bartholomew, 1959). We conclude that factors other than soil N supply were important in our experiment.
The two [sup.15]N methods did not differ in estimating [N.sub.2] fixed at Becker and Rosemount (P > 0.05). The [sup.15]N methods estimated [N.sub.2] fixation of about 70 kg [ha.sup.-1] across accessions at Becker and <10 kg [ha.sup.-1] fixed at Rosemount. At Lamberton, estimates of [N.sub.2] fixed were different for the two methods (P < 0.01), with an estimate only one-half as large with the [sup.15]N natural abundance method as with the [sup.15]N enrichment method.
Only 12 kg N [ha.sup.-1] was fixed at Rosemount according to the total N difference method, vs. 67 kg [ha.sup.-1] at Becker and 142 kg [ha.sup.-1] at Lamberton. The total N difference method produced higher estimates of [N.sub.2] fixed than did the [sup.15]N natural abundance method at Lamberton (P < 0.001), but the methods were similar at Becker (P > 0.05). In contrast, the total N difference method gave lower estimates of [N.sub.2] fixation than the [sup.15]N enrichment method at Becker (P < 0.001), but higher at Lamberton (P < 0.001), with no difference at Rosemount (P > 0.05). Except for the results at Rosemount, these estimates of [N.sub.2] fixation by Illinois bundleflower compare favorably to those of birdsfoot trefoil in the north-central USA, which generally have ranged from 50 to 205 kg N [ha.sup.-1], with a median of 90 to 100 kg N [ha.sup.-1] annually (Seguin et al., 2000). Under irrigated, high-yield conditions at Becker, others have estimated that established alfalfa fixed between 240 and 340 kg N [ha.sup.-1], depending on the method used to estimate [N.sub.2] fixation (Lamb et al., 1995).
Plants at Rosemount were visibly chlorotic in the second year. The relatively wet spring and dry summer in 2001 may have adversely affected [N.sub.2] fixation at Rosemount. Brandon and Date (1998) found that growth of Desmanthus virgatus was limited by S and Mo in a relatively fertile soil, and recommended a critical leaf S concentration of 2.0 g [kg.sup.-1]. Sulfur concentrations in Illinois bundleflower leaves were much lower at Rosemount (1.8 g [kg.sup.-1]) than at Becker (4.8 g [kg.sup.-1]) or Lamberton (5.2 g [kg.sup.-1]). Molybdenum concentrations were also lower in leaves from Rosemount (0.25 mg [kg.sup.-1]) than in leaves from Becker (2.1 mg [kg.sup.-1]) or Lamberton (3.0 mg [kg.sup.-1]). These results are quite surprising, given the high levels of soil organic matter at Rosemount (Table 1). Sulfur and Mo fertilization research may help clarify whether these nutrients limit Illinois bundleflower growth.
The inoculant rhizobial strains accounting for almost all nodules at Rosemount and Becker may have been less than optimal in symbiotic efficiency considered. Dinitrogen fixation was highest at Lamberton, the only site within the native range of Illinois bundleflower. Over one-half of the nodules sampled were occupied by noninoculant indigenous strains at this site. The relative effectiveness of the inoculant and indigenous strains has not been characterized and is a promising direction for further investigations into the [N.sub.2]-fixation capabilities of Illinois bundleflower.
Another factor that deserves investigation is the role of mycorrhizal fungi in enhancing uptake of nutrients and [N.sub.2] fixation in Illinois bundleflower. After conducting these studies, we learned that Illinois bundleflower is a host of Gigaspora spp. mycorrhizae (C. Picone, 2002, personal communication). If mycorrhizal inoculum present at Lamberton, where Illinois bundleflower is native, was not available at the other locations where the plant is introduced, differences in [delta][sup.15]N among locations might result. Infection by mycorrhizae generally leads to lower [delta]sup.15]N than noninfected plants (Handley and Scrimgeour, 1997), and would thus lead to higher estimates for %Ndfa. We did not inoculate with mycorrhizae in the experiment designed to determine ISN fractionation (Table 3). Inoculation with appropriate mycorrhizae also may improve access to nutrients that otherwise may limit plant growth.
Overall, %Ndfa was lowest at Rosemount in both years. Little or no [N.sub.2] was fixed at Becker and Rosemount the first year, whereas 30 to 43 kg [ha.sup.-1] was fixed at Lamberton, depending on the measurement method. Dinitrogen fixation did not occur at Rosemount in the second year, whereas apparent [N.sub.2] fixation was 60 to 79 kg [ha.sup.-1] at Becker and 67 to 142 kg [ha.sup.-1] at Lamberton, depending on the measurement method.
Large differences in herbage yield among locations, but similar yields among accessions, indicate that conditions at one or two of the sites were suboptimal for Illinois bundleflower forage production. We expected to find larger differences among accessions in herbage yield, N yield, and apparent [N.sub.2] fixation. We confirmed that indigenous rhizobial strains reduced [delta][sup.15]N in herbage, compared with the inoculant strains. This implies that nodule occupancy should be determined when the natural abundance [sup.15]N method is used at locations that may differ in rhizobial strain presence and/or competitiveness. On the basis of our analysis of nodule occupancy, we recommend that additional rhizobial strains be sought for use as inoculum for this potential forage crop.
Table 1. Soil characteristics at three experimental sites in Minnesota. ([dagger]) Site Year pH SOM ([double Total N dagger]) g [kg.sup.-1] Becker 2000 6.9 ([double 24 0.49 dagger]) 2001 6.7 25 0.46 Rosemount 2000 6.9 47 1.88 2001 6.7 48 1.95 Lamberton 2000 7.4 41 1.64 2001 7.0 46 1.55 Site N[O.sup.-.sub.3]-N P K mg [kg.sup.-1] Becker 6.2 54 121 6.1 50 94 Rosemount 16.3 31 98 16.8 30 100 Lamberton 32.6 19 156 32.6 16 138 ([dagger]) Values reported are means across six replicates at each site. Samples were taken to a depth of 15 cm at planting time in 2000 and when shoots were first visible in 2001. ([double dagger]) SOM, soil organic matter. Table 2. Monthly precipitation and average air temperature means and deviation (dev) from 30-yr (1971-2000) mean (dev) at Lamberton, Rosemount, and Becker, MN, May to September 2000 and 2001 (Minnesota Climatology Working Group, 2002). Becker 2000 2001 Mean Dev Mean Dev Precipitation, mm May 61 -32 86 -7 June 109 5 102 -2 July 106 -1 50 -57 August 40 -67 68 -39 September 10 -80 68 -22 Air temperature, [degrees]C May 15.4 1.4 14.4 0.4 June 17.6 -2.4 19.7 -0.3 July 22.1 0.1 23.0 1.0 August 22.2 2.2 22.6 2.6 September 15.2 0.2 14.9 -0.1 Rosemount 2000 2001 Mean Dev Mean Dev Precipitation, mm May 104 4 116 16 June 109 0 113 4 July 230 128 26 -76 August 79 -22 42 -59 September 15 -61 95 19 Air temperature, [degrees]C May 16.6 1.6 14.7 -0.3 June 18.3 -1.7 19.4 -0.6 July 21.4 -0.6 22.8 0.8 August 21.2 0.2 22.4 1.4 September 15.7 -0.3 15.4 -0.6 Lamberton 2000 2001 Mean Dev Mean Dev Precipitation, mm May 165 86 58 -21 June 60 -28 57 -31 July 117 23 167 73 August 91 20 39 -32 September 14 -76 87 -3 Air temperature, [degrees]C May 15.5 0.5 15.3 0.3 June 19.2 -0.8 20.3 0.3 July 21.0 -0.1 23.1 1.1 August 21.3 0.3 21.5 0.5 September 15.4 -0.6 15.3 -0.7 Table 3. Isotopic fractionation ([[delta].sup.15][N.sub.a]) in whole shoots of 72-d-old Illinois bundleflower plants grown in the absence of soil N. Plants were thus dependent solely on fixed [N.sub.2], except for N originally present in the seed. Rhizobial strains Accession Inoculant Inoculant + indigenous ([dagger]) [per thousand] PNL534 -0.13 [+ or -] 0.15a -0.64 [+ or -] 0.11ab ([double dagger]) PNL539 -0.74 [+ or -] 0.11b -1.59 [+ or -] 0.10c PNL541 -0.65 [+ or -] 0.13ab -0.93 [+ or -] 0.13b Strain mean -0.51 [+ or -] 0.08p -1.05 [+ or -] 0.07q Accession Accession mean [per thousand] PNL534 -0.38 [+ or -] 0.09i PNL539 -1.16 [+ or -] 0.08j PNL541 -0.79 [+ or -] 0.09k Strain mean ([dagger]) Indigenous rhizobial strains obtained from Lamberton, MN. ([double dagger]) Values reported are least squares means [+ or -] standard error. Means followed by the same letter are not significantly different by the Tukey-Kramer approximate honestly significant difference (P < 0.05), with a, b, c for interaction means, i, j, k for overall accession means, and p, q for overall rhizobial strain means. The number of replicates per accession x rhizobial strain combination varied from 5 to 11, depending on plant survival. Table 4. Herbage yield and aboveground N yield for first-year (2000) Illinois bundleflower. ([dagger]) Herbage yield Accession Becker Rosemount Lamberton Mg [ha.sup.-1] PNL534 0.65a 2.30i 3.45p ([double dagger]) PNL539 0.99a 3.24j 3.68p PNL541 1.42b 3.04ij 3.93p Location mean 1.02 2.86 3.69 Wild senna 1.99 3.29 1.84 Aboveground N yield Accession Becker Rosemount Lamberton kg [ha.sup.-1] PNL534 11a 45i 76p PNL539 15a 53i 77p PNL541 24b 60i 87p Location mean 17 53 80 Wild senna 16 46 30 ([dagger]) Values reported are least squares means of six replicate plots. ([double dagger]) Means followed by the same letter are not significantly different by the Tukey-Kramer approximate honestly significant difference (P < 0.05), with a, b for Becker means, i, j for Rosemount means, and p for Lamberton means. Table 5. Herbage vield and aboveground N yield for second-year (2001) Illinois bundleflower. ([dagger]) Herbage yield Accession Becker Rosemount Lamberton Mg [ha.sup.-1] PNL534 4.73a 2.89i 8.03p ([double dagger]) PNL539 4.36a 3.20i 8.61p PNL541 4.44a 2.89i 8.18p Location mean 4.54 2.99 8.27 Wild senna 2.91 3.07 3.68 Aboveground N yield Accession Becker Rosemount Lamberton kg [ha.sup.-1] PNL534 97a 49i 179p PNL539 92a 46i 175p PNL541 97a 48i 186p Location mean 95 48 180 Wild senna 28 35 38 ([dagger]) Values reported are least squares means of six replicate plots. ([double dagger]) Means followed by the same letter are not significantly different by the Tukey-Kramer approximate honestly significant difference (P < 0.05), with a for Becker means, i for Rosemount means, and p for Lamberton means.
The Land Institute (Salina, KS) supported this research through a graduate research fellowship to the senior author, as did the Department of Agronomy and Plant Genetics at the University of Minnesota. Special are thanks are due Becki Tlusty (University of Minnesota) for performing the rep-PCR analysis, Karena Schmidt (USDA-ARS) for assistance with total N analysis, Dr. Lee DeHaan (The Land Institute) for recommending accessions and supplying seed, and Doug Swanson, Eric Ristau, Jim Halgerson, Donn Vellekson, Joshua Larson, Kristin Knopke, Aaron Lorenz, Margaret Bury, and Kaydi Strickler for field and laboratory assistance. The Statistical Consulting Service at the University of Minnesota and Dr. Frank Martin in particular helped with the analysis of this experiment, with funding provided by the Minnesota Agricultural Experiment Station and the College of Liberal Arts.
Abbreviations: DAP, days after planting: PCR, polymerase chain reaction; PNL, Perennial Native Legume; %Ndfa, percentage of N derived from the atmosphere.
(1) Names are necessary to report factually on available data. However, the USDA and the University of Minnesota neither guarantee nor warrant the standard of the product, and use of the name by the USDA and the University of Minnesota implies no approval of the product to the exclusion of others that may be suitable.
Allen, O.N., and E.K. Allen. 1981. The Leguminosae: A sourcebook of characteristics, uses, and nodulation. Univ. of Wisconsin Press, Madison, WI.
Allos, H.F., and W.V. Bartholomew. 1959. Replacement of symbiotic fixation by available nitrogen. Soil Sci. 87:61-56.
Boddey, R.M., M.B. Peoples, B. Palmer, and P.J. Dart. 2000. Use of the [sup.15]N natural abundance technique to quantify biological nitrogen fixation by woody perennials. Nutr. Cycling Agroecosyst. 57:235-270.
Brandon, N.J., and R.A. Date. 1998. Nutrient limitations of clay soils for Desmanthus virgatus. I. What is the cause of chlorosis in field-grown desmanthus on a black earth soil? Trop. Grassl. 32:1-5.
Brandon, N.J., R.A. Date, R.L. Clem, B.A. Robertson, and T.W.G. Graham. 1998. Growth responses of Desrnanthus virgatus to inoculation with Rhizobium strain CB3126. II. A field trial at 4 sites in south-east Queensland. Trop. Grassl. 32:20-27.
Chalk, P.M. 1985. Estimation of [N.sub.2] fixation by isotope dilution: An appraisal of techniques involving [sup.15]N enrichment and their application. Soil Biol. Biochem. 17:389-410.
de Faria, S.M., G.P. Lewis, J.I. Sprent, and J.M. Sutherland. 1989. Occurrence of nodulation in the Leguminosae. New Phytol. 111: 607-19.
DeHaan, L.R., N.J. Ehlke, C.C. Sheaffer, R.L. DeHaan, and D.L. Wyse. 2003. Evaluation of diversity among and within accessions of Illinois bundleflower. Crop Sci. 43:1528-1537.
Dovel, R.L., M.A. Hussey, and E.C. Holt. 1990. Establishment and survival of Illinois bundleflower interseeded into an established kleingrass pasture. J. Range Manage. 43:153-156.
Handley, L.L., and C.M. Scrimgeour. 1997. Terrestrial plant ecology and [sup.15]N natural abundance: The present limits to interpretation for uncultivated systems with original data from a Scottish old field. Adv. Ecol. Res. 27:133-212.
Hojjati, S.M., W.C. Templeton, Jr., and T.H. Taylor. 1978. Nitrogen fertilization in establishing forage legumes. Agron. J. 70:429433.
Irwin, H.S., and R.C. Barneby. 1982. The American Cassiinae: A synoptical revision of Leguminosae tribe Cassieae subtribe Cassiinae in the New World. The New York Botanical Garden, New York.
Kulakow, P.A., L.L. Benson, and J.G. Vail. 1990. Prospects for domesticating Illinois bundleflower, p. 168-171. In J. Janick and J.E. Simon (ed.) Advances in new crops. Timber Press, Portland, OR.
Lamb, J.F.S., D.K. Barnes, M.P. Russelle, C.P. Vance, G.H. Heichel, and K.I. Henjum. 1995. Ineffectively and effectively nodulated alfalfas demonstrate biological nitrogen fixation continues with high nitrogen fertilization. Crop Sci. 35:153-157.
Ledgard, S.F. 1989. Nutrition, moisture and rhizobial strain influence isotopic fractionation during [N.sub.2] fixation in pasture legumes. Soil Biol. Biochem. 21:65-68.
Luckow, M. 1993. Monograph of Desmanthus (Leguminosae-Mimosoideae). Syst. Bot. Monogr. 38. Am. Soc. of Plant Taxonomists, Ann Arbor, MI.
Minnesota Agricultural Experiment Station. 2002. Minnesota varietal trials results. Publ. 106. Minnesota Agric. Exp. Stn., Univ. of Minnesota, Minneapolis.
Minnesota Climatology Working Group. 2002. Monthly temperature and precipitation summary tables [Online]. Available at: http:// climate.umn.edu/cawap/monsum/monsum.asp (cited 1 Dec. 2002; modified 17 Nov. 2003; verified 17 Nov. 2003). Minnesota Climatology Working Group, MN.
Oelhert, G.W. 2000. A first course in design and analysis of experiments. W.H. Freeman and Co., New York.
Posler, G.L., A.W. Lenssen, and G.L. Fine. 1993. Forage yield, quality, compatibility, and persistence of warm-season grass-legume mixtures. Agron. J. 85:554-560.
Rennie, R.J. 1984. Comparison of N balance and [sup.15]N isotope dilution to quantify [N.sub.2] fixation in field-grown legumes. Agron. J. 76:785-790.
SAS Institute. 1996. SAS system for mixed models. SAS Inst., Cary, NC.
Schneider, M., and F.J. de Bruijn. 1996. Rep-PCR mediated genomic fingerprinting of rhizobia and computer-assisted phylogenetic pattern analysis. World J. Microbiol. Biotechnol. 12:163-174.
Seguin, P., M.P. Russelle, C.C. Sheaffer, N.J. Ehlke, and P.H. Graham. 2000. Dinitrogen fixation in Kura clover and birdsfoot trefoil. Agron. J. 92:1216-1220.
Shearer, G., and D.H. Kohl. 1986. [N.sub.2] fixation in field settings: Estimations based on natural [sup.15]N abundance. Aust. J. Plant Physiol. 13: 699-56.
Somasegaran, P., and H.J. Hoben. 1994. Handbook for rhizobia: Methods in legume- rhizobium technology. Springer-Verlag, New York.
Springer, T.L., G.E. Aiken, and R.W. McNew. 2001. Combining ability of binary mixtures of native, warm-season grasses and legumes. Crop Sci. 41:818-823.
Streeter, J. 1988. Inhibition of legume nodule formation and [N.sub.2] fixation by nitrate. Crit. Rev. Plant Sci. 7:1-23.
Veraslovic, J., M. Schneider, F.J. de Bruijn, and J.P. Lupski. 1994. Genomic fingerprinting of bacteria using repetitive sequence-based polymerase chain reaction. Methods Mol. Cell. Biol. 5:25-40.
Jaehyun Byun, Craig C. Sheaffer, * Michael P. Russelle, Nancy J. Ehlke, Donald L. Wyse, and Peter H. Graham
J. Byun, C.C. Sheaffer, N.J. Ehlke, and D.L. Wyse, Dep. of Agronomy and Plant Genetics, 411 Borlaug Hall; M.P. Russelle, USDA-ARS Plant Science Res. Unit and Dep. of Soil, Water, and Climate, 439 Borlaug Hall; and P.H. Graham, Dep. of Soil, Water, and Climate, 439 Borlaug Hall, Univ. of Minnesota, 1991 Upper Buford Circle, St. Paul, MN 55108. Received 22 Nov. 2002. * Corresponding author (email@example.com).
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Crop Physiology & Metabolism|
|Author:||Byun, Jaehyun; Sheaffer, Craig C.; Russelle, Michael P.; Ehlke, Nancy J.; Wyse, Donald L.; Graham, P|
|Date:||Mar 1, 2004|
|Previous Article:||Soybean [N.sub.2] fixation estimates, ureide concentration, and yield responses to drought.|
|Next Article:||Comparison of uptake and distribution of cadmium in different cultivars of bread and durum wheat.|