Biological Nitrogen Fixation, Carbon assimilation and plant performance of Lotus tenuis, contribute to dene a strategic role in the lowlands in the Salado River Basin (Argentina)

Background and Aims Forage production constitutes a great challenge for the Salado River Basin (Argentina), the major area devoted to livestock in the country. The successful naturalization of the European legume Lotus tenuis has been a productive and environmental relevance for the region. This study aims to evaluate its strategic contribution, reporting for the rst time the B value for this species in these marginal environments for agriculture. Methods The 15 N natural abundance method was used to evaluate the BNF of L. tenuis at soil plots and non-leguminous weed species in the same plots were utilized as reference plants. The assays included determinations of the 13 C isotope, as well. Simultaneously, evaluations were carried out in the greenhouse for the determination of the B value of L. tenuis and the relative reproducibility of the eld experiments. Results The results obtained demonstrated that the L. tenuis promotion is accompanied by an increase in forage quality, due to the predominance of this C3 legume species, over C4 species. Moreover, its contribution to N inputs to the system, through the BNF with native rhizobia demonstrated to be about 80%, a highly relevant percentage for these constrained agroecosystems. This work supports the hypothesis that L. tenuis promotion plays a strategic role in the sustainability of the ecosystem, especially in soil constrained areas. Joined with data previously published, the results obtained contribute to support the criteria that dene the naturalized legume playing a strategic role in the sustainability of agriculture marginal soils.


Introduction
The Flooding Pampa (Argentina) constitutes a heterogeneous ecosystem covering approximately 90000 Km 2 area in South America (Imbellone et al. 2021). Most of the region is still covered by grasslands that are the main forage resource for cattle breeding (León et al. 1984;Cid et al. 2011) There are few native legumes in these grasslands because this area is characterized by severe phosphorus de ciency, high alkalinity and salinity levels, together with periodic exposure to waterlogging (Antonelli et al. 2019). Lotus tenuis Waldst et Kit (ex Lotus glaber Mill) is an introduced European legume species that has added sustainability improving the quality of lowlands in the Flooding Pampa (Escaray et al. 2012). Due to its ability of adaptation to restrictive conditions, L. tenuis became quickly naturalized and spread into the plains and humid lowland plant communities (Miñón et al. 1990; Escaray et al. 2012). L. tenuis not only contributes to forage production and in uences the growth of associated plant species in this area, but also contributes to restoration and phytoremediation of degraded environments (Vignolio et al. 1999).
This bene cious effect on the vegetal community observed justify its inclusion within the keystone species because of the large effects on community structure and ecosystem function (González-Robles et al. 2020;Campestre et al. 2020). This effect is large relative to abundance of forage supply (i.e., high fodder community importance) (Bailleres et al 2020). Moreover, L. tenuis have a lower phosphorus requirement than other legumes of temperate climate, reaches its maximum biomass in summer with high forage quality and maintain symbiotic associations with rhizobia, transferring the xed N to the accompanying grasses (Vignolio et al. 2010). Due to aforementioned characteristics, Lotus promotion has been conceived as an appealing alternative to meet the needs for cattle production during summer season in soils of the Flooding Pampa (Antonelli et al. 2016;Bailleres et al. 2020). This agricultural practice consists of removing broad leaf weeds or aggressive grass weeds using herbicides, and therefore improving Lotus species establishment ability (Díaz et al. 2005). This practice does not alter soil bacterial communities and led to increased fungal diversity (Nieva et al. 2016(Nieva et al. , 2018(Nieva et al. , 2019. The Rhizobium-legume symbiosis provides an alternative to N fertilizers to balance N losses in the environment. Several works were carried out on the diversity of native rhizobia and their ability to nodulate this naturalized L. tenuis species in the Flooding Pampa (Fulchieri et al. 2001;Estrella et al. 2009;Sannazzaro et al. 2011). However, the symbiotic capacity of the Mesorhizobium -L. tenuis association to x biological N 2 in the eld using the N isotope discrimination technique, has never been determined in this area.
The overall bene ts of including a N 2 -xing L. tenuis promotion in marginal areas of the Flooding Pampa cannot be assessed unless a reliable and accurate eld measurement is made of the levels of xation achieved. The 15 N natural abundance method, discussed in detail elsewhere (Peoples et al. 1989;Shearer and Kohl 1989;Guimarães et al. 2008;Pauferro et al. 2010), is the most appropriate technique recommended for crops and pasture legumes because it gives an overall estimate of the contribution of biological N 2 xation (BNF) up to the time of sampling (Cadisch et al. 2000;Boddey et al. 2009). In addition, particularly in natural ecosystems, disturbance of the system is not required and measurements may be made on samples dried in the eld (Shearer and Kohl 1989), determining the proportion of N in the legume plant derived from the air (%Ndfa) and comparing its 15 N natural abundance value (δ 15 N) with that of a non-xing companion species also denominated as reference plant (Pate et al. 1994).
Furthermore, in naturalized species like L. tenuis, the δ 15 N in N 2 xing plants grown with N 2 in air as the only N source ('B' value), must be determined simultaneously in order to account for discrimination against 15 N during N 2 xation (Carlsson et al. 2006;. As Lotus species are suggested to belong to the C 3 group of plants, determination of the natural abundance of 13 C (δ 13 C) may provide useful information on carbon assimilation and plant performance during the growth period (Kurdali and Al-Shamma'a 2009). Despite the importance that L. tenuis implies for livestock activity and restoration of degraded environments in the Flooding Pampa, there is a lack in the bibliography of reference plants and 'B' values for this species in combination with rhizobia isolated from elds. For these reasons, our aims in this study were to: (1) identify and select a potential reference plant species for the application of the 15 N natural abundance technique to assess the N 2 xing performances of a L. tenuis promotion; (2) calculate the 'B' value of L. tenuis species in different phenological stages; (3) determine the δ 13 C to understand variations in carbon assimilation; (4) evaluate the symbiotic capacity of L. tenuis and native rhizobia through the determination of %Ndfa and the amount of N 2 xed in the eld site in the Flooding Pampa. As well as determine the %Ndfa of a promoted L. tenuis in the eld, we aimed to evaluate the possibility of making future determinations in L. tenuis under greenhouse assays for the speci c study of the effect on %Ndfa with commercial strains of rhizobia and co-inoculation tests with growth-promoting bacteria. For this reason, the determination of %Ndfa under greenhouse was carried out in parallel to the eld study to check the reproducibility between both evaluations.

Study sites description
The studies were conducted simultaneously in two sites during the experimental period September 2017-March 2018: 1-In a 10-years promoted L. tenuis pasture at the Chacra Experimental Integrada Chascomús (CEICh-Ministerio de Desarrollo Agrario de la Provincia de Buenos Aires -Instituto Nacional de Tecnología Agropecuaria, Argentina, 35º45'27"S, 58°3'18"W), which is located in the Flooding Pampa region. This region has a temperate sub-humid climate with mean temperatures averaging 8.5°C in winter and 21.5°C in summer, and annual rainfall 850-1050 mm. Short oods of 5-7 cm depth occurs at the beginning of almost every spring, sometimes followed by severe droughts may in summer. The experimental area is mostly covered by a Typic Natraquoll, (US Soil Taxonomy), whose A horizon has 35 g kg -1 organic matter content and 0.22 mg kg -1 of extractable Fe. Its natric and clays horizon appears at 0.17 m and contains 533 g kg -1 clay (Lavado and Taboada 1988). Before promotion, plant communities were mainly composed by grasses (Festuca arundinacea, Thinopyrum ponticum, Cynodon dactylon, L. tenuis and Sporobolus indicus) and exotic dicots, mainly Compositae species. Native legumes were largely absent. by. 2-In a naturally-lit greenhouse located in the Instituto Tecnológico Chascomús (INTECh), twenty kilometers separated from the CEICh. Temperature and irradiance were recorded throughout the experiment. Mean temperature was 26.3 ± 8.2°C and mean maximum irradiance per day was 1150 ± 225 µmol m -2 s -1 .

Pasture measurement and oristic composition
Plant biomass was harvested from ve 0.25 m 2 quadrants randomly located in the paddock. Plant samples were collected during three seasons (spring, summer and autumn) by clipping approximately 1 cm above the soil surface. Samples were dried at 70°C to constant weight and dry biomass was calculated per hectare.
For forage quality determinations, L. tenuis plants were clipped in spring, summer and autumn. The following parameters were evaluated: Crude Protein (CP) % by Kjeldahl method and Digestibility of Dry Matter (DDM) % by (Tilley and Terry 1963). Samples were analyzed at the Laboratory of Animal Nutrition (INTA Balcarce).
For determination of oristic composition, a linear transect with 25 points on a 10-m-long line was used (Gaucherand and Lavorel 2007). For this, a stick was planted vertically every 40 cm and at that point each individual plant was classi ed as Annual grass, Perennial grass, weed, L. tenuis or Trifolium repens.
Experiment 1: Promotion of L. tenuis pasture and determination of total N, C, δ 13 C and δ 15 N On a surface area of 2 ha, Glyphosate (N-(phosphonomethyl) glycine; 3.5 L/ha), followed by two applications of 2,4 DB(4-(2,4-dichlorophenoxy) butyric acid, 1 L/ha) and a single dose of Quizalofopp-ethyl (Ethyl(R) − 2-[4-(6-chloro-2-quinoxalyloxy) phenoxy] propionate; 1.2 L/ha) were applied, in six or seven annual cycles from June to August. After 4 or 5 years of herbicide application, plant species composition in promoted paddocks shifted, and L. tenuis became the dominant species (Nieva et al. 2016(Nieva et al. , 2018. L. tenuis shoots were harvested in three different phenological stages: 1-Early bloom (EB), 90 days after germination, 2-Full bloom (FB), 120 days after germination and 3-Regrowth (RG), 60 days post-full bloom (after simulating grazing). Shoots were placed into paper bags and oven-dried at 60°C for 72 hours to constant weight. Samples were reduced into a ne powder using a roller mill and homogenized for the analysis of total N, C, δ 13 C and δ 15 N.
For the estimation of the %Ndfa is necessary to compare its δ 15 N value with that of a non-xing species (Pate et al. 1994) and the selection of more than one non-xing species is recommended . In this study, Conyza bonariensis and Lythrum sp were tested as reference species because of their close similarity with the legume in terms of phenology and rooting pro le with soil depth, adjacent closeness to the legume in the site and for its accompaniment throughout the growth cycle. Shoot reference materials were harvested at the same time and at the same three phenological stages as the legume and were processed in the same way for total N and δ 15 N determination. Identical treatment was performed to process the samples used for 13 C estimation. The results of δ 13 C and δ 15 N were expressed as parts per thousand (‰) deviations, in the ratio of the heavy to the light isotope of each element, from the international standards (Vienna Pee Dee Belemnite, V-PDB, for 13 C and atmospheric N 2 for 15 N).

Experiment 2: Determination of L. tenuis 'B' value for the Flooding Pampa Region
The 'B' value is the δ 15 N of shoots of legumes that are fully dependent upon N 2 xation and sampled at the same growth stage and with the same rhizobia strains as the eld plants (Unkovich and Baldock 2008). It is best determined on plants grown in a greenhouse, so the determination of L. tenuis 'B' value was performed in a naturally-lit greenhouse simultaneously with the eld trial and estimated for each phenological period sampled (EB, FB and RG). Harvested seeds from the promoted L. tenuis were scari ed and disinfected with sulphuric acid for 3 min, washed 10 times with sterile distilled water, and sown in Petri dishes containing water/agar (0.8%). They were incubated for 7 d in a growth chamber with a 16/8 h day/night cycle at 24/21 ± 2°C and 55/65 ± 5% relative humidity. Light (at an intensity of 250 µmol m -2 s -1 ) was provided by Grolux F 40 W uorescent lamps. L. tenuis seedlings were grown in 4 L pots with sterile sand and irrigated with N-free nutrient solution (Rigaud and Puppo 1975). Inoculation was achieved as ) using a soil suspension with the native rhizobia present in the promoted eld site. For this, soil was collected from the top 20 cm layer from many points in the promoted area and mixed thoroughly. Then, 1 g of soil was mixed with 100 ml sterile water, and 3 ml of soil suspension per seedling were used as inoculum. At each harvest period, shoot material was placed into paper bags and oven-dried, grounded, homogenized and analyzed for δ 15 N. Experiment 3: Total N, C, δ 13 C and δ 15 N determination of L. tenuis grown in pots under greenhouse Harvested seeds from the promoted L. tenuis were scari ed and incubated in a growth chamber until germination. Seedlings were transferred to the greenhouse into 4 L pots containing soil obtained from the top 20 cm of the promoted L. tenuis land horizon. Seedlings were irrigated with water from a rain water harvesting system. In parallel, other 4 L pots were lled with promoted soil, irrigated and stimulated to germinate its seed bank. When the seeds germinated, all species except the ones previously selected as reference plants were removed, so that the ones kept developed accompanying the L. tenuis seedlings. Then, reference plants and L. tenuis shoot materials were harvested at the same three phenological stages, placed into paper bags, oven-dried, grounded and homogenized for total N, C, δ 13 C and δ 15 N determination.

Estimates of Biological Nitrogen Fixation
For %Ndfa estimation, samples were reduced into a ne powder using a roller mill and homogenized for analysis. Aliquots were loaded into a Thermo Delta Advantage isotope ratio mass spectrometer couple with a Flash 2000 Elemental Analyser at the Laboratorio de Isótopos Estables en Ciencias Ambientales (LIECA, Argentine) to obtain the N and C isotope ratio and the total N and C content. Natural abundances of 15 N and 13 C were expressed using 'delta' notation (‰): Where; R sample and R standard are the 15 N: 14 N and 13 C: 12 C ratios of samples and of the standards, which are atmospheric N 2 , for N, and Pee Dee Belemnite, for C.
Although δ provides information on the 13 C/ 12 C of tissues, it is often preferable to express the values as the leaf carbon isotope discrimination (∆) (Kurdali and Al-Shamma'a 2009): ∆ = (δ 13 C air − δ 13 C sample ) / (1 − δ 13 C sample /1000) Where δ 13 C air is the δ 13 C value of air (− 8‰) and δ 13 C sample is the measured value in the plant.
Estimates of the %Ndfa were made using the 15 N natural abundance technique. The proportion of xed N in the plant was calculated using the following formula (Shearer and Kohl 1989)

Experimental Design And Statistical Analysis
A completely randomized experimental design was used in all cases and all measurements were performed on 4 plants (= 4 biological replicates for each experiment and species). Data were subjected to t-tests for differences between experimental sites and one-way ANOVA analysis for differences between harvest times for each experimental site. Duncan`s test was used for multiple comparisons (P < 0.05). All statistical analysis was performed using the INFOSTAT statistical software package (Di Rienzo et al. 2010) (Grupo InfoStat, FCA, Universidad Nacional de Córdoba, Argentina,).

Determination of plant biodiversity in pastures and forage quality
Floristic diversity was modi ed in both, the promoted Lotus and in the grassland sites at all seasons (Table 1). In the grassland, the presence of legumes was scarce during Spring and Summer, but a slight appearance of Trifolium repens (8.00%) and L. tenuis (8.66%) was observed in Autumn. Perennial grasses clearly prevailed (67.00% -75.00%) over annual grasses (8.66% -19.33%). As expected, in the promoted Lotus, L. tenuis was the predominant species during the three studied seasons (61.00% -74.00%). T. repens was only observed during Spring and Autumn (2.33% and 11.33%, respectively). No differences in biomass were observed between promoted Lotus and grassland in Spring and Summer (Table 2). Unlike, higher biomass was measured in grassland in Autumn. When summed, no difference was observed in annual biomass between promoted Lotus and grasslands sites (p= 0.286). However, better forage quality was observed in the promoted Lotus situation (Table 3). L. tenuis not only provided more CP compared to the grassland, but also a resource with higher DDM. Moreover, higher than 17% CP values were observed at promoted Lotus at all seasons, while only 12% CP value was observed at grassland site during spring. Differences in CP (17.53% vs 8.33% respectively) and DDM (71.83% vs 60.50%) were greater in Summer (Table 3).
Modi cation of δ 13 C and ∆ 13 C in greenhouse and eld assays according to Lotus tenuis phenology.
As observed in Figure 1, during EB and FB, no δ 13 C differences were obtained between sites for L. tenuis (δ 13 C approximately -28‰). For them, differences were only observed during RG stage with more negative δ 13 C values in the eld (-29‰). Regarding greenhouse assays, a more negative value was obtained during FB (-29.41‰) compared to its value during EB (-28.36‰) and RG (-28.76‰), while for eld assays a more negative δ 13 C was obtained during RG (-29.32‰) compared to EB (-28.70‰) and FB (-28.75‰). On the other hand, the same trend was observed in Figure 2 for the ∆ 13 C, where differences were only observed during RG stage. For greenhouse assays a higher ∆ 13 C was observed during FB stage compared to EB and RG, while no differences were observed in the eld during the three phenological stages.
Modi cation of δ 15 N in greenhouse and eld assays according to Lotus tenuis phenology.  were obtained for C. bonaerensis in this site. δ 15 N of the B value did not differ between EB and FB (-1.9‰; p= 0.910). In L. tenuis, δ 15 N values were negative in the greenhouse (-0.26‰) and positive in the eld (1.92‰) and did differ between them ( Figure 3).
Proportion of N derived from the air (%Ndfa) in greenhouse and eld assays according to Lotus tenuis phenology.
Regarding N concentration (Figure 4 A) L. tenuis values did not differ between the greenhouse and the eld in the EB and the FB stages ( Figure 4). In contrast, in the RG stage, N concentration was higher in the eld than in greenhouse. N was accumulated differently according to the phenological stage. In the greenhouse, it was signi cantly (P < 0.05) lower in the RG than in the EB and FB stages. Unlike, a higher 15 N concentration was observed in the eld in the RG stage. With regard to C accumulation, similar values were observed in EB and RG stages, while lower C accumulation was observed in the eld in the RG stage (Figure 4 B). The resulting C/N ratios were similar between greenhouse and eld sites at EB, slightly and not signi cantly (P= 0.320) higher in the eld at FB, and signi cantly (P = 0.001) higher in the greenhouse at RG (Figure 4 C). δ 15 N values in the two non-leguminous reference species.
The two non-leguminous reference species selected -C. bonaerensis and Lythrum sp.-achieved mean δ 15 N values between 4-10‰ depending on the site, phenological stage and plant species, while the mixture of both reference species achieved mean δ 15 N values between 5-7‰, depending on site and phenological stage (Figure 3). %Ndfa values in L. tenuis and reference plants according to phenological stages There was no signi cant effect from both reference plant species (%Ndfa= 80%) in the EB and FB phenological stages, as well as from the mixture of both (Figure 5 A, B). For this reason, the use of some of the two species or the mixture of both was proved indistinct in these stages. Because Lythrum sp. was not found in the RG phenological stage, C. bonaeriensis species was used as the only reference plant (Figure 5 C). Interesting to note, %Ndfa results for L. tenuis in the greenhouse were similar during EB, FB and RG stages ( Figure 6). The same was observed in the eld during EB and FB. Unlike, during RG lower %Ndfa was observed in the eld than in the greenhouse ( Figure 6).

Discussion
Improvement of grasslands forage contribution, by L. tenuis promoting Temperate pastures of South America regions are comprised mainly by a few C3 grass and legume species, offering a quite good biomass production in winter and providing forage for raising cattle in a season when the productivity of native grasslands is scarce (Bresciano et al. 2019). The main ecosystem limitation is their poor ability to maintain high levels of productivity during the summer season (Tejera et al. 2016). Thus, the presence of C4 species in these environments plays an underlying role providing a chance to improve biomass stability ensuring year-round productivity, particularly, in the Flooding Pampa where speci c edaphic limitations such as high levels of halo-hydromorphism are very common. As shown by Table 1, the grassland was mainly composed by annual and perennial C4 grasses such as Paspalum dilatatum, Sporobolus indicus and Cynodon dactylon. This agrees with previous works carried out in the region, showing that the prevalence of C4 over C3 grasses is mainly due to edaphic conditions and specially, the topsoil pH gradients (Perelman et al. 2001(Perelman et al. , 2005Cid et al. 2011) and its relative abundances can change at a particular site and such changes may have direct consequences for ecosystem processes such as net primary production along with many other ecosystem services (Lattanzi 2010;Kim et al. 2015). Despite its contributions, this forage resource (C4 grasses) has poor quality, as shown by its low CP % and DDM % values (Table 3), which does not cover an e cient cattle production in the region, causing either a considerable increase in the need for feed supplements or a decrease in their reproductive performance (Grigera et al. 2007). Plant physiological status and their photosynthetic pathways can be assessed by considering its carbon isotope discrimination values (Brüggemann et al. 2011;Luo et al. 2021). In this sense, the L. tenuis δ 13 C and ∆ 13 C values obtained from around − 28‰ and 20‰ respectively (Figs. 1 and 2, respectively) are considered typical of C3 plants following the Calvin cycle (Farquhar et al. 1989;Sun et al. 2011). The introduction of C3 forage legumes such as L. tenuis can modify the C3:C4 grass cover ratio, thus resulting in a richer nutritional forage resource (Table 1 and Table 3, respectively). It is also well reported that legumes can contribute improving not only the nutritive value of the pasture, but also the herbage yield of the improved pastures (Del Pino et al. 2016;Vignolio et al. 2016) and sustainability beyond current extents (Muir et al. 2014). Besides, carbon sequestration can be achieved after legume introduction (overseeding) in grazed natural grasslands, depending on grazing management practices, as was reported by Bondaruk (2020) in a case study in Argentina on commercial farms using Lotus subbi orus. Successful naturalized L. tenuis stands (Insausti et al. 2001;García and Mendoza 2008), as well as the consociation between L. tenuis and Paspalum dilatatum (Striker et al. 2008;Vignolio and Fernández 2011) were found as keys for future forage contributions for these highly constraining environments. However, a trade-off between increased pasture production and decreased vegetation stability (long-term integrity) may be operating in legumeimproved grasslands, thus requiring further studies on the effects of other mechanisms such as grazing management options (Jaurena et al. 2016;Vecchio et al. 2019).

Parameters for the determination of the BNF by Lotus tenuis
The correct estimation and quanti cation of BNF by legumes depends on the applied methodology (Hardarson 1993). Until date, most of the work undertaken on the evaluation of N 2 xation by L. tenuis species has been performed in the framework of forage mixtures and with the acetylene reduction technique or nitrogenase activity throwing a theoretical N 2 xation (Re and Escuder 1998). In this work we have obtained the required parameters to allow us an accurate value of N 2 xed by this legume, and through it, obtain a better understanding of its economic and environmental importance for the region.
The %N accumulated by eld plants was approximately 2.7% during Spring (47 kg N ha − 1 ) and 2.15% during Summer (62 kg N ha − 1 ). This N 2 xation potential is in the range of previous works using the nitrogenase activity method, such as 27-42 kg N ha − 1 per year in grazed tall fescue / L. tenuis swards growing on an alkaline hydromorphic soil, or 14-59 kg N ha − 1 per year tall fescue / white clover swards (Díaz et al. 2005). Furthermore, Danso et al. (1991) reported a 91 % of Ndfa by L. corniculatus when tall fescue was considered as the reference plant. In our study, the amount of Ndfa was determined to be 80% without signi cant differences between the reference plants considered in the calculation. The amount of nitrogen in the soil can also be increased with the presence of L. tenuis, as found by (Vignolio et al. 2010) two years after planting, with the soil N being increased from 30.5 mg/kg to 66.4 mg/kg.
A very important aspect of N 2 xation estimations using the 15 N natural abundance technique is a good selection of the reference plants. In this selection is important to take in mind that the δ 15 N of soil can vary between N pools, soil depths, and over time, the reference plant should follow the same dynamics in N uptake and take N from the same soil N pools as the N 2 xing species (Carlsson et al. 2006;). In our trials, there were no signi cant differences according to the reference plant species used in the contribution of BNF of L. tenuis inoculated with native rhizobia (% Ndfa = 80%) for the EB and FB phenological stages (Fig. 3). Furthermore, no differences were observed using the species mix either (Fig. 3). For this reason, the use of some of the two species or the mixture of both is indistinct in these stages. However, although the bibliography recommends using a mixture of reference species, Lythrum sp. was not found in the RG phenological stage and therefore the species C. bonaerensis was used as the reference plant. Besides, quantifying biologically xed nitrogen (BNF) by legumes through the 15 N natural abundance technique requires the correct determination of the B value which can vary among species (also among accessions of the same plant species), environmental conditions and plant physiological status (Boddey et al. 2000;Peoples et al. 2009;Nebiyu et al. 2014). Most of the B values determined for legumes are usually in the range of 0 and − 2‰ with few exceptions (Okito et al. 2004). In our study, the B values obtained for L. tenuis in two phenological stages were − 2.1 and − 1.7 for EB and FB, respectively, being these within the range previously mentioned. The aforementioned B values constitute the rst report for L. tenuis in two phenological stages in these speci c environments of the Salado River Basin.
BNF from L. tenuis, in different phenological states in greenhouse and eld sites Except for greenhouse assays during RG stage, N concentration in L. tenuis leaves was always higher than that in both non-xing plants or the mixture of them (data not shown). This result demonstrated that the 15 N natural abundance method to estimate %Ndfa should be circumscribed to EB and FB phenological stages in greenhouse conditions. The δ 15 N value was − 1.37‰ ± 0.6, which was close to the atmospheric value but different from that of the available soil N determined in a large number of non-N 2 -xing tree (+ 6.4‰ ± 0.6) pointing to a major contribution of N 2 xing in this plant species. As has been discussed in a previous publication , there can be some issues with applying the 15 N natural abundance technique to regrowth because of the potential complication of N recycling in the N 2 -xing species. Our results also con rm the importance of forage management, which also impacts on N xation, especially grazing management practices that in uence plant regrowth (Vecchio et al. 2019). In this sense, protocols should be designed on the basis of the legume "nutritional cycles", avoiding the overgrazing in critical phenological periods where the relationship sink/source of C and N must be considered in a framework of grasslands conservation (Abdalla et al. 2018). As was reported in Medicago sativa, total or partial removal of the photosynthetic area results in the mobilization of C and N reserves from roots to shoots generating an inversion of source and sink organs (Teixeira et al. 2007). The management system imposed and specially the defoliation periods will affect the patter of reserves accumulation and subsequent regrowth (Vignolio et al. 2018;Mitchell et al. 2020). After shoot removal, regrowth of the new shoots in L. tenuis plants must be supported by non-structural carbohydrates along with N compounds (such as proteins and amino acids) stored in the taproots (Striker et al. 2011), specially during early regrowth (Avice et al. 1996). These non-structural carbohydrates are used to support respiration of the crown and taproots until photosynthesis is re-established. Although a larger reduction in non-structural carbohydrates in comparison with N root reserves has been observed, regrowth is linked to N reserves in roots rather than C reserves (Striker et al. 2011;Aranjuelo et al. 2014). Moreover, is accepted that a reduction in N 2 -xation happens in response to a lack of nodulation ability or increased N availability from regrowth stage because of a mayor input of N from the soil. Studies conducted in the past show that, after shoots are removed, nodule functioning is reduced during the rst days of regrowth and is concomitant with an 88% decrease in N xation 24 h after cutting (Aranjuelo et al. 2014). In addition, as was reported previously (Skinner et al. 1999) nutrient uptake and photosynthesis can be greatly reduced following defoliation.

Conclusions
We consider that L. tenuis could aid rehabilitation of disturbed or marginal edaphic areas by adding xed atmospheric nitrogen that can be used by other plants (non-legumes) including typical C4 species growing in association. However, many leguminous species do not grow well under adverse soil conditions such as ooding soils. The knowledge that some leguminous plants, mainly Lotus spp., are growing well in these constrained soils, focused our attention on studying the BFN contribution in these harsh conditions. The results obtained con rm that in the Salado River Basin, the naturalization of L. tenuis to more constrained areas where there is no signi cant presence of native legumes could substantially modify the volume of N cycling, in uencing, also, the carbon sequestration rates.
As far as we know, this is the rst study where the N 2 contribution of L. tenuis in symbiotic association with native rhizobia is effectively determined in the Flooding Pampa. Also, no information on seasonal variation or levels of N 2 xation of L. tenuis pastures was available. The observation that L. tenuis promotion is an appealing forage alternative in marginal areas to meet the nutritional requirements of livestock in the summer season has been con rmed by the actual data. This, along with data previously published, brings another parameter to validate L. tenuis playing a signi cant role in the constrained areas in the Salado River Basin and, in consequence, on the cow productivity in the most important region devoted to cattle production in Argentina.

Declarations
Competing Interest statement: The authors declare that they have no competing interests.  Figure 1 Modi cation of δ13C in Greenhouse (white bars) and Field (grey bars) assays according to Lotus tenuis phenology (Early Bloom, Full Bloom and Regrowth). Bars with the same uppercase (Greenhouse analysis) and lowercase (Field analysis) letter are not statistically different (Duncan test; P < 0.05). Bars with asterisks represent signi cant differences between study sites in a speci c phenological stage (T test; P < 0.05).

Figure 2
Modi cation of ∆13C in greenhouse (white bars) and eld (grey bars) assays according to Lotus tenuis phenology (Early Bloom, Full Bloom and Regrowth). Bars with the same uppercase (Greenhouse analysis) and lowercase (Field analysis) letter are not statistically different (Duncan test; P < 0.05). Bars with asterisks represent signi cant differences between study sites in a speci c phenological stage (T test; P < 0.05). 0.05). Bars with asterisks represent signi cant differences between study sites in a speci c phenological stage (T test; P < 0.05). . Bars with asterisks represent signi cant differences between study sites (T test; P < 0.05).