These problems necessitate the development and use of alternative sources of N. One such alternative is the use of biological nitrogen fixation (BNF), a process in which atmospheric N is converted into a plant available form of N by free living and symbiotic N₂ fixing microorganism. BNF is a cost effective, sustainable and eco-friendly means of supplying N to plants (Hungria and Vargas, 2000). The scientific exploitation of BNF can greatly decrease our dependence on artificially produced commercial N fertilizer and improve the quality and quantity of internal resources (Arujo et al., 2012).

The use of leguminous green manure crops and trees is one of the mechanisms of taking advantage of BNF. Legumes belonging to the family of Fabaceae form symbiotic association in their roots with a group of microorganisms which belong to the genus Rhizobium. In the process of association the roots of legumes and Rhizobia form a structure known as a nodule in which atmospheric N₂ fixed and converted into plant useable form them and for none N₂ fixing crop (Peoples and Craswell, 1992). Sesbania (Sesbania sesban (L) Merr.) is one of the most important and widely used green manure and improved fallow species to replenish N depleted soils by farmers in eastern and southern Africa (Makatiani and Odee, 2007). It is a fast growing, leguminous N₂-fixing, multi-purpose tree adapted to subtropical and tropical environments (Makatiani and Odee, 2007). When sesbania biomass is incorporated into the soil as green manure, it decomposes fast and release substantial amounts of crop available N and organic carbon into the soil (Nigussie and Alemayehu, 2013). For instance, Weerakoon (1989) reported that green biomass of Sesbania sesban applied at 4.4 tha^-1 with N equivalence value 83 kg N ha^-1 increased the grain yield of maize from 1.9 t ha^-1 in the unfertilized control plot to 3.9 t ha^-1 in Sri Lanka.

Sesbania was introduced to Ethiopia in the 1970s and thought as the most promising species in the highlands due to its N₂-fixing ability, deep rooting tree with good quality foliage to be used as feed for animals and protein supplement (Mekoya et al., 2009) and for erosion control and soil fertility restoration (Degefu et al., 2011). It can be integrated into the farming system for soil fertility improvement as alley cropping, where its biomass can be pruned and incorporated into the soil, or for the improvement of the fallow system.

The high value of sesbania as indicated above as an organic N fertilizer source for soil fertility improvement and/or as forage is due to its N-fixing ability in symbiotic association with bacteria, commonly known as rhizobia (Wolde-meskel et al., 2004a). Through this association sesbania can fix up to 542 kg N ha^-1 year^-1 (Shaheen et al., 2004). Similarly, Degefu et al. (2011) reported that N-fixation by sesbania is in the range of 500 to 600 kg N ha^-1 year ^-1.

However, the amounts of N fixed by tree legumes, including sesbania among others, depends on the symbiotic effectiveness of rhizobia nodulating the particular legume. This is due to the fact that there are wide variations among legumes in their specificity for a particular rhizobia strain. Soils may contain several types of rhizobia and yet a strain specific and effectively nodulating a particular legume species may be absent or the population of the appropriate strain could be too small for nodulation to occur (Dart et al., 1991). Thus, in such soils there is a need to inoculate leguminous crops and trees with the effective rhizobia strain which is specific to a particular legume species. There are also promiscuous types of rhizobium strains which are able to effectively nodulate different species of leguminous trees and crops. If such strains are obtained, it will be more advantageous than specific strain nodulating and inducing N-fixation in only one species. In any case, developing a specific or/and promiscuous rhizobia strain for enhanced N-fixation by legumes requires isolation and screening such bacteria from diverse agro ecologies and soil types in the laboratory, greenhouse and fields.

Once such effective rhizobium strains are obtained they can be commercially produced and supplied to farmers or any user so that they can be used as inoculants of sesbania. In this regard, Wolde-meskel et al. (2004a) isolated 241 different rhizobium strains from roots of 15 woody species most of them being indigenous to Ethiopia and three leguminous crops. They further reported that they were belonging to genera of Agrobacterium, Bradyrhizobium, Mesorthizobium, Rhizobium and Sinorhizobium based on bio-chemical and molecular methodologies. However, information on nodulating ability of these strains and their effects on the growth and N content of legumes is lacking. Such information is important to identify effective rhizobium strains that can be used as inoculants of a particular legume which ultimately helps to produce high quality sesbania green manures that will be used to fertilize the soil. Thus, an experiment was conducted to investigate the ability of selected indigenous strains of rhizobia in inducing nodulation in sesbania and to evaluate the effectiveness of each strain on the growth and N content of sesbania in the greenhouse. 

The list of strains along their geographic origin and host plants from which they were first isolated are indicated in Table 1.

The experiment was conducted in the greenhouse using modified Leonard jars, assembled as described in Vincent (1970). The jars were made of two plastic jugs of different sizes. The upper part of the jar had a diameter of 6 cm wide at the mouth and was filled with sterilized sand. The lower part of the jar was used to fill the N free nutrient solution (Figure 1).The cotton wick in the middle makes the nutrient available to the growing plant.

Parallel to preparing the Leonard jars, the laboratory sterilization and pre-germination of sesbania seeds were conducted. Visually healthy looking seeds of sesbania were selected and the surface sterilized with 96% alcohol followed by surface disinfection with 3% sodium hypochlorite for one minute (Somasegaran and Hoben, 1994). Then, the seeds were rinsed with distilled water five times. The disinfected seeds were aseptically transferred into petri-dishes containing 1% water agar medium and allowed to germinate for 12 days at room temperature.

The germinated seeds of sesbania were transplanted to the Leonard jar using sterilized forceps. Two seedlings were planted per jar and later thinned to one seedling per jar. The seedlings in the Leonard jar were fertilized with quarter strength of Modified Jensen N- free solution (Broughton and Dilworth, 1970) twice a week. In the meantime, each strain indicated in Table 1 was grown on yeast extract manitol broth (YMB) for 3 to 5 days, depending on the genera specific of strain. After the emergence of the first leaf (5 days after transplanting), the seedlings of were inoculated with each YMB cultured strains containing approximately 10⁹ cell ml^-1 as per the treatment. Un-inoculated and un-fertilized (-N) control treatment was also supplied with N fertilizer free medium. The un-inoculated but N fertilized (+N) treatment plants were supplied with 1 g of KNO₃ L^-1 of the nutrient solution (Broughton and Dilworth, 1970). The experiment consisted of a total of 42 treatments and was laid out in completely randomized design (CRD) with three replications. The plants were grown for 9 weeks.

Two weeks after seedlings were inoculated with rhizobium strains, data on plant height (PLHT) and number of leaves (LN) were collected every week until the end of the experiment. The plants were harvested at the end of 9^th week. Tops were severed at sand level dried at 70°C in an oven for 48 h and weighed. The roots were carefully separated from the sand from each jar and immediately brought to the nearby soil laboratory where they were washed over 1 mm sieve by  gently  flowing  tap  water  in  order  to remove adhering sand and other dust particles. Then nodules were counted. Both the nodules and roots were then weighed after oven drying at 70°C as per the treatment, and the replications were taken.

The dried shoots of each seedling were further processed to demine the total tissue N content. The dried tissues of each seedling were pass through a 0.5mm sized sieve. Then the processed samples were analysed for total N content (NC) by using Kjeldhal digestion procedure as described in Rowell (1994).

Determination of symbiotic efficiency of Rhizobia isolates

Symbiotic efficiency (SE) of each isolate was calculated using the following formula:

Symbiotic effectiveness (SE) values were rated as ineffective (< 35%), effective (35-85%) and highly effective (>85%) as described in Beck et al. (1993).

Statistical analysis

Data on the number of nodules per plant (NN), nodule size (NS), nodule dry weigh (NDW), shoot dry weight (SDW), root dry weight (RDW), PLHT, NL and tissue N contents (NC) were subjected ANOVA using SAS software version 8.1 (SAS, 2000). Further, mean separations were done using least significant difference method at 0.05 probability level. Cluster analysis of the rhizobium strains studied in experiment was performed using the same software based on nodulation, growth parameters and the NC of sesbania (functional diversity).

The majority of strains in Table 4 or 5 which were found to be effective inoculants of sesbania were found to belong to the genera of Rhizobium and Sinorhizobia. The test strains in the two genera accounted for 53.3 and 33.3% of all the test strains. This is in line with Wolde-meskel et al., (2004a) who reported that most strains that produced higher plant height and dry matter in sesbania belonged to the genera of Rhizobia and Sinorhizobia. The apparent large number of N₂ fixing bacteria which we found to be effective in nodulating and enhancing the growth and N content of sesbania suggests that the tree is promiscuous both for nodulation and effectiveness. Similarly, Boivin et al. (1997) reported that sesbania species can enter into symbiosis with many other species of rhizobia.

It is concluded that out of the forty rhizobium strains screened for their symbiotic effectiveness in the greenhouse, 37.5% of them produced dry matter yield, plant height and tissue N content in sesbania similar or higher than that produced in +N fertilizer treatment. Of this number only eight (20%) of them produced significantly higher dry matter, plant height and tissue N content in sesbania than that produced in +N control treatment. This shows that there is a high potential to identify effective rhizobium strains from indigenous sources from the diverse agro ecologies of Ethiopia. This, further implies that some of the efficient strains identified in this study can be used as inoculants of sesbania for enhanced production and productivity of the tree which in turn can be exploited as green manure or/and forage. Most of the rhizobium strains identified as effective symbionts of sesbania were originally isolated from trees and low land food leguminous crops other than sesbania. This implies that  there is possibility to use one or more of the rhizobium strains identified as efficient in this study as inoculants of many legumes. 

The authors have not declared conflict of interest.

We would like to thank the legume Rhizobium symbiosis research project (NUFU) funded by the Norwegian government for financing this study. The technical support from N2Africa-Ethiopia is well appreciated.