The cotton zone of North-Cameroon is characterized by a climate of sudanean type (precipitation is 1200 mm, temperature is 28˚C, evapotranspiration is 168 mm/month, insolation annual is 2800 hours, cycle cultural cycle of 175 days) within the meridional part comprising the Touboro and Garoua regions. In the north to which belong Guider and Maroua regions, the Sudano-Sahelian type of climate is predominant (precipitation is 700 - 850 mm, temperature is 27˚C - 28˚C, evapotranspiration is 173 mm/month, insolation annual is 2800 hours, cycle cultural cycle of 130 days) [6] .

The tow soybean varieties used in this study were tested and vulgarized by the institute of Agricultural Research and Development (IRAD), and bear the following characteristics (Table 1).

Soils used to trap rhizobia in a pot experiment were sampled in the four cotton zones of North-Cameroon represented by fields at Touboro, Garoua, Guider and Maroua. A total of 15 soils samples were collected respectively at Yagoua, Mouhour, kodeck, Zibou for Maroua zone, at Soukoundou, Sorawel, Doundehi for Guider zone, at Gashiga, Adoumri, Pitoa, Sanguere-Paul for Garoua zone, and at Touboro, Ouro-Barka, Vogzom, Sackdje for Touboro zone. Soil samples from the same zone were mixed in a composite sample. Rhizobia were trapped from composite soils by sowing soybean seeds in pots and allowing plants to grow. At 45 days after sowing, plants were removed and the efficient nodules selected for rhizobia isolation [7] . Rhizobia were isolated from deshydrated root nodules on YEMA (Yeast Extract Mannitol Agar) medium composed of Yeast Extract Mannitol Agar (YEMA) medium was prepared as previously described with the following composition [8] : 10 g mannitol; 0.2 g MgSO₄∙7H₂O; 1 g yeast extract; 0.5 g KH₂PO₄/K₂HPO₄; 0.1 g NaCl; pH = 6.8. The medium was supplemented with 0.25% Congo red to restrict possible contaminants. Isolatation of rhizobia was carried out after sterilization and grinding of root nodules in a Petri dish using a sterile carpel, aliquot of the macerate was taken and aseptically struck onto the YEMA solid medium. Inoculated petri dishes were incubated at 28˚C ± 2˚C temperature. Three days after the first streak, different colonies grown separately were sub-cultured onto the YEMA-Congo Red solid medium under the same conditions, in order to screen all the single colonies. Three days after sub-culture, the morpho-cultural characteristics of isolated colonies were described before the study of their physiological and biochemical characteristics [9] .

Three days after incubation at 28˚C ± 2˚C on YEMA growth medium in the dark, different typical colonies were identified, counted and their morphological criteria (shape, size, appearance, elevation, surface, color and edge) assessed as reported [11] . Rhizobia strains were screened through Congo red test, Bromothymol blue tests, Gram staining and cell observation under photonic microscope [8] .

Isolated rhizobia were studied for their resistance to different stresses, by streaking on YEMA media at different NaCl concentrations (1%, 2%, 3%, 5% and 10%), pH (4, 7, 8, 9, 11) and temperatures (28˚C, 30˚C, 35˚C, 38˚C and 42˚C) [8] [11] . After growth of isolates on YEMA solid media at 28˚C ± 2˚C for 3 days, the catalase test was carried out by adding 2 - 3 drops of hydrogen peroxide (3%) on cultures. The formation of gas bubbles was an indication of the presence (+) or the absence (−) of catalase produced by isolates [12] . The ability of rhizobia isolates to hydrolyze starch was assessed by adding 0.2% starch powder to nutrient agar medium plates of each isolates. Labelled petri dishes were then incubated at 28˚C ± 2˚C for 3 days, after which the isolates competences to hydrolyze starch through amylase was detected by flooding plates with Gram’s iodine. The presence (+) or absence (−) of halos around the rhizobia colonies was recorded [13] . The triple sugar ion agar test was performed on triple sugar iron agar medium (3 g/L beef extract, 3 g/L yeast extract, 15 g/L peptone, 5 g/L NaCl, 10 g/L lactose, 10 g/L sucrose, 1 g/m dextrose, 0.2 g/L ferrous sulphate, 0.3 g/L thiosulphate, 0.24 g/L phenol red, 15 g/L agar, pH 7.0) to determine the behavior of isolates to use various carbohydrate sources for their growth. The test was carried out to distinguish among rhizobia groups able to ferment glucose, while producing acid and hydrogen sulphide [14] . Ammonia production by isolates was assessed, while adding 2% urea and 0.012% Phenol Red to YEMA medium [15] . Following sterilization by filtration, 10 μL of each isolate was transferred by streaking on the solid medium, and incubated at 28˚C ± 2˚C for 3 days. The shift of the initial medium color from red to pink was an indication of the hydrolysis (+) of urea to form ammonia and carbon dioxide in an alkaline environment, while a yellowish color was a sign of a negative reaction (−) in an acidic environment.

Ammonia production by rhizobia isolates was assessed on 0.5 mL Peptone broth medium (1 g peptone, 0.5 g NaCl, 0.5 g potassium nitrate, and 1000 mL distilled water), dispersed in each labelled test tubes with a control. Following incubation for 3 days, 1 mL of Nessler’s reagent was added to each of the rhizobia culture in the test tubes, then development of color from orange to brown was recorded as an indication of the presence (+) of ammonia [16] . The qualitative analysis of Indole Acetic Acid (IAA) production by rhizobia isolates was carried out in tubes culture in each of which was added 50 μg/mL filtered tryptophan broth solution. Following incubation at 28˚C ± 2˚C for 3 days of tubes containing different isolates including the negative control, 4 mL of the reagent was added to 1 mL of the supernatant, and the solution was thoroughly mixed and incubated for 30 min to allow development of pink color as an indication of IAA production (+) by isolates [17] .

From the six rhizobia strains isolated from pot experiment, seven rhizobia inoculants namely S0, IS1, IS2, IS3, IS4, IS5, IS6 and M were prepared as liquid formulations, 1 mL of each containing averagely 1 × 10⁹ cells [7] .

The experimental fields were established at Soukoundou, and Sanguere-Paul, belonging respectively within the Guider and Garoua cotton field zones. The fields trials were established in a completely randomized block design for each of the soybean Houla1 (V1) and TGX1910 14F (V2) varieties, each comprising 7 treatments represented by the 7 rhizobia formulations IS0, IS1, IS2, IS3, IS4, IS5, IS6 and M. IS0 was the negative control, while M was a mixture of the six rhizobia inoculants at equal volume. Inoculation at sowing consisted of introducing two seeds in a sowing hole, and wet it with 2 mL of the corresponding inoculant. Seed of the control treatment were sown without inoculation.

At 35 days (at flowering) 60 days (at podding) after sowing, 30 randomly selected plants per treatments were carefully removed from soil, and their total root nodules counted. The efficiency of a nodule to fix atmospheric nitrogen was assessed by cutting with a laser blade all root nodules from a plant to observe the color of the leghemoglobin that respectively indicates if the host plant is not fixing (whitish) or is fixing (greenish) atmospheric nitrogen [7] . The dry weigh of root nodules was obtained after measurement on an electronic balance.

The effects of different treatments were analyzed by One-way ANOVA, using the plant weights and the nodule weights as variables. Means between treatments were graded using the Duncan multiple range test, through the STATGRAPHICS 5.0 program.

At the end of the 72 hours incubation at 28˚C on YEMA medium of struck culture, 16 rhizobia colonies (Figure 1) were identified and grouped into 6 isolates based on their morphological characteristics, of which isolates IS1, 1S2, 1S3 and IS4, IS5, IS6 were respectively from Houa1 and TGX 1910 14F rhizospheres (Table 2). Isolate 1 (IS1) was a composite of three (circular, translucid and convex) colonies representing 18.75% of all the colonies. Isolate 2 (IS2) was represented by three (circular, white and convex) colonies (18.75%). Isolate 3 (IS3) was composed of six colonies (37.5%) circular, milklish and convex. Isolate 4 (IS4) was made up of two (12.5%) ovoid, milklish and convex colonies. Isolate 5 (IS5) and 6 (IS6) were represented each by only one colony, respectively ovoid, yellowish, flat or circular, whitish and flat.

Of all the sixteen identified colonies, 81.25% were circular, whereas 87.5% were convex in elevation. Translucid colonies were less represented. These morphological characteristics referred to as slow growing rhizobia colonies, which were reported to be mostly circular and convex, and rarely translucid [18] . The average size of colonies ranged from 2 to 7 mm after 3 days of incubation on YEMA medium. The present size range of colonies was higher than the one previously pointed out, which reported the size of slow growing rhizobia colonies between 1 and 2 mm after 3 - 5 days of incubation [18] [19] .

Figure 1. Rhizobia colonies growing on YEMA medium.

TB: Touboro; GA: Garoua; GD: Guider; MA: Maroua.

The six identified rhizobia isolates were incubated for 3 days at 25˚C, 28˚C, 30˚C and 42˚C each, to test their ability to resist to heat (Table 3).

All the isolates grew well at 28˚C, confirming previous reported results indicating that optimal growth of rhizobia occurs at between 25˚C - 30˚C [20] . The optimal growth of isolates IS1, IS3 and IS5 was still maintained at 25˚C, compared to that of isolates IS2, IS4 and IS6. Above 28˚C, rhizobia isolates maintained a good growth, except isolate IS4 for which a weak growth was observed at 42˚C, in accordance to other results from which some rhizobia strains were able to display growth at between 40˚C - 45˚C [21] . Moreover, slow growing rhizobia were reported to be more thermo-tolerant than fast growing rhizobia [22] . When rhizobia isolates were tested for growth at different pH, no growth was observed at pH 3. Although all isolates did grow well at pH 4, the maximum growth was observed at pH 7 for all the isolates, confirming the neutroplilic behavior of most of the isolated rhizobia [20] . In contrast, the optimum growth of rhizobia found in Calycotome spinose rhizosphere was said to be located between pH 4 - 6.8 [23] . These differences could be ascribed not only to the intrinsic characters of isolates, but also to the environmental condition factors. In this study, all the isolates expressed growth for a wide pH spectrum extending from

(+++): very good growth; (++): Good growth; (+): weak growth; (−): no growth.

4 - 11, thus close to pH range 4 - 10 reported to be tolerated by some rhizobia [24] . Fast growing rhizobia were revealed to be more sensible to acidity than slow growing rhizobia [12] . As far as tolerance to salt concentrations is concerned, all the identified rhizobia isolates showed a good growth from 1% to 5% NaCl, but an alteration of growth at 10% NaCl, in agreement with results of other authors [25] .

Isolated bacteria were characterized on YEMA medium, supplemented either with Bromothymol Blue (BBT) or Congo Red (CR) as shown on Table 4. After BBT reaction, all the isolates shifted the medium from green (BBT color) to blue, indicating that they were all not acid producers, thus were slow growing rhizobia [26] . These results obtained from a different agro-eco-zone are opposite to those of Gomoung et al. [27] , who isolated fast growing rhizobia (acid producers) associated to soybean in the guinea-savannah zone of Cameroon. These results indicate the ability of soybean to establish symbiosis either by slow/or fast-growing rhizobia or both. The microscopic obervations of cells after Gram staining revealed all the isolated rhizobia to be of rod-shape, and pink color, indicating that they were Gram negative bacteria [7] [8] . These results line with those recently reported and indicating Gram negative rhizobia-associated to nodulation of Senegalia senegal, Vachellia seyal, Cajanus cajan L. and Vigna unguiculate. [28] .

RC: Congo Red; BBT: Bromothymol Blue; Gram staining types. The negative (−) signs for CR signify that isolates absorb very weakly RC and are Gram negative.

Isolates IS2 and IS4 used starch, whereas IS3, IS4 and IS5 were the only isolates able to utilize triple sugars for their growth (Table 5). Citrate was used for growth by almost all the six rhizobia isolates, except isolate IS6. All the isolates were unable to produce catalase for detoxification of hydrogen peroxide, but were able to produce ammonia (NH3). Isolates IS1, IS3 and IS5 were able to provide a good production of the growth regulator IAA, against a weak production by isolates IS2, IS4 and IS6.

(+): good production; (±): weak production; (−): no production, NH3: Ammoniac; IAA: Indole Acetic Acid.

Not only all inoculated plants with rhizobia formulations were stimulated to produce root nodules, but also control plants that were not inoculated (IS0), respectively at Sanguere-Paul (Figure 2(a)) and Soukoundou (Figure 2(b)). The formation of root nodules by uninoculated plants was an indication of the presence of native rhizobia strains which could be specific and efficient or/not to the soybean varieties [7] [29] . The greatest efficiency of nodules expressed by uninoculated plants would indicate the high competitiveness of indigenous isolates compared to the introduced ones [30] . Defined as the capacity of a bacteria strains to form root nodule with its host partner, the infectivity was measured by the total number of nodules found on roots. In contrast, the efficiency of isolates was their ability to effectively fix atmospheric nitrogen, and was measured in this case by the greenish color of the leghemoglobin pigment inside the root nodules [31] [32] .

At Sanguere-Paul, soybean plants of the variety Houla 1 inoculated with IS4 and IS5 isolates were revealed to significantly (p = 0.019) bear more efficient root nodules than control plants, meaning that these two strains were more competitive than the native strains, thus able to improve biological nitrogen fixation following inoculation. When inoculated with IS6 isolate, soybean variety TXG 1910 14F formed more efficient root nodules than the native isolates IS0, or IS1, IS2 and the mixture of isolate (M). These findings suggest that the ability of a rhizobia isolate to efficiently fix atmospheric nitrogen may differ for the same host plant from one host variety to another, and from one rhizobia isolate to another even for the same host variety. At Soukoundou, soybean varieties Houla 1 and TGX 1910 14F were very much affected by inoculation with rhizobia isolates.

For soybean variety Houla 1, the native rhizobia were more infective and efficient as the mixture of rhizobia isolates (M) than the other isolates. As far as soybean variety TGX 1910 14F is concerned, the rhizobia isolates mixture significantly increased the number of root nodules per plant (p = 0.0001), as well as

the effective root nodules per plant (p = 0.002), compared to the control and other inoculated plants. This increment would indicate that rhizobia acting as biofertilizers have fixed the atmospheric nitrogen during the symbiotic association with its host plant [29] [33] . It looks like the synergistic action of isolates was the main factor when several isolates were fused together before inoculation of the soybean variety TGX 1910 14F, instead of competition. This mixed infection with several rhizobia isolates was previously revealed with the host Vigna unguiculata or other crop legumes [34] . The mixed infection is in line with the theory of nodule occupancy which determines the different rhizobia isolates that are found within the same root nodule [35] . Surprisingly, rhizobia isolates IS5 and IS6 that were more competitive for the two soybean Houla1 and TGX 1910 14F varieties at Sanguere-Paul became less performant at Soukoundou. The fact that these tow isolates originated from Touboro cotton zone, which is close to the Sudano-guinea savannah zone, justifies the differences in the pedo-climatic conditions at Soukoundou (belonging to the Sudano-sahelian zone) that can affect the behavior and efficiency of isolates. These results revealed that the isolates promiscuity, defined as the symbiotic affinity of a host plant to at least two rhizobia strains and attributed to several cultivated crop legumes [27] , was less pronounced for the Houa1 soybean variety at Soukoundou. Moreover, the specificity of symbiosis was shown to be restricted in soybean, a given strain of rhizobia being able to associate only with very few hosts [36] . In a previous study, root nodule efficiency was shown to be correlated to nitrogen fixation and soybean seed yield at harvest [37] .

Surprisingly, rhizobia isolates IS5 and IS6 that were more competitive for the two soybean Houla1 and TGX 1910 14F varieties at Sanguere-Paul became less performant at Soukoundou. The fact that these tow isolates originated from Touboro cotton zone, which is close to the Sudano-guinea savannah zone, justifies the differences in the pedo-climatic conditions at Soukoundou (belonging to the Sudano-sahelian zone) that can affect the behavior and efficiency of isolates. These results revealed that the isolates promiscuity, defined as the symbiotic affinity of a host plant to at least two rhizobia strains and attributed to several cultivated crop legumes [27] , was less pronounced for the Houa1 soybean variety at Soukoundou. Moreover, the specificity of symbiosis was shown to be restricted in soybean, a given strain of rhizobia being able to associate only with very few hosts [36] . In a previous study, root nodule efficiency was shown to be correlated to nitrogen fixation and soybean seed yield at harvest [37] .

At flowering (Figure 3(a)), inoculated soybean of the variety Haoua 1 significantly formed more efficient root nodules (p = 0.0001), increased the nodule dry weight (p < 0.0001) at Soukoundou than at Sanguere-paul. There was no significant difference on the total number of root nodules of the variety Haoua 1 or the number of efficient nodules of the variety TGX1014F whether plants were cultivated at Soukoundou or Sanguere-Paul. Conversely, the total number of root nodules was significantly very high (p = 0.0002) at Sanguere-Paul, whereas the Soukoundou site mostly and significantly (p < 0.0001) enhanced the dry weight of root nodules of soybean variety TGX1014F. At podding (Figure 3(b)), Sanguere-Paul site was significantly favorable to the root nodule formation efficiency and dry weight of nodules (p < 0.0001) compared to Soukoundou site when both soybean varieties Haoua 1 and TGX1014F were sown. The salt stress of the growing site has been reported to negatively influence the symbiotic nitrogen fixation through inhibition of nodulation by reducing the infection sites [38] . We therefore are pointing out the differences in the soil salt concentrations Soukoundo and Sanguere-Paul sites as the causes of the various results obtained.