Selection of Efficient Indigenous Rhizobia Inoculants for the Production of Selected Tropical Legumes in Limpopo Province (South Africa)

Abstract

Tropical legumes are important food crops for human and animal nutrition as well as the improvement of soil fertility. In Southern Africa, tropical legumes are cultivated mostly by smallholder farmers partly because of their ability to thrive in poor soils and adverse weather conditions. Tropical legumes are useful in these cropping systems because of their ability to fix nitrogen (N) thus minimizing the necessity for chemical nitrogenous fertilizers. Soil rhizobia (as bio-inoculants) can enhance the productivity of these legumes through the improvement of soil fertility. However, both the compatibility and competency of individual rhizobial strains is important for attaining optimum crop productivity. Therefore, it is imperative to identify the best efficient rhizobial strain x legume genotype combinations for successful nodulation and optimum legume yield for the benefit of resource-limited farmers in smallholder farming systems and end-users. The aim of this study was to improve the productivity of tropical legumes. The specific objectives of the study were to (i) determine efficient rhizobial strains that combine with tropical legume species to produce optimum crop productivity and (ii) determine efficient tropical legume species x rhizobial genus combinations that produce optimum crop productivity. The study consisted of two experiments both of which were carried out at the Agricultural Research Council (ARC), Plant Health and Protection (Pretoria) greenhouse and laboratory facilities (25° 61’ 547” S, 28° 36’ 435” E). The conditions in the greenhouse were set at a 14 h day temperature of 28° C and 10 h night temperature of 15 °C. In the first experiment, four tropical legumes (pigeonpea; soybean; tepary bean; bambara groundnut) were used in the study. At planting, each legume species was inoculated separately with each of 15 rhizobial strains and allowed to grow for six weeks after which a range of N fixation variables were evaluated. The experiment was laid out in a split plot design with legume species as the main factor and rhizobial strain as the sub-factor. The quantitative data sets on leaf color score (LCS), nodule dry weight (NDW), shoot dry weight (SDW) and root dry weight (RDW) were subjected to standard analysis of variance (ANOVA) procedures and GGE biplot analysis. In the second experiment, three tropical legumes and four rhizobial strains (each from a distinct genus) which were selected from experiment 1 were used. A split-split plot design with the rhizobial strain as the main factor, legume species as the sub-factor and legume variety as sub-sub factor with two replications was used and the data sets of the N fixation variables were analyzed following the same procedure as described above. The results showed that there was variation in legume species x rhizobial strain compatibility and the pattern of root nodulation varied within the legume species depending on the specific strain used. Tepary bean nodulated poorly with most of the rhizobial strains. There were significant differences among the legume species for all the traits that were measured. However, there were no significant (P < 0.05) differences among the rhizobial strains for NDW, RDW and SDW. The highest NDW (0.05 g) and SDW (0.42 g) were attained by bambara groundnut. The mean LCS for the trial was (19.26) while the highest (23.65) and lowest (14.23) LCS were associated with the rhizobial strains Bradyrhizobium elkanii (R4) and Phyllobacterium leguminum (R11), respectively. Among the Paraburkholderia species, the rhizobial strain Paraburkholderia phenolyruptix (R1) was associated with the highest (22.08) LCS. The results also revealed that the rhizobial strains Rhizobium leucaenae (R8) and Rhizobium sp (R6) induced the highest (20.89) and lowest (15.24) LCS, respectively. However, the rhizobial strain Bradyrhizobium elkanii (R14) which attained a relatively high LCS, was associated with the heaviest NDW among all the strains. In contrast, there were detectable nodules associated with four rhizobial strains. Two distinct rhizobial species, namely Paraburkholderia sp (R2) (designated N362) and Bradyhizobium lupini (R5) were associated with the heaviest shoots across the legume species. In contrast, the control strain achieved the lowest (0.25 g) SDW. In addition, all the three strains from the genus Paraburkholderia and the single strain from the genus Phyllobacterium as well as all the five from the genus Bradyrhizobium (R4; R5; R12; R13 and R14) were associated with significantly (P< 0.05) heavier SDW than the control. The results also revealed highly significant (P < 0.01) positive correlations between the LCS and NDW. However, the LCS was negatively but significantly (P < 0.05) correlated with the RDW. In addition, there were highly significant positive correlations between the SDW and each of NDW and RDW. The variety x legume species interaction was highly significant (P < 0.001) for all the attributes that were measured except for NDW. Inoculation with Bradyrhizobium sp (33a-PP4) showed varietal differences in the pattern of N fixation indicators in bambara groundnut and pigeonpea. Nonetheless, some of the soybean varieties formed no nodules after inoculation with the Paraburkholderia sp and Phyllobacterium leguminum strains. Similarly, soybean responded poorly to inoculation with Rhizobium sp. (34a-PP5) and Bradyrhizobium sp (33a-PP4). In pigeonpea, all the four varieties that were used in the study showed similar LCS values. The SDW was used for determining the ideal rhizobial genus and legume species by applying the GGE biplot method in which the legume species and rhizobial strains were coded as environment scores and genotypes scores, respectively. The GGE biplot analysis indicated that 60.0% of the rhizobial strains were distributed in the top left quadrant but none in the bottom right quadrant. The rhizobial strains ‘R2’ (Rhizobium sp; 34a2-PP5) and ‘R1’ (Paraburkholderia sp; N362), were positioned far away from the origin suggesting that they uniquely influenced the legume species. The biplot analysis also revealed that the legume species (coded as environment scores), particularly ‘E1’ (pigeonpea) and ‘E2’ (soybean), were separated by acute angles between them and grouped in the same top right quadrant. In contrast, the remainder of the pairs of legume species were separated by obtuse angles with each suggesting that they were negatively related to each other in terms of the SDW trait. The legume species ‘E3’ (tepary bean) showed the shortest absolute projection suggesting that it was the most stable in performance across the rhizobial strains. In determining the ‘which- won where’, the biplot analysis explained 95.23% total variation of which PC1 and PC2 accounted for 84.41% and 10.82%, respectively. The results also revealed that among the rhizobial strains (depicted as genotypes) on the vertices of the polygon ‘R2’ (Rhizobium sp; 34a2-PP5) and ‘R4’ (Bradyrhizobium elkanii; 33a-PP4), performed best with pigeonpea (E1) and soybean (E2), respectively (Fig. 10). Bambara groundnut (E4’) showed the longest vector, suggesting that it had a high discriminating ability. The rhizobial strain ‘R2’ (Rhizobium sp; 34a2-PP5) was positioned in the innermost concentric circle, thus representing the ideal and most stable strain for SDW among the strains. The study findings provided new information in the patterns of N fixation among a set of tropical legumes that were inoculated with distinct rhizobial genera. The information will be useful in future for formulating bio-inoculants that may improve legume productivity in Limpopo Province (South Africa) where all the legume species that were used in this study are cultivated. The validation of the symbiotic efficiency of Bradyrhizobium elkanii (33a-PP4) and Phyllobacterium leguminum; 31b-PP4) with more diverse bambara groundnut germplasm will be merited.