COMPETITIVENESS AND PERSISTENCE OF STRAINS OF RHIZOBIA FOR FABA BEAN IN ACID AND ALKALINE SOILS

Two ecological aspects of the rhizobial inoculation of faba beans were examined, namely their persistence in acid soils and their competitiv eness in soils with high naturally-occurring populations of rhizobia. Of 8 Rhizobium leguminosarum by. viciae strains introduced into 3 acid soils (pH range 4.8-5.7), only 3 consistently maintained soil populat ions in excess of 100 g(-1) 2 years after planting. This poor persiste nce was reflected in nodulation scores and plant growth of uninoculate d faba bean sown 2 years after the initial introduction of inocula. On ly plants in plots originally inoculated with WSM1014 had yields that were comparable to that of plots sown with freshly-inoculated (commerc ial strain SU303) seed. In the same area in southwest Victoria, soils of 11 commercial faba bean crops that had been inoculated with SU303, were sampled 1 year later and the total number of rhizobia in the soil s were compared with those in soil from adjacent paddocks with no grai n legume cropping history. Below pH 6, numbers in the sown soils were low, with little or no increase over numbers in unsown soil. Above pH 6, the numbers in the unsown soil varied from ''not detectable'' to 10 (4) g(-1), and the sown soils showed a variable increase. In only 3 in oculated and 2 uninoculated soils (pH 6.08-8.25) did rhizobia equal or exceed 10(3) g(-1). Serological examination of isolates from MPN plan ts indicated poor establishment of the inoculant strain. In conclusion , the persistence of R. leguminosarum by. viciae strains in acid soils was poor, however strains with improved persistence may be selected. Five potential inoculant strains were introduced as inoculants on faba bean seed into 7 neutral-to-alkaline soils in the Northern Wimmera, w hich possessed high populations of competitive native strains. At 12 w k, all plants, including the uninoculated controls, were well-nodulate d. At 16 wk, serological examinations indicated only a low level of in fection by the inoculant strains. At harvest, there were no yield incr eases due to inoculation. At one site, isolates from nodules in uninoc ulated plots were serological identical to one of the inoculant strain s, but the application of molecular biology techniques showed they wer e not the inoculant strain. Another strain showed a loss of infectivit y in the peat inoculant. These results demonstrated the difficulty of demonstrating a response to inoculation in the presence of a substanti al native population.