Abstract:Banana Fusarium wilt disease, caused by the fungus Fusarium oxysporum f. sp. cubense race 4, has been reported to be the most limiting factor in Cavendish-banana production worldwide since 1996. Furthermore, the Cavendish-banana production comprised about 90% of the banana growing areas in south China, and more than 200,000 ha banana orchards in this area have been attacked. Among the managements for controlling Fusarium wilt disease, pre-planted fumigation of soil is one of the effective and stable means for this disease prevention. However, methyl bromide, one widely used soil fumigant, will be banned internationally in 2015 due to the destruction of the ozone layer. In this study, ammonia, to our knowledge, has been first applied as soil fumigant to control this soil borne disease as a chemical alternative to methyl bromide in banana planting field. Traditional plate counting, Real-Time PCR and PCR-DGGE methods were used to determine the effects of ammonia fumigation on soil microflora, disease control efficiency and banana yield in a banana orchard with serious Fusarium wilt disease. Through the estimation by plate counting method, the number of Fusarium in the soil collected after fumigation with 130 L/667 m2 of ammonia decreased 1 and 0.5 log unit, respectively, when compared with pre-fumigation (BF) and the non-fumigation control (CK). The numbers of soil culturable bacteria in the soils from fumigation treatment (AMO) and CK were significantly lower than that from BF treatment, while there was no significant difference could be detected among AMO and CK. CK and BF treatments showed significantly higher numbers of soil culturable fungi than AMO, while no significant difference could be detected among BF and CK treatments. The ratio of bacteria to fungi (B/F) in AMO was significantly higher than that in CK, but no significant difference was observed among pre-fumigating treatment (BF) and ammonia fumigating treatment (AMO). Through estimation by Real-Time PCR, the numbers of total bacteria and fungi in AMO were significantly lower than that in the BF and CK, while no significant difference of total B/F was found. Based on fungi PCR-DGGE fingerprints, the intensities of bands A1, A2 and A3 strengthened and bands of B1, B2 and B3 weakened in the AMO treatment whereas band F related to Fusarium sp. was obviously weakened in AMO treatment compared to BF and CK treatments. Soil bacteria PCR-DGGE fingerprints also revealed that the intensities of bands a1, a2 and a3 strengthened in AMO treatment and bands b1, b2, b3 and b4 weakened in the AMO treatment. However, in soil bacteria or fungi community, the three-replicate cluster of AMO treatment distinguished separately to the other clusters of BF and CK treatments while the clusters of BF and CK grouped together. Based on DGGE results, the soil bacterial and fungal structure for AMO treatment was obviously altered after fumigation compared to that of BF and CK. Field experiment showed that the Fusarium wilt disease incidence in AMO decreased about 20% compared to that in the CK and the banana yield per 667 m2 for AMO increased 1.25 t due to the disease incidence decrease. In conclusion, all the results indicated that using ammonia fumigation could effectively improve soil microflora and control the banana Fusarium wilt disease.