DOSE-RESPONSE MODELS IN BIOLOGICAL-CONTROL OF PLANT-PATHOGENS - AN EMPIRICAL VERIFICATION

Two models, based on the hyperbolic saturation relationship and probit function, that relate the densities of the biocontrol agent and patho gen with disease response were developed. The hyperbolic saturation (H S) and probit (PB) models and a modification of the negative exponenti al (MNE) model recently proposed by Johnson were evaluated. In the NE model, parameters c and k supply information on the efficiency of the pathogen and biocontrol agent, respectively, and provide endpoint valu es such as the maximum proportion of pathogen inoculum inactivated by the biocontrol agent (B). The HS model gives information on the 50% ef fective dose (ED(50)) for both the pathogen (K-x) and the biocontrol a gent (K-z), asymptotic disease levels without biological control (Y-ma x), and the maximum proportion of pathogen the biocontrol agent can in activate (I-max). The PB model provides information on the relative vi rulence of the pathogen (tau) and relative efficiency of the biocontro l agent (sigma) and on the ED(50) for the pathogen (lambda) and biocon trol agent (mu). Two patho-systems (an aerial and a root disease) and two types of biocontrol agent (antagonistic bacteria and nonpathogenic isolates of the pathogen) were compared. The data from Mandeel and Ba ker on biological control of Fusarium oxysporum f. sp. cucumerinum on cucumber with nonpathogenic isolates as well as our data on biological control of Stemphylium vesicarium on pear with selected isolates of P seudomonas and Erwinia spp. were used. Data sets of the effect of seve ral densities of the biocontrol agent and pathogen on disease levels w ere fitted to the models by nonlinear regression. Estimated parameters permitted quantitative comparisons among biocontrol-pathogen-host sys tems. The most valuable parameters obtained from the data sets that fi tted adequately to the models were the proportion of pathogen potentia lly inactivated by the biocontrol agent, the relative efficiency and E D(50) of the biocontrol agent, and the ED(50) biocontrol/pathogen rati o. The values of B ranged from 0.79 to 0.98, and the values of I-max r anged from 0.96 to 1.04, indicating that a high proportion of the path ogen inoculum was susceptible to inactivation by the biocontrol agents . The values of sigma ranged from 0.7 to 1.4 and were consistent with the independent action of the biocontrol agents on the pathogen. The m ean ED(50) for the biocontrol agent (HS and PB models) for the fungus- fungus pathogen-biocontrol system was 2 x 10(3) CFU/g of soil, and for the bacteria-fungus pathogen-biocontrol system, it was 6 x 10(6) CFU/ ml. The ED(50) biocontrol/pathogen ratio ranged from 1 to 10 for the f ungus-fungus biocontrol-pathogen system and from 77 to 435 for the ant agonistic bacteria-fungus system, indicating the existence of a lower number of targets for fungus-fungus competition for common sites than with bacteria-fungus antagonism.