Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA

Background and objectives
Fire blight of pear and apple, caused by Erwinia amylovora, is controlled biologically when epiphytic, antagonistic bacteria (Pseudomonas fluorescens or E. ;herbicola) become established on the stigmatic surfaces of blossoms prior to colonization by E. ;amylovora . In the USA, the antagonist P. ;fluorescens strain A506 is available commercially and an experimental use permit has been granted for the antagonist E. ;herbicola strain C9-1 for suppression of fire blight. Understanding the factors affecting establishment of these bacteria in blossoms and their population dynamics and effectiveness after introduction in orchards have been the goals of our research. Recent studies have evaluated (1) use of honey bees for primary establishment of antagonistic bacteria in blossoms, and (2) the importance of secondary colonization of blossoms by these bacteria in the efficacy of biocontrol.

Results and conclusions
Two bee hives, one to disperse the antagonist, P. ;fluorescens A506, and the other to disperse the pathogen, E. ;amylovora 153, were placed in apple orchards during bloom. Bees were forced to walk through a freeze-dried preparation of either bacterium as they left these hives. Both the antagonist and pathogen were dispersed at 104-106 colony-forming units/bee. Apple blossoms were sampled over time at various distances from the hives. In general, recovery of P. ;fluorescens A506 and E. ;amylovora 153 from sampled blossoms increased over time, but maximum recoveries were variable ranging from 23 to 81% of sampled blossoms. Several concerns with using bees for primary establishment of antagonistic bacteria were identified: rates of blossom colonization were variable among seasons, bee foraging activity depended greatly on weather conditions and other flowering plants competed for inoculated bees. Because biocontrol of fire blight requires that populations of antagonists become established on stigmas prior to establishment of the pathogen, we concluded that variation in antagonist establishment in blossoms, when vectored by honey bees, is too high when compared with establishment obtained when bacteria are applied by sprayers. Nonetheless, based on this study, the use of honey bees to disperse bacteria has proven valuable as a method to inoculate pears flowers with realistic doses of the pathogen for epidemiological studies.

Secondary dispersal and colonization of pear blossoms by P. ;fluorescens A506 and E. ;herbicola C9-IS were monitored in orchard blocks in which the centre rows of trees were sprayed with these antagonists at midbloom. Immediately after spraying, both antagonists were detected only on treated blossoms. As bloom progressed, blossoms colonized by P. ;fluorescens A506 and E. ;herbicola C9-IS were detected on untreated trees up to 10 ;m from treated trees. Each year, honey bees were used to inoculate all trees with E. ;amylovora at 80% bloom. After full bloom, the proportion of blossoms with high populations of the pathogen was greatest in the outermost rows and decreased linearly with proximity to rows treated with antagonists. In one experiment, diseased blossom clusters also decreased significantly (P<0.05) from the outermost rows to the treated rows. Both P. ;fluorescens A506 and E. ;herbicola C9-IS were isolated from about 50% of bees sampled near trees in the experimental block, suggesting that bees are involved in the movement of antagonists from blossom-to-blossom.

We concluded that the distinction between primary establishment and secondary colonization of bacterial antagonists in blossoms is important and analogous to the primary and secondary phases of polycyclic disease epidemics. In polycyclic processes, rates of secondary colonization are highest when colonized and noncolonized sites are present in roughly equal proportions. Spray applications, in contrast to bees, achieve primary establishment of antagonists quickly and on a large proportions of blossoms. The high levels of primary antagonist establishment obtained with sprayers compresses the time until secondary colonization of blossoms, which is facilitated by bees (as well as other mechanisms), can occur at maximum rates.