SUMMARY
The progress of infection of American foul-brood (AFB), Paenibacillus larvae larvae, in honey bee (Apis mellifera) larvae, prepupae, and pupae were studied in in vitro rearings, micro colonies and full-size colonies. P. l. larvae spores in various known numbers (3–1384 spores per larva) were inoculated at a larval age of 24–28 h. The in vitro study suggested that survival time decreased with increasing spore inoculation dose. No larvae inoculated with three spores died before day 6, but 30% of the larvae inoculated with 1384 spores had died at day 4, 36 h before the time of capping (in a bee colony). On day 4, approximately 480 000 bacterial colonies were cultured per larvae inoculated with 1384 spores at a larval age of 24 h. Viable counts of P. l. larvae per larvae from inoculation and four weeks onwards fitted (R2 = 0.917) a standard model for bacterial growth: lny = b ln (1+ exp (a—rt)), where y is predicted viable count, r is growth rate, t is larval age in hours, and b and a are constants. Bacterial growth rate in the four weeks infection period was estimated to be r = 0.179±0.030 h−1. In the queen-less micro colonies and full-size colonies the first signs of AFB were not visible to human eyes until day 4. By day 3 nurse bees removed 40% and 50 % of the inoculated larvae, respectively, indicating that they are able to detect infected larvae before disease symptoms are visible. An early removal behaviour probably is a very important trait to focus on when breeding for resistance against AFB. The removal behaviour of nurse bees in micro colonies was well correlated with removal in full—size colonies. We conclude that, the micro colonies may serve as a labour and time saving model for full—size colonies when testing the removal behaviour of selected bee lines.