Hazard ratios of death relative to controls are predicted by the time workers spent in aggregations. Groups of 12 workers were challenged by brief immersion in the conidial suspensions of nine different strains of Metarhizium (for two colonies, n = 18). The strains corresponded with three different species. Open circles, M. brunneum; open squares, M. roberstsii; open triangles, M. guizhouense; solid circles, controls. The line of regression does not include controls. The time in seconds that workers spent in aggregations of greater than eight was log-transformed. After controlling for colony effects, standardized β = −0.580, p = 0.013.