Antagonism between two natural enemies improves biological control of a coffee pest: The importance of dominance hierarchies Theresa W. Ong ⇑ , John H. Vandermeer University of Michigan, Department of Ecology and Evolutionary Biology, 2019 Kraus Nat. Sci. Bldg., 830 North University, Ann Arbor, MI 48109-1048, United States highlights  Infection of Coccus viridis by Lecanicillium lecanii is reduced in the presence of Azya orbigera.  Predation has a negative effect on pathogenesis.  Predator was dominant control agent in field survey.  Pathogen was dominant control agent in experiment.  Presence of both enemies improves biological control only when predator is dominant. graphical abstract NEUTRALISM (0/0) (a) CONTROL AGENT 1 CONTROL AGENT 2 PEST Arrow head = positive effect Ball head = negative effect Thickness = strength of effect Legend CONTROL AGENT 1 CONTROL AGENT 2 PEST ANTAGONISM (-/+) (b) CONTROL AGENT 1 CONTROL AGENT 2 PEST (c) article info Article history: Received 4 November 2013 Accepted 3 June 2014 Available online 12 June 2014 Keywords: Diversity Natural enemies Antagonism Dominance Pathogen Predator Coccus viridis abstract The relationship between two natural enemies of Coccus viridis (green coffee scale), an important hem- piteran coffee pest was determined using a combination of experimental and observational approaches. Adult and larval forms of Azya orbigera, a coccinellid beetle predator were included on leaves of coffee plants with healthy scale populations resulting in lower proportions of scales infected with the second natural enemy, an entomopathogenic fungus (Lecanicillium lecanii). C. viridis populations on leaves where A. orbigera were excluded exhibited twice as much fungal infection by L. lecanii. In addition, field surveys of C. viridis populations on whole coffee plants corroborated experimental findings with eight times less fungal infection for coffee plants where A. orbigera was present than for plants where the predator was absent a month prior to surveys of L. lecanii. Despite a reduction in fungal infection in both the experi- ment and survey, the presence of the beetle reduced overall biological control of the pest only in the experiment where the receiver of the antagonism (L. lecanii) was more dominant in controlling C. viridis than the instigator of the antagonism (A. orbigera). In the survey, A. orbigera was dominant over L. lecanii, resulting in equal to greater levels of biological control depending on the degree to which A. orbigera was dominant over L. lecanii. Our results indicate that a negative relationship exists between A. orbigera and L. lecanii, but that contrary to expectations, this antagonism may in some cases improve overall biological control of the shared pest target. Ó 2014 Elsevier Inc. All rights reserved. 1. Introduction In the field of biological control, managers often focus on the successful implementation of a single control agent. Although a single agent may be introduced, the receiving environment is likely to contain a diversity of natural enemies that share the same target pest host. The effects of natural enemy diversity on the success of biological control programs are various, with no clear consensus reached to date. Recent reviews examining the relationship between natural enemy diversity and biological control conclude that although natural enemy diversity often decreased the density http://dx.doi.org/10.1016/j.biocontrol.2014.06.002 1049-9644/Ó 2014 Elsevier Inc. All rights reserved. ⇑ Corresponding author. Fax: +1 734 763 0544. E-mail address: weiyingo@umich.edu (T.W. Ong). Biological Control 76 (2014) 107–113 Contents lists available at ScienceDirect Biological Control journal homepage: www.elsevier.com/locate/ybcon