Frenemy at the gate: Invasion by Pheidole megacephala facilitates a competitively subordinate plant ant in Kenya TODD M. PALMER, 1,2,9 CORINNA RIGINOS , 3,4 PATRICK D. MILLIGAN, 1,2 BRANDON R. HAYS, 2,4 ALEJANDRO G. PIETREK, 1,2,5 NELLY J. MAIYO, 6 SAMUEL MUTISYA, 6 BENARD GITUKU, 6 SIMON MUSILA, 7 SCOTT CARPENTER, 8 AND JACOB R. GOHEEN 2,4 1 Department of Biology, University of Florida, Gainesville, Florida 32601 USA 2 Mpala Research Centre, Box 555, Nanyuki, Kenya 3 The Nature Conservancy, Lander, Wyoming 82520 USA 4 Department of Zoology and Physiology, University of Wyoming, Laramie, Wyoming 82071 USA 5 Instituto de Bio y Geociencias del NOA (IBIGEO), Salta, Argentina 6 Conservation Department, Ol Pejeta Conservancy, Private Bag-10400, Nanyuki, Kenya 7 Mammalogy Section, National Museums of Kenya, Nairobi, Kenya 8 School of Forestry and Environmental Studies, Yale University, New Haven, Connecticut 06520 USA Citation: Palmer, T. M., C. Riginos, P. D. Milligan,B. R. Hays, A. G. Pietrek, N. J. Maiyo, S. Mutisya, B. Gituku, S. Musila, S. Carpenter, and J. R. Goheen. 2020. Frenemy at the gate: Invasion by Pheidole mega- cephala facilitates a competitively subordinate plant ant in Kenya. Ecology. 00(00):e03230. 10.1002/ecy. 3230 Abstract. Biological invasions can lead to the reassembly of communities and understand- ing and predicting the impacts of exotic species on community structure and functioning are a key challenge in ecology. We investigated the impact of a predatory species of invasive ant, Pheidole megacephala, on the structure and function of a foundational mutualism between Acacia drepanolobium and its associated acacia-ant community in an East African savanna. Invasion by P. megacephala was associated with the extirpation of three extrafloral nectar-de- pendent Crematogaster acacia ant species and strong increases in the abundance of a competi- tively subordinate and locally rare acacia ant species, Tetraponera penzigi, which does not depend on host plant nectar. Using a combination of long-term monitoring of invasion dynamics, observations and experiments, we demonstrate that P. megacephala directly and indirectly facilitates T. penzigi by reducing the abundance of T. penzigi’s competitors (Cremato- gaster spp.), imposing recruitment limitation on these competitors, and generating a landscape of low-reward host plants that favor colonization and establishment by the strongly dispersing T. penzigi. Seasonal variation in use of host plants by P. megacephala may further increase the persistence of T. penzigi colonies in invaded habitat. The persistence of the T. penzigi–A. drepanolobium symbiosis in invaded areas afforded host plants some protection against her- bivory by elephants (Loxodonta africana), a key browser that reduces tree cover. However, ele- phant damage on T. penzigi-occupied trees was higher in invaded than in uninvaded areas, likely owing to reduced T. penzigi colony size in invaded habitats. Our results reveal the mecha- nisms underlying the disruption of this mutualism and suggest that P. megacephala invasion may drive long-term declines in tree cover, despite the partial persistence of the ant–acacia symbiosis in invaded areas. Key words: Acacia drepanolobium; ant-plant; biodiversity; elephants; facilitation; herbivory; invasive species; mutualism breakdown; Pheidole megacephala; plant defense. INTRODUCTION Biological invasions are a leading driver of biodiver- sity decline worldwide (Murphy and Romanuk 2014, Bellard et al. 2016, Mollot et al. 2017, Blackburn et al. 2019, Kortz and Magurran 2019) and invasive species can strongly alter both native species assemblages (Mack et al. 2000) and the ecological processes that supply ecosystem services (Vila et al. 2011). A key challenge for invasion ecology is thus to understand and predict the effects of species invasions on both the structure and function of communities. While comparisons of assem- blages and ecological processes in invaded vs. non-in- vaded habitats can reveal the outcomes that are correlated with species invasions, such comparisons do not establish causality, and can thereby obscure whether community changes are driven by the invader or by other environmental factors correlated with the occur- rence of the invader (MacDougall and Turkington 2005). Pinpointing the mechanisms by which invasive Manuscript received 1 May 2020; revised 17 July 2020; accepted 14 September 2020. Corresponding Editor: David A. Holway. 9 E-mail: tmp@ufl.edu Article e03230; page 1 Ecology , 0(0), 2020, e03230 © 2020 by the Ecological Society of America