1. Context: A mechanistic understanding of the drivers of competition between species at a community level can improve invasive species management by helping identify where and when impacts are likely to be greatest. Invasive cavity-breeding birds provide a way to test shared traits and resource requirements are related to intensity of competition. Australia is home to one of the richest cavity-nesting communities globally with over 100 native and introduced bird species requiring hollows: but the impact of invasive species on most native communities is not well understood.  
2. Methods: We examined aggressive interactions between birds in large hollow-bearing trees. To explore the drivers of aggression, we explored whether more similarly sized birds interacted more frequently, whether larger species won aggressive interactions more often, and whether cavity-breeding species with similar preferences for nesting sites (breeding niche space) interacted more frequently. 
3. Results: We recorded a total of 48 bird species of which 20 are cavity-nesters, and 410 aggressive interactions. These interaction networks are dominated by the invasive common myna, the native noisy miner (a non-cavity-breeder), and the native rainbow lorikeet (Trichoglossus moluccanus), but the common myna won the largest total number of interspecific interactions. While on average larger birds won aggressive interactions more frequently, there were some important exceptions to this finding; the common myna (113 ± 30 g) won more interactions (26 interactions won) against the larger native rainbow lorikeet (126 ± 44 g; 3 interactions won). Among the cavity-breeding birds, species with more similar nest-site preferences were observed interacting more frequently. 
4. Synthesis and applications: The impact of the invasive common myna on the cavity-nesting community is greatest for species that are close to the common myna in body size and share nest site preferences in tree hollows.  Myna control efforts should focus on birds that nest in natural tree hollows. Additionally, cavity-nesting species are not immune from the impact of native noisy-miner aggression. 
 

Interaction surveys were conducted during 10, one hour visits to each of eight survey sites. Sites were chosen based on the present of large-hollow bearing trees. 

During a visit to a site, one observer recorded all species that visited the focal trees and all interspecies interactions that occurred within the focal trees for one hour. Observers were 10-30 m from the study trees and used 8 x magnification binoculars. At all sites, we observed evidence of nesting including carrying nesting materials into cavities, cavity excavation, carrying food into cavities and the presence of fledglings coming and going from cavities. We recorded all interactions between two or more species, with interactions defined as one species flying at or within 50 cm of another species. For each interaction, we recorded the species and the number of individuals involved, which species initiated the interaction and which species was the recipient. For each individual in each interaction, the behaviour was recorded (swoop, contact, chase, fight or physical contact, threat, displace, avoid), and an outcome was determined for each species (win, lose). A loss was determined if a species flew from its location. If the recipient species showed no response (no alarm call, no movement toward or away from the initiating species) it was considered a “win” for that species and a “loss” for the species that initiated the interaction.

The data is organized as interacting species pairs in order to caclulate the win-lose network graph, with the winner and loser of each interaction are noted. The data was organized following recommendations for the open source program "Gephi", the documentation of which can be found in the following link (https://gephi.org/publications/gephi-bastian-feb09.pdf).