Plastic has been widely integrated into the day-to-day lives of people throughout the world. However, its use has been growing in an unsustainable fashion, leading to mass production at an alarming rate. Most plastics do not easily break down; when they do, they create smaller plastics, known as secondary microplastics. Microplastic contamination in lakes has been recorded worldwide, but the impacts of microplastic exposure on freshwater organisms require further study. Many studies are laboratory-based, include only one or two trophic levels, or use unrealistically high concentrations of microplastics. Our study aims to help address these knowledge gaps and limitations by conducting an experiment in a more realistic system that involved multiple trophic levels in outdoor 1,200L mesocosms. Each plastic-treated mesocosm included three different types of microplastics: two sizes of spherical fluorescent orange polyethylene primary microplastics (45-53 μm and 355-425 μm in diameter), and 300-600 μm secondary polyethylene microplastics. These plastics were used at a field-realistic combined concentration of 1.5 particles/L. Mesocosms were established to represent a shallow coastal lake food web, with the introduced red-finned perch (Perca fluviatilus) as the top predator. In a 2x2 design, the presence of perch was crossed with the addition of microplastics, with seven replicate mesocosms per treatment combination. In the presence of plastic, there were changes in perch diet and invertebrate community composition. 28.5% of perch in plastic treatments ingested the added plastics. Prey electivity indices showed that plastics were selected against relative to a wide range of invertebrate prey. However, compared to past experiments, fewer plastics overall were ingested, and changes in fish behaviour and foraging seemed minimal in plastic treatments. This is likely because of the lower concentration of microplastics in our mesocosms, which is expected to be more realistic for New Zealand lakes.