Garlic mustard (Alliaria petiolata) is an invasive, spring-germinating, obligately biennial plant. In a central Pennsylvania forest, two distinctive plant patch types exist simultaneously: patches with mixed juveniles and adults and patches that are exclusively juvenile. We experimentally tested two hypotheses to explain this peculiar distribution. The habitat limitation hypothesis proposes that garlic mustard, like other biennials, is limited to rarely available habitats and can only reach maturity in a subset of the patches where its seeds germinate. The intraspecific competition hypothesis proposes that juveniles only survive to maturity in patches lacking adults. We mapped the natural distribution of ten juvenile and ten mature patches and used transects to assess juvenile and adult density in 2 y. We found that patches cycled annually from juvenile to mixed, supporting the intraspecific competition hypothesis. We looked for biotic (interspecific competition) and abiotic (shading) factors that might contribute to habitat limitation, but none affected juvenile performance. We also looked for evidence of intraspecific competition by monitoring juvenile growth in juvenile, mixed and mixed with adults-removed patches. We found that juveniles from juvenile patches have a higher probability of survival than juveniles from mixed patches and adult removal improved juvenile survival in mixed patches. We conclude that the spatial and temporal segregation of juvenile and adult age classes is maintained by intraspecific competition.