The prevailing aetiological paradigm for schizophrenia and other psychotic syndromes suggests these disorders are a characteristic of the human condition and ubiquitous with respect to the environment. This is based on an erroneous belief that the incidence of schizophrenia does not vary geographically. Since the seminal work of Faris and Dunham (1939), the increased risk of schizophrenia associated with urbanicity has presented one of the most frequently replicated findings within psychiatric epidemiology. Despite the exclusion of several potential confounders, few specific socioenvironmental risk factors, which presumably underpin the 'urbanicity' effect, have yet to be identified. Findings for the affective psychoses are contradictory. Ethnic minority status also increases the risk of psychoses, independent of 'urbanicity', suggesting risk factors associated with migration or post-migratory experiences are also aetiologically relevant. It is unclear whether this risk is independent of socioeconomic status or interacts with generation status or neighbourhood-level risk factors. Using two large, first episode UK datasets I tested whether the incidence of psychoses varied between and within urban areas. I investigated heterogeneity in the incidence of psychoses by age, sex, ethnicity, socioeconomic status and place. I hypothesised that neighbourhood-level socioenvironmental risk factors, including social capital, ethnic density, ethnic fragmentation and deprivation were associated with the incidence of psychoses; perhaps interacting with individual-level risk factors, such as ethnicity. Considerable variation in the non-affective psychoses was demonstrated between and within urban areas. This variation was independent of age, sex and ethnicity. Approximately one quarter of this variation was attributable to neighbourhood-level risk factors, including social capital, ethnic density and ethnic fragmentation, independent of socioeconomic deprivation. No evidence of variation in the affective psychoses by place was observed. Ethnic minority status independently increased the risk of all psychoses, irrespective of socioeconomic status, but this risk was conditional upon generation status. Non-affective psychoses were unlikely to be solely a characteristic of the human condition. It was proposed that socioenvironmental factors at the neighbourhood-level, including social capital, mediated the risk of psychoses for the individual, probably interacting with genetic susceptibility.