The usefulness of a prominent infraorbital skin crease as a marker of atopic dermatitis (AD) was examined in 160 consecutive children aged 3 to 11 years in a population setting of a primary school in London, England. Infraorbital crease was recorded by two trained observers according to a strict protocol, and AD was determined by an independent dermatologist who was blinded to the study design. A prominent infraorbital crease was present in only four (27%) of 15 children with AD, compared with 49 (34%) of 145 children who did not have AD (P = .80). A prominent crease was a common finding in black children, even in the absence of AD, affecting 49% (34/69) of normal black children and 25% (11/44) of white children (P = .02). Interobserver agreement for the presence of infraorbital crease was low, with a kappa value of 0.38 (95% confidence interval, 0.23 to 0.53). While infraorbital crease may be of some use in diagnosing individual cases of AD in a hospital setting, it may be less useful in population-based studies because of its poor validity and repeatability. Studies that still use this sign as an indication of allergy need to take ethnic group differences into account.