Background: Cockroach allergen, Bla g 1, is an important indoor allergen. Although household exposure has been documented, little is known about the potential for exposure outside the home. Objective: We investigated the settled dust concentration of Bla g 1 in 147 samples collected from classrooms, kitchens, cafeterias, and other sites in four primary schools in the city of Baltimore. Methods: School authorities were questioned about characteristics of schools, teachers, and students, as well as cockroach control and cleaning procedures. Settled dust samples were collected with a hand-held vacuum cleaner from the floors of all classrooms, food-related areas, and other sites of the schools over a 3-week period. A sample collection in each school took 1 to 2 days. Dust samples from each room were pooled and analyzed as a single sample for Bla g 1 by using a two-site monoclonal ELISA. Results: One hundred two (69%) of the 147 samples had detectable Bla g 1 and were within the range reported by other investigators in inner city homes. There was no difference between the median levels of Bla g 1 in three schools: school 1 (5.2 U/gm), school 2 (3.0 U/gm), and school 4 (2.7 U/gm); but school 3 had a significantly lower level (<0.8 U/gm, p < 0.001). The median level from the food-related areas was significantly higher than the median classroom level ( p = 0.048). School 3 had fewer students on subsidized lunch, fewer African-American students, and fewer students per teacher. Bla g 1 levels were compared in the different schools while controlling for potential confounding variables by a stepwise multiple regression analysis with a logit model for ordinal responses. On the basis of this analysis, Bla g 1 levels in schools 1, 2, and 4 differed significantly from levels in school 3 ( p < 0.001 in each case). Food-related areas had significantly higher levels than classrooms ( p = 0.048). Floor level, the presence of a sink, and the presence of carpeting did not have significant effects. Conclusions: We conclude that Bla g 1 is detectable at potentially significant concentrations in some inner city schools. Furthermore, the level of exposure is different between different schools and between sites within individual schools. (J Allergy Clin Immunol 1997;99:486-90.)