Programmes for the surveillance of Aedes aegypti (L.) often focus on residential areas, ignoring non-residential sites. Between November 2003 and October 2004, pupal/demographic surveys were therefore conducted in non-residential sites in the Peruvian city of Iquitos. The sampled sites included schools, factories, ports, public markets, petrol stations, commercial zones, airports, government buildings, animal-production areas, and recreational areas. Compared with the residential sites that had been surveyed a few years earlier, the non-residential sites generally had fewer pupae/ha, even though pupae were found in a high percentage of the sites investigated. Nonetheless, although <56 pupae/ha were observed in the industrial, commercial, recreational and school sites, the river boats in the ports and the areas in and around public markets sometimes had pupal abundances (of 122-213 pupae/ha) that were comparable with those previously recorded in the residential sites. When the relative production of Ae. aegypti was calculated by container type and characteristic (lidded/lidless, indoors/outdoors, and water-use patterns), no single container category was found to be a major producer of Ae. aegypti, with the exception of flower vases in cemeteries. In general, almost all (97%) of the pupae collected in the non-residential sites came from unlidded containers, although 91% of those collected in river boats came from lidded storage areas. With the exception of lumber mills, plant nurseries and markets (where only 39%-60% of the pupae were collected outdoors), >70% of pupal production was outdoors. In commercial areas, 41% of the pupae came from manually-filled containers, compared with <12% in residential sites. These results indicate that non-residential sites can be highly productive for Ae. aegypti and that the role of such sites in dengue transmission requires further investigation.