A mosquito larval–pupal survey was conducted in 1,160 households of the Mexican city of Mérida during the rainy season of 2003 to determine their differential productivity for Aedes aegypti. Larvae and pupae were detected in 15 broad categories of container types. All breeding sites were found in the patios (backyards) and were potentially rain filled. Ae. aegypti pupae were produced from all categories of breeding site, and no single container type was predominately responsible for pupal production. The most productive buckets comprised 42% of the pupae-positive containers and provided 34% of the total pupae collected. Pupal production in buckets, together with plastic rubbish, pet dishes and basins, utensils for cooking and washing, tires, and flowerpots, accounted for almost 87% of pupal production. However, the most important pupal producers had low infestation indices for immature forms, illustrating that the use of positive-container indices can underestimate the importance of certain breeding sites. Overall, 40% of containers that were observed harboring Ae. aegypti pupae were classified as disposable. The remaining containers were considered useful, although some were seldom used. The discussion focuses on the potential utility of the pupal survey for targeting control, and its resulting pupae-per-person entomological indicator, both for comparison with a theoretical threshold for dengue transmission and for targeting vector control in this Mexican city.