In interventions aimed at the control of the immature stages of Aedes aegypti (L.), the principal vector of the dengue viruses, attempts are often made to treat or manage all larval habitats in households. When there are resource-constraints, however, a concentration of effort on the types of container that produce the most pupae may be required. Identification of these 'key' container types requires surveys of the immature stages and particularly - since these give the best estimates of the numbers of adults produced - of the numbers of pupae in local containers. Although there has been no clearly defined or standardized protocol for the sampling of Ae. aegypti pupae for many years, a methodology for 'pupal/demographic' surveys, which may allow the risk of dengue outbreaks in a given setting to be estimated, has been recently described. The consistency and practicality of using such surveys has now been investigated in three cities in the Mexican state of Chiapas, Mexico. Using a combination of 'quadrat'- and transect-sampling methods, 600 houses in each city were each sampled twice. Containers within each study household were searched for pupae and larvae. Although 107,297 containers, belonging to 26 categories, were observed, only 16,032 were found to contain water and 96% and 92% of these 'wet' containers contained no pupae and no third- or fourth-instar larvae, respectively. Although the random 'quadrat' sampling gave similar results to sampling along transects, there were statistically significant differences in the numbers of pupae according to container type and locality. The most important containers for pupal production were found to be large cement wash basins, which were present in almost every household investigated and from which 84% (10,257/12,271) of all pupae were collected. A focus on this class of container could serve as the basis of a targeted intervention strategy. When traditional Stegomyia indices were calculated they appeared to be correlated with the assessments of pupal abundance. The methodology for pupal/demographic surveys appears to be practical and to give consistent results, although it remains to be seen if monitoring of pupal productivity can adequately reflect the impact of vector-control interventions.