ABSTRACT: ŒCasitas¹ are artificial habitats that mimic large crevice-type shelters sought by juvenile Caribbean spiny lobsters Panulirus argus. Despite the widespread use of casitas for lobster fishing in the Caribbean, the ecological performanceof casitas is not yet fully understood. We examined the mode of operation of casitas in a habitat-limited population of juvenile P. argus in the reef lagoon at Puerto Morelos, Mexico. The density of juveniles in the reef lagoon was estimated withtransects as 11 ha^-1. After the transect surveys, a year-long baseline study on the lobster population, consisting of 13 visual censuses, was conducted at 9 fixed sites, each of area 100 x 100 m (1 ha). All lobsterswere juveniles (10.7 to 56.5 mm carapace length, CL), and only 12% measured >35 mm CL. No movement of tagged lobsters was recorded among sites. Overall density in the 9 sites was 5.0 ha^-1, but density varied widely in space and time. Twogroups of sites were evident: a group of 5 sites with virtually no lobsters (NL sites), and a group of 4 sites with lobsters (WL sites). Lobster density was related to the abundance and size of natural shelters in each site. Casitas were deployed in 2 ofthe 4 WL and 3 of the 5 NL sites (treatment sites), and further bimonthly censuses were conducted in all sites for an additional year. Despite temporal variations, lobster density in both WL and NL treatment sites increased significantly compared tocontrol sites and to their respective baseline values. The mean size of lobsters and the biomass of all lobsters became gradually larger in treatment sites compared to control sites and to the baseline study, but both mean size and biomass fluctuated intime. These results support the hypothesis that artificial habitats increase the abundance and biomass of lobsters in habitat-limited environments, and the notion of an attraction-production continuum effect of casitas on lobsters.