Phytotelmata are temporary water bodies, even if the habitats themselves are permanently available. The duration of single phytotelms depends on the development of the bromeliads, where new leaves in the centre of the plants form new habitats. Older outer leaves entrap water to a greater extent, but water-holding capacity is lost when the leaves die. Here, the question about the mechanisms of immigration into a newly formed leaf and its phytotelm or even in phytotelmata of different bromeliads raises. Colonization of phytotelmata can be described as "the result of a series of interlinked events - dispersal, immigration, and establishment" (Maguire 1991, p. 461). Dispersal depends on the mode of transportation, being either passive or active. Active dispersal and selection of a suitable habitat is already described for the bromeliad-crab Metopaulias depressus (Diesel 1989). Passive dispersal includes disseminules, for instance resting eggs, blown with the wind. Common in algae, this transport mechanism might also hold for Phyllognathopus viguieri, which was found even in a rain meter near our field station (Janetzky, pers. obs.). However, it should kept in mind that passive dispersal is of high risk to fail. Phoresy is still a passive, but more certain way to reach a nearby habitat, because animals, such as frogs visiting phytotelmata for water uptake, are used by smaller ones for transportation. Examples are ostracods found to be carried by birds or frogs (e.g., Roy 1931, Seidel 1995).
Colonization experiments performed in Jamaica (bromeliads were cleaned, refilled and exposed to various environmental conditions; Janetzky 1997) have shown that the cyclopoid Tropocyclops jamaicensis, a species widespread in bromeliads of the study site is a rapid colonizer of phytotelmata in terrestrial bromeliads. Two weeks after start of colonization experiments, T. jamaicensis was already found in the manipulated bromeliads. In that particular plant the second observation was six weeks later (Reid & Janetzky 1996), which might be caused by the sampling technique. The discontinuity in copepod records in all bromeliads under investigation led to the assumption that different phytotelmata even in the same bromeliad form distinct habitats with their own "history" in the colonization and establishment of aquatic communities. The re-establishment of stable populations of T. jamaicensis required several months, most probably due to the need to build up organic matter as food resource. The lack of sufficient food supply could have led to strong competition and to the failure of some early populations (Reid & Janetzky 1996). This suggestion is supported by the observation that higher individual numbers of harpacticoid copepods were observed only when the aquatic fauna was depauperate (Janetzky 1997) and in investigation of aquatic communities in Puerto Rican bromeliads (Maguire 1970). Maguire (1970) has shown that harpacticoid copepods and ostracods as well as cyclopoid copepods and mosquito larvae exclude each other because of similar feeding habitats (negative associations). In contrast, cyclopoids and harpacticoids or ostracods could be found together (positive associations).