Introduction

How mangrove forests develop and establish has been the subject of numerous studies. Classical mangrove investigations reported species-specific zonation patterns in mangrove forest (Davis 1940, Lugo and Snedaker 1974, Chapman 1976). There have been numerous observations for and against mangrove forest zonation patterns in many parts of the world (Smith 1987, Smith 1992, Bunt 1996, Chen and Twilley 1998). Typically mangrove species specific sorting have been related to physical or biological gradients (Rabinowitz 1978a, Clarke and Allaway 1993, Smith 1992, Smith et al 1994, Chen and Twilley 1999, Clarke and Kerrigan 2000). More recently, interest has turned towards gap dynamics and applying concepts from upland terrestrial systems to mangrove forest dynamics in order to understand development within both mixed species and zonal patterned forest (Smith 1992, Feller and McKee 1999, Clarke and Kerrigan 2000, Sherman et al. 2000, Duke 2001, Ellison 2002).

Gaps generally are thought to provide an altered environment, which may facilitate conditions that can shift species-specific survivorship, recruitment, and growth of the flora both among and within a species across life stages (Brokaw 1985, Denslow 1987, Hubbell et al. 1999). Mangrove canopy gaps have been found to alter important physical factors and biogeochemical processes important for regeneration: humidity, evapotranspiration, light levels, and soil properties (salinity, temperatures, and nutrients) (Smith 1987a, Smith 1992). The two previous studies of naturally occurring small-scale mangrove canopy gaps found preferential facilitation of species-specific saplings (R. mangle, Avicennia marina) in the gaps as opposed to the surrounding forest (Sherman 2000, Smith 1987a). However, how gaps affect recruitment and mortality into other life stages were not addressed. Presumably recruitment, survivorship, and growth change for the different forest stages, which may allow a particular species to prevail at a particular life stage within the gap successional process. The objective of this paper was to determine how survival (mortality), growth, and recruitment (both as density and specific rates) varied across three successional stages of mangrove forest development (newly initiated lightning gaps, closing gaps and intact forest) for the four dominant life phases (propagules, seedlings, saplings, and adult) of the three mangrove species (Avicennia germinans, Laguncularia racemosa, Rhizophora mangle) in the Everglades mangrove forest of Florida. In this way I was able to follow change in density of stems but also change in population structure at these different successional stages as gaps progressed to closed canopy conditions.