Our data show that individuals more than 5 cm DBH of all three mangrove species suffered significant initial mortality (Figure 1). Mangroves less than 5 cm DBH had less than 10% mortality. Mortality did not, however, increase linearly with increasing DBH. Maximal mortality was recorded in the 20-25 cm DBH class for R. mangle and A. germinans and in the 15-20 cm DBH class for L. racemosa. Mortality decreased for trees more than 30 cm DBH. Comparisons among species reveal that A. germinans had significantly less mortality than either L. racemosa or R. mangle (F_2,59 = 4.27, p < 0.05; Figure 2).

These initial estimates of tree mortality do not reflect the full impact of Hurricane Andrew on the forest, however. During the September 1992 assessment, we found many red mangroves with obvious hurricane damage (e.g., cracked stems and missing bark) that were still bearing green leaves and thus still alive. Numerous individuals of A. germinans and L. racemosa were observed to be resprouting after complete defoliation from the storm. Although these trees appeared to have survived the hurricane, subsequent remeasurements of marked individuals indicated that they eventually succumbed to hurricane damage. This delayed mortality is particularly pronounced for L. racemosa, larger individuals of A. germinans, and the smallest size class for all three species (Figure 1). For certain species and size classes, such delayed or continuing mortality has added 50% to our initial estimates of tree death (Figure 1).

The delayed mortality is continuing. More than two years after Hurricane Donna, trees were still dying from the storm's damage (Craighead and Gilbert 1962). Additionally, this continuing mortality is not restricted to sapling and tree-sized individuals. We encountered large numbers of mangrove propagules throughout the affected forests, often more than ten per square meter. The hurricane dispersed a large proportion of these propagules. Many of these seedlings have subsequently perished, especially in regions of sediment deposition and areas subjected to increased concentrations of pore water sulfide.

Figure 1. Mortality by diameter at breast height class for Avicennia germinans (Ag), Laguncularia racemosa (Lr), and Rhizophora mangle (Rm), measured in September-October 1992 (solid) and in May-July 1993 (cross-hatched). The figure is based on measurements taken on more than 4000 individuals. The diameters are given in centimeters. [larger image]

---

Figure 2. Differential mortality for individuals growing in the canopy (C) and in pre-existing canopy gaps (G) for the three species of mangrove trees. Standard error is indicated. [larger image]