fixation has been found in association 

 with mangrove leaves, both living and 

 dead, mangrove sediment surfaces, the 

 litter layer in mangrove swamps, and man- 

 grove root systems (Gotto and Taylor 1976; 

 Zuberer and Silver 1978; Gotto et al. 

 1981). In virtually all cases, nitrogen 

 fixation appears to be limited by the 

 availability of labile carbon compounds. 

 Perhaps for this reason, the highest rates 

 of mangrove nitrogen fixation have been 

 measured in association with decaying 

 mangrove leaves; presumably, the decaying 

 leaves act as a carbon source and thus 

 accelerate nitrogen fixation. Macko 

 (1981), using stable nitrogen ratio 

 techniques, has indicated that as much as 

 25% of the nitrogen associated with black 

 mangrove peat in Texas is derived from 

 nitrogen fixation. 



Zuberer and Silver (1978) speculated 

 that the nitrogen fixation rates observed 

 in Florida mangrove swamps may be suf- 

 ficient to supply a significant portion of 

 the mangrove's growth requirements. Al- 

 though this hypothesis is impossible to 

 test with present information, it might 

 explain why moderately productive mangrove 

 stands occur in waters which are severely 

 nitrogen depleted. 



In summary, knowledge of nutrient 

 cycling in mangrove swamps is highly 

 speculative. These ecosystems appear to 

 act as a sink for many elements, including 

 nitrogen and phosphorus, as long as a 

 modest input occurs. Nitrogen fixation 

 within the swamp may provide much of the 

 nitrogen needed for mangrove growth. 



3.4 LITTER FALL AND DECOMPOSITION 



Unless otherwise stated, litter fall 

 refers to leaves, wood (twigs), leaf 

 scales, propagules, bracts, flowers, and 

 insect frass (excrement) which fall from 

 the tree. Mangrove leaves are shed con- 

 tinuously throughout the year although a 

 minor peak occurs during the early part of 

 the summer wet season in Florida (Heald 

 1969; Pool et al. 1975). Sporadic litter 

 fall peaks may follow periods of stress 

 from cold air temperatures, high soil 



salinities, and pollution events. Litter 

 fall typically can be partitioned as 68% 

 to 86% leaves, 3% to 15% twigs and 8% to 

 21% miscellaneous; the latter includes 

 flowers and propagules. 



Litter fall is an important ecosystem 

 process because it forms the energy basis 

 for detritus-based foodwebs in mangrove 

 swamps (see sections 3.5 and 3.6). The 

 first measurements of litter fall in man- 

 grove swamps were made by E.J. Heald and 

 W.E. Odum, working in the North River 

 estuary in south Florida in 1966-69. 

 This was subsequently published as Heald 

 (1969), Odum (1970), and Odum and Heald 

 (1975a). They estimated that litter pro- 

 duction from riverine red mangrove forests 

 averaged 2.4 dry g of organic 

 matter/m /day (or 876 g/m /year or 8.8 

 metric tons/ha/year). 



Subsequent studies agreed with this 

 early estimate (Table 3), although varia- 

 tion clearly exists between different 

 types of communities. Scrub forests with 

 scattered, very small trees have the 

 smallest amount of leaf fall. Basin and 

 hammock forests, which appear to be 

 nutrient limited, have intermediate leaf 

 fall values. Not surprisingly, the 

 highest values occur in the highly produc- 

 tive fringing, overwash, and riverine 

 forests. Odum and Heald (1975a) suggested 

 that the relatively uniform litter fall 

 values from productive mangrove forests 

 around the world result from the shade 

 intolerance of the canopy leaves and the 

 tendency for the canopy size to remain the 

 same in spite of increasing height. If 

 detailed information is lacking, red man- 

 grove forests of south Florida, which are 

 not severely limited by lack of nutrients, 

 can be assumed to produce litter fall of 

 2.0 to 3.0 g/m /day of dry organic matter. 

 Pure stands of black mangroves usually 

 have a lower rate of 1.0 to 1.5 g/nr/day 

 (Lugo et al. 1980). 



Decomposition of fallen Florida man- 

 grove leaves has been investigated by a 

 number of researchers including Heald 

 (1969), Odum (1970), Odum and Heald 

 (1975a), Pool et al. (1975), Lugo and 

 Snedaker (1975), Twilley (1980) and Lugo et 



32 



