areas to 0.5 km offshore, or to depths of approximately 2 m where shoals extended well 

 offshore (Pasco and Hernando Counties). 



Sighting locations on the flight record charts were categorized into three habitat 

 types: (1) offshore: the Gulf of Mexico, (2) bay-estuary: bays, estuaries, and large rivers 

 with direct access to the Gulf of Mexico, and (3) marsh-river: complex marsh habitats 

 (see Leatherwood and Platter 1979), inland bays (Monroe County), and narrow rivers. 

 Using criteria from Remane and Schlieper (1971), salinity at each sighting location was 

 subsequently classified as fresh ( < 0.5°/oo salt), brackish (0.5 to 30°/oo salt), or marine 

 ( >30°/oo salt) based on available reports (E.P.A., Wang and Raney 1971, U.S. Dept. of 

 Commerce 1973, Weinstein et al. 1977, Schmidt and Davis 1978). Offshore habitats were 

 always categorized as marine, even though salinities in some areas might have been 

 influenced by tide and fresh water runoff from recent storms. Relative survey effort 

 was estimated as the percentage of total flight time in each habitat and salinity type. 



Patterns of relative abundance and mean herd or group size were evaluated using 

 chi-square and analysis of variance (ANOVA) procedures (Sokal and Rohlf 1969). Multiple 

 comparisons among means were analyzed w ith Duncan's M ultiple Range Test (Steel and 

 Torrie 1960). A square root transformation (/herdsize + 0.5) was applied to the counts to 

 make them suitable for parametric analysis (Steel and Torrie 1960). Computations were 

 performed with programs of the Statistical Analysis System (Helwig and Council 1979) at 

 the University of Florida, Gainesville, Florida. 



