above data were collected to determine if several hypotheses concerning effects of 
stress on organisms, assemblage, community and faunal composition were consistent 
with data on assemblage structure. These hypotheses were that: (1) stress should 
reduce the average size of organisms; (2) shorten food chains; (3) reduce predation 
intensity; (4) reduce species richness and diversity; and (5) increase the relative 
abundance of predator-susceptible ancestral species (i.e., Archaeogastropoda). Water 
quality data suggest that the two most likely forms of stress in deeper (> I m) areas of 
Florida Bay adjacent to the Keys are cold water temperatures associated with winter 
cold fronts and low predawn oxygen associated with warm summer temperatures, high 
salinity, and periodic algal and seagrass drift buildups. Seagrass sites had high 
population densities and low diversity due to the dominance of Astraea americana 
Gmelin (American star shell) in Florida Bay and Modulus modulus L. in Hawk Channel 
seagrass habitats. Florida Bay sites had high species richness on a small spatial scale, 
but Hawk Channel sites had more species and greater encounter rates of new species on 
a larger scale. Predawn oxygen measurements taken during July in four habitats were 
positively correlated with prosobranch species richness and diversity. Faunal data, 
analyzed on a population density basis, fit the above hypotheses of body size, trophic 
level, and evolutionary age of the species. Attempts to measure predation on an 
experimental prosobranch (A. americana) were unsuccessful but a tethering experiment 
with a sea urchin ( Echinometra lucunter L.) indicated higher predation in the less 
stressful Hawk Channel than Florida Bay hard substratum sites. Stress appears to 
reduce the abundance of higher trophic levels (both prosobranch and finfish predators) 
resulting in the dominance of ancestral forms not adapted to predation but tolerant of 
environmental stress. Eutrophication or increased oxygen demands in Florida Bay could 
result in further species richness and diversity declines. 
1992 0 
McClanahan, T. R., and N. A. Muthiga (1992) Comparative sampling methods for subtidal 
epibenthic gastropods. J. Esp. Mar. Biol. EcoL 164:87-101. 
[DATE OF SAMPLING UNKNOWN OR NOT APPLICABLE.] A comparative survey of patchily 
distributed prosobranchs inhabiting seagrass and hard substratum (live, dead or 
Pleistocene coral) in both the Florida Bay and Hawk Channel environments of the Florida 
Keys was undertaken to compare a l-hr search versus a quadrat method (5 m' 2 ) of 
sampling. We tested the hypotheses that (1) there are differences in observer search 
ability and calculated community structure parameters, (2) large-bodied species will 
be over-sampled compared to small-bodied species, (3) abundant species will be under¬ 
sampled due to observer habituation towards abundant species, (4) individuals with 
cryptic or nocturnal habits will be under-sampled during daytime sampling, and (5) 
there are differences in search efficiency among habitats. Two independent observers 
using the search method had less than 20% variation in all community structure 
parameters and 10% variation in community composition similarity (Bray-Curtis 
Index) suggesting that observer-bias is small for experienced observers. In three 
habitats, Hawk Channel seagrass, Florida Bay seagrass and hard substratum, there was 
no evidence of over-sampling large-bodied species or of habituation to abundant 
species. In Hawk Channel hard substratum sites, Strombus gigas L. (Queen Conch) 
appeared to be the single species over-sampled due to its large body size, and some 
evidence suggests that Cerithium literatum Born, which buries itself in the sand, was 
under-sampled by the search method. Nocturnal sampling indicated that two species C. 
stratum Born and Marginalia apicina Menke may have been under-sampled in the Hawk 
Channel seagrass habitat during the daytime while no species appeared to be under¬ 
sampled in the Florida Bay site. These nocturnally active species were patchily 
distributed, produce population estimates with high variation, and, therefore, day-night 
population density comparisons were not statistically different. The search method 
353 
