4. Discussion
We were unable to detect a statistically significant effect of
mangrove trimming on the density, biomass, or species composition
of the understory assemblage of fishes in Rookery Bay, FL.
Specifically, we were unable to reject any of the 5 hypotheses that
we erected that considered a variety of possible responses. Retrospective
power analysis suggests that it was unlikely to detect a
response by fish, even if any had occurred. It is not our intention
to use power analysis to maintain that some response by fish was
present but statistically indistinguishable (Morrisey, 1993), as we
did not find any trends even suggestive of trimming effects. Instead,
our work illustrates large change in fish abundance was required to
statistically demonstrate perturbations to the intertidal mangrove
fishes in Rookery Bay.