Shertzer and Williams: Reef fishes off the the southern United States 
261 
Figure 2 
Average silhouette width (lines with circles) from /z-medoid 
cluster analysis of species in (A) the headboat sector and 
(B) the commercial sector. Lower and upper lines (without 
circles) represent 5 th and 95 th percentiles, respectively, from 
n = 1000 bootstrap replicates. Average silhouette width mea- 
sures goodness of clustering; higher values indicate better 
concordance with data. 
In general, indices of abundance were not syn- 
chronous (Table 2). Within the deepwater as- 
semblage, snowy grouper was positively but not 
significantly correlated with blueline tilefish, 
and neither species was strongly correlated with 
speckled hind. Within the southern assemblage, 
correlations between species were mostly negative; 
however, those between yellowtail, lane, and gray 
snappers were positive and significant, indicating 
synchrony in this subset. Within the northern as- 
semblage, about half of the correlations between 
species were negative, and only the correlation 
between vermilion snapper and black sea bass 
was positive and significant. These results offer 
little evidence of synchrony in population trends 
within assemblages. 
Similarly, first-differenced indices of abundanc- 
es were out of synchrony (Table 3). Correlations 
between species were both positive and nega- 
tive; only one was significantly negative (between 
snowy grouper and speckled hind), and one was 
significantly positive (between gray triggerfish 
and vermilion snapper). These results from first- 
differenced time series do not support the hypoth- 
esis of synchrony in annual population growth 
rates within assemblages. 
Discussion 
It is unlikely that sufficient resources will ever be 
available to monitor, assess, and manage every fish 
stock individually. Thus, managing assemblages 
by means of indicator species has intuitive appeal. 
It begins a shift from single-species management 
toward ecosystem-based approaches and provides a 
scientific and managerial shortcut by supplanting 
the need to monitor and assess every managed stock. In 
the United States, for example, the Magnuson-Stevens 
Fishery Conservation and Management Reauthorization 
Act of 2006 (MSRA, 2006) requires that annual catch 
limits be established to end and prevent overfishing by 
2011 in all fisheries (by 2010 for fisheries where overfish- 
ing is occurring), yet many stocks cannot be assessed 
and their status is therefore unknown. Conceivably, 
setting catch limits by assemblage rather than stock- 
by-stock could satisfy the statute 1) without substantial 
new resources devoted to both data collection programs 
and stock assessment and 2) within the time frame 
allowed. 
Despite its possible appeal, the use of indicator spe- 
cies to extrapolate trends of other species should be 
viewed with considerable skepticism. From the per- 
spective of niche theory, fishes that coexist are able 
to do so, in part, because they have adapted to use 
different niches in their shared environment (May and 
MacArthur, 1972; Leibold, 1995). Consequently, species 
within assemblages differ in reproductive character- 
istics, foraging behavior, habitat requirements, and 
population-level responses to such factors as competi- 
tion, predation, disease, and environmental variation 
(Landres et al., 1988). Because of these differences, 
population trends of one species (or stock) do not readily 
extrapolate to others in the assemblage (e.g., Niemi et 
ah, 1997; Shaul et al., 2007). For empirical and theo- 
retical reasons, several authors have concluded that the 
use of indicator species should be avoided, unless sup- 
ported by strong evidence from the system in question 
(Landres et al., 1988; Niemi et al., 1997). 
From another perspective, even without strong evi- 
dence of synchrony, indicator species may still be use- 
ful if applied in a restrictive sense. That is, if fishing 
effort occurs at the level of assemblages, regulations 
to reduce effort on one species (the indicator) could 
transmit to others of unknown status. The cost of this 
approach would be the forgone yield of any species that 
could sustain increased rates of exploitation. Ideally, 
the indicator species should be the weakest link of the 
assemblage, although defining weakest link could be 
problematic, along with choosing the correct species 
(Simberloff, 1998). Furthermore, there may be limited 
data for the species that is the weakest link of a ma- 
rine fish assemblage. If achievable, however, such a 
