Contente and Rossi-Wongtschowski: Fish assemblages in the southeastern Brazilian Bight 
233 
2014) have indicated that the reproductive activity and 
larval growth rates of Brazilian sardinella were higher 
at higher temperatures (~21-27°C) of CW in the SBB. 
Although some species, such as the flying gurnard, 
Brazilian sardinella, and piquitinga anchovy, have 
been more abundant in a specific water mass, they co- 
occurred more frequently with species that were more 
abundant in another water mass. For example, al- 
though the Brazilian sardinella was more abundant in 
CW+M, it occurred more frequently with species more 
abundant in SACW. This result reveals that the species 
show a high flexibility with environmental variability 
and that their abundance patterns are characterized 
by gradients (instead of fixed points) that follow the 
environmental gradient defined from one water mass 
through another. 
This study had 3 limitations. First, the number of 
fish samples from different water masses was not bal- 
anced within each survey and among the ECOSAR 
cruises; this difference in sampling may be one of the 
reasons for the lack of significant differences in fish 
structure among the different cruises (i.e., different 
years) and along the latitudes. 
Second, although our large data set ensured a gener- 
al overview of the species composition and abundance 
of pelagic fish aggregations in the SBB, we understand 
that our data set may be biased with regard to esti- 
mates of these attributes for the whole SBB pelagic 
nekton (i.e., the pool of organisms both inside and out- 
side aggregations) because 1) it is unclear whether the 
adopted spatial scale for sampling best represented the 
variance in the biomass of populations because there 
was no previous study to inform it, 2) sampling ran- 
domness was not adopted, and 3) many pelagic large 
fast-swimming predators, such as sharks, the dolphin- 
fish ( Coryphaena hippurus), and the skipjack tuna 
( Katsuwonus pelamis), were poorly represented or not 
captured; despite being common in SBB aggregations 
(Rossi-Wongtschowski et al., 2014), such predators 
were not susceptible to midwater trawls (Brodeur et 
al., 2005). Therefore, it is strongly recommended that a 
future study should be based on information that helps 
to specify the best sampling spatial scale, on multiple 
sampling gear, and on a sampling design balanced in 
relation to cruises, latitudes, and water masses. 
Third, although significant, the effect of the spatial 
variation of water masses explained little (7%) of the 
total variation in fish assemblage structure. The large 
body of undetermined effects in the data may result 
from important but unconsidered factors for recruit- 
ment success, for maintenance of stock size, and ulti- 
mately for spatiotemporal distribution of assemblages, 
such as turbulence, currents (Thomson et al., 1992), 
zooplankton availability, abundance, and productivity 
(Longhurst and Pauly, 2007), and especially the age 
of a population and its reproductive structure (Fablet 
et al., 2012; Carvalho and Castello, 2013; Dias et al., 
2014). 
Satellite-derived data for surface levels of chlo- 
rophyll-a during the periods of this study, used as a 
proxy for phytoplankton abundance, did not show sig- 
nificant effect on fish assemblage structure and there- 
fore contributed to the large body of undetermined 
effects. For future studies, we recommended that phy- 
toplankton be measured concomitantly with the tows, 
at the same depth of tow or as close as possible. Such 
simultaneous measurements should also be valid for 
other food resources, such as zooplankton. This concur- 
rent measurement of phytoplankton would increase the 
possibility of detecting the recurrent and structuring 
role of plankton on pelagic assemblages of fishes (Lon- 
ghurst and Pauly, 2007), of which the bulk of species 
are planktivorous. Other potential causes for this lack 
of significant effect are 1) that the spatial resolution 
of chlorophyll-a measurement may have been low and 
incompatible with the spatial scale of tows or 2) that 
fish populations may have a delayed response to phy- 
toplankton and subsequent zooplankton peaks (Litz et 
al., 2014). 
In conclusion, small pelagic fish populations domi- 
nated the multispecific aggregations over the SBB in 
spring and summer, and the spatiotemporal variability 
of the aggregations fitted well with the spring-sum- 
mer, hydrographically complex structure of the SBB. 
Two species groupings were reliably linked to either 
the CW+M or the SACW. Such conclusions may support 
the adoption of an EBFM for the SBB, which has been 
heavily affected by fishing and activities related to oil 
and gas extraction. 
For conservation purposes, for example, the map- 
ping of the species assemblages in SBB water masses 
through remote sensing may be a useful indicator of 
the spatial variability of pelagic fish assemblages. The 
identification of areas with high density of Argentine 
anchoita will be particularly important for manag- 
ing the potential, near-future exploitation of the SBB 
stocks of this species, stocks that are in almost-virginal 
condition (Madureira et al., 2009). The management of 
this stock will deserve particular attention because the 
Argentine anchoita population of the SBB exhibits a 
lower individual growth rate and a higher natural mor- 
tality rate than does the population that inhabits the 
southern Brazilian shelf (Carvalho and Castello, 2013). 
To better understand and forecast human-induced im- 
pacts in the SBB, an effective EBFM is needed and 
should include future efforts 1) to accurately estimate 
pelagic fish biomass and identify pelagic trophic groups 
to model food-web dynamics and 2) to investigate the 
relationship of fish assemblages and population struc- 
ture and dynamics with water masses, physical forc- 
ings, and distribution of food availability to improve 
the predictability of multispecies multivariate models. 
Acknowledgments 
We are especially grateful to the Brazilian Institute 
of Environment and Renewable Natural Resources for 
funding the cruises for this study, and we thank the 
officers and crew of the ship Atlantico Sul, the staff 
