Contente and Rossi-Wongtschowski: Fish assemblages in the southeastern Brazilian Bight 
223 
m. Pre-established, fixed stations were determined on 
the grid to map the salinity and temperature patterns 
for the entire study area. To interpolate suitably among 
stations, the coverage level of the cruises was calcu- 
lated as the ratio between the prospected distance, in 
kilometers, and the total area (Soares et al., 2005). The 
acoustic data were collected with a Simrad EK500 5 sci- 
entific split-beam echosounder (Simrad Fisheries, Lyn- 
nwood, WA) operating at 38 and 120 kHz. Such data 
were not used in this study and are available in Soares 
et al. (2005) and Madureira et al. (2005). 
When schools were detected (during the day or 
night), the midwater net was deployed at the mean 
depth of the school. The tension and the length of the 
cables of the midwater net and the boat speed were 
changed frequently whenever the target reacted, and 
therefore the position of the net in the water column 
was correspondingly modified. The net performance 
and the mean depth of the net operation were mea- 
sured with a Simrad FR500 probe (Trawl Eye system; 
Simrad Fisheries) located at the upper part of the net. 
The mean depth of net operation was considered the 
mean net depth between the beginning and the end 
of the haul (i.e., mean depth of the fish aggregation 
at the water column) and is referred to in this article 
as Z a . The hauls were performed against the wind and 
current at, on average, speeds between 1.5 and 2 m/s 
(3 and 4 knots) for a period of time depending on the 
size and reaction of the schools. Boat speed was mea- 
sured in relation to the sea bottom. For more details 
on the determination of grid sizes, sampling methods, 
and procedures of fisheries operations, see Soares et al. 
(2005) and Madureira et al. (2005). 
A net designed especially for fishing small pelagic 
fish was used. It had a squared opening (sides of 15 m) 
and mesh of 400 mm between knots in the wings and 
square. The mesh gradually decreased from 50 mm in 
the tunnel to 20 mm in the bag, which was covered in- 
ternally by a panel of mesh. The net was kept open by 
2 doors (Suberkrub type) of 4 m 2 (3 mxl.35 m), weigh- 
ing approximately 380 kg each. 
The salinity and temperature were measured with 
a conductivity, temperature, and depth (CTD) profiler: 
SBE 19 SeaCAT (Sea-Bird Electronics Inc., Bellevue, 
WA) during the ECOSAR III cruise and Ocean Seven 
316 CTD (IDRONAUT S.r.l, Brugherio, Italy) during 
the ECOSAR IV-VII cruises. Measurements were tak- 
en at each 5-m depth during the ECOSAR III cruise 
and at each 1-m depth during the other cruises. 
During the ECOSAR IV, V, VI, and VII cruises, the 
local depth and geographic positions were taken at the 
beginning and at the end of the hauls. For the ECO- 
SAR III cruise, only the geographic position and the 
local depth at the beginning of the haul were taken. 
For all cruises, the duration of tow (At in seconds) was 
taken. Because of a technical problem, the mean tow 
5 Mention of trade names or commercial companies is for iden- 
tification purposes only and does not imply endorsement by 
National Marine Fisheries Service, NOAA. 
speed ( vt in meters per second) for all tows was avail- 
able only for ECOSAR III cruise. 
The catch of each haul was weighed, and a ran- 
dom subsample was taken and frozen on board when 
large hauls were captured (>500 individuals). In this 
case, the weight of each species was estimated on the 
basis of their proportions in the subsample. Benthic 
and demersal species typically found on the seafloor 
were infrequently caught when the bottom of the net 
reached the seafloor, and they were found in low abun- 
dance. Because this net was not designed specifically to 
sample such fish, abundance estimates for these spe- 
cies may have been biased. Therefore, they were not 
considered in this study. The taxonomic identification 
guides used are listed in Rossi-Wongtschowski et al. 
(2014). Vouchers were stored at the Zoology Museum of 
the University of Sao Paulo. Rossi-Wongtschowski et al. 
(2014) provided a complete list of species, including the 
records from the benthic and demersal species, records 
from hauls not considered in the study discussed in 
this article (see below), and records from purse seine 
operations concurrently operated with a midwater net 
during some trawls. 
Analysis procedures 
Abundance estimates and classification of fish samples 
Fish biomass was the abundance metric, and values of 
fish biomass were standardized to a volume (cubic me- 
ters) for all hauls to incorporate the differences in tow 
distance and duration (Harding et al., 2011). The vol- 
ume of water was calculated as follows: distance of haul 
(D in meters)xarea of the opening of the net (268 m 2 ). 
In the ECOSAR IV, V, VI, and VII cruises, D was calcu- 
lated by using loxodromic lines based on the latitude 
and longitude at the beginning and at the end of the 
hauls (Freire and Vasconcellos, 2011). Because we did 
not have these geographic variables for the ECOSAR III 
cruise, D was estimated on the basis of vt as D-vtxAt. 
We used the salinity and temperature values that 
were estimated through interpolation from the records 
of the nearest neighboring stations at the same depth 
of tow. Interpolations were conducted with nominal 
kriging, a geostatistical procedure that weights values 
collected at surveyed points according to a covariance 
function to predict a value for an unmeasured loca- 
tion, by using GeoR (Ribeiro and Diggle 6 ; Diggle and 
Ribeiro, 2007) in R software, vers. 3.1.0 (R Core Team, 
2014). Because of technical issues during the ECOSAR 
V cruise, salinity and temperature were not regularly 
measured along the grid (Fig. 1). As a result of this 
lack of data, we removed fish samples from the south- 
ern SBB. 
The fish samples from the 78 tows considered in 
this study were not equally distributed over the entire 
length of the SBB (Fig. 1) or over the cruises and water 
6 Ribeiro, P. J., Jr., and P. J. Diggle. 2001. GeoR: a package 
for geostatistical analysis. R News 1(2): 15-18. [Available 
at website.] 
