72 
Fishery Bulletin 1 13(1) 
TabSe 1 
Environmental and ecological variables used in the generalized additive model (GAM) of environmental effects on growth of 
juvenile European hake (Merluccius merluccius) in the northwestern Mediterranean Sea. European hake were sampled in 
June 2011 in the Tyrrhenian Sea (Tyr) and Ligurian Sea (Lig). Latitude (Lat.) and longitude (Long.) of the mean point of 
each haul are shown, as well as the number of specimens (otoliths), mean bottom temperature (Bottom temp.), mean depth, 
and recruit density (number of individuals per square kilometer) recorded in each haul, the mean sea-surface temperature 
(SST), mean wind scalar speed, and mean concentration of chlorophyll-a (chl-a) used in the GAM. 
Recruit 
Haul 
code 
Area 
Lat. 
Long. 
Number 
of 
otoliths 
Bottom 
temp. 
(°C) 
Mean 
depth 
(m) 
density 
(individuals 
km -2 ) 
SST 
(°C) 
Mean scalar 
wind speed 
(m s -1 ) 
Chl-a 
(mg m -3 ) 
42 
Tyr 
4208.07 
1121.43 
14 
13.9 
156 
3214.0 
14.00 
6.53 
0.449 
43 
Tyr 
4211.32 
1116.55 
22 
14.0 
124 
3571.0 
14.00 
6.53 
0.443 
59 
Tyr 
4227.59 
1052.53 
24 
13.9 
131 
8319.0 
13.80 
6.31 
0.438 
62 
Tyr 
4237.86 
1051.43 
22 
13.9 
109 
6631.0 
13.87 
6.31 
0.413 
63 
Tyr 
4232.51 
1047.81 
23 
13.8 
164 
7377.0 
13.83 
6.31 
0.424 
91 
Lig 
4255.79 
1015.80 
21 
13.7 
113 
1008.0 
13.87 
5.57 
0.457 
94 
Lig 
4313.1 1 
958.73 
21 
13.6 
251 
3159.0 
13.50 
7.74 
0.448 
98 
Lig 
4332.80 
952.79 
19 
13.5 
241 
6835.0 
13.42 
7.74 
0.441 
100 
Lig 
4312.68 
1003.93 
20 
13.6 
153 
1819.0 
13.63 
7.74 
0.408 
103 
Lig 
4306.27 
1006.32 
28 
13.6 
150 
1378.0 
13.62 
7.74 
0.426 
115 
Lig 
4344.53 
950.26 
18 
13.4 
205 
3197.0 
13.42 
8.54 
0.449 
126 
Lig 
4309.57 
1020.49 
22 
13.8 
101 
1072.0 
13.59 
7.74 
0.434 
131 
Lig 
4359.29 
941.66 
17 
13.4 
140 
2003.0 
13.48 
8.54 
0.414 
supporting SeaStar software (Star-Oddi, Gardabaer, 
Iceland), attached to the trawl net. European hake ju- 
veniles were identified according to the morphological 
features described by Murua (2010). At each station, 
density (number of individuals per square kilometer) of 
juvenile European hake was computed as the ratio be- 
tween the number of recruits and the swept area (unit 
of measurement km 2 ). 
Otolith reading 
All specimens of European hake that were caught 
at the 13 selected stations were measured in TL to 
the nearest 0.5 cm; otoliths (sagittae) were removed 
from a subsample of 318 individuals that ranged in size 
from 4.5 to 18.0 cm TL. The upper size limit of 18.0 
cm TL was used to define European hake recruits (fish 
in their first year of life) (Belcari et al., 2006). The 
length-frequency distribution of the specimens caught 
at the 13 stations was broken down into normal compo- 
nents by using Batthacharya’s method (Bhattacharya, 
1967). 
Left sagittae were ground on wet sandpaper, polished 
on abrasive cloth with alumina slurry, and mounted 
external-side up on glass slides with a 2-component 
epoxy resin; a second grinding and polishing procedure 
was performed to obtain thin frontal sections (Belcari 
et al., 2006). Microstructure analysis (counting of daily 
growth rings) of otolith sections was carried out with 
a compound green-light, polarizing microscope with 
plan apochromatic objectives. The number of daily in- 
crements deposited within the primordium, the central 
area of the otolith (Morales-Nin and Moranta, 2004), 
was recorded to estimate the duration of the presettle- 
ment period of this species (Arneri and Morales-Nin, 
2000; Morales-Nin and Moranta, 2004; Belcari et al., 
2006). 
Otolith readings were used to back-calculate the 
hatching-date distribution by subtracting the fish age 
from the date of capture. An indirect validation of the 
periodicity of increment formation was obtained by 
comparing the estimated hatching-date distribution 
with the spawning period of the species (Arneri and 
Morales-Nin, 2000; Panfili et ah, 2002; Belcari et al., 
2006). To estimate the age-frequency distribution and 
to back-calculate the hatching-date distribution, the 
age-length key that resulted from the readings was 
applied to the length-frequency distribution that was 
obtained from the 13 stations that were sampled. In 
addition, a subsample of 40 otoliths was reread by the 
same operators for evaluation of the match between 
the readings (Belcari et al., 2006). 
Data analysis 
A multivariate GAM with Gaussian distribution was 
used to describe the growth of juvenile European hake, 
specifically the length-age relationship and environ- 
mental covariates. For the initial model that was used 
